Encyclopedia of Linguistics 2 Volume Set

  • 37 1,812 9
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up

Encyclopedia of Linguistics 2 Volume Set

e n c y c l o p e d i a o f LINGUISTICS V o l u m e 1 Philipp Strazny Editor e n c y c l o p e d i a o f LINGUI

5,704 365 9MB

Pages 1274 Page size 612 x 792 pts (letter) Year 2007

Report DMCA / Copyright

DOWNLOAD FILE

Recommend Papers

File loading please wait...
Citation preview

e n c y c l o p e d i a

o f

LINGUISTICS

V o l u m e

1

Philipp Strazny Editor

e n c y c l o p e d i a

o f

LINGUISTICS V o l u m e

1

A-L

FITZROY DEARBORN An Imprint of the Taylor & Francis Group New York • Oxon

This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”

Published in 2005 by Fitzroy Dearborn An Imprint of the Taylor & Francis Group 270 Madison Avenue New York, NY 10016 Published in Great Britain by Fitzroy Dearborn An Imprint of the Taylor & Francis Group 2 Park Square Milton Park, Abingdon Oxon, OX14 4RN UK Copyright © 2005 by Taylor & Francis Books, Inc. All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage and retrieval system, without permission in writing from the publishers. Library of Congress Cataloging-in-Publication Data Encyclopedia of linguistics / Philipp Strazny, editor. p. cm. Includes bibliographical references and index. ISBN 1-57958-391-1 (set : alk. paper) -- ISBN 1-57958-450-0 (v. 1 : alk. paper) -- ISBN 1-57958-451-9 (v. 2 : alk. paper) 1. Linguistics--Encyclopedias. I. Strazny, Philipp. P29.E483 2005 410'.3--dc22 ISBN 0-203-31920-6 Master e-book ISBN

2004014173

Board of Advisers Robert Beard Bucknell University Bill Foley University of Sydney John Goldsmith University of Chicago David Ingram Arizona State University Francis Katamba Lancaster University Michael Kenstowicz Massachusetts Institute of Technology Darlene LaCharité Université Laval Sydney Lamb Rice University Vincenzo Lombardo Universita di Torino Brian MacWhinney Carnegie Mellon University Horst Müller Universität Bielefeld Emanuel Schegloff University of California, Los Angeles Sarah Thomason University of Michigan Elly van Gelderen Arizona State University

Contents Alphabetical List of Entries ix Thematic List of Entries xv Contributors xxi Introduction xxix Entries A to Z 1 Index 1211

Alphabetical List of Entries Volume 1 A Acoustic Phonetics Acquisition Acquisition Theories Aerodynamics of the Vocal Tract Affixation African American Vernacular English Afroasiatic Age and Language Agreement Ainu Akan and Nyo Languages Albanian Algeria Altaic Ambiguity American Sign Language Amharic and Ethiopian Semitic Languages Analogical Change Anaphora Anatomy of the Articulatory System Anatomy of the Auditory System Ancient Egyptian Ancient Greek Animals and Human Language 1: Overview Animals and Human Language 2: Dolphins Animals and Human Language 3: Parrots Animals and Human Language 4: Primates Aphasia Applied Linguistics: Overview Arabic Arabic Traditional Grammar

Aramaic Arawak Archaeology and Language Armenian Artificial Intelligence Artificial Languages Aspect Assimilation and Coarticulation Assimilation and Dissimilation Austin, John Langshaw Australia Austria Austroasiatic Austronesian Auxiliaries B Babylonian Traditional Grammar Balkans Baltic Languages Bambara, Mandenkan and the Mande languages Bar-Hillel, Yehoshua Basque Baudouin de Courtenay, Jan Ignacy NiecisBaw Belgium Benveniste, Emile Bever, Thomas Bilingual Acquisition Bilingual Mixed Languages Bilingualism Biosemiotics Bloomfield, Leonard Boas, Franz Body Language, see Paralanguage

ix

Alphabetical List of Entries Bopp, Franz Brain Organization and Auditory Pathway Bresnan, Joan British Sign Language Brugmann, Karl Bühler, Karl Burmese Burushaski C Canada Cape Verdean Creole Carib and Cariban Languages Caribbean Case Causation Celtic Languages Chafe, Wallace Chao Yucn Ren China Chinese (Mandarin) Chinese and Japanese Traditional Grammar Chinese Pidgin English Chinese Pidgin Russian Chinook Jargon Chomsky, Noam Clark, Eve V. Clause Clause-Type Indicators Clinical Linguistics Code-Switching Coherence in Discourse Color Terms Communication Theory Comparative Method Compositional Semantics Compounding Computational Linguistics Computer-Assisted Language Learning (CALL) Comrie, Bernard Configurationality Connectionism Constituency Test Context Conversation Analysis Coordination Coptic Egyptian Corpus Linguistics Courtroom Discourse Crioulo, Gulf of Guinea Crioulo, Upper Guinea

x

D Dakota and Siouan Languages Deep Structure Definiteness Deixis Determiner Developmental Stages Dialectology Diglossia Discourse Analysis Discourse Strategies Dravidian Dutch Dyslexia E Emeneau, Murray Barnson Emotion and Language Empty Categories Empty Morphemes Endangered Languages English Epenthesis and Syncope Eskimo-Aleut Ethnicity and Language Ethnography of Communication Etymology Euphemism European Traditional Grammar Evolution of Language 1: Overview Evolution of Language 2: Cognitive Preadaptations Evolution of Language 3: Physical Preadaptations Evolution of Language 4: Social Preadaptations Ewe and Gbe Languages F Fanakalo Farsi Feature Theory Ferguson, Charles Albert Field Methods Figurative Speech Fillmore, Charles John Finite-State Morphology Finnish and Finnic Languages Firth, John Rupert Fishman, Joshua A. Focality Forensic Linguistics France French Language

Alphabetical List of Entries Fromkin, Victoria Alexandra Function Words Functional Approaches G Gender and Language Gender: Class Marking Generation Generative Grammar Genericity Genetic Relationship Genre Georgian and Caucasian Languages German Germany Givón, Talmy Gothic Grammar, Traditional Grammar, Theories Grammatical Function Grammaticalization Great Britain Great Vowel Shift Greek, Modern Greenberg, Joseph Harold Grice, H. Paul Grimm, Jacob Guaymí and Chibchan Languages Gullah Gumperz, John Joseph H Haas, Mary Rosamond Haitian Creole Hale, Kenneth Halle, Morris Halliday, Michael Alexander Kirkwood Handwriting Harris, Zellig Sabbetai Haugen, Einar Hausa and Chadic Languages Hawaiian Creole English Hebrew: Biblical Hebrew: Modern Hindi-Urdu Hiri Motu Historical Linguistics History of Linguistics: Overview Hittite Hjelmslev, Louis Hmong and Hmong-Mien (Miao-Yao) Languages

Hockett, Charles F. Humboldt, Wilhelm von Hungarian and Ugric Languages Hymes, Dell Hathaway I Iconicity Identity and Language Ideology and Language Idiomaticity Idioms Igbo and Igboid Languages India Indian Ocean Creoles Indian Traditional Grammar Indo-European 1: Overview Indo-European 2: Germanic Languages Indo-European 3: Indo-Iranian Languages Indo-European 4: Romance Indo-European 5: Slavic Indo-Pakistani Sign Language Inflection and Derivation Information Retrieval Interpreting Irony Israel Italian Italy J Jakobson, Roman Japanese Japanese Sign Language Javanese Jespersen, Otto Jones, Sir William Juba and Nubi Arabic K Kayardild and the Tangkic Languages Khmer and Mon-Khmer Languages Khoisan Kinship Terms Kituba Korean Krio (and West African Pidgin English) Kriol (Roper River Creole) L Labov, William Ladefoged, Peter

xi

Alphabetical List of Entries Lakoff, George Langacker, Ronald Language Change Language: Contact—Overview Language Death Language Planning Language Socialization Lateralization and Handedness Latin Lehmann, Winfred Philipp Leskien, August Lesser Antillean French Creole Lévi-Strauss, Claude Lexical Borrowing Lexicalization Lexicon: Overview Lingua Franca Literacy Localization of Linguistic Information Long-Distance Dependency Long-Range Comparison Louisiana Creole Lyons, John Volume 2

Modification Mohawk and Iroquoian Languages Mongolian Montague, Richard Mood Moore and the Gur Languages Morpheme Morphological Typology Morphology Murrinh-Patha and Daly Languages N Nahuatl and Uto-Aztecan Languages Naming Natural Classes Navajo and Athabaskan-Eyak Languages Neurolinguistics New Guinea Nida, Eugene Albert Niger-Congo Nigeria Nilo-Saharan Languages Nootka and Wakashan Languages Northwest Caucasian Languages Noun Incorporation Number Marking

M Machine Translation Malagasy Malay-Indonesian and Malayic Languages Malkiel, Yakov Manner of Articulation Ma¯ ori and Polynesian Languages Marshallese and Micronesian Languages Martinet, André Mass Media and Language Mathematical Linguistics Maya and Mayan Languages McCawley, James David Meaning Medicine and Language Meillet, Antoine Metaphor Metathesis Metonymy Mexico Middle (Classical) Japanese Middle English Migration and Language Miskito and Misumalpan Languages Modern Linguistics

xii

O Official Language Selection Ojibwe and Algonquian Languages Okanagan and Salishan Languages Old Chinese Old Church Slavonic Old English Old French Old High German Old Irish Old Japanese Old Norse Old Tibetan Onomatopoeia Oto-Manguean Languages P Pacific Pama-Nyungan Pa¯ nini " Papiamento Paralanguage Parsing Paul, Hermann

Alphabetical List of Entries Peirce, Charles Sanders Personality and Language Philippine Spanish Creoles Philology Philosophy of Language Phoneme Phonetic Transcription Phonetics Phonology Phrase Structure Pidgins and Creoles Pike, Kenneth Lee Pintupi and Pama-Nyungan Languages Pitcairnese Place of Articulation Plurality Polish and West Slavic Languages Politeness Possessives Pragmatics Predication Professions for Linguists Proficiency Testing Pro-forms Proper Nouns Prosody Psycholinguistics Q Quantification Questions Quine, Willard van Orman R Reading Reading Impairment Reference Register Reichenbach, Hans Relevance in Discourse Rhetoric and Linguistics Romanian Rules vs. Constraints Russell, Bertrand Arthur William Russian and East Slavic Languages S Sandhi Sanskrit Sapir, Edward Sapir-Whorf Hypothesis

Saramaccan Saussure, Ferdinand de Scandinavia Searle, John R. Second Language: Acquisition Second Language: Learning Second Language: Teaching Semantic and Discourse Typology Semantics Semiotics Semitic Languages Serbo-Croatian and South Slavic Languages Serial Verb Constructions Sign Relationships Signed Languages Sindhi Sino-Tibetan Languages Slobin, Dan Isaac Sociolect Sociolinguistics Soundwave Analysis South Africa South America: Argentina South America: Brazil Southeast Asia Soviet Union Soviet Union: Successor States Spain Spanish and Iberoromance Languages Spectral Analysis Speech Acts Speech Perception Speech Processing Speech Processing (Neurobiology) Speech Production Speech Production: Neurobiology Speech Synthesis Sprachbund Sranan Sri Lanka Portuguese Standard Language Story Grammar Structural Interference Structuralism Stylistics Sumerian Swadesh, Morris Swedish and Scandinavian Languages Sweet, Henry Switch-Reference Switzerland

xiii

Alphabetical List of Entries Syllable: Structure Syntactic Categories Syntactic Typology Syntax

Universal Grammar Utterance-Centered Linguistics

T

Variation Vietnamese Visual Word Recognition Voice Vowel Harmony

Taboo Tagalog and Philippine Languages Tamil Tarascan Teaching: Curricula Teaching: Methods Telugu Tense and Aspect Marking Tense: Syntax Tewa and Kiowa-Tanoan Languages Text Linguistics Text Understanding Thai and Tai Languages Thematic Structure Tibetan Time and Tense Tiv and Tivoid Languages Tocharian Tok Pisin Tone Languages Translation Trubetzkoy, Nikolai Sergeyevich Tuareg and Berber Languages Tungusic Turing, Alan Turkey Typology U Udmurt United States

xiv

V

W Wackernagel, Jacob Warao Wayampi and Tupi-Guarani Languages Weinreich, Uriel Western Caribbean Creole(s) Whorf, Benjamin Lee Wittgenstein, Ludwig Wolof and Atlantic Languages Word Word Order Word Sense Disambiguation Working Memory Writing Systems Y Yémba and the Grassfields Bantu Languages Yiddish Yoruba and Yoruboid Languages Z Zipf, George Kingsley Zulu and Southern Bantu Languages Zuni

Thematic List of Entries ACQUISITION Acquisition Acquisition Theories Bilingual Acquisition Developmental Stages Language Socialization Second Language: Learning AFROASIATIC LANGUAGES Amharic and Ethiopian Semitic Languages Arabic Aramaic Biblical Hebrew Coptic Egyptian Hausa and Chadic Languages Modern Hebrew Tuareg and Berber Languages ALTAIC LANGUAGES Mongolian Tungusic AMERICAN INDIAN LANGUAGES Arawak Carib and Cariban Languages Chibchan Languages Dakota and Siouan Languages Maya and Mayan Languages Miskito and Misumalpan Languages Mohawk and Iroquoian Languages Nahuatl and Uto-Aztecan Languages Navajo and Athabaskan-Eyak Languages Nootka and Wakashan Languages

Ojibwe and Algonquian Languages Okanagan and Salishan Languages Oto-Manguean Languages Tarascan Tewa and Kiowa-Tanoan Languages Wayampi and Tupi-Guarani Languages APPLIED LINGUISTICS Applied Linguistics: Overview Computer-Assisted Language Learning (CALL) Forensic Linguistics Genre Language Planning Literacy Professions for Linguists Proficiency Testing Reading Second Language: Teaching Stylistics Teaching: Curricula Teaching: Methods ARTIFICIAL LANGUAGES Artificial Languages AUSTRALIAN LANGUAGES Kayardild and the Tangkic Languages Murrinh-Patha and Daly Languages Pintupi and Pama-Nyungan Languages AUSTRO-ASIATIC LANGUAGES Khmer and Mon-Khmer Languages Vietnamese

xv

Thematic List of Entries AUSTRONESIAN LANGUAGES Javanese Malagasy Malay-Indonesian and Malayic Languages Ma¯ ori and Polynesian Languages Marshallese and Micronesian Languages Tagalog and Philippine Languages COMMUNICATION THEORY Animals and Human Language 1: Overview Animals and Human Language 2: Dolphins Animals and Human Language 3: Parrots Animals and Human Language 4: Primates Biosemiotics Interpreting Paralanguage Translation COMPUTATIONAL LINGUISTICS Artificial Intelligence Finite State Morphology Generation Information Retrieval Machine Translation Parsing Speech Processing Text Understanding Visual Word Recognition Word Sense Disambiguation DEAD LANGUAGES Ancient Egyptian Ancient Greek Gothic Hittite Latin Middle English Middle (Classical) Japanese Old Chinese Old Church Slavonic Old English Old French Old High German Old Irish Old Japanese Old Norse Old Tibetan Sanskrit Sumerian Tocharian

xvi

DRAVIDIAN LANGUAGES Tamil Telugu HISTORICAL LINGUISTICS Analogical Change Archeology and Language Comparative Method Dialectology Etymology Evolution of Language: Overview Evolution of Language: Cognitive Preadaptations Evolution of Language: Physical Preadaptations Evolution of Language: Social Preadaptations Genetic Relationship Great Vowel Shift Language Change Long-Range Comparison Migration and Language HISTORY OF LINGUISTICS Arabic Traditional Grammar Babylonian Traditional Grammar Chinese and Japanese Traditional Grammar European Traditional Grammar History of Linguistics: Overview Indian Traditional Grammar Modern Linguistics INDO-EUROPEAN LANGUAGES Albanian Armenian Baltic Languages Celtic Languages Dutch English Farsi French Language German Greek, Modern Hindi-Urdu Italian Polish and West Slavic Languages Romanian Russian and East Slavic Languages Serbo-Croatian and South Slavic Sindhi Spanish and Iberoromance Languages Swedish and Scandinavian Languages Yiddish

Thematic List of Entries ISOLATES (LANGUAGES) Ainu Basque Burushaski Japanese Korean Warao Zuni LANGUAGE CONTACT Bilingual Mixed Languages Bilingualism Endangered Languages Language: Contact—Overview Language Death Lexical Borrowing Official Language Selection Pidgins and Creoles Sprachbund Structural Interference LEXICON Clause Lexemes Grammaticalization Idiomaticity Lexicalization Lexicon: Overview LINGUISTIC APPROACHES Communication Theory Computational Linguistics Connectionism Corpus Linguistics Field Methods Functional Approaches Generative Grammar Grammar, Theories Grammar, Traditional Historical Linguistics Mathematical Linguistics Morphology Neurolinguistics Philology Philosophy of Language Phonetics Phonology Pragmatics Psycholinguistics Rhetoric and Linguistics Semantics

Semiotics Sociolinguistics Structuralism Syntax Typology Writing Systems MAJOR LANGUAGE FAMILIES Afroasiatic Altaic Austroasiatic Austronesian Dravidian Eskimo-Aleut Hmong and Hmong-Mien (Miao-Yao) languages Indo-European 5: Slavic Indo-European 4: Germanic Languages Indo-European 3: Indo-Iranian Indo-European 2: Overview Indo-European 1: Romance Khoisan Niger-Congo Nilo-Saharan Languages Pama-Nyungan Semitic Languages Sino-Tibetan Languages MORPHOLOGY Affixation Agreement Compounding Empty Morphemes Gender: Class Marking Inflection and Derivation Morpheme Number Marking Sandhi Switch-Reference Tense and Aspect Marking Vowel Harmony Word NEUROLINGUISTICS Aphasia Brain Organization and Auditory Pathway Clinical Linguistics Dyslexia Lateralization and Handedness Localization of Linguistic Information Reading Impairment

xvii

Thematic List of Entries Second Language: Acquisition Speech Processing (Neurobiology) Speech Production: Neurobiology Working Memory NIGER-CONGO LANGUAGES Akan and Nyo Languages Bambara, Mandenkan, and the Mande Languages Ewe and Gbe Languages Igbo and Igboid Languages Moore and the Gur Languages Tiv and Tivoid Languages Wolof and Atlantic Languages Yémba and the Grassfields Bantu Languages Yoruba and Yoruboid Languages Zulu and Southern Bantu Languages NORTH CAUCASIAN LANGUAGES Georgian and Caucasian Languages Northwest Caucasian Languages PERSONS Austin, John Langshaw Bar-Hillel, Yehoshua Baudouin de Courtenay, Jan Ignacy NiecisBaw Benveniste, Emile Bever, Thomas Bloomfield, Leonard Boas, Franz Bopp, Franz Bresnan, Joan Brugmann, Karl Bühler, Karl Chafe, Wallace Chao, Yuen Ren Chomsky, Noam Clark, Eve V. Emeneau, Murray Barnson Ferguson, Charles Albert Fillmore, Charles John Firth, John Rupert Fishman, Joshua A. Fromkin, Victoria Alexandra Givón, Talmy Greenberg, Joseph Harold Grice, H. Paul Grimm, Jacob Gumperz, John Joseph Haas, Mary Rosamond Hale, Kenneth Halle, Morris

xviii

Halliday, Michael Alexander Kirkwood Harris, Zellig Sabbetai Haugen, Einar Hjelmslev, Louis Hockett, Charles F. Humboldt, Wilhelm von Hymes, Dell Hathaway Jakobson, Roman Jespersen, Otto Jones, Sir William Labov, William Ladefoged, Peter Lakoff, George Langacker, Ronald Lehmann, Winfred Philipp Leskien, August Lévi-Strauss, Claude Lyons, John Malkiel, Yakov Martinet, André McCawley, James David Meillet, Antoine Montague, Richard Nida, Eugene Albert Pa¯ nini " Paul, Hermann Peirce, Charles Sanders Pike, Kenneth Lee Quine, Willard van Orman Reichenbach, Hans Russell, Bertrand Arthur William Sapir, Edward Saussure, Ferdinand de Searle, John R. Slobin, Dan Isaac Swadesh, Morris Sweet, Henry Trubetzkoy, Nikolai Sergeyevich Turing, Alan Wackernagel, Jacob Weinreich, Uriel Whorf, Benjamin Lee Wittgenstein, Ludwig Zipf, George Kingsley PHONETICS Acoustic Phonetics Aerodynamics of the Vocal Tract Anatomy of the Articulatory System Anatomy of the Auditory System Assimilation and Dissimilation

Thematic List of Entries Manner of Articulation Phonetic Transcription Place of Articulation Prosody Soundwave Analysis Spectral Analysis Speech Synthesis PHONOLOGY Assimilation and Dissimilation Epenthesis and Syncope Feature Theory Metathesis Natural Classes Phoneme Rules vs. Constraints Syllable: Structure Tone Languages PRAGMATICS Code-Switching Coherence in Discourse Conversation Analysis Deixis Discourse Analysis Discourse Strategies Irony Politeness Relevance in Discourse Speech Acts Taboo Text Linguistics Utterance-Centered Linguistics PSYCHOLINGUISTICS Age and Language Handwriting Speech Perception Speech Production PIDGINS AND CREOLES Cape Verdean Creole Chinese Pidgin English Chinese Pidgin Russian Chinook Jargon Crioulo, Gulf of Guinea Crioulo, Upper Guinea Fanakalo Gullah Haitian Creole Hawaiian Creole English Hiri Motu

Indian Ocean Creoles Juba and Nubi Arabic Kituba Krio (and West African Pidgin English) Kriol (Roper River Creole) Lesser Antillean French Creole Lingua Franca Louisiana Creole Papiamento Philippine Spanish Creole(s) Pitcairnese Saramaccan Western Caribbean Creole Sranan Sri Lanka Portuguese Tok Pisin REGIONS Algeria Australia Austria Balkans Belgium Canada Caribbean China France Germany Great Britain India Israel Italy Mexico New Guinea Nigeria Pacific Scandinavia South Africa South America: Argentina South America: Brazil Southeast Asia Soviet Union Soviet Union: Successor States Spain Switzerland Turkey United States SEMANTICS Ambiguity Causation

xix

Thematic List of Entries Color Terms Compositional Semantics Context Coordination Definiteness Figurative Speech Genericity Idioms Kinship Terms Meaning Metaphor Metonymy Modification Naming Onomatopoeia Plurality Proper Nouns Quantification Reference Register Sapir–Whorf Hypothesis Sign Relationship Time and Tense SIGN LANGUAGES American Sign Language British Sign Language Indo-Pakistani Sign Language Japanese Sign Language Signed Languages SINO-TIBETAN LANGUAGES Burmese Chinese: Mandarin Tibetan SOCIOLINGUISTICS African American Vernacular English Courtroom Discourse Diglossia Emotion and Language Ethnicity and Language Ethnography of Communication Euphemism Gender and Language Identity and Language Ideology and Language Mass Media and Language Medicine and Language

xx

Personality and Language Sociolect Standard Language Variation SYNTAX Anaphora Aspect Auxiliaries Case Clause Configurationality Constituency Test Deep Structure Determiner Empty Categories Focality Function Words Grammatical Function Iconicity Long-Distance Dependency Mood Noun Incorporation Phrase Structure Possessives Predication Pro-forms Questions Serial Verb Constructions Story Grammar Syntactic Categories Tense: Syntax Thematic Structure Universal Grammar Voice Word Order TAI LANGUAGES Thai and Tai Languages TYPOLOGY Morphological Typology Semantic and Discourse Typology Syntactic Typology URALIC LANGUAGES Finnish and Finnic Languages Hungarian and Ugric Languages Udmurt

Contributors Abazova, Alfia. Jersey City, New Jersey.

Baptista, Marlyse. Linguistics Program, University of Georgia, USA.

Ackema, Peter. University of Edinburgh, UK. Aikhenvald, Alexandra. Research Center for Linguistic Typology, Institute for Advanced Study, La Trobe University, Melbourne, Australia. Altarriba, Jeanette. Department of Psychology, University at Albany, State University of New York, USA. Altmann, Lori J.P. Department of Communication Sciences and Disorders, University of Florida, USA. Anderson, Warren D. Department of Foreign Languages and Anthropology, Southeast Missouri State University, USA.

Barcelona, Antonio. Departamento de Filología Inglesa, Universidad de Murcia, Spain. Barnard, Roger. Department of General and Applied Linguistics, University of Waikato, New Zealand. Barron, Anne. Englisches Seminar, Universität Bonn, Germany. Barry, Betsy. University of Georgia, USA. Barss, Andrew. Department of Linguistics, University of Arizona, USA. Barton, Ellen. Linguistics Program, Department of English, Wayne State University, Michigan.

Antieau, Lamont. Linguistics Program, University of Georgia, USA.

Bashir, Elena. University of Chicago, Illinois.

Archibald, John. Department of Linguistics, University of Calgary, Alberta, Canada.

Bastiaanse, Roelien. University of Groningen, The Netherlands.

Arends, Jacques. Department of Linguistics, University of Amsterdam, The Netherlands.

Basturkmen, Helen. Department of Applied Language Studies and Linguistics, Auckland University, New Zealand.

Aronoff, Mark. Department of Linguistics, Stony Brook University, Stony Brook, New York, USA.

Bat-El, Outi. Department of Linguistics, Tel-Aviv University, Israel.

Attardo, Salvatore. English Department, Youngstown State University, Ohio.

Battaner-Moro, Elena. Departamento de Lengua Española y Lingüística General, Universidad Nacional de Educación a Distancia, Madrid, Spain.

Axelrod, Melissa. Department of Linguistics, University of New Mexico, USA.

Bauer, Laurie. Linguistics and Applied Language Studies, Victoria University of Wellington, New Zealand.

Baldauf, Richard B., Jr. School of Education, University of Queensland, Australia.

Baumgartner, Joanne Marie. Australian National University, Canberra, Australia.

xxi

Contributors Beckwith, Christopher I. Indiana University, USA.

Carter, Michael G. Department of Oriental and East European Studies, Oslo University, Norway.

Belnap, R. Kirk. Brigham Young University, Utah. Benner, Allison. Victoria, British Columbia, Canada. Benremouga, Karima. Fine Arts and Language Arts, San Jacinto College Central, Pasadena, TX. Berge, Anna. Alaska Native Language Center, University of Alaska, Fairbanks, USA. Bernhardt, Karl. University of Buckingham, United Kingdom. Bhatia, Tej K. Syracuse University, New York. Bird, Steven. Linguistic Data Consortium, University of Pennsylvania, Philadelphia.

Clankie, Shawn M. Center for Language Studies, Otaru University of Commerce, Japan. Cloutier, Robert A. Amsterdam Center for Language and Communication/Department of Dutch Studies, University of Amsterdam, Amsterdam, The Netherlands. Cockerton, Camilla M. Happy Valley, Rakaia, New Zealand. Cogill-Koez D. Department of Linguistics, University of New England, Armidale, Australia. Cohen, Ariel. Department of Foreign Literatures and Linguistics, Ben-Gurion University of the Negev, Beer Sheva, Israel.

Blume, Kerstin. Bergische Universität Wuppertal, Germany. Bodomo, Adams. Department of Linguistics, University of Hong Kong, China.

Colarusso, John. McMaster University, Hamilton, Ontario, Canada.

Bolotin, Naomi. Cambridge, Massachusetts.

Cole, Jennifer. Department of Linguistics, University of Illinois at Urbana-Champaign, USA.

Born, Renate. Department of Germanic and Slavic Languages, University of Georgia, USA.

Colley, Michael. Linguistics Program, University of Georgia, USA.

Bowden, John. Linguistics Department, Research School of Pacific and Asian Studies, The Australian National University, Australia.

Comrie, Bernard. Max-Planck-Institut für evolutionäre Anthropologie, Leipzig, Germany.

Bretones, Carmen. International Computer Science Institute, Berkeley, California. Brown, Jeffrey M. Texas A&M International University, USA. Branigan, Holly. School of Philosophy, Psychology, and Language Sciences, University of Edinburgh.

Corré, Alan D. Hebrew Studies, University of WisconsinMilwaukee, USA. Cots, Josep Maria. Departament d’Anglès i Lingüística, Universitat de Lleida, Catalunya, Spain. Croft, William. Department of Linguistics, University of Manchester, United Kingdom.

Burley, Lynn Ann. Department of Writing and Speech, University of Central Arkansas, USA.

Crowley, Terry. Department of General and Applied Linguistics, University of Waikato, New Zealand.

Camdzic, Amela. University College London, United Kingdom.

Daniels, Peter T. Bronx, New York.

Campana, Mark. Kobe, Hyogo-ken, Japan.

Daniliuc, Radu. Department of Linguistics, Australian National University, Canberra, Australia.

Carnie, Andrew. Department of Linguistics, University of Arizona, USA.

Daniliuc, Laura. Department of Linguistics, Australian National University, Kambah, Australia.

Carstairs-McCarthy, Andrew. Linguistics Department, University of Canterbury, Christchurch, New Zealand.

Daza, Jaime Luis. Linguistics Department, Unversity of California, Los Angeles, USA.

xxii

Contributors Dimitriu, Rodica. Department of English, University of Iasi, Romania.

González-Romero, Luisa. Departamento de Filología Inglesa, Universidad de Huelva, Andalucía, Spain.

Ding, Picus Sizhi. Macao, Hong Kong, China.

Gordon, Lynn. Department of English, Washington State University, USA.

Dooley Collberg, Sheila. Department of Linguistics, University of Arizona, USA. Douglas, Dan. Department of English, Iowa State University, USA. Dumas, Bethany K. Department of English, University of Tennessee, USA. Ebermann, Erwin. Department of African Studies, University of Vienna, Austria. Echeruo, Michael J.C. Department of English, Syracuse University, Syracuse, New York, USA. Eira, Christina. Adelaide Graduate Centre, University of Adelaide, Australia. Engelberg, Stefan. University of Wuppertal, Germany.

Graesser, Arthur. Department of Psychology, University of Memphis, USA. Green, Ian. Graduate School, University of Tasmania, Hobart, Australia. Greenberg, Marc L. Slavic Department, University of Kansas, USA. Grenoble, Lenore A. Program in Linguistics and Cognitive Science, Dartmouth College, New Hampshire. Gribble, Charles. Slavic Department, Ohio State University, USA. Grohmann, Kleanthes K. Department of English Studies, University of Cyprus, Nicosia.

Ennaji, Moha. Fez, Morocco.

Grondona, Veronica M. Department of English, Eastern Michigan University, USA.

Essegbey, James. Department of African Linguistics, Leiden University, The Netherlands.

Guillaume, Jacques. Paris, France.

Evans, Nicholas. Department of Linguistics & Applied Linguistics, University of Melbourne, Australia. Falk, Julia S. La Jolla, California, USA. Feldman, Laurie Beth. University of Albany, State University of New York, USA.

Gutiérrez-Rexach, Javier. Department of Spanish & Portuguese, Ohio State University, USA. Haberlandt, Karl. Department of Psychology, Trinity College, Connecticut, USA. Hahn, Udo. Linguistische Informatik/Computerlinguistik, Freiburg University, Germany.

Fettes, Mark. Simon Fraser University, Ottawa, Ontario, Canada.

Hajicova, Eva. MFF UK, Prague, Czech Republic.

Flier, Michael S. Department of Slavic Languages and Literatures, Harvard University, Massachusetts, USA.

Hall, David R. Linguistics Department, Macquarie University Sydney, Australia.

Foster-Cohen, Susan H. The Champion Centre and University of Canterbury, Christchurch, New Zealand.

Hall, Kira. Department of Linguistics, Stanford University, California.

Gagné, Christina L. Department of Psychology, University of Alberta, Edmonton, Canada.

Hammond, Michael. Department of Linguistics, University of Arizona, USA.

Gildea, Spike. Department of Linguistics, University of Oregon, USA.

Harlow, Ray. Department of General and Applied Linguistics, University of Waikato, New Zealand.

Goatly, Andrew. Department of English, Lingnan University, N.T., Hong Kong, China.

Harrington, Jonathan. Institute of Phonetics and Digital Speech Processing, University of Kiel, Germany.

xxiii

Contributors Harrison, K. David. Linguistics Department, Swarthmore College, Swarthmore, Pennsylvania, USA.

Jaggar, Philip J. School of Oriental & African Studies, London, United Kingdom.

Hartig, Matthias. Sprach- und Literaturwissenschaft, Universität Paderborn, Germany.

Jensen, Cheryl. Summer Institute of Linguistics, Belém, Pará, Brazil.

Hartsuiker, Robert J. Department of Experimental Psychology, Ghent University, Belgium.

Johnson, E. Marcia. Department of General & Applied Linguistics, University of Waikato, New Zealand.

Heine, Bernd. Institut für Afrikanistik, Universität zu Köln, Germany.

Johnson, Ellen. Department of English, Rhetoric, and Writing, Berry College, Georgia.

Hendriks, Petra. Center for Language and Cognition Groningen, University of Groningen, The Netherlands. Heredia, Roberto R. Texas A&M International University, USA. Herrmann, Stefanie. University of Tübingen, BadenWürttemberg, Germany. Holtgraves, Thomas M. Psychological Sciences, Ball State University, Indiana. Hongladarom, Krisadawan. Department of Linguistics, Faculty of Arts, Chulalongkorn University, Bangkok, Thailand. Hopple, Paulette. Summer Institute of Linguistics, University of Texas, Arlington, USA. Howe, Darin. Department of Linguistics, University of Calgary, Alberta, Canada. Hua, Zhu. Department of Speech, University of Newcastle upon Tyne, United Kingdom. Hudson, Grover. Linguistics Department, Michigan State University, USA. Hurford, James R. Department of Linguistics, University of Edinburgh, Scotland, United Kingdom.

Jones, Francis R. School of Modern Languages, University of Newcastle, United Kingdom. Jongman, Allard. Department of Linguistics, University of Kansas, USA. Justus, Carol F. Classics Department, University of Texas, Austin, USA. Karimi, Simin. Department of Linguistics, University of Arizona, USA. Katz, Leonard. Haskins Laboratories, New Haven, Connecticut. Kaye, Alan S. Linguistics Department, California State University-Fullerton, USA. Kemper, Susan. Department of Psychology, University of Kansas, USA. Kibrik, Andrej. Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany. King, Robert D. Department of Linguistics, University of Texas, Austin, USA. Kirtchuk-Halevi, Pablo Isaac. Beer Sheva, Israel. Knight, Chris. School of Social Sciences, University of East London, Dagenham, Essex, United Kingdom.

Hutchins, W. John. Norwich, United Kingdom.

Kroskrity, Paul V. Department of Anthropology, University of California, Los Angeles, USA.

Ibarretxe-Antunano, Iraide. Facultad de Filosofía y Letras, Universidad de Deusto - Deustuko Unibertsitatea, Bizkaia, Spain.

Kull, Kalevi. Department of Semiotics, University of Tartu, Estonia.

Inchaurralde, Carlos. Departamento Filologia Inglesa y Alemana, Universidad de Zaragoza, Spain. Insler, Stanley. Department of Linguistics, Yale University, Connecticut.

xxiv

Kullavanijaya, Pranee. Department of Linguistics, Faculty of Arts, Bangkok, Thailand. Labeau, Emmanuelle. School of Languages and Social Sciences, Aston University, Birmingham, West Midlands, United Kingdom.

Contributors Lahousse, Karen. Katholieke Universiteit Leuven (Catholic University Leuven) & Fonds voor, Wetenschappelijk Onderzoek–Vlaanderen (Fund of Scientific Research–Flanders).

Mahmoodi-Bakhtiari, Behrooz. Center for the Great Islamic Encyclopedia, Tehran, Iran.

Lainio, Jarmo. Centre for Finnish Studies, Mälardalen University, Eskilstuna/Västerås, Swedan.

Martín Arista, Javier. Departamento de Filologías Modernas, Universidad de La Rioja, Logroño, Spain.

Lang, George. Dean of Arts, University of Ottawa, Canada.

Martín Camacho, José Carlos. Facultad de Filosofía y Letras, Universidad de Extremadura, Cáceres, Spain.

Lasnik, Howard. University of Maryland, USA. Leiber, Justin. Philosophy Department, University of Houston, Texas. Leitner, Gerhard. Institut fuer englische Philologie, Freie Universität Berlin, Germany. Levis, John M. Department of English, Iowa State University, USA. Lieberman, Philip. Department of Cognitive and Linguistic Sciences, Brown University, Rhode Island.

Marret, Carine. Nice, France.

Mattina, Anthony. Department of Anthropology, University of Montana, USA. Mattissen, Johanna. Institut für Linguistik (Department of Linguistics), University of Köln (Cologne), Germany. Maxwell, Judith M. Department of Anthropology, Tulane University, Louisiana. Mc Laughlin, Fiona. African Languages and Linguistics, University of Florida, USA. McDonald, David. Arlington, Massachusetts.

Llurda, Enric. Departament d’Anglès i Linguistica, Universitat de Lleida, Catalunya, Spain.

McGregor, William B. Department of Linguistics, University of Aarhus, Denmark.

Loengarov, Alexander. Research Assistant of the Fund for Scientific Research–Flanders (Belgium). Department of Linguistics, K.U. Leuven, Belgium.

McMenamin, Gerald. Fresno, California.

Lohr, Marisa. Trinity College, Cambridge, United Kingdom. Louwerse, Max M. Department of Psychology/Institute for Intelligent Systems, University of Memphis, USA. Luo, Yongxian. Melbourne Institute of Asian Languages and Societies, University of Melbourne, Australia. Luraghi, Silvia. Dipartimento di Linguistica, Università di Pavia, Italy.

Meisel, Jürgen M. Institut für Romanistik, Universität Hamburg, Germany. Melançon, Megan E. Department of English, Speech, and Journalism, Georgia College and State University, USA. Menchetti, Angiolo. Dipartimento di Scienze Storiche del Mondo Antico, Università di Pisa, Italy. Miller, Catherine. Iremam-MMSH, University of Aix en Provence, France. Minkova, Donka. Department of English, University of California, Los Angeles, USA.

Mackert, Michael. German-English Language Services, Morgantown, West Virginia, USA.

Mithun, Marianne. Department of Linguistics, University of California, Santa Barbara, USA.

MacMahon, Michael K.C. Department of English Language, University of Glasgow, Scotland, United Kingdom.

Moore, John. Department of Linguistics, University of California, San Diego, USA.

Mahanta, Deepshikha. Central Institute of English and Foreign Languages, Hyderabad, India. Mahdi, Waruno. Berlin, Germany.

Morgan, Michael. Kobe City University of Foreign Studies, Kobe, Japan. Mufwene, Salikoko. Department of Linguistics, University of Chicago, Illinois, USA.

xxv

Contributors Mukherjee, Joybrato. Department of English, Justus Liebig University, Giessen, Germany.

Paap, Kenneth R. Department of Psychology, New Mexico State University, USA.

Müller, Horst M. AG Experimentelle Neurolinguistik, Fakultät für Linguistik und Literaturwissenschaft, Universität Bielefeld, Germany.

Paltridge, Brian. OE Faculty of Education, University of Sydney.

Murray, Robert W. Department of Linguistics, University of Calgary, Alberta, Canada. Murray, Denise. National Centre for English Language Teaching and Research, Macquarie University, Ryde, Australia. Nash, David. Australian National University, Canberra, Australia. Nau, Nicole. General Linguistics, Kiel University, Germany. Newmark, Leonard. University of California, San Diego, USA. Nichols, Lynn. Department of Linguistics, University of California, Berkeley.

Pandey, Anjali. Department of English, Salisbury University, Maryland, USA. Pandey, Anita. Department of English & Language Arts, Morgan State University, Maryland Papangeli, Dimitra. Faculteit der Letteren, Utrecht Institute of Linguistics, OTS, The Netherlands. Papen, Robert A. Linguistique et de didactique des langues, Université du Québec, Montréal, Canada. Parasher, Shree Vallabh. Central Institute of English and Foreign Languages, Hyderabad, India. Parkvall, Mikael. Institutionen för Lingvistik, Stockholms Universitet, Sweden.

Nifadopoulos, Christos G. Bristol, United Kingdom.

Pensalfini, Rob. School of English, Media Studies and Art History, University of Queensland, Brisbane, Australia.

Noma, Hideki. Department of Korean Studies, Tokyo University of Foreign Studies, Japan.

Pepperberg, Irene M. Department of Psychology, Brandeis University, Waltham, MA 02454.

O’Rourke, Sean. Department of Linguistics, Yale University, Connecticut.

Perkins, Michael. Human Communication Sciences, University of Sheffield, South Yorkshire, United Kingdom.

Oakes, Michael P. University of Sunderland, United Kingdom.

Pierce, Marc. Department of Germanic Languages and Literatures, University of Michigan, USA.

Obeng, Samuel Gyasi. Linguistics Department, Indiana University, USA.

Piller, Ingrid. Linguistics Department, University of Sydney, Australia.

Ohala, John J. Department of Linguistics, University of California at Berkeley, USA.

Pineda, Baron. Department of Anthropology, Oberlin College, Ohio.

Oliverio, Giulia. Alaska Native Language Center, University of Alaska-Fairbanks, USA. Ongstad, Sigmund. Oslo University College, Norway. O . la Orie, O . lanike. . Linguistic Program & Department of Anthropology, Tulane University, Louisiana. Orr, Robert. Gloucester, Ontario, Canada. Otero, Carlos. University of California, Los Angeles, USA. Oventile, Robert S. Pasadena City College, California.

xxvi

Pires, Acrisio. Department of Linguistics, University of Michigan, USA. Ponchon, Thierry. Université de Reims, France. Prakasam, Vennelakanti. Central Institute of English and Foreign Languages, Lucknow, India. Quesada, J. Diego. Escuela de Literatura y Ciencias del Lenguaje. Universidad Nacional. Heredia. Costa Rica. Radev, Dragomir R. School of Information, Department of Electrical Engineering and Computer Science, and Department of Linguistics, University of Michigan, USA.

Contributors Rajagopalan, Kanavillil. State University at Campinas, Campinas, Brazil. Rathert, Monika. Seminar für Sprachwissenschaft, Universität Tübingen, Baden-Württemberg, Germany. Reh, Mechthild. Department of African Languages and Ethiopian Studies, Hamburg University, Germany.

Sayahi, Lotfi. Department of Languages, Literatures and Cultures, The University at Albany, State University of New York, USA. Schneider, Klaus P. Englisches Seminar, Universität Bonn, Germany. Scott, Suzanne. Dunedin, New Zealand.

Reid, Nicholas. School of Languages, Cultures and Linguistics, University of New England, Armidale, Australia.

Sew, Jyh Wee. Singapore.

Reid, Lawrence A. Tokyo, Japan.

Sgall, Prague Petr. Faculty of Mathematics and Physics, Prague, Czech Republic.

Reintges, Chris H. Center for Linguistics, Leiden University, The Netherlands. Reiss, Charles. Linguistics Program, Concordia University, Montreal, Canada.

Sezer, Engin. Bilkent University, Ankara, Turkey.

Sgarbas, Kyriakos N. Department of Electrical and Computer Engineering, University of Patras, Greece. Shekar, Chandra. California State University, Fresno.

Riedinger, Edward A. Department of Spanish and Portugese, The Ohio State University, USA.

Sicoli, Mark. Departments of Anthropology and Linguistics, University of Michigan, USA.

Rodríguez-Arrizabalaga, Beatriz. Departamento de Filología Inglesa. University of Huelva, Andalucía, Spain.

Siegel, Jeff. School of Languages, Cultures and Linguistics, University of New England, Australia, and Department of Second Language Studies, University of Hawai’i, USA.

Rogers, Henry. Department of Linguistics, University of Toronto, Ontario, Canada. Romaine, Suzanne. Merton College, University of Oxford, United Kingdom. Rubio, Gonzalo. Department of Classics and Ancient Mediterranean Studies, Pennsylvania State University, USA. Rutkowski, PaweB. Department of Polish, Warsaw University, Poland. Sackmann, Robin. Freie Universität Berlin, FB PhilGeist, Freie Universität Berlin, Germany. Sanders, Alton. Department of Computer Science and Systems Analysis, Miami University, Ohio.

Signorini, Inês. Department of Applied Linguistics, Universidade Estadual de Campinas (UNICAMP), São Paulo, Brazil. Silzer, Peter J. Department of TESOL and Applied Linguistics, Biola University, USA. Siple, Patricia. Department of Psychology, Wayne State University, Michigan. Sitzmann, Alexander. Institut für Germanistik, Universität Wien, Austria. Sivell, John N. Department of Applied Language Studies, Brock University, St. Catharines, Ontario, Canada.

Sanders, Carol. Department of Linguistic, Cultural and International Studies, University of Surrey, Guildford, United Kingdom.

Smith, Ian R. Department of Languages, Literatures and Linguistics, York University, Toronto, Ontario, Canada.

Sanders, Ruth H. Department of German, Russian, East Asian and Hebrew, Miami University, Ohio.

Smith, Allison. Department of English, Louisiana Tech. University, USA.

Sansò, Andrea. Dipartimento di Linguistica, Universita’ di Pisa, Italy.

Snoj, Marko. Institut za slovenski jezik ZRC SAZU, Slovenia.

Savova, Lilia. Department of English, Indiana University of Pennsylvania, USA.

Sotillo, Susana M. Department of Linguistics, Montclair State University, New Jersey.

xxvii

Contributors Steinkrüger, Patrick O. Zentrum für allgemeine Sprachwissenschaft, Berlin, Germany.

Vovin, Alexander. Department of East Asian Languages and Literatures, University of Hawai’i, USA.

Strazny, Philipp. Monona, USA.

Vrooman, Michael. Department of Modern Languages and Literatures, Grand Valley State University, Michigan.

Sujoldzic, Anita. Institute for Anthropological Research, Zagreb, Croatia. Sunderland, Jane. Department of Linguistics and English Language, Bowland College, Lancaster University, Lancashire, United Kingdom. Tadmor, Uri. Department of Linguistics, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany/Jakarta, Indonesia. Tambovtsev, Yuri. Department of English and Linguistics, KFNPU, Novosibirsk, Russia.

Wagner, Richard. Department of Psychology, Florida State University, USA. Walker, Douglas C. Department of French, Italian and Spanish, University of Calgary, Alberta, Canada. Walker, James A. Department of Languages, Literatures, and Linguistics, York University, Toronto, Ontario, Canada. Wang, William S.-Y. Department of Electronic Engineering, City University of Hong Kong, China.

Tatham, Mark. Colchester, United Kingdom.

Watt, William C. Social Science, University of California, Irvine, USA.

ten Hacken, Pius. School of European Languages, University of Wales Swansea, UK.

Waugh, Linda R. Department of French, Italian and English, University of Arizona, USA.

Thomason, Sarah G. University of Michigan, USA.

Wehmeyer, Ann. Department of African and Asian Languages and Literatures, University of Florida, USA.

Thomassen, Arnold J.W.M. University of Nijmegen, The Netherlands. Toribio, Josefa. Department of Philosophy, The University of Edinburgh, Scotland. TSE, Grace Yuen Wah. School of Arts and Social Sciences, The Open University of Hong Kong, China. Tsedryk, Egor. Department of French Studies, University of Western Ontario, London, Canada and Department of French Studies, Concordia University, Montreal, Canada. Tserdanelis, Giorgos. The Ohio State University, USA.

Weiss, Sabine. Brain Research Institute, Cognitive Neuroscience Group, Medical University of Vienna, Austria. Wexler, Paul. Department of Linguistics, Tel-Aviv University, Israel. Whaley, Lindsay. Linguistics & Cognitive Science, Dartmouth College, New Hampshire. Wilcox, Phyllis Perrin. Department of Linguistics, University of New Mexico, USA.

Turovski, Aleksei. Tallinn Zoo, Tallinn, Estonia.

Wilson, Andrew. Linguistics Department, Lancaster University, United Kingdom.

van Gelderen, Elly. English Department, Arizona State University, USA.

Wireback, Kenneth J. Department of Spanish and Portuguese, Miami University, Ohio.

van Wieringen, Astrid. Lab. Exp. ORL, Katholieke Universiteit Leuven, Belgium.

Woll, Bencie. Sign Language & Deaf Studies, Department of Language and Communication Science, City University, London, United Kingdom.

Vaux, Bert. Department of Foreign Languages and Linguistics, University of Wisconsin-Milwaukee, USA.

Wong, Andrew. Department of Linguistics, University of Hawai’i at Manoa, USA.

Veselinovic, Elvira M. University of Cologne, Germany. Viereck, Wolfgang. University of Bamberg, Germany.

Woodman, Karen. School of Languages, Cultures and Linguistics, University of New England, Armidale, Australia.

Voeltz, F.K. Erhard. Institut für Afrikanistik, Universität zu Köln, Germany.

Zeshan, Ulrike. Research Centre for Linguistic Typology, La Trobe University, Bundoora, Australia.

xxviii

Introduction The study of language goes far back in recorded history. Almost two-and-a-half millenia ago, the Indian grammarian Panini wrote his formal treatises on Sanskrit, Xun Zi appeared as China’s first major philosopher of language, and Plato and Aristotle initiated the Greek philosophy of language. Since the renaissance, there has been an increasing focus on the description of individual languages, the exploration of familial relationships between languages, and the formulation of increasingly general theories of language structure. Now, almost every university or college has a language department or even a specialized linguistics department, which means that an immense number of researchers are working in the field and have published an enormous body of primary literature.

Given this wealth of published data, it is not surprising to meet beginning graduate students of linguistics who already identify themselves as ‘syntacticians’ or ‘phoneticians’. This early specialization reflects genuine interest, but is also in part a mechanism by which students block out a large number of possible inputs to be able to concentrate on a more manageable few. In their subsequent research, most researchers narrow down their field of interest even more; they become immersed in the highly conceptual and terminological world of their specialty, and they often write articles comprehensible only to their small group of peers. In other words, linguistics is a mature science and, as in other scientific fields, there can be a communications gap both within the field and, even more so, between the active researchers and the general public. The Encyclopedia of Linguistics provides an accessible overview of and introduction to the multiple facets of the study of language. To bridge this gap between professional linguists and the general public, my editorial colleagues

and I made this encyclopedia very readable by eliminating technical terminology as far as possible and by making each essay self-contained.

How to Use This Book The Encyclopedia of Linguistics is organized into a series of 508 free-standing essays, between 1000 and 3000 words in length. They range from factual narrative entries to thematic and analytical discussions, and combinations of all these. Where debates and controversies occur, these are indicated and discussed. As far as possible, this book takes the field of linguistics up to the present, at least to the opening years of the twenty-first century. Perhaps the most significant feature of the encyclopedia is the easily accessible A to Z format. Cross-referencing in the form of See Alsos at the end of each entry refers the reader to other related essays. A thorough, analytical Index complements the accessibility of the entries, easing the reader’s entry into the wealth of information provided. References at the end of each entry refer the reader to seminal writings as well as some of the most recent work on the subject. Other special features include 12 language-distribution maps and a thematic Table of Contents in addition to an alphabetical Table of Contents. In addition, more than 100 illustrations are dispersed throughout. A total of 288 scholars from 34 countries have contributed their expertise to this encyclopedia.

Contents These essays have been written by scholars who provide a general introduction to the material without presupposing knowledge about the subject and without going into a

xxix

Introduction theoretical depth that would raise questions that cannot be answered in the space given. The entries fall into the following thematic categories: linguistic topics (for example, code switching) (50%), languages (for example, Sumerian) (30%), persons (for example, Noam Chomsky) (15%), and regions of the world (for example, Algeria) (5%).

how this particular situation came about historically, and discuss language-political issues relevant in this region. We have made a concerted effort to cover languages around the globe. Should you note any imbalance in favor of ‘western’ regions, languages, persons, or topics, this simply reflects that the European languages are the best-studied languages in the world.

Linguistic Topics Among these essays are general introductions to major fields of inquiry, such as semantics, historical linguistics, and neurolinguistics. Other articles concentrate on issues within those fields, introducing concepts that are important in linguistics regardless of theoretical perspective, such as ‘affix’, or ‘reference’, or which are so commonly used that the concept becomes unquestioned, such as ‘phoneme’ or ‘deep structure’. Other articles describe crucial phenomena that any theory of language has to explain, such as ‘acquisition’, ‘tone’, or ‘aphasia’.

Languages These essays cover the major language families of the world and discuss how individual languages are related to each other. Articles concentrating on specific languages from Ainu to Zuni explain where these languages are spoken, by whom, and under what sociopolitical circumstances. They provide a glimpse of the language’s structure and highlight particularly interesting characteristics with examples.

Persons These essays highlight the major theoretical contributions of noted linguists, discuss the influences that led up to their work, and put the contributions into historical perspective, in addition to providing basic biographical sketches.

Regions These essays map out the linguistic landscape of a particular region. They show which languages are spoken, explain

Acknowledgments First of all, I thank Steve Larue and Paul Schellinger of Fitzroy Dearborn Publishers for giving me the opportunity to participate. And to their staff Christy Prahl, Heather Sabel, and Peter Daniels, who did most of the hard work to start this project. At Routledge Reference, a team led by Marie-Claire Antoine—including Kate Aker, Susan Gamer, and Josh Pasternak—successfully completed this work. Deciding which topics to cover was one of the hardest aspects in shaping this reference work, and I am deeply indebted to our advisers who helped shape the book in multiple rounds of suggestions and constructive criticism. Robert Beard and John Goldsmith unfortunately resigned from the Advisory Board when the book took a direction they could not agree with. Still, many thanks to them for their invaluable input while they were part of the project. Among our authors, I would like to specifically acknowledge Alexandra Aikhenvald, Jim Hurford, Mikael Parkvall, and Ulrike Zeshan for their helpful suggestions. And many thanks to Elly van Gelderen, Marc Greenberg, and Ray Harlow, for helping to get the project off the ground with their sample articles. When I first invited potential advisers, a well-known linguist declined, saying that ‘linguists can’t write for lay people’. I firmly believe that our authors proved this statement wrong. Thus, greatest thanks is due our authors, who did an impressive job of introducing their specialty without recourse to the precise terminology they are accustomed to using. I thank both authors and advisers for staying the course during the years from inception to publication of the Encyclopedia of Linguistics. Last but not least, I am deeply indebted to Karin, Isabel, Max, and Mia, who supported me while I disappeared into the basement for months on end to work on this book.

Philipp Strazny

xxx

A Acoustic Phonetics Acoustic phonetics is the study of the acoustic characteristics of speech. Speech consists of variations in air pressure that result from physical disturbances of air molecules caused by the flow of air out of the lungs. This airflow makes the air molecules alternately crowd together and move apart (oscillate), creating increases and decreases, respectively, in air pressure. The resulting sound wave transmits these changes in pressure from speaker to hearer. Sound waves can be described in terms of physical properties such as cycle, period, frequency, and amplitude. These concepts are most easily illustrated when considering a simple wave corresponding to a pure tone. A cycle is a sequence of one increase and one decrease in air pressure. A period is the amount of time (expressed in seconds or milliseconds) that one cycle takes. Frequency is the number of cycles in one second, expressed in hertz (Hz). An increase in frequency usually results in an increase in perceived pitch. Amplitude refers to the magnitude of vibrations, with larger vibrations resulting in greater peaks of pressure (greater amplitude), which usually result in an increase in perceived loudness. Unlike pure tones, which rarely occur in the environment, speech sounds are complex waves with combinations of different frequencies and amplitudes. However, as first stated by the French mathematician Fourier (1768–1830), any complex wave can be described as a combination of simple waves. A complex wave has a regular rate of repetition, known as the fundamental frequency (F0). Changes in F0 give rise to differences in perceived pitch, whereas changes in the number of constituent simple waves and their amplitude relations result in perceived differences in timbre or quality.

Fourier’s theorem enables us to describe speech sounds in terms of the frequency and amplitude of each of its constituent simple waves. Such a description is known as the spectrum of a sound. A spectrum is visually displayed as a plot of frequency vs. amplitude, with frequency represented from low to high along the horizontal axis and amplitude from low to high along the vertical axis. The usual energy source for speech is the airstream generated by the lungs. This steady flow of air is converted into brief puffs of air by the vibrating vocal folds, two muscular folds housed in the larynx. The dominant way of conceptualizing the process of speech production is in terms of the source-filter theory, according to which the acoustic characteristics of speech can be understood as a result of a source component and a filter component. The source component is determined by the rate of vocal fold vibration, which in turn is affected by a number of factors, including the rate of airflow and the mass and stiffness of the vocal folds. The rate of vocal fold vibration directly determines the F0 of the waveform. The mean F0 for adult women is approximately 220 Hz, and approximately 130 Hz for adult men. In addition to their role as properties of individual speech sounds, F0 and amplitude also signal emphasis, stress, and intonation. For speech, the source component itself has a complex waveform, and its spectrum will typically show the highest energy at the lowest frequencies and a number of higher frequency components that systematically decrease in amplitude. This source component is subsequently modified by the vocal tract above the larynx, which acts as the filter. This filter

1

ACOUSTIC PHONETICS 3000 heed hayed

2500

had

hid

Frequency (Hz)

head

2000 heard

hod

hud

1500

hood hawed

who'd hoed

1000 hod head

500 heed

hid

hayed

had

hawed

hud heard

hoed

hood

who'd

0

enhances energy in certain frequency regions and suppresses energy in others, resulting in a spectrum with peaks and valleys, respectively. The peaks in the spectrum (local energy maxima) are known as formant frequencies. The lowest-frequency peak is known as the first formant, or F1, the next lowest is F2, and so on. The vocal tract filter is determined by the size and shape of the vocal tract and is therefore directly affected by the position and movement of the articulators such as the tongue, jaw, and lips. Vowels are typically characterized in terms of the location of the first two formants, as illustrated in Figure 1 for the vowels of American English. For a given speaker, each vowel typically has a unique formant pattern. However, variation in vocal tract size among speakers often leads to a degree of formant overlap for different vowels. Consonants can also be described in terms of their spectral properties. These sounds are produced with a complete or narrow constriction in the vocal tract, essentially creating a vocal tract with two sections: one behind and the other in front of the constriction. The length of the section in front of the constriction is one of the primary determinants of the spectra of these sounds. The longer this section (i.e. the farther back the constriction), the lower the frequency at which a concentration of energy occurs. For example, consonants like k and g, which are produced at the back of the mouth, are typically characterized by a concentration of energy between approximately 1,500 and 2,500 Hz, whereas more anterior consonants like t and d typically have a concentration of energy above 3,000 Hz. Similarly, the sibilants [ʃ,] produced in the middle of the mouth have major energy around 2,500 to 3,500 Hz, whereas the more anterior ones [s, z] have major energy well above 4,000 to 5,000 Hz. However, in the 2

Figure 1. Frequencies of the first two formants of 12 vowels of American English, averaged across 48 adult female speakers. F1 is in black, F2 is in gray. (Source: Hillenbrand et al., J. Acoust. Soc. Am., 1995)

case of consonants with a constriction toward the very front of the vocal tract, the extremely short section in front of the constriction does not result in clearly defined spectra. As a result, bilabial [b, p] and labiodental [f, v] consonants are described as having diffuse spectra, without any clear concentration of energy. From a linguistic point of view, a detailed description of speech sounds in terms of their frequency, in addition to amplitude and duration, can elucidate the factors that shape sound categories and determine phonological processes both within and across languages. In addition, acoustic phonetic analysis may serve to quantify atypical speech patterns produced by nonnative speakers or speakers with specific speech disorders.

References Asher, R., and Eugenie Henderson (eds.) 1981. Towards a history of phonetics. Edinburgh: Edinburgh University Press. Fant, Gunnar. 1960. Acoustic theory of speech production. The Hague: Mouton. Flanagan, James L. 1965. Speech analysis synthesis and perception. Berlin: Springer-Verlag. Hardcastle, William, and John Laver (eds.) 1997. The handbook of phonetic sciences. Oxford and Cambridge, MA: Blackwell. Helmholtz, Hermann von. 1877. Die Lehre von den Tonempfindungen als physiologische Grundlage fur die Theorie der Musik. Braunschweig: F. Vieweg; as On the sensations of tone as a physiological basis for the theory of music, translated by Alexander J. Ellis, London: Longmans, Green, 1885. Hillenbrand, James, Laura A. Getty, Michael J. Clark, and Kimberlee Wheeler. 1995. Acoustic characteristics of American English vowels. Journal of the Acoustical Society of America 97(5). 3099–111. Jakobson, Roman, Gunnar Fant, and Morris Halle (eds.) 1952. Preliminaries to speech analysis. Cambridge, MA: MIT Press.

ACQUISITION Johnson, Keith. 1997. Acoustic and auditory phonetics. Oxford and Cambridge, MA: Blackwell. Kent, Raymond D., Bishnu S. Atal, and Joanne L. Miller (eds.) 1991. Papers in Speech Communication: speech production. Woodbury, New York: Acoustical Society of America. Ladefoged, Peter. 1962. Elements of acoustic phonetics. Chicago and London: The University of Chicago Press, 2nd edition, 1996.

Lehiste, Ilse. 1970. Suprasegmentals. Cambridge, MA: MIT Press. Lieberman, Philip, and Sheila E. Blumstein. 1988. Speech physiology, speech perception, and acoustic phonetics. Cambridge: Cambridge University Press. Stevens, Kenneth N. 1998. Acoustic phonetics. Cambridge, MA: MIT Press.

ALLARD JONGMAN

Acquisition Language acquisition is the study of the development of a person’s language. It generally refers to the way people learn their native, first, second, or other languages. More specifically, it may refer to the time a language feature has been acquired. This may vary from the first emergence or onset of a language item to the time of its accurate use. As a field of study, it is the subject of linguistics, psychology, and applied linguistics. Its object is to study (1) how languages are learned, (2) what are the developmental stages in this process, and (3) what is the nature of language. To find answers to these questions, researchers apply longitudinal and cross-sectional methods. In the first of these, they study specific developments in the language of individuals or groups over a period of time. In the second, they research a particular feature in the language of a group at a given point in time.

First Language Acquisition First language acquisition is the child’s learning of his or her first or native language. Traditionally, and especially in monolingual societies, ‘first’ and ‘native’ language were used synonymously. With the expansion of cross-cultural communication, the two terms become more distinct. For example, children may acquire some knowledge of another language from a nurse or a relative before they acquire their native language, e.g. the language of the country they live in. Thus, a Chinese child born in the United States may first learn Chinese from her parents, and learn English later from English-speaking children and adults. To avoid the confusion arising from the use of ‘first’ and ‘native’, another term, ‘primary’, is sometimes used to indicate a child’s first language chronologically. First Language Acquisition and the Language Acquisition Device Noam Chomsky’s work aroused interest in the way children learn their native language. He believes

that children are born with the ability to learn a language, i.e. they are born with a ‘language acquisi-tion device’. The latter is species-specific or only for humans, language-specific or only for their first/ native language, and innate or only inborn. He also claims that this ability is unconscious and children learn their native language by exposure to it and by using it, and not by being taught or corrected. He argues that as children acquire their native language, they are able to produce sentences that they have not heard before. While early work on children’s language acquisition focused on the development of children’s ability to produce novel sentences, more recently, researchers have emphasized children’s acquisition of word meanings and their linguistic and cognitive development, their acquisition of the phonology of their native language, and their language development in relation to their interaction with parents and peers. Some researchers also see a parallel between the stages of children’s language development regardless of the specific language they are learning. First Language Acquisition and Cognitive Development A child’s language development is closely related to his or her cognitive development. Here, the ability to identify and form categories and concepts is of crucial importance. ‘Categorization’ involves the treatment of distinct linguistic phenomena, such as ‘worked’, ‘studied’, ‘saw’, and ‘went’, as if they were part of the same phenomenon, or the same grammatical category, i.e. past tense. Young children do not have fully developed abilities in categorization. Many childhood first language errors, for example, ‘*I eated it’, point to the gaps in their ability to form categories. Furthermore, even seemingly correct utterances do not imply that the child has achieved an adult stage in the mastery of the corresponding language category.

3

ACQUISITION Closely related to the ability to categorize is the ability to differentiate a category, for example, tense, from the mental structure, which it represents, in this case, time. These mental structures are known as ‘concepts’. To learn a language, a child must acquire the concepts that underlie linguistic structures. It is not possible to master grammatical categories, such as tense, in any language without mastering concepts such as time, space, modality, causality, and number. Young children’s errors in tense indicate that they do not grasp the concept of time. First Language Acquisition and Social Development Children’s social adjustment is as important as their cognitive growth to their language development. As they acquire various language categories and the concepts they represent, children also learn about the cultural, moral, religious, and other conventions of the society they live in. They learn how to express their thoughts, feelings, and wishes in a socially acceptable manner. For example, children learn that it is not always advisable to speak their minds. As they come to realize that words can serve to make friends as well as enemies, they learn that it is not always possible to tell the truth. In this way, while acquiring their first language, children also develop a social identity. How-ever, their progress is slow and not devoid of some rather amusing or even embarrassing errors. Children learning their first language, therefore, have a long way to go, even after they have acquired the basic concepts and their corresponding language categories. By comparison, adult foreign language learners, who are knowledgeable about the sociocultural aspects of their native language, are a step ahead of child learners, even though they may also be prone to similar social blunders because of sociocultural differences. First Language Acquisition and the Critical Period The ‘critical period hypothesis’ claims that there is a period in child development during which language can be acquired with native-like proficiency. Some, like the biologist Lenneberg, believe that this period lasts until puberty, after which the brain loses some of its ability to adapt due to its laterization, i.e. the establishment of specific language functions in particular parts of the brain. After that, the decreased plasticity of the brain makes the acquisition of another language a psychologically different and more difficult process. While there is compelling evidence that supports those claims, there are also important facts that undermine their veracity. First, the strictly biological evidence is by no means conclusive. Second, other factors, such as lack of motivation, may explain nonnative pronunciation.

4

First Language Acquisition and Bilingualism Cognitive and social development, as well as the language acquisition device and the critical period affect the language development of the bilingual as well as the monolingual child. Bilingual first language acquisition is defined as the parallel acquisition of two languages, which is, supposedly, an evenly paced process. However, such a perfect balance can rarely be achieved. Commonly, the child would use one language in one environment, and another in a different setting. Thus, inevitably, one language gains dominance over the other. This dominance may extend to some or all areas of communication. As a result, the child’s other language may become secondary in both development and use. Furthermore, there may be some interference from the dominant language that causes errors in the child’s secondary language. However, there is no evidence that this results in massive confusion in one or both languages. Furthermore, there is no evidence that bilingual children differ from other children in their cognitive, social, or language development.

Second Language Acquisition Second language acquisition (SLA) is defined as the process of becoming competent or proficient in a second or foreign language, from the first use of a language item to its advanced applications at a later stage. As a field of research, SLA is a fairly new interdisciplinary subject with most of its empirical research done since the 1960s. It is largely based on theories and research methods developed in the fields of education, psychology, linguistics, anthropology, foreign languages, English as a second language, and linguistics. In the United States, researchers study the way nonnative speakers acquire English phonology, syntax, and pragmatics. The purpose of SLA studies is to describe and explain the way second languages are learned in terms of both linguistic and communicative competence. To do this, researchers study learners’ performance and their intuitions about correct and incorrect use of language. The object of second language acquisition is to find more effective ways of teaching and learning foreign languages, and assumes that such research can affect the way foreign languages are learned. The Meaning of ‘Second Language’ in Second Language Acquisition There are different interpretations and uses of both ‘second’ and ‘acquisition’ in SLA. ‘Second’ may be used to distinguish it from ‘foreign’ or ‘third’ language acquisition. Traditionally, the terms ‘second’ and ‘foreign’ have been used alternatively to refer to any language other

ACQUISITION than the first. More recently, with the emergence of English as a global language and the establishment of Teachers of English to Speakers of other Languages (TESOL) as a worldwide professional organization, a distinction is made between the two. ‘Second’ language acquisition refers to the study of English by foreigners in countries where English is the native or the official language, whereas ‘foreign’ language acquisition refers to the study of English everywhere else. Furthermore, this distinction extends to differences in what is learned and how it is learned. Learners of English as a foreign language (EFL) prefer standard varieties of English, whereas learners of English as a second language (ESL) try to blend with their sociolinguistic environment. All of these differences are reflected in the goals and methods of EFL and ESL. A further distinction is made between ‘second’ and ‘third’ language acquisition, which marks the learner’s relative proficiency rather than the order in which he or she acquired these languages. Sometimes, the term ‘alternative’ is used to refer to any nonnative language. The Meaning of ‘Acquisition’ in Second Language Acquisition Acquisition is often used to refer to different aspects of the process and results of learning a second language. While trying to find out about the process, i.e. how second languages are learned, researchers often compare different learning experiences that lead to SLA, such as learning a language through organized instruction or in an immersion situation. From a sociolinguistic perspective, acquisition through organized instruction occurs in classrooms with the help of teachers and instructional materials. Acquisition through immersion occurs in social situations using contextual clues. Yet another distinction is made from the psycholinguistic perspective. Klein identifies ‘spontaneous’ and ‘guided’ acquisition. The first focuses on everyday communication, whereas the second targets the mastery of the language system. Similarly, from a psychological point of view, Krashen distinguishes between ‘acquisition’ and ‘learning’. In his analysis of the process of mastering a second language, he reserves ‘acquisition’ for the subconscious process of learning a language by being exposed to it. ‘Learning’, according to him, is the conscious process of mastering a language by studying it. Ellis finds this distinction problematic and considers its demonstration difficult. Furthermore, he states that researchers disagree about what kinds of performance constitute evidence of ‘acquisition’. For some, such evidence can be found in the ways learners speak and write. For others, it is the learner’s intuitions about the second language that matter. Yet another group of scholars seek evidence of acquisition by assessing the learner’s introspections.

Other researchers analyze what it means to know a second language. From a linguistic perspective, Chomsky focuses on the results of SLA, which he defines as ‘competence’ and ‘performance’. According to him, ‘competence’ in a second language is the mastery of the internalized grammar that the ideal speaker or hearer, not a real one, has of the whole language. Such mastery enables him/her to produce grammatically correct sentences as well as recognize existing and nonexisting sentences. For example, knowing the rule, which makes ‘I speak English’ possible, a person can produce ‘I speak French’ even though he or she may not have seen this before. Furthermore, he or she would know that the form ‘*I speaking English’ is nonexistent. ‘Performance’ in Chomsky’s Generative Transformational Grammar, on the other hand, refers to a person’s actual use of a language in the understanding and production of sentences. Unlike communicative competence, which is internal and invisible, performance is external and observable. Furthermore, performance does not mirror competence, since people may know how to produce a sentence but may err when they try to do so. Thus, performance can also be defined as the grammar that a person uses to understand and produce language, which is both correct and appropriate. Performance could be used to investigate competence through the analyses of samples of spoken or written discourse. Within performance, Widdowson distinguishes ‘usage’, which refers to the learner’s ability to apply grammar rules accurately in the production of grammatically correct sentences. ‘Use’, on the other hand, signifies their ability to apply linguistic and sociolinguistic knowledge appropriately and communicate effectively in diverse contexts. Second Language Acquisition Research A large part of SLA research is learner based. It describes and analyzes the nature of learner language and learner differences, learning processes, and pedagogical input and output. It does so to provide answers to important second language research questions, which may offer effective solutions to crucial second language classroom problems. Learner Language in Second Language Acquisition Research Researchers study learner language by examining samples of oral and written texts. Their goal is to identify errors, establish developmental patterns and sequences, trace variability, and explore use. Errors were first believed to be the result of native language ‘transfer’ or ‘interference’. This view was promoted by numerous contrastive analyses conducted from the 1940s to the 1960s. Such studies compared two languages to find out what similarities and differences

5

ACQUISITION existed between the two. Lado thought that similar elements would be easy to learn, while dissimilar ones would be difficult to master. The belief that linguistic difference could be a predictor of difficulty gave rise to the ‘contrastive analysis hypothesis’. That and the behaviorist approach to learning, which claimed that learning is a process of habit formation, led to the belief that SLA should be a process of overcoming habits from the native language and consolidating correct habits in the target language. In 1967, Pitt Corder proposed a new definition of errors. He thought they were systematic deviations from the norm, which reflect the learner’s current stage of second language development. Errors, he claimed, are different from mistakes, which can easily be self-corrected. ‘Error analysis’ treated errors as a sign of the learner’s hypothesis testing, which would ultimately lead to the formation of the correct form and its underlying rule. Thus, errors were seen as part of the learner’s language at every stage of its development. To emphasize its unique features, Corder referred to learner language as ‘idiosyncratic dialect’. Nemser called it an ‘approximative system’ and Selinker coined the term ‘interlanguage’. Thus, the notion of learner language evolved from a faulty, deviant product of the target language to a continuous, approximative progress towards its mastery. Both contrastive and error analysis were criticized for their exclusive reliance on the analysis of a linguistic product, i.e. errors, to yield insights into a psycholinguistic process, i.e. second language acquisition. Another feature of learner language is its passage through a sequence of developmental stages, which are universal. Thus, many of the initial utterances that learners produce may be simple formulae, for example, ‘What’s this’. These are followed by structures of greater morphological and syntactic complexity, for example, ‘I wonder what this might be’. The existence of developmental stages in SLA, which are similar to those in first language acquisition, along with some variations in the specific order in which particular features occur, have renewed interest in grammar instruction. In its systematic development, learner language also exhibits certain variability. For example, learners may use the third person singular ‘s’ correctly sometimes and omit it at other times. In addition to lexical and syntactic variability, they often have problems on the pragmatic level, i.e. they may use language or act in a socially inappropriate manner. Factors in Second Language Acquisition SLA is also influenced by the environment in which it occurs. Social factors, language input, and interaction

6

affect the way learners acquire a second language. Ellis contends that social factors shape learners’ attitudes, which, in turn, may affect motivation and learning outcomes. Social factors include natural and educational settings. For example, natural settings, where English is the native or the official language, offer opportunities different from educational settings, such as the foreign language classroom where the native language is the medium of instruction. While social factors influence second language acquisition indirectly, input, output, and interaction seem to have a direct impact. ‘Input’ is the learner’s access and exposure to the second language, both written and oral. Exposure to the foreign language may engage learners in ‘interactions’ with native or nonnative speakers, or it may involve them in listening to tapes, films, radio, and TV programs. Researchers vary in their assessment of the importance of input and interaction. Behaviorist theories emphasize the importance of input. Chomsky, on the other hand, claims that there is no necessary correlation between language input and learner output. Krashen believes that learners acquire language in a natural order as a result of being exposed to ‘comprehensible input’ addressed to them. In contrast to Krashen, Swain proposes the ‘comprehensible output hypothesis’, which claims that comprehending input alone will not prepare students to produce language. According to him, it is correct production resulting from challenging practice in speaking and writing that facilitates acquisition. Both the comprehensible input and comprehensible output hypotheses have been criticized on the grounds that the processes of comprehensible input and output and the process of SLA are not the same. General factors, such as social setting, input, output, and interaction, result in a variety of individual differences in SLA. Furthermore, individual factors, such as age, language aptitude, motivation, cognitive style, and learning strategies can have similar effects. These factors affect second language learning in ways that are mostly independent of the learner. For example, a learner can do nothing about his or her age, or language aptitude. Few learners may have the opportunity to switch from one educational setting to another. Given the appropriate guidance, however, some learners may be able to improve their motivation and learning strategies over time. For example, ‘extrinsic motivation’, which derives from external rewards, may evolve into ‘intrinsic motivation’, which derives from personal interests. Learner strategies, which contribute to the learner’s conscious efforts to learn, may also change. For example, learners may expand their ‘cognitive strategies’ by learning new concepts. They may also perfect their ‘metacognitive strategies’ by

ACQUISITION THEORIES developing their study skills, or enhance their ‘social strategies’ by practicing their knowledge in authentic social settings. References Bennet-Castor, T. 1988. Analyzing children’s language. Oxford: Blackwell. Brown, Douglas H., and Susan Gonzo. 1995. Readings on second language acquisition. Englewood Cliffs, NJ: PrenticeHall Regents. Chomsky, Noam. 1965. Aspects of the theory of syntax. Cambridge, MA: MIT Press. Cook, Vivian. 1993. Linguistics and second language acquisition. New York: St. Martin’s Press. Ellis, Rod. 1994. The study of second language acquisition. Oxford: Oxford University Press. Fletcher, P. 1985. A child’s learning of English. Oxford: Blackwell. Fletcher, P., and M. Garman (eds.) 1979. Language acquisition. Cambridge: Cambridge University Press.

Hatch, E. (ed.) 1978. Second language acquisition. Rowley: Newbury House. Klein, Wolfgang. 1986. Second language acquisition. Avon: Cambridge University Press. Krashen, Steven. 1981. Second language acquisition and second language learning. Oxford: Pergamon. Larsen-Freeman, Diane, and Michael Long. 1991. An introduction to second language acquisition research. London and NewYork: Longman. Lenneberg, E. 1967. Biological foundations of language. New York: Wiley. Long, Michael. 1995. The least a second language theory needs to explain. Readings on second language acquisition, ed. by Douglas H. Brown and Susan Gonzo. Englewood Cliffs, NJ: Prentice-Hall Regents. Richards, Jack, et al. 1992. Longman dictionary of applied linguistics. Harlow, Essex: Longman. Widdowson, Henry. 1978. Teaching language as communication. Oxford: Oxford University Press.

LILIA SAVOVA See also Acquisition Theories

Acquisition Theories The goal of acquisition theories is to explain how it is that any normal child, born into any linguistic community, learns the language (or languages) of that community. For many theorists, the challenge is also to explain what appears to be the relatively short time period in which acquisition is achieved, the fact that it appears to be done without either overt teaching or sufficient information from the input (what the child hears [or sees, in the case of sign language]), and to follow a path that seems remarkably similar in all children, despite variation in early childhood experiences and in the types of languages they are exposed to. There is also a consensus that language acquisition is largely independent of cognitive development, despite the fact that some deficits in cognitive development can have an effect on certain aspects of language development. Whether the language is a spoken language or a sign language, whether the language is highly inflected like Finnish or uninflected like Mandarin, whether the child is raised in poverty or luxury, by highly educated or illiterate adults, or even other children, it seems that normally developing children pass through roughly the same stages in the same sequence, and achieve the steady state of acquired language by about the same age. We know from unfortunate natural experiments in which children are raised in isolation (or near isola-

tion) from language-using older members of our species (i.e. they are severely neglected, or are raised by other animals) that language does not emerge without at least some linguistic exposure—input—during the first few years of life. This strongly suggests that there is a critical or sensitive period during which the mechanism or mechanisms responsible for language development is/are primed to receive input. However, the resilience of language development to quite wide variations in input within any given language community, as well as the similarities among children learning quite different languages, suggest that these mechanisms, whatever they are, must either be quite tolerant of such variation, or be primed in such a way that the crucial input for language acquisition is always made available. There are important roles for both ‘experiencedependent’ (nurture) and ‘experience-expectant’ (nature) learning in language development, and theories are distinguishable in terms of the relative contributions they see for these two types. On the one hand, there are those researchers who see a large didactic role for input (experience-dependent learning), and on the other, there are those who see a much smaller role for input and a much larger role for genetic predispositions that are triggered by linguistic experience (experienceexpectant learning). Theories are also distinguishable

7

ACQUISITION THEORIES in terms of whether they are trying to explain how language emerges on a day-by-day basis in any given child or whether they are trying to explain how what is perceived as a gulf between experience as a child and adult knowledge of language could be bridged in principle. The latter are engaged in trying to solve the ‘logical problem of language acquisition’. Another dimension of difference between theories concerns the nature of the experience-expectant (innate) aspects of language acquisition. There are those researchers, most notably the generative linguists in a broadly Chomskyan paradigm, who argue for a dedicated language acquisition device (LAD), that has evolved to serve the precise purpose of language acquisition. This device is primed specifically to receive linguistic input, and requires a minimal amount of it to set to work building mental representations for language in the mind of the child. The fact that the required triggering input is so minimal provides an explanation for the consistency of language acquisition paths across otherwise fairly widely varying life experiences. The most elaborated version of the LAD account— the Principles and Parameters approach—suggests that children are born with a Universal Grammar (UG), which means they are (unconsciously) anticipating those features that are common to all languages (the principles), as well as limited options for those things that differ among languages (the parameters). Upon exposure to actual input from a given language, children are able to ‘decide’ which sort of language they have encountered. So, for example, some languages have basic subject–verb–object organization in which complements are attached to the right of the heads of phrases (thus objects follow verbs, relative clauses follow noun heads, and noun phrases follow prepositions), while other languages are subject–object–verb where the reverse order of complements is found. A child exposed to a language of the first type need only process a simple structure (say one with a verb followed by an object) and it will trigger the expectation that all the other head-complement structures will be in the same order. When all the open parameters have been set, the child possesses the ‘core’ grammar of the specific language he or she is exposed to. At the same time, however, the child has also been acquiring those aspects of the language that are not anticipated by UG, using experience-dependent learning. These aspects are often together referred to as ‘peripheral’ grammar. Some researchers in this paradigm have assumed that all principles and parameters are operational or sensitive to the input from the beginning of life. Others have suggested that some may at least emerge with maturation.

8

In the course of acquisition, generalization (and overgeneralization) allows new knowledge to permeate across-the-board, and ‘bootstrapping’ allows learning in one area of language to trigger new learning in another. Semantic bootstrapping involves understanding an utterance in context and using it in conjunction with innate expectations of language principles to crack the code of the syntax. (For example, a child who does not yet know the required complements of the verb ‘put’, can work these out from understanding utterances such as ‘Put the cup on the table’ in context.) Syntactic bootstrapping involves working from an already understood structure to fill in meanings, semantic information, by deduction. (For example, if you hear ‘John glopped his friend on the head and he fell down’, you may not know exactly what ‘glopped’ means, but you can work out a lot of what it must mean.) The Principles and Parameters (P&P) model has been a highly influential linguistic approach in language acquisition research, even while researchers in psychology and anthropology have been pursuing significantly different lines of investigation. In linguistics, the P&P model lies at the intersection of generative (specifically Chomksyan) linguistic theorizing about the nature of adult mental representations for language and accounts of how children acquire language. It has evolved as an account of how language could actually develop across time, even while its roots are in the ‘logical problem of language acquisition’ because it assumes that what cannot be learned from the input must be genetically prespecified. It advances various arguments in support of the thesis that the input is in fact incapable of providing sufficient information for language to be learned entirely through experience-dependent mechanisms, and thus that there is an essential problem of the ‘poverty of the stimulus’. The account is also strongly modular in the sense that it sees language as being acquired by a specially dedicated mechanism (the LAD), independent of other types of mental representations or mental functioning. Despite its power within linguistics, accounts predicated on some version of the UG story actually attract only a minority of adherents within the broad field of child language research. Most researchers are convinced that language in its entirety can be worked out by the child on the basis of the input, coupled with innate (nonspecific) predispositions for analyzing their environment. As a result of their analysis, children possess the capacity to produce and comprehend language. They may also have stored mental representations for how language works, but, unlike the P&P account, most of the accounts that rely on processing of the input as the way in which acquisition occurs also regard the endpoint as processes for production

ACQUISITION THEORIES and comprehension, without independent mental representations. Ann Peters and Dan Slobin, for example, have argued that children possess strategies, operating principles, for carving up the speech stream into repeating bits, storing those bits, recognizing the commonalities across them, and thereby building up a performance grammar for comprehension and production called Basic Child Grammar, which subsequently becomes elaborated into an adult grammar. Another of the processing theories of language acquisition is Elizabeth Bates’ and Brian Mac Whinney’s Competition Model. Unlike the operating principles approach, this does not argue for a steady state in the form of a grammar at the end of acquisition, but rather for a permanently dynamic response to input throughout life. The impression that language has been definitively learned comes simply from the fact that new input changes the child’s system very little if a child remains in the same speech community (although it will usually change with exposure to a new dialect or to a new language). Acquisition takes place as children respond to the distribution of various cues to meaning (word order, inflectional information, intonation and stress, etc.) and respond probabalistically to conflict among them. Since each child processes the input independently, individual differences between children are expected, and advocates of this approach argue that the differences found among children support the model. As they learn more of their language, they pay attention to more and more of the cues and let the stronger ones win out over the weaker ones for their language. So, at an early stage, a child learning English may assume that nouns at the beginning of sentences are agents, but when they begin to pay attention to passive morphology, they will have to adjust their assumptions accordingly. As should be clear from this example, the ability to derive meanings for language from context, in advance of actually understanding how language is structured is crucial (as it is in semantic bootstrapping). Active application of this and other distributional models, aimed at demonstrating that the input is sufficient to account for language acquisition, is seen in the computer-modeled connectionist networks. Attracted by the architectural similarity between computer networks and neurons, these researchers argue that experience with language teaches the network so effectively and so quickly that it gives the impression of prior knowledge. As already indicated, children frequently seem to rely on pragmatic expectations of what language ought to mean in a given situation as a basis for learning how it is structured. Some theories of language acquisition place an even greater emphasis on the role of the sociopragmatic environment. Jerome Bruner,

for example, argues that the behavior of a child’s caretaker provides all the cues that the inherently social child needs to acquire language. The way in which language is used in the here and now, in conjunction with actions that match what is being said, ‘scaffolds’ the child’s understanding of the language. Others see difficulties with this approach, not only because not all successful language learners receive the kind of careful scaffolding it seems to require but also because the parental complexity of language seems to follow rather than lead increasing complexity in the child’s language. In these and other ways, the account seems too simple to account for the complexity of the task. A similar approach was advocated by Jean Piaget, who saw language development as the logical extension not of social behavior, but of cognitive development. In Piaget’s initial proposal, embedding of sentences was seen as analogous to nesting boxes, and the former dependent on the latter. Although most of his specific predictions have not been supported by subsequent research, it is clear that at least certain aspects of language are intimately related to cognitive development, although the direction of influence is not clear. One area of current research concerns the emergence of the capacity to make informed guesses about what other people know (‘theory of mind’), and its relationship to language development. There is considerable discussion about whether the capacity to embed clauses under main verbs such as ‘think’ and ‘know’ drives or is driven by the capacity to understand that others may not share the same assumptions as oneself. Finally, it is worth noting that significant debate surrounds the issue of whether, when an individual learns a second language, they go about it in the same way as a first language learner, or whether it is a fundamentally different process. Evidence suggests that there is some kind of critical period for second language development, as children seem to be able to do it so much better than adolescents and adults. However, it is not yet known when bilingual first language development should be seen as having given way to second language development. Nor is it clear that adults, when given the kind of input and motivation of children, are always incapable of the same level of success. It is also unclear whether second language learners are able to reaccess the learning capacities they had as children learning their first language, or whether these are permanently overridden and made unavailable by the presence of the first language. References Crain, Stephen, and Diane Lillo-Martin. 1999. An introduction to linguistic theory and language acquisition. Oxford: Blackwell.

9

ACQUISITION THEORIES de Boysson-Bardies, Benedicte. 1999. How language comes to children. Cambridge, MA: MIT Press. Fletcher, Paul, and Brian MacWhinney (eds.) 1997. The handbook of child language. Oxford: Blackwell. Foster-Cohen, Susan. 1999. An introduction to child language development. Harlow, Essex: Longman/Pearson Education. Hoff-Ginsberg, Erika. 2000. Language Development, 2nd edition. Pacific Grove, CA: Brooks/Cole Publishing Company. Jackendoff, Ray. 1995. Patterns in the mind: language and human nature. New York: Basic Books. Karmiloff, Kyra and Annette Karmiloff-Smith. 2001. Pathways to language: from fetus to adolescent. Cambridge, MA: Harvard University Press.

Owens, Robert. 1996. Language development: an introduction, 4th edition. Boston: Allyn and Bacon. Pinker, Steven. 1994. The language instinct: how the mind creates language. New York: William Morrow and Company Inc. Ritchie, William and Tej Bhatia (eds.) 1999. Handbook of child language acquisition. New York: Academic Press. Smith, Neil. 1999. Chomksy: ideas and ideals. Cambridge: Cambridge University Press.

SUSAN H. FOSTER-COHEN See also Acquisition; Developmental Stages; Slobin, Dan Isaac

Aerodynamics of the Vocal Tract Speech production may be viewed as a means of converting slow pressure variations in the vocal tract into the rapid pressure variations that constitute sound. When there is a difference in air pressure across a valve and it opens, sound is created by the rapid movement of air. A stop burst is created by the sudden release of air under pressure, the continuous noise of a fricative is created by a similar but slower release, and repetitive release of air through the glottis by the vibrating vocal cords creates voicing. Speech aerodynamics studies in detail as to how these sounds are created, how the air pressures and airflows are generated, and their phonological consequences. From an aerodynamic point of view, the vocal tract is a system of chambers connected to each other by valves, with some of them having access to the air outside. Most of the chambers are equipped with piston-like structures that can change the volume of their chamber. There are a few general physical principles that govern the generation of pressures and flows in the vocal tract: 1. For a given mass of air, pressure varies inversely with the volume of the chamber containing it (Boyle–Mariotte’s Law). 2. For a given volume of a chamber, pressure varies directly with the mass of air in it. 3. The quantity of airflow (the ‘volume velocity’) through an aperture varies as a function of the size of the aperture and the pressure difference across it. 4. The velocity of airflow (the so-called ‘particle velocity’) varies directly with the volume velocity

10

and inversely with the size of the channel it is flowing through. (This principle is important because turbulence, the source of acoustic noise, is a function of air velocity.) 5. When air is flowing through a channel, the pressure it exerts on the walls of a channel at 90° with respect to the direction of airflow varies inversely with the velocity of the airflow (this is the Bernoulli effect). Four ‘air stream mechanisms’ for generating the pressures needed in speech are used in languages of the world. They are named for the principal piston-like structure creating the volume change, plus ‘egressive’ or ‘ingressive’ to denote whether air moves out of the cavity or into it when the pressure is released to the atmosphere: Pulmonic egressives: The air in the lungs is compressed by the expiratory muscles acting to decrease the volume of the rib cage; the pressure increases (by principle 1). Sound is created when this air is released at the glottis and/or at any supraglottal constriction, e.g. [p], [l], [m], [o]. All languages use pulmonic egressive sounds and the vast majority like English, Japanese, Finnish, and Hawai’ian use them exclusively. Glottalic egressive (‘ejectives’): With air trapped in the oral cavity between a closed glottis and a supraglottal constriction (and, of course, velic closure), the air is compressed by the elevation of the larynx (the glottis acting like a piston) and some pharyngeal constriction. Such sounds may include stops, affricates, and fricatives, e.g. [p’], [ts’], [f’]. Ejectives are found in approximately 18% of the world’s languages, e.g. Amharic, Quechua, and Navaho.

AFFIXATION Glottalic ingressive (‘implosives’): The oral cavity is enlarged (by lowering the larynx and possibly also the jaw) so that the pressure in the oral cavity decreases. Most implosives are voiced, the lowering larynx thus acting like a leaky piston, but voiceless implosives, although quite rare, do exist. Voiced implosives occur in about 10% of the world’s languages, e.g. Kalabari, Sindhi, and Xhosa. Velaric ingressive (‘clicks’): Air is trapped between the tongue and the palate (or the lips) and the small cavity is enlarged by the tongue moving down and/or backwards. The air pressure is greatly lowered and when released by the tongue or lips, producing a ‘pop’ or fricative sound. The sound that comic strips try to convey with ‘tsk’! or ‘tut-tut’ (a sound expressing disapproval or regret) is the alveolar click [!]. The sound of a ‘kiss’ is the bilabial click [>]. Clicks are found widely in different cultures, but usually as interjections or calls to animals. As speech sounds, they are quite rare and are found only in certain Southern African languages such as Zulu, Xhosa, Ndebele, and !Xóõ. Given that clicks’ sound-making structures are so localized in one part of the vocal tract, clicks can be combined simultaneously with sounds made by the other air stream mechanisms, e.g. they can be accompanied by voicing or voicelessness, nasalization, etc. Voicing and trills: Trills and voicing constitute aerodynamically driven, self-sustaining oscillations. The cycle of vocal cord vibration could start with the vocal cords lightly adducted in the presence of airflow. This causes the subglottal pressure to build up behind the glottal closure, the increased pressure forces the cords apart, and the resulting high rate of airflow through the cords creates a negative pressure between them (by principle 5, above), which ‘sucks’ the cords together again (along with an elastic recoil force of the vocal cord tissue); the cords close again, and the cycle repeats itself. Trills are created in a similar way. Obstruent and voicing: Obstruents inherently favor voicelessness for the following reason: voicing

requires that there be air flowing between the vocal cords. Obstruents partially or completely block the airflow exiting the oral cavity. But if glottal airflow continues, then the pressure in the oral cavity will rise (by principle 2, above). Then, the difference in the pressure across the glottis will be reduced and may fall below the value needed to maintain the airflow required for vocal cord vibration (by principles 3 and 5, above). Passive yielding of the vocal tract walls to the impinging pressure permits voicing to persist for around 65 msec (more for labial stops, which have more surface area behind the constriction and less for velars). For voicing to be maintained beyond this, some active expansion of the oral cavity is required, e.g. by lowering the larynx and/or the jaw. This leads to many phonological consequences, among them the greater incidence of ‘missing [•]’ in the stop inventories of languages utilizing a voicing contrast on stops, e.g. as in Dutch, Czech, and Thai, the greater tendency for geminate stops to be voiceless or to undergo devoicing morphophonemically. Many other common sound patterns in language, e.g. the development of affricates from stops, as well as the way specific speech sounds are modified in certain cases of speech disorders, e.g. cleft palate, laryngectomy, can be explained by reference to speech aerodynamics. References Ladefoged, P., and I. Maddieson. 1996. The sounds of the world’s languages. Oxford: Blackwell. Ohala, J.J. 1983. The origin of sound patterns in vocal tract constraints. The production of speech, ed. by P.F. MacNeilage, 189–216. New York: Springer-Verlag. Ohala, J. J. 1997. Aerodynamics of phonology. Proceedings of the 4th Seoul International Conference on Linguistics [SICOL], 92–7. 11–15 August 1997.

JOHN J. OHALA See also Speech Production

Affixation Affixation is a morphological process that adds phonological material to a word in order to change its meaning, syntactic properties, or both. Some examples of affixation in English are given in (1).

(1) a. fond

fondness

b. start

restart

c. car

cars

11

AFFIXATION In this article, an overview of different types of affixation processes is given, followed by a discussion of how to distinguish affixation from other processes. Then, a number of theoretical problems related to affixation are presented.

Types of Affixation

(3) steen (‘stone, rock’)

The phonological material added in affixation is called the affix. The affix is attached to the base. Different types of affixation can be distinguished on the basis of the position of the affix with respect to the base or on the basis of how affixation affects the meaning of the base. In this section, the emphasis is on the former. The latter is discussed in the section on delimitation issues. Suffixation is the most common type of affixation. In suffixation, the affix is added to the end of the base. Examples are (1a) and (1c) above. In (1a), the suffix-ness is added to the adjective fond to produce the noun fondness. In (1c), the suffix -s is added to the noun car to produce the plural of the noun. In most languages, suffixation is the most widespread form of affixation. In languages such as Turkish and Finnish, it is the only type of affixation. In many respects, prefixation is the mirror image of suffixation. In prefixation, the affix is attached to the beginning of the base. An example is (1b) above, where the prefix re- is added to the base start. Prefixation is less widespread than suffixation, but some languages, e.g. Khmer, only have prefixation. In languages that have both suffixation and prefixation, the former usually has a larger range of functions. Thus, in English, all inflection is expressed by suffixation. In word formation, we find both, but prefixation only rarely changes the syntactic category of the base, as when the verb enrich is formed from the adjective rich. In most languages, affixation only involves suffixes and prefixes. Infixation, in which the affix is attached inside the base, is much rarer. An example is the infix -um- in Tagalog, illustrated in (2). (2) a. bilih (‘buy’) b. gradwet (‘graduate’)

bumilih (‘bought’) grumadwet (‘graduated’)

The infix in (2) is used to express the past tense of the verb. It attaches itself after the first consonant or cluster of consonants of the base. This is its anchor point. In general, the anchor point of an infix is at most one syllable removed from the left or right boundary of the base. Whether other types of affixation exist depends on the theoretical framework adopted. Thus, circumfixation is the simultaneous addition of a prefix and a suffix. The mere observation of a form such as enrich-

12

ment is not sufficient evidence for the existence of a circumfix in English, because the form can be analyzed transparently as a result of suffixation of -ment to the verbal base enrich, which is in turn the result of prefixation of en- to the adjective rich. A better candidate for a circumfix is Dutch ge…te, exemplified in (3). gesteente (‘type of stone/rock’)

The change in meaning in (3) can be expressed as a generalization to a type. It can be argued that ge…te is a circumfix, because neither *gesteen nor *steente are possible Dutch words. A theory that excludes circumfixes would have to postulate one of these nonexisting forms as an intermediate form in the derivation.

Delimitation Issues As affixation is the most salient morphological process, it is no surprise that many of the foundational discussions in morphology are directly related to affixation. A first set of issues concerns the function of affixation. Traditionally, inflection and derivation are distinguished. Inflection adapts a word to its syntactic environment. A prototypical example is the agreement of a verb with its subject in person and number, as illustrated for sleep in (4). (4) a. The child sleeps. b. The children sleep. Derivation is a type of word formation that involves affixation. It forms new lexemes, marked by a different syntactic category and/or a different lexical meaning, as illustrated in (1a-b). Another contrast is the one between inflection and stem formation. An example of the latter can be observed in Dutch kinderstoel (‘highchair’), a compound of kind (‘child’) and stoel (‘chair’). The form kinder is different from the singular ( kind) and the plural ( kinderen). The suffix -er can be said to produce a secondary stem used in compounds. Another set of delimitation issues concerns the nature of the affix as a phonologically dependent, morphological unit. Affixes contrast on the one hand with clitics, and on the other with stems used in compounding. Clitics are phonologically dependent particles, i.e. particles that must attach to a host word much like an affix, whose distribution is, however, determined by syntactic rules rather than morphological rules. An example is the Italian clitic pronoun -ti (‘you’) in incontrarti (‘to meet you’). While these clitics are generally not considered as affixes, Turkish

AFFIXATION case and number endings as in (5) are more of a borderline case. (5) Ankara ve I˙zmire gideceg˘im ‘Ankara and Izmir I go’, i.e. I go to Ankara and Izmir In (5), Ankara does not have any case ending, whereas I˙zmir is followed by the dative ending -e. Nevertheless, the dative ending applies to the coordinated noun phrase Ankara ve I˙zmir. Therefore, it can be argued that the dative marker is not an inflectional affix but a clitic. Similar observations can be made about the genitive marker in the queen of England’s hat. Compounding usually combines items that occur as independent words, e.g. bookshelf. Words such as philanthropic and anthropomorphic share many properties with compounds, but their constituent parts do not occur as independent words (in English). They are often called neoclassical compounds. In fact, anthropo is the Ancient Greek word for human being, a sense it also has in English words. Although it is not a word in English, it does not behave like an affix either. It has a stem-like meaning and can appear in different positions in a compound. Other elements of the same general type, e.g. pseudo, occur almost exclusively as a prefix, often with a reduced meaning more typical of an affix than of a stem, e.g. pseudocultured. It is very difficult to draw a clear borderline between stems and affixes among such neoclassical elements. The term affixoid is sometimes used to refer to items for which it is difficult to determine whether they are stems or affixes. An example from outside neoclassical word formation is German reich (‘rich’) in ertragreich (‘productive, profitable’). In English, -ful in successful is similar.

Affix or Process? One of the central questions in morphology is whether affixation should be seen as a rule applying to two elements, a base and an affix, or as a process applying to a base. Charles Hockett (1954) calls the first position Item & Arrangement, and the second Item & Process. Although all morphological processes can be described in either framework, some decriptions are more natural than others. In a strict Item & Arrangement position, such as defended by Rochelle Lieber (1992), stems and affixes have separate entries in the mental lexicon. The suffix -ness is described as a noun that requires an adjective to its left. Given an appropriate concept of headedness (cf. below), this explains why fondness is a noun. The Item & Arrangement approach works particularly well as long as the form of the resulting word is the concatenation of the forms of the stem and the affix, as in fondness. Special provisions have to be

made for cases where phonological processes interact with the concatenation, as in (6–8). (6) a. b. (7) a. b. (8) a. b.

active intend a house to read live extend

activity intention to house asylum seekers a good read life extent

In (6), suffixation triggers a phonological change of the base, stress shift in (6a), and a change of the final consonant in (6b). In an Item & Arrangement account, one might say that these phonological changes are triggered by the affix. In (7), there is no affix, but the relationship between the nouns and verbs is very similar to what we find in cases of genuine affixation such as encase and entertainment. In an Item & Arrangement account, one might either say that (7) is not affixation but something else (e.g. conversion), or that (7) involves zero affixes, i.e. affixes that do not have a phonological realization. The examples in (8) show that phonological changes can occur without an affix. This can be interpreted as phonological rules triggered by a zero affix or by conversion. The general approach in Item & Arrangement is that concatenation of base and affix is taken as central. Cases such as (6–8) are treated as exceptions to be fitted in. In an Item & Process approach, as argued for by Stephen Anderson (1992), the process of affixation takes priority over the affix. In (7), the word formation process changes the syntactic properties of the input without affecting its form. In (8), the process modifies the form of the input by changing the quality of the final consonant and, in (8a), of the vowel. In (6), the process affects the form of the input by changing the stress position or the voicedness of the final consonant and appending some further phonological material, the suffix. Affixation of the type illustrated by fondness, the prototypical case in Item & Arrangement, is a special case in Item & Process where there is a perfect match between the input and a part of the output of the affixation process. It is difficult to find conclusive arguments for either Item & Arrangement or Item & Process approaches to affixation. Proponents of Item & Arrangement usually claim that their approach is theoretically more restrictive than Item & Process. Proponents of Item & Process typically argue that processes such as infixation, reduplication, vowel change, consonant change, and conversion are too frequent, especially in non-European languages, to treat them as exceptional compared to ‘pure’ affixation. Linguistic data alone cannot determine as to which of the two approaches is correct, because every Item & Arrangement account can be reformulated into Item & Process and vice versa.

13

AFFIXATION

Headedness The concept of head of a word only makes sense in an Item & Arrangement-based account of morphology. The head determines the syntactic category of the resulting word as well as other properties such as gender in languages that distinguish them. Thus, fondness is a noun because -ness is the head, and French activité (‘activity’) is a feminine noun because -ité is the head. Different methods have been proposed to determine the head. Anna Maria Di Sciullo and Edwin Williams (1987) propose that the head is always the rightmost element. This concurs with the observation that suffixes often change the syntactic category of their base, but prefixes, as in (1b), usually do not. Exceptions include cases such as enrich and debug, where the prefixes make verbs out of an adjective and a noun. An alternative proposed by Lieber (1990) is that the last affix to attach is the head. This generalization makes re- the head in (1b) and accounts correctly for enrich and debug, but has problems with prefixes such as counter- in (9). (9) a. intuitive b. example c. sign

counterintuitive counterexample countersign

In (9), counter- is the only (hence last) affix attached, but the syntactic category depends on the word it attaches to. The Spanish diminutive, illustrated in (10), is problematic for both. (10) a. el pintor el pintorsito b. la cancyón la cancyonsita

‘the painter’ ‘the little painter’ ‘the song’ ‘the little song’

The final -o and -a in (10) are not a part of the affix, but so-called word markers, whose form is determined in this case by the grammatical gender of the word (-o for masculine, -a for feminine). Although the suffix -sit is the rightmost element and the only affix, it is the base that determines the gender as indicated by the article. In an Item & Process approach, there is no need for a head, because its function is subsumed in the affixation process. References Anderson, Stephen R. 1992. A-morphous morphology. Cambridge: Cambridge University Press. Aronoff, Mark H. 1994. Morphology by itself: stems and inflectional classes. Cambridge, MA: MIT Press. Booij, Geert, Christian Lehmann, and Joachim Mugdan (eds.) 2000. Morphologie—Morphology: Ein Internationales Handbuch zur Flexion und Wortbildung—An international handbook on inflection and word-formation, Vol. 1. Berlin: Walter de Gruyter. Di Sciullo, Anna Maria, and Edwin Williams. 1987. On the definition of word. Cambridge, MA: MIT Press. ten Hacken, Pius. 1994. Defining morphology: a principled approach to determining the boundaries of compounding, derivation, and inflection. Hildesheim: Olms. Harris, James W. 1991. The exponence of gender in Spanish. Linguistic Inquiry 22. 27–62. Hockett, Charles F. 1954. Two models of grammatical description. Word 10. 210–31. Lieber, Rochelle. 1990. On the organization of the lexicon. New York: Garland. Lieber, Rochelle. 1992. Deconstructing morphology: word formation in syntactic theory. Chicago: University of Chicago Press.

PIUS TEN HACKEN See also Compounding; Inflection and Derivation; Word

African American Vernacular English African American Vernacular English (AAVE) has been known by various names in linguistics, including Negro Nonstandard English, Black English Vernacular, and more recently Ebonics. Except for the last, each name corresponds to a particular time in American history when it was more fashionable to refer to African Americans as Negroes or Blacks, among other names.

14

The terms have been used to designate nonstandard English varieties spoken by less educated African Americans other than those in coastal Georgia and South Carolina, who speak Gullah, a creole. The average speakers of both vernaculars do not have specific names other than English for the ways they speak; scholars have coined the names to distinguish their

AFRICAN AMERICAN VERNACULAR ENGLISH ethnolects (ethnic dialects) from standard English and from European-American vernaculars. Ebonics, which was first used in print in the work of Robert Williams (1975), is probably the most confusing of the terms. It subsumes Gullah and thus corresponds to what may be identified broadly as African American English. However, in the way Williams defines it, it also refers to other language varieties of the black diaspora, including English and non-English creoles of the Atlantic (e.g. Jamaican and Haitian creoles and Nigerian Pidgin English). The name becomes quite useless when it is also extended to cover NigerCongo languages. The average AAVE speaker might be puzzled as to why their English is lumped with a multitude of other languages they hardly understand, if at all, and why the fact that it is first of all an American linguistic phenomenon is virtually overlooked. On the basis of literary sources, the predecessors of AAVE must have become distinct from other American English vernaculars already in the early eighteenth century (Brasch 1981). Literary representations of the time are similar to those about the Caribbean, where larger plantations with African majority populations were already booming. It is not clear whether the literary representations (mostly by British, rather than American, authors) reflect the speech of African Americans of the time or, instead, generalized stereotypes that were based on what was developing in the Caribbean. Newspaper advertisements about runaway slaves reveal that slaves born in the colonies (a large proportion by the early eighteenth century) or who had lived there for a long time did not speak differently from the European indentured servants with whom they interacted on a regular basis, before the institutionalization of segregation in tobacco and cotton plantation ex-colonies in the late nineteenth century. This history suggests that, unlike Gullah, AAVE could not have developed as a distinct variety before the late nineteenth century. Aside from letter-writing evidence, more reliable literary representations of AAVE probably also date from the nineteenth century, when most of the authors were also American born. AAVE has become central to research in American quantitative sociolinguistics and historical dialectology since the 1960s. William Labov, William Stewart, and J.L. Dillard, together with Walt Wolfram and Ralph Fasold, focused on subsets of the following questions: (1) Is AAVE a rule-governed vernacular, and is it a full-fledged language? (2) Is it an American English dialect or is it, instead, an erstwhile creole (similar to those of the Caribbean), and therefore a separate language, which has acquired English-like features by ‘decreolization’? (Putatively, this evolution would be caused by approximations by African

Americans of standard English structural features, supposedly because of the mass media, mass education, and socioeconomic mobility since the abolition of slavery.) (3) How different is AAVE’s system from those of standard English and European American vernaculars such as Appalachian and Ozark English? (4) How does variation work in its system and what is its significance from an evolutionary point of view? Overall, a very large proportion of the research has been devoted to a small subset of structural features that makes AAVE peculiar, at least by its statistical distribution. This includes the following: 1. AAVE allows the loss of consonant clusters at the end of words, and this means that the final consonant is often dropped in words such as desk, passed, and old (pronounced des’, pass’, and ol’). 2. Diphthongs (double vowel sounds) such as [aw] and [ai], as in south side, become long monophthongs (simple vowels): [a:]. 3. The ‘r’ is dropped in front of consonants and at the word end, as in Lord and car (pronounced Lo’d, ca’); the ‘r’ is pronounced before vowels, as in the words arrive and road. 4. The copula ‘be’ is often absent in sentences such as She cute and He gon come. 5. Be is used to denote invariant or repeated states of affairs, as in He be tired/reading every time I visit him. 6. AAVE uses multiple negation, as in He ain go nowhere (nohow). 7. The negative auxiliary (e.g. didn’t) is moved to the front of the sentence, as in Didn’t nobody tell me ‘hush yoh mouth’, ‘Nobody told me, ‘hush your mouth’.’ 8. AAVE differs from standard English in various aspects of time reference, as in I been there, ‘I have been there’, and They BEEN married (with emphasis on been), ‘They have been married for a long time and still are’, and He done lef when I come, ‘He had already left when I came’. 9. Plural is expressed with dem (in alternation with the presence or absence of the standard English plural marker -s), as in dem boys, and an dem signals an ‘associative plural’, as in Yolanda an dem, ‘Yolanda and her friends/relatives’. 10. Many words and word meanings are peculiar to African American speech, as in You are bad (with bad pronounced emphatically and with an extra-long vowel), ‘you are impressive/ something else’, and you look clean, ‘you look sharp(ly dressed)’. Interest is now growing in describing not only isolated features but also integrated subsystems of the

15

AFRICAN AMERICAN VERNACULAR ENGLISH vernacular, highlighting, on the one hand, similarities with and differences from other American English varieties and, on the other, how subsystems such as time reference or the noun phrase are internally organized. This is a corollary of the fact that today, studies of AAVE’s structure are no longer almost exclusively in the quantitative sociolinguistics paradigm, but also in theoretical linguistics. There has also been research in African American discourse differences, especially on discourse genres, the structure and contents of some narratives, the form of language used in them, and whether or not the meanings of particular utterances are transparent. Particularly outstanding in this subarea are playing the dozens, in which male participants exchange witty, fictional insults directed especially at female relatives and tall tales or toasts, rhyming epics in which street culture and its heroes are celebrated in a street verbal style.

Dillard, J. L. 1972. Black English: its history and usage in the United States. New York: Vintage Books. Labov, William. 1972. Language in the inner city: studies in the Black English vernacular. Philadelphia: University of Pennsylvania Press. Mufwene, Salikoko S., et al. (ed.) 1998. African-American English: structure, history, and use. London: Routledge. Poplack, Shana (ed.) 1999. The English history of African American English. Malden, MA: Blackwell. Poplack, Shana, and Sali Tagliamonte. 2001. African American English in the diaspora: tense and aspect. Oxford: Blackwell. Rickford, John R. 1999. African American vernacular English. Malden, MA: Blackwell. Rickford, John Russell, and Russell John Rickford. 2000. Spoken soul: the story of Black English. New York: John Wiley & Sons, Inc. Smitherman, Geneva. 2000. Talkin that talk: language, culture and education in African America. London and New York: Routledge. Williams, Robert L. (ed.) 1975. Ebonics: the true language of Black folks. St. Louis, MO: Robert L. Williams & Associates, Inc.

SALIKOKO MUFWENE

References Brasch, Walter M. 1981. Black English and the mass media. Lanham, MD: University Press of America.

See also Caribbean; Ethnicity and Language; Gullah

Afroasiatic The Afroasiatic (formerly called Hamito-Semitic, also Semito-Hamitic, more recently Afrasian) phylum unites five or six language families that are clearly related, but the relations among the families are far from clear. The confusion—and the profusion of names—result from the history of the study of these languages. Known from antiquity, of course, were three of the principal Semitic languages: Hebrew, Arabic, and Aramaic in its Christian (Syriac) and Jewish literary forms. The Ethiopic literary language Ge‘ez became known to European scholars during the Renaissance and was immediately recognized as a close relative of the other three. Beginning in the mid-eighteenth century, a number of ancient languages that survived only in inscriptions were deciphered, starting with Palmyrene, a close relative of Syriac, and Phoenician, which is close to Hebrew. In 1781, the name ‘Semitic’ was suggested for these languages as a family, on the basis of the ‘Table of Nations’ in Genesis 10:21–4, where Asshur, Aram, and Eber are among the descendants of Shem. The ninteenth century brought two supremely important decipherments, those of Egyptian (in the

16

early 1820s) and Akkadian (in its two main varieties, Assyrian and Babylonian, in the late 1840s). Certain peculiarities of the Semitic language structure, when noticed at an early stage in the decipherment of Akkadian by Edward Hincks, proved to be very helpful in making further progress, and there was soon no doubt that another major Semitic language had been discovered. With Egyptian, although Jean-François Champollion already suggested a Hebrew comparison in his initial announcement of the decipherment, and Hincks used Egyptian loanwords in the Hebrew Bible in his analyses of Egyptian, the resemblances were less obvious, and the actual kinship was suggested by the pioneering Egyptologist and linguist Richard Lepsius by 1880. Also during the mid-nineteenth century, explorers, notably Lepsius, traveling through Abyssinia (modern Ethiopia, biblical Cush) gathered information on the languages they encountered there. Some (notably Amharic, the language of the ruling dynasty) proved to be Semitic languages, others (notably Oromo, which was then called by the insulting name Galla, and

AFROASIATIC Somali) resembled each other fairly closely and Semitic only remotely; this group was soon dubbed Cushitic. Only in the past generation has it been suggested that ‘West Cushitic’ is actually a separate branch of Afroasiatic (renamed Omotic), perhaps very distant from Cushitic; this was widely accepted for some time, but opinion may be shifting back to the earlier view. (The principal Omotic language is Wolaytta of Ethiopia, with some 2 million speakers, but very little is known about most of its relatives.) As Africa was partitioned among various colonial powers, naturally the investigation of Africa’s indigenous languages was carried out largely by scholars from the respective metropolitan nations. Thus, much of the initial work on Cushitic was done by German and Italian linguists, while French investigators provided most of the descriptions of Berber tongues from North Africa (Morocco to Libya). The last of the five main branches of Afroasiatic to receive attention is called Chadic, from its location near Lake Chad; the principal language of Nigeria, Hausa, is its main representative, and what little investigation was done was largely carried out by British linguists. The large number of Chadic-speaking communities, their small populations, and their remote locations, mostly in northern Nigeria, are factors that have made Chadic the least known of these groups overall. Returning to the ‘Table of Nations’, we find that Ham is listed as the father of Cush and Egypt, so not unreasonably these languages that were clearly, although distantly, related to Semitic came to be called ‘Hamitic’. But this proved to be a suboptimal choice for two reasons: linguistically, it implies that all the other families form a unit as against Semitic (implying that the Semitic family was the first to branch off from the others), which is not the case; and ethnographically or anthropologically, it caused the classification of languages to become enmeshed with the spurious notion of a ‘Hamitic race’ of dark-skinned but European-featured Africans. (In fact, of course, there is no necessary connection between language family relationship and ‘race’ or ethnic group.) For this reason, the name ‘Afroasiatic’ is now widely used for the phylum, since it simply names the two continents where it is found. This can be contracted to the convenient ‘Afrasian’. What, then, are the relationships among the branches of Afroasiatic? The very nature of the data makes it difficult to achieve consensus. It is as if scholars of Indo-European were faced with only Gothic and the other Germanic languages, Greek, modern Romance and Celtic, Armenian, and the modern Indic languages—analogous, respectively, to Akkadian and the other Semitic languages (an ancient tongue not ances-

tral to any of the modern ones), Egyptian (a single language attested over thousands of years), Cushitic and Omotic (whose common ancestry is disputed), Berber (a close-knit group of dialects with some ancient materials), and Chadic (a vastly differentiated family with a widely used, politically dominant variety). (The IndoEuropean analogy can also remind us that we have no way of knowing how many languages or entire branches of Afroasiatic have vanished without a trace—what ‘Avestan’ or ‘Umbrian’, what ‘Tajik’ or ‘Albanian’ there might have been.) Thus, it is no wonder that there are almost as many views of Afroasiatic relationships as there are Afroasiaticists. One view divides the northern groups—Berber, Egyptian, Semitic—from the southern ones—Chadic, Cushitic, Omotic. Another groups Semitic and Cushitic, plus Berber, against Egyptian and Chadic, plus Omotic. Some characteristics that are widely shared among Afroasiatic languages are as follows. Consonants tend to be fairly numerous and vowels very few. Stops and sibilants appear in three series, voiced (b d g, z), voiceless (p t k, s), and ‘emphatic’, customarily marked with a dot (t., .s). The ‘emphasis’ appears differently in different languages: in Semitic, it is mostly represented by velarization or pharyngealization (where constriction at the back of the mouth colors the consonant and adjacent vowels), but elsewhere it appears as glottalized or even as implosive varieties of these sounds. There are also a fairly large number of consonants made in the back of the mouth (including the notorious Arabic ‘ayn /ʕ/). Another peculiarity is a tendency to laterals (sounds produced with air flowing past the sides of the tongue) and interdentals (with the tongue tip between the front teeth as in English th []). Conversely, Afroasiatic languages tend to have the near-minimal vowel set /a i u/ (both long and short). The best evidence that the Afroasiatic languages are related (rather than reflecting much borrowing of vocabulary and grammatical features) is the shared inflectional paradigms: the personal prefixes and suffixes self-evidently have a common origin throughout. The elements used to derive one verb from another also appear throughout the phylum (Table 1). It is noteworthy that both the earliest and the latest comparative grammars of Semitic, those of Heinrich Zimmern (1898) and Burkhart Kienast (2001), systematically include Afroasiatic comparisons. (Meanwhile, Carl Brockelmann, who compiled the definitive collection of comparative grammatical information on Semitic as then known (1908–1913), apologizes in the preface for excluding ‘Hamitic’ materials, as the data were inadequate at the time. Other writers on Semitic have followed his lead.) The well-known ‘triconsonantal root’ structure of the Semitic languages, however, does not appear to be

17

AFROASIATIC TABLE 1

Some Afroasiatic Comparisons Common Semitic

Ancient Egyptian

Tuareg (Berber)

Afar (Cushitic)

Hausa (Chadic)

ø ttytnt-…-n t-…-n y-…-n y-…-n

náa káa kín yáa táa mún kún kún sún sún

ø -t -t ø -t -tVn -tVn -tVn -Vn -Vn

níi kái kée -šíi ’ítá múu kúu kúu súu súu

s m t

— — —

Prefix-conjugation markers (Hausa: aspect pronouns) 1 sing. 2 m. sing. 2 f. sing. 3 m. sing 3 f. sing 1 pl. 2 m. pl. 2 f. pl. 3 m. pl. 3 f. pl.

’atata-…¯¹ yatanata-…u¯ ta-…a¯ ya-…u¯ ya-…a¯

— — — — — — — — — —

ø-…g´ t-…-d t-…-d ytnt-…-m t-…-mt -n -nt

Suffix-conjugation or pronominal affixes 1 sing. 2 m. sing. 2 f. sing. 3 m. sing. 3 f. sing. 1 pl. 2 m. pl. 2 f. pl. 3 m. pl. 3 f. pl.

-ku¯ -ta¯ -t ¯¹ ø -at -nu¯ -tunu¯ -tina¯ -u¯ -a¯

-w’¹ -k -J -f -s′ -n -Jn -Jn -s′n -s′n

-i -k -m -s -s -näg´ -wän -kmät -sän -snät Verb derivation markers

causative passive reciprocal

š n t

— — —

shared throughout Afroasiatic; elsewhere in the phylum, ‘roots’ tend to have just two consonants. The only attempts at description of the grammar of the hypothetical ancestral language ‘ProtoAfroasiatic’ are the two small volumes by I. M. Diakonoff (1965, 1988), and he changed his mind considerably between the two. A summary, quite close to the earlier version, may be found in his article in the Encyclopædia Britannica (1974). Diakonoff must be considered quite brave for even attempting to reconstruct a proto-language, given the nature of the available materials: contrast the situation in Indo-European (the real situation, not the truncated version offered above as a parallel to that in Afroasiatic). Proto-IndoEuropean can be reconstructed in considerable detail (and with considerable agreement among scholars) because the hypothetical community that would have spoken a language similar to what is reconstructed thrived only a couple of thousand years or so before the earliest available evidence (Hittite, Vedic, Avestan, Mycenean). We can even verify the accuracy of comparative and reconstructive methodology over such a time span because we have the precious case of the Romance languages—reconstructed Proto-Romance is pleasingly similar to Vulgar Latin as it is known

18

s t m

from the first centuries AD. Proto-Semitic may be considered quite reliable for the same reasons—the large amount of data and the small time depth. Contrast the situation in Afroasiatic: one branch (Semitic) wellknown, one branch (Egyptian) known in considerable detail over a long time—but with no vowels known, and one branch (Berber) with a scattering of meager inscriptions from Roman times. Thus, four of the six branches are known only or primarily from modern data—and the time depth of Chadic alone is estimated as about the same as that of Indo-European. It is no wonder, then, that among the three presentations of comparative Afroasiatic vocabulary—those of Marcel Cohen (1947; omitting Chadic, with an occasional Hausa comparison), Christopher Ehret (1995; omitting Berber), and Vladimir Orel and Olga Stolbova (1995)—there is precious little agreement. (Nor has the preliminary requirement of comparative dictionaries of each of the constituent families been accomplished, even for Semitic.) Part of the difficulty is the shortness of words in the languages, with their roots consisting of just two consonants. Thus, it is all the more difficult to proceed to the other main consideration regarding an ancestral language: the homeland of its speakers.

AGE AND LANGUAGE There are three ways of determining where a language family may have originated. One, which is hypothetical, suggests that the area with the greatest concentration of diverse, related languages is likely to be where the family originated. (It is not likely that several related language groups would all happen to move to the same locality!) For Afroasiatic, this points to somewhere in northeast Africa, near where Chadic and Cushitic (and Omotic?) are found, say Ethiopia/Sudan. Further support for this suggestion might be found in the evidence that the Egyptians entered the Nile Valley from the south and settled gradually northward. There is simply nothing but speculation as to why and whence successive populations of Semitic-speakers appeared on the scene in southwest Asia. A second technique is to examine the vocabulary that can be traced back to the protolanguage. A dishearteningly limited number of words appear to have survived from so long ago, but they seem to point to a preagricultural, nonmaritime environment. The third source of information is the accidents of discovery owing to archaeology. Terrain, climate, and politics conspire to make expeditions unlikely; but there is reason to suppose that an Afroasiatic homeland is to be sought in regions that are now the eastern Sahara Desert, but were hospitably temperate over 10,000 years ago. An archaeological horizon in the western Arabian peninsula has been suggested as coordinating with ProtoSemitic speakers, but in the absence of written records, no assignment of languages to archaeological sites can be considered certain. What of wider connections? From time to time, Afroasiatic is claimed to be part of a still older, vastly extended superstock called ‘Nostratic’, which is said to incorporate Indo-European, Uralic, Altaic, and various other phyla of Asia. Any perceptible reflection of a Proto-Nostratic must be uninterpretably dim; but a question that seems not to be addressed by Nostraticists is the location of the Proto-Nostraticspeaking community—a location from which all the postulated homelands must be reachable. From this point of view, the suggestion that Afroasiatic is not a

member of Nostratic, but is coordinate with the languages that make up ‘Eurasiatic’, is slightly more tenable. An independent suggestion, made by archaeologists who investigated Indo-European origins, that ‘Afroasiatic’—from which they seem to be aware only of Egyptian and Semitic—originated on the southern coast of the Black Sea, is ipso facto untenable. The careful study of Afroasiatic languages goes back to antiquity. The study of the Afroasiatic phylum is nearly two centuries old. The tasks that remain for researchers are largely the same as those faced by specialists in other languages: intensive fieldwork, especially in Chadic, Cushitic, and Omotic, to record ample information about these languages, and then to bring to bear all the analytic techniques of linguistic theory and comparative method. A difference with Afroasiatic is that it also includes some of the oldest written records on earth, providing a unique cross-section of human linguistic behavior across 5,000 years. References Brockelmann, Carl. 1908–1913. Grundriss der vergleichenden Grammatik der semitischen Sprachen. Berlin: Reuther & Reichard. Cohen, Marcel. 1947. Essai comparatif sur le vocabulaire et la phonétique du chamito-sémitique. Paris: H. Champion. Diakonoff, I. M. 1965. Semito-Hamitic languages. Moscow: Nauka. Diakonoff, I. M. 1988. Afrasian languages. Moscow: Nauka. Ehret, Christopher. 1995. Reconstructing Proto-Afroasiatic (Proto-Afrasian): vowels, tone, consonants, and vocabulary. Berkeley and Los Angeles: University of California Press. Kienast, Burkhart. 2001. Historische semitische Sprachwissenschaft. Wiesbaden: Harrassowitz. Orel, Vladimir, and Olga Stolbova. 1995. Hamito-Semitic etymological dictionary: materials for a reconstruction. Leiden: Brill. Zimmern, Heinrich. 1898. Vergleichende Grammatik der semitischen Sprachen. Berlin: Reuther & Reichard.

PETER T. DANIELS See also Ancient Egyptian; Hausa and Chadic Languages; Semitic Languages; Tuareg and Berber Languages

Age and Language Comprehension and production impose many simultaneous demands on the reader to process information on a number of levels, including syntactic, semantic,

and pragmatic. It has been generally assumed that these simultaneous processing demands are handled by a human’s working memory and that working

19

AGE AND LANGUAGE memory capacity declines with age, affecting language processing. Support for this hypothesis is largely correlational. Older adults are typically found to have smaller working memory spans than young adults, and such span measures are found to correlate with language-processing measures. A particular source of difficulty for older adults is the production and comprehension of complex syntactic structures involving multiple levels of sentence embeddings that impose high processing loads on working memory. The spontaneous use of such constructions declines with advancing age, and their use contributes to the breakdown of reading and listening comprehension. Text processing is particularly vulnerable to the effects of aging; both cross-sectional and longitudinal research shows that older adults’ reading comprehension is impaired, as is their listening comprehension. Further, word and name retrieval is disrupted by aging, contributing to an increase in tip-of-the-tongue difficulty in finding words. It is unclear as to whether working memory affects the immediate processing of words and sentences, or whether working memory exerts its affects on the processes involved with the storage, recall, and recognition of linguistic information. The general slowing of cognitive processes with age may also contribute to older adults’ processing problems by affecting the speed of lexical access and other component processes. Neural inhibitory mechanisms also appear to weaken with age and permit the intrusion of irrelevant thoughts, personal preoccupations, and idiosyncratic associations during language processing. These irrelevant thoughts compete for processing resources, such as working memory capacity, and impair older adults’ comprehension and recall. A controversial issue in the study of aging and language is ‘off-target verbosity’. A minority of older adults have been observed not only to talk a great deal but also to drift from topic to topic, weaving into their conversations many unrelated and irrelevant topics. Off-target verbosity appears to be related to poor performance on tests of the function of particular brain regions, as well as to psychosocial stress, extraversion, limited social support, and decreased social interaction. Whereas off-target verbosity has been cited as providing strong support for inhibitiory deficit theory, others have argued that this speech style is limited to social settings in which older adults construe their task differently than do young adults—as a monolog, responsive to an internal chain of associations. The debates over working memory capacity limitations and inhibitory deficits are central to the general study of gerontology. These core research issues have been supplemented by two emerging topics: research examining the effects of Alzheimer’s disease on language production and comprehension, and research

20

examining how to improve older adults’ comprehension of language. Task demands, lapses in attention, response biases, or other cognitive deficits contribute to the syntactic comprehension problems of adults with dementia. Semantic and lexical processes are particularly disrupted by Alzheimer’s dementia. Many researchers have found that tests of implicit word knowledge reveal few differences between dementing adults and normal older adults, and lexical knowledge is commonly found to be preserved. Other researchers have concluded that the structure of semantic knowledge is destroyed by Alzheimer’s dementia. As a result, performance on verbal-fluency tasks, such as generating examples of categories or words beginning with a specific letter, is impaired, picture and object naming is hindered, and word associations are destroyed. The semantic network of adults with Alzheimer’s dementia may be intact but inaccessible, perhaps due to the general, task-independent slowing of cognitive processes, including lexical access. General slowing may increase progressively with disease severity. A breakdown of inhibitions may also contribute to many of the word retrieval problems experienced by adults with Alzheimer’s disease if they are unable to inhibit inappropriate word associations. A special speech register, sometimes termed ‘elderspeak’, has been described as an accommodation in communicating with older adults, especially those with dementia. This accommodation is characterized as involving a simplified speech register with exaggerated pitch and intonation, simplified grammar, limited vocabulary, and a slow rate of delivery. Many of the characteristics of elderspeak, such as its slow rate, exaggerated prosody, and simplified syntax and vocabulary, resemble the characteristics of other speech registers such as ‘motherese’, or the speech used by adults when communicating with small children. Elderspeak may be evoked by the actual communicative needs of the elderly individual as well as by (culturally based) negative stereotypes of older adults, and it is often viewed by older adults as insulting and patronizing. Older adults, especially those in nursing homes, may adapt to situational demands by becoming more accepting of elderspeak. Some form of elderspeak may enhance communication with older adults, especially those with Alzheimer’s dementia, but little systematic research has been undertaken to assess intervention and training programs. In sum, the study of language and aging has focused on the effects of processing limitations on older adults’ comprehension and production. Nonetheless, it is important not to lose sight of three general points. First, many aspects of language processing do not change with age, particularly lexical access and semantic

AGREEMENT memory, but are vulnerable to the ravages of Alzheimer’s disease. Other aspects of language, particularly syntax, may be more resistent to Alzheimer’s disease, yet susceptible to age-related decline due to working memory limitations or inhibitory deficits. Second, there has been a gradual shift toward the use of sophisticated methodologies such as the study of selfpaced reading times, eye movement patterns, and brain activation patterns to study how aging and dementia affect language processing. Third, the shift toward developing and evaluating practical applications, as seen in the growing body of research on elderspeak, continues. The linkage of basic research on language processing with everyday practicalities has led to the development of consumer standards and guidelines for presenting medical information, and for electronics and telecommunications. References Arbuckle, Tannis, and Deloris Pushkar Gold. 1993. Aging, inhibition, and verbosity. Journal of Gerontology: Psychological Sciences 48. 225–32. Caplan, David, and Gloria Waters. 1999. Verbal working memory and sentence comprehension. Behavioral and Brain Sciences 22. 114–26. Charness, Neil, Denise C. Parks, and Barbara A. Sabel (eds.) 2000. Communication, technology, and aging. Doylestown, Pennsylvania: Springer. Hasher, Lynn, and Rose T. Zacks. 1988. Working memory, comprehension, and aging: a review and a new view. The

psychology of learning and motivation, Vol. 22, ed. by Gordon H. Bower. New York. Academic Press. Kemper, Susan. 1992. Language and aging. The handbook of aging and cognition, ed. by Fergus I. M. Craik and Timothy A. Salthouse. Hillsdale, NJ: Lawrence Erlbaum Associates. Kemper, Susan, and Tamara Harden. 1999. Disentangling what is beneficial about elderspeak from what is not. Psychology and Aging 14. 656–70. Kemper, Susan, and Reinhold Kliegl (eds.) 1999. Constraints on language: aging, grammar, and memory. Norwell, MA: Kluwer Academic. Light, Leah L., and Deborah M. Burke (eds.) 1988. Language, memory, and aging. New York: Cambridge University Press. Park, Denise C., Roger W. Morrell, and Kim Shifren (eds.) 1999. Processing of medical information in aging patients: cognitive and human factors perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Ryan, Ellen B., Howard Giles, Giampiero Bartolucci, and Karen Henwood. 1986. Psycholinguistic and social psychological components of communication by and with the elderly. Language and Communication 6. 144–66. Snowdon, David A., Susan J. Kemper, James A. Mortimer, Lydia. H. Greiner, David. R. Wekstein, and William R. Markesbery. 1986. Cognitive ability in early life and cognitive function and Alzheimer’s disease in late life: findings from the Nun study. Journal of the American Medical Association 275. 103–13. Wingfield, Arthur, and Elizabeth A. L. Stine-Morrow. 2000. Language and speech. The Handbook of Aging and Cognition, 2nd edition, ed. by Fergus I. M. Craik and Timothy A. Salthouse. Mahwah, NJ: Lawrence Erlbaum Associates.

SUSAN KEMPER See also Aphasia; Psycholinguistics; Working Memory

Agreement Agreement is a means by which languages signal the presence of a grammatical relation—usually the subject of a sentence, but sometimes also the object and/or the indirect object. Typically, it manifests itself as a prefix or a suffix on the verb, as with the English thirdperson singular marker -s (e.g. John lives in Maryland). Some languages have so-called rich agreement—that is, separate forms for nearly every person/number combination. French represents such a case: (1)

Rich agreement (French) 1SG 2SG 3SG

je parl-e tu parl-es il parl-e

‘I speak’ ‘You speak’ ‘He speaks’

1PL 2PL 3PL

nous parl-ons vous parl-ez ils parl-ent

‘We speak’ ‘You speak’ ‘They speak’

The features most often associated with agreement are person, number, and gender. Given the logistics of speech, there are at least three persons: those referring to the speaker (first), the hearer (second), and everything else (third). Some languages have a fourth person—obviative—reserved for entities further removed from the speaker/hearer’s point of view. Inuktitut (Eskimo) and Algonquian are known for this. Number usually breaks down into singular and plural, but ‘dual’ represents another possibility. English recognizes two numbers in its pronominal system, but

21

AGREEMENT as far as agreement is concerned, only third-person subjects are marked: cf. He understands but We understand. In addition to person and number, agreement may also involve the gender features of a subject and/or objects. French nouns are divided into two grammatical genders—masculine and feminine. When noun phrases headed by these nouns appear in subject position, agreement can be discerned on predicate adjectives and particles (M = masculine, F = feminine): (2) Gender agreement (French) a. Le soleil est monté b. La lune est montée

‘The sun (M) has risen’ ‘The moon (F) has risen’ [PARTICIPLE]

c. Le soleil est brillant ‘The sun (M) is bright’ d. La lune est brillante ‘The moon (F) is bright’ [ADJECTIVE] Feminine forms are marked with an extra vowel that—although not pronounced today—were presumably audible at an earlier stage of the language. Of course, grammatical gender is not to be confused with biological gender; the division of nouns into gender classes should be considered as an aid to reference and/or acquisition (German has three genders, for example).

Agreement and Case As a grammatical category, agreement appears to function like case-markers—special morphemes that attach to noun-phrases (NPs) (rather than verbs) that signal their grammatical function within the sentence (e.g. as subject, object, etc.). In Japanese, for instance, NPs marked with [-ga] are interpreted as subjects, while those marked with [-o] are understood as objects (TOP = topic; COMP = complementizer): (3)

Case-marking and grammatical relations (Japanese) Ichiro-wa Taro-ga hirugohan-o Ichiro-TOP [Taro-SUBJ lunch-OBJ tabeta-to iita ate]-COMP said ‘Ichiro said that Taro ate lunch’

22

In traditional grammar, the case of the subject is called nominative, while that of the direct object is accusative. Since only subjects are associated with agreement in English, this may be taken as a sign of nominative case. Objects are not overtly marked with accusative case, but their grammatical function can be determined by their position following the verb. Another way in which agreement appears functionally similar to case is through its interaction with transitivity. In languages like Japanese, transitive and intransitive subjects take the same case-marker, and in opposition to objects. This is known as a nominative/accusative case-marking system. English is essentially the same with respect to agreement—that is, it pertains to both types of subject. In some languages, however, intransitive subjects and transitive objects pattern in the same manner, in contrast to transitive subjects. Dyirbal (Australian) has a case-marking system of this type, known as ergative/absolutive (transitive subjects are marked with ergative case). Languages that employ agreement, rather than case to signal grammatical functions can exhibit the same pattern, as in Mam (Mayan) below (ABS = absolutive; ASP = aspect; ERG = ergative): (4)

Ergative/absolutive agreement (Mam) a. ma chiin-x-a ASP 1SG/ABS-go ‘I went’ [INTRANSITIVE] b. ma chin-ok t-tzeeq’an-a ASP 1SG/ABS 2/ERG-hit ‘You hit me!’ [TRANSITIVE]

Why is it important to indicate grammatical functions? Many linguists believe that languages use these devices to identify the participants in the event specified by the verb. In a sentence like ‘The President hires young assistants’, for example, the verb calls for two participants: someone who initiates the action (the President), and someone/something affected by it (young assistants). The first is sometimes called an Actor, the second (in this sentence) a Patient or Undergoer. Grammatical functions signaled by subject agreement and accusative case (the latter by virtue of its position following the verb) ensure that each participant is interpreted correctly, or that the sentence does not mean ‘Young assistants hire the President’. Languages that employ neither case-marking nor agreement must obviously find some other way of interpreting such potentially ambiguous sentences.

AGREEMENT

Agreement and Inflection Agreement is often called an inflectional category, as opposed to lexical ones like nouns, verbs, and adjectives. Other inflectional categories include tense, mood, and aspect. Sometimes, agreement is fused with one or more of these to form a single entity. This is the case with the English suffix -s: not only does it signal a third-person singular subject but also present tense and indicative mood. Agreement disappears in the past tense and/or indicative mood: cf. John lives in Maryland now, but last year he lived in Maine; Mary wants to stay, but it is necessary that she leave (NOT: leaves). Still, there is some debate as to whether agreement is truly absent from these forms or simply covert—that is, ‘present but invisible’. In German, for example, agreement is perfectly compatible with the past tense, and in Palauan (Austronesian) with the irrealis mood (similar to nonindicative): (5)

Agreement and past tense (German) 1SG ich glaub-t-e 2SG du glaub-t-est 1SG wir glaub-t-en

(6)

‘I believed’ ‘You believed’ ‘We believed’

Agreement and irrealis mood (Palauan) REALIS 1SG ak2SG ke3SG ng-

IRREALIS kuomole-

Object agreement is often associated with the inflectional category aspect. Again, in Palauan, objects of completed events trigger ‘perfective’ agreement on the verb, whereas objects of ongoing ones do not (IMPF = imperfective; NM = noun-marker; P-preposition; PF = perfective; R = realis): (7) Agreement based on aspect (Palauan) a. ng-chillebed-ii a bilis R3SG-hit-PF3SG NM dog ‘S/he hit the dog’ [PERFECTIVE] b. ng-milengelebed er a bilis R3SG-hit (IMPF) P NM dog ‘S\he was hitting the dog’ [IMPERFECTIVE] The relationship between agreement—which basically pertains to nouns—and tense, mood, and aspect (a property of verbs) is therefore sometimes quite

intricate, and poses a major challenge for linguists attempting to explain it.

Nonstandard Cases While subjects normally trigger agreement in Englishtensed clauses, there are a number of cases in which it is suppressed and/or the status of the subject itself is called into question. One case involves ‘subjunctive’ mood, following verbs like suggest and demand: cf. I demanded that he be on time; We quickly suggested that she leave (compare: …that she should leave, which is also acceptable but does not show signs of agreement). Other languages show reduced forms of agreement under negation, or when one grammatical function or another is questioned: (8) Reduction of agreement (Chamorro/ Austronesian) a. Ha-fa’gasi si Juan i kareta 3SG-wash Juan car ‘Juan washed the car’ [NORMAL AGREEMENT] b. *Hayi ha-fa’gasi i kareta? who 3SG-wash car ‘Who washed the car?’ [WH-QUESTION/NORMAL AGR] c. Hayi fuma’gasi i kareta? who wash+UM car ‘Who washed the car?’ [WH-QUESTION/SPECIAL AGR] The existential there construction in English arguably has two subjects: one occurring in clauseinitial position (filled by there), the other post-verbally and signaling agreement: cf. There’s a man standing on the corner vs. There are three men standing on the corner. Nevertheless, speakers sometimes take the clause-initial subject as the agreeing one, resulting in sentences like There’s four things that can be said about him. Other nonstandard (but occurring) breakdowns involve plural complements of singular subjects (Neither one of you guys are going to win), and coordinate (hence plural) subjects with singular marking on the verb: The fact that we were bombed, and that the markets reopened so quickly suggests that the foundations of our country are quite strong.

Exotic Forms of Agreement Agreement in natural language can deviate significantly from single forms that encode features of person and number. In Potawatomi (Algonquian), each of

23

AGREEMENT these features can be associated with a separate morpheme (OBV = obviative): (9)

Split agreement (Potawatomi) 1SG n-wapm-a 2SG k-wapm-a 3SG w-wapm-a-n 1PL

n-wapm-a-mun

2PL 3PL

k-wapm-a-wa w-wapm-a-wa-n

‘I see him’ ‘You see him’ ‘He sees him (OBV)’ ‘We see them/him (OBV)’ ‘You see him’ ‘They see them/him (OBV)’

Here, the prefixes [n-], [k-], and [w-] represent first, second-, and third-person subjects, regardless of their number. The suffixes [-wa] and [-mun] indicate plurality (the latter also encodes features of the object, i.e. plurality and/or obviation). Interestingly, the same prefixes and suffixes also appear on nouns to indicate possession and number: (10) Nominal agreement (Potawatomi; c = alveopalatal affricate) 1SG n-ciman ‘my canoe’ 2SG k-ciman ‘your canoe’ 3SG w-ciman ‘his/her canoe’ 1PL 2PL 3PL

n-cimanwa-nan k-ciman-wa w-ciman-wa

‘our canoe’ ‘your canoe’ ‘their canoe’

In the larger picture then, agreement must be regarded as more than just a property of verbs. ‘Subjects’ of NPs (possessors) corrolate with sentential subjects: it is no accident that the possessive marker [-s] in English has the shape it does (cf. John’s dog). In Irish and other Celtic languages, prepositions inflect for the person/number features of their objects; this grammatical relation roughly correlates with sentential objects. To fully appreciate how complex agreement can be, consider that objects of transitive verbs in Potawatomi can also be associated with separate morphemes of person and number. In (9) above, the suffix [-a] signals that the object is a third person, while [-n] registers its plurality and/or obviation; also notice how the subject plural marker [-wa] intervenes between these two suffixes. Inuktitut (Eskimo) takes a different tack in the distribution of agreement features, using single forms called portmanteau to encode the person and number of both subject and direct object (ABS = absolutive; AL = allative; CMT = comitative; PRT = participle; REL = relative):

24

(11) Portmanteau (Inuktitut) a. tuni-jara nutarar-mut give-PRT/1SG/3SG child-AL arna-up taku-jaa-nut woman-REL see-PRT/3SG-AL ‘I gave it to the child that the woman saw’ b. arna-up pilauti-milk woman-REL knife-CMT angut tuni-jaa man (ABS) give-PRT/3SG/3SG ‘The woman gave the man the knife’ In (11a), the form [-jara] signals a first-person subject and third-person object, while in (11b) [-jaa] indicates two third persons.

Evolution of Agreement One might well ask why agreement is so pervasive in language, or where it comes from. Many morphemes appear to originate as independent pronouns that slowly undergo a process of phonological reduction and subsequent attachment. A comparison of pronouns and agreement from Udi (Caucasian) demonstrates this most clearly: (12) Pronouns and agreement (Udi) AGREEMENT 1SG -zu, -z 2SG -nu, -n, -ru, -lu 3SG -ne, -le, -re, -n

PRONOUNS zu un meno, kano, seno

1PL 2PL 3PL

yan van, efan met’ovon, kat’ovon, set’ovon

-yan -nan, -ran, -lan -q’un

Although the match-up is far from complete, there is enough similarity of form to allow for a theory of language change. Independent/emphatic pronouns may pass through an intermediate stage of cliticization on their way to becoming agreement morphemes. Pronominal clitics also share the fundamental features of person, number, and gender. French has a series of subject-markers traditionally regarded as pronominal: je, tu, il, etc. Nevertheless, these forms cannot be separated from the (auxiliary) verb like their English counterparts; cf. *Il vraiment est retourné vs. He really has gone home. This suggests that je, tu, and il are clitics, and in the frequent presence of emphatic pronouns may

AINU someday evolve into a new series of agreement morphemes (cf. Moi, je vais retourné). Also, note that (except for first- and second-person plurals) the traditional agreement suffixes of modern French are no longer pronounced.

Theory of Agreement How do linguists analyze agreement? Usually, a local relationship is assumed to hold between the element that triggers it (e.g. the subject NP), and an inflectional head, which may be comprised of tense, mood, or aspect. Agreement is the visible sign of this relationship. This explains why only verbs in tensed clauses show agreement—at least in English. In structural terms, the subject NP is called the specifier of the clause, attached to but outside the immediate domain of inflection: [NP [verb+inflection, etc.]]. This also parallels the internal structure of the NP, where articles like a, the, and some are attached to but outside the immediate domain of the noun: [the [dog]]. In fact, possessed NPs (e.g. John’s dog) can be assigned essentially the same structure as a clause if the possessor (John) is regarded as a specifier, and the sign of possession ([-’s]) as a kind of inflection. It is probably no accident that English uses the same sound to express both specifier–head relationships. The function of agreement—that is, as a means of highlighting one participant in the event described by the verb—can be explained in the following way. First, a verb’s meaning is more closely comprised of its nonActor participants than its Actor participants. The verb

to kill, for instance, can involve essential information that is separate from the person who initiates the action (cf. kill an hour, kill a motion, kill the mood, kill the light—in addition to the standard meaning ‘cause to die’). Languages typically recognize this distinction through the mechanism of case-marking, and nonActor participants are often made visible through cases selected by the verb. Actors, on the other hand, are less likely to be singled out by case. Instead, they must rely on some other mechanism to be rendered visible, such as agreement. Inflection then—the head of every clause—provides a specifier position that can accommodate one non-case-marked participant for every verb. References Anderson, Steven. 1992. A-morphous morphology. New York: Cambridge University Press. Aronoff, Mark. 1994. Morphology by itself: stems and inflectional classes. Cambridge, MA: MIT Press. Comrie, Bernard. 1989. Language universals and linguistic typology, 2nd edition. Chicago: University of Chicago Press. Croft, William. 1990. Typology and universals. New York: Cambridge University Press. Givon, Talmy. 1984. Syntax: a functional–typological introduction. Amsterdam: John Benjamins Press. Marantz, Alec. 1992. On the nature of grammatical relations. Cambridge, MA: MIT Press. Pinker, Stephen. 1984. Language learnability and language development. Cambridge, MA: Harvard University Press.

MARK CAMPANA See also Austronesian; English; French Language; German; Grammar, Traditional; Japanese

Ainu The Ainu are an indigenous group found mainly on the northern Japanese island of Hokkaido. Physically and linguistically distinct from the neighboring Japanese, at one time the territorial range of the Ainu extended throughout the northern Japanese island of Hokkaido (Yezo), including the Kurile islands to the northeast and southern Sakhalin (Karafuto) to the north. Evidence from place names, however, suggests that at one time the Ainu had also settled in northern Tohoku, on the main Japanese island of Honshu, and on Russia’s Kamchatka Peninsula as well. Due to resettlement of the Sakhalin Ainu (1875) and the Kurile Ainu (1884) today, however, the majority of the

remaining Ainu are situated on the island of Hokkaido. It should be noted that the Ainu language found in northern Japan is unrelated to the Turkic language bearing the same name that can be found in western China. The genetic and linguistic history of the Ainu has been the topic of much debate. Physical features such as increased body hair (when compared to the Japanese) and more distinct facial features make the Ainu appear Caucasian to many people and as such the Ainu have sometimes been considered Caucasian (Caucasoid) in ancestry. Other scholars have sought to draw a relationship between the Ainu and native

25

AINU American groups. Recent studies conducted with DNA testing, however, suggest the possibility that the Ainu are Mongoloid in origin. Numerous attempts have been made to classify Ainu linguistically, including theories that place Ainu in the Malayo-Polynesian, Paleo-Asiatic, and UralAltaic language families. At present, there are several competing theories on the relationship of Ainu with other languages. The first and most widely offered explanation is that Ainu is a language isolate, with at present no apparent relationship with other languages. The second theory rests on the assumption that Ainu is related to the Altaic language family, and languages such as Japanese and Korean. Some studies arguing the latter position suggest that if Ainu is at all connected to other Altaic languages, then the splitting of Ainu from Japanese and other Altaic languages must have occurred at least 10,000 years ago. However, Alexander Vovin (1993), in his work on reconstruction of Proto-Ainu (the hypothesized ancestor language), suggests a connection to the Austro-Asiatic languages of the south. Further complicating matters has been the historical lack of a writing system, and as a consequence, a written record of Ainu over time. The relatively rapid transition from a viable, functioning language with monolingual native speakers to one that is, at present, nearly extinct has further confounded attempts to conclusively place Ainu among the world’s languages. The Ainu have had a continuous presence in Hokkaido for at least the past 1300 years. The first record of the Ainu on Yezo occurs in the Japanese Kojiki (A.D. 720) calling them ‘Emishi’, or those outside of the Japanese law. The Ainu refer to themselves as Utari or ‘compatriot’ in Ainu, and this term is in more frequent use at present because of the sometimes negative connotations that have been leveled on the term ‘Ainu’ by the Japanese. The word Ainu simply means ‘human’. While no written form of Ainu existed until recently (Ainu is now written in Romanized form or with the Japanese katakana script), the Ainu culture has developed a rich oral tradition most well known for their epic poems called Yukar. Yukar follow traditional patterns in their subject matter with some of the Yukar told from the perspective of an animal, others based on activities of ancestors, and still others are hero tales. The Yukar differ from the spoken language, tending to exhibit more complex (polysynthetic) word structures and more conservative features than spoken Ainu. These stories have served as a rich source of linguistic material as they maintain archaic forms, providing a rich source of data attesting to earlier forms of the Ainu language. Yukar has come to be a collective term

26

for all forms of epic story in Ainu but as Masayoshi Shibatani notes, In the strict sense of the term yukar refers only to the heroic verse, mythic epics being more specifically referred to as kamuy yukar, mat yukar, or oyna. There are as well prose-style old stories and folktales.

As noted above, Ainu is a polysynthetic language, where words tend to be composed of multiple morphemes. This is particularly the case in classical Ainu. In contrast, colloquially spoken Ainu tends to be more analytical, with more of the morphemes appearing as independent words, suggesting the possibility of a long-term change in progress. In terms of its grammar, Ainu is a S(ubject)– O(bject)–V(erb) language. But, as seen in the following example, word order is important in determining the grammatical relationship in Ainu. Kamuy aynu rayke. bear person kill ‘The bear killed the man.’ Aynu kamuy rayke. person bear kill ‘The man killed the bear.’ Phonologically, Ainu has a five-vowel system (/a/, /i/, /e/, /o/, /u/) with syllable-initial vowels preceded by a glottal stop. The consonants of Ainu, 12 in all, include the stops of /p/, /t/, /k/, and the glottal stop, sibilant /s/, glottal fricative /h/, affricate /c/, semivowels /w/,/y/, nasals /m/ and /n/, and the liquid /r/. Ainu has a pitch accent system in which syllables are pronounced with high or low pitch. Its syllable structure is C(onsonant)V(owel) or CVC. The exception to this, however, is the Sakhalin dialect, which contained long and short vowel contrasts, also allowing for CVV to occur. Vowel sequences tend to be avoided in other forms of Ainu. Vocalic euphony or vowel harmony, the assimilation of a vowel to another vowel in the word or phrase, is one of the more notable aspects of the sound system of Ainu. The Ainu vocabulary is rich with references to nature, reflecting the close relationship held between the people and natural phenomena. Most Ainu villages (called kotan) were located alongside rivers and coasts and former Ainu settlements are easily recognizable in place names today with endings such as -nai or -pet (-betsu is the Japanese representation), both meaning ‘river’, as found in the place names of Wakkanai, Noboribetsu, and Monbetsu. The importance of salmon and bears to Ainu culture is represented by the

AINU abundance of lexical items related to these two animals. Particularly, the role of the bear was an important part of Ainu culture, and they were often kept in villages for ceremonial events and as food. Within Ainu, three distinct dialect groups are generally recognized, the Kurile, Sakhalin, and Hokkaido dialects, each representing the distinct regions where the Ainu once inhabited. The last speaker of the Sakhalin dialect is believed to have died in 1994. Within these dialect groups, further distinctions can be made, such as that between southern and eastern Hokkaido Ainu. At least 19 dialects of Ainu have been identified, most of which are now extinct. It is estimated that the number of Ainu descendants (including those of mixed Ainu-Japanese heritage) number around 24,000, with those of solely Ainu descent being very small. Postwar census figures classify the Ainu as Japanese for census purposes; therefore, only estimates of the number of remaining Ainu in Hokkaido are possible. Discrimination aimed at the Ainu by Japanese as well as a Japanese desire to view Japan as a single homogeneous society have further caused many Ainu to conceal their heritage. It is believed that the actual number of Ainu in Hokkaido may be double or triple the figure of the government estimate of 24,000 (1995). The Ainu face regular discrimination in employment and marriage opportunities. Such discrimination and fear of disclosing their Ainu ethnicity have, over time, contributed significantly to the loss of the Ainu language. The Ainu language and culture have been impacted dramatically by the influx of the Japanese into Hokkaido, which began as early as the 1300s. The peak of the Ainu culture occurred in the thirteenth and fourteenth centuries, prior to the large-scale arrival of the Japanese (wajin), which began in the fifteenth century. From this point onward, battles with the Japanese in 1457, 1669, and 1789, as well as diseases such as smallpox, measles, cholera, and tuberculosis all took their toll on the Ainu. In 1550, an agreement was reached between the Ainu and the Japanese to allow a Japanese settlement area on Yezo. This would be a turning point in AinuJapanese relations. Over the next three centuries, Japanese control of Hokkaido increased and finally in 1869 the official government-sanctioned colonization of Hokkaido began. The Japanese population of Hokkaido quickly grew to over 1 million, placing the Ainu in a minority. From this point onward, the Ainu were systematically stripped of their cultural and linguistic identity, land, and hunting rights through a forced assimilation policy of the Japanese government. Also during this period, disease again struck the Ainu, leading to a decrease in their population.

By 1871, many Ainu customs, including the tattooing of women’s mouths and the burning of a familial household upon the death of a family member, were banned. The Ainu language was also banned as the Japanese government set out to assimilate the Ainu into becoming Japanese and officially classifying them as ‘former aborigines’. In 1899, the Hokkaido Aborigine Protection Act was created; yet, despite its name, the act did little more than increase the efforts to assimilate the Ainu. The Japanese government encouraged the Ainu to shift from their traditional role as hunter/gatherers to farming. At the same time, Ainu lands were ceased. When land was redistributed to the Ainu, they were given the least fertile lands. In 1901, Ainu children were forced to attend Japanese primary schools and Ainu children began learning Japanese as their first language. The assimilationist policies of the Japanese government resulted in a near elimination of the Ainu language and culture by the mid-twentieth century. Today, there are no monolingual native speakers left. Today, the Ainu are fluent Japanese speakers. In spite of the past Japanese efforts to eliminate the Ainu language, bilingual speakers of Ainu and Japanese do exist, however, and recently modest efforts have begun to continue the teaching of the Ainu language and traditions to those willing to learn. Ainu language use also continues in ceremonial and ritualistic use and in events performed for tourists. A preserved Ainu village at Shiraoi is maintained for educational and tourism purposes. In 1946, the Ainu Association of Hokkaido was founded, seeking to bring increased attention to the fight for Ainu rights, the preservation of the language and culture, and a reevaluation of the Hokkaido Aborigine Protection Act. In 1994, Shigeru Kayano became the first Ainu member of the Japanese Diet (parliament). Subsequently, in 1997, the Japanese government finally and officially recognized the Ainu as the indigenous people of Hokkaido and created the Ainu rights law. This law was designed to replace the Hokkaido Former Aborigine Protection Act. It requires local governments in traditional Ainu areas to support promotion of the Ainu culture, including promotion of the Ainu language. As a result, government-subsidized weekly Ainu language radio broadcasts are offered in Hokkaido with free texts that are widely available. There are also 13 state-supported language classrooms in traditional Ainu villages throughout Hokkaido that seek to provide opportunities for students to learn Ainu. A bi-weekly Ainu language newsletter has also been produced. The 1997 Ainu rights law provided official recognition to the Ainu and a degree of legimaticy to the

27

AINU efforts aimed at preservation and revitalization of the language. Whether the Ainu language will undergo a successful revitalization is dependent upon the ability of the Ainu to increase awareness among the Japanese of what might be lost, should the Ainu language and culture disappear. References Batchelor, John. 1889. An Ainu–English–Japanese dictionary. Tokyo: Y. Kumata; Iwanami-Syoten, 4th edition, 1938. Brochure on the Ainu. Sapporo, Japan: The Ainu Association of Hokkaido, n.d. Chiri, Mashiho. 1952. Ainugo ni okeru boin chouwa. Hokkaido Daigaku Bungaku Kiyou 1. 103–18.

Hattori, Shiro. 1959. Nihongo no Keitou. Tokyo: Iwanami. Kayano, Shigeru. 1994. Our land was a forest: an Ainu memoir. Boulder, CO: Westview Press. Maher, John, and Kyoko Yashiro. 1995. Multilingual Japan: an introduction. Journal of Multilingual and Multicultural Development 16(1). 1–17. Patrie, James. 1982. The genetic relationship of the Ainu language. Oceanic Linguistics Special Publication No. 17. Honolulu: University of Hawai’i Press. Shibatani, Masayoshi. 1990. The languages of Japan. Cambridge: Cambridge University Press. Vovin, Alexander. 1993. A reconstruction of Proto-Ainu. Leiden: E. J. Brill.

SHAWN M. CLANKIE See also Altaic; Japanese

Akan and Nyo Languages Nyo, like Left Bank, belongs to the (New) Kwa subfamily of the Niger–Congo group. Nyo has more languages than the Left Bank. There are five daughter languages in Nyo: Agnéby, Attié, Avikam–Alladian, Potou–Tano (which accounts for over half of New Kwa in terms of both the speakers and geographical area), and Ga–Dangme. Potou–Tano is composed up of six language groups: Ega, Potou, Tano (previously called Volta–Comoé), Lelemi (and its related languages), Logba, and Basila–Adele. Westermann classified Avikam–Alladian, Agnéby, Attié, and Potou under the Lagoon subfamily; and Lelemi (and its related languages), Logba, and Basila Adele as Togo–Remnant languages. Agnéby: The Agnéby languages, a Central Nyo group, are spoken in the Ivory Coast in the Agnéby and Bandama river basins by 230,000 people. Adioukrou, Abiji, and Abbey are the three languages in this group. Adioukrou (Adjukru) is spoken in the north of the Ebrié Lagoon westward of the Agnéby river by about 70,000 people. Abidji is spoken by about 80,000 people in the Abidjan District and in the subdistricts of Sikensi and Dabou. Abidji has two main dialects: Enyembe and Ogbru. Speakers of Abiji tend to be bilingual in Adioukrou, Baoule, and Jula. Important linguistic features include tone and nasalization on syllables. Abbey, spoken by about 80,000 people, is spoken around Agboville and to the north of the Abiji and Krobou areas. Examples of words and their plurals taken from Adioukrou are as follows:

28

libr ‘red’ - ebr ‘reds’ lufu ‘white’ - efu ‘whites’ lorn ‘snake’ - morŋn ‘snakes’ Attié: Attié is spoken in Côte d’Ivoire by about half a million people living in Abidjan, Anyama, Alep, Adzope, Affery, Agou, Akoupe, and Yakasse– Attobrou. The three main dialects of Attié are Anaindin (Nindin), Ketin, and Bodin (the most prestigious and populous). Most Attié speakers are bilingual in Abbey, Anyin, Baoule, Ebrie, and Jula. Avikam–Alladian: It is a Western Nyo language spoken in the Western Ivory Coast. It has two languages: Avikam and Alladian. Avikam is spoken in the basin of the Bandama River by 12,000 speakers, whereas Alladian is spoken by about 15,000 people on the east of Avikam. Potou–Tano: The six language groups identified under Potou–Tano are Ega, Potou, Tano, Lelemi (and its related languages), Logba, and Basila–Adele. Ega: Ega was originally classified as Kru, but Richards (1982) proved that it is Kwa. It is located inside the Kru area. Like other Kwa languages, Ega has a reflex of the cognate (words maintained by languages of common ancestry) nyɔ /υ ˜ / ‘two’. Potou: It is spoken in the Eastern Ivory Coast Lagoon area by about 75,000 people. Potou has two languages:

AKAN AND NYO LANGUAGES Ebiré (spoken by 60,000 people in the north of the Ebiré Lagoon from the Agnéby River to the Potou Lagoon in the east) and Mbatto (spoken by about 15,000 between the Potou Lagoon and the Comoé River). Tano: Tano, previously called Volta–Comoé, has more speakers than any of the subfamilies of the Potou–Tano group. Tano has four subfamilies: Krobou, Western Tano (Abouré and Eotilé), Central Tano (Bia, Akan), and Guang (Southern and Northern). Krobou has about 6000 speakers and is spoken within the Agnéby language area in the Ivory Coast. Abouré and Eotilé are also spoken in the Ivory Coast. Abouré is spoken by about 40,000 people in the Comoé River basin near the sea, whereas Eotilé is spoken by 3000 people in Vitre #I and Vitre #II on the Ebiré Lagoon to the north of Grand Bassam. Bia: The Bia languages, sometimes referred to as Nzema–Anyi–Baule, are spoken by nearly two and a half million people in the Bia, Tano, Nzi, Bandama, and Comoé River basins and along the south coast of Ghana and the Ivory Coast. The two main subgroups are: Northern Bia (comprising Anyi, Baule, Chakosi (Anufo)); and Southern Bia (Nzema and Ahanta). Anyi and Nzema are spoken in Ghana and in the Ivory Coast, Baule in the Ivory Coast, Ahanta in Ghana, and Chakosi in Ghana and Togo. Nzema has two dialects: Nzema and Evalue. Anyi also has two dialect groups: Anyi (Aowin) and Sehwi (Sanvi). Anyi has other dialects like Indene, Moronou, Bona, Djoablin, Ano, and Bini (all in the Ivory Coast), and Afema and Brossa in Ghana. Baule, with nearly one and a half million speakers, is the largest language group in the Bia subfamily. Chakosi is found in an enclave in the Gur territory on both sides of the Ghana–Togo border. In Togo, Chakosi is found in Sansanne–Mango and in Ghana, in the Chereponi area. Most speakers of Nzema (about 120,000) are found in Ghana; in the Ivory Coast, there are about 50,000 speakers. Important linguistic features in the Bia group include consonant mutation (in which consonants change some of their phonetic qualities in certain phonetic or grammatical contexts. For example, in Nzema /k/ changes to /h/ between vowels. /k/ in kɔ ‘go’ becomes /h/ in jihɔ ‘she’s gone’) and vowel harmony (cooccurrence restrictions in the distribution of vowels, i.e. the vowels are grouped into two sets and in words of more than one syllable; only vowels from one set occur. For example, in Nzema, words otosu mean ‘he still cries’ and ɔtɔkυ ‘he still fights’; all the vowels in otosu are produced with an advanced tongue root position, whereas those in ɔtɔkυ are produced with a retracted tongue root). Other linguistic features include double articulated sounds

(single sounds with two different places of articulation, e.g. /kp, tp, gb/ . For example, in Chakosi, the words kpiε ‘old man’ and gblaki ‘to faint’ have /kp/ and /gb/, respectively). Bia languages also have tone terracing, where pitch is lowered toward the word end. Plural is formed through affixation. For example, in Baule, the plural marker {-mu ˜} can be added to any count noun. For example, wa ‘child’ wamu ˜ ‘children,’ ti ‘head’ timu ˜ ‘heads.’ There are also independent nasalized vowels (phonemic nasalization of vowels). For example, in Baule, wu means ‘see’, whereas wu ˜ means ‘die’.

Akan Akan is the largest of the Central Tano group of languages. Until the 1950s, these clusters of mutually intelligible varieties did not have a single name. Some of the dialects (such as Fante, Bono, and Wasa) were viewed as different languages. However, a unified Akan orthography was designed by the Akan Orthography Committee in the 1950s and since then, the name Akan has been adopted as the name of the language. Akan is spoken by over ten million people in Ghana and in the Ivory Coast. In Ghana, Akan is spoken between the rivers Tano and Volta, stretching from the coast to the inland. Abron (Bono) or Dormaa is spoken in Ghana and in Côte d’Ivoire by about a million people. In Côte d’Ivoire, Akan (Abron) is spoken in the northwest by about 60,000 people. Speakers of Abron speak and understand Asante Twi. Wasa is spoken by nearly 300,000 people in Southwestern Ghana. It has two main dialects called Amenfi and Fianse. Wasa and Abron are mutually intelligible. In Ghana, Akan is spoken by nearly half of the population as their mother tongue and by two-thirds of the population as a second language. Akan has 11 dialects: Fante (Mfantse), Akuapem, Asante, Agona, Dankyira, Asen, Akyem, Kwawu, Ahafo, Bono (Abron), and Wasa. The Akuapem, Fante, and Asante dialects have different officially recognized orthographies and are used on radio and TV. Quite recently, a unified orthography for all the Akan languages has been adopted. The Twi (Asante–Akyem–Kwawu) dialects are the largest dialects, with a total population of over seven million. Fante, spoken along the coast, has over two million speakers and four important dialects: Agona, Gomua, Abura, and Anomabu. Akuapem is spoken on the Akwapim Hills between the Akyem and Ga areas. Important linguistic features include a Subject– Verb–Object word order, example: S V O Kofi nom nkwan ‘Kofi drinks soup’ Ama dɔ Kwame ‘Ama loves Kwame’

29

AKAN AND NYO LANGUAGES Akan languages also have two types of vowel harmony: cross-height and rounding. With respect to cross-height harmony, we observe that the vowels are grouped into two sets: advanced /i, e, æ, o, u/ and unadvanced /¹, ε, a, ɔ, υ/. In any word of two or more syllables, only vowels of one harmonic set occur. For example, obewu, ‘She’ll die’, has vowels from the advanced set, whereas ɔbεwυ, ‘She’ll give birth’, has vowels from the unadvanced set of vowels. With respect to rounding harmony, in some of the Akan languages, especially in Fante, vowels of verbal affixes agree with those of a verb root by being advanced or unadvanced, rounded or unrounded. For example, muruwu, ‘I’m dying’, has only rounded advanced vowels, whereas in miridzi, ‘I’m eating’, all the vowels are both advanced and unrounded. Rounding harmony is also found between vowels of nominal roots and affixes in the Asante and Akyem dialects. For example, the vowel of the noun suffix /o/ in owuo ‘death’ agrees in rounding with the vowel /u/ of the noun root, owu ‘die’. The vowel /e/ suffix of the noun root esie ‘hill’ harmonizes with those of the root /e, i/ by being unrounded. The Akan languages have uncommon sounds like the alveolo-palatal semivowel /ɥ/, labio palatalized alveolo-palatal affricate /t ɥ/, occurring in words like we /ɥi/’chew’ and Twi /t ɥi:/ ‘name of some Akan dialects’. Another important feature of the Akan languages is tone terracing (especially downdrifting, where, during the production of long utterances, there is a gradual drop in pitch heights if high tones are interspersed with low tones. The drop in pitch could be such that an utterance final high tone could be lower in pitch than an utterance initial low tone. The final drop in pitch could also reach the bottom of a speaker’s pitch range. For example, in: Pàpá Kòfí dìdí wìé kòtìé ‘Father/Mr. Kofi finishes eating and goes to listen’, the initial low tone on /pà/ of Pàpá is higher in pitch than the final high tone on /é/ of kòtìé. Guang: The Guang subfamily, which is spoken by about half a million people in Ghana, The Ivory Coast, Togo, and Benin, has more languages than the Central Comoé group. However, the languages in this subfamily are for the most part spoken by small numbers of people. This group of languages subclassify into: North, North-East, Central, and South. North Guang languages include Gonja, Nkonya, Nawuri, and Choroba. Gonja (Ngbanito) is spoken in a large area from Bole to Kamabuie near Salago and between the White Volta, Volta, and Daka rivers in Ghana. In the Ivory Coast, Gonja is spoken in the Bondoukou and in Togo it is spoken at Semere and is called Bazantcé. The Gonja variety spoken in the Republic

30

of Benin is referred to as Chombulon or Tshummbuli (similar to a Ghanaian Central Guang variety called Nchumburu). Southern Guang subdivides into Coastal Guang (a dialect cluster—Awutu, Efutu, and Senya). Awutu is spoken in Winneba, Senya Breku, Obutu, and Bawjiase in Ghana, spoken by about 200,000 people, and Hill Guang is spoken by about 150,000 people on the Akwapim Hills and on the left bank of the Volta Lake. Important dialects include Cherepong, Larteh, and Anum (Gwa). Other Guang subfamilies include Lelemi-Lefana (Buem), Akpafu (Siwu), and Santrokofi spoken in Ghana and Togo by about 45,000 people; Logba spoken by about 4,000 people on the Ghana–Togo border; and Basila–Adele, comprising the Basila and Adele dialects, is spoken by about 10,000 in both Ghana and Togo. Basila has three subdialects: Gokolodjya, Gilempla, and Giseda. Adele has more speakers in Togo (7,000) than in Ghana (3,000). Like the Akan group of languages, the Guang languages have advanced tongue root vowel harmony. For example, in the following Efutu words, all the vowels are either advanced oi ‘man’or unadvanced ɔ`kυ´ɔ` ‘again’. The harmony may operate over entire sentences. For example, muudi nu ‘She is eating (food)’; mυ´n`wɔ´ ‘She did not go’. Negation is formed through affixation and consonant alternation. For example, in Siwu, in negating commands, voiceless stops change into voiced stops and voiced stops become nasals. Also, the morphemic particle daa is preposed to the main verb. Example, kε`lε` ‘go’ dàá ε´ lε` ‘don’t go’; kpε` ‘weed’; dàá bε´ ‘don’t weed’. Ga–Dangme: Ga–Dangme has two languages: Ga and Dangme. The Ga–Dangme subfamily is spoken in Ghana on the southeastern coast around Accra and inland. Ga is spoken in Ghana along the coast in and around Accra (Ghana’s capital), and in the Densu and Kpone basins by about 500,000 speakers. Dangme (Adangme) is spoken along the coast (to the east of Ga) and at the mouth of the Volta and also in the Akwapim Hills. Like Ga, Dangme has over half a million speakers. Dangme has six dialects: Ada, Ningo, Prampram, Shai, Krobo, and Osudoku. Krobo is the largest dialect, with about 300,000 speakers. Structurally, Ga has two tones, High and Low, and a downstepped high tone (a high tone whose pitch has been lowered. Downstep is marked with an exclamation mark placed before the syllable). For example, ékòmé ‘one’, gbé!kε´ ‘boy’. Plurals are formed through affixation, (suffixation). For example, tsò ‘tree’, tsèi ‘trees’; tʃu`‘ house’ tʃu` i ‘houses’. There are three tones, High, Mid, and Low, in Dangme.

ALBANIAN References Dakubu, Mary Esther Kropp. 1988. The languages of Ghana. London: Kegan Paul International. Greenberg, Joseph. 1955. Studies in African linguistic classification. New Haven: Compass Publishing Company.

Stewart, John. 1989. Kwa. The Niger–Congo languages, ed. by John Bendor-Samuel. New York: University Press of America.

SAMUEL OBENG See also Niger-Congo

Albanian Albanian is the name English speakers use to refer to the language of Albania (Albanians call the language ‘shqipe’), as well as for the branch of the Indo-European family to which it belongs. As a branch of IndoEuropean, Albanian is a sister to many language groups from Europe to South Asia: Germanic (including English, Swedish, and German, for example), Romance (including French, Spanish, and Italian), Slavic (including Russian, Serbian, and Polish), Indic (including Hindi, Gujarati, and Urdu), Hellenic (Greek), Armenian, and Celtic (including Irish and Scottish Gaelic). Persuasive evidence for the kinship of Albanian with other Indo-European languages was first advanced in the middle of the nineteenth century and continues to be amassed. In addition to regular correspondence of sounds between particular words and parts of words in Albanian with those in several already established IndoEuropean languages, the evidence lies in features of Albanian verb conjugation and noun declension that reflect a common origin with those languages: for example, the irregular forms of certain verbs (e.g. Albanian ështëEnglish isLatin estSanskrit asti, while Albanian jamEnglish amLatin sumSanskrit asmi). Albanian linguists claim that the predecessor of modern Albanian is the ancient Illyrian language, but since little is left of Illyrian, verification for this claim is elusive. The earliest texts clearly identifiable as Albanian are scarce and go back only to the fifteenth century; we have no record of the characteristics of the language during earlier periods, although extensive documentary evidence exists for many other IndoEuropean languages. To complicate the task of tracing its origins, Albanian, like English, has adopted and adapted a great substance of the vocabulary from other languages throughout its history—in earlier times, basic vocabulary from Latin and Turkish and more recently, technical vocabulary from Italian, Russian, and English. Today, Albanian is spoken by more than six million people—about three million in Albania itself, another

two million in Kosovo, some 300,000 in Macedonia, and at least 700,000 more living mainly in Montenegro, Greece, Turkey, Italy, and the United States. In recent years, both voluntary and forced emigration of Albanians has greatly increased the presence of speakers of the language outside their homelands. During its known history since the fifteenth century CE, Albanian has been associated with two main dialects: Tosk in the south and Gheg in the north. After World War II, the government made increasingly successful efforts in Albania to adopt a unified written form of the language, called Standard Literary Albanian, and in 1968 this form of Albanian was also adopted for schools and mass media by Albanians in Kosovo. However, for ordinary conversation, few Albanians in Albania and Kosovo completely abandoned their own dialects. In rural areas, ‘unified Albanian’ may be scarcely used and known only from television and radio, and even well-educated, urban Albanian speakers exhibit regional and social differences in their language, which appear in pronunciation, vocabulary, and grammar. The speech of Albanian speakers living outside Albania reflects their Gheg or Tosk origins. A modern alphabet was devised for Albanian in 1908. This 36-letter alphabet includes six digraph letters (th, dh, sh, zh, gj, nj) and two special characters (ë and ç), and was adopted quite readily by speakers. For Gheg words, a carat (^) is used to mark nasalized vowels, but the vowel length so distinctive to Gheg dialects remains unmarked in written Albanian. In addition to its genetic membership in the IndoEuropean language family, Albanian shares certain specific features with other Indo-European languages in the Balkan area. For example, like Romanian, Albanian marks ‘definiteness’ (in English, marked by the word the) by endings on nouns: Albanian gjysh, ‘grandfather’ becomes gjyshi, ‘the grandfather’;

31

ALBANIAN similarly, Romanian bunic, ‘grandfather’ becomes bunicul, ‘the grandfather’. Particular features of Albanian seem particularly striking to non-Albanians. Already mentioned is the vast number of loan words from the international vocabulary—such as heroik, burokrat, standardizon—as well as many words from Turkish (during the 400-year occupation of Albania under the Ottoman Empire), such as penxhere (window), fukara (poor or pauper), and çoban (shepherd). Foreign names are generally spelled in Albanian to approximate their sound—to an Albanian ear—in the language from which they derive, so that Heather Lockwood would become Hedhër Lakvud. As in many other languages, Albanian indicates the grammatical function, or case, of a noun with noun endings. Thus, gjyshi, ‘the grandfather’, is the subject of a verb, but as the direct object of a verb it takes the form gjyshin, and as the indirect object of the verb, it takes the form gjyshit. The verb system in Albanian is quite complex. Simple verbs not only have different endings to indicate whether their subject is the person speaking (‘first person’), the person spoken to (‘second person’), or someone or something else (‘third person’), but also whether that subject includes the person speaking (‘first person plural’), the person addressed (‘second person plural’), or is entirely composed of others (‘third person plural’). Those endings also indicate whether the time is present or past; if past, the ending indicates whether the verb applies over time or at a particular moment.

Another set of endings indicate whether the subject is itself the object of the verb. These ‘medio-passive’ endings are used for a range of functions that are quite different in English. For example, e rruan means ‘they shave him’, but rruhen can mean, ‘they are shaved’ (passive), ‘they shave themselves’ (reflexive), or ‘they shave one another’ (reciprocal). Still other sets of verb endings can express the speaker’s surprise or dismay, and yet others may express a wish. Pronouns also take different forms for varying grammatical functions. For example, as subjects of verbs, objects of verbs, and objects of certain prepositions, first- and second-person pronouns have the accented forms shown in Table 1. Albanian lacks corresponding forms for thirdperson pronouns, but instead uses a complex system of ‘pointing’ words that convey the perceived distance of the referent, as well as the referent’s number, gender, and case (see Table 2). Unaccented pronoun forms indicate that the verbs immediately following have an object (direct or indirect) (see Table 3). These forms appear whether or not an accented object is also present in the sentence, so that ‘Agim told me’ is më tha Agimi, and ‘Agim told me’ is më tha Agimi mua; similarly, ‘Agim told her’ is i tha Agimi, and ‘Agim told her’ is i tha Agimi asaj. Unaccented third-person direct objects may occur in combination with any unaccented indirect object, yielding the forms shown in Table 4. The position of the unaccented pronoun (or combination) in Albanian is unique among Indo-European

TABLE 1

First person Singular Plural Second person Singular Plural

English equivalent

Subject

Direct object

Indirect object

Prepositional object

I, me we, us

unë ne

mua ne

mua neve

meje nesh

you (thou) you all (you)

ti ju

ty ju

ty juve

teje jush

TABLE 2 English equivalent Third person Singular Masculine Feminine Plural Masculine Feminine

32

Subject

Direct object

Indirect object

Prepositional object

Far

Near

Far

Near

Far

Near

Far

Near

that one: he, it, him that one: she, it, her

ai ajo

ky kjo

atë atë

këtë këtë

atij asaj

këtij kësaj

atij asaj

këtij kësaj

they, them they, them

ata ato

ky kjo

ata ato

këta këto

atyre atyre

këtyre këtyre

asosh asosh

kësosh kësosh

ALBANIAN TABLE 3 Person

Number

English equivalent

Direct object

Indirect object

Singular Plural

me us

më na

më na

Singular Plural

you (thou) you all (you)

të ju

të ju

Singular Plural

him, her, it them, ‘em

e i

i u

First Second

much like English of to indicate possession—djali i Gjonit (the son of John, or John’s son)—but more often to indicate some looser relationship in which the noun after the attributive article narrows the scope of its referent. For example, such a construction may be used to identify: ●

Third ●



TABLE 4 ●

Indirect Direct object object

Combination Example

Meaning



e

= ma



i

= m’i



e

= ta



i

= t’i

na

e

= na e

na

i

= na i

ju

e

= jua

ju

i

= jua

i

e

= ia

i

i

= ia

u

e

= ua

u

i

= ua

Agim gave it to me Agim gave them to me Agim gave it to you Agim gave them to you Agim gave it to us Agim gave them to us Agim gave it to you Agim gave them to you Agim gave it to him/her Agim gave them to him/her Agim gave it to them Agim gave them to them



ma dha Agimi m’i dha Agimi ta dha Agimi t’i dha Agimi na e dha Agimi na i dha Agimi jua dha Agimi jua dha Agimi ia dha Agimi ia dha Agimi ua dha Agimi ua dha Agimi

languages in one particular aspect: in positive commands, it may come directly after the verb, but if the command is addressed to a plural ‘you’, then it comes before the verb ending. Finally, Albanian is unusual among languages in its use of a set of ‘attributive articles’, which indicate that the word after it refers to a noun or pronoun, its ‘referent’, which may or may not be present in the sentence. The choice of which of the four members of this set (i, e, të, së) is to be used is determined by the grammatical form of the referent. Pure adjectives must always be preceded by one of these articles: këndim i mirë ‘good singing’ with the adjective i mirë ‘good’ —contrast the adverb mirë ‘well’ as in këndon mirë ‘He or she sings well’. Any noun with a ‘genitive’ case ending can be preceded by one of these attributive articles, functioning







the substance of which something is made: buka e grurit (the wheat bread, or literally, ‘the bread of the wheat’) a quality that characterizes the referent: aktet e trimërisë (acts of bravery) the object of an action: shkelja e rregullit (the violation of the rule) the subject of an action: ardhja e miqve (the arrival of friends) the whole of which the referent is part: gjysma e bukës (half of the loaf) the identifying element of the referent: muaji i korrikut (the month of July) a characterizing quality after the verb ‘to be’: është i gojës (he is articulate, or literally, ‘is of the mouth’) the universe of a superlative: më i madhi i djemve (the biggest of the boys)

Among the small national languages of Europe, Albanian can claim a special pride in the consistency of its standard orthography and grammar, and the richness of its vocabulary. References Camaj, Martin. 1984. Albanian grammar. Wiesbaden: Otto Harrassowitz. Hamp, Eric. 1972. Albanian. Linguistics in Western Europe. Current Trends in Linguistics, Vol. 9, ed. by Thomas A. Sebeok. The Hague: Mouton. Hubbard, Phillip L. 1985. The syntax of the Albanian verb complex. New York & London: Garland Publishing. Huld, Martin E. 1984. Basic Albanian etymologies. Columbus, OH: Slavica Publishers. Newmark, Leonard. 1957. Structural grammar of Albanian. Bloomington: Indiana University Publications of the Research Center in Anthropology, Folklore, and Linguistics. Newmark, Leonard. 1980. Spoken Albanian. Ithaca, NY: Spoken Language Services; revised edition, 1997. Newmark, Leonard. 1998. Oxford Albanian–English dictionary. Oxford: Oxford University Press. Newmark, Leonard, Philip Hubbard, and Peter Prifti. 1982. Standard Albanian: a reference grammar for students. Stanford: Stanford University Press. Newmark, Leonard, and Vladimir Dervishi. 1999. Albanian handbook with English and Albanian glossaries. Kensington, MD: Dunwoody Press. Zymberi, Isa. 1991. Colloquial Albanian. London and New York: Routledge.

LEONARD NEWMARK See also Indo-European 1: Overview

33

ALGERIA

Algeria Algeria, officially the Democratic and Popular Republic of Algeria, is the second largest country in Africa with 919,595 square miles after Sudan, stretching from the shores of the Mediterranean to the Sahara Desert. It has a population of 31,736,053 (2001 estimate) and is divided into 48 provinces called wilayat and 700 local communes. The population consists of Arabs, Berbers (Kabyles being the largest group followed by Chaouia) and people of mixed Arab and Berber ancestry. Arabs form 80–83%, Berbers 16–20%, and others 1%. Linguistic independence came rather late to Algeria. During the Muslim expansion of the seventh and eighth centuries, the Arabic language spread from the Arabian Peninsula to North Africa, supplanting the local languages. During the Ottoman Empire, the Arabic language was suppressed, as elsewhere in the Arab World. Later, the French conquest of Algeria in 1830 further complicated the linguistic situation. The authorities introduced French as the sole official language. Their main goal was to replace the Arabic culture by a more ‘civilized’ culture. The use of Arabic was generally limited to a small minority of religious scholars, and to the rural regions, where Arabic religious schools were set up to teach the Qur’an. After gaining its independence from France in 1962, the Algerian government declared Arabic as the official language, which was a difficult task as very few Algerians were actually literate in Arabic. Teachers were imported from Arab countries such as Egypt and Syria. And in the 1970s, the Arabization policy required that street and shop signs be in Arabic, despite the fact that 60 % of the population could not write or read Arabic. Over the years, Algeria struggled between the pressure to eliminate any legacy from the French colonialism, to meet the demands of the Kabyles to adopt Tamazight as a second official language and to teach in schools, and to bear the costs of rapid Arabization. Arabic is held in high regard because it is the language of the Qur’an. With the spread of Islam, Arabic has become the most used language among the living Semitic languages. It is related to Hebrew, Aramaic, and Amharic. Subgroups of Arabic are Classical Arabic, Eastern Arabic, Western Arabic, and Maltese. The modernized version of Classical Arabic, which is referred to as Modern Standard Arabic, is the language that Arab countries use as their official language.

34

The linguistic situation in Algeria is complicated, and is a matter of intense discussion within the country itself. Algeria’s population diversity is reflected in the large array of languages spoken. Seventeen languages are believed to be spoken in Algeria, Arabic and, most recently, Tamazight, being the official languages. The majority of Algerians speak Arabic, followed by French and Berber. Modern Standard Arabic, a simpler version of Classical Arabic, is taught at schools and is used in formal meetings and in the media, but is not used for ordinary conversation. The Algerian dialect, known as Darja, is used in everyday life. And increasingly, the vernacular is being used in theater and in novels because it is believed to reflect the Algerian culture. After gaining independence, Algeria proceeded to ‘arabize’ education in order to rid the country of French influence. However, the fact that the country was trying to rediscover its roots while forcing its population into ‘modernity’ has pushed it to maintain French as a language of science and technology. Thus French became associated with high social and educational status. Arabization on the other hand was pursued for ideological reasons related to nationalism and religion. Despite the efforts to introduce Arabization to rid the country of the cultural traces left by the long French colonialism, Algerians continue to use French for formal and informal conversations. In fact, French is considered by many the ‘unofficial’ official language, as it is used in most formal administrative meetings, gatherings, and various other functions. Modern Standard Arabic is used only in formal education, by the media, and for written purposes. The policy that the Algerian government has adopted toward minority language groups has resulted in tension among the different language users. Arabization, the language policy aimed at expanding the role of Modern Standard Arabic to the detriment of French and local dialects and languages, led to dissatisfaction and resentment among the Berber ethnic minority and the French educated population. Opponents of this policy argued against making Arabic the exclusive language of instruction in public schools when French remains the language of science instruction at the universities. Most Algerians speak Algerian Arabic, with educated and clerical masses speaking Modern Standard Arabic in formal situations. Arabs usually do not

ALGERIA speak Berber because it is not taught in schools. However, most Berbers speak Arabic and French, especially the younger generation. Currently, there is an active movement to regain the Berber language and teach it in schools. As of recently, Tamazight is being taught in the Kabyle major cities, but not in the nonBerber areas, although there are pockets of Berbers in every major city. Even though the Algerian census does not gather information about household languages, the number of Berber speakers has been estimated to be as high as 6,000,000 speakers. The Berbers are considered to be the original inhabitants of North Africa. They have lived in Algeria since 3000 BC, or, according to some historians, as early as 8000 BC. The word Berber comes from the Greeks, who used this term to refer to the people of North Africa. Although Berber uses an ancient script called tifhagh, it is primarily a spoken language. Currently, Roman script is used to write the Berber language. Most Berbers, especially the Kabyles, maintain their own language and a sense of ethnic identity. The Chaouia, however, tend to be more assimilated to Arab culture than the Kabyles. The Kabyles refer to themselves as Imazighen, and to their language as Tamazight. The linguistic division between Arabic speakers and Tamazight speakers stems mainly from the government policies that favor the Arabic language over the other local languages such as Tamazight. For years, the Algerian government suppressed the Berber languages, depriving those who did not speak, write, or read Arabic from basic human rights such as reading the newspaper and watching television. The Berbers, also known as Amazigh in North Africa, are the original inhabitants of North Africa. Because the Berbers have mingled with other ethnic groups over the centuries, especially the Arabs, their identification is purely linguistic and not racial. The Berber language (Tamazight) is primarily oral. Its written form is little known and rarely used. Like Arabic, the Berber language is a branch of the AfroAsiatic language family. Many languages derive from the Berber language family. Some of the Berber languages, the number of speakers, and the geographical areas where they are mostly spoken are given in table. As of 2003, 41 years after independence, French is still taught as a second language and is introduced at the second-grade level. French also remains the language of science instruction in all universities after a semi-unsuccessful attempt to arabize the educational system, including higher education. Attempts to substitute English for French as the second national language in the mid-to-late 1990s also failed. However, since 2003, English is being taught at the seventh-grade level, a year earlier than in previous years.

Berber Language

Number of Speakers

Geographical Areas

Kabyle

Up to 6,000,000 (1998)

In the Djudjura moun tain range, and along the northern central and eastern coastal region, east of Algiers Main cities: TiziOuzou, Dellys, Bejaia

Chaouia

1,400,000 (1993)

South and southeast of the Grand Kabylie region and south of Constantine, in the Aurès Mountains Main cities: Batna, Ain el Baidha, Ain Mlila

Chenoua

15,000–75,000 (1996)

Small towns east of Algiers

Tachelhit

Unknown

Southern Algeria near the Moroccan border

Tahaggart

25,000 (1987)

Southern Algeria in the Hoggar region Major cities: Djanet and Tamanrasset

Taznatit

40,000 (1995)

Around the city of Timimoun

Tumzabit

70,000 (1995)

Mzab region, 330 miles south of Algiers Main city: Ghardaia

Because the spoken Algerian Arabic is a simplification of Modern Standard Arabic, there is less distinction regarding number and gender. For example, the dual form and the feminine plural do not exist in the Algerian dialect. The Modern Standard Arabic masculine plural form is used in Algerian Arabic for dual, feminine and masculine, and for the feminine plural. Algerian Arabic consists of three major dialects spoken in three major cities: Algiers (center), Constantine (east), and Oran (west). These major dialects differ considerably from the dialects spoken by nomadic people in the Sahara.

Algerian Arabic vs. Modern Standard Arabic Phonological Differences Modern Standard Arabic is written from right to left and uses a root system, usually consisting of three consonants. For example, the English word ‘merge’ would be made up of the root ‘m-r-g’. In Arabic, the word ‘islam’ is made up of the root ‘s-l-m’. The Arabic script uses symbols for only consonants and

35

ALGERIA long vowels. The short vowels are not written and are represented by diacritics. However, although morphological and grammatical meaning is indicated by vowels, the vowel diacritics are usually omitted. The reader has to infer its particular meaning and pronunciation from the context. Even though the Algerian dialect is a variant of Arabic, its vocabulary differs considerably from Modern Standard Arabic. Many words originate from sources other than Modern Standard Arabic, such as Turkish, Spanish, Italian, and most of all Tamazight (Berber) and French. However, most of the vocabulary comes from Modern Standard Arabic with phonetic, phonological, and semantic changes. The examples below show Modern Standard Arabic words with Algerian Arabic correspondences. The elements in which the Algerian Arabic words differ from the Algerian Arabic ones are indicated in bold. Modern Standard Arabic Assimilation: taji:’u tazu:ru tadu:mu Sadr niSf

Algerian Arabic dji (she comes) dzu:r (she visits) ddu:m (it lasts—fem.) zdar (chest) nuSS (half)

pronunciation. For example, French has nasal vowels, but Arabic does not. Therefore, the nasal vowels in borrowed French words are denasalized (1). Also, French sounds that do not exist in Arabic are substituted by Arabic sounds (2). As the Algerian dialect relies heavily on emphatism, loan words are no exception (3). Arabic has three short vowels, a, i, and u, and three corresponding long vowels, a:, i:, and u:. As a result, French vowels that do not exist in Arabic are replaced by one of the three vowels (short or long) (4). French influence extends to syntax as well (5). French (1) bouchon (2) il roule

une serviette une savate (3) il sonne une place (4) un bureau (5) une table

Metathesis: yartaεid laεana

yatrεid (he shivers) nεal (he cursed)

Dissimilation: finja:n ba:dhanja:n

finja:l badhanja:l

Phone Substitution: shajara

sajra (a tree)

Monophthongization: zawj Sayf zayt

zu:dj (two) Si:f (summer) zi:t (oil)

Semantic Changes: yudi:ru ydi:r (he directs) (he does) ‘ajlis ‘uqεud (sit down—when someone is standing up) tgaεεad ‘uqεud (sit down—when someone is lying down)

Algerian Arabic bushu:n (cork) yru:li (he wanders) The French r is replaced by the Arabic r. serbita (towel) The sound v is replaced by b. Sabba:t (a shoe) ySu:ni (he rings) The French s is replaced by the emphatic S. bla:Sa (a place) bi:ru (a desk, an office) The French ü is replaced by i:. Tabla (a table) The Arabic feminine indicator a is added.

The linguistic situation in Algeria depends on the future political situation. The Algerian people believe that if the Muslim fundamentalists come to power, the trend will be toward arabization of all sectors of life; on the other hand, if democratization is successful, bilingualism will be stressed with the French language used for the technical and hard sciences, and Arabic for social sciences and civil services. After years of political unrest, the political and cultural situations are far from being resolved. While a solution is being sought for the linguistic conflict, Modern Standard Arabic remains a foreign language for some of the older Algerians, especially the Berbers. However, recent attempts to make Tamazight a second national language have been somewhat successful. In 2002, President Bouteflika announced that the Tamazight language would become an official language in Algeria alongside Arabic. References

The French Influence The Algerian dialect contains a good number of borrowed French words. However, the borrowed vocabulary is subject to Arabic rules of grammar and

36

Benremouga, Karima. 1987. A survey of a typical eastern dialect of Algerian Arabic: phonology and morphology. Master’s Thesis. Lawrence, KS: University of Kansas. Benremouga, Karima. 1997. Spoken Algerian ‘Arabic’: An overview. [Ed. by In Xingzhong Li, Luis, López,

ALTAIC and Tom, Stroik, Papers from the 1997 Mid-America Linguistics Conference.] Columbia, MO: University of Missouri. Grimes, Barbara F. (ed.) 1992. Ethnologue, languages of the world. Dallas, TX: Summer Institute of Linguistics. Harik, Elsa M., and Donald G. Schilling. 1984. The political of education in colonial Algeria and Kenya. Athens, OH: Ohio University Center for International Studies. Ruhlen, M. 1987. A guide to the world’s languages. Vol. 1: classification. London: Edward Arnold.

Saad, Zohra. 1993. Language planning and policy attitudes: a case study of Arabization in Algeria. Doctoral Dissertation. New York: Columbia University. Talmoudi, Fathi. 1984. The diglossic situation in North Africa: a study of classical Arabic/dialectal Arabic diglossia with sample text in mixed Arabic. Göteborg, Sweden: ACTA Universitas Gothoburgensis.

KARIMA BENREMOUGA See also Arabic; French Language

Altaic Altaic, in its broadest conception, is a family of languages composed of four subfamilies (JapaneseOkinawan, Mongolic, Tungusic, and Turkic) and two language isolates (Ainu and Korean). However, the status of this family remains a matter of controversy. On the one hand, some scholars who accept the genetic association of Mongolic, Tungusic, and Turkic (hereafter, Altaic proper) do not agree that there is a demonstrable relationship between Altaic proper and Japanese-Okinawan, Korean, or Ainu. On the other hand, some scholars question even the notion of Altaic proper. The skepticism often rests on purely methodological grounds; that is, although they accept the likelihood that Mongolic, Tungusic, and Turkic languages all stem from a common ancestor, some linguists doubt that enough evidence exists to prove this connection conclusively. In other instances, linguists reject even the likelihood of Altaic proper; they argue that the similarities among these language families are better explained by borrowing and contact-induced changes than by a genetic association. Ainu is spoken on Hokkaido, the northernmost island of Japan, and on Sakhalin Island, which is part of the Russian Federation. Ainu is spoken only as a second language, primarily by older people. Only about 1% of the 18,000 to 25,000 ethnic Ainus claim an ability to speak the language, and the number of fluent speakers is probably a fraction of this small number. Ainu, as is typical of Altaic languages, has a basic word order of subject–object–verb, but the connection with other Altaic languages is tenuous at best. Such a connection was proposed by John Street (1962) and was argued more forcefully by James Patrie (1982), but few accept Ainu as being part of the Altaic family. Japanese-Okinawan, a subfamily of Altaic, consists of two branches. The first branch consists solely of Japanese, which is spoken by approximately 130 mil-

lion people worldwide. The second branch, called Okinawan or Ryukyuan, consists of a set of ten languages, most of which are mutually intelligible to some degree. There are roughly 900,000 speakers of Okinawan languages, although very few of the languages are being transmitted to children at this point. Japanese and the Okinawan languages share roughly 70% of their vocabulary; they are not mutually intelligible. The earliest written records of Japanese date back to the early eighth century and are composed in Chinese characters. Since then, two different syllabic writing systems derived from Chinese characters have been developed for Japanese: Hiragana and Katakana. Both are syllabic; i.e. each symbol stands for a syllable. Additionally, there is now a phonetic writing system, Romaji, which uses the Roman alphabet. Here, the symbols stand for individual sounds. Japanese is unlike other Altaic languages in that it possesses a very simple syllable structure; most syllables consist of just a vowel or a consonant followed by a vowel. Also, somewhat unusually among the Altaic languages, Japanese has only five vowels and a small number of consonants. As in other Altaic languages, the basic word order is subject–object–verb, and it uses suffixes to indicate the grammatical function of a noun, i.e. whether the noun is the subject (-ga) or an object (-o, -ni). Perhaps the most careful attempt to link Japanese to Altaic proper and to Korean was that presented by Roy Andrew Miller (1971). Despite his efforts, few scholars readily accept the Japanese connection with Altaic, although the affiliation between Japanese and Korean has gained a wider consensus. Korean is spoken by 65 million people worldwide, the vast majority of them living in North and South Korea. The earliest examples of written Korean used Chinese characters, a practice that continued into the

37

ALTAIC twentieth century. However, a Korean alphabet, called Hangul, was invented in the fifteenth century. Hangul is now the primary means by which Korean is written. Chinese has had an obvious effect on Korean. In fact, more than half the Korean words have been borrowed from Chinese. Even so, the structure of Korean remains quite distinct from Chinese, and superficially at least, it shares many properties with Altaic proper. For example, the basic constituent order is subject– object–verb, it has no definite articles, and it exhibits ‘vowel harmony’; i.e. vowels within a word tend to assimilate to one another. A family relationship between Korean and Altaic proper was advanced by two leading Altaicists in the twentieth century: G.J. Ramstedt (1952–1957) and Nicolas Poppe (1965). The connection between Korean and Japanese has received more attention in the intervening years, sometimes being examined within the context of Altaic (for example, by Roy Andrew Miller [1972]) and sometimes independently of Altaic (for example, by Samuel Martin [1966, 1991]). Mongolic languages are spoken in the central Asian regions of the People’s Republic of China, the Russian Federation, and the Republic of Mongolia. Although there has been much debate regarding the internal structure of the family, it is perhaps most often depicted as having two primary branches. Western Mongolic consists of a single language called Moghul. The language, which is mutually unintelligible with all other Mongolic languages, is spoken by fewer than 200 people found in two villages in Afghanistan. Eastern Mongolic is broken down further into three subbranches: Dagur (which consists solely of the Dagur language, spoken by around 60,000 people in China), Monghuor (a set of closely related languages spoken in China, all of which have been deeply influenced by Chinese and, in some cases, Tibetan languages), and Oirat-Khalka (the subbranch containing most Mongolic languages). Within the Oirat-Khalka branch, the three most widely spoken languages are Oirat, spoken by approximately 300,000 people in the Kalmyk Autonomous Region in Russia; Buriat, spoken by 350,000 people in the Buriat-Mongol Autonomous Region in Russia; and Khalka, the official language of the Mongolian Republic, with more than 2.3 million speakers. It is this last language that most people refer to when they talk about Mongolian. The oldest written records of Mongolian date back to the early thirteenth century and take the form of inscriptions, but the literary text known as The Secret History of the Mongols had appeared already in 1240. The script used for this early writing was borrowed from the Uighurs, a neighboring Turkic group that had been using the script for many centuries.

38

Like the other Altaic proper families, Mongolic languages typically use vowel harmony. The basic word order tends to be subject–object–verb, the grammatical function of nouns is signaled by suffixes, and there are no definite articles (“the”). Tungusic languages are spoken over a geographic expanse that goes all the way from the Arctic regions of the Russian Federation to China’s Xinjiang Province. Most of these languages were traditionally spoken by small nomadic groups. As nomadism has come to an end in China and Russia, Tungusic speakers have Russified and Sinicized rather rapidly, and most Tungusic languages will probably start to die out in the near future. Tungusic has traditionally been divided into two branches: Northern Tungusic and Southern Tungusic. The best-known member of the Southern branch is Manchu, which was the official language of China’s last dynastic empire. Although the language was used for official purposes well into the twentieth century, very few fluent speakers remain. Sibe, another Southern Tungusic language, which, in fact, might be better labeled a dialect of Manchu, has 27,000 speakers in Xinjiang Province, making it the largest Tungusic language. The Northern branch of Tungusic consists of Even and a cluster of very closely related languages— Evenki, Negidal, Solon, and Orochen. Many scholars suggest that some of the Southern languages should be set off into a third branch of the family, Central Tungusic. The Central Tungusic languages, for example, Nanai, tend to be similar to Southern Tungusic in their pronunciation but similar to Northern Tungusic in the ways in which words are formed. Only two Tungusic languages have written records that extend back before the twentieth century: Manchu and Jurchen. The Jurchen data date to the twelfth century, but its writing system mixes logographic symbols, which stand for concepts, with symbols for pronunciation, and it is still not completely deciphered. The Manchu script was derived from Mongolic script at the end of the sixteenth century. Today, we have a huge body of Manchu literature spanning the seventeenth to the twentieth centuries. Tungusic languages are characteristically Altaic in having vowel harmony, case suffixes, and possessive suffixes. Also in line with other Altaic languages, they lack definite articles and have a weak distinction between nouns and adjectives. The Turkic language family covers a large area from the Balkans and Asia Minor in the southeast to the Central Asian regions of both the Russian Federation and the People’s Republic of China and up into the Arctic region of Russia. Turkic is split into two main branches: Bulgar and Common Turkic. The main language in the Bulgar branch is Chuvash, which is

ALTAIC spoken by about 1.5 million people in the Chuvash Autonomous Region of Russia. Chuvash, although grammatically similar to the rest of the Turkic languages, is marked by a series of historical changes in pronunciation that are hard to interpret in the context of Turkic and Altaic. Therefore, some scholars have suggested that Chuvash is actually a sister to proto-Turkic rather than a daughter. That is, instead of a branch of the Turkic family per se, Chuvash and the Turkic family may both be branches of a larger family unit. Common Turkic is generally divided into four or five subbranches, most of which have languages spoken by more than a million people. The Eastern subbranch includes, among other languages, Uzbek (the national language of Uzbekistan, 15 million speakers) and Uighur (6.7 million speakers, mostly in China). The Southern subbranch includes Turkish (50 million speakers), Azeri (14 million speakers, half in Iran and half in Azerbaijan), and Turkmen (3 million speakers). The Northern subbranch includes Tuvan (200,000 speakers, mostly in Russian but also in Mongolia) and Yakut (300,000 speakers). The Western subbranch constitutes Bashkir (1 million speakers) and Tatar (5.5 million speakers). Many scholars separate out of Western Turkic a fifth subbranch called Central Turkic. Within Central Turkic, one finds the national languages of Kazakhstan, Kazakh (7.6 million speakers), and Kirghiz (2 million speakers). The earliest written record of a Turkic language was found in a set of inscriptions in Mongolia dating to the early eighth century. These inscriptions were written in Orkhon Turkic and use what is called a runiform script. Some scholars have argued that the script is an invention of Orkhon Turkic speakers, although there is growing consensus that the script is actually of Semitic origin. Turkic languages are characterized by vowel harmony. Typically, Turkic languages manifest a subject–object–verb word order and heavy use of suffixes. Nouns and adjectives are only weakly distinguished. Speculation about some kind of genetic relationship among Mongolic, Tungusic, and Turkic stems back to the eighteenth century, and suggestions about the link between these languages and Korean and Japanese arose in the early nineteenth century. Despite a steady stream of scholarship with respect to the classification of these languages over the past 200 years, there is still widespread disagreement about the merits of an Altaic language family and, even granting its existence, which subfamilies should be contained therein. With respect to Altaic, no one denies that there are a host of putative cognates, i.e. words that are quite obviously related. The issue is whether these cognates reflect shared ancestry or whether they have arisen through extensive borrowing among the languages.

Certainly, there is plenty of evidence that Altaic languages have been in contact for many centuries, and everyone involved in the Altaic debate agrees that much borrowing has taken place. The point of contention is whether these cognates should be dismissed en masse as borrowings or whether a subset of the cognates does in fact reflect a family association. Adding fuel to the notion that putative Altaic cognates are not good indicators of a family relationship is the fact that among the basic vocabulary in the languages, for example, the lower numerals, body parts, and natural phenomena, the cognate evidence is not compelling. Typically, one expects better cognate evidence in these vocabulary realms. One of the strongest pieces of evidence in favor of Altaic, or at least Altaic proper, is the first-person singular pronoun (‘I’). In Tungusic languages, the nominative form of the pronoun is bi, but for all other cases, the stem is min-; e.g. Oroqen min-ŋi, ‘my’, is the genitive (possessive) first-person singular. Precisely the same contrast in stems is found widely in Mongolic. Hence, Buriat has bi, ‘I’, for the nominative first-person singular and min-ni, ‘my’, for the genitive. Although the correspondences in Turkic are not perfect, one does find similar patterns in some modern Turkic languages (e.g. Chuvash), and Old Turkic allowed variability in the pronominal stems; this, if not fully corresponding to the Tungusic and Mongolic patterns, shows enough similarity to suggest a connection among all three families. References Binnick, Robert I. 1987. On the classification of the Mongolian languages. Central Asiatic Journal 42. 178–95. Johanson, Lars, and Éva Ágnes Csató, (eds.) 1998. The Turkic languages. London and New York: Routledge. Kornfilt, Jaklin. 1987. Turkish and the Turkic languages. The world’s major languages, ed. by Bernard Comrie. New York and Oxford: Oxford University Press. Martin, Samuel E. 1966. Lexical evidence relating Korean to Japanese. Language 42. 185–251. Martin, Samuel E. 1991. Morphological clues to the relationships of Japanese and Korean. Patterns of change, change of patterns: linguistic change and reconstruction methodology, ed. by Philip Baldi. New York: Mouton de Gruyter. Miller, Roy Andrew. 1971. Japanese and other Altaic languages. Chicago: University of Chicago. Patrie, James. 1982. The genetic relationship of the Ainu language. Honolulu: University of Hawaii Press. Poppe, Nicholas. 1965. Introduction to Altaic linguistics. Wiesbaden: Otto Harrasowitz. Ramstedt, G.J. [1952–1957.] Einführung in die altaische Sprachwissenschaft (Introduction to Altaic linguistics), 2 vols. Helsinki. Street, John 1962. Review of Vergleichende Grammatik der altaischen Sprachen by Nicholas Poppe. Language 38. 92–8.

LINDSAY WHALEY See also Ainu; Chinese (Mandarin); Japanese; Korean; Tungusic

39

AMBIGUITY

Ambiguity ‘Ambiguity’ is the term for the existence of at least two separate, incompatible interpretations of a stretch of speech. Two kinds of ambiguity are recognized: if it is due to words, it is called ‘lexical’ ambiguity; if it is due to sentence structure, it is called ‘structural’ ambiguity. Lexical ambiguity can be either ‘polysemy’ or ‘homonymy’. Homonyms are distinct units in the vocabulary with identical pronunciation (‘homophones’) and/or spelling (‘homographs’). Each of the homonyms in a set has its own meaning, for example, bank ‘financial institution’ vs. bank ‘slope’. In this example, homophony and homography co-occur. But this need not always be the case; compare peak [pi:k] ‘summit’ and peek [pi:k] ‘glance’, which are homophones but not homographs. From a historical point of view, homonyms cannot be traced back to a common etymological origin, as with bat: the word for the flying mammal has a Swedish origin, the word for the implement is related to battle. From the contemporary viewpoint, homonyms have no shared meaning either. These two viewpoints are not watertight, however. A very competent speaker knows more about the meaning of a given pair of homonyms than a less competent speaker—but whose competence should be taken as the yardstick? Nor is it clear at which point etymological reconstruction should come to an end. Multiple meanings distributed over several words with the same pronunciation is called ‘homonymy’; multiple meanings within a single word is called ‘polysemy’. The different meanings of a polysemous expression have a base meaning in common. Furthermore, the meanings of a polysemous term are often related by metaphor or metonymy, as in point: ‘punctuation mark’, ‘sharp end’, ‘detail, argument’, etc. Here, we observe several meanings associated with one word, the shared base meaning of which could be something like ‘smallest unit’—either a concrete one (‘punctuation mark’, ‘sharp end’) or a metaphorical one (‘detail, argument’). There are cases of idiosyncratic polysemy, such as green ‘a certain color’ and ‘inexperienced’, but there are also cases of systematic polysemy, as in the actual/dispositional distinction (as with fast in this is a fast car) or the distinction between a building and the institution housed in it (as with school). Structural ambiguity has two subcases as well: ‘attachment’ ambiguity and ‘scope’ ambiguity. Attachment ambiguity refers to the possibility that the same sequence of words may be assigned different structures. In The policeman observes the lady with

40

the telescope, the prepositional phrase with the telescope modifies either the lady (thus, the lady is a lady with a telescope) or observes (thus, the policeman observes by means of a telescope). The ambiguity of the sentence is an attachment ambiguity—the prepositional phrase may be ‘attached’ to or associated with different elements in the sentence. Scope ambiguity refers to the possibility of assuming different logical forms of a sentence. An example is the sentence Every man loves a woman, which has two distinct readings: for each man there is ‘his’ woman, and he loves her; or alternatively there is a specific woman who is loved by all the men. With the first reading, every man ‘has scope over’ a woman, i.e. the sentence is primarily about ‘every man’. With the second reading, it is the other way round, i.e. the sentence is primarily about ‘a woman’. Ambiguity, whether lexical or structural, must be distinguished from vagueness. A vague expression is imprecise, whereas an ambiguous expression has several precise meanings. An example of a vague expression is the predicate red in Mary owns a red skirt. A dark pink or a dark orange skirt would be borderline cases for this sentence, due to the intrinsic vagueness of red. Vagueness is ‘intrinsic’ in the sense that it has nothing to do with lack of knowledge (we know what the color red looks like). There is a close connection between vagueness and context dependence. The existence of borderline cases is not necessarily a bad thing, as different usages may be salient in different situations; cf. Mary owns an expensive skirt. Depending on the financial situations of Mary and the listeners, the vague predicate expensive will have quite different readings. In sum, vagueness may be defined as follows: a sentence is vague if—despite the knowledge of all the circumstances in a given situation—one cannot determine with certainty whether it is true or false. An expression is vague if it occurs in a sentence in such a way that it is responsible for the sentence’s vagueness. Vague sentences violate the principle of classical logic demanding that a sentence be either true or false. Thus, either there is a third truth value (or many) besides ‘true’ and ‘false’—this is assumed in the many-valued logic approach (three-valued logics and fuzzy logics). Or there are gaps in truth-value assignment—this is the thesis of supervaluation semantics. In sum, vagueness and ambiguity have different communicative statuses. Vagueness is part of the common understanding of the speaker and hearer, accepted

AMERICAN SIGN LANGUAGE by both. With ambiguity, the situation is different. Usually, the speaker has a certain meaning in mind— there is no vague reading of an ambiguous sentence. Instead, any vague reading of an ambiguous sentence is a communicative breakdown, as ambiguous sentences are always in need of disambiguation. References Ballmer, Th.T., and M. Pinkal (eds.) 1983. Approaching vagueness. Amsterdam: North-Holland.

Lyons, J. 1977. Semantics. Cambridge: Cambridge University Press. Pinkal, M. 1991. Vagheit und Ambiguität. Semantics, ed. by A.v. Stechow and Dieter Wunderlich. Berlin and New York: de Gruyter. Rescher, N. 1969. Many-valued logic. New York: McGrawHill. Williamson, T. 1993. Vagueness. London: Routledge. Zwicky, A., and J. Sadock. 1975. Ambiguity tests and how to fail them. Syntax and semantics, Vol. 4, ed. by J. Kimball. New York: Academic Press.

MONIKA RATHERT

American Sign Language American Sign Language (ASL) is a naturally occurring language with a syntax and culture distinct from spoken English. Geopolitically, ASL covers a large territory, serving between 100,000 and 500,000 signers in both Canada and the United States. ASL is not a ‘universal sign language’, nor does it have close etymological ties with British Sign Language, another distinct signed language. Instead, French Sign Language, or langue des signes française (LSF), has historically given ASL its closest linguistic ties. Throughout early history deaf people lived isolated lives, possibly communicating with their hearing parents and siblings through the use of idiosyncratically created home signs. As world populations grew, larger metropolitan groups began to provide the contact for visual gestural dialects to emerge. In the late 1760s, the French priest Charles-Michel de l’Epée, concerned for the religious salvation of deaf Parisians, gathered a small group of deaf children together and began teaching them through the use of an invented manual French that he created as an educational counterpart to spoken French. This gathering in Paris is now recognized to be the first public school for the deaf in the world. The home signs and various Parisian dialects that the deaf children brought to their school merged with de l’Epée’s ‘methodological’ signs and subsequently evolved into the language now known as LSF. The reputation of the school and the emerging language flourished. In the early nineteenth century, a Protestant minister Thomas Hopkins Gallaudet was sent by wealthy American philanthropists to Europe to investigate a school for the deaf in London. Unable to obtain any assistance from England, Gallaudet visited

France and eventually persuaded Laurent Clerc, one of the young deaf teachers from de l’Epée’s school, to travel back to America with him. Together, they established the first American School for the Deaf in Hartford, Connecticut. The French language taught by Clerc mingled with the indigenous dialects brought in by his young American students from across the developing country. Mingling with a couple of fledging signed languages found in several small established deaf communities such as Martha’s Vineyard in Massachusetts (descendants of British immigrants), and Henniker, New Hampshire, a new national signed language began to evolve. In 1864, a liberal arts college now known as Gallaudet University was established in Washington, DC through the efforts of Abraham Lincoln and Edward Miner Gallaudet, Thomas Hopkins’s son. Graduating deaf students, many of whom had previously attended Clerc’s Connecticut school, returned to their native homes and frequently established or became responsible for establishing schools for deaf children in their respective states. The unique language mixture of Clerc’s LSF signs and indigenous signs brought in or naturally created by Gallaudet students gradually evolved into modern ASL. Spoken languages can usually trace their history back through many centuries. Even including its historical French ties, ASL has only 250 years of documented language use. Yet, ASL is already following a process of grammatical evolution that spoken languages follow. Cross-linguistic studies of the historical relationships between LSF and ASL reveal paths of lexical and grammatical change taking place in this young signed language.

41

AMERICAN SIGN LANGUAGE One example can be found in the semantic changes that have occurred between the LSF nominal sign for ‘coin’ and a modern ASL discourse marker used to stop arguments or to change the subject (Wilcox 2000). The old LSF classifier noun for money, ARGENT (sign words are conventionally capitalized), semantically evolved into the sign for GIVE (donner), a verb in LSF that can also be found in modern ASL. The verb stem phonology of both ‘give’ signs has a common etymology, but diverge functionally. Following a path of lexicalization, one GIVE sign became an inflected verb of both literal and metaphorical nature; the other GIVE sign grammaticalized and extended into the word GIVE-IN-ARGUMENT. This verb is used metaphorically only and functions solely as a discourse marker. Typical paths of grammaticalization in spoken languages reveal that language evolves from the mapping of literal objects onto nouns (nominalization) to semantic shifts of abstraction. The original concrete reference of a French money sign became increasingly abstract as it evolved into the metaphorical discourse marker in ASL. This example, plus the documented grammaticalization of modals such as WILL and MUST, demonstrate that ASL is following a linguistic path of subjectivity taken by many spoken languages. ASL has an underlying s(ubject)–v(erb)–o(bject) sentence structure. In addition, it is strongly characterized by topic-comment structure. Topics act as discourse pivots to link previous information with the following and also indicates a shift in the point of view that the signer takes of the discourse events (Janzen et al. 1999). Topic marking in ASL is generally a marker with explicit discourse function. The following example also demonstrates the flexible word order of ASL: [KIP BUY NEW COMPUTER]-topic, CURIOUS ‘I wonder if Kip bought a new computer.’ Signed words consist of parts and can be broken down into components called parameters: the handshape, the location, the movement, and the orientation of the palm. These components exhibit both sequentiality and simultaneity when the signed words are put together, differing from spoken languages that are highly sequential. The iconic nature of ASL, with its strong link between form and meaning, is pervasive and is documented throughout its phonological and morphological analysis. The morphology of ASL verbs is especially complex, more so than spoken English. The verb root is often expressed in the handshape and grammatical

42

information is expressed in the movement, as Ronnie Wilbur describes: …in a multimorphemic verb sign, the movement can carry the information regarding the transfer of a themeobject (e.g. GIVE), the starting location can carry the source-agent-subject, the ending location can carry the goal-recipient-indirect-object, the handshape can carry the physical characteristics of the object itself (size and shape indicated by ‘handle’ classifier), and each piece of information is phonologically distinct from the others. (2000:216)

ASL has a large set of signs called classifiers that can be used to specify something without actually naming the specific referent. Classifiers can represent pronouns and describe in detail the size, shape, depth, or texture of an object, and they can function as a noun and indicate the location of that noun and its actions (Baker-Shenk and Cokely 1999). Classifiers are abundant in ASL and most signed languages. Another critical component of ASL is the use of nonmanual signs. Grammar can be produced and distributed across the face, the head, and the shoulders. Adverbial and adjectival information can be produced on the mouth, tongue, and cheeks. For example, one visual gesture, ‘th’, is made with a slack jaw and a protruding tongue pushing through slightly separated lips, and indicates carelessness or incorrectness. Other nonmanual markers such as the eyebrows, the nose, head movements, and eyegaze can coarticulate with signed words or be produced in isolation. These nonmanuals mark wh-questions, yes/no questions, rhetorical questions, conditionals, negations, topics, and relative clauses. Rhetorical questions, uncommonly used in spoken English, are found quite frequently in ASL conversations. A rhetorical question is accompanied by a brow raise and a tilting of the head and often involves the use of the signs WHY, WHAT, WHO, HOW, and REASON. Nonmanual facial expressions can also indicate mood and affect. Fingerspelling makes use of 26 handshape configurations that correspond to the English alphabet. Fingerspelling is more prevalent in ASL than in other signed languages of the world. It is primarily used to indicate proper nouns and technical terms. Fingerspelling is more than a sequence of canonical handshape-alphabet letter correspondence, however, since the articulatory movements of segments within the fingerspelled word influence each other. Forward (carry-over) and anticipatory assimilation affects the actual shaping of fingerspelled words, creating a fluid transition between letters that is prosodic and complex (Wilcox 1992). Language borrowing involving fingerspelling takes place through the physical proximity of English and

AMERICAN SIGN LANGUAGE ASL. Fingerspelled English words are often modified so that they fit better the structure of a sign (Battison 1978). For example, the word A-L-L can be fingerspelled letter by letter, or it can serve as a borrowed loan sign and will be restricted semantically in several ways. Instead of fingerspelling the three letters in the conical signing space in front of the body, a signer can hold up an imagined list with one hand while using the other hand to fingerspell the letters, literally running down the entire length of the arm to show ‘all’ of the items on the list. Or a twohanded loan sign can fingerspell an abbreviated A-L while sweeping the letters in front of the signer to designate everyone in the room including the signer. This particular lexicalized loan sign is highly productive and limits the semantic meaning according to specific modifications that are applied to the three letters. The culture of deaf people endures not through artifacts such as food or clothing, but through the powerful use of the language of ASL. Historically, the hearing world has not condoned the language choice of deaf people. A defining event in 1880 still exerts influence on deaf people and their use of signed languages. The representatives of the World Conference for the Deaf in Milan, Italy, voted unanimously (except the American delegates) to ban signed languages from all the schools in Europe and replace it with an ‘oral’ method of instruction (Lane 1984). Even in America, this edict had a profound effect on the educational systems. Deaf teachers were fired or prevented from educating younger deaf children, teachers who signed were not promoted or respected, older students who might ‘contaminate’ the younger children were hidden in separate quarters, and curricula were changed to speech-based activities. The most influential leader of the oral method of educating deaf children was Alexander Graham Bell, the inventor of the telephone. His great influence and wealth had a lasting impact upon the movement. Organizations that were founded by his proponents still work to oppress the use of ASL in the United States today. In the early 1900s, with schools no longer administered by deaf adults, the language was kept alive by the use of deaf children born of deaf people, a heritage unlike that found in spoken languages. Only 8–10% of deaf children are born to deaf parents who sign, which means that around 90% of the ASLusing deaf children learn it from their ‘native’ peers. The proponents of the oral method were relatively successful in keeping signed language out of the educational arena until the early 1960s when William C. Stokoe’s (1965) seminal work on linguistic analysis of ASL appeared. His pioneering work ignited an explosion of research on ASL that now permeates linguistic inquiry. Because of the tenacity of ASL users, along with the backing of research that provides linguistic legitimacy, ASL is now finding rewarding acceptance

in schools, universities, courtrooms, and social and business settings across America. The past 200 years have witnessed a strong expansion of ASL throughout the world due to religious or educational agendas. Also, thousands of deaf students from around the world have attended Gallaudet University (the world’s only liberal arts university for deaf people); thus, the signed language can be found throughout Asia, South America, Africa, and Europe — almost every country has deaf alumni who use ASL in addition to their own native signed language. Many deaf people have a working knowledge of ASL and the language is used freely at most of the international deaf conferences and workshops. However, its pervasive use of fingerspelling and marked initialization of English words has thus far prevented it from being accepted as an official signed lingua franca. Nevertheless, ASL users provide incentive to the world’s remaining deaf populations to maintain their cultural and linguistic identities as their respective signed languages advance and evolve. References Armstrong, David F., Michael A. Karchmer, and John Vickrey Van Cleve. 2002. The study of signed languages: essays in honor of William C. Stokoe. Washington, DC: Gallaudet University Press. Baker-Shenk, Charlotte, and Dennis Cokely. 1980. American sign language: a teacher’s resource text on grammar and culture. Washington, DC: Clerc Books, Gallaudet University Press. (fourth printing, 1999). Battison, Robbin. 1978. Lexical borrowing in American sign language. Silver Spring, MD: Linstok Press. Janzen, Terry, B. Shaffer, and Sherman Wilcox. 1999. Signed language pragmatics. Handbook of pragmatics, ed. by Jef Verschueren, Jan-Ola Ostman, Jan Blommaert, and Chris Bulcaen, 1–20. Amsterdam: John Benjamins Publishing Company. Klima, Edward S., and Ursula Bellugi. 1979. The signs of language. Cambridge, MA: Harvard University Press. Lane, Harlan. 1984. When the mind hears. New York, NY: Random House. Lane, Harlan, Robert Hoffmeister, and Ben Bahan. 1996. A journey into the deaf-world. San Diego, CA: DawnSignPress. Padden, Carol, and Tom Humphries. 1988. Deaf in America: voices from a culture. Cambridge, MA: Harvard University Press. Stokoe, William C., Dorothy Casterline, and Carl Croneberg. 1965. A dictionary of American sign language on linguistic principles. Washington, DC: Gallaudet College Press. Reprint, Silver Spring, MD: Linstok Press, 1976. Wilbur, Ronnie. Layering of nonmanuals in ASL. 2000. The signs of language revisited: an anthology to honor Ursula Bellugi and Edward Klima, ed. by Karen Emmorey and Harlan Lane, 215–44. Mahwah, N J: Lawrence Erlbaum Associates. Wilcox, Phyllis Perrin. 2000. Metaphor in American sign language. Washington, DC: Gallaudet University Press. Wilcox, Sherman. 1992. The phonetics of fingerspelling. Philadelphia, PA: John Benjamins Publishing Company.

PHYLLIS PERRIN WILCOX See also British Sign Language; Grammaticalization

43

AMHARIC AND ETHIOPIAN SEMITIC LANGUAGES

Amharic and Ethiopian Semitic Languages There are perhaps 11 Ethiopian and Eritrean Semitic (EES) languages now spoken, although the presence of dialect continua and lack of information on intelligibility between varieties makes a fully confident count difficult. Two other languages are known, but are now extinct. These 13 languages are (spelling of names of lesser known varieties varies in the literature): 1. Tigre, of western Eritrea, with dialects on the Red Sea coast and Dahlak islands; 2. Tigrinya, spoken in Eritrea and northern Ethiopia, but especially in Ethiopia’s Tigray region; 3. Ge‘ez, the ancient language known in epigraphy since about 500 BC, the language of the ancient kingdom of Axum centered on the modern Tigrayan city of the same name, unused as a mother tongue for many centuries, but continuing in regular use as the liturgical language of the Ethiopian Orthodox Christian Church; 4. Gafat, of the Gojjam region of central Ethiopia, extinct now for some 50 years; 5. Amharic, the world’s most populous Semitic language after Arabic, with over 16 million speakers in Ethiopia, and long that nation’s official language; 6. Argobba, near mutually intelligible neighbor to Amharic, in the northwestern Shewa region; 7. Soddo (also known as Kestane ‘Christian’), spoken just west of Addis Ababa, with a dialect Dobbi (or Gogot); 8. Mesqan, south of Soddo, whose linguistic separateness from neighboring varieties is least confident; 9. Chaha and mutually intelligible varieties including Ezha, Gomara, Gura, and Muher, spoken south of Mesqan territory; 10. Inor and mutually intelligible varieties including Endegenya, Enner, Gyeta, Magar, and Mesmes (the latter probably now extinct), southern neighbors of the Chaha group; 11. Silte and mutually intelligible varieties including Enneqor, Ulbareg, and Walane, southeastern neighbors of the Inor group; 12. Zay, about Lake Zway some 70 km south of Addis Ababa; and 13. Harari (Adare), spoken in and about the ancient eastern Ethiopian city of Harar.

44

Some 76 languages are spoken in Ethiopia and Eritrea, in addition to the 11 Semitic languages, some 21 Cushitic languages, 23 Omotic languages, 20 NiloSaharan languages, and one as yet unclassified, Ongata. The presence of so many Semitic, Cushitic, and Omotic languages, three of the six subgroups of the Afroasiatic family (the other three Egyptian, Berber, and Chadic), suggests that the region is one of very early Afroasiatic settlement. Amharic, with over 16 million mother tongue speakers and, according to the 1994 Ethiopian census, over five million second-language speakers, is by far the most important Ethiopian language, and the second most populous Semitic language after Arabic. Amharic is spoken throughout Ethiopia but particularly in urban areas, and is taught in all public schools. Regional varieties or dialects of Amharic are recognized in regions of Begemder, Gojjam, Wello, and Shewa. The Ethiopian capital city of Addis Ababa (in Shewa) is nowadays the focus of Ethiopian economic and cultural life, and Addis Ababa Amharic has naturally become the prestige dialect. The numbers of speakers of most Ethiopian languages may be estimated from counts of the 1994 Census of Ethiopia. The Census, however, like much of the linguistic literature, grouped as ‘Gurage’ the six languages numbered 7–12 in the list, and provided separate figures for only the Soddo and Silte varieties. Amharic has three times more speakers than Tigrinya, with perhaps five and a half million including two million in Eritrea. Tigre may have a million speakers, all in Eritrea, followed by Ethiopian Silte (800,000), Soddo (200,000), and Chaha (perhaps 100,000). Speakers of the others are much fewer in number. The census reported Harari to have 22,000 speakers, and Argobba only 10,000. The region is naturally characterized by bilingualism and multilingualism, often involving Amharic or Oromo, a Cushitic language (known in the old literature as ‘Galla’) perhaps as populous as Amharic, and more populous when Oromo speakers of Kenya are added. Research in the late 1960s showed, for example, that in Jimma town of western Ethiopia some 90% of speakers of languages other than Amharic knew Amharic, and 42% of Amharic speakers knew Oromo. Today, perhaps the majority of town and city-dwelling Ethiopians, except in largely Tigrinya-speaking Tigre province, are at least second-language speakers of

AMHARIC AND ETHIOPIAN SEMITIC LANGUAGES Amharic, and knowledge of the language certainly continues to spread throughout the region. After a largely Tigrayan army overthrew the 1991 derg (‘committee’) communist government in Addis Ababa, whose predecessors had deposed the Amhara emperor Haile Sellassie in 1974, Ethiopian political power lay in the hands of Tigrinya speakers, who have encouraged considerable linguistic freedom of expression and identity including, rather controversially, the use of local-majority languages in public education. Oromo, backed by a nationalist movement, became a growing rival to Amharic-language dominance. With its many second-language speakers and long literary development, however, Amharic has largely continued its dominance in literary culture, education, and public life. Written in Ge‘ez, royal chronicles and religious writings are known from the fourteenth century, and shortly after this time writings in Amharic also began to appear, but especially in the seventeenth century. Publication in Amharic has increased steadily, and Ethiopian publications in Amharic today include writings of all kinds: newspapers, literary and news magazines, drama, novels, history, textbooks, and poetry. Amharic-language magazines are published in Europe and the United States to serve the Ethiopian expatriate populations there. Publication in Tigrinya as well as in Eritrea and Ethiopia’s Tigray region has flourished. Amharic and Tigrinya, and often nowadays other EES languages, are written in the Ge‘ez or ‘Ethiopic’ writing system, an adaptation of the ancient South Arabian writing system, itself the southern adaptation of evolved Egyptian writing sometimes termed ‘Sinaitic’, whose northern adaptations include Aramaic, Arabic, and Hebrew writing (and Phoenician, perhaps that transformed by the Greeks into their alphabet). The epigraphic record suggests that during the fifth century reign of the Axumite king Ezana, Ethiopic writing was significantly modified to include representation of vowels, as regular extensions or shapings of its consonantal characters, approximately as early as in Indic Kharosthi writing. Probably since ancient times, Amharic has spread in territory earlier populated by speakers of other, especially Cushitic, languages, and as a result Amharic has come to share a number of features with these other languages. Amharic has long borrowed from Ge‘ez, a favored practice for the modern satisfaction of needs for philosophical, technical, and other new vocabulary. Borrowings from Italian entered the language in the Italian Colonial era, but today the principal source of new words is English, the language of Ethiopian and Eritrean secondary and higher education. The obvious Semiticness of Amharic may be suggested by a comparison of a selection of five basic words of Hebrew, Classical Arabic, and Amharic.

‘all’ Hebrew ko¯l Arabic kull Amharic hullu

‘die’ ma¯t ma¯ta motə

‘eye’ ayin ayn ayn

‘house’ bayit bayt bet

‘name’ še¯ m ism s m

While a different selection of words might show no similarities, and many similarities that exist are far from obvious, the systematic prevalence of such comparisons as these confirms that Hebrew, Arabic, and Amharic are all descendant languages of that spoken by the first Semitic peoples. Beyond lexical comparisons, Semitic languages, including EES, are particularly characterized by threeconsonant roots filled out by vowels and affixes in word derivation (so-called ‘root-and-pattern’ morphology), and other characteristics including the presence of a set of subject-agreeing verb prefixes including yfor third-person masculine ‘he’, and t- for both second-person ‘you’ and third-person (‘she’) feminine:

Hebrew Arabic Amharic

3m.sg. yi-kbad ya-kbud y -kbəd

3f.sg. ti-kbad ta-kbud t -kbəd

2m.sg. ti-kbad ta-kbud t -kbəd

The root (kbd heavy’) and the prefixes are cognate; a difference is that the Amharic verbs are subjunctives, vs. the present tense of the Hebrew and Arabic. Within the Semitic family of languages, Amharic and its kin in Eritrea and Ethiopia are typically grouped as ‘South Semitic’ with the ancient and modern Semitic languages of South Arabia, such as epigraphic Sabaen of Yemen and modern Soqotri of Soqotra island. The unity of EES is suggested by several features which they tend to share, including a nonfinite verb conjugation for all but the last of verbs in sequence (Amh. səbro ‘he having broken’), verbs expressed as compounds of particle + ‘say’ (z mm alə ‘he was quiet’), and a special verb of presence (Amh. alləhu ‘I am present’). EES languages are divided into two groups, North (Tigre, Tigrinya, and Ge‘ez) and South (Amharic and the rest), which differ in a number of characteristic features, of which two are the presence of so-called ‘broken’ noun plurals in the North (for example, Tigrinya sor ‘ox’, pl. ’aswar), formed by modification of root structure instead of affixation, and present-tense doubling of the second root consonant of triconsonantal verbs (compare Tigrinya y səbb r ‘he breaks’ vs. Amh. y səb r). The South EES languages are bifurcated as an eastern group of Amharic and Argobba plus the Siltegroup, Zay, and Harari, against a western group of the Chaha-and Inor-group languages plus Mesqan and Soddo. The eastern group typically augments presenttense main verbs with a suffixed auxiliary verb, as in

45

AMHARIC AND ETHIOPIAN SEMITIC LANGUAGES Amharic main verb y səbr-al ‘he breaks’ vs. subordinate y səbr. Argobba differs from Amharic in small but numerous ways, probably resulting in at least slight mutual nonintelligibility. The divergence of Amharic and Argobba may reasonably have resulted from the separation of these peoples upon the adoption of Islam by the Argobbas in about the sixteenth century. The typical divergence between pairs of EES languages is much greater, on the order suggested by 50% or so of cognates in a hundred words of core vocabulary between North and South EES languages, and about 60% between South EES languages. Ge‘ez, because it is the only anciently known EES language, has sometimes been considered the ancestor language from which all the others descend, as Latin is ancestor to the Romance or Italic languages, but this does not seem to be the case, because there are old linguistic features of the other languages unreasonably absent in Ge‘ez, pervasive Ge‘ez features unreasonably absent in the others, and because the diversity of the other languages seems too great to date only from the early Axumite times of likely Ge‘ez dispersal. Amharic, like other EES languages, constructs typical transitive sentences with the main constitutent order subject–object–verb (SOV), in contrast to the VSO typicality of Semitic Arabic and Hebrew. An Amharic topicalized object can precede a subject, but the verb is rigidly final. Amharic, however, has a number of wordorder characteristics inconsistent with the basic SOV type, including prepositions instead of postpositions (bə-bet ‘in house’, bə- ‘in’) and verb prefixal clause subordinators (s -tsəbr ‘when you break’, s - ‘when’). Typical SOV characteristics of Amharic and EES languages are the linear precedence of adjectives, genitives, and relative clauses before their modified nouns. The syntactic peculiarities of EES are typically attrib-

uted to the influence of Cushitic-language neighbors, which are mainly of the consistent SOV type, with postpositions and case suffixing absent in EES. Amharic and most other EES languages usually casemark only definite objects, Amharic with -n: w ša-w l K -u-n nəkkəsə ‘the dog bit the boy’. Amharic and other EES languages have the phonological characteristic of a series of glottalized ejective consonants, as in Amharic k’utt’a ‘anger’ (with ejectives k’ and t’). These differ from the so-called ‘emphatic’ velarized/pharyngealized cognate consonants of e.g. Arabic. Some EES languages preserve the Semitic pharyngeals  and ʕ. As in the other Semitic languages, the voiceless labial stops p and p’ are marginal, or limited to loan words such as ityopp’ya ‘Ethiopia’, from Greek. Like most of the EES languages, Amharic has the seven-vowel system i, e, a, , ə, u, o; others have a five-vowel system with length. References Bender, M. L., J. D. Bowen, R. L. Cooper, and C. A. Ferguson (eds.) 1976. Language in Ethiopia. London: Oxford University Press. Hetzron, Robert. 1972. Ethiopian Semitic: studies in classification (Journal of Semitic Studies monograph 2). Manchester: Manchester University Press. Hetzron, Robert (ed.) 1997. The Semitic languages. London: Routledge. Leslau, Wolf. 1978. English-Amharic context dictionary. Wiesbaden: Otto Harrassowitz. Leslau, Wolf. 1987. Comparative dictionary of Ge‘ez. Wiesbaden: Otto Harrassowitz. Leslau, Wolf. 1995. Reference grammar of Amharic. Wiesbaden: Otto Harrassowitz.

GROVER HUDSON See also Afroasiatic; Arabic; Hebrew: Biblical; Hebrew: Modern; Semitic Languages; Writing Systems

Analogical Change In a general sense, analogy is the correspondence of two or more entities with respect to certain properties. In linguistics, the term ‘analogy’ is most often used to describe processes in which a new linguistic object is created by aligning an existing linguistic object A with another existing linguistic object B, which serves as a model. A and B usually have some properties in common. For instance, Early Modern English pigs and kine

46

are both plural nouns but differ in the way the plural is formed: -s vs. -n + vowel change. The form pigs— respectively horses, stones, and the like—is considered the model here and the plural of cow is aligned with the plural of pig and thus, by analogy, kine becomes cows. Sound change and analogical change have already been considered by Neogrammarians as the fundamental processes of language change. The two important

ANALOGICAL CHANGE types of analogical change are analogical leveling and proportional analogy. Analogical leveling results in the reduction of allomorphy within paradigms. The paradigms of the Old English and Old High German verb meaning ‘to choose’ contain forms with a fricative [z] or [s], where other forms have an [r]. In both languages, the paradigms were leveled out by adopting the same consonant throughout the whole paradigm. While Modern English chose [z], Modern High German dropped the fricative in favor of [r]. Old Engl. ce-o[z]an ce-a[s] cu[r]on (ge-)co[r]en

> > > >

Mod. Engl. choo[z]e cho[z]e cho[z]e cho[z]en

PRES PAST-(SG) PAST-(PL) PAST-PART

The above example also illustrates the interaction of sound change and analogical change. The alternation of [s] and [z] results from a sound change referred to as ‘Verner’s Law’ and [r] is the result of a subsequent change from [z] to [r] in certain positions (‘rhotacism’). Sound change is conditioned by phonetic factors and is not sensitive to the needs of grammatical and semantic functioning. It often obliviates the principle ‘one form— one meaning’ (‘Humboldt’s universal’). Analogical leveling is always conditioned by nonphonetic factors and often ‘fixes the damage’ caused by sound change. In contrast to sound change, analogy has been considered to be an unsystematic process. This has given rise to the formulation of ‘Sturtevant’s Paradox’: sound change is regular but creates irregularity, whereas analogy is irregular but creates regularity. However, in some cases analogy is fairly systematic. For instance, [z]–[r] alternations of the above type have been leveled out in English quite systematically, as within the paradigms of loose, freeze, and rise. There are only very few cases where the [r] has been retained (e.g. was vs. were). By proportional analogy—also called ‘analogical extension’ or ‘four-part analogy’—an already existing morpheme or relation becomes generalized to other linguistic forms. In most cases, proportional analogy creates or assimilates derived forms on the basis of another morphological pattern. Proportional analogy proceeds like the solution of an equation of the form: A : B stone : stones

= C : X = cow : X

solve for X X = cows (< kine)

The solution of this equation with respect to the phenomenon mentioned above yields the replacement of kine by cows as the plural form of cow. As with analogical leveling, proportional analogy is not fully systematic: the analogy did not carry over to words like foot: feet / *foots.

Other less productive forms of proportional analogy are backformation and hypercorrection. With backformation, it is the morphologically less complex item that is created. In Early Modern English, pease was a mass noun, which ended in a sequence that resembled a plural morpheme. Its reanalysis led to the creation of a singular form pea by dropping the final fricative. Backformation also occurs as a word-formation process.

Old High Germ. kiu[s]an ko[s] ku[r]un (gi-)ko[r]an

bean

> > > >

Mod.H.G. kü[r]en ko[r] ko[r]en geko[r]en

: beans [-z] = X

depression : depress

: pease [-z] (X = pea) = aggression : X (X = aggress)

Hypercorrection consists in imposing a pattern that relates a colloquial low-prestige variant and a standard higher-prestige variant onto another word presumed to be low prestige. Thereby, an allegedly high-prestige variant of the word is created: fella’ : fellow

= umbrella : X (X = umbrellow)

Other phenomena that have been discussed under the heading of ‘analogy’ are blending, contamination, reanalysis, and folk etymology. Blending or ‘portmanteau formation’ consists in combining parts of existing words where these parts are usually not morphemes. Examples are brunch from breakfast and lunch or infotainment from information and entertainment. In the case of contamination, a similar fusion of two words occurs but the resulting expression retains the meaning of one of the source words. For example, the [t] of Proto-Indo-European pəte-r should have turned up as [d] by Verner’s Law and a subsequent sound change, yielding fader. Instead, contaminated by the regularly derived brother, we get father. In reanalysis, a deviant structure is assigned to an expression. The words napkin and apron both derive from French nape ‘cloth’. In the second case, the Middle English sequence that results from the combination of the word with an indefinite article anapron was reanalyzed as an apron, interpreting n as part of the article. A related phenomenon is folk etymology. Sometimes, words change on the basis of what a speaker assumes is the etymological origin of the word. For instance, French crévisse should occur in English as something like crevisse, but appears as

47

ANALOGICAL CHANGE crayfish because the second part got misinterpreted as referring to some kind of fish or aquatic creature. Theoretical approaches to analogical change manifest themselves in attempts to construct a theory of analogical processes, which is explicit and restrictive and unifies the different forms of analogy (e.g. within generative phonology; cf. Kiparsky 1982). Some researchers formulated a number of principles (sometimes called ‘laws’) in order to explain why and how analogical changes occur (e.g. KuryAowicz 1949, Man´czak 1980). These principles are usually subject to counterexamples and should therefore rather be understood as tendencies (cf. McMahon 1994). Among these principles were the following: (i) Analogy proceeds from a basic to a derived form or from a shorter to a longer form. Counterexample: backformations. (ii) More overt and complex markers are favored by analogy. For example, the Old High German plural forms of Gast ‘guest’ and Topf ‘pan / pot’ are Gästi and Topfa. By analogy Topfa adapted to the form Gästi, which has the more complex plural marking (umlaut and suffix) yielding Gäste and Töpfe in Modern High German. Counterexample: Old English lang : leng-ra turned into long : long-er analogous to warm : warm-er, not into long : *leng-er. (iii) In processes of proportional analogy, it is usually the more productive pattern that serves as a model. Counterexample: extension of strong verb forms to weak ones as in the change from dived to dove as past tense of dive analogous to ride : rode. (iv) Allomorphy gets reduced by analogy (in particular by analogical leveling). Counterexample: the change of the stem Topf- to Töpf- in German mentioned above. (v) When a new analogical formation becomes accepted (e.g. brother) and the older form also survives (brethren), the analogical form takes over the basic meaning. Counterexample: blendings. Analogy has also been considered an important tool in other branches of linguistics like theories on

language acquisition, syntax, and orthography: (i) Analogy is assumed to be at work when children overregularize morphological forms, like goed instead of went. (ii) An example of syntactic analogy is the change of impersonal constructions like me hungreth, me thinks to the more widespread personal constructions, i.e. to I hunger, I think. (iii) Modern High German lieb ‘kind / nice’ orthographically reflects a Middle High German diphthong, which has been monophthongized to [i:] in Modern High German. Words like Wiese ‘lawn’ adopted this spelling by proportional analogy, although the Middle High German spelling and pronouncation wise showed a monophthong. References Antilla, Raimo, and Warren A. Brewer. 1977. Analogy: A basic bibliography. Amsterdam: Benjamins. Hock, Hans Heinrich. 1991. Principles of historical linguistics, 2nd revised and updated edition. Berlin, New York: de Gruyter. Hock, Hans Heinrich, and Brian D. Joseph. 1996. Language history, language change, and language relationship. An introduction to historical and comparative linguistics. Berlin, New York: Mouton de Gruyter. Kiparsky, Paul. 1982. Explanations in phonology. Holland: Dordrecht / Foris: Cinnaminson. KuryAowicz, J. 1949. La nature des procès dits >analogiques form’. In general, Chafe is unique among the American linguists of his generation in never being tempted to accept the generative Chomskyan approach. This orientation toward semantics was also the chief element of his well-known book Meaning and the structure of language (1970), later translated into five languages. One of the main components of that book is a semantic classification of verbs. Along with Charles Fillmore, Chafe was among the first who realized the importance of semantic roles. The book also clearly demonstrated a typical feature of Chafe’s subsequent publications — a very clear and simple style of writing that is of great persuasive force. In the 1970s, the main focus of Chafe’s theoretical research interests shifted to natural discourse. At that time, the trend that later acquired the label ‘functionalism’ and became a radical alternative to generativism started to take shape in the United States. Although Chafe was never an ideolog, he became one of the leading functionalists. From the very beginning of his discourse studies, Chafe relied on a cognitive approach and tried to reconstruct the mental processes of a speaking person. In that respect, he was far ahead of the other students of text. In a widely known 1976 article, Chafe approached such familiar phenomena as referential devices, word order, and choice of subject from a cognitive perspective. His studies of the 1970s culminated in the 1980 monograph The pear stories in which an unusual method of data collection was used: Chafe and his younger associates (D. Tannen, J. du Bois, P. Clancy, among others) showed a short film of a boy collecting pears to speakers of different languages and ages who later retold what they saw. Thus, the input was controlled, and the different versions of the pear stories were available for comparison. Many innovative conclusions concerning processes of the verbalization of visual experience, the dynamics of human consciousness, cultural differences in emphasizing different layers of information, and the cognitive basis of grammatical choices were drawn in that study. The following stage of Chafe’s work is reflected in his recent (1994) book Discourse, consciousness, and time. This work looks into relationships between language and consciousness, the latter understood by Chafe as the mental system crucially responsible for discourse production. According to Chafe, language and consciousness cannot be understood and

explored in isolation from each other. Chafe evaluates both disciplines studying those two domains, that is, modern linguistics and psychology, in a very critical vein, in particular, for their reliance upon artificial data and distrust toward introspective methods of research. Chafe emphasizes the use of introspection as a source of insights into language and cognitive processes, as well as the priority of natural data. The latter concern results from Chafe’s inherent connection with the American ethnographic tradition that has always emphasized the importance of natural empirical data. In this 1994 book, Chafe explores two main problems. The first is the explanation of linguistic phenomena on the basis of processes occurring in the speaker’s consciousness. Consciousness can be immediate (a reflection of what is here and now) and displaced (memories, imagination). Immediate consciousness is more basic and elementary, while displaced consciousness can only be understood as a modification or complication of immediate consciousness. The second problem is the comparison of spoken and written language. Spoken language is more basic and universal, despite the fact that for decades linguists have directed their attention to written or quasiwritten language only. Chafe gives priority to the spoken mode, but has also studied properties that distinguish written language. The fundamental unit of discourse, according to Chafe, is the intonation unit, which is a quantum of discourse corresponding to one focus of consciousness. Each intonation unit typically contains one element of new information. The opposition ‘given vs. accessible vs. new’ is responsible for prosodic (stressed vs. unstressed) and lexical (pronoun vs. noun) realizations of referents. The status of syntactic subject, central for English grammar, is explained on the basis of the notion of the starting point for the expression of a focus of consciousness. The important light subject constraint suggests that subjects are normally given or accessible, and can be new only under very special conditions. The intermediate status of accessible information is the base for defining a discourse topic, which, in its turn, underlies a definition of sentence. Among various modifications related to written language and the displaced mode of consciousness, prose written in the third person and quoted speech are discussed. At present, Chafe is engaged in producing descriptions of the Seneca and Caddo languages, and he is also investigating the uses of prosody in ordinary speech, particularly for the expression of emotions and humor, and relations between language and music.

185

CHAFE, WALLACE Among the main characteristics of Chafe’s style of scientific thinking are the ability to look afresh at ordinary phenomena and to integrate a broad range of facts into a coherent picture, as well as independence from the dominant trends of thought. There is no doubt that Chafe is among the leading figures and a classic of modern linguistics, but like Sapir he has not created a circle of immediate followers and his works are not in the mainstream of modern linguistics.

Biography Wallace Chafe was born in Cambridge, Massachusetts on September 3, 1927. He served in the US Navy, 1945–1946, and Central Intelligence Agency, 1950–1954. He received his B.A. in German from Yale University in 1950 and M.A. in linguistics from Yale University in 1956. He received his Ph.D. in linguistics from Yale University in 1958 for a dissertation on the Seneca language. He was Assistant Professor of Modern Languages at the University of Buffalo, 1958–1959, and linguist in the Bureau of American Ethnology, Smithsonian Institution, 1959–1962. In 1962, he moved to the University of California at Berkeley and worked there in the Department of Linguistics until 1986. From 1986 onward, he worked in the Department of Linguistics at the University of California at Santa Barbara. At present, he is Professor Emeritus at UCSB. He was Fellow of the Center for Advanced Study in the Behavioral Sciences, Palo, 1976–1977. He held a number of additional teaching appointments, including the Linguistic Institute of the Linguistic Society of America, in the summer of 1995, Albuquerque, New Mexico. Chafe was editor of the Mouton Grammar Library, 1984–1991; Topic Editor, Oxford International Encyclopedia of Linguistics; President of the Society for the Study of the Indigenous Languages of the Americas, 1985–1986; Associate Director of the Linguistic Institute of the Linguistic Society of America, Georgetown University, 1985; co-Director of the Center for the Study of Writing, Berkeley, 1985–1986; and member of the Linguistic Society of America Executive Committee, 1987–1989. He was also Member of the advisory and editorial boards of: International Journal of American Linguistics; Anthropological Linguistics; John Benjamins Typological Studies in Language Series; International Pragmatics Association; Humor: International Journal of Humor Research; Text: An Interdisciplinary Journal for the Study of Discourse; and Consciousness and Cognition. In addition, he was a member of the following profes-

186

sional societies: Linguistic Society of America; American Anthropological Association; American Psychological Association; American Psychological Society; Society for Linguistic Anthropology; Society for the Study of the Indigenous Languages of the Americas; International Society for Historical Linguistics; and International Pragmatics Association. He was Skomp Distinguished Lecturer, Indiana University, 1997, and received the medal of the University of Helsinki in 1998.

References Chafe, Wallace. 1962. Estimates regarding the present speakers of North American Indian languages. International Journal of American Linguistics 28. ———. 1967. Seneca morphology and dictionary. Washington: Smithsonian Press. ———. 1970. A semantically based sketch of Onondaga. International Journal of American Linguistics, Memoir 25 (Supplement to 36 (2)). ———. 1970. Meaning and the structure of language. Chicago: University of Chicago Press. ———. 1974. Language and consciousness. Language 50. ———. 1976. The Caddoan, Iroquoian, and Siouan languages. The Hague: Mouton. ———. 1976. Givenness, contrastiveness, definiteness, subjects, topics, and point of view. Subject and topic, ed. by C.N. Li. New York: Academic Press. ——— (ed.) 1980. The pear stories: cognitive, cultural, and linguistic aspects of narrative production. Norwood, NJ: Ablex. Chafe, Wallace, and Johanna Nichols (eds.) 1986. Evidentiality: the linguistic encoding of epistemology. Norwood, NJ: Ablex. Chafe, Wallace. 1994. Discourse, consciousness, and time: the flow and displacement of conscious experience in speaking and writing. Chicago: University of Chicago Press. ———. 1996. Sketch of Seneca, an Iroquoian language. Handbook of North American Indians. Vol. 17, Languages, ed. by W.C. Sturtevant. Washington, DC: Smithsonian Institution. ———. 1998. How a historical linguist and a native speaker understand a complex morphology. Historical linguistics 1997, ed. by M.S. Schmid, J.R. Austin, and D. Stein. Amsterdam/Philadelphia: John Benjamins. ———. 1998. Language and the flow of thought. The new psychology of language: cognitive and functional approaches to language structure, ed. by M. Tomasello. Mahwah, NJ: Lawrence Erlbaum. ———. 2000. Loci of diversity and convergence in thought and language. Explorations in linguistic relativity, ed. by M. Pütz, and M. Verspoor. Amsterdam/Philadelphia: John Benjamins. Chafe, Wallace. Autobiography. Historiographia Linguistica (in press). Parret, Herman (ed.) 1974. Interview with Wallace Chafe. Discussing language.The Hague: Mouton. Kibrik, Andrej A. 1994. Cognitive studies in discourse. Voprosy jazykoznanija 5 [in Russian].

ANDREJ KIBRIK

CHAO YUEN REN

Chao Yuen Ren A Chinese–American linguist and one of the leading figures in American Structuralism, Chao Yuen Ren made important contributions in the general theory of language and by introducing the Chinese language(s) as an object of research to a broader linguistic community. His work covers a wide range of issues both in theoretical and in applied linguistics, including fieldwork on Chinese dialects (usually conceived today as separate languages), the grammar of Modern Chinese, Chinese lexicography, practical material for Chinese language teaching, language standardization in China, the general theory of language, and studies in sociolinguistics. Born into a family belonging to the old scholar and mandarin class, Chao’s interest in linguistic variability was aroused during early childhood, when he came into contact with, and learned to speak, a number of different dialects. His family originally came from the town of Changzhou (between Nanjing and Shanghai), where the westernmost variety of the Wu dialect group is spoken. During the first half of his childhood, however, the family lived in various places in Hebei province, Northern China, and frequently moved around between the then provincial capital Baoding and smaller places. In those years, Chao was therefore exposed not only to the Wu dialects spoken by the older generations of his family but also to different varieties of Northern Mandarin, which he heard from his surroundings. Although he majored in mathematics and physics at Cornell University, he increasingly took an interest in philosophy and was influenced in particular by the writings of Bertrand Russell. However, he kept up his strong interest in languages, further stimulated by a course in phonetics and other linguistic courses. During those years, he also learned German and French, and picked up more Chinese dialects from Chinese fellow students. During his graduate studies in the Department of Philosophy at Harvard, he attended an introductory course in linguistics and took quite a few language courses on the side, among them a course in Sanskrit. During the winter of 1920/1921, Chao served as an interpreter to Bertrand Russell during his lecture tour in China and made full use of his mastery of the dialects. This experience provided a challenge to Chao in a way physics never had, and it was then that he decided upon linguistics as a major field of study and research. Therefore, while teaching philosophy at Harvard during the next three years, he took linguistic courses at the same time, and subsequently went to

Europe to work and study with the phoneticians Daniel Jones and Stephen Jones, the linguists Joseph Vendryes and Antoine Meillet, and the sinologists Bernhard Karlgren, Paul Pelliot, and Henri Maspero. After his return to China, primarily to teach Chinese phonology, he spent more than ten years making and organizing surveys of the Chinese dialects, which resulted in a number of pioneering studies: the first one was on the Wu dialects of Jiangsu and Zhejiang provinces (1926–1927), published as his ground-breaking Studies in the modern Wu dialects (1928), followed by surveys on the Yue dialects of Guangdong and Guangxi provinces (1928–1929), Minnan dialects of Shantou and Chaozhou and the dialects of Anhui (1934), Jiangxi and Hunan (1935), and Hubei provinces (1936). During the late 1920s, he was also active in the Committee on Unification of the National Language and headed a group of Chinese linguists who developed a new system of Latinization for Chinese, National Language Romanization (Gwoyeu Romatzyh), which was unique in using spelling variants instead of diacritics to express the tones of Chinese. Chao had plans for a dialectal survey of the whole country, but his fieldwork was eventually interrupted by the political development in China and the outbreak of war with Japan. Following these events, in 1938 he came to the United States for good and subsequently taught at the University of Hawaii; at Yale, where he met Leonard Bloomfield, Edward Sapir, and Bernard Bloch; at Harvard; and, finally, at the University of California, from where he retired in 1960. Chao’s pioneering fieldwork on the Chinese dialects is clearly one of his main contributions to linguistics. Before him, the European sinologists Henri Maspero (1883–1945) and Bernard Karlgren (1889–1978) had used Chinese dialect data in comparative studies for reconstructing earlier stages of Chinese (in particular, what Karlgren called ‘Ancient Chinese’), but they were interested in Chinese dialects only to the degree that they had preserved the phonetic features of those earlier stages. Chao, however, was the first to look at the Chinese dialects from a truly modern perspective, recognizing them as objects worth studying for their own sake. He carried out wide-ranging dialect surveys using the most modern sound recording devices of the times, and he included real dialect words found in nonlearned, colloquial speech. Although Chao’s fieldwork was disrupted by the historical events in China and thus remained

187

CHAO YUEN REN fragmentary, it provided a solid foundation for later work, much of which was done by his students. In the field of Chinese grammar and language teaching, Chao wrote two epoch-making books in which he applied the grammatical approach of American structuralism to the description of Modern Chinese: Mandarin primer, published in 1948 as a textbook for American students learning Chinese, contains an introductory grammatical sketch that was translated into Chinese and had a tremendous impact in China, effectively introducing the methodology of structuralism to a wide Chinese audience. The grammatical sketch of the Primer provided the best linguistic description of Modern Chinese until 1968, when Chao published A grammar of spoken Chinese, a monumental work that was the first comprehensive reference grammar of Chinese written in English, and that is arguably the best description of Modern Chinese to this day. In Chinese lexicography, Chao and Yang’s Concise dictionary of spoken Chinese (1947) made a number of important lexicographic innovations, restricting entries to those morphemes that occur in the modern spoken or written language, marking the morphemes as ‘free’ or ‘bound’ from the point of view of the modern language, and as ‘literary’ if they occur in the written language only. Pronunciation of the entries is given in National Language Romanization. Among Chao’s works on theoretical phonology, the most influential one was his classical paper ‘The nonuniqueness of phonemic solutions of phonetic systems’ (1934/1957), which deals with the definition of the phoneme and tries to show that the principles for establishing a system of phonemes for a given language depend on a number of possible objectives set up by the linguist, some of which are external to the language system itself. Chao also introduced a convenient device of representing tones in phonetic or phonological transcriptions, which became known as ‘tone letters’ and was later adopted for the International Phonetic Alphabet. Over the years, a number of important articles on the problems of phonetics and transcription followed. More than anyone else, Chao laid the foundation for studying Chinese within the international perspective of modern linguistics, and played a key role in introducing the ideas of Western linguistics to linguistic scholars inside China. Outside the area of linguistics, Chao is remembered as an outstanding translator (Alice’s adventures in wonderland, 1922) and as a composer of Chinese and Western music.

Biography Chao was born in Tianjin, China, on November 3, 1892. From the age of seven, he was given a traditional Chinese education in the family school at home. The

188

family returned to Changzhou in 1901, and after both his parents died in 1904, he spent a year in his aunt’s home in Suzhou. After attending a primary school in Changzhou in 1906, he was sent to Nanjing for three years to continue his studies. In 1910, he passed a government examination, which earned him a scholarship for studying in the United States of America, and he went to the Tsing Hua College in Beijing to prepare himself. Chao thus studied mathematics and physics at Cornell University, Ithaca, B.A. (1914), followed by physics and philosophy at Harvard, where he earned his Ph.D. degree in 1918 with Continuity, a study in methodology, which was supervised by the philosopher and logician Henry Maurice Sheffer. Chao received the postdoctorate Sheldon Fellowship, 1918–1919 and went on to teach physics at Cornell in 1919–1920. He then went back to China to teach mathematics at Tsing Hua College in Beijing, but was soon asked to interpret for Bertrand Russell on his lecture tour of China, 1920–1921. Afterwards, he returned to Harvard to teach philosophy and Chinese, at the same time taking linguistic courses (1921–1924). He went to Europe to visit the sinologist Bernhard Karlgren in Gothenburg, took courses in phonetics with Daniel Jones and Stephen Jones in London, and attended lectures by Joseph Vendryes, Antoine Meillet, Henri Maspero, and Paul Pelliot in Paris (1924). Returning to China, he taught Chinese phonology and music at Tsing Hua College’s newly established Institute of Sinology (1925), and then conducted and organized fieldwork on Chinese dialects between 1926 and 1936. He was an active member of the Committee on Unification of the National Language and the chairman of the Linguistics Section of the Institute of History and Philology, Academia Sinica (1929–1938). He was visiting professor at the Sunyatsen University, Canton (1928—1929) and then temporarily returned to the United States of America as the director of the Chinese Educational Mission in 1932–1933. He obtained his American citizenship in 1938 and then taught literary Chinese as a visiting professor at the University of Hawaii (1938–1939) and Yale (1939–1941) before returning to Harvard once again, as a researcher with the Harvard-Yenching Institute’s Chinese–English Dictionary Project. He also taught Chinese and Cantonese, and served as head of Harvard’s wartime Chinese language program (1941–1947). He became Professor of Oriental Languages and Linguistics at the University of California, Berkeley in 1947 and was made Agassiz Professor of Oriental Languages and Linguistics in 1952. A guest lecturer at National Taiwan University and Kyoto University, Japan in 1959, he retired the following year, in 1960. Chao was also President of the Linguistic Society of America, 1945; full member of the Academia Sinica, 1947; Fellow of the American Academy of Arts

CHINA and Sciences, 1948; and President of the American Oriental Society, 1960. He earned honorary degrees from Princeton University (1946), the University of California (1962), and Ohio State University (1970) and was named Honorary Professor at Beijing University in 1981. Chao died in Cambridge, Massachusetts, on February 24, 1982. References Bibliography of Yuen Ren Chao’s works, In Chao. 1976. 402–15. Chao Yuen Ren. 1928. Xiandai Wuyu de Yanjiu (Studies in the modern Wu dialects). ———.1930. A system of tone letters. Le Maître Phonétique 45. 24–7. ———.1934. The non-uniqueness of phonemic solutions of phonetic systems. Bulletin of the Institute of History and Philology (Academia Sinica) 4. 363–97; reprinted in Readings in linguistics, ed. by Martin Joos, 38–54. Washington, DC: American Council of Learned Societies, 1957.

———.1947. Cantonese primer. ———.1948. Mandarin primer. ———.1968. A grammar of spoken Chinese. ———.1968. Language and symbolic systems. ———.1968. Readings in sayable Chinese. ———. 1975. Yuen Ren Chao’s autobiography: first 30 years, from 1892–1921. Ithaca, New York: Spoken Language Services. ———. 1976. My linguistic autobiography. In Chao. 1–20. ———. 1976. My fieldwork on the Chinese dialects. In Chao. 26–33. ———. 1976. Aspects of Chinese sociolinguistics, ed. by Anwar S. Dil. Stanford: Stanford University Press. Chao, Yuen Ren. and Yang Lien-Sheng. 1947. Concise dictionary of spoken chinese. Dil, Anwar S. 1976. Introduction. In Chao. xi–xiv. Egerod, Søren, 1996. Chao Yuen Ren. Lexicon Grammaticorum, ed. by Harro Stammerjohann et al., 178–9. Tübingen: Niemeyer. Wang, W. S.-Y. 1983. Chao Yuen-Ren. Language 59. 605–7.

ROBIN SACKMANN See also Structuralism

China The region of China discussed in this essay covers the entire territory claimed by the People’s Republic of China, including Taiwan. It consists of 23 provinces, five autonomous regions, four municipalities, and two special administrative regions in an area of over 9,630,000 square kilometers. With a population of 1,295,330,000 (in the 2000 census), China is the most populous country and one of the most diversified regions in terms of languages and cultures in the world. No accurate number for languages spoken in China is available, but at least 235 languages (including signed languages and ten extinct languages) have existed in this region, cf. Grimes’s Ethnologue (2000). Barring a few exceptions, they belong to one of the following groups (parenthetical figures indicate number of members and the group’s percentage of total living languages): Sinitic (13, 5.78%), Tibeto-Burman (72, 32%), Tai-Kadai (31, 13.78%), Hmong-Mien (32, 14.22%), Altaic (25, 11.11%), Austro-Asiatic (24, 10.67%), Austronesian (18, 8%), and Indo-European (4, 1.78%). Figure 1 outlines their approximate distribution in China. The map shows that Sinitic sits in the center, but it has spread out to every corner of China. There has been a general migration flow of people from the north to the

south throughout much of the history of China. Southward invasion by Altaic speakers sometimes leads to the spreading of Chinese further north, as in the case of Manchu’s government. On the other hand, southward expansions and escapes by Sinitic speakers are responsible, in part, for the migration of Hmong-Mien and TaiKadai speakers to their present-day habitation. Scholars in China generally consider the SinoTibetan family to consist of four branches: Sinitic, Tibeto-Burman, Tai-Kadai, and Hmong-Mien (MiaoYao). While the typological profiles of the latter two are rather similar to that of Sinitic, their genetic relationship with the family remains to be proven. Table 1 summarizes three aspects of the major language groups as to whether their members (confined to those in China) are tone languages, whether morphemes (the smallest unit that carries a meaning) in the languages are predominantly monosyllabic, and in what order the Subject (S), Verb (V), and Object (O) appear in canonical sentences. The boundary between language and dialect is fuzzy in nature, and highly controversial when political factors are taken into consideration. The case of Chinese is a good example of this. The Chinese identify themselves ethnically as Han, and Hanyu is the

189

CHINA

Altaic

Altaic River Yellow

itic

Sin

Tibeto-Burman

River

Yangtze Hmong-Mien Austro-Asiatic

Tai-Kadai

TABLE 1 A Brief Typological Comparison of the Major Language Groups in China Language Group

Tone Language

Monosyllabicity of Morphemes

Word Order

Sinitic TibetoBurman Tai-Kadai HmongMien Altaic AustroAsiatic Austronesian

All Most

Yes Yes

S–V–O S–O–V

All All

Yes Yes

S–V–O S–V–O

None Few

No Most

S–O–V S–V–O

None

No

V–S–O

official language of China. From a strictly linguistic point of view (based on mutual intelligibility), however, at least 13 Chinese languages are spoken in modern China (see Figure 2). In the Chinese tradition, these are construed as dialects of Hanyu or subdialects therein. Prior to the twentieth century, Classical Chinese had been used in writing for all purposes, serving as the official language of the emperor for many dynasties. At the May 4th movement in 1919, student protesters in Beijing held among their goals a standard written language based on the modern Chinese speech. The then Government of the Republic of China responded by designating Gwoyu ‘national language’, based on Beijing Mandarin, as

190

Austronesian

Figure 1. The distribution of major language groups in China.

the standard language of the new republic. Later in the 1950s, the Government of the People’s Republic of China promoted Putonghua ‘common language’ (also based on Beijing Mandarin) as the national language. Gwoyu (still in use in Taiwan) and Putonghua have little difference in phonology. Their divergence has largely resulted from further language standardization introduced on the mainland, notably the replacement of phonetic characters by pinyin in the Roman alphabet and especially the use of simplified characters. The latter has compromised the antediluvian tradition of a unified written form for the Chinese since the unification of the warring states in China in 221 BCE. Encyclopedia Sinica (1988) estimates the number of speakers of seven major dialect groups of Hanyu in the early 1980s as follows: over 700 million for Mandarin, 70 million for Wu, 40 million for Yue, 40 million for Min, 25 million for Xiang, 30 million for Gan, and 37 million for Hakka. Figure 2 shows the relation between 13 identified Sinitic languages. The tentative placement of Wan Nan is geographically motivated. The Yangtze River is the watershed between Northern Chinese and Southern Chinese. Mandarin stretches from the vast region north of the Yangtze River to southwestern China provinces of Sichuan, Yunnan, Guizhou, western Hunan, and northwestern Guangxi. Following the escape of the Nationalist government from the mainland, a sizable number of Mandarin speakers settled down in Taiwan in the

CHINA Mandarin Northen Chinese

Jinyu

Wu Sinitic

Min Nan (Southern Min) Pu-xian

Yue Min Dong (Eastern Min) Min Southern Chinese

Min Bei (Northern Min) Xiang Min Zhong (Central Min) Gan Hakka Wan Nan (Huizhou)

1950s. More than 870 million people (in the late 1990s) speak Mandarin as a first language, including nearly all Manchu and the entire population of Hui, the second and third largest minority nationalities of China. The figure rises well beyond one billion if the number of fluent speakers of Mandarin as a second language is taken into account. Mandarin has four major dialects: Northern Mandarin, Northwestern Mandarin, Southwestern Mandarin, and Lower Yangtze Mandarin. Found in coastal provinces, Wu is spoken on the east coast, Yue on the south coast, and Min on the southeast. The Yangtze Delta can be considered as the homeland of Wu. Wu is spoken nowadays in Zhejian province and much of the area south of the Yangtze River in Jiangsu province. It has five dialects: Taihu (which includes Shanghai as one of its six subdialects), Taizhou, Wenzhou, Wuzhou, and Liqu. The homeland of Yue (Cantonese) is in the Pearl Delta; from there, it spreads to a large part of Guangdong province and as far as southeastern Guangxi. Yue dialects include Yuehai (which is spoken in Guangzhou, Hong Kong, and Macau), Siyi, Gaoyang, and Guinan. Min is regarded as the fourth largest dialect of Hanyu, and also the most diversified one with respect to dialectal difference. Indeed, the diversion is so significant that it warrants the treatment of Min as a linguistic group comprising five consanguineous, but distinct, languages. Li’s Dialects of Fujian (1997) describes the distribution of these languages as follows: excluding a small Wu-spoken area in the northern corner and about one third of the province in western Fujian (where Hakka and Gan are spoken), the rest of Fujian can be divided into six areas: the south-

Figure 2. A linguistic tree of identified Sinitic languages in China. ern half contains Min Nan and Pu-xian (the latter occupies only a corner in the east), and the northern half contains Min Dong (in the eastern half), Min Bei (about 60% of the remainder in the north), and Min Zhong (the western half of the leftover). A transitional language between Min Nan and Min Dong exists in Youxi county, the very center of the province. Min Nan is the largest Min language, with 25.7 million speakers (in 1984) in southern Fujian, eastern and western parts of Guangdong, parts of Hainan and Zhejiang, and an additional 15 million speakers (in 1993) in Taiwan. Major Min Nan dialects include Hokkien, Chao-Shan, Leizhou, Hainan, and Zhenan Min. Many can be subdivided, e.g. Hokkien has four subdialects: Amoy, Quanzhou, Zhangzhou, and Longyan. Pu-xian and Min Zhong occupy the smallest areas among the Sinitic, but Pu-xian has around three million speakers, whereas speakers of Min Zhong are less than one million. They each have two dialects, and so do Min Dong and Min Bei. Xiang is mainly spoken in much of the eastern half of Hunan province. With Mandarin dominant in the north and in the west, Xiang has been under great pressure of assimilation. The influence of Mandarin represents a primary factor for variation between its two dialects: Northern Xiang is subject to more innovation, while Southern Xiang is more conservative. Gan is largely spoken in Jiangxi province, with five major dialects: Chang-Jing, Yi-Liu, Ji-Cha, Fu-Guang, and Ying-Yi. It is the closest relative of Hakka. The distribution of Hakka is unique among Southern Chinese in that it covers the widest area in China, but often with a low concentration; it does not have a home base in any province. While the border area between northeastern Guangdong, southern

191

CHINA Jiangxi, and southwestern Fujian has a higher concentration of Hakka speakers, small Hakka-speaking communities are scattered in parts of Taiwan, Hunan, Sichuan, Guangxi, and throughout Guangdong. Major Hakka dialects include Min-ke, Yue-Tai, Yuezhong, Yuebei, Yugui, Tonggu, etc. The majority of She nationality also speak Hakka as their first language. Jinyu and Wan Nan are two of the less-known Chinese languages. Spoken in Shanxi province and adjacent areas in Shaanxi and Henan provinces, Jinyu is generally assumed to be a subdialect of Northwestern Mandarin. Wan Nan is mainly spoken in an area called Huizhou in southern Anhui province. It is usually treated as a subdialect of Lower Yangtze Mandarin. The Chinese government recognizes only 55 minority nationalities. Their total population is about 106.8 million, or 8.25% of China’s population (in the 2000 census). Around 60% of them are native speakers of over 94% of languages in China. Table 2 shows the top 12 minority languages, each with one million or more speakers (in the early 1980s or 1990s). Spreading across Xizang Tibetan Autonomous Region, southeastern Qinghai, western Sichuan, and much of the Yunnan provinces, Tibeto-Burman accounts for almost one third of identified languages in China. The three largest groups within it are YiBurmese (28 languages), Himalayish (17 languages), and Qiangic (15 languages). Speakers of Yi-Burmese in China exceed 6.4 million, including such nationalities as Yi, Hani, Lisu, Lahu, Naxi, Nu, Jino, and Achang. Chinese scholars consider that Yi nationalities speak six rather distinct “dialects” (just as the Chinese themselves do). Yi nationalities use an ideographic system for writing their languages, but characters vary considerably from district to district over southeastern Sichuan, Guizhou, and Yunnan. A set of

standard characters based on Sichuan Yi was introduced in 1975. Tomba is the traditional writing developed by Naxi nationality. The hieroglyphic writing, however, is restricted to religious texts only. Himalayish languages have about 4.1 million speakers in China. Over 82% of them speak Tibetan, Khams, or Amdo. These three consanguineous languages (regarded as ‘dialects’ of Tibetan in China) share a common writing script adapted from the Indic syllabic writing system. Besides Tibetan nationality, the other official nationality for speakers of Himalayish is Moinba (whose population was less than 8,000 in the 1990 census). Exclusive to China, Qiangic languages have about 490,000 speakers in total. Approximately one fifth of Qiangic speakers are Qiang or Pumi nationalities; the rest identify themselves officially as being of Tibetan nationality. Historically, their languages have no written form, with the exception of Tangut (an extinct relative), which had a logographic writing modeled after Chinese characters. Bai, Tujia, Jingpo, Drung, and Lhoba are among the 17 official nationalities speaking Tibeto-Burman languages in China. Spoken in much of Guangxi Zhuang Autonomous Region, Hainan, southern Guizhou, and southwestern Yunnan, Tai-Kadai has the second greatest number of speakers in China. Chinese scholars regard Northern Zhuang and Southern Zhuang as dialects of Zhuang, which was once written in logographic characters akin to Chinese characters, but a new Roman orthography was introduced in the late 1950s. Together with 20 others, the Zhuang languages belong to the Kam-Tai group. Seven nationalities are designated for the 20 million speakers of Kam-Tai, namely Zhuang, Buyi, Dai, Dong, Sui, Mulam, and Maonan. The Dai have adapted Brahmi script (a syllabic Indic script) for writing Lü and Tai Nüa for many centuries. Kadai (81,000

TABLE 2 Top Twelve Minority Languages of China Language Name

Linguistic Affiliation

Nationality

Northern Zhuang Uyghur Southern Zhuang Mongolian Bouyei Korean Sichuan Yi Khams Southern Dong Kazakh Tibetan Hmong Njua

Tai-Kadai, Kam-Tai Altaic, Turkic Tai-Kadai, Kam-Tai Altaic, Mongolic Tai-Kadai, Kam-Tai Altaic Tibeto-Burman, Yi-Burmese Tibeto-Burman, Himalayish Tai-Kadai, Kam-Tai Altaic, Turkic Tibeto-Burman, Himalayish Hmong-Mien, Hmongic

Zhuang Uygur Zhuang Mongol Buyi Korean Yi Tibetan Dong Kazak Tibetan Miao

192

No. of Speakers 10,000,000 7,200,000 4,000,000 3,381,000 2,000,000 1,920,000 1,600,000 1,487,000 1,480,750 1,100,000 1,066,200 1,000,000

CHINA speakers in five languages) and Hlai (747,000 speakers in two languages) are two smaller groups within Tai-Kadai. Gelao and Li are the two official nationalities speaking Kadai and Hlai, respectively. All 32 Hmong-Mien languages found in the world have speakers in China, scattered in various parts of Guizhou, Hunan, Guangdong, Guangxi, and Yunnan. The three official nationalities—Miao, Yao, and She— correspond to the three language groups within Hmong-Mien. Hmongic is the largest, comprising 26 languages spoken by 4.2 million people who are mostly Miao, but some are Yao. Over 1.1 million Yao speak five Mienic languages, but less than 1,000 She speak the singleton language in the Ho Nte group. HmongMien languages have no traditional writing, but Roman orthographies have been devised for several larger languages since the 1950s. With 11 Turkic languages, eight Mongolic languages, and five Tungusic languages, all three major branches of the Altaic family are present in China. Spoken along the northern and northwestern border of China—mostly in Heilongjiang province, Inner Mongolian Autonomous Region, and Xinjiang Uygur Autonomous Region—by 18 nationalities, the Altaic languages amount to 25, if Korean is included. Khakas, Nanai, and Manchu each have less than 50 speakers in China. According to estimates of the 1990s, Turkic has over 8.49 million speakers among Uygur, Kazak, Kirgiz, Salar, Bonan, Ozbek, Tatar, and Yugur nationalities; Mongolic has nearly 4.15 million speakers among Mongol, Tu, Daur, Dongxiang, and Yugur nationalities; and Tungusic has about 49,700 speakers among Xibe, Ewinki, Oroqen, Hezhen, and Manchu nationalities. Arabic script is used for writing Uyghur, Kazakh, and Kirghiz, whereas traditional Mongolian script is used for writing Mongolian and Xibe. Koreans write in Hangul, a phonetic-based spelling system created in 1444. Va, Blang, and De’ang nationalities in western Yunnan, plus Jing nationality on the south coast of Guangxi, are speakers of Austro-Asiatic languages. All 24 Austro-Asiatic languages of China belong to the Mon-Khmer branch, with around 381,000 speakers in total. Only one sixth of them have more than 10,000 speakers. Parauk is the largest, with about 180,000 speakers (in 1990). The mainland government designates an Austronesian-speaking minority in Taiwan as Gaoshan nationality, but on the island they are further divided into nine ethnic groups, collectively referred to as Aborigines. The total number of Austronesian speakers in China is around 339,800 for 18 Austronesian languages. Those spoken on the Taiwan island belong to the Formosan branch, which comprises three groups: Paiwanic (ten languages), Tsouic (four languages), and Atayalic (two languages). The largest is

Amis (with 130,000 speakers) of the Paiwanic group, which lost seven members during the colonial rule of Japan (from 1895 to 1945). Grimes’s Ethnologue reports that another four Paiwanic languages and two Tsouic languages are nearly extinct. Unlike mainland China, Taiwan has witnessed language extinction similar to that found in Australia and North America, in the violent way of supplanting an aboriginal language with a colonial one. Under the rule of the Nationalist government in the 1950s, linguistic imperialism was imposed on the Taiwanese people with Mandarin in lieu of Japanese. As the Taiwan government has advocated multiculturalism in recent years, the environment has began to change to one propitious for linguistic diversity. Yet, many educated Chinese parents in Taipei (the capital city) prefer speaking Mandarin to their young children, even though they are native speakers of Min Nan. Mandarin has undoubtedly established its superior status in Taiwan over the half century. The firmly established status of Mandarin is also unchallengeable in mainland China. While the Chinese government does not adopt any aggressive language policies to wipe out linguistic minorities, policies that would sustain linguistic diversity are equally lacking. It is stated in the Chinese Constitution that minority nationalities have the right to maintain their own languages, but this by no means implies active promotion of minority languages. In spite of the availability of newly devised orthographies for languages of a few larger nationalities, the priority for mastering Putonghua, which is often the sole language in education and mass media, cannot be mistaken. Ethnic minorities have been in a plight parallel to that of new immigrants in countries such as the United States. In order to become part of the mainstream society, they have to give up their native languages, sometimes eagerly. Interethnic marriage (not necessarily with Han-Chinese) in cities is another factor for prompting children to become monolingual in Mandarin. This kind of language shift is usually completed within several generations, characterized by bilingual speakers during the peaceful transition. Bilingual speakers of autochthonous languages are legion in China but the bias toward Mandarin is conspicuous. Most of the bilinguals, except for a small fraction of minority nationalities, are literate only in standard Chinese based on Mandarin. Linguistic minorities surrounded by larger communities of different languages are likely to become multilingual in several oral languages, e.g. some Zhuang nationalities speak Yue and Mandarin in addition to their native language and some Pumi nationalities are fluent in Prinmi, Mandarin, and Sichuan Yi. While considerable ethnic minorities (especially those above the age of 50) are fluent in two or more languages, the ratio of

193

CHINA monolingual Mandarin speakers in minority nationalities is steadily rising. Ethnic minorities are not alone under the pressure of linguistic absorption. All languages other than Mandarin are facing varied degrees of endangerment, as Putonghua is upheld as the national language at the expense of the others. For instance, vernacular pronunciation of Chinese characters is no longer taught in school (except in Hong Kong and Macau). In the past, Wu, Yue, Min Nan, and other Sinitic languages were featured in vernacular operas in different parts of the country, but nowadays popular songs are sung exclusively in Putonghua in mainland China. These suggest an ever-shrinking room for the survival of other languages in China. The linguistic landscape in China is byzantine. Many languages are awaiting in-depth studies and proper classification, while some are yet to be discovered. Hopefully, these would be accomplished before language death accelerates under the momentous impact of the on-going economic development and modernization.

References Grimes, Barbara (ed., Joseph Grimes, consulting editor) 2000. Ethnologue: languages of the world, 14th edition. Dallas, TX: Summer Institute of Linguistics. Ji, Xianlin (editor-in-charge) 1988. Zhongguo Da Baikequanshu: Yuyan, Wenzi (Encyclopedia sinica: languages and scripts). Beijing: China Encyclopedia Publisher. Li, Rulong. 1997. Fujian Fangyan (Dialects of Fujian). Fuzhou: Fujian People’s Publisher. Norman, Jerry. 1998. Chinese. Cambridge: Cambridge University Press. Ramsey, Robert. 1987. The languages of China. Princeton, N J: Princeton University Press. Wang, William S.-Y. (ed.) 1991. Languages and dialects of China. Berkeley, CA: Journal of Chinese Linguistics. Wurm, Stephen, et al. (general editors) 1987. Language atlas of China. Hong Kong: Longman. Wurm, Stephen, Peter Muhlhausler, and Darrell Tryon (eds.) 1996. Atlas of languages of intercultural communication in the Pacific, Asia, and the Americas. Berlin and New York: Mouton de Gruyter. PICUS SIZHI DING

See also Altaic; Austronesian; Chinese (Mandarin); Chinese and Japanese Traditional Grammar

Chinese (Mandarin) The Chinese language is represented by groups of dialects that are approximately as distant from each other as English from German, or French and Spanish. The traditional classification has been into seven major groups: Mandarin, Wu, Xiang, Gan, Min, Yue, and Hakka. Collectively, these dialects are spoken by more speakers than any other language in the world. Of these, Mandarin has the special status of having been the official language of China for most of Chinese history. As such, Mandarin is the first dialect learned by the great majority of people in China, and is acquired to various degrees of proficiency by speakers of all other dialects. Thus, there are well over a billion people who use Mandarin in some way or other in their lives. Whereas the other dialects are concentrated mostly along the southeastern coast, Mandarin is distributed widely all over China. The Mandarin spoken in Harbin in the northeast is clearly different from the Mandarin spoken in Chengdu in the southwest, a thousand miles away. Because Beijing has been the capital of China for most of the past millennium, its speech has become an official linguistic standard, recognizing of course

194

that the term Mandarin in its wide sense actually covers many varieties of speech. There are many ways to refer to this language. Traditionally, when there is no need to differentiate it from the other major dialects, Mandarin is simply called ‘Chinese’. We will follow this tradition in the present article. The word ‘Mandarin’ actually means ‘official’, corresponding to the Chinese word ‘guan’. Thus, the Chinese counterpart to ‘Mandarin’ as a language is ‘Guanhua’, or ‘official speech’. To emphasize its wide usage and enhance its populist intent, the term currently preferred is ‘Putonghua’, where ‘putong’ means ‘common’. The term ‘Guoyu’, or ‘national language’, is used in Taiwan. The term ‘Huayu’ is used in Singapore, where ‘hua’ is an auspicious word that has been used to designate ethnic Chinese since the ancient Xia dynasty, some 4,000 years ago. In fact, the two elements ‘hua’ and ‘xia’ are often used together as a compound to refer to the long ethnic heritage. Lastly, the term ‘Sinitic’ is used in more technical discussions, such as ‘the Sinitic language’, meaning ‘the Chinese language’. The stem ‘Sin-’ probably derives from the name of the Qin dynasty (221–206

CHINESE (MANDARIN) BCE), when China was first unified politically. ‘SinoTibetan’ refers to the family of languages one branch of which is Sinitic; another branch is Tibeto-Burman, which contains several hundred languages spoken in Western China and around the Himalaya mountains. The traditional classification of Sino-Tibetan is that proposed by F.K. Li; it includes a third branch of Zhuang-Dong languages, and a fourth branch of MiaoYao languages. This four-branch hypothesis has recently become controversial as more data on minority languages became available, although the relation between Chinese and Tibeto-Burman has never been disputed. It has been estimated that the Chinese and Tibet-Burman retained their unity until about 6,000 years ago, when extensive migrations radiated outward from its homeland, presumably in northern China, which ultimately gave rise to the hundreds of languages we see today. There have been efforts to connect the Sino-Tibetan languages to much more distant origins. In the early part of the twentieth century, the American linguist Edward Sapir remarked on the deep similarities between Tibetan and some of the native languages of North America. More recently, research is being done on the DeneCaucasian hypothesis, which posits that groups as distant as the Na-Dene languages of North America, the Yeniseian languages in Siberia, Basque in Western Europe, among others, are all related to Sino-Tibetan. A competing hypothesis links Chinese to the Austronesian languages, thus implying an ancient homeland more to the south. Such hypotheses on distant prehistories are intrinsically fascinating, although naturally they are more difficult to verify, often for lack of adequate data or precise methodology. An encouraging development in recent years is the increase in interdisciplinary collaboration, particularly among linguists, anthropologists, and geneticists, which significantly broadens our perspectives on human evolution, and brings new promise that the controversies surrounding the ancestry of the Chinese language can eventually be resolved. The earliest specimens of Chinese date back some 3,400 years, in the form of inscriptions on animal bones and on bronze artifacts. These inscriptions were made for the purpose of divination or for various historical or ritual purposes. Although both the inscribed characters and the language they represented have changed over the centuries, their ancestral relations with modern characters and contemporary Chinese can be clearly established. Thus of all the modern languages, Chinese can boast of the longest period of continuous use. Since the characters often contain a phonetic component, they also gave valuable indirect clues for the reconstruction of the spoken language. Efforts to understand the shape, sound, and semantics of the ancient characters

have accumulated over the centuries into an impressive body of philological scholarship in China. This scholarship received a major impetus from abroad some 2000 years ago, when Buddhism first came into China. With the religious teachings came the Sanskrit language and its quasialphabetic spelling system, which in turn stimulated the phonetic analysis of the Chinese language. Under this influence, Chinese scholars divided into the syllable for the time in terms of the fanqie system. A character X can be spelled by two other characters, A and B, where A indicates the initial consonant of X and B indicates the remainder of the syllable of X. The fanqie system fails to help when the reader does not know either A or B. Around 1,700 years ago, Chinese scholars first began writing about the four tones that their language had, which were named ping, shang, qu, and ru. Rime dictionaries were compiled around 1,500 years ago, phonetic charts were constructed 1,000 years ago, and by 400 years ago, scholars had achieved a reasonable understanding of the history of the language. Chinese and western traditions began to come together around the beginning of the twentieth century. In 1898, the first western style grammar of Chinese was published by Ma Jianzhong, who studied several European languages while in Shanghai, and who later received training in Paris. Early in the twentieth century, the Swedish scholar Bernhard Karlgren traveled to China to conduct fieldwork on Chinese dialects, which he published as part of his dissertation, 1915–1926. Karlgren was the first to integrate the philological achievements of the Chinese scholars with the real sounds of modern dialects to arrive at systematic reconstructions. Although many aspects of these reconstructions have been superseded, his work has provided an invaluable stimulus for integrating western ideas into Chinese scholarship. A very important source of data for reconstructions, investigated by Chinese scholars over many centuries, is the Shijing, a body of some 300 poems and songs that date back to around 3,000 years ago. This work is variously known in English as the Book of Poetry, the Book of Songs, or the Book of Odes. By investigating which sets of characters rimed in the Shijing, and by integrating this knowledge with the phonetic information contained in the individual characters, we can go a long way toward inferring how they were pronounced. The language of the Shijing is referred to as Old Chinese. Thanks to the early rime dictionaries, the best known of which is the Qiejun, we can also reconstruct the language of the seventh century, which is referred to as Middle Chinese. Mandarin can be traced back to another famous rime dictionary, the Zhongyuan Yinyun, compiled in the early fourteenth century. Another important source of information on the sounds of this period is the

195

CHINESE (MANDARIN) ‘Phags-pa alphabet, invented by a Tibetan lama in 1260, by order of Kubli Khan. The language revealed by these sources is called Old Mandarin.

Phonology Some of the distinct aspects of Mandarin can be discussed by reference to the structure of the syllable, as shown in the following chart: Tone Initial

Final Medial

Rime Nucleus

Ending

All the dialects of Chinese can be described according to the above chart, since they all share the same syllable structure. The three basic building blocks are the tone, the initial, and the final. In European languages, a syllable is distinguished from other syllables by its vowels, such as ‘me’ vs. ‘moo’, or by its consonants, such as ‘me’ vs. ‘bee’. In a tone language, a syllable is distinguished along an additional dimension, as illustrated in the illustration below. A Chinese syllable is also distinguished by its vowels and consonants. These are illustrated for Mandarin as spoken in Beijing in Figure 1 by ‘ma’ vs. ‘mi’ for vowel distinctions, and by ‘ma’ vs. ‘pa’ for consonant distinctions. Additionally, ‘ma’ can carry any of four distinct tones, yielding four completely unrelated words with the meanings of ‘to scold’, ‘hemp’, ‘mother’, and ‘horse’. Tone is produced primarily by shaping the pitch contour of the voice by varying the rate of vibration of the vocal folds. The pitch contours of these four words, as pronounced by the present author, are shown in the computer traces in Figure 1. We can see, for example, that the pitch contour for ‘to scold’ is primarily falling while for ‘hemp’ it is rising. For other varieties of Mandarin, both the number of tones and the shapes of their pitch contours will be different. Going back to the chart of the syllable, the initial represents at most one consonant. Thus, there are no consonant clusters in Chinese, such as in English ‘spy’ or ‘pry’. The medial may be any one of the three onglides: ‘i’, ‘u’, or ‘ü’. The nucleus in Mandarin is a simple vowel, although it may be a diphthong in other dialects. The ending may be any one of the three consonants ‘r’, ‘n’, or ‘ng’. Thus, a Chinese syllable ‘liàng’, which means ‘light’, contains each of the following elements: Tone: high falling, Initial: l, Medial: i, Nucleus: a, Ending: ng.

196

Morphology In Chinese, by far the great majority of morphemes are represented by single syllables. Since single syllables are written by single characters, there is good correspondence between the minimal units of the sound system, the writing system, and the grammatical system, i.e. syllable = morpheme = character. In contrast to European languages, word formation in Chinese does not involve any inflectional morphology to speak of, such as one finds in ‘books’, ‘walked’, etc. This poses no problem, of course, since information regarding gender, number, and case in nouns and tense in verbs is usually provided by the context. The speaker is not forced by the grammar, for instance, to choose ‘he’ or ‘she’ when referring to a friend, since both of these are ta in Chinese. On the other hand, Chinese does make use of derivational morphology, such as: Prefixes— Suffixes—

di wu ‘the fifth’, tou wu ‘the first five’, zhuantou ‘brick’, zhitou ‘finger’, haizi ‘child’, maozi ‘hat’.

There is an interesting compounding process, whereby opposites are conjoined to form a word. For example, da ‘big’ and xiao ‘small’ are conjoined to mean ‘size; lai ‘come’ and wang ‘go’ combine to mean ‘social visitation’. A particularly interesting word formed by this process is from dong ‘east’ and xi ‘west’; the word dongxi now has the very general meaning of ‘thing’ and is used very productively. While English has similar words like ‘flip-flop’ and ‘zig-zag’, the two processes are different in that the two members of the Chinese compounding are not phonetically related. Another interesting process of derivational morphology is reduplication. The process works in several different ways. One area of reduplication has to do with the formation of kinship terms. Whereas most languages have ‘papa’ and ‘mama’, such terms are much more abundant in Chinese, e.g. didi ‘younger brother’, nainai ‘grandmother’, jiujiu ‘maternal uncle’, etc. Another area of reduplication has to do with certain nouns to convey the meaning of ‘every’, e.g. tiantian ‘every day’, renren ‘every person’, jiajia ‘every family’, etc. Verbs reduplicate, to convey a range of meanings, such as transitory action, e.g. kankan ‘to take a look’, changchang ‘to have a taste’, tantan ‘to have a chat’, etc. When the verb is disyllabic, the reduplication applies to the two syllables together, e.g shangliangshangliang ‘to discuss a little’, or yanjiuyanjiu ‘to investigate a little’. Adjectives also reduplicate to form adverbs by taking the particle de. Thus, kuai means ‘quick’ and kuaikuaide pao means ‘run quickly’; qing

CHINESE (MANDARIN) means ‘light’ and qingqingde fangxia means ‘put down lightly’. However, when the adjective is disyllabic, the pattern of reduplication is different from that of the verb. The reduplication here is syllable by syllable, e.g. an’anjingjingde from anjing means ‘quietly’ and gaogaoxingxingde from gaoxing means ‘happily’. In addition to the compounding and reduplication discussed above, another distinctive aspect of Chinese morphology is the use of classifiers. In English, one uses measure words to quantify mass nouns, in expressions such as ‘a grain of sand’, ‘a head of cattle’, or ‘a piece of cloth’. Many classifiers in Chinese, however, have very little semantic content to speak of, such as in yige ren ‘one-classifier person’, yipi ma ‘one-classifier horse’, yijian dayi ‘one-classifier coat’, yiben shu ‘one-classifier book’, etc. In learning how to use a noun in Chinese, one needs to learn the appropriate classifier for it.

Syntax Since the language has very little inflectional morphology, grammatical relations among the words are primarily indicated by word order. The basic syntax in Chinese is subject–verb–object, much as in English. A typical declarative sentence with the perfective aspect marker le would be: [1] Ta mai-le shu.

‘He sell-aspect book.’

A major feature of Chinese syntax is that modificational structures precede the modified. This means that the head is always at the end of phrases. We have seen above that adverbs precede the verb in the verb phrase kuaikuaide pao ‘run quickly’. In noun phrases, numerals, possessive, adjectives all precede the noun, such as tade sanben hong shu ‘his three red books’. This applies to relative clauses as well, with the help of the particle de, as in zuotian zai xianggang kande shu ‘yesterday in Hong Kong read-particle book’. There are two syntactic variants of a declarative sentence: the disposal form with the particle ba shown in [2], and the passive form with the particle bei shown in [3]: [2] Ta ba shu mai-le. [3] Shu bei ta mai-le. In these variants, the object is fronted to precede the verb. The meaning of the passive sentence [3] is comparable to that of passive sentences in English. However, the disposal form has no ready counterpart in the syntax of European languages. Although it is a form that is frequently used, its meaning and range of use are topics for which there is no general agreement as yet among Chinese grammarians.

In forming yes–no questions, the typical process for European languages is the inversion of word order, such as ‘can he sell books?’ formed from ‘he can sell books’. English is unusual in the respect that a periphrastic ‘do’ is used to carry the tense, such as ‘did he sell books?’ formed from ‘he sold books’. Rather than inversion, the typical process for Chinese is to conjoin an affirmative verb phrase with its negative counterpart. The negative counterpart to [1] is [4]: [4] Ta mei-you mai shu. ‘He neg-aspect sell book’ Conjoining [1] and [4], we obtain either the question forms [5] or [6] below, depending on whether we delete repeated materials from the affirmative or negative verb phrase: [5] Ta mai-le shu mei-you? ‘He sell-aspect book neg-aspect?’ [6] Ta you mei-you mai shu? ‘He aspect neg-aspect sell book?’ Corresponding to the WH question words in English, i.e. ‘who, what, when’, Chinese has shei, shemme, and shemme shihou, respectively. These have been called SH words. In English, WH words need to be moved to the front of the sentence; in Chinese there is no such movement. Going back to [1], we may form the questions [7] and [8] as follows: [7] Shei mai-le shu? [8] Ta mai-li shemme?

‘Who sell-aspect book?’ ‘He sell-aspect what?’

Because of the required movement, the English translation of [8] would be ‘What did he sell?’ Every language has a stock of prefabricated phrases that have a richer cultural content, sometimes by means of their historical allusions. Such phrases are learned as complete units, rather than formed spontaneously for the moment. Thus, in English, we might say ‘He is the early bird that caught the worm this time’; or ‘Aren’t you the hostess with the mostess?’ Chinese is extremely rich in this regard, particularly in phrases formed with four syllables. These range from colloquial expressions such as you tou hua nao ‘oily head slippery brain’, used to describe someone devious, to forms that are more literary such as pao zhuan yin yu ‘throw brick induce jade’, used as a gesture of humility to say that what you have to offer is not worthy of what you are about to receive. Because they can have a wide range of cultural associations, prefabricated phrases are sometimes difficult to understand, especially when they are used as part of language games. A striking case of misunderstanding took place when the American writer Edgar Snow interviewed Mao Zedong. Mao described himself with the expression heshang da shan, which literally means

197

CHINESE (MANDARIN) ‘monk with umbrella’. This led Snow to evoke a poignant image: ‘He was, he said, only a lone monk walking in the world with a leaky umbrella’ [Life Magazine, April 4, 1971]. In actuality, the second part of this four-syllable expression is wu fa wu tian, which could mean ‘no hair no sky’ or ‘no law no heaven.’ The concept ‘heaven’ is especially important in this context because in traditional Chinese culture, the ruler governs by tianming, i.e. ‘mandate of heaven’. The key to this expression is the fact that fa is a homonym that means either ‘hair’, which a monk lacks, or ‘law’. The true message Mao was conveying is one of bravado— defying heaven—which Snow completely missed because of the cultural and linguistic gap. There are many other interesting aspects of the Chinese language: Boltz (1994) gives a good survey of the Chinese writing system, for example, and Huang et al. (1996) discuss the special challenges that Chinese poses to natural language processing—interfacing the language with the computer. References Baxter, William H. 1992. A handbook of old Chinese phonology. The Hague: Mouton. Boltz, William G. 1994. The origin and early development of the Chinese writing system. American Oriental Society.

Chao, Y.R. 1968. A grammar of spoken Chinese. Berkaley, CA: University of California Press. Huang, Churen, et al. (eds.) 1996. Readings in Chinese natural language processing. Journal of Chinese Linguistics Monograph 9. Huang, James C.-T., and A. Y.-H. Li (eds.) 1996. New horizons in Chinese linguistics. Dordrecht: Kluwer. Karlgren, Bernhard. 1954. Compendium of phonetics in ancient and archaic Chinese. Bulletin of the Museum of Far Eastern Antiquities 26. 211–367. Li, F.K. 1973. Languages and dialects of China. Journal of Chinese Linguistics 1. 1–13. Masini, Federico. 1993. The formation of modern Chinese lexicon and its evolution toward a national language. Journal of Chinese Linguistics Monograph 6. Norman, Jerry. 1988. Chinese. Cambridge. Peyraube, Alain. 1996. Recent issues in Chinese historical syntax, ed. by Huang and Li, 161–214. Sun, Chaofen (ed.) 1997. Studies in the history of Chinese syntax. Journal of Chinese Linguistics Monograph 10. Tzeng, O.J.L., and William S.-Y.Wang. 1983. The first two R’s. American Scientist 71. 238–43. Wang, William S.-Y. 1973. The Chinese language. Scientific American, February issue. Wang, William S.-Y. (ed.) 1991. Languages and dialects of China. Journal of Chinese Linguistics Monograph 3. Wang, William S.-Y. (ed.) 1995. The ancestry of the Chinese language. Journal of Chinese Linguistics Monograph 8.

WILLIAM S.-Y. WANG See also Austronesian; China

Chinese and Japanese Traditional Grammar The Chinese linguistic tradition bequeathed an analysis of tone, foreshadowed twentieth-century views of syllable structure, and contributed a moral and political perspective on the question of discrepancy between word meaning and word usage. In addition, explication of the formation of Chinese graphs made an important contribution to the understanding of writing systems. The Japanese tradition, on the other hand, offers unique perspectives on defining parts of speech, and analyzing the ways in which the parts combine to form larger wholes. Both traditions bear a debt to the linguistic tradition of India, some of which was transmitted through the powerful vehicle of Buddhism, in the intersection of phonetic analysis and sacred sound. Also common to both was the impetus of early poetry, in the desire to know the precise sounds and meanings of the past.

198

The Chinese debate on ‘rectification of names’ originates in several passages from the Analects of Confucius, wherein the belief that use of words should reflect reality is asserted. When it does not, the people cannot be properly governed. Xun Zi (c. 313–238 BCE), regarded as China’s first major philosopher of language, contributed to this debate in a work titled Zheng Ming (Rectification of names), which addresses the origin of names, in terms similar to those found in Plato’s Cratylus. Unlike Cratylus, however, the main focus does not concern whether language is to be trusted as epistemology. In Xun Zi’s view, names are established by convention, and have no inherent correctness. They thus cannot represent absolute standards for human behavior. Hsü Shen (58–147 CE) compiled the Shuo wen jie zi (first century CE, Explanations of simple graphs and

CHINESE AND JAPANESE TRADITIONAL GRAMMAR analyses of composite graphs), the first major lexicon, and one of the most significant works of the Chinese tradition. The 9,353 graphs contained in the work are organized under 540 radicals. One reason for its landmark status derives from the fact that this is the first surviving account of the six principles of formation of the Chinese graph, including that of semantic radical plus phonetic element, which accounts for more than 90% of modern graphs. A poet, Shen Yue (441–513), is traditionally credited with the first analysis of Chinese tones, recognizing and naming four types. The tradition of rhyme books is thought to have originated as an aid to composition of poetry, and the technique of fan qie (reverse cutting) used therein prefigures twentieth-century notions of the syllable as composed of an internal structure, with onset and rhyme. The purpose of the rhyme books was to indicate the pronunciation of the monosyllabic graphs of Chinese, and did so by first making four major tonal divisions, and then within each tonal section, listing graphs in charts arranged by initial sound, and by the remainder, or rhyme. The so-called ‘cutting’ involved indicating pronunciation of a given graph by providing two other graphs having the same initial and rhyme portions, respectively. The Qie yun (601 CE), compiled by Lu Fayan, is the earliest major example of the genre. Rhyme tables, the earliest known example of which is the Yunjing (twelfth century, Mirror of rhymes), provided a more detailed classification of all of the syllables of Chinese. In these works, initials are arranged according to articulatory categories such as ‘tongue sounds’, or dentals, and characterized in terms of phonation by such terms as ‘clear’ (voiceless) or ‘muddy’ (voiced). In Japan, early linguistic studies were concerned with determining the sound correlates to different sets of phonograms in Old Japanese, which had merged into single sounds in later stages of Japanese. Notable in their efforts in this regard were the poet Fujiwara Teika (1162–1241) and Buddhist monk Keichû (1640–1701). Analysis of morphology and syntax in traditional Japanese grammar is largely a concern with tactics: the combinatorial properties and functions of various verbal and adjectival suffixes, on the one hand, and postpositions or particles, on the other. Fujitani Nariakira (1738–1779) created the first comprehensive grammatical framework, using the metaphor of clothing: he categorized nouns as ‘names’, verbs and adjectives as ‘clothes’, adverbs as ‘hairpins’, and postpositions and suffixes as ‘binding cords’. His bestknown works are the Kazashishô (1767, On hairpins) and the Ayuishô (1773, On binding cords). In his Teniwoha himokagami (1771, Survey of particles), Motoori Norinaga (1730–1801) explicated the

ways in which certain emotive particles trigger concord with inflectional endings in the verb, and coined the term kakari-musubi, still in use today, to refer to it. In Mikuni kotoba katsuyôshô (1782, On inflections in our language), Motoori laid the groundwork for understanding the patterns of verbal conjugation. Suzuki Akira (1764–1837), student of Motoori Norinaga, in his Gengo shishûron (1824, On the four categories in language), breaks down parts of speech into nouns, verbs (activities), and adjectives (states), on the one hand, and particles and bound suffixes, on the other. Underlying this categorization is a conceptualization of content vs. function words, said to have been drawn under the influence of such a classification for Chinese. These grammatical studies from the early modern period fall under the scholarly domain of kokugaku (nativism), while the work of those continuing in the same vein in the twentieth century is termed kokugogaku (national language studies), as separate from the western-influenced discipline of linguistics (gengogaku). The most influential twentieth-century figure to continue along nativist lines was Tokieda Motoki (1900–1967), who expanded the framework to encompass all languages, without restriction to Japanese. He added a subjective, performative dimension to the bifurcation of content words, or shi (in his view, objectively conceptualized) vs. function words, or ji, which express the emotions, attitude, or judgments of the speaker. Under this view, content words are nested within function words in an iconic display, which leaves the subjective elements rightmost at each level of grammatical construction. Major work on Japanese dialects was carried out by the folklorist Yanagita Kunio (1875–1962) in his Kagyûkô (1927, Thoughts on ‘Snail’), where he posited the ‘circle theory’ of linguistic diffusion, and linguist Hattori Shirô (1908–), in his investigations of the geographical distribution of accent patterns. The National Language Research Institute has published two major dialect atlases: the Linguistic atlas of Japan (1981–1985) and the Grammar atlas of Japanese dialects (1989–). References Baxter, William H. 1992. A handbook of old Chinese phonology. Berlin and New York: Mouton de Gruyter. Chaudhuri, Saroj Kumar. 1998. Siddham in China and Japan. Sino-Platonic Papers, No. 88. Furuta, Tôsaku, and Hiroshi Tsukishima. 1972. Kokugogaku shi (History of Japanese language studies). Tokyo: Tokyo Daigaku Shuppan Kai. Hansen, Chad. 1983. Language and logic in ancient China. Ann Arbor: The University of Michigan Press. Malmqvist, Göran. 1994. Chinese linguistics. History of linguistics, Vol.1: The Eastern traditions of linguistics, ed. by Giulio Lepschy. London and New York: Longman.

199

CHINESE AND JAPANESE TRADITIONAL GRAMMAR Miller, Roy Andrew. 1975. The Far East. Current trends in linguistics, Vol.13: Historiography of linguistics, Part 2, ed. by Thomas A. Sebeok. The Hague and Paris: Mouton. Motoori, Norinaga. 1997. Kojiki-den, Book 1, translated by Ann Wehmeyer. Ithaca, New York: Cornell East Asia Series. Ramsey, S. Robert. 1982. Language change in Japan and the odyssey of a Teisetsu. Journal of Japanese Studies 8.

Ramsey, S. Robert. 1987. The languages of China. Princeton, NJ: Princeton University Press. Sakai, Naoki. 1992. Voices of the past: the status of language in eighteenth-century Japanese discourse. Ithaca, New York: Cornell University Press.

ANN WEHMEYER See also China; Chinese (Mandarin); Japanese

Chinese Pidgin English Were Chinese Pidgin English not widely known by that name, it would be preferable to use the alternative, China Coast Pidgin. While having significant influence, English alone did not provide the core of conventions that came to constitute the language. Instead, there was convergence of English and Cantonese features in a process of mutual accommodation. Indeed, the hybrid nature of the structures of Chinese Pidgin English makes it a crucial piece of evidence in debate within creole linguistics, reflecting a shift away from particularist notions of pidgins and creoles to one in which they are considered merely relative, although related, forms of contact languages. Chinese Pidgin English originated during the first locally regulated phase of trade between Chinese and English in the late seventeenth century (c. 1689–1748). It found a stable form in and around Canton for about another century (1748–1842), and was thereafter disseminated further up the Chinese coast during the remainder of the nineteenth century, and beyond by the Chinese diaspora to the Pacific and even the United States (1842–1890). Finally, Chinese Pidgin English fell into obsolescence during the twentieth century when it was heavily anglicized and lost many of its original structures as numerous speakers came to target English as a lingua franca. Unfortunately, with the exception of slender second-hand evidence from the first period of genesis, mostly in European travel accounts, almost all the data we have fall between 1836 and 1901, and, heavily anglicized, are not reliable. The early data, however tenuous, nonetheless reflect a process of mutual accommodation in which speakers had to make guesses about what their interlocutors would understand and ‘right’ guesses would be incorporated into the grammar of the developing contact language. Balanced against the predominance of English etymons in Chinese Pidgin English is the relative impor-

200

tance of Cantonese syntax and phonology, such that to most English speakers, Chinese Pidgin English was hardly more intelligible than Chinese itself. There is convincing documentary evidence that the right guesses that set the parameters of Chinese Pidgin English took place within the trade factories at Canton between 1699 and 1748, where its inchoate forms were the instrument of limited exchange between the ‘supercargoes’ or British East India Company property agents, and the Cantonese-speaking domestics assigned to them by their authorized ‘Hong merchant’ interlocutors. One objection to such a scenario is the undeniable impact of Portuguese on Chinese Pidgin English, one that might lead us to suspect that its origins go back to the arrival of the Portuguese along the south China coast in 1557—English ‘Canton’ is derived from the Portuguese pronunciation of Guangzhou. Some of the earliest fragments of putative Chinese Pidgin English appear to contain more Portuguese than English, which would tend to point toward a Portuguese base, e.g. this phrase from 1748: (1) Carei grandi hola, pickenini hola? Want large whore, small whore Quer grande [puta], pequena [puta] (Ptg) ‘Do you prefer a big or a small prostitute?’ There is no doubt that Portuguese or some pidginized form of it was spoken along the China coast up to and throughout the time Chinese Pidgin English was formed. In Macao, there eventually formed Macaista, a variety of creolized Portuguese. Maritime and pidginized Portuguese were used from the Indian subcontinent across through Malaysia and Indonesia. As for Canton itself, we know that the first requirement of a supercargo on English ships trading to China was a knowledge of Portuguese. Yet, in the final analysis it is unlikely that a prior Portuguese pidgin served as the base of incipient

CHINESE PIDGIN RUSSIAN Chinese Pidgin English, not so much because the Portuguese contribution was limited to an ever-decreasing percentage of its lexicon but rather because the role of Cantonese is predominant both in pronunciation and word structure, from which fact we can suppose that the majority of early pidgin learners were native Cantonese speakers who adopted the vocabulary of the supercargo masters to whom they were assigned. The Canton-based pattern of trade that formed Chinese Pidgin English continued until the end of the eighteenth century. In 1796, the Emperor banned the opium trade, at which point the English turned to smuggling. In 1840, a second attempt was made to ban opium, which led to the Opium War and the Nanking Treaty of 1842, opening the doors to China. The focus of contact shifted from Canton to the Treaty cities and to Hong Kong. As the monopoly of the Hong Kong merchants broke down, contacts began to proliferate between aspiring Chinese entrepreneurs and the quickly increasing number of foreigners in south China. The habit of resorting to Chinese Pidgin English was carried into new sites of trade where the status and practice of the language changed. As speakers of standard English began entering the Chinese community from the late nineteenth century onward, Chinese Pidgin English became increasingly associated with the servants of foreigners, and was deprecated. Eventually, Chinese Pidgin English was conveyed first as far north along the coast as Shanghai and Tientsin, British

traders taking their Cantonese servants with them as China was forced open to trade. After economic conditions worsened throughout the nineteenth century, there was a large Cantonese diaspora to California, Hawaii, Australia, and Singapore. At the same time, the core of Chinese Pidgin English, learned and transmitted as an auxiliary second language for several generations, began to dissolve. It was this late period of its history, roughly from 1842 onward, that gave rise to stereotypes and misunderstandings, which have tended not only to mask the originality of this language but also to confuse it with all forms of imperfectly learned English, which it was not. References Bauer, A. 1975. Das Kanton-Englisch: ein Pidginidiom als Beispiel für ein Kulturkontaktphänomen. Bern: Peter Lang. Holm, John. 1988–1989. Pidgins and Creoles, 2 vols. Cambridge: Cambridge University Press. Lang, George. 2000. Hardly more intelligible than Chinese itself: a brief account of Chinese Pidgin english. Asian Englishes 3(1). 21–38. Morse, Hosea Ballou. 1966. The chronicles of the East India Company trading to China, 1635–1834. Taipei: Ch’eng-Wen Publishing Company. Shi, Deng-Xi. 1991. Chinese Pidgin English: its origins and linguistic features. Journal of Chinese Linguistics 19. 1–40. Thomason, Sarah Grey, and Terrence Kaufman. 1988. Language contact, creolization, and genetic linguistics. Berkeley: University of California Press.

GEORGE LANG

Chinese Pidgin Russian Not surprisingly, the Russian expansion across the Ural mountains into Siberia and America led to a number of contact languages to a greater or lesser extent lexically based on Russian. We have documentation of such varieties from four locations—Kyakhta (on the Mongolian border), Manchuria, the Taymyr Peninsula, and in the Ussuri area just north of Machuria. The respective pidgins are hereafter abbreviated as KPR, MPR, TPR, and UPR. It is also possible that further varieties exist or have existed in Siberia, as well as in areas of (attempted) Russian colonization in Alaska, California, and Hawaii. In addition, mention will be made of Russenorsk (RN), a trade pidgin of the Arctic Ocean based partly on Russian and partly on Norwegian. I will not, however, discuss Mednyj Aleut, spoken in

the Strait of Bering, since, although a contact language, it is not a pidgin (but rather a variety of Aleut with Russian verbal morphology). MPR emerged in the very beginning of the twentieth century, when Russians managed the Manchurian railway. It began to decline after World War II, when both parties took an increasing interest in learning the other’s language, and definitively dropped out of use when the new political climate put an end to Sino-Russian friendship. People who remember the pidgin are still alive, however; thus, research is still possible, but urgent. KPR emerged as a result of trade between Russians and Chinese in the eighteenth century. The pidgin has been considered extinct, but in 1990 some speakers were found among Chinese merchants in the Mongolian capital of Ulan Bator.

201

CHINESE PIDGIN RUSSIAN TPR, also called Govorka, is used between speakers of Samoyedic, Turkic, and Tungusic languages on the Taymyr Peninsula in northernmost Siberia. It has been declining in favor of Russian since World War II, and is now spoken only by elderly people. Fortunately, research on TPR is currently carried out by German linguist Dieter Stern. UPR is the least well documented of the Russian pidgins, at it is known virtually exclusively through accounts of Russian expeditions to the area in the late nineteenth and early twentieth centuries. Russenorsk, finally, was used in the trade between Russians and Norwegians in the border area until the beginning of the twentieth century. There is a relative wealth of material on this pidgin, which has been analyzed primarily by Norwegian linguists. Although Russenorsk and TPR are no doubt local developments, it should be emphasized that the relationship between the Far Eastern varieties is unclear, and, indeed, many authors treat them all as ‘Chinese Pidgin Russian’. In most of the pidgins, there is a tendency toward syllabic simplification, favoring CV syllables. Other phonological aspects are highly variable, as they display influences from surrounding languages, such as the devoicing of Russian plosives in MPR, the replacement of /f/ with /p/ in KPR, and the merger of /s, z, ʃ, tʃ  / in TPR. Lexically, some of the varieties are remarkably mixed. This is particularly true for RN, in which only about half of the vocabulary was of Russian origin, and MPR, where approximately one third of the lexicon was non-Russian (mainly from Mandarin). Some lexical influence from local languages is reported for the other varieties as well. In addition to the lexical items themselves, some calquing from substrate languages has also been observed. As in all pidgins, the small lexicon entails an exceptional degree of polysemy and multifunctionality, the serious semantic underspecification leading to a strong reliance on context for disambiguation. As for the derivation of the Russian forms, it is noteworthy that the verbs are often developed from imperatives rather than from infinitives, as in pidgins lexically based on Western European languages. The basic word order is of particular interest, for only MPR was consistently SVO, like both Russian and Mandarin. RN was basically SVO, but used SOV with adverbials. The other varieties are predominantly SOV, presumably due to substrate influences. TPR is also spectacular in having the order REL N. Another un-Indo-European feature is the optional lack of fronting of interrogative pronouns. As would be expected from SOV languages, postverbal auxiliaries are attested.

202

Tense/Mode/Aspect (TMA) marking is usually rather restricted. TPR encodes tense, but not aspect. If the little data available there are indeed reliable, UPR is remarkable, and possibly unique among pidgin langauges, in having a grammaticalized evidentiality marker. This, however, seems to have been the only TMA category in the language. Russenorsk had no grammaticalized TMA marking at all, whereas the other Far Eastern varieties seem to have made optional use of nonbound morphemes for this purpose. Negation is Russian style in MPR, but postverbal in the other Asian varieties. Russenorsk had a preverbal negation, or rather two of them, which usually occupied the second position in the sentence. Nominal number is usually left unmarked, although TPR has optional morphemes derived from Russian words meaning ‘much’ and ‘they all’. Grammatical gender has not survived in any of the varieties. One of the shibboleths of pidginhood is a limited adpositional inventory. Russenorsk had po as its only preposition, corresponding for the most part to za in the Far East. These items expressed more or less any spatial relationship imaginable. The same job is carried out by mesto in TPR, which, however, also has a sociative adposition meste. It is noteworthy that adpositions in TPR are post- rather than prenominal. Reduplication is absent from Russenorsk and MPR, but has been reported for KPR. As in many other pidgins, juxtaposition is frequently used to indicate possessive relationship, and in a similar manner, clauses are to a great extent joined paratactically without any overt conjunction or subjunction (although such clauses do exist). Again, as one would expect from pidgins, bound morphology is scarce, although there are a couple of potential candidates for the status of bound morphemes. The pronominal systems also merit mention. The forms themselves, invariable with regard to syntactic function, are mostly derived from Russian genitives, possibly a carryover from Russian foreigner talk. Exceptions to this are the TPR forms, and the 2sg form elsewhere, which derive from Russian accusatives. They have been subject to some rather far-reaching restructuring, however, in particular in TPR. There, only the singular forms have been taken over, with additional morphemes marking number. Even more conspicuous is the presence of an inclusive/exclusive distinction for nonsingular pronouns. References JabBonska, Alina, and Anatole Lyovin. 1969. The Sino-Russian mixed language in Manchuria. University of Hawaii Working Papers in Linguistics 1. 135–64.

CHINOOK JARGON Neumann, Günther. 1966. Zur chinesisch-russischen Behelfssprache von Kjachta. Die Sprache 12. 237–51. Nichols, Johanna. 1980. Pidginization and foreigner talk: Chinese Pidgin Russian. Papers from the 4th International Conference on Historical Linguistics, ed. by Elizabeth Closs Traugott, Rebecca Labrum, Susan Shepherd, and Paul Kiparsky, 397–407. Amsterdam: Benjamins. Nichols, Johanna. 1986. The bottom line: Chinese Pidgin Russian. Evidentiality: The linguistic coding of epistemology, ed. by Wallace Chafe and Johanna Nichols, 239–57. Norwood: Ablex. Peter Mühlhäusler, Darrell T. Tryon, and Stephen A. Wurm (eds.) 1997. Atlas of languages of intercultural communica-

tion in the Pacific, Asia, and the Americas. Berlin and New York: Mouton de Gruyter. Wurm, Stephen. 1992. Some contact languages and Pidgin and Creole languages in the Siberian region. Language Sciences 14(3). 249–85. Wurm, Stephen. 1996. The Taimyr Peninsula Russian-based pidgin. Language contact in the Arctic. Northern Pidgins and contact languages, ed. by Ernst Håkon Jahr and Ingvild Broch, 79–90. Berlin and New York: Mouton de Gruyter. Xelimskij, Evgenij. 1996. ‘Govorka’ — the pidgin Russian of the Taymyr Peninsular Area, In Mühlhäusler Tryon and Wurm, pp. 1033–4.

MIKAEL PARKVALL

Chinook Jargon Chinook Jargon is a pidgin language of the Pacific Northwest that is first attested reliably from the first decade of the nineteenth century, in the journals of Lewis and Clark; the earliest extensive documentation is by Horatio Hale (1846). Its lexifier language—the language from which most of its vocabulary is drawn—was Lower Chinook (Shoalwater), the language of a once-powerful tribe at the mouth of the Columbia River. At its peak, Chinook Jargon was spoken as far north as southern Alaska, as far south as the northern border of California, and as far east as the Idaho panhandle in the United States and interior British Columbia in Canada. It flourished especially between c. 1850 and 1950, when it was a primary medium of communication between Whites, on the one hand, and Native Americans and Native Canadians on the other. Its use declined sharply in the late twentieth century as English replaced it everywhere in the Northwest. As late as 1980, however, monthly sermons were delivered in Chinook Jargon in at least one British Columbian church, and a few elderly fluent speakers on the Grand Ronde Reservation in Oregon contributed to its survival by teaching it to younger tribal members. Efforts are currently under way to revitalize Chinook Jargon, especially at Grand Ronde. By the late twentieth century, Chinook Jargon was the main or only Native language spoken on the reservation, where members of six Native tribes descended from speakers of three completely different indigenous language families. As Henry Zenk has noted (1984), Chinook Jargon became ‘an important factor in the sense of identity and solidarity that many Natives of the reservation period came to feel as “Grand Ronde Indians”.’ Chinook Jargon thus per-

formed its traditional function as a lingua franca even toward the end of its continuous existence. Like other classic trade pidgins, Chinook Jargon has a limited vocabulary. Over 600 Chinook Jargon words are reliably attested—that is, they occur in at least two independent sources, and usually in more than two. Of these, perhaps a third come from French or English, reflecting the widespread use of Chinook Jargon as a lingua franca between Whites and Natives in the late nineteenth and early twentieth century. The earliest Chinook Jargon sources, from the first half of the nineteenth century, also list English and French words, but not nearly as many as in the later documentation. A rough count of c. 650 Chinook Jargon words must certainly omit many that were in common use: even the most ambitious dictionaries inevitably omit many of a language’s words, and the published Chinook Jargon dictionaries (e.g. Shaw 1909) are not especially ambitious. Chinook Jargon phonemes are typical for an indigenous language of the Pacific Northwest. More than 30 consonant phonemes are attested, in a few to numerous words, in at least two independent sources each. Many of the consonants are unfamiliar to speakers of English and other European languages: glottalized (specifically, ejective) stops and affricates, uvular as well as velar obstruents, a glottalized lateral affricate, and a lateral fricative. Moreover, Chinook Jargon words often contain consonant clusters that are completely foreign to European languages, e.g. in /tk’up/ ‘white, light in color’, /ptSix/ ‘thin’, and /tL’m@n/ ‘soft, ground up’. Chinook Jargon thus presents a striking counterexample to the often-repeated claim that pidgin structures are maximally simple, and the consistency of many of its phonological features

203

CHINOOK JARGON across Native speakers from numerous tribes offers a counterexample to the common claim that pidgin structures show more internal variation than nonpidgin languages. To judge by the surviving documentation and the few existing tape recordings from the midtwentieth century, most White speakers of Chinook Jargon did not learn either the non-European sounds or the non-European consonant clusters of the pidgin (but see Demers et al. (1871) for a notable exception). Also, like other classic trade pidgins, Chinook Jargon has limited morphosyntactic resources. It entirely lacks the complex morphological (word) structures that characterize Native languages in the region, and its range of syntactic constructions is not large. As with the sound system, its syntactic structures closely resemble those of Native Northwest languages—with one possible major exception, the dominant S(ubject)–V(erb)–O(bject) word order. Most Northwest languages are verb-initial; Chinook Jargon syntax is verb-initial only with adjectival predicates, e.g. _Hayas ulu tsuq nayka_ ‘I am very thirsty’ (lit. ‘much hungry water I’). Besides the SVO word order, the constructions are clearly Native, not based on either English or French. They include, among others, sentence-initial negation, yes/no questions formed with an optional question particle, and an imperative construction (‘it would be good if you would do X’). The question of the pidgin’s origin remains highly controversial. One hypothesis is that Chinook Jargon predates extensive contact with Whites in the Northwest—that it was used as a means of intertribal communication, perhaps at first between speakers of Lower Chinook and their Native slaves. (‘Slave jargons’ are reported elsewhere in the Northwest, for instance among the Nez Perce.) Based on this theory, Chinook Jargon achieved its later spread when Whites adopted it for use as a lingua franca, shortly after 1806. The second origin hypothesis is that Chinook Jargon arose as a lingua franca only after Whites arrived in large numbers in the Northwest. A common feature of this theory is the proposal that a Nootka trade jargon (or pidgin) arose first on Vancouver Island, at the end of the eighteenth century or early in the nineteenth century, and then spread to the mouth of the Columbia River when Whites ventured there for trade; the Nootka trade jargon/pidgin was then relexified through the replacement of most Nootka words by Chinook words. Evidence adduced in support of the second theory is primarily lexical. Most of the lexicon of early Chinook Jargon, including the bulk of the basic vocabulary, comes from Lower Chinook, but two or three dozen words (some of them quite basic) are from Nootka, and a smaller number of words come from Salishan and other languages of the region. In addi-

204

tion, quite a few words entered the pidgin from French and then English, especially after about 1850. The hypothesized contribution of the Nootka Jargon to the initial formation of Chinook Jargon is predicated on the assumption that such basic words would necessarily have been in Chinook Jargon from the beginning, not added later after Chinook Jargon was fully formed. Against this assumption, however, is the undoubted fact that, while the Nootka words in Chinook Jargon show clear signs of transmission from Whites, the Chinookan and Salishan words in Chinook Jargon were clearly transmitted from Natives to other Natives. Words of Chinookan and Salishan origin show all the elaborate features of typical Northwest phonological systems, including glottalized stops, velar vs. uvular dorsal obstruents, and lateral fricatives and affricates. This is not the case with Nootka-origin words: they are significantly distorted in comparison to their Nootka source words, with virtually no sounds that would be foreign to English and French speakers. The transmission of some of the Nootka jargon/pidgin words to Chinook Jargon is therefore much more likely to have occurred after, not before, the crystallization of Chinook Jargon as a pidgin language. Both origin theories are plausible. The postcontact theory has been popular in part because Chinook Jargon, from the time it was first documented, already had the Nootka words and a fair number of French and English words. Also, of course, all Chinook Jargon documentation is necessarily postcontact, since Natives in the Northwest had no writing before Whites arrived. The precontact origin theory is preferable if one adopts the standard simplicity criterion of historical linguistics: the Native phonology and syntax are easily accounted for if Natives created Chinook Jargon without significant participation by Whites, but if the pidgin arose postcontact, with some French- and/or English-influenced structural features, those features must have been lost before the pidgin was documented. Based on this criterion, the precontact origin is the simpler hypothesis. But not all pidgin/creole specialists accept this criterion; hence, the controversy continues. The future of Chinook Jargon is in some doubt, because the Grand Ronde elders and others who spoke the language as part of their ordinary daily lives are now gone. Its fate rests with the younger enthusiasts, especially younger tribal members, who are now working to prevent the pidgin from disappearing. References Demers, Modeste, F. N. Blanchet, and L. N. St. Onge. 1871. Chinook dictionary, catechism, prayers, and hymns. Montreal: Quebec Mission.

CHOMSKY, NOAM Hale, Horatio. 1846. United States exploring expedition during the years 1838–1842; ethnography and philology. Philadelphia: Sherman. Samarin, William J. 1986. Chinook Jargon and pidgin historiography. Canadian Journal of Anthropology 5. 23–34. Shaw, George C. 1909. The Chinook Jargon and how to use it. Seattle: Rainier Printing Co., Inc. Splawn, A.J. 1944. Ka-Mi-Akin: last hero of the Yakimas, 2nd edition. Yakima, WA, Caldwell, Idaho: The Caxton Printers; 1st edition, 1917.

Thomason, Sarah G. 1983. Chinook Jargon in areal and historical context. Language 59. 820–70. Zenk, Henry Benjamin. 1984. Chinook Jargon and Native cultural persistence in the Grand Ronde Indian community, 1856–1907: a special case of creolization. Eugene, OR: University of Oregon dissertation.

SARAH G. THOMASON See also Language: Contact—Overview; Pidgins and Creoles

Chomsky, Noam Noam Chomsky is one of the most profound and influential thinkers of our time, ‘arguably the most important intellectual alive’ (New York Times Book Review, February 25, 1979), an evaluation in accord with the fact that he is the world’s most cited living author in several citation indexes. In his early years, Chomsky read Hebrew literature with his father, a professor of Hebrew and Jewish education at Gratz College in Philadelphia (and president of its faculty for 45 years), considered ‘one of the world’s foremost Hebrew grammarians’ (New York Times obituary, July 22, 1977). His immersion in the Jewish cultural tradition was deep. Out of this experience, his budding political interests converged toward what was then mainstream Zionism (now widely considered ‘anti-Zionism’). Politics brought Chomsky into linguistics. As a teenager, he was deeply interested in radical politics with an anarchist or left-wing Marxist flavor (strongly antiLeninist, and more generally anti-Bolshevik). Through these political interests, he met Zellig Harris, whom Chomsky has described as ‘a person of unusual brilliance and originality’ and very broad interests. Harris was not only an acute left-libertarian thinker and analyst but also a leading figure in linguistics, teaching at the University of Pennsylvania, and the most rigorous practitioner of the reigning structuralist methodology. Harris was surely the linguist best prepared to initiate someone with Chomsky’s mind and inclinations into the field of linguistics (as then understood). He also interested him in the study of philosophy, logic, and mathematics, fields that were to open for him the avenue to major discoveries. The first reading Chomsky did in linguistics, before he had taken any courses, was the proofreading of Harris’s important Methods of structural linguistics (1951), the most exacting exposition of immediately preChomskyan linguistic theory. The Preface (dated

January 1947) records that ‘N. Chomsky has given much-needed assistance with the manuscript.’ Since the appearance of his first published book, Syntactic Structures, in 1957, Chomsky has been recognized as ‘an eminent and revolutionary scholar in the field of linguistics’, to quote from a representative blurb. Chomsky almost single-handedly assimilated psychology to the natural sciences by formulating ‘transformational generative grammar’ (a model for the ‘cognitive sciences’, of which it was the first). This theory takes for granted the now widely (although not unanimously) held view that all human languages operate under the same general principles in spite of their superficial diversity. Particularly important in this context is that the universal foundation for language is taken to be inherent in (and an essential part of) ‘human nature’, this being the central notion in both of Chomsky’s main endeavors. Thus, Chomsky rejected behaviorism, unequivocally so in his renowned 1959 review of B.F. Skinner’s 1957 Verbal behavior, a review that demonstrated the inability of the behaviorist approach to account for any but the most superficial aspects of human language. Chomsky redefined the nature and scope of linguistics. As the Encyclopedia Britannica (1992) says, ‘... there is no major theoretical issue in linguistics today that is debated in terms other than those in which he has chosen to define it’. A recent book on the future of science describes him, accurately, as ‘the most important linguist who has ever lived’. As John Lyons says in Chomsky (1970), he has spoken ‘with unrivaled authority in all aspects of grammatical theory’ since Syntactic structures ‘revolutionized the scientific study of language’, fundamentally changing the then current understanding of language and the mind/brain. Chomsky’s most fundamental contribution has been to open the way for the cognitive natural sciences. His

205

CHOMSKY, NOAM specific model for human language (transformational generative grammar) can be regarded as a confluence of traditional (and long-forgotten) concerns of the study of language and the mind (as in the work of Wilhelm von Humboldt or Otto Jespersen) and new understanding provided by the formal sciences in the late 1930s, particularly recursive function theory. The basic idea is that a sentence is the result of a computation producing a ‘derivation’, beginning with an abstract structural representation sequentially altered by structure-dependent operations (‘transformations’). Chomsky’s unusual knowledge of philosophy, logic, and mathematics allowed him to make this model precise, and to go on to develop, in the late 1950s, algebraic linguistics, a branch of abstract algebra, which is now part of computer science. (For Israeli logician and mathematician Yehoshua Bar-Hillel, one of its most eminent practitioners, Chomsky was not just the founder of algebraic linguistics but also ‘by far the best man in this exciting new field’.) Chomsky’s written output is truly prodigious, encompassing, in linguistics alone, dozens of books and nearly 200 articles (and several times that in political analysis). His first substantial contribution to the development of generative grammar was his B.A. thesis (1949), an examination of certain morphophonemic alternations in Modern Hebrew, revised and expanded in his M.A. thesis (1951), and revised once more that year (that version was eventually published in 1979). Although this work had no impact on the field of linguistics at the time, it did introduce numerous technical devices that would eventually be of major significance, including abstract underlying syntactic and phonological forms and crucially ordered rules deriving surface forms from these underlying forms. Significantly, already in this early work, Chomsky was emphasizing that a grammar is a finite characterization of an infinite set of sentences. This theme continues to be of profound importance in linguistics. Over the next few years, while a Junior Fellow of the Society of Fellows at Harvard University, Chomsky wrote The logical structure of linguistic theory (LSLT), a monumental work (completed in 1955, but published, and only in part, 20 years later) that laid out the formal basis for a complete theory of linguistic structure. The concepts and technical notions that were to become central to theoretical linguistics for the next several decades were developed in this paper, including the crucial idea of abstract underlying structure. Made precise are such concepts as level of representation, phrase structure rule, phrase marker, grammatical transformation, and derived constituent structure. LSLT is also rich in conceptual and empirical arguments, and it provides an extraordinarily detailed account of the syntax of English. Many of the patterns

206

Chomsky explained had never even been noted before. Further, many of the analyses he presented have yet to be surpassed. However, like the M.A. thesis, LSLT had no direct impact on the field at the time. Within two years, though, LSLT had a revolutionary indirect influence, since the material contained in it formed the basis for the lecture notes that were published in 1957 as Syntactic structures, a book whose unprecedented influence was noted above. The year 1965 witnessed the publication of Chomsky’s next major book, Aspects of the theory of syntax, which made fully explicit the role of linguistics in the investigation of the human mind. Its first chapter, originally written during the year Chomsky spent at the Institute for Advanced Studies at Princeton (1958–1959), remains one of the clearest and most forceful expositions of the major goal of generative theorizing: providing an account of how the child presented with limited evidence develops a computational system making possible the production and comprehension of an unlimited number of brand new sentences (‘explanatory adequacy’). The book laid out what came to be called the ‘standard theory’ of syntax. In this theory, the syntactic derivation of a sentence begins with deep structure (DS), a representation constructed via phrase structure rules and lexical insertion rules. Grammatical relations are fully determined at this level. Transformations then successively, and cyclically (‘bottom up’), modify this representation, eventually producing surface structure (SS), the input to the phonological component. This contrasts with the LSLT model, wherein the recursive component is the transformational one: ever larger structures are created not by phrase structure rules but by generalized transformations, which embed one structure inside another. (Interestingly, in Chomsky’s ‘minimalist program’ of the 1990s, there is a return to generalized transformations.) Versions of the standard theory (including the ‘extended standard theory’ and the ‘revised extended standard theory’) dominated syntactic theorizing for more than two decades. While Chomsky is most closely associated with syntactic theory, much of his early work was concerned with phonology, culminating in the groundbreaking book in 1968, The sound pattern of English (co-authored by Morris Halle), which was a detailed study of the phonology of English, but, just as importantly, a full-blown theory of the phonological component of linguistic theory and its interaction with the syntactic component. Also noteworthy is the elaborated outline of a theory of markedness, which sought to explain why certain phonological processes are much more common than others. Chomsky’s concern with explanatory adequacy led in due course to the ‘Government-Binding’ model,

CHOMSKY, NOAM articulated in Lectures on government and binding (1981). In this model, the language (and dialect)-specific and construction-specific rules of earlier frameworks are replaced by operations of great generality whose functioning is constrained by universal conditions. The rules and conditions are ‘principles’, and the limited range of variation available for them constitutes the ‘parameters’. (Chomsky came to prefer the name ‘Principles and Parameters’ for the model, reasoning that government and binding are just two among many technical devices in the theory, and not necessarily the most important ones.) The P&P model—the first one ever to suggest a substantive solution to the fundamental problem of language acquisition—represents a radical break from the rich tradition of thousands of years of linguistic inquiry. It is true, though, that the entire (modern) generative grammar period is in many ways a new era. The P&P model was in turn further refined in Chomsky’s ‘minimalist program’ of the 1990s and early 2000s (see Chomsky (1995), the second chapter of which was written and circulated in 1988, based on lectures in 1986 and 1987). The P&P model had DS, SS, Logical Form (LF), and Phonetic Form (PF) as significant levels of representation. Given that a human language is a way of relating sound (or, more generally, gesture, as in sign languages) and meaning, minimalism seeks to establish that there are no levels except the ‘interface’ levels PF and LF. In the most recent developments of the theory, Chomsky suggests, in another partial return to an earlier formulation, that even PF and LF, as specific levels of representation in the technical sense, do not exist. Rather, throughout the derivation, the syntactic structure thus far created is encapsulated and sent off to the interface components for phonetic and semantic interpretation. The minimalist program further maintains that derivations and representations conform to an ‘economy’ criterion demanding that they be minimal in a sense determined by the language faculty: no extra steps in derivations and no extra symbols in representations. A major technical goal is to reduce all constraints on representation to ‘bare output conditions’, determined by the properties of the mental systems that the linguistic computational system must interface with that are external to the computational system itself. A direct consequence of Chomsky’s scientific work is that it provides strong evidence in support of epistemological rationalism, as he was quick to point out. Early on, he turned to the serious study of the Cartesian tradition, which he was to revive and update, shortly after he made his initial discoveries. These discoveries made it possible for him to go well beyond the programmatic insights of the Cartesians, and give substance to their central claims, in the process reconstructing the enduring ideas of the first phase of the

age of modern philosophy (a not always recognized antecedent of the cognitive revolution of the 1950s), on which Chomsky has shed much light as an intellectual historian. It seems fair to describe him as the scholar who is to the period initiated by the cognitive revolution of the mid-1950s what Descartes was to the first phase of the age of modern philosophy.

Biography Noam Chomsky was born in Philadelphia, Pennsylvania on December 7, 1928. His father, William Chomsky, was a noted Hebrew scholar who wrote Hebrew, the eternal language. While studying at the University of Pennsylvania under the supervision of Zellig Harris, he was also a Junior Fellow of the Harvard Society of Fellows from 1951 to 1955. After receiving a Ph.D. in linguistics from the University of Pennsylvania in 1955, Chomsky joined the faculty at MIT under the sponsorship of Morris Halle and was promoted to full professor of Foreign Languages and Linguistics in 1961, appointed Ferrari Ward Professor of Linguistics in 1966, and Institute Professor in 1976. Since 1967, when the University of London and the University of Chicago awarded him his first two honorary doctorates, he has been the recipient of scores of honorary degrees throughout the world. In 1969, he delivered the John Locke Lectures at Oxford University, and in 1970, the Bertrand Russell Memorial Lecture at Cambridge University. Among the almost countless other honors that he has received is the 1988 Kyoto Prize in basic science, created in 1984 for the purpose of recognizing outstanding achievements in categories not named by the Nobel Prizes. In 1984, he received the Distinguished Scientific Contribution Award of the American Psychological Association ‘for enlarging our definition of scientific psychology’. References Anthony, Louise M., and Norbert Hornstein (eds.) 2003. Chomsky and his critics (Philosophers and their critics). Oxford: Blackwell Publishers. Barsky, Robert F. 1997. Noam Chomsky: a life of dissent. Cambridge, MA: MIT Press. Chomsky, Noam. 1957. Syntactic structures. The Hague: Mouton. ———. 1959. A review of B.F. Skinner’s verbal behavior. Language 35. 26–58. ———. 1965. Aspects of the theory of syntax. Cambridge, MA: MIT Press. ———. 1966. Cartesian linguistics. Cambridge, MA: MIT Press. ———. 1972. Problems of knowledge and freedom: the Russell lectures. New York: Random House. ———. 1975. The logical structure of linguistic theory. New York: Plenum.

207

CHOMSKY, NOAM ———. 1979. The morphophonemics of modern Hebrew. New York: Garland. ———. 1981. Lectures on government and binding. Dordrecht: Foris. ———. 1986. Knowledge of language. New York: Praeger. ———. 1995. The minimalist program. Cambridge, MA: MIT Press. ———. 2000. New horizons in the study of language and mind. Cambridge: Cambridge University Press. ———. 2000. Minimalist inquiries: the framework. Step by step: essays on minimalist syntax in honor of Howard Lasnik, ed. by Roger Martin, David Michaels, and Juan Uriagereka, 89–155. Cambridge, MA: MIT Press. Chomsky, Noam, and Morris Halle. 1968. The sound pattern of English. New York: Harper and Row.

Harris, Zellig. 1951. Methods in structural linguistics. Chicago: University of Chicago Press. Jenkins, Lyle. 2000. Biolinguistics: exploring the biology of language. Cambridge: Cambridge University Press. Lyons, John. 1970. Chomsky, 3rd edition. London: Fontana. McGilvray, James. 1999. Chomsky: language, mind, and politics. Cambridge: Polity Press. Otero, C. P. (ed.) 1994. Noam Chomsky: critical assessments (8 tomes). London: Routledge. Smith, Neil. 1999. Chomsky: ideas and ideals. Cambridge: Cambridge University Press.

HOWARD LASNIK AND CARLOS OTERO See also Bar-Hillel, Yehoshua; Computational Linguistics; Grammar, Theories; Structuralism

Clark, Eve V. Educated in France and the United Kingdom, Eve V. Clark started her linguistics career in the United States at Stanford University. She worked on the Language Universals Project led by Joseph Greenberg and Charles Ferguson and became one of the first faculty members of the Linguistics Department. She has worked extensively in various subfields of linguistics including pragmatics, lexical semantics, and psycholinguistics. Her greatest contribution to linguistics and psychology is her groundbreaking research on meaning acquisition—in particular, her earlier work on the semantic feature hypothesis, her comparative studies of children’s word formation, and her current research on adults’ offers of and children’s uptake of lexical and semantic information. In the 1970s, Eve Clark’s research focused mostly on the semantic feature hypothesis. Simply put, the semantic feature hypothesis states that in learning a new word, children only learn some of its semantic features and then add to this knowledge as they find out more about what the word means. In one of her earliest studies, Clark looked at the acquisition of the meaning of before and after. The acquisition order of these two words is as follows: (1) children did not understand either word; (2) they understood before but not after; (3) they interpreted after as if it meant before; and finally, (4) they understood both words correctly. Notice that the meaning of these two words is made up of two components—namely, Time and Prior, and these components can have either a positive

208

or negative value. A positive value of Time shows that the word refers to some aspect of time. Similarly, the values of Prior indicate that one event precedes (positive) and follows (negative) the other. In light of the semantic analysis of these two words, one can see that children do not understand the component Time at first. Once they understand that both before and after contain the component Time, they may still not have acquired the feature Prior. At the third stage, they understand the positive member of Prior (i.e. before) but not the negative member (i.e. after)—in other words, the positive value of Prior is learned before the negative one. Finally, they understand both. Clark showed that this type of analysis could explain the findings of other studies on the acquisition of other relational terms such as more–less, high–low, and tall–short. In the 1980s, Clark’s research focused on word formation in acquisition. With colleagues Ruth Berman, Barbara Hecht, and Randa Mulford, she conducted extensive comparative studies of children’s word formation in English, French, and Hebrew. She proposed several principles—semantic transparency, productivity, and conventionality—that influence children’s acquisition of word-formation devices (e.g. compounding, affixation) in language. For instance, younger English-speaking children (e.g. aged three) often rely on simple compounds (e.g. plant-man) to coin new nouns; their use of -er for agents (e.g. wall-builder) is rather inconsistent.

CLARK, EVE V. This finding is in accordance with the principle of semantic transparency: while the head noun -man is semantically transparent to them (i.e. they know what the noun man means), younger children still need to analyze -er to obtain the agentive meaning of the suffix. As children get older, they replace words that they have coined through compounding with those conventionally used (e.g. plant-man with gardener). This provides evidence for the principle of conventionality: ‘For certain meanings, a conventional word or word formation device exists that should be used in the language community’ (Clark and Berman 1984:549). Another principle is productivity: [T]hose word-formation devices used most often by adults in word innovations are preferred in the language for constructing new word forms’ (ibid.: 548). Empirical evidence shows that children rely on this in acquiring word-formation devices. For instance, among the agentive suffixes -er, -ist, and -ian in English, -er is the most frequent. According to the principle of productivity, children should acquire -er before the others, and this turns out to be the case (see Clark and Cohen 1984). Cross-linguistic studies have shown that regardless of the language being acquired, these principles play a significant role in children’s acquisition and use of word formation. Since the 1990s, Clark’s research has centered on the pragmatics of child-directed speech, especially its relation to children’s meaning acquisition. In acquiring vocabulary, children need to learn to link specific terms with their conventional meanings. In the last few decades, some researchers have proposed that children observe certain built-in constraints in mapping meanings onto word forms, e.g. the mutual exclusivity constraint where each referent is picked out by just one word. However, studies have shown that the constraint approach underestimates the importance of adult contribution. Children as young as two years old can and do make use of lexical and semantic information offered to them by adults. For example, two-year-olds readily accept and use multiple terms for the same referent (e.g. dog and spaniel) when adults provide them with this information. An alternative to the constraint approach is to examine the role of adult-directed speech in meaning acquisition and children’s uptake of adults’ offers of words and relations among words. Clark’s current research goals are to (1) establish the range of adults’ offers of words and relations and (2) determine how effective different types of offers are for word learning, as measured by children’s uptake. The central themes in much of Clark’s research are the emphases on general cognitive principles and on social interaction in language acquisition. While many

studies on meaning acquisition focus on specific constraints, the principles that Clark proposed for meaning acquisition apply to other facets of language acquisition and to adult language use as well. For instance, the pragmatic principles of conventionality and contrast (i.e., a difference in meaning is marked by a difference in form) are not only essential to children’s acquisition of meaning but also guide adult speakers in their uses of language. The explanatory power of these principles reflects the strength of Clark’s research.

Biography Eve V. Clark was born in Camberley, UK, July 26, 1942. She received her M.A. Hons. (1965) in French Language and Literature, with minors in Spanish and Phonetics, University of Edinburgh; Postgraduate Diploma in General Linguistics (1966), University of Edinburgh; and Ph.D. in Linguistics (1969), University of Edinburgh. She was Assistant Instructor and Instructor in the Department of French at the University of Pittsburgh (1967–1969); Research Associate, Language Universals Project, Stanford University (1969–1970); Lecturer, Assistant Professor, Associate Professor, and Full Professor, Stanford University (1970–present); Visiting Scientist, MaxPlanck-Institut für Psycholinguistik Nijmegen, The Netherlands (1981, 1983–1984, 1990–1991, 1997–1998); Fellow, Center for Advanced Study in the Behavioral Sciences, Stanford, California (1979–1980); Guggenheim Fellow, John Simon Guggenheim Memorial Foundation (1983–1984); and Foreign Member, Royal Netherlands Academy of Arts and Sciences (KNAW) (elected 1991). References Clark, Eve. 1971. On the acquisition of the meaning of before and after. Journal of Verbal Learning and Verbal Behavior 10. 266–75. ———. 1973. What’s in a word? On the child’s acquisition of semantics in his first language. Cognitive Development and the Acquisition of Language, ed. by Timothy E. Moore, 65–110. New York: Academic Press. ———. 1983. Meaning and concepts. Handbook of child psychology, Vol. 3: Cognitive development, 4th edition, ed. by John H. Flavell and Ellen M. Markman, 787–840. New York: Wiley. ———. 1991. Acquisitional principles in lexical development. Perspectives on thought and language: interrelations in development, ed. by Susan A. Gelman and James P. Byrnes, 31–71. Cambridge: Cambridge University Press. ———. 1993. The lexicon in acquisition, Cambridge and New York: Cambridge University Press. ———. 1999. Acquisition in the course of conversation. Studies in the Linguistic Sciences 29. 1–18.

209

CLARK, EVE V. Clark, Eve, and Ruth Berman. 1984. Structure and use in the acquisition of word-formation. Language 60. 547–90. Clark, Eve, and Sophia Cohen. 1984. Productivity and memory for newly formed words. Journal of Child Language 11. 611–25. Donaldson, M., and G. Balfour. 1968. Less is more: a study of language comprehension in children. British Journal of Psychology, 59. 461–471. Donaldson, M., and R. Wales. 1970. On the acquisition of some relational terms. Cognition and the development of language, ed. by J.R. Hayes, 235–268. New York: Wiley. Gelman, S., J. Coley, K. Rosengren, E. Hartman, and A. Pappas. 1998. Beyond labeling: the role of maternal input in the acquisition of richly structured categories. Monographs of the Society for Research in Child Development. 63.

Markman, E. 1987. How children constrain the possible meanings of words. Concepts and conceptual development: econological and intellectual factors in categorization, ed. by U. Neisser, 255–287. Cambridge: Cambridge University Press. Markman, E., and G. Wachtel. 1988. Children’s use of mutual exclusivity to constrain the meanings of words. Cognitive Psychology 20. 121–157. Merriman, W., and L. Bowman. 1989. The mutual exclusivity bias in children’s word learning. Monographs of the Society for Research in Child Development. 54.

ANDREW WONG See also Acquisition; Pragmatics; Psycholinguistics; Semantics

Clause Sentences can be of arbitrary length. Every wellformed sentence, however, can be adequately described in terms of its internal structure: the words it contains, the phrases these build and the morphemes they are made up from, and other such relations. Beyond the regular phrase level—noun phrase (NP), verb phrase (VP), etc.—sentences can be divided into clauses. Above all, a sentence is always a clause; but some sentences may consist of more than one clause. In traditional grammar, the distinction is made between main (subordinate) and subordinate (dependent) clauses. In a sentence like John likes Mary, the whole structure is also one clause, but in the slightly longer sentence John likes Mary when she plays football, the boldfaced part is a separate clause all by itself. In generative grammar, the term embedded is used to indicate that the subordinate clause is embedded under the main clause, but still part of the same sentence. This is more than mere notation: traditional grammar does not consider subordinate clauses to be clauses proper; in the generative approach, however, all well-formed sentences are clauses, and every (grammatical) embedded clause is by definition a well-formed structure on the sentential level, hence a clause. The sentence–clause distinction in generative terms (as in the Principles-and-Parameters theory) is often one of categorial projection, selection, and complementation. Two main categories are decisive for the type of structure: the Inflectional Phrase (or IP for short, formerly S) and the Complementizer Phrase (CP or S’). IP denotes the traditional concept that a well-formed sentence consists of a subject and a predicate (often the

210

verb phrase VP). In X’-theory, the canonical subject position is the specifier of IP. The sentence John likes Mary can adequately be described in terms of an IP (and further internal structure). The CP-level is invoked for additional material, such as the question Who does John like?, which is transformationally derived from the former sentence. Here, the subject John occupies the same position in a structural description (i.e. the tree or the phrase-marker) of the sentence as the original sentence. But it is assumed that the object who has moved from its canonical predicate- or VP-internal position to the beginning of the sentence, the specifier of CP. We can thus see that IP and CP can be well-formed main clauses. Finiteness plays a role for, but is no essential identifier of, main vs. subordinate clauses. In principle, any clause can be finite or nonfinite. As a rule of thumb, unmarked, declarative sentences—what have sometimes been dubbed kernel sentences—are IPs, while more complicated or derived structures (including interrogatives, imperatives, or exclamatives, such as Does John likes Mary?, Play better soccer now!, or What a great game she played!) are usually CPs. Another rule of thumb is that a well-formed clause contains only one subject and one predicate (i.e. one main verb and its potential complements, depending on whether the verb is an intransitive, a transitive, or a ditransitive verb, for example). But virtually every clause can be embedded under or subordinated to a higher clause. This depends on the type of clause and/or on the potential main clause predicate. Addressing the former first, subordinate or embedded clauses come in various types. We commonly

CLAUSE-TYPE INDICATORS distinguish between adverbial, complement, and relative clauses. An adverbial clause (also oblique clause) relates to the main clause through the semantic meanings expressed by adverbs, such as time, manner, place, instrument, circumstance, concession, purpose, result, cause, or condition. In English, adverbial clauses are typically introduced by a subordinating conjunction (or subordinator), an element that links the adverbial clause to the higher main clause. An example is John likes Mary because she plays soccer well. Other subordinators are after, when, whenever, while, as, although, or if. In generative analysis, subordinators invariably head a CP-projection; adverbial clauses are thus CPs. Complement clauses are characterized as clauses that serve as a complement to a lexical item, i.e. without which that lexical item would not be complete. This addresses the role of the potential main clause predicate mentioned above. Two major classifications of complement clauses exist, namely noun-complement and verbcomplement clauses. Just as some physical object must be expressed to form a complete structure of the verb throw (as in The goalkeeper threw the ball), as part of the definition or subcategorization frame of that verb, some clause must be expressed to satisfy the subcategorization frame of a verb like tell. The boldfaced part in Mary told John (that) she played soccer is a complement clause, namely complement to the main verb tell. (Naturally, tell can also take an NP as a complement, as in Mary told John a lie, but we concentrate on clauses here.) In this case, the complementizer that is optional; it can be sued or left out. In this instance, other elements could also be used instead, such as why or how. The important relation is that between the main verb tell and the proposition expressed by the following clause. A similar relationship can be found with certain nouns. A noun like fact takes a complement clause. For example, we say John likes the fact that Mary plays soccer. It is thus part of the subcategorization frame of a noun like fact that it must be followed by a complement clause. Relative clauses, finally, modify, describe, or further specify a noun, but are structurally part of the entire noun phrase, i.e. the head noun, any other modifiers of

it, and the relative clause form one constituent. Moreover, relative clauses are optional, not obligatory. Thus, while in a noun-complement clause, the complement clause must be used, relative clauses need not be. Consider, for example, the sentence John likes the girl who plays soccer . The boldfaced part is a relative clause, adding further information about the NP, the girl. This type is a restrictive relative clause, which is characterized by being essential for identification (otherwise we would not know which girl John likes). A nonrestrictive relative clause is not required for identification of the NP, as in Mary, who plays soccer, is a nice girl. To sum up, the property of human language that sentences can be of infinite length does not mean that longer sentences cannot be analyzed. Rather, a property of language is that sentences can consist of more than one clause, but each clause can be clearly identified. We distinguish between a (unique) main clause and (any number of) subordinate clauses. The latter come in three types: adverbial, complement, and relative clauses. References Haegeman, Liliane. 1991. Introduction to government and binding, 1st edition. Oxford: Blackwell; 2nd edition, 1994. Huddleston, Rodney. 1984. Introduction to the grammar of English. Cambridge: Cambridge University Press. Huddleston, Rodney, and Geoffrey K. Pullum. 2002. The Cambridge grammar of the English language. Cambridge: Cambridge University Press. Keenan, Edward. 1985. Relative clauses Vol. 2: Complex constructions, ed. by Shopen, pp. 141–70. Noonan, Michael. 1985. Complemenatation. Language typology and syntactic description, Vol. 2: Complex constructions, ed. by Timothy Shopen, 42–140. Cambridge: Cambridge University Press. Radford, Andrew. 1988. Transformational grammar. Cambridge: Cambridge University Press. Shopen, Timothy (ed.) 1985. Language typology and syntactic description, Vol.2: Complex constructions. Cambridge: Cambridge University Press. Thompson, Sandra A., and Ronald E. Longacre. 1985. Adverbial clauses. ed. by Shopen, pp. 171–234.

KLEANTHES K. GROHMANN See also Grammar, Traditional

Clause-Type Indicators Communicating by language is a highly structured and goal-directed social activity that serves many purposes: making contact, delivering a speech, telling a joke,

expressing surprise, gathering information, giving orders, and so forth. For verbal interaction to be effective, the speaker has to formulate what he or she wants

211

CLAUSE-TYPE INDICATORS to say in such a way that its communicative intention is recognizable or identifiable to the other participants in the speech situation. Making a statement will naturally take a different form than posing a question or issuing a command. In language theory, the systematic relation between the grammatical form of sentences and their conventionalized conversational use is called clause type. The three major clause types that occur most frequently in the world’s languages are declaratives, interrogatives, and imperatives, which correlate with statements of facts, inquiries, and directives (a cover term for commands, requests, orders, and the like), respectively. Clause types represent a grammatical system in at least two respects. First, one can easily construct triplets of corresponding declaratives, yes–no questions, and imperatives (e.g. John left in a hurry; did John leave in a hurry?; leave us alone!). Second, clause types are mutually exclusive, no sentence belonging simultaneously to two different types. To distinguish one clause type from another on a formal basis, languages resort to a large variety of grammatical devices, which include changes in the shape of verbs, word-order alternations, different intonation patterns, and special particles and affixes. The concern here is with such dedicated affixes and particles, which are put to use as clause-type indicators. Clause-type indicators constitute a subset of function words with a primarily classificatory function that mark the sentence they modify as a declarative or an interrogative construction. Out of the ordinary as sentence-type indicators may look at first sight, they are also attested in English. Consider, for instance, the subordinating conjunctions that and whether, which introduce embedded declarative and interrogative sentences, respectively (e.g. John said that there would be enough food at the party vs. John wondered whether there would be enough food at the party). As many linguists have observed, declarative sentences tend to be the most unmarked of these clause types: they typically occur without any clause type indicator or specific ordering and impose the fewest restrictions on what verbal categories can be selected. There are, however, many languages with marked declaratives. For example, Welsh (a Celtic language of the British Isles) has a clause-initial particle y(r), which is placed in declarative constructions with periphrastic verb forms, which are composed of the present and the imperfect forms of bod ‘to be’ (third-person masc. sing. mae ‘he is’) and a verbal noun (e.g. Y mae Siôn yn gweld draig [PARTICLE is John PROGRESSIVE.PARTICLE see VERBAL-NOUN] ‘John sees (lit. is seeing) a dragon’). The declarative particle y(r) contrasts with other sentence-type indicators like the negative particle ní(d) and the question particle

212

a, neither of which can occur together with y(r). Yet, the particle y(r) is absent in embedded and verb-initial declarative constructions. In Maale (an Omotic language of Southern Ethiopia), all clause types are marked by a special verbal affix. Thus, the main verb of affirmative declarative clauses contains the affix –ne (e.g. ʔ–atsí ziginó mukk–é–ne [man yesterday come-PERFECTIVE-AFFIRMATIVE. DECLARATIVE] ‘The man came yesterday’), while the corresponding interrogatives are characterized by the affix –iya when the main verb is marked for Perfective aspect, which indicates the completion of the event that is described (e.g. ʔ–atsí ziginó mukk–é–iya? [man yesterday come-PERFECTIVE-INTERROGATIVE] ‘Did the man come yesterday?’). A more complex situation is obtained in languages with so-called evidential systems. Evidential particles (or affixes) indicate the source reliability of the speaker’s knowledge and the kind of evidence he or she adduces for what is being communicated. In languages with elaborate evidential marking, such as Tuyuca (a Tucanoan language spoken in Colombia and Brazil), there are no unmarked declaratives, since every sentence must contain an evidential marker qualifying the information on which an assertion is based. A simple declarative like he played soccer corresponds to five different verbal constructions, each associated with a specific evidential marker. Compare: díiga apé–wi [soccer play-VISUAL] ‘he played soccer’ (I say him play), díiga apé–ti [soccer play-NON-VISUAL] ‘he played soccer’ (I heard the game and him, but I didn’t see it or him), díiga apé–yi [soccer play-APPARENT] ‘he played soccer’ (I have evidence that he played soccer, but I didn’t actually see the game), díiga apé–yig! [soccer play-HEARSAY] ‘he played soccer’ (I obtained the information from someone else), díiga apé–hîyi [soccer play-DEDUCTIVE] ‘he played soccer’ (it is reasonable to assume it). Interrogative sentences come in two varieties: yes–no questions, where the truth of the questioned statement is at issue (e.g. Was John invited to the party?) and the constituent question, where the question word or phrase signals the missing piece of information (e.g. Who was invited to the Party?). In Polish, the question marker . czy is used to distinguish main and embedded yes-or-no questions from declarative sentences (e.g. Pan Kowalski by w Austrii ‘Mr. Kowalski has been to Austria’ vs. czy Pan Kowalski by w Austrii ‘Has Mr. Kowalski been to Austria?’ and nie pytaem czy Pan Kowalski by w Austrii ‘I did not ask whether Mr. Kowalski had been to Austria’). In Mandarin Chinese, question words like shei and shemne convey not only a question interpretation ‘who’ and ‘what’ but may also be interpreted as indefinite pronouns ‘someone, anyone’ and ‘something, anything’ (e.g. shei mai–le yi–ben-shu? [who

CLINICAL LINGUISTICS buy-ASPECT one-CLASSIFIER -book] ‘who bought a book?’ vs. ta bu xiang jian shei [S/he not want see who] ‘She (or he) does not want to see anyone’). If, on the other hand, the question particle ne occurs in sentence-final position, only the interrogative and not the indefinite interpretation of the question word is available (e.g. Hufei mai–le shemne ne? [Hufei buy-ASPECT what INTERROGATIVE.PARTICLE] ‘What did Hufei buy?’ [NOT ‘Hufei bought something’]). Languages with dedicated interrogative particles like Chinese leave the question word in exactly the same syntactic position as the corresponding declarative sentence, while languages without such clause-typing devices, like English, place the question word in front of the clause and require the inversion of the auxiliary verb and subject (and insertion of do if there is no auxiliary verb). With evidential particles and affixes being a relatively recent discovery, more research needs to be done to clarify the functional role of such clause-type indicators. References

Barnes, Janet. 1984. Evidentials in the Tuyuca verb. International Journal of American Linguistics 50. 255–71. Bielec, Dana. 1998. Polish—an essential grammar. London and New York: Routledge. Cheng, Lisa Lai-Shen. 1991. On the typology of wh-questions. Ph.D. Dissertation, Massachusetts Institute of Technology. (reprinted in 1997, New York and London: Garland). Cinque, Guglielmo. 1999. Adverbs and functional projections: a cross-linguistic perspective. Oxford and New York: Oxford University Press. Palmer, Frank R. 1986. Mood and modality. Cambridge and New York, Cambridge University Press. Sadock, Jerrold M., and Arnold M. Zwicky. 1985. Speech act distinctions in syntax. Language typology and syntactic description, Vol. I: Clause structure, (ed. by Timothy Shopen,) 155–96. Cambridge and New York: Cambridge University Press. Sproat, Richard. 1985. Welsh syntax and VSO structure. Natural Language and Linguistic Theory 3. 173–216. Willett, Thomas. 1988. A cross-linguistic survey of the grammaticalization of evidentiality. Studies in Language 12. 51–97. Yule, George. 1996. Pragmatics, Oxford introductions to language study. Oxford and New York: Oxford University Press.

Amha, Azeb. 2001. The Maale language. Ph.D. Dissertation. Leiden University, CNWS Publications 99.

See also Clause

CHRIS H. REINTGES

Clinical Linguistics Clinical linguistics, a branch of applied linguistics, is the use of linguistics to describe, analyze, and treat language disability. The study of linguistic aspects of communication disorders is of relevance to a broader understanding of language and linguistic theory. Practitioners of clinical linguistics range from academic linguists with research and teaching interests in language disability to practicing professionals such as speech and language pathologists/therapists, educational and clinical psychologists, and neurologists. Research in the field tends to be multifaceted, drawing on a wide range of disciplines in addition to linguistics, such as psycholinguistics, cognitive science, neuroscience, and biomedical science. Although systematic phonetic descriptions of speech disorders have been routine since the 1950s, clinical linguistics did not emerge as a coherent discipline inclusive of phonology, grammar, semantics, and pragmatics until the late 1970s. This was in part due to the pioneering work of David Crystal whose book Clinical linguistics (1981) has been particularly influential in defining

the area. Other milestones were the establishment of the journal Clinical linguistics and phonetics in 1987 and the founding of the International Clinical Phonetics and Linguistics Association (ICPLA) in 1991. Rather than diagnosing using medical criteria, clinical linguistics focuses on the linguistic manifestations of the disorder, aiming to provide a comprehensive typology of language disorders based on their linguistic characteristics. This approach aids in diagnosing and treating impairments of unknown underlying causes. For example, there is as yet no agreed medical explanation for Specific Language Impairment (SLI), a condition found in children who have problems with spoken language but no other obvious cognitive or neurological deficit. In spite of this, it is still possible to describe the linguistic characteristics of SLI precisely enough for research purposes and for devising remedial programs. There appears to be no level of language organization that is immune to impairment. Language disorders are equally likely to be found in adults and in children who are still acquiring language, and in both

213

CLINICAL LINGUISTICS the production and comprehension of spoken, written, and signed language. Anatomical deficits such as cleft palate may affect a speaker’s phonetics and phonology, resulting in an inability to differentiate pairs of words such as ‘bat’ and ‘mat’. Grammar, semantics, and pragmatics may be impaired as a result of developmental disorders in children or a stroke in adults. A child with SLI might repeat ‘Has the mouse been chased by the cat?’ as ‘A mouse chasing cat’. A stroke patient might see nothing wrong with adding a past tense suffix to nouns, resulting in words like ‘towned’ and ‘faithed’, due to problems with word formation. People with autism find it difficult to make use of context to infer nonliteral meaning. Any linguistic theory that purports to throw light on the nature of the human mind should describe both the normal and the pathological. Some have argued that certain syntactic anomalies found in the language of people with Broca’s aphasia, a condition resulting from damage to Broca’s area of the brain, can be explained using Chomsky’s ‘Principles and Parameters’ theory of syntactic structure. These researchers believe that such anomalies provide evidence that syntactic abilities are separate from other cognitive skills. In addition, data from children with SLI and Williams Syndrome, a genetic condition in which linguistic proficiency develops despite poor cognitive abilities, have been use d to support a similar view of syntax as an autonomous mental ‘module’. Others, however, cite evidence from a range of language disorders in support of a nonmodularist, functionalist account of language suggesting a considerable degree of codependency between linguistic and cognitive processes. Clinical linguistic research will continue to play a key role in debates of this kind. The range of analytical methods used in clinical linguistics is comparable to that found in other areas of linguistics. Both theory-based and data-driven approaches are used, with a wide selection of materials produced for purposes of diagnosis, assessment, and

intervention. Nevertheless, clinical linguistics is still a new field and a great deal of exploratory work remains to be done, particularly in the form of case studies and in areas such as semantics. Often, it is only when a complex system goes wrong that we become aware of the contribution—and even the existence—of its component subsystems. One potential growth area for clinical linguistics, therefore, is its role in informing and evaluating linguistic theory and in illuminating our understanding of normal language structure and use. References Ball, Martin J. 1989. Phonetics for speech pathology. London: Taylor & Francis, 2nd edition, London: Whurr, 1993. Ball, Martin J., and Raymond D. Kent (eds.) 1997. The new phonologies: developments in clinical linguistics. San Diego: Singular. Chiat, Shula. 2000. Understanding children with language problems. Cambridge and New York: Cambridge University Press. Crystal, David. 1980. Introduction to language pathology. London: Edward Arnold, 4th edition (co-author Rosemary Varley), London: Whurr, and San Diego: Singular, 1998. Crystal, David. 1981. Clinical linguistics. Vienna: SpringerVerlag, London: Whurr, 1989. Grodzinsky, Yosef. 1990. Theoretical perspectives on language deficits. Cambridge, MA: MIT Press. Grundy, Kim (ed.) 1989. Linguistics in clinical practice. London: Taylor & Francis; 2nd edition, London: Whurr and San Diego: Singular, 1995. Hewlett, Nigel, Louise Kelly, and Fay Windsor (eds.) 2001. Themes in clinical linguistics and phonetics. New York: Lawrence Erlbaum. Howard, Sara J., and Barry C. Heselwood. 2005. Clinical phonetics and phonology: analysing speech in a clinical context. New York: Academic Press. Perkins, Michael R. 2005. Pragmatics and communication impairment. Cambridge and New York: Cambridge University Press. Perkins, Michael R., and Sara J. Howard (eds.) 1995. Case studies in clinical linguistics. London: Whurr and San Diego: Singular.

MICHAEL PERKINS See also Aphasia; Applied Linguistics: Overview; Psycholinguistics

Code-Switching Speakers of more than one language (e.g. bilinguals) are known for their ability to code-switch or mix their languages during communication. This phenomenon occurs when bilinguals substitute a word or phrase from one lan-

214

guage with a phrase or word from another language. While some linguists suggest that people code-switch as a strategy in order to be better understood and to enhance the listeners’ comprehension, code-switching among

CODE-SWITCHING bilinguals has traditionally been viewed as a strategy to compensate for diminished language proficiency. The premise behind this theory is that bilinguals code-switch because they do not know either language completely. This argument is also known as semilingualism, meaning bilinguals ‘almost’ speak both languages correctly. However, language proficiency is not clearly defined, as it is unclear whether reading and writing language skills should take precedence over spoken skills in determining language proficiency. This reliance on reading and writing is problematic because most bilinguals receive their formal education in one language, while a majority of their social interactions take place in the other language. Thus, when their reading and writing abilities are tested in both languages, the language in which bilinguals received more formal education will usually fare better. Recent psycholinguistic research has focused on how code-switching is a natural product of the interaction of the bilingual’s two languages. Early researchers viewed code-switching as evidence that the bilinguals’ two languages were organized in separate and distinct mental dictionaries. For example, a general finding throughout the literature is that bilinguals take longer to read and comprehend sentences containing code-switched words as compared to monolingual sentences. Apparently, this time-consuming process is due to a ‘mental switch mechanism’ that determines which of the bilingual’s two mental dictionaries are ‘on’ or ‘off’ during the course of language comprehension. This mental switch is responsible for selecting the appropriate mental dictionary to be used during the comprehension of a sentence. Other research shows that bilinguals comprehend codeswitched words faster when there is an overlap between the two languages’ sound systems. For example, Chinese–English bilinguals, where Chinese is the native language, take longer to recognize English code-switched words in Chinese sentences, but only if the English words begin with consonant–consonant clusters (e.g. block), as opposed to consonant–vowel clusters (e.g. big), because the Chinese language words do not begin with consonant–consonant clusters. Other important factors reported to influence the recognition of code-switched words include context, phonetics, homophonic (e.g. words pronounced the same) and homographic (e.g. words spelled the same) overlap between the two languages. Another current view suggests that language dominance, which language is used more frequently, plays an important role in code-switching. For example, Spanish–English bilinguals report more linguistic interference (code-switching) when they communicate in Spanish, their first language, and little or no codeswitching when they communicate in English, their

second language. In other words, these bilinguals code-switch more when they communicate in Spanish than when they use English. Psycholinguistic evidence also suggests that bilinguals retrieve English codeswitched words faster when they listen to Spanish sentences, whereas they are slower to retrieve Spanish code-switched words as they listen to English sentences. This evidence suggests that the bilingual relies on the second language as opposed to the first. How are these findings explained? The general premise behind this view is that after a certain level of fluency and frequent use of the second language, a language shift occurs. That is, during early stages of bilingualism, Spanish–English bilinguals rely on their first language when they communicate in their second language. As a result, bilinguals are more likely to code-switch to Spanish, when they communicate in English. However, as the second language becomes the dominant language, bilinguals rely on the second language when they communicate in the first language. In this case, bilinguals code-switch to English when they communicate in Spanish. The second language becomes more readily accessible and bilinguals come to rely on it more. Regardless of which language the bilingual learned first, the more active (dominant) language determines which mental dictionary is going to be accessed faster. This argument is reasonable since most bilinguals in the United States, whose first language is Spanish, obtain their formal education in English. Likewise, many of their everyday interactions involve the second language. As a result, words and concepts in English, the second language, become more accessible than words in Spanish, the first language. Thus, code-switching is not the same for both languages. Rather, it depends on language dominance. In short, code-switching among bilinguals may be indicative of difficulties in retrieval (access) affected by a combination of closely related factors such as language use (i.e. how often the first language is used) and word frequency (i.e. how much a particular word is used in the language). Examination of codeswitching behavior can contribute to a further understanding of second-language acquisition, as well as language acquisition and development, and general linguistic theory. References Lederberg, Amy R., and Cesareo Morales. 1985. Code switching by bilinguals: evidence against a third grammar. Journal of Psycholinguistic Research 14. 113–36. Li, Ping. 1996. Spoken word recognition of code-switched words by Chinese–English Bilinguals. Journal of Memory and Language 35. 757–74.

215

CODE-SWITCHING Macnamara, John, and Seymour Kushnir. 1971. Linguistic independence of bilinguals: the input switch. Journal of Verbal Learning and Verbal Behavior 10. 480–87. Heredia, Roberto R. 1997. Bilingual memory and hierarchical models: a case for language dominance. Current Directions in Psychological Science 6. 34–9. Kolers, Paul. 1966. Reading and talking bilingually. American Journal of Psychology 3. 357–76.

Myers-Scotton, Carol. 1993. Duelling languages: grammatical structure in code-switching. New York: Oxford University Press.

ROBERTO R. HEREDIA AND JEFFREY M. BROWN See also Acquisition; Bilingual Acquisition; Bilingualism

Coherence in Discourse Discourse is a communicative event in which language plays a prominent role. It minimally requires a sender (writer, speaker), a receiver (reader, listener), and a message that is being communicated. This message is not just a concatenation of clauses; it forms a unified, coherent whole. Both the sender and receiver normally have the implicit agreement that the message being communicated is coherent. Coherence in discourse has been studied in a range of disciplines, including linguistics, philology, sociology, philosophy, psychology, and computer science. Linguists identify and analyze inventories of the linguistic markers of coherence that are available in a language. Sociologists explore the production and comprehension of coherent discourse in naturalistic conversations that involve different groups and cultures. Psychologists collect data in experiments that test hypotheses about the effect of coherence on cognitive processing and representations. Computer scientists design and test computer models that attempt to produce and test coherent text. The term coherence has been defined in various ways. Some researchers apply the term cohesion to the surface structure of the text and the term coherence to the concepts and relations underlying its meaning. Cohesion has sometimes been applied to smaller units of language in the text, and coherence, to some general overall interrelatedness in the text. Other researchers have defined cohesion as continuity in word and sentence structure, and coherence as continuity in meaning and context. As in the case of coherence, discourse has been defined in different ways. Several years ago, the term discourse was reserved for dialogue, and text was reserved for monologue. In contemporary research, discourse covers both monologic and dialogic spoken and written language. Somewhat more subtle distinctions are sometimes made. One can distinguish between discourse-as-prod-

216

uct (the linguistic construct) and discourse-as-process (the communicative event). Coherence can be reserved for the conceptual relationships that comprehenders use to construct a coherent mental representation accommodated by what is said in the discourse. Cohesion is limited to the linguistic markers that cue the comprehender on how to build such coherent representations. Cohesion emphasizes discourse-as-product, and coherence emphasizes discourse-as-process. Cohesion alone is not sufficient for the interpretation of the discourse. Comprehenders generate inferences on the basis of background knowledge and discourse constraints. Much of the background knowledge is experiential; hence, it involves common procedures and activities (called scripts), social interactions, and spatial settings. For instance, a narrative usually describes a setting, an action sequence with a conflict and plot, and an outcome. A script for eating in a restaurant would furnish inferences and help coherently tie together the explicit content of a narrative about a bad restaurant experience. Although cohesion alone cannot fully account for coherence in discourse, the psycholinguistic literature has shown that cohesion facilitates coherence. Cohesion and coherence can be divided into local (microstructure) and global (macrostructure). Local cohesion and coherence are related to the interrelatedness between adjacent discourse segments. Global cohesion and coherence are related to the interrelatedness of larger spans of discourse. For instance, scripted action sequences are globally coherent. Also, there are the rhetorical structures of narrative (such as setting + conflict + plot + resolution), expository (such as claim + evidence, problem + solution), and other discourse genres. Cohesion and coherence can be grammar driven and vocabulary driven. Grammar-driven cohesion refers to sentence structure, word structure, and the intonation of the discourse segments. Vocabulary-driven cohesion

COHERENCE IN DISCOURSE refers to the lexical vocabulary of the discourse segment. These cohesion cues activate vocabulary-driven (pregrammatical, knowledge-based) and grammar-driven (syntax-based) coherence. Vocabulary-driven and grammar-driven coherence are not necessarily mutually exclusive but often support each other, as illustrated below. Consider the sentence The dean (i) read the New York Times (ii) in his office (iii). A paraphrase with grammar-driven cohesion would reduce the discourse elements to the grammatical necessities: He (i) always reads it (ii) there (iii). A vocabulary-driven paraphrase, on the other hand, would find meaningful lexical alternatives, as in The man (i) always reads the newspaper (ii) behind his desk (iii). In addition to the distinctions between local and global and between grammar- and vocabulary-driven cohesion, the types of cohesion discussed below have often been recognized. Conjunctions relate adjacent discourse segments. There have been several classifications of these conjunctions in virtually every field. Most of these classifications include additive (and, but), temporal (before, until), and causal (because, although) conjunctions that are either extensive (and, before, because) or adversative (but, until, although). Coreference specifies that two expressions refer to the same entity. Often, the coreference is grammar driven by the use of pronouns, both pronominal (he, she) and reflexive (himself, herself). The interpretation of these pronouns is determined by their antecedents, i.e. previously mentioned words referring to the same person or object. The coreference can be both forward and backward. Anaphoric reference is a backward reference to an antecedent noun phrase or clause that was introduced earlier in the discourse (John kissed Mary because he loved her). Cataphoric reference is a forward reference to a noun phrase or clause that will be mentioned later in the text (Because he loved her, John kissed Mary). With substitution, repeated forms, and ellipsis, a constituent of one expression is replaced by a constituent of another (substitution), is repeated (repeated forms), or is omitted (ellipsis). The intended meanings can be reconstructed from the preceding discourse and from world knowledge. Will we make it on time? (1) I think so (substitution of we will make it in time by so). (2) Yes, we will make it on time (repeated forms). (3) If we hurry (ellipsis: we will make it on time is omitted). With lexical relationships, the type of cohesion is vocabulary driven. Two lexical items are related to each

other insofar as they mean the same thing (synonyms) or the opposite thing (antonyms), stand in a superset/subset relationship (hypernym vs. hyponym, respectively), or have some other conceptual relationship. (1) (2) (3) (4)

The tax collector sent another letter. I don’t like this guy. That monster never leaves us alone. The sweetheart keeps asking for more each year

With comparison, a constituent in an expression is compared with a constituent in another expression (I like the oak cabinet. The pine desk is much nicer). Discourse psychologists have extensively investigated five cohesion and coherence relations that are related to the previously mentioned seven: referential, spatial, causal, temporal, and additive relationships. They answer the questions of the who, where, why, when, and what of the events described by the discourse. Explicit markers facilitate the comprehension process. Several classifications of relations have been proposed. Some focus only on the closed set of grammardriven cohesion, whereas others include vocabularydriven relations and relations that are reconstructed from world knowledge and the unique situation conveyed in the text. Those classifications that go beyond grammar consider the intentions of the producer of the communicative event. In written monologic discourse, comprehenders can rely on linguistic cues to a great extent (although not completely). However, in oral dialogic discourse, there are conversational cues that go well beyond print, such as intonation, gestures, and the physical environment. A complete theory of discourse coherence requires a harmonious layering of several levels, including vocabulary, sentence structure, meaning, discourse context, style, and world knowledge. When these levels lack coordination, the coherence is more difficult. To get the message across, the sender will try to coordinate the levels. The receiver assumes that the sender’s message is intended to be well formed and will make every attempt to construct a coherent interpretation. References Beaugrande, Robert de, and Wolfgang U. Dressler. 1981. Introduction to text linguistics. Harlow: Longman. Brown, Gillian, and George Yule. 1983. Discourse analysis. Cambridge: Cambridge University Press. Clark, Herbert H. 1996. Using language. Cambridge: Cambridge University Press. Costermans, Jean, and Michel Fayol (eds.) 1997. Processing interclausal relationships. Studies in the production and comprehension of text. Mahwah: Lawrence Erlbaum. Gernsbacher, Morton Ann. 1994. Handbook of psycholinguistics. New York: Academic Press.

217

COHERENCE IN DISCOURSE Givón, Talmy. 1995. Coherence in the text and coherence in the mind. Coherence in spontaneous text, ed. by Morton Ann Gernsbacher and Talmy Givón. Amsterdam: John Benjamins. Graesser, Arthur C., Gernsbacher, Morton Ann, and Goldman, Susan (eds.) 2003. Handbook of discourse processes. Mahwah, NJ: Erlbaum. Graesser, Arthur C., and Gordon H. Bower (eds.) 1990. Inferences and text comprehension. San Diego: Academic Press. Halliday, Michael A. K., and Ruqaiya Hasan. 1976. Cohesion in English. London: Longman.

Knott, Alistair, and Chris Mellish. 1996. A feature-based account of the relations signalled by sentence and clause connectives. Journal of Language and Speech 39. 143–83. Mann, William C., and Sandra A. Thompson. 1987. Rhetorical structure theory: toward a functional theory of text organization. Text 8. 243–281. Van Dijk, Teun A., and Walter Kintsch. 1983. Strategies of discourse comprehension. New York: Academic Press.

MAX M. LOUWERSE AND ARTHUR C. GRAESSER

Color Terms Color terms have been used in anthropological and cognitive linguistics to investigate whether certain aspects of meaning are universally found in all languages. The classic study of color terminology, Berlin and Kay 1969, is often cited as proof against the Sapir–Whorf hypothesis of ‘linguistic relativity’ or ‘linguistic determinism.’ This hypothesis argues that language affects or even determines cognition—that is, that the categories present in the language a person speaks will influence the way that person thinks, the way he or she perceives the world. Languages vary as to which ‘basic color terms’ occur in the vocabulary. Berlin and Kay set out to test whether the presence or absence of a color category in the language affects the way in which speakers see colors. In the course of their research, they found evidence favoring a ‘prototype’ theory of meaning and evidence for universal tendencies in the structuring of color vocabulary in the world’s languages. Berlin and Kay gave color charts to speakers of 98 different languages and asked them to sort the 329 colored squares into categories that corresponded to the black and > white

red

>

“grue” or > yellow

yellow or > “grue”

basic color terms in their language. They put three limits on what could count as a basic color term: (1) It had to be a simple word; e.g. reddish-brown is not a basic color term. (2) It had to apply generally, not in a specific domain (like blonde or sienna). (3) It could not count as a ‘kind of ___’, in the way that emerald and olive are kinds of green. They found that the most basic color terms any language has is 11, as in English (also Tagalog, Zuni, and 218

others): white, black, red, yellow, green, blue, brown, purple, pink, orange, and gray. Interestingly, these terms occurred in predictable patterns, and Berlin and Kay proposed a scheme of ‘implicational universals’ to explain the evolution of color terminology. Some languages have only two basic color terms, and these are always black and white. If a third is added, as in the Arawak language of the Caribbean or in Swahili of East Africa, it is always red. After red, languages with four color categories contain either yellow or a term that encompasses both blue and green (Kay devised the term ‘grue’); languages with five colors include both yellow and grue, with differentiation between green and blue occurring in the six-color system. Brown is added as the seventh, then purple, pink, orange, and gray, which do not follow a specific ordering. An example of an implicational universal that follows from this discovery would be that languages with a term for blue will always have words for black, white, red, yellow, and green. The sequence can be depicted as shown below: blue and > green

brown

>

purple pink orange or gray

The orderly sequence for the development of color terms leads us to the idea that they might reflect cognitive universals, and this is indeed what Berlin and Kay found in their research on the focal meanings of the color terms. When they asked people to choose the best example of each of the basic color terms from the chart, choices were remarkably similar around the world. For example, people whose language contained no term for orange chose the same focus for yellow as those whose language had terms for both yellow and orange; people whose language contained only three

COMMUNICATION THEORY color terms had the same focus for red as those whose language had all 11 terms. This is the finding that refutes the so-called strong version of the Sapir–Whorf hypothesis (a version espoused by neither Edward Sapir nor Benjamin Lee Whorf), that our perception will be determined by the categories of our language. For color, at least, perception seems to follow universal human neural response patterns in a way that is unaffected by our language habits, although there have been dissenting claims suggesting that Berlin and Kay’s experimental methods may not accurately reflect the way speakers apply color terms to objects in the real world. The existence of universals for the core meanings of color terms does not contradict the fact that culture does play a role both in determining the wider meaning of color terms for speakers and in the inventory of non-basic terms for colors in one’s speech repertoire. Color terms may be closely associated metaphorically with objects in our experience, for example, the sun with yellow, or fire with red. Color vocabulary may be highly elaborated in some cultures and for some speakers, according to the importance of color in various arenas of daily life. A painter or an interior decorator, for example, will have an extensive lexicon of color terms. In American society, women have larger color vocabularies than men, due to the sexual division of labor in purchasing clothing and household furnishings and the proliferation of color terms as a form of marketing in American consumer culture. In applying color terms to objects and to parts of the color chart that are farther away from the focal colors, speakers seem to compare the color in question with the focal color and to judge both similarities and differences. The task of drawing boundaries on the color chart around the limits of each color produced more varied answers than determining the best example of a color did. The focal color seems to act as a prototype for that color. Prototype theory views word meaning as

the set of referents that can be designated by that word, which is determined by how those referents compare to a prototypical exemplar of the category. This produces ‘fuzzy sets’, with some uncertainty about whether referents that are marginal really belong to the category or not. Is turquoise a kind of blue or a kind of green? Is burgundy a kind of red or a kind of purple? Different people will give different answers. Researchers continue to debate the ways in which meanings of color terms are encoded in our minds. There are still questions about how language and culture relate to cognition, including whether cognitive images or concepts of color are based solely on neurologically determined perceptions that are universal for humans or whether they are molded in part by linguistic and cultural experience that includes using color words in ways quite different from an experimental setting. References Kay, Paul, and Chad K. McDaniel. 1978. The linguistic significance of the meanings of basic color terms. Language 54. 610–46. Kay, Paul, Brent Berlin, and William Merrifield. 1991. Biocultural implications of systems of color naming. Journal of Linguistic Anthropology 1. 12–25. Berlin, Brent, and Paul Kay. 1969. Basic color terms: their universality and evolution. Berkeley: University of California Press. Lucy, John A. 1992. Language diversity and thought: a reformulation of the linguistic relativity hypothesis. Cambridge: Cambridge University Press. MacLaury, Robert E. 1991. Social and cognitive motivations of change: measuring variability in color semantics. Language 67. 34–62. Whorf, Benjamin. 1956. Language, thought, and reality: selected writings of Benjamin Lee Whorf, ed. by John B. Carroll. Cambridge, MA: MIT Press. Wierzbicka, Anna. 1990. The meaning of color terms: semantics, culture, and cognition. Cognitive Linguistics 1. 99–150.

ELLEN JOHNSON

Communication Theory The contents of the terms ‘communication’ and ‘communication theory’ are notoriously vague. A straightforward psychologist’s definition is the following: ‘communication is the discriminatory response of an organism to a stimulus’. Thus, Pavlovian conditioning would also be an instance of communication. This is strange to the linguist, because

it is generally accepted that not even the learning of languages is a simple conditioning process. How could it be possible that after such a complicated process like language acquisition human beings use their languages for communicative purposes that are as simple as Pavlovian conditioning? The stimulus–response theory does not seem to be on the right track. 219

COMMUNICATION THEORY Another way to define communication is by information. Often, the term ‘communication theory’ is used in the strict sense of information theory. Information theory originated in problems around telecommunication techniques and was developed by Claude Shannon and Warren Weaver in 1949. This theory of communication suggests the following schema for communicative acts: Source ——> Transmitter ——> Channel ——> Receiver ——> Destination In effect, the source sends some message to the destination. But let us look at the details between source and destination. First, the message is encoded into a signal; this is done by the transmitter. This point is important in telecommunication; however, in ordinary face-to-face talk, source and transmitter coincide. Leaving the receiver, the signal is sent through the channel. In ordinary face-to-face talk, the channel would be the air channel. But in telecommunication, it would be the electric channel. In the channel, signals are usually changed, or, rather, distorted by the noise. In ordinary face-to-face talk, the surrounding sounds that are not part of the conversation would be the noise of the channel. After the channel, the signal reaches the receiver. There, the decoding of the signal takes place, and we end up with a message again. This message is finally stored in the destination. In ordinary face-to-face talk, receiver and destination coincide. As precise as this schema is, and as useful as it is for telecommunication, it neglects several factors of natural communication. First, the message lacks interpretation. It seems odd to say that the message is finally just ‘stored’ in the destination. This is not even true for the simplest acts of communication. If my neighbor at the dinner table utters ‘could you please pass me the salt’, I would not just store this message in my brain. Instead, I would interpret the message as nonliteral, and furthermore as an instruction to pass the salt to my neighbor. This simple example demonstrates that, often, people communicate because they want others to do certain things. This leads to the next factor of communication that is neglected in the Shannon–Weaver model: the dialogic nature of communication. Natural communication is not just one way as in the Shannon/Weaver schema; normally, something goes back to the sender. Considerable research has been done to capture the necessary dialogism of communication. The most interesting schools are the Bakhtin Circle, the Prague school, and some versions of functionalism. Let us focus on the Bakhtin Circle. In the pre-Stalin era, a group of three scholars emerged with some common views on communication and cognition: the literary critic Mikhail Bakhtin, the psychologist Lev

220

Vygotsky, and the philosopher Valentin Voloshinov. Due to political reasons, it took some time before their work reached western research. But since the 1960s and 1970s, their ideas have influenced western psychology and text linguistics. The main idea of these scholars is that the language sign is inherently dialogic, and that all manifestations of the signs are inherently dialogic. Language is formed in the process of social interaction. While the ruling stratum tries to posit a single discourse as exemplary, the subaltern classes are inclined to subvert this monologic closure. In the sphere of literature, poetry and the epic represent the centripetal forces within the cultural arena while the novel is the structurally elaborated expression of popular ‘ideologiekritik’ (ideology criticism). In Bakhtin’s now famous study Problemy tvorchestva Dostoevskogo (‘Problems of Dostoevskii’s work’), it is argued that Dostoevskii’s work is imbued with a profoundly democratic spirit. Bakhtin argues that Dostoevskii’s creative method is not Hegelian. In a Hegelian scheme, two positions struggle for ascendancy but are transformed into a synthesis at the end. According to Bakhtin, Dostoevskii does not merge the voices into one final, authoritative voice as in the Hegelian absolute. Instead, Dostoevskii presents an unmerged dialogue of voices, each given equal rights. Bakhtin calls this type of dialogue ‘polyphonic’. The voice of the narrator resides beside the voices of the other characters. Voices intersect and interact, mutually illuminating their ideological structures, potentialities, biases, and limitations. Polyphony is also observed across different texts: for instance, literary texts show ‘intertextuality’, which means that they refer to other texts, thus yielding a huge larger text together. ‘Voicing’ is not only a characteristic of literary texts. In ordinary speech, several factors also lead to a multivoiced message: the cultural context, the social relations between the persons involved in the communication, the context of the situation — all these factors have an effect on the message. Linguistics, however, usually concentrates on only some aspects of communication, leaving cultural and social factors aside. So far, we have concentrated on the two forms of verbal communication: written and oral communication. But human communication may also be nonverbal. There are many different channels of nonverbal communication: facial expressions, hand gestures, body movements (‘kinesics’), and touch (‘haptics’). All these modes of nonverbal communication make use of signs, movement, or touch. In animal communication, there are — in addition to nonverbal and (partly) verbal communication — other ways of communication, namely by smell and taste. Human beings do not make much use of smell and taste in the communicative

COMPARATIVE METHOD domain. Humans receive information about the world through the channels of smell and taste, for instance, while eating. But they do not use smell and taste for communication. Body smells seem to have a communicative function only in sexuality. We have seen that communication has many aspects, and as many as communication theory has. If the focus of attention is more on social factors, the dialogism of communication will be important; if information structure plays a greater role, the theory will be more technical. Little attention has been paid to integrate the findings on nonverbal communication into the general frameworks existing for verbal communication. This might be an interesting field of study for the future.

References Infante, Dominic A., et al. 1997. Building communication theory. Prospects Heights, IL: Waveland. Miller, Katherine. 2001. Communication theories: perspectives, processes, and contexts. Mountain View, CA: Mayfield Publishing. Neuliep, James W. 1996. Human communication theory: applications & case studies. Boston: Allyn & Bacon. Phillipsen, Gerry. 1992. Speaking culturally: explorations in social communication. Albany, NY: State University of New York Press. West, Richard, and Lynn H. Turner. 2000. Introducing communication theory: analysis and application. Mountain View, CA: Mayfield Publishing.

MONIKA RATHERT See also Text Linguistics

Comparative Method The comparative method is the most important of the various methods and techniques used to recover linguistic history. It is also of interest in language classification, in linguistic prehistory, and in research on distant genetic relationships, determining whether seemingly interrelated languages descend from a common ancestor after all. It is generally recognized that 6,000 languages are currently spoken all over the world (one quarter with fewer than 1,000 speakers), and they are grouped into more than 30 widely recognized families. It is estimated that half of those languages will die in the next century. The comparative method is therefore of much help in clarifying their status and evolution through time. The comparative method is a set of principles and a methodology for comparing forms from related languages to reconstruct an earlier (proto-) language or an earlier state of a language. The aim of reconstruction is to recover as much as possible of the protolanguage from a comparison of the descendant languages and to determine what changes have taken place in the various languages that developed from the proto-language. The comparative method was developed in nineteenth-century Germany for the reconstruction of Proto-Indo-European and was afterward applied to the study of other language families. It is considered the most outstanding achievement of linguistic scholarship in the nineteenth century and was originally stimulated by the discovery by an English orientalist, Sir

William Jones, in 1786, that Sanskrit was related to Latin, Greek, and German. His famous speech to the Asiatic Society of Bengal marks the beginning of comparative philology. Furthermore, the year 1816, when the German scholar Franz Bopp presented the linguistic public with language material from Sanskrit compared with some other Indo-European languages, remains a historical date in linguistics, because Bopp was the first to realize that the question of the mutual relations of Indo-European languages was worth becoming the subject of specialized inquiry. Indo-European was posited as a hypothetical ancestor language, and comparative studies were almost exclusively focused on the Indo-European language family. Six years later, another German scholar, Jacob Grimm, included a systematic survey of the relationships between Germanic consonants and their correspondents in Greek, Latin, and Sanskrit in the second edition of his comparative grammar of Germanic languages. He established the existence of a circular ‘soundshift’ (Lautverschiebung) governing these relationships in the prehistory of Germanic. Grimm observed that when other Indo-European languages such as Latin and Greek have a voiced unaspirated stop (b, d), Gothic has the corresponding voiceless unaspirated stop (p, t), and that, when other IndoEuropean languages have a voiceless unaspirated stop, Gothic has a voiceless fricative (f,θ). In other words, the voiced stops inherited from Proto-Indo-European

221

COMPARATIVE METHOD became voiceless, and the voiceless stops became fricatives. His discovery is known in linguistics under the name of Grimm’s Law (or the Germanic consonant shift), although Grimm himself did not use the term law. Grimm’s Law was probably first worked out in detail by the Danish linguist Rasmus Rask (whose studies, being written in Danish, were less accessible to most European scholars). However, linguists found at least two classes of exceptions to Grimm’s Law with respect to IndoEuropean voiceless stops: there are cases in which Proto-Indo-European voiceless stops remain voiceless stops in Germanic and cases in which Proto-IndoEuropean voiceless stops become voiced stops in Germanic. Grimm’s Law was thus corrected in 1875 by the Danish linguist Karl Adolf Verner, who examined words in Vedic Sanskrit and Attic Greek that are related to Germanic words that reflect Indo-European voiceless stops as voiced stops. He discovered that in all such cases, the Proto-Indo-European word had intonational stress after the voiceless stop. Verner understood that what seems to be an exception to sound laws may be governed by some other regularity and that sound changes may be unconditioned or may depend on the environment in which the sound occurs. This was a very important step in linguistics, and ancestral common forms could be reconstructed by using the principle of regular sound change and postulating a number of different sound laws that operated independently in the different branches of the IndoEuropean family. The comparative method was often criticized because it is based on a misleading genealogical metaphor. In the 1850s, the German linguist August Schleicher transferred the Darwinian view of evolution to language and introduced the model of the evolutionary ‘family tree’ to comparative linguistics. As at that time it was customary to concentrate on the reconstruction of what was supposed to be the source of all Indo-European languages, Schleicher ventured an actual reconstruction of a short text, the ‘fable of the sheep and the horses’. During the second half of the nineteenth century, a group of four young scholars known collectively as the Junggrammatiker (‘young grammarians’ or ‘neogrammarians’) proclaimed the ‘absolute exceptionlessness of sound laws’, i.e. all changes in the sound system of a language as it developed through time were subject to the operation of regular sound laws. The assumption that sound laws were in principle absolutely regular in their operation was quite generally accepted and had become the landmark of the comparative method. Realizing that sound changes occur in defined geographical regions that do not coincide with the

222

spread of an official standard language, the German scholar Johannes Schmidt criticized the family-tree theory as unrealistic and misleading. According to his theory, known in linguistics as ‘the wave theory’ (1872), each sound law has its own territory, and different linguistic changes will spread, like waves, from a politically, commercially, or culturally important center along the main lines of communication. These innovations will not necessarily cover exactly the same area. A real achievement of the method of reconstruction was the Laryngeal Hypothesis presented by the famous Swiss linguist Ferdinand de Saussure in his 300-page Mémoire sur le Système Primitif des Voyelles dans les Langues Indo-Européennes (1879; Report on the Primitive System of Vowels in Indo-European Languages): there had to be two hypothetical sounds for Proto-Indo-European, not attested in any of the Indo-European languages known at that time, but sufficient to solve several puzzling irregularities in the development of certain vowels in a number of languages. He was fairly conservative about claiming what they must have been, but he called them laryngeals, indicating that these sounds must have been articulated with the larynx (voice box) as the main articulator, and pointed out the precise locations in words where they must have occurred. Fifteen years after Saussure’s death, when the ancient Anatolian language Hittite, the oldest attested Indo-European language, was deciphered, linguists found with amazement that Saussure was right: Hittite had the laryngeal sounds, and they were placed exactly where Saussure had predicted just on the basis of careful reconstruction. The history of comparative philology (this was the name for linguistics at that time) shows that the comparative method addresses directly only material in the related languages that is inherited from the proto-language and has no means of its own for dealing with borrowings. All things considered, there are four basic assumptions on which the comparative method is based: (1) the proto-language was uniform, with no dialect variation; (2) language splits are sudden; (3) after the split up of the proto-language, there is no subsequent contact among the related languages; and (4) sound change is regular. Nowadays, the work of reconstruction usually begins with phonology, i.e. with an attempt to reconstruct the sound system; this in turn leads to reconstruction of the vocabulary and grammar of the proto-language. In the nineteenth century, comparative studies were entirely word based: it was words or lexical word stems that were reconstructed and followed through history. Only after the knowledge of phonetics

COMPARATIVE METHOD had begun to increase did linguists realize that they should start reconstruction with sounds, not with lexical items. Consequently, historical linguistics has often enjoyed great success in reconstructing ancestral phonological systems and vocabularies. But for grammatical systems, the situation is slightly more complicated: grammatical patterns are often seriously perturbed by independent and cointeracting processes of change, and grammar usually undergoes more dramatic changes than the sound system and even the vocabulary. Generally, studies in the comparative method start with the assumption that certain languages are related and hence descended from a common ancestor. The first step in the work of reconstruction is to find sets of putatively related words in the languages or dialects being compared, i.e. placing side by side a number of words with similar meanings from the languages under comparison, for example, Sanskrit pita, Latin pater, Greek pater, and Gothic fadar, all of them meaning ‘father’. Such sets of ‘cognate’ forms should be examined for what seem to be systematic correspondences in their sound structure. Each correspondence must yield a plausible-looking sound in the ancestral language, one that could reasonably have developed into the sounds that are found in the several daughter languages; phonological change should be taken into account, because any difference in the form of cognates is a consequence of sound change in one or more of the languages. Establishing regular correspondences between languages permits the assumption that the languages in question are genetically related; these correspondences can also be used to formulate likely hypotheses about what changes gave rise to those correspondences, and to reconstruct antecedent unattested forms in the proto-language from those changes. A set of ‘reconstructed’ sounds can then be postulated (marked with an asterisk by the standard convention) to which the sounds in the attested languages can be systematically related by means of sound laws. With these results, each word surviving in the various daughters may be put face to face with its form in the ancestral language. The reconstructed Proto-IndoEuropean word for ‘father’ is *pəter. If neither chance nor borrowing can account for observed similarities, shared inheritance (i.e. ‘genetic’ relatedness) is the only possibility. Languages replace lexical items over time; the more the time, the fewer the items retained from the ancestral language, and the fewer the potential cognates. This is why linguists agree that it is convenient to start with cognates belonging to the basic vocabulary,

because they resist borrowing more than other components of the vocabulary. The final step is to determine what system of sounds the ancestral language apparently had and what the rules were for combining these sounds. The success of any reconstruction depends on the material at hand and on the ability of the comparative linguist to decipher what happened in the history of the languages under comparison. Generally, the longer in the past the proto-language split up, the more linguistic changes will have occurred and the more difficult the task of reconstruction with full success is going to be. The choice of languages is also very important. If we simply pick some arbitrary languages and compare them, we cannot expect to see systematic correspondences appear, even if the languages selected are indeed genetically related. The comparative method has to be applied carefully and thoughtfully, and every piece of information must be taken into consideration as possibly relevant. Linguists have to be aware of a number of potential difficulties that might lead the reconstruction into error. Perhaps the most obvious point is that the comparative method cannot recover any feature of the ancestral language that has disappeared without a trace in all the attested daughters. Other problems are loanwords, pure coincidences, nursery words, imitative words, and phonaesthetic words (sound symbolism), because these tend to be somewhat similar even in unrelated languages. References Beekes, Robert S.P. 1995. Comparative Indo-European linguistics: an introduction. Amsterdam, Philadelphia: Benjamins. Brogyanyi, Bela, and Reiner Lipp (eds.) 1993. Comparativehistorical linguistics: Indo-European and Finno-Ugric. Amsterdam, Philadelphia: Benjamins. Collinge, N.E. 1985. The laws of Indo-European. Amsterdam, Philadelphia: Benjamins. Meillet, Antoine. 1953. Introduction à l’étude comparative des langues indo-européennes, 8th edition. Paris: Hachette. Meillet, Antoine. 1967. The comparative method in historical linguistics, translated from the French by G.B. Ford, Jr. and H. Champion Paris. Oslo: Instituttet for Sammenlignende Kulturforskning. Saussure, Ferdinand de. 1968. Mémoire sur le système primitif des voyelles dans les langues indo-européennes (1879; Report on the Primitive System of Vowels in Indo-European Languages) (Reprografischer Nachdruck der Ausg. Leipzig 1879.) Hildesheim: Olma. Sihler, Andrew. 1995. New comparative grammar of Greek and Latin. New York: Oxford University Press.

LAURA DANILIUC See also Bopp, Franz; Grimm, Jacob; Historical Linguistics; Saussure, Ferdinand de

223

COMPOSITIONAL SEMANTICS

Compositional Semantics Compositional semantics includes all approaches to the meaning of natural language that subsume, in one formulation or another, the principle of ‘compositionality of meaning’. Very generally put, this principle reads: ‘The meaning of a complex expression is a function of the meanings of its parts.’ This is often called Frege’s principle, although it cannot be found in explicit form in his writings. Almost any theory of meaning for natural language is based on this principle. There is a psychological motivation for this, because the principle is able to explain how a speaker can understand sentences never heard before. A more explicit formulation of the principle is: ‘The meaning of an expression is a function of the meanings of its parts and of the way they are syntactically combined.’ This formulation makes it explicit that the meanings of the parts are not sufficient for determining the meaning of the whole, and that the way the parts fit together is responsible for differences in meaning (John hit Mary vs. Mary hit John). In this form, the principle is rather uncontroversial, although when one looks at the way the relation is a function of is spelled out in different semantic theories, there is considerable diversity. This principle is the fundamental one in Montague Grammar, an application of the essential methods of formal logic to the study of natural language semantics: it helps in designing a finite system (with a limited number of words and ways of combining them) with infinite output (meanings for all possible sentences). One of the basic tenets of Montague Grammar is the ‘rule-to-rule’ correspondence between syntax and semantics: the syntax contains several rules that provide several ways to form a complex expression from parts, each with a specific semantic effect (i.e. for each syntactic rule, there is a corresponding semantic rule). Put informally, the meaning of a complex expression results from the application of a

semantic rule to the meanings of the parts of the expression. Some natural language expressions have been claimed to be difficulties for a compositional approach to semantics. One such difficulty is represented by compound nouns such as bookworm (which is not normally a worm but a person) or ashtray (which is not a tray). However, most of these challenges have been successfully met by insightful compositional analyses. Due to its psychological plausibility, the compositionality principle characterizes not only formal approaches to semantics but also one major area of cognitive linguistics. In Adele Goldberg’s Construction Grammar, for instance, syntactic constructions are considered to be substantive principles of semantic composition: since the constructions per se are taken to be meaningful, the meaning of linguistic expressions is the result of integrating the meanings of the words into the meanings of constructions. References Goldberg, Adele E. 1995. Constructions. Chicago, IL: University of Chicago Press. Janssen, Theo M.V. 1997. Compositionality. Handbook of logic and language, ed. by Johan van Benthem and Alice ter Meulen. Oxford: Elsevier and Cambridge; MA: MIT Press. Katz, Jerrold J., and Jerry A. Fodor. 1963. The structure of a semantic theory. Language 39. Partee, Barbara H. (ed.) 1976. Montague Grammar. New York: Academic Press. Partee, Barbara H. 1984. Compositionality. Varieties of formal semantics, ed. by Fred Landman and Frank Veltman. Dordrecht: Foris. Thomason, Richmond H. 1974. Formal philosophy. Selected papers of Richard Montague. New Haven, CT: Yale University Press.

ANDREA SANSÒ See also Meaning; Montague, Richard; Semantics

Compounding Compounding is the process of creating new words whose elements are smaller words. For example, the relatively new English word chatline is created by compounding (a process sometimes termed ‘composi224

tion’) from the two simpler words chat and line. A word formed in this way is called a ‘compound’. The ‘words’ that are the constituent parts of compounds may be subject to specific formal requirements,

COMPOUNDING which depend on the language concerned. In the Latin compound liquefacio, ‘to dissolve’, facio means ‘I make’, but lique, although clearly related to liquidus, ‘liquid’, cannot occur by itself, but only in compounds and other derivative word-forms. In the Danish compound sovesofa, ‘sleep sofa, sofa bed’, the verb in the first element of the compound is in the infinitive. In other languages, although it is rare, finite verb-forms can be used, as in Arabana-Wangkangurru yanhi·rnda· tharka·kura ‘speak-pres-stand-past.continuous = they were standing talking’. In languages like Finnish and Sanskrit, both basic and inflected forms of nouns can be found in the first element of compounds. Compounds usually consist of two parts: a modifying word and a modified one. The modified word is called the ‘head’ of the compound because it contributes the dominant meaning and determines whether the compound is, e.g. a noun or a verb. Whatever the form of the modifying element of the compound, it is usually fixed in the compound and does not change depending on the environment in which the compound occurs. The head element, however, may be inflected, e.g. for tense (in the case of verbs) or number (in the case of nouns). Typically, compounds denote hyponyms, or special cases, of their head. A chatline, for example, is a type of line, not a type of chat. In English, the head is almost always the right-hand element in the compound. In a language like French, however, the head is frequently on the left: a leçon-cuisine, literally ‘lesson cookery’ (cooking lesson), is a type of leçon and not a type of cuisine. Rolls-Royce is quite correctly translated into French as Royce-Rolls. A remarkable number of languages show both patterns of headedness in compounds. In either case, compounds are ideal constructions for creating names for subtypes. Because compounds are typically hyponyms of their head element, they also take their part of speech from their head element: for example, chatline has a noun in its head and is hence a noun, user-friendly has an adjective in its head and is thus an adjective, and stage-dive has a verb in its head and is consequently a verb. Compounds in which this is true are called ‘endocentric compounds’ because their center of meaning lies within the compound itself (Greek endo, ‘inside’). There are also ‘exocentric compounds’ (Greek exo, ‘outside’), which are not hyponyms of their head element (an egghead is a kind of person, not a kind of head) or, more radically, whose part of speech is not deducible from that of their elements (for example, put-down consists of a verb, put, and a preposition, down, but it acts as a noun). Although exocentric compounds are not hyponyms of their own heads, they can still be seen to be headed, in that one element modifies the other. A redneck may not be a type of neck, but it is still clear that red modifies neck and that redneck thus has its own right-hand head;

a bas-bleu (French) ‘bluestocking’ (literally, ‘stocking blue’) may not be a stocking, but not only does bleu obviously modify bas, the construction as a whole takes its gender from bas. The same is not so obviously true of another type of compound, sometimes called ‘coordinative compounds’. Here, two elements are put together, and the compound either denotes the unity of the two elements or is used as the superordinate term for things that include the two named elements. __ _ In the first case, we get things such as Marathi ai -ba p, ‘parents’ (literally ‘mother–father’); in the latter case, we get things like Kannada bassu karu, ‘vehicles’ (literally ‘bus–car’). The borderline between coordinative and endocentric compounds is often blurry, and classifications may vary for different scholars: several distinguishable classes of compounds similar to coordinative compounds are found, but they are not always kept clearly apart. All the examples that have been considered so far are examples of what are usually called ‘root’ or ‘primary’ compounds. One of the typical characteristics of primary compounds is the wide range of possible meaning relations between the elements. For example, a flour mill grinds flour, but a water-mill is powered by water, whereas a sawmill uses a saw as an instrument in its work. There is nothing overt in these words to show that they have different interpretations, and the same form may have different interpretations on different occasions: the Washington flight may be the flight to Washington or the flight arriving from Washington. A contrasting set of compounds, usually called ‘synthetic’ or ‘verbal’ compounds (some scholars distinguish between the two labels), has a very restricted interpretation. These are words such as bus driver, in which the head element is a derivative of a verb and in which the modifying element must be interpreted as an argument of that verb. The argument concerned is usually the direct object of the verb (as in bus driver, ‘someone who drives buses’) but may on occasions be other arguments, such as instrument in hand-sewn and subject in consumer spending. Precisely where compounding stops and sentence formation begins is a matter of some dispute. In particular, it is not clear whether structures like Onondaga waʔ·ha·hwist·ahtu·ʔt·aʔ, ‘he lost money’ (in which the direct object money is directly attached to the verb lost) should be regarded as a compound or some other construction. References Adams, Valerie. 1973. An introduction to modern English word-formation. London: Longman. Bauer, Laurie. 2001. Compounding. Language typology and language universals, ed. by Martin Haspelmath, Ekkehard

225

COMPOUNDING König, Wulf Oesterreicher, and Wolfgang Raible, 695–707. Berlin and New York: de Gruyter. Botha, Rudolf P. 1984. Morphological mechanisms. Oxford: Pergamon. Coates, Richard. 1999. Word structure. London and New York: Routledge.

Ryder, Mary Ellen. 1994. Ordered chaos. The interpretation of English noun–noun compounds. Berkeley, CA: University of California Press. Williams, Edwin. 1981. On the notions ‘lexically related’ and ‘head of a word’. Linguistic Inquiry 12. 245–74.

LAURIE BAUER

Computational Linguistics Computational linguistics is a discipline between computer science and linguistics, for which the main concern is the computational aspects of human language. These are aspects that can be put into the form of a sequence of instructions that a computer can understand. For example, if we have a sentence like John runs, we can set up some simple grammar rules. These rules state that a sentence is made up of a noun and a verb, a noun is always John, and a verb is always runs. Of course, this grammar only produces our sentence, but it is a grammar that can be used very easily in a computer program. We only have to put three instructions (Sentence = Noun + Verb, Noun = John, and Verb = runs), in a format that the machine understands, to produce the sentence. We say then that this grammar is fit for computation. We do not include in our definition those tools that computer science gives us for the study of language. Sometimes these are also presented under the generic label, but they are normally not considered to be part of this field. Two main practical goals in computational linguistics have been machine translation and natural language interaction with computers and machines. When we refer to ‘natural language’, we are doing so in contrast to ‘artificial language’. Language can be called ‘natural’ when it can be used for communicating with people in the same way as a human being would do. Machine translation, that is, translation carried out by a computer, was the first goal of computational linguistics. When scientists started to work in this field, it looked like an easy task. It was thought to be just a question of making words and expressions in the source language correspond to words and expressions in the target language. Very soon it became clear that it was not so simple, since language has a structure that is more complex than just a lexicon database. There is more to language than just words. However imperfect the process still is, though, automated translation now gets much better results than in the past. And it has achieved them through a

226

systematic approach to how to perform the task. From the very early years of machine translation studies, we have two main approaches to the problem of automating translation. The first method involves carrying out a direct translation from one language into the other. The second method is the ‘transfer’ grammar approach, in which everything is translated to a kind of intermediate language. From here, it is easier to make a translation to many other languages. This intermediate grammar may have the form of a real human language (Esperanto or English are examples), or a certain symbolic language. A simple example is the following: we have the French sentence Marie aime Jean and then we can make a translation to a simplified English sentence of the form Marie(S)-loveJean(O), with S marking the subject and O the object. From this intermediate sentence, we translate again, now into Spanish: Marie ama a Jean. The technique is very useful when translating to and from many different languages, since we have an intermediate expression from which we can easily make translations that are faithful to the same extent to the basic structure. Both techniques have their advantages and drawbacks. In addition, even if there should not be any problem at all with the translation strategy adopted, there are other facts that machine translation has not properly accounted for to date, the most important being the resolution of ambiguity in language. The phrase The cover of the book which I like has syntactic ambiguity, since we cannot be sure whether what I like is the cover or the book. Similarly, The seal is on the rock has semantic ambiguity, since we do not know whether what is on the rock is an animal (a seal) or some kind of stamp (a seal). We may also have pragmatic ambiguity, of which a very simple case is the difficulty to translate the English pronoun you into languages like Spanish, French, German, or Japanese, where there are several different ways of referring to a second person according to degrees of formality or to the relation between the speaker (or writer) and the

COMPUTATIONAL LINGUISTICS hearer (or reader). Obviously, the computer cannot infer this type of information, which depends on the communicative situation. Of all these types of ambiguity, the most important for machine translation is the last one, since it cannot be resolved by looking at the immediate linguistic context. Another main area that computational linguistics has always been interested in is the interaction with machines using natural language. This can be done through the use of the written word or through the use of oral speech. In the latter case, the system needs to incorporate a module for speech recognition, in addition to the other modules used for linguistic processing that we have mentioned. The usefulness of mechanisms that react to ordinary speech in activities where free hands are needed has led to the development of such systems for cars, planes, or control of processes in factories, which are typical examples. In all these cases, in addition to the procedures of language analysis and generation, there are special calculations that take into account the noise that may accompany ordinary language. In particular, there has been a special concern in trying to produce language in a manner as human-like as possible in expert systems. An expert system is a system capable of reasoning in a way similar to a specialist in a certain field, and it is therefore necessary, or at least very convenient, that the queries can be carried out in a language as close as possible to natural human language. This is so because expert systems are intended for use in many kinds of settings, and may have to be operated by persons with different degrees of expertise. It is also important to try to reproduce the typical interactive client–expert exchange. In natural language interaction there are two main strategies. The oldest and the easiest is to check for certain words and expressions, ignoring the rest of the message. This is the direct, crude approach. According to the expressions the system recognizes, it gives certain answers. This was the approach used in Eliza, a program created by Joseph Weizenbaum in 1964–1966. It was designed to take on the role of a Rogerian psychotherapist, and it produced responses according to the trigger expressions that appeared. Thus, to a sentence with the word friends, it could react saying something like Why are you thinking of your friends now? so that to the sentence I have many friends, or My wife has many friends, we would get the question, Why are you thinking of your friends now? The rigidity of this approach is obvious, although it is the most economical by far. On the other hand, we have systems where there is a special module for language recognition, in which sentences are parsed (analyzed) and semantic inter-

pretations are assigned to them. This is the true computational linguistics approach, which has also found problems similar to the ones mentioned for machine translation. It is this approach that shall be explained below. In computational linguistics, the structure of sentences needs to be represented in some way. In order to do this, the computer has some internal arrangement of data that corresponds to the semantic structures of the knowledge that we have in our mind, and that helps clarify the way in which the linguistic message organizes the information it conveys. This arrangement is in the form of a semantic representation. The computer also knows about the rules of the language for putting words and sounds (or letters) together. These are syntactic and phonetic (or graphemic) rules. The semantic representations used in computational linguistics try to make the most of the resources that the machine offers, but they also try to be faithful to the structure of concepts. Let us take the example of the concept ‘bird’ to illustrate how these representations can be made. When we think of a bird, we immediately visualize an animal with wings and feathers. In our mind, we have a link between the concept ‘bird’ and the concepts ‘wing’ or ‘feather’. If we write the words ‘bird’, ‘wing’, and ‘feather’, and draw two lines from the label ‘bird’, one to the label ‘wing’ and another to the label ‘feather’, we will have a representation of the above image by using a labeled diagram. If we now add more words to the diagram, such as, for instance, ‘beak’, ‘egg’, ‘nest’, or ‘to fly’, the diagram becomes more complex, but now the links start to be of different kinds. A bird normally ‘has’ a beak, ‘lays’ eggs, ‘builds’ nests, or just ‘flies’. These different types of relation also need to be labeled. Finally, we now know that not all these features are necessary and sufficient, but rather prototypical. It is difficult to imagine a bird without a beak or wings, but certainly not all of them fly or build nests. This is more difficult to put in our diagram, but we can try to use some numbers that denote to which extent these features can be applicable. When our diagram becomes larger, it should look like a net. In fact, it is a way of depicting a semantic net, a kind of representation that is being used in computational linguistics since the late 1960s. Semantic nets are incorporated in the data of computational linguistic applications, so that the semantic structure of sentences is interpreted or generated by building small nets that match the connections that appear in the larger ones. As a formalism, a semantic net consists of a collection of points (the labels ‘bird’, ‘wing’, ‘feather’, etc. in our example) called nodes, each representing a concept. A node may be

227

COMPUTATIONAL LINGUISTICS connected to any other node in the net by means of relations that in turn carry labels (‘have’, ‘build’, ‘fly’, etc.). There are some conventions: two very typical relations normally referred to in computational linguistics books are the labels IS-A (x ‘is a’ y) and H-A-P (x ‘has as part’ y). But any other that corresponds to common English verbs or adjectives (LIKE, COLOUR, etc.) can also be used. Let us look at another example. A semantic net for the sentence Mary likes Peter could be the following: GIRL IS-A MARY H-A-P EYES COLOUR BLUE

LIKES

BOY IS-A PETER

assign several statements as conditions for the truth of another one, and the resulting structure is called a rule. An example will help to understand this. Below there is PROLOG rule: aunt (X, Y) :female (X), brothers (X, Z), parent (Z, Y). Let us translate these statements into understandable English (on the right): aunt (X, Y) :female (X), brothers (X, Z), parent (Z, Y).

This net provides a semantic representation for the message conveyed. However, most of the information is not in the sentence itself, which only carries labels. We retrieve all the additional information from structures of knowledge that are either in our mind or in the computer’s data. There is a matching process in order to see which parts of the sentence’s semantic structure and the program’s data coincide. It is like having two pictures in two transparencies, A and B, of a face and a nose. The nose has more detail in B, but it coincides partially with what is drawn in A; hence, we know that it corresponds to that face. In addition, when we put the two transparencies together, the face receives a better drawing, with the addition of a more detailed nose. This matching process is currently referred to in computational linguistics as unification, and it is considered a very powerful resource for natural language processing. When a computer programmer uses a programming language such as PROLOG, for instance, there is an extensive use of unification. We can illustrate this by having a close look at how PROLOG works. As has just been stated, PROLOG is a programming language, and all programming languages have a vocabulary and a syntax, exactly in the same way as natural languages do. In the vocabulary, we have strings of symbols that stand for variables (with capital letters, e.g. X, YW, AA) or for constants (without capital letters, e.g. a, bb, person, mother). The corresponding items in a natural language for variables would be pronouns (which stand for other terms). For constants, we have proper nouns (which, in ideal cases, refer to unique entities). Statements in PROLOG normally have a structure in which there is one predicate and several arguments. If A is the predicate and x, y the arguments, the convention is to write A (x, y). The symbol :- is used to 228

‘X (a variable) is Y’s (another variable) aunt if...’ ‘X is a female’ ‘X and Z (a third variable) are brothers/sisters’ ‘Z is a parent of Y’

In plain English, this would be something like ‘Some entity X is the aunt of some other entity Y if the entity X is a female, the entities X and Z are brothers or sisters, and the entity Z is one of Y’s parents.’ The computer programmer writes this rule as part of a larger program, and we can also add some information to the program’s database by writing the following statements: female (mary) ‘Mary is a female’ brothers (mary, john) ‘Mary and John are sister and brother’ parent (john, jim) ‘John is one of Jim’s parents’ (Notice that the terms ‘female’, ‘mary’, ‘john’, etc. have no capital letters, because they are constants.) The computer operator may wish at some stage to know who is Mary’s nephew, or to whom she is an aunt. We can then make use of the ‘aunt’ rule. The operator writes: ?- aunt (mary, Y)

‘Mary is Y’s aunt’

As this is a plain statement (not a rule), the PROLOG system looks for a suitable term that can change the variable Y into a constant. If you look at the information provided in the statements that have been incorporated into the program so far, this term is jim. Accordingly, it answers: Y = jim We then say that jim instantiates the variable Y. The PROLOG system has carried out a unification process in order to come to the right solution. It has

COMPUTER-ASSISTED LANGUAGE LEARNING (CALL) unified the rule and the statements in the following way: aunt (mary [=X], Y) :female (mary [=X]) brothers (mary [=X], john [=Z]) parent (john [=Z], jim [=Y]) So that Y = jim. Unification is a very important procedure in computational linguistics. It simulates on a computer the processes that take place when we compare information from our mental store with the information in the linguistic message. Another well-known programming language for artificial intelligence is LISP, of which many different versions have been created afterwards (schema, xlisp, etc.). LISP stands for ‘List Processing’, and it is a complementary approach to PROLOG. Whereas in PROLOG we had a deductive mechanism that unified knowledge structures, in LISP we have a powerful engine for processing lists. This is useful in computational linguistics, where syntactic parsing is an important step in the analysis of sentences. An additional important feature, which is very convenient for many artificial-intelligence approaches is that LISP programs are able to self-modify, making it possible to write code that ‘learns’, or that adapts itself to changing circumstances. PATR (Shieber 1986) is a declarative language that has eventually become a standard in computational linguistics. This is so because it is very easily adaptable to many different theories. In fact, many current developments use feature structures and unification in a PATR style. The most prominent characteristics of the other two languages that we have seen here, unification (PROLOG) and string concatenation (LISP), are used in this language. With respect to semantic nets, computer scientists soon found that representations like the ones shown above fail to give detailed information about the situations that are represented. Location, time, setting, etc., were notions that did not appear in the structures. This led to more complex representations, in which concept

nodes were used for events as well as objects, and cases appeared as labels in semantic nets. Parallel to the relation IS-A for objects, there is also now ACTOF. In the sentence John throws the ball, John is linked through an AGENT relation to a node. ‘The ball’ is also linked to that same node by the relation OBJECT. And the node is linked to ‘throw’ by means of ACT-OF, which therefore links particular instances of events to the general node for a certain action in the knowledge database. An example of how the semantic net paradigm can be realized is R.C.’s (1973) theory of conceptual dependency. This author’s graphs follow the formal structure of a dependency grammar, and have four different types of nodes. These are symbolized by PP, ACT, PA, and AA, corresponding very roughly to noun, verb, adjective, and adverb, respectively. Schank makes use of a small set of primitive actions (PROPEL, MOVE, INGEST, EXPEL, GRASP, PTRANS, MTRANS, ATRANS, SPEAK, ATTEND, and MBUILD) and the default action DO. With the help of a small number of states, this author claims that the meaning of all verbs can be represented. It is interesting to see how the semantic primitive approach is used here. In addition, we also have rules that must be applied for well-formed conceptualizations. Sets of semantic primitives have been used in many CL systems. However, it is important to clarify here that there is no claim about their universality or a very strong psychological justification. The use of a limited set has more to do with a question of convenience. References Schank, R.C. 1973. Conceptual dependency: a theory of natural language understanding. Cognitive Psychology 3. 552–631. Shieber, Stuart M. 1986. An introduction to unification-based approaches to grammar. CSLI Lecture Notes, Vol. 4. Stanford. Wilks, Yorick. 1973. Preference semantics. Memoranda from the Artificial Intelligence Laboratory 206. Stanford, CA: Stanford University.

CARLOS INCHAURRALDE See also Ambiguity; Artificial Machine Translation

Intelligence;

Computer-Assisted Language Learning (CALL) Computer-assisted language learning (CALL) has been an important aspect of language teaching and learning since the 1960s, and its developments have been influ-

enced and shaped by those occurring more generally in the fields of education and applied linguistics. The first applications of computers in language teaching were 229

COMPUTER-ASSISTED LANGUAGE LEARNING (CALL) based on large-scale, specialized mainframe computers, and hundreds of person-hours of programming were needed to develop the instructional software. These computer-assisted instruction (CAI) systems were meant to be ‘stand-alone’, which meant that all aspects of teaching and learning were contained within the hardware and software. It was very difficult (or often impossible) for teachers to tailor the software to their classroom teaching, and some of the earliest complaints about CAI were related to this problem. The teaching approach used in CAI systems was tutorial (drill and practice) and was based on behaviorist learning theories that were popular at the time. Among the early CAI projects, perhaps the two most influential for language teachers, were PLATO (Programmed Logic for Automatic Teaching Operations) and TICCIT (Timeshared, Interactive Computer-Controlled Information Television), developed at the University of Illinois and Brigham Young University, respectively. In the 1970s and 1980s, the change to more communicative approaches to language teaching and learning coincided with the development of microcomputers and the corresponding proliferation of software. Many programs were still tutorial (drill and practice), but more general learning tools, including, for example, word-processing, database, spreadsheet, and, by the 1980s, communications software, were adopted from the business environment. In addition, authoring software (high-level programming languages) helped teachers write computer programs to complement their other classroom activities. Consistent with new communicative approaches to language teaching, teachers began to group students so that they could work together in teams to write simple software programs. Not only did this new instructional approach help students to become more independent learners, but also team-programming provided them with opportunities to practice language in authentic communicative contexts. Since the 1990s, with the advent of inexpensive memory and other hardware components, for example, resident CD-ROM or DVD (digital video disk) players, teachers have had access to a range of commercially developed, multimedia language software. Such software is usually created as a hypertext environment in which text, sound, video, and graphics are creatively linked to one another. Students can point and click on different words, ideas, or branches of a program and move through instructional material at their own pace, according to their personal interests. However, multimedia software can be expensive and cannot always be tailored to individual classroom learning needs. Some researchers have also questioned whether reading and writing in hypertext environments, disjointed as it can be, helps or hinders language learning.

230

Also since the early 1990s, access to global communication networks and the World Wide Web, and the availability of point-and-click Web-browsing software, has become increasingly easy to use and inexpensive. The Internet has made global communication, through electronic mail (e-mail), commonplace and has led to a variety of tandem language-learning projects. For example, English-speaking students studying French in the United States could be paired electronically with French-speaking students studying English in Canada. Both groups write to each other in the target language and then send, read, or reply to messages at their convenience; they do not need to be on-line at the same time (asynchronous communication). Learning is aided as students work at their own pace and help each other to notice problems in grammatical forms or learn how to express ideas in more natural, idiomatic language. In addition, students can provide insights to their e-mail partners about respective cultural practices and norms. Such peer-based sharing of ideas to help learners develop a deeper understanding of language is called ‘scaffolding’ and has been influenced by learning theories from education and psychology. In contrast, synchronous communication, in which all people are on-line and writing (typing) at the same time, is also used in CALL classrooms. Synchronous communication, for example, in on-line chatrooms, requires students to log into centralized ‘rooms’ on the Web, where they can adopt pseudonyms or even assume different on-line identities. The pace of written exchanges is fast, and the style of language is often informal. Many language teachers believe that unless there is a well-developed curriculum structure and educational purpose to synchronous communication, students may not learn appropriate linguistic forms. In fact, some researchers believe that on-line chatroom communication involves a new, hybrid form of language that incorporates features of both speaking and writing. Although the Internet supports the exchange of written text through e-mail or in chatrooms, it also provides access to a variety of other language-teaching resources. These include, for example, access to Web-based software (much of which can be downloaded) and specialized Web sites containing organized links to language-learning resources around the world. For example, students can read daily newspapers, listen to live radio broadcasts, or watch and listen to television news from a variety of countries. As a result, access to ‘authentic’ language resources is becoming easier for teachers, but how to evaluate, organize, and integrate online materials so that students can learn language in a structured manner have become important issues in teacher training. Certainly, there is a growing acceptance

COMRIE, BERNARD that the role of the teacher and student in CALL classrooms is changing, with teachers assuming more facilitative, guiding roles and students becoming more autonomous learners. The latest technological changes that are affecting CALL include wireless communication, which relies on radio signals, rather than physical wiring, to connect to the Internet, and a variety of new software tools. For example, it is now possible for students to use wireless laptop computers, video cameras, and word-processing and movie-editing software to write, film, and review their own language materials. Such a project-based approach to teaching and learning could effectively integrate current theories of autonomous learning, peer scaffolding of learning, and authentic communication. Regardless of the technology used, however, the overriding issue for teachers will remain how to coordinate learning experiences so that students have meaningful access to language—however sophisticated their computers and whichever software they use.

References Bush, Michael D., and Robert M. Terry (eds.) 1997. Technology-enhanced language learning. Lincolnwood, IL: National Textbook Company. Egbert, Joy, and Elizabeth Hanson-Smith (eds.) 1999. CALL environments: research, practice, and critical issues. Alexandria, VA: TESOL Publications. Levy, Michael. 1997. Computer-assisted language learning: context and conceptualization. Oxford: Clarendon Press and New York: Oxford University Press Inc. Murray, Denise E. 1995. Knowledge machines: language and information in a technological society. New York: Longman. Pennington, Martha C. 1996. The power of CALL. Houston: Athelstan Publications. Taylor, Todd W., and Irene Ward (eds.) 1998. Literacy theory in the age of the internet. New York: Columbia University Press. Warschauer, Mark. 1999. Electronic literacies: language, culture, and power in online education. Mahwah, NJ: Lawrence Erlbaum Associates. Warschauer, Mark, and Richard Kern (eds.) 2000. Networkbased language teaching: concepts and practice. Cambridge: Cambridge University Press.

E. MARCIA JOHNSON

Comrie, Bernard Bernard Comrie is one of the leading figures in language typology. As a fieldwork typologist, Comrie has carried out studies on Fula, Tunisian Arabic, Malayalam, Moroccan Arabic, Khmer, Tsez, Sudanese Arabic, Maltese, and Basque, on which he has taught field methods classes. He has also worked in depth in the field on two additional languages: Haruai and Bezhta. Comrie has also taught and researched Historical Linguistics and Diachronic Syntax. Although typological linguistics is the overall approach that is closest to his own, his work has very close ties with cognitive linguistics. Even if as a formal framework, Relational Grammar is now virtually dead, in terms of formal approaches to grammar, Comrie believes that linguists have learned a lot about the grammars of different languages and about the general properties of language from that framework. Comrie has been influential in the recent upsurge of interest in linguistic typology, both through his general books on the topic and through individual articles. Within typology, the areas in which he has had most impact are relative clauses, causative constructions, and grammatical relations. In addition, his books on tense and aspect have had an important role in gener-

ating interest in these topics, even among linguists who end up disagreeing with his approach. Some common linguistic terms are actually used for the first time in his papers. The term ‘accessibility hierarchy’ came with Keenan and Comrie (1977). His book Aspect played a major role in getting the distinction widely accepted between ‘perfect’ and ‘perfective’. He is perhaps the first linguist to introduce, among other terms, the ‘S, A, P’ terminology (as opposed to Dixon’s similar ‘S, A, O’). One of the main features in his thinking has been the integration of links between language and other phenomena, in particular cognitive and social. He is in search of correlating linguistic and other (genetic, archeological) evidence for human prehistory. He insists that these different disciplines provide us with different windows into human prehistory that must then be reconciled with one another, rather than presupposing, for instance, a close connection between language families and biological populations. Comrie’s ideas have not remained static. In some of his later articles (e.g. ‘Rethinking the typology of relative clauses’, 1998), he has argued for substantial revisions to the ways in which typologists and other

231

COMRIE, BERNARD linguists look at relative clauses, in particular, paying more attention to semantic and pragmatic factors. With regard to aspect, he gives more import to the role of lexical properties of verbs than he did in the mid-1970s. Comrie’s first publications were articles derived from his doctoral dissertation, and thus dealing with generative grammar. However, his first articles in the field of universals and typology include ‘Noun phrase accessibility and universal grammar’, written jointly with Edward L. Keenan (1977). He has a remarkable number of publications, mostly on the typology of ergativity, aspect, causative constructions, and nominalizations. His major publications include Aspect (1976), Lingua descriptive studies: questionnaire (with Norval Smith, 1977), The Russian language since the Revolution (with Gerald Stone, 1978), Language universals and linguistic typology: syntax and morphology (1981), The languages of the Soviet Union (with B.G. Hewitt and J.R. Payne 1981), Tense (1985), and The Russian language in the twentieth century (with Gerald Stone and Maria Polinsky 1996). To this list, we should add his several important edited volumes (like Classification of grammatical categories (1978), Studies in the languages of the USSR (1981), The world’s major languages (1987), The Slavonic languages (with Greville G. Corbett 1993), and some others, along with dozens of articles on language typology.

Biography Bernard Comrie was born on May 23, 1947 in Sunderland, England. By 14 years of age, he knew French and had already started learning Latin, German, Ancient Greek, and Russian. Coming across R.H. Robins’ textbook on linguistics left a deep impression on him, showing him not only that there is a field called linguistics, but that it is also a systematic, scientific endeavor. Comrie studied Modern and Medieval Languages (mainly French, German, Russian, Linguistics) at the

University of Cambridge, receiving a B.A. (1968) and a Ph.D. in Linguistics (1972). His dissertation ‘Aspects of sentence complementation in Russian’ was written under the supervision of Pieter A.M. Seuren. Comrie taught linguistics as a University Lecturer in Linguistics at the University of Cambridge (1974–1978). In 1978, he joined the Department of Linguistics at the University of Southern California, Los Angeles, and worked there as a Professor of Linguistics until 1998. Since 1997, he is director of the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, in addition to being a part-time Distinguished Professor of Linguistics at the University of California, Santa Barbara (since 2002). In addition, he holds honorary positions as a Research Professor of Linguistics, University of Southern California (since 1998) and as a Professor of Linguistics, University of Leipzig (since 1999). References Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press. Comrie, Bernard. 1979. Noun phrase accessibility and universal grammar. Linguistic Inquiry 8. 63–99. Comrie, Bernard. 1989. Language universals and linguistic typology: syntax and morphology, 2nd edition. Oxford: Blackwell and Chicago: University of Chicago Press. Comrie, Bernard. 1998. Rethinking the typology of relative clauses. Language Design 1. 59–86. Dahl, Östen. 1987. Comrie’s tense. Folia Linguistica 21. 489–502. Declerck, Rennat. 1986. From Reichenbach (1947) to Comrie (1985) and beyond: towards a theory of tense. Lingua 70. 305–64. Kulikov, Leonid, and Heinz Vater (eds.) 1998. Typology of verbal categories: papers presented to Vladimir Nedjalkov on the occasion of his 70th birthday (Linguistische Arbeiten 382). Tübingen: Max Niemeyer. Rakhilina, Ekaterina V., and Yakov G. Testelets (eds.) 1999. Typology and linguistic theory: from description to explanation, for the 60th birthday of Aleksandr E. Kibrik, 319–39. Moscow: Jazyki russkoj kul´tury.

BEHROOZ MAHMOODI-BAKHTIARI

Configurationality English is a configurational language because the grammatical functions subject and object appear compulsorily in particular syntactic positions: nearly every English clause must have an overt preverbal subject and every English transitive clause must have an overt

232

postverbal object. Of the three main constituents of the clause—subject, verb, and object—the syntactic relationship between verb and object is bounded, as opposed to the syntactic relationship between subject and verb, which is unbounded. The strong syntactic

CONFIGURATIONALITY ties between verb and object, which together constitute a phrasal unit, the verb phrase (hereafter VP), are evidenced by tests like VP deletion (Mark washed his car and Bill did [wash the car] too), VP pronominalization (Mark washed his car and so did Bill), and VP fronting (Mark said that he would wash his car and wash his car he did). What these tests show is that subjects appear outside the verb phrase in syntactically unmarked English sentences (that is, less complex and more frequent than their unmarked counterparts), whereas objects are immediately contained in the VP in English. Subjects and objects, therefore, exhibit different properties of coding and behavior (coding properties are morphological properties such as agreement with the verb, while behavior properties are distributional and include, for example, the omissibility of the subject in a coordinate subject construction like Matthews complained angrily and left the room): for instance, in English the verb agrees with the subject— not with the object—in number and person (He writes vs. They write) and the subject—not the object—of dependent clauses can be raised to—becomes—subject or object of the matrix (main) clause under certain conditions (They consider that she is bright/I consider her bright/She is considered bright). This correspondence between grammatical functions and syntactic positions or phrase structure configurations has motivated the configurational definition of subject and object according to which the subject is the noun phrase directly dominated by the node sentence and the object is the noun phrase directly dominated by the node VP (subconstituents are immediately dominated by the constituents where they belong). This configurational definition puts aside functional relations like subject and object to the benefit of structural aspects of constituency, including, at least dependence (constituents have heads on which subconstituents depend), hierarchy (constituents consist of compulsory and optional subconstituents), and linearization (constituents take up a clausal position relative to the verb). In this approach, case assignment (as in who–whom–whose–to whom) and agreement (as in I eat vs. she eats) crucially depend on the asymmetry between the noun phrases that bear subject and object: the object requires syntactic continuity with respect to the verb; the subject does not. Syntactic continuity is defined as a relation of adjacency between a head and one of its dependents or between two heads: in Peter sent them to the Greens, the preposition and the noun phrase it governs are adjacent to each other in the linear order of the sentence, whereas in Whom did Peter send them to? the sequence noun phrase-stranded (that is, final) preposition is interrupted by the subject, the verb, and the object, thus constituting an instance of syntactic

discontinuity. Syntactic continuity is often referred to in the literature as local dependency, as opposed to nonlocal dependency or syntactic discontinuity. Theoretically speaking, the problem arises of how to account for syntactic discontinuity in grammatical phenomena like case assignment and agreement. In general, two solutions have been adopted: monostratal theories recognize a single level of syntactic representation, or, more clearly, a single linguistic representation, in which the order and form of the elements of the expansion (including grammatical case and phonological properties of accent, rhythm, and intonation) do not necessarily reflect the order and form of the elements of the sentence that the speaker utters, whereas multistratal theories recognize more than one level of syntactic representation, or, in other words, more than one linguistic representation, in such a way that every level represents the sentence that the speaker utters in a less abstract way than the preceding level (abstractness is understood as the distance between the linguistic expression and its representation). Multistratal theories justify the necessity of defining multiple levels of representation on the grounds of the existence of syntactically marked constructions (that is, more complex and less frequent than their unmarked counterparts) like passives, questions, preposition stranding, etc., which result from the application of a number of structurechanging operation rules or transformations between the deep (more abstract) and the surface (less abstract) levels of syntactic representation. Structure-changing operations fall into three types: first, operations of deletion of specified elements, as in I met a man who was wearing a yellow coat vs. I met a man wearing a yellow coat; second, operations of substitution of one specified element by another specified element, as in These are the shoes [I bought the which in Paris] vs. These are the shoes which I bought in Paris; and third, operations of permutations of specified elements, as in Jim doesn´t like bananas vs. Bananas Jim doesn´t like. The existence of the node VP, which is criterial for configurationality, has been questioned on a cross-linguistic basis. The interlinguistic evidence provided includes: first, the lack of adjacency between the object and the verb in VSO and OSV languages; second, the syntactically unconstrained permutability of the subject and object in languages such as Latin, where the order of the major elements of the clause can be SVO as well as OVS; and, third, the optionality of overt subject and object in languages like Navajo, a language in which some clauses consist simply of the element V. In consequence, a division into configurational (with VP) and nonconfigurational (without a VP) languages is proposed. Although nonconfigurational languages often coincide with relatively free constituent order languages, this is not always

233

CONFIGURATIONALITY the case: there are languages, like Navajo, with a relatively rigid constituent order that qualify as nonconfigurational. Navajo has a rich system of verbal prefixes, including prefixes of person and number of the subject and object, thus allowing for null nominals (phonologically empty, or nonexistent) in the function of both subject and object. Another dimension has been explored in which languages can be configurational or nonconfigurational: the discoursive dimension. A language is discourse configurational if it assigns the discourse functions of topic and focus. The discourse function topic is used for foregrounding a discourse topic about something that is predicated. Such a discourse topic may or may not coincide with the grammatical subject and is usually placed in a particular structural position, typically clause-initial. The discourse function of focus is used to highlight particular aspects of the predication about the topic. The constituent that bears focus either appears in a particular structural position, typically clause-final, or receives special morphosyntactic coding or prominence, or participates in marked syntactic constructions.

References Baker, Mark C.2001. The natures of nonconfigurationality. The handbook of contemporary syntactic theory, ed. by Mark Baltin and Chris Collins. Oxford: Blackwell. Dik, Simon C. 1989. The theory of functional grammar. Part 1: the structure of the clause, ed. by Kees Hengeveled. Berlin: Mouton de Gruyter, 2nd edition, 1997. Dik, Simon C. 1997. The theory of functional grammar. Part 2: complex and derived constructions, ed. by Kees Hengeveld Berlin: Mouton de Gruyter. Hale, Ken. 1983. Walpiri and the grammar of non-configurational languages. Natural Language and Linguistic Theory 1. Hale, Ken. 1989. On nonconfigurational structures. Configurationality. The typology of asymmetries, ed. by László Marácz and Pieter Muysken. Dordrecht: Foris. Kiss, Katherine E. (ed.) 1995. Discourse configurational languages, Oxford: Oxford University Press. Siewierska, Anna. 1988. Word order rules. London: Croom Helm. Van Valin Jr., Robert D. 2001. An introduction to syntax. Cambridge: Cambridge University Press. Van Valin Jr., Robert D., and Randy LaPolla. 1997. Syntax: structure, meaning and function. Cambridge: Cambridge University Press.

JAVIER MARTÍN ARISTA See also English; Navajo and Athabaskan-Eyak Languages; Syntax

Connectionism In the 1980s, the connectionist view, an alternative to rule-based approaches to language, gained increasing acceptance among cognitive scientists. According to this view, our knowledge of language is the result of our interaction with the universe of linguistic utterances that we encounter. The principal tool of connectionist research to study the acquisition and use of linguistic knowledge is the neural network. A neural network is a computer simulation inspired by neural processes that govern the brain. Neural networks consist of primitive nervelike processing units, connected at multiple points. The strength of these connections depends on past experience. Accumulated experiences shape the connections, using weights that represent the knowledge of the network. Environmental events are registered via input units as patterns of activation, where the activation is expressed numerically. Thus, the vector (0, 1) indicates activations of 0 and 1 for two input units. Networks vary in complexity in the number of layers of units and patterns of connectivity. A unit transforms input signals into an activation function and

234

passes the activation forward to other units; therefore, information in neural networks is represented by patterns of activation spread over many neurons and the links between them. The networks are said to represent knowledge in a distributed manner, as a single network can accommodate multiple pairs of independent input–output patterns. Learning is a natural property of connectionist networks. Learning involves a change of the connection weights in the network as a result of a specific training regimen. Connectionist networks are applied to a wide variety of linguistic phenomena, including recognition, pronunciation, morphology, and syntax, typically with an eye toward simulating empirical effects observed in psychological research. The interactive nature and sensitivity to multiple inputs of connectionist networks are well suited to simulating the recognition of various linguistic structures. An early interactive model by McClelland and Rumelhart (1986) was capable of simulating the word superiority effect, the finding that letters are recognized better when they occur in the context of words rather than in isolation.

CONSTITUENCY TEST Connectionist researchers have also trained networks to acquire spelling-to-sound correspondences by associating orthographic input signals with corresponding phonological outputs. Networks have been trained to mimic the acquisition of morphological knowledge of speakers, for example, of the regular and irregular past tense of English verbs, indicating that children more easily learn those irregular verbs that use the stem for the past tense, as in hit. At least two classes of models of syntactic parsing, or sentence construction are advanced through connectionist modeling. One interprets sentences via case role assignments, while the other constructs phrase structure trees. In both approaches, multiple constraints change as a listener progresses through a sentence. In the first model, each successive word in a sentence imposes constraints on the interpretation of the current case assignment and the entire sentence. The second approach uses a recurrent network to acquire and to discover the phrase tree structure of sentences. Although increasingly popular for these and other applications (e.g. the lexicon, language impairment, and language production), neural networks remain controversial. According to Pinker (1999), the models suffer from an inherent design flaw, as they do not take advantage of the computational device of a variable, which can stand for a class of words regardless of the particulars of context. The connectionist models are limited by their contradiction between the averaging mechanism and the value of information about the individual nuances so common in language. Finally, although networks acquire their knowledge without programmer intervention, the programmer has

great latitude in shaping the architecture and the processing details of a network. Frequently, in a manner not apparent to others, a theorist determines all of the features of a model, including the number of input units, hidden units, and output units, as well as the pattern of connections between units. While the networks can simulate many empirical findings, there have been both failures of prediction and very powerful predictions, for example, in the domain of speech recognition. In spite of these criticisms, the neural network approach has become a part of the landscape in cognitive science and—due to its psychological plausibility—will continue to be a significant player in the field. References Anderson, James. 1995. An introduction to neural networks. Cambridge and London: MIT Press. Bechtel, William, and Abrahamsen, Adele. 1994. Connectionism and the mind: an introduction to parallel processing in networks. Oxford (UK) and Cambridge (USA): Blackwell. Ellis, Rob, and Glyn Humphreys. 1999. Connectionist psychology: a text with readings. Hove (UK): Psychology Press. McClelland, Jay L., David. E. Rumelhart, and the PDP Research Group (eds.) 1986. Parallel distributed processing: explorations in the microstructure of cognition, Vol. 2: psychological and biological models. Cambridge, MA: MIT Press. Pinker, Steven. 1999. Words and rules—the ingredients of language. New York: Basic Books. Rumelhart, David E., Jay L. McClelland, and the PDP Research Group (eds.) 1986. Parallel distributed processing: explorations in the microstructure of cognition, Vol. 1: foundations. Cambridge, MA: MIT Press.

KARL HABERLANDT

Constituency Test Constituency and Phrase Structure Constituency has played an important role in grammatical theory, especially regarding the empirical evaluation of different approaches to phrase structure. A constituent corresponds to a unit in the syntactic structure. Consider the sentence The janitor read magazines in the morning. One can account for it as representing a sentence S, containing a noun phrase (NP) the janitor and a verb phrase (VP) read magazines in the morning.

Both NP and VP are units of S, they are constituents of S. The VP in turn can be analyzed as containing the verb read, the NP magazines as well as the prepositional phrase (PP) in the morning, which are all constituents of the VP. A formal representation in which the major constituents of the sentence are labeled would thus be: (1) [S [NP The janitor] [VP read [NP magazines ] [PP in the morning]]]

235

CONSTITUENCY TEST Constituency Tests Many grammatical operations can only target strings of words that form constituents. This has given rise to a number of widely recognized Constituency Tests. The rationale is that a string of words that are taken to form a constituent has to pass one or more constituency tests. The most common tests are summarized below. Replacement or Substitution Test A constituent can be replaced by a proform (e.g. a pronominal — it, he — or expressions such as there, then, do so). The replacing expression can be interpreted in a context so that the meaning of the replaced constituent is completely recovered. For instance, both the underlined subject NP ( a janitor) and the entire VP in (4) can be replaced by proforms (he and did so, respectively), showing that they are constituents.

English; hence, the embedded clause in (6a) cannot be easily topicalized. (7) *[That he is on vacation this week] the janitor said. This restrictive character of topicalization yields an important conclusion. The fact that a given string of words satisfies a constituency test (i.e. any test) is in general sufficient to indicate that this string is a constituent. However, this does not mean that all constituency tests can equally felicitously be applied, since each type of sentence structure may be subject to certain constraints (as, e.g. in topicalization).

Coordination Test

(4) [NP The janitor]j [VP read magazines in the library]k. [ He]j [ did so]k everyday.

The coordination test shows that only constituents can be conjoined (8a), with the additional requirement that they correspond to the same kind of phrase, as shown by the ungrammaticality of (8b):

On the other hand, there is no proform that can substitute for a string of words that spans over parts of different constituents, such as [the janitor read].

(8) a. The janitor read magazines [PP in the morning] and [PP in the afternoon]. b. *Sue drove [NP a car] and [PP to Ottawa].

Displacement Tests A constituent can be realized in a position that is different from its canonical position in a sentence, i.e. it appears to be ‘displaced’. There are different types of displacement, depending on what additional elements are added to the sentence. In Clefting, a structure like (5a) is generated with the addition of it was…that supporting dislocation of in the morning. However, magazines in the morning, which is not a constituent, cannot be clefted, as (5b) shows (the asterisk shows ungrammaticality): (5) a. It was [PP in the morning] that the janitor read magazines. b. * It was [magazines in the morning] that the janitor read. In Preposing, the material added to the dislocated constituent is usually a form of the copula be and a whword (e.g. what, who, where). The bracketed (embedded) sentence in (6a) is a constituent, and thus it can be displaced, as the preposing structure in (6b) shows: (6) a. The janitor said [S that he is on vacation this week]. b. [S That he is on vacation this week] is what the janitor said. In Topicalization, no material is added to the dislocated constituent. Topicalization is fairly restricted in

236

Relevance of Constituency Constituency has been a powerful test to evaluate and select among different approaches to phrase structure and grammatical theory. For example, it empirically supported the adoption of X’-Theory in the 1980s, which claimed that the major sentence constituents in turn have internal structure. For instance, read magazines in (1) can be shown in (9) to behave as a constituent, excluding in the morning. (9) [Read magazines] was what the janitor did in the morning. This may be taken as evidence for a more finegrained phrase structure with an additional level between V(erb) and VP, called V’ (hence X’-Theory): (10) [VP [V’ read [NP magazines] ] [PP in the morning]]] References Bunt, Harry, and Arthur van Horck. 1996. Discontinuous constituency. Berlin: Mouton de Gruyter. Carnie, Andrew 2002. Syntax: a generative introduction. Oxford: Blackwell. Chametzky, Robert. 2000. Phrase structure: from GB to minimalism. Malden, MA: Blackwell. Chomsky, Noam 1981. Lectures on government and binding. Dordrecht: Foris.

CONTEXT Hawkins, John A. 1994. A performance theory of order and constituency. Cambridge: Cambridge University Press. Huck, Geoffrey J., and Almerindo E. Ojeda. 1987. Syntax and semantics, Vol. 20: discontinuous constituency. Orlando: Academic Press. Kayne, Richard. 1984. Connectedness and binary branching. Dordrecht: Foris. Lasnik, Howard. 1999. Minimalist analysis. Malden, MA: Blackwell.

Postal, Paul. 1964. Constituency structure: a study of contemporary models of syntactic description. International Journal of American Linguistics 30 (1, part 3). Siewierska, Anna. (ed.) 1998. Constituent order in the languages of Europe. Berlin: Mouton de Gruyter.

ACRISIO PIRES See also Phrase Structure; Pro-forms; Syntactic Category

Context Contexts are texts’ ‘nontexts’ and thus relational phenomena. They are caused by figure/ground mechanisms where texts are focused figures and their contexts function as their (back)grounds. Contexts are accordingly always changing. Particular academic fields prioritize particular kinds of researched objects, and accordingly particular kinds of contexts. An overall workable definition of context consequently may not be found, since it would be too general to cover the diverse and specific needs of subdisciplines. This tension between specific and general is captured in the aphorism Meaning is context bound, but context is boundless (Culler 1981:24). Nevertheless, contexts as phenomena have become increasingly important within linguistics, and in general. Everyday configurations and common-sense understandings, reflected in different dictionary definitions, are commonly twofold. Contexts are seen as textual elements embedding particular utterances, and as circumstances surrounding particular situational events. Yet, new insights into how language functions challenge these definitions. First, the notion of text is expanding, incorporating even the meanings associated with pictures, films, and bodies. Thus, an extended concept of text destabilizes a seemingly safe ground for a definition of context. Second, contexts are seen as interactive and not just as passive ‘embedding’ phenomena, a position that triggers new questions. When is a context? To report observed situations implies both decontextualizing and recontextualizing. Through decontextualization, focused events are separated from their original environment, and hence their original contexts are destroyed. In order to reconstruct meaning for receivers of reports that are separated from the original context, a text needs to be sufficiently (re)contextualized. However, the report now consists of elements of both the ‘original’ text and elements of its

new context. Hence, texts, traveling through time, space, and culture, will always occur in new contexts. The when question thus leads to a where question. A simple answer to Where is context? is: internal and/or external. An internal perception implies seeing contexts as general premises for interpreting utterances, while an external perception perceives the material world as constitutive for contexts. However, both views tend to be static. A dynamic view implies that contexts are not just around texts as pre-fixed situations, but are dynamically woven into utterances while uttering. Hence, there are contextual elements or hints within most texts, which assist the interpretation of the intended meaning. The merging of given (G) and new (N) in texts is regulated by such hints: Once upon a time there was a boy [boy N]. The boy [boy G] had a nut [nut N]. The nut [nut G] had a little hole [little hole N], etc. However, the opening phrase, Once upon a time..., only seems to imply a fairy tale. Rather, it is an example of just another context, namely the genre and context of an entry in the genre and context of an encyclopedia. The outcome is an inclusive embeddedness characteristic for contexts as phenomena [encyclopedia [entry [example [fairy tale hint]]]]. Hence, utterances are doubly contextual, relying on existing contexts for their production and interpretation and being, in their own right, events that shape new contexts for following actions. How ‘given’ and ‘new’ are combined depends on the kinds of communication, such as discourses, registers, or genres. Concrete texts and utterances focus on the micro level, while discourses, registers, and genres are active as contexts at a macro level. They can create or are presuppositions for contexts of situations. Some text theories try to specify contextual constituents more systematically, as M.A.K. Halliday and

237

CONTEXT his followers (e.g. C.M.I.M. Matthiessen and Jim Martin) do in Systemic Functional Grammar. This approach is contextual and sees language principally as meaning-potential for combining text and context and suggests that language has three interactive main functions: textual, ideational, and interpersonal. Similarly, Jürgen Habermas (1984) sees each person’s communicative context as a ‘lifeworld’ consisting of a triad of three basic interactive elements: inner nature (self), outer nature (world), and others (society). Although he is not arguing that lifeworld is context, his triad can be linked to Halliday’s so that the textual function is respectively related to communicators, the ideational to world, and the interpersonal to society, so that all elements define each other mutually. Following Halliday and Habermas and others on this matter, a triadic understanding will imply that both language (or utterances and texts) and context (genres and discourses) will be triadic. They are all interrelated and constitute a systemic contextual lifeworld. Consequently, contexts are always and simultaneously embodied, world-related, and societal. Embodied contexts, however, will in turn function as discursive resources for uttering and as such distribute power asymmetrically, even in a context that is seemingly equal for all the participants. Thus, classroom episodes with students and teacher, medical consultations between patients and doctor, command situations between soldiers and officer, will represent not only different embodied communicational resources, but different contextual power relationships. If contexts are systemic and interrelated, academic fields have to find their place in such a system, by developing awareness of similarities and differences vis-à-vis other fields. Subdisciplines within linguistics are likely to focus on textually established contexts. The nearest contexts for sounds are words, the nearest contexts for words are sentences, etc. Fields such as applied linguistics and anthropological linguistics, for instance, will need to take the larger, more general contextual aspects, such as genres and discourses, into consideration as relevant for the (re)production of meaning.

Context can be seen as a relational concept in transition and hence open to reconceptualization. To rethink something means to recontextualize it. Thus, reconceptualization is recontextualization. References Bakhtin, Mikhail. 1986. Speech genres and other late essays. Austin: University of Texas, Press. Bateson, Gregory. 1972. Steps to an ecology of mind. New York: Ballantine Books. Bourdieu, Pierre. 1989. Outline of a theory of practice. Cambridge: Cambridge University Press. Culler, Jonathan. 1981. The pursuit of signs: semiotics, literature, deconstruction. Ithaca: Cornell University Press. Duranti, Alessandro, and Charles Goodwin (eds.) 1992. Rethinking context. Language as an interactive phenomenon. Cambridge (UK): Cambridge University Press. Erickson, Fredric, and Jeffrey Schultz. 1981. When is a context? Ethnography and language in educational settings, ed. by Judith Green and C. Wallat. Norwood, NJ: Ablex. Fairclough, Norman. 1992. Discourse and social change. Cambridge (UK): Polity Press. Foucault, Michel. 1972. The archeology of knowledge. London: Tavistock. Goodwin, Charles, and Alessandro Duranti. 1992. Rethinking context: an introduction. Rethinking context. Language as an interactive phenomenon, ed. by Alessandro Duranti and Charles Goodwin. Cambridge (UK): Cambridge University Press. Habermas, Jürgen. 1984. The theory of communicative action, Vol 1. Boston: Beacon. Halliday, M.A.K. 1978. Language as social semiotic. London: Arnold. Halliday, M.A.K. 1994. An introduction to functional grammar, 2nd edition. London: Arnold. Malinowski, Bronislav. 1935. Coral garden and their magic, 2 vols. London: Allen & Unwin. Martin, Jim. 1997. Analysing genre: functional parameters. Genre and institutions. Social processes in the workplace and school, ed. by Frances Christie and Jim Martin. London and Washington: Cassell. Matthiessen, C.M.I.M. 1993. Register in the round: diversity in a unified theory of register analysis. Register analysis: theory and practice, ed. by Mohsen Ghadessy. London: Pinter.

SIGMUND ONGSTAD See also Halliday, Michael Alexander Kirkwood

Conversation Analysis Conversation analysis (CA) involves the study of spoken interactions. In CA, naturally occurring talk is recorded, transcribed, and then analyzed. The term

238

conversation analysis, however, may be misleading, as a range of spoken interactions are studied under the heading ‘conversation’, from casual conversation to

CONVERSATION ANALYSIS more formal talk such as courtroom interactions or medical consultations. As a result, CA is increasingly referred to as the study of ‘talk-in-interaction’, although this term has also been criticized for ignoring nonverbal issues such as pauses, gestures, or gaze. Traditionally associated with sociologist and others working in the ethnomethodological tradition, CA is used in different ways by different scholars. Scholars from a variety of fields, such as linguists, anthropologists, and communication specialists, also analyze spoken interactions, using a range of methods, and based on varying assumptions. Conversation analysis began in the 1960s with the work of American sociologists Harvey Sacks, Emanuel Schegloff, Gail Jefferson, and others, influenced by Erving Goffman’s focus on everyday interactions and by Harold Garfinkel, the founder of ethnomethodology. Ethnomethodology is concerned with uncovering participants’ (as vs. researchers’) perspectives on the activities in which they are involved. Ethnomethodologists seek to understand how people produce and understand daily activities, including conversation. Within this tradition, CA sees talk as social action, and attempts to uncover the structural properties of talk in terms of the sequential development of talk. CA provides a substantive body of evidence indicating that talk is rule-oriented. One area of investigation is turntaking. CA revealed that speakers tend to obey a rule of one-speaker-at-a-time. Several features of talk are related to this preference, including when speakers chose to begin their turns. Another finding is the adjacency pair, a two-turn sequence of adjacent utterances by successive speakers. The first utterance sets up an expectation for the second. An example is greeting– greeting. When a speaker greets another speaker, there is an expectation that the next utterance will be a return greeting. Other areas of investigation in CA have included the concept of repair, or what happens when a speaker misspeaks or is not understood, laughter, and storytelling. A number of assumptions underlie this methodology used within CA, the ideas that data should consist of multiple instances of naturally occurring talk, not constructed examples, that talk should be transcribed in such a manner as to attempt to capture the sense of the sequential development of the talk, of how the interaction unfolded, and that no item is too small to be investigated. As a result, transcriptions include such items as overlapping speech of multiple speakers talking at once, pauses, laughter, and changes in volume. In addition, CA tends to dismiss assumed categories of sociocultural context such as status or gender, claiming that the relevance of these categories must be demonstrated within the data analyzed.

The term ‘context’ is used to refer to the textual environment, rather than the sociopolitical environment. Talk can be both context bound, in that we understand utterances because of their sequential placement within a conversation, and context free. Conversationalists seek to find meaning, interpreting utterances in relation to prior utterances or knowledge, and in this sense, all utterances are context bound. Conversation is context free in that there are rules of conversation that are general and are not constrained by changes in such things as the participants or setting: participants in talk still take turns, repair misunderstandings, and the like. Context is also both retrospective and prospective. What is said in a given moment emerges in part because of what was said in immediately prior utterances, and it also influences what will be said next. In this manner, participants mutually construct talk. CA within the ethnomethodological tradition has a number of limitations. Much of the work in the field has been carried out in English. Cross-cultural studies have not always concurred with its findings. CA has also been criticized for ignoring the larger influences of sociocultural context, and for not acknowledging the benefits of quantitative analyses of spoken interactions. Scholars from a variety of fields have undertaken analyses of spoken interactions, often addressing the limitations mentioned above. Dell Hymes’s ethnography of speaking, also known as the ethnography of communication, is one such area that explicitly examines larger sociocultural contexts such as participant roles or setting. One goal of this approach is cross-cultural comparisons of interactions. Work on speech acts focuses on talk as performing social actions such as complimenting or apologizing. Analysts in the Gricean tradition also tend to use constructed examples of talk in order to determine the cooperative principles of conversation. Interactional sociolinguistics provides another avenue into the analysis of spoken interactions, as does variation analysis. Much of this work examines how people vary their talk as a result of changes in sociocultural contexts. Conversation analysis has contributed to linguistics, in part by making talk itself a focus of serious academic investigation. We now understand a great deal about the structure of talk, and therefore how deviations from our expectations of how talk should proceed lead us to interpret—rightly or wrongly—our speech partners’ intentions. Fail to answer a greeting with a greeting, for example, and you may well be judged as rude in certain speech communities. CA is also combined with other linguistic disciplines and approaches, such as second-language acquisition and syntactic analysis, investigating how syntactic choices

239

CONVERSATION ANALYSIS are affected by the emerging nature of discourse. As CA is incorporated into wider circles of linguistic research, a more complete understanding of the various components and functions of language is possible. References Drew, P., and J. Heritage (eds.) 1992. Talk at work: interaction in institutional settings. Cambridge: Cambridge University Press. Eggins, S., and D. Slade. 1997. Analysing casual conversation. London: Cassell. Hutchby, I., and R. Wooffitt. 1998. Conversation analysis: principles, practices, and applications. Oxford: Polity Press.

Markee, N. 2000. Conversation analysis. Malwah, NJ: Lawrence Erlbaum. Psathas, G. 1995. Conversation analysis: the study of talk-ininteraction. Thousand Oaks, CA: Sage. Sacks, H., E. Schegloff, and G. Jefferson. 1974. A simplest systematics for the organization of turn-taking in conversation. Language 50. Tannen, D. 1984. Conversational style: analyzing talk among friends. Norwood, NJ: Ablex. ten Have, P. 1999. Doing conversation analysis: a practical guide. London: Sage.

SUZANNE SCOTT See also Context; Discourse Analysis; Hymes, Dell H.; Sociolinguistics; Speech Acts

Coordination A coordination joins two sentence elements, called conjuncts. In a coordinate structure like cats and dogs, the conjunction coordinates the conjunct cats with the conjunct dogs. In many languages, conjunctions like and or or can conjoin words or phrases of virtually every category, under the condition that the categories being conjoined are of the same sort. It might appear as if coordination was a relatively simple phenomenon. However, coordination is notoriously difficult for linguistic theory to define. Although a wide variety of structures can be conjoined, not all coordinations are acceptable. One of the first generalizations regarding coordination is Ross’s Coordinate Structure Constraint (1967). This constraint states that coordination does not allow for asymmetrical constructions. For example, the sentence This is the man whom Kim likes and Sandy hates Pat is unacceptable, because only the first conjunct is relativized. The sentence This is the man whom Kim likes and Sandy hates is acceptable, because both conjuncts are relativized. The Coordinate Structure Constraint might be explained by the requirement that the conjuncts in a coordinate construction must be of the same ‘sort’. This requirement is sometimes referred to as the Law of Coordination of Likes. Linguists are uncertain as to the relationship between ‘sort’ and syntactic category. The sentence Pat is stupid and a liar shows that being of the same syntactic category is too strong a requirement for conjuncts in a coordinate construction, since an adjective phrase (stupid) can be conjoined with a noun phrase (a liar). It is therefore unclear what it means for two conjuncts to be of the same sort.

240

Linguists are further concerned with which material is allowed as a conjunct in a coordinate construction. The second example showed conjoined sentences, but coordination is also possible for noun phrases as in the apples and the pears, verb phrases like run fast or jump high and adjectival phrases such as rich and very famous, etc. Both sentences and phrases intuitively form meaningful units within a sentence, called constituents; however, not all sentence elements can be constituents. Subject and verb do not form a constituent in some frameworks of generative grammar. However, they can occur together as a conjunct in the sentence Kim bought, and Sandy sold, three paintings yesterday. The possibility of this ‘nonconstituent coordination’ has led a number of linguists to relax the notion of constituency. In categorial grammar, for example, subject and verb can form a constituent. Coordination phenomena can therefore provide a testing ground for even basic theoretical notions such as constituency. Another important question concerns the way coordination of phrases is interpreted, as phrasal coordination seems strongly related to sentential coordination. The sentence Kim ran and jumped, in which two verb phrases are conjoined, has the same interpretation as the coordination Kim ran and Kim jumped, in which two sentences are conjoined. The dominant approach in generative syntactic theories indicates that phrasal coordination can be derived from sentential coordination by means of reduction rules. This approach states that Kim ran and jumped is the result of a reduction rule having deleted the subject of the second conjunct, Kim. Much syntactic research focuses on formulating appropriate reduction rules. This has turned out to be

COPTIC EGYPTIAN quite difficult due to the potential for phrasal constructions such as Kim and Sandy are similar, which lack a sentential source. The work of Richard Montague in the early 1970s ushered in the method of deriving the interpretation of conjoined phrases directly from their surface form. Semantic explanations of phrasal coordination differ from strictly syntactic ones. In semantic analysis, conjuncts are interpreted as functions that require certain semantic arguments to make a sentence. The semantic type of ran requires a subject to yield an interpretable sentence. Because jumped is of the same semantic type as ran, ran and jumped can be conjoined according to the Law of Coordination of Likes. If this conjoined verb phrase is applied to the subject Kim, the resulting interpretation will be that Kim ran and Kim jumped. Under this semantic approach, the interpretation of phrasal coordination is related to, but not derived from, the interpretation of sentential coordination. Further, a problem particular to syntactic analysis concerns how to formally represent coordination, as the conjunction structure appears to contradict current models of sentence construction. The term ‘coordination’ implies that the conjuncts are located at the same structural level. However, this view is incompatible with the assumption in generative syntax that syntactic structures are formed by repeatedly unifying two elements at a time. The argument that a conjunction unifies its two conjuncts at the same level entails the conclusion that syntactic structures can unify three elements. Some linguists attempt to remedy this apparent contradiction by suggesting a subordinating structure for coordination, in which the conjunction combines with one conjunct first and the resulting constituent then combines with the other conjunct. Others suggest the hypothesis that both conjuncts basically stand in

exactly the same relationship to the conjunction, and that the two conjuncts essentially introduce a third structural dimension. Although various solutions have been proposed, there is not yet a satisfactory explanation for all of the problems discussed here. Coordination occurs both in phrase structure and sentence structure, but the relationship between the two remains unclear. The role of coordination in sentence construction is also largely undetermined. Coordination therefore remains a phenomenon that is difficult to explain for any formal linguistic theory. References Goodall, Grant. 1987. Parallel structures in syntax. Cambridge: Cambridge University Press. Hartmann, Katharina. 2000. Right node raising and gapping: interface conditions on prosodic deletion. Philadelphia/ Amsterdam: John Benjamins. Johannessen, Janne Bondi. 1998. Coordination, New York: Oxford University Press. Montague, Richard. 1973. The proper treatment of quantification in ordinary English. Approaches to natural language: Proceedings of the 1970 Stanford Workshop on Grammar and Semantics, ed. by Jaakko Hintikka, J.M.E. Moravcsik and Patrick Suppes. Dordrecht: Reidel. Oirsouw, Robert R. van. 1987. The syntax of coordination. London: Croom Helm. Ross, John Robert. 1967. Constraints on variables in syntax. Ph.D. Dissertation, MIT, reprinted as Infinite syntax; Norwood, NJ: Ablex Publishing Corporation, 1986. Steedman, Mark. 2000. The syntactic process. Cambridge, MA: MIT Press. Winter, Yoad. 1998. Flexible Boolean semantics: coordination, plurality and scope in natural language. Ph.D. Dissertation, Utrecht University.

PETRA HENDRIKS See also Constituency Test; Montague, Richard

Coptic Egyptian In the course of its productive language history, which spans a period of over 4,000 years, Ancient Egyptian went through several developmental stages. Its latest stage is Coptic Egyptian (not to be confused with Egyptian Arabic), which is the vernacular of lateantique and medieval Christian Egypt (fourth to fourteenth centuries AD). The modern term Coptic is derived from Arabic q_u bt.¯ι, itself a corruption of the Greek word (ai)gypt(ios) ‘Egyptian’.

The Copto-Greek alphabet Coptic, like many other ancient languages of literature, has been passed down to us through large corpora of texts. These texts were written in a highly standardized notational system of alphabetic signs representing the different sounds of the Coptic language. The origin of the Coptic writing system lies in occasional Greek transcriptions of native words in Egyptian texts of the Hellenistic and Roman periods. In the first three

241

COPTIC EGYPTIAN centuries AD, the use of such transcriptions became increasingly common and entire corpora of texts with a predominantly magical character were written in a Greek-derived alphabet. The Christianization of the country in the fourth century AD constituted a turning point: the abandonment of the pagan literary tradition and culture manifested itself in the replacement of hieroglyphic writing and Demotic, its cursive variant, by Greek script. In its present form, the Copto-Greek alphabet consists of 32 letters, 24 of which are taken from Greek and eight from Demotic writing. With the exception of the monosyllabic grapheme ti, the Demotic-based signs represent phonemes that were absent in Greek, but which are part of the native Coptic Egyptian sound system.

Bilingualism and linguistic borrowing The emergence of Coptic was the result of intensive language contact in a bilingual (Egyptian–Greek) speech community. Greek was not only the language of the literate elite but also the language of the Holy Scriptures and the new religion, and hence a language of great cultural importance. The impact of this prestigious language on the native vernacular was all-pervasive. Although no clear statistics are available at present, it is estimated that approximately 40% of the Coptic vocabulary consists of Greek loanwords. The transfer of Greek lexical material into the Coptic vocabulary was not only restricted to content words (nouns, verbs, adjectives) with a clear link to Hellenistic and Christian culture (e.g. te-psikhE ‘the soul’, klEronomei8 ‘to inherit’, hagios ‘holy’), but also involved a variety of Greek function words (i.e. grammatical words with no descriptive-lexical content), such as sentence conjunctions, discourse markers, and some types of adverbs and prepositions (e.g. hOste ‘such that’, E ‘or’, oude ‘and not’, alla ‘but’). Despite the massive influx of Greek items, grammatically organized words (pronouns, articles, numbers, and the like) are all drawn from the native stock. To fit into Coptic phrase structure, Greek loanwords underwent minor changes in the course of borrowing.

Dialect variation Coptic Egyptian is actually a dialect cluster, consisting of at least six regional varieties, two of which gained supraregional importance: Sahidic, the language of the whole Nile valley above the Delta, and Bohairic, the language of the Nile Delta. Sahidic and Bohairic Coptic differ significantly from one another in a number of grammatical features (sound system, verbal morphology, syntax) and the amount of lexical and grammatical borrowing from Greek. Both language

242

varieties also differ with respect to the time depth of their attestation. Prior to the Arabic conquest in AD 641, Sahidic was the predominant literary dialect of Coptic. Its supremacy became challenged by Bohairic Coptic from the ninth century onward. By the eleventh century, Bohairic had replaced Sahidic as the official church language and had become the sole representative of Coptic Egyptian, which survived as the liturgical language of the present-day Coptic Orthodox Church. The language material of the following typological sketch is exclusively drawn from Sahidic Coptic, the main reference dialect.

Phonology The sound system of Sahidic Coptic consists of 20 consonants (p, t, k, d, g, kj, f, s, S, h, B, z, tS, m, n, l, r, w, y, /) and eight vowels (i, e, E, u, o, O, a, ´). Its most striking property is the absence of voiced consonants (i.e. consonants produced with a vibration of the vocal cords) in the class of stops (consonants produced by making a closure at the lips), fricatives (consonants where the airflow is constricted to form a turbulence but is not interrupted), and affricatives (sounds produced by an initial closure that is released gradually, making it sound like a fricative), the main exception being the voiced fricative /B/. It is worthwhile pointing out that the voiced stops /d/ and /g/ as well the voiced fricative /z/ have a special status as loan phonemes that are by and large restricted to Greek borrowings. The glottal stop /// (a sound produced by a complete but brief contraction of the vocal cords) has no alphabetic sign of its own. Yet, its presence in the Sahidic sound system can be deduced from a sequence of two identical vowel letters that indicates ‘broken’ vowels, i.e. long vowels that are pronounced with a brief interruption (e.g. maatSe /ma/atSe/ ‘ear’, mEESe /mE/ESe/ ‘crowd’). Coptic has pairs of vowels: /e/ vs. /E/ and /o/ vs. /O/. The schwa /´/ (a colorless, unstressed vowel) is graphically expressed not by a letter, but rather a special diacritic, the so-called supralinear stroke (a vertical line above a letter), e.g. n@ /´n/ ‘of’. Stress (indicated as ") is lodged only on those syllables that either contain a long vowel (…) (e.g. noute /nu…".te/ ‘god’) or a vowel–consonant sequence (e.g. anaS /a.na"S/ ‘oath’). Moreover, there is at most one main stress per word, regardless of its length.

Noun morphology Unlike Pre-Coptic Egyptian, Coptic has only a few word formation processes for the makeup of nouns. Thus, there are only a few lexical items where number (singular vs. plural) and gender (masculine vs. feminine) are marked on the noun itself (e.g. son ‘brother’,

COPTIC EGYPTIAN

sOne ‘sister’, sneu8 ‘brothers, fellow monks’). In the vast majority of cases, gender and number distinctions are expressed on the definite article, which is placed in front of the noun it modifies (e.g. p-rOme ‘the (sing. masc.) man’, te-shime ‘the (sing. fem.) woman’, neshime ‘the (pl.) women’). The Sahidic determiner system offers a three-way contrast between a definite, an indefinite, and a zero article (e.g. p-rOme ‘the (sing. masc.) man’; u-rOme ‘a man’, hen-rOme ‘men’; rOme ‘man’). Besides the indefinite and definite articles, there are demonstratives articles, so-called because they indicate closeness or distance to the location of the speaker: pei8-rOme ‘this (sing. masc.) man’ vs. ‘that man’ (lit. the man who is (over) there). Both demonstrative articles have corresponding demonstrative pronouns, which have a word-like status: pai8 ‘this one’ vs. p-et-´mmau8 ‘that one’. Reference to grammatical person (the speaker, the addressee, and a third party) is expressed by two sets of pronouns: independent pronouns and bound pronouns. Independent pronouns have a word-like status, while bound pronouns form an inseparable unit with the sentence element they modify. Different forms exist for pronominal prefixes that precede and pronominal suffixes that follow their host. Table 1 presents the pronominal paradigm of Sahidic Coptic (-º indicates a null morpheme, i.e. a grammatical element that is not pronounced). Independent pronouns have an emphatic meaning and place the entity or object they refer to at the center of attention. Pronominal prefixes and suffixes, on the other hand, lack such a highlighting function. While independent pronouns are restricted to peripheral syntactic positions (e.g. ntos de a-s-onk-s ehrai8 ‘(as for) her (ntos), she leaped up herself’ (Eudoxia 50:17)), bound pronouns are compatible with all nominal positions of the clause (e.g. a-f-ent-s ‘he brought it’, nE-t´n ‘for you’).

Verb morphology The basic principle of Coptic verb formation, which it shares with Semitic languages, is that of ‘root-and-patTABLE 1 Pronominal System Bound Pronouns

1st sing. 2nd sing. masc. 2nd sing. fem. 3rd sing. masc. 3rd sing. fem. 1st pl. 2nd pl. 3rd pl.

Prefixes tikte-, terfstentet´nse-

Suffixes -i, -t -k -e, -º -f -s -n -t´n, -tEu8t´n -se, -su…, -u…

Independent Pronouns anok ´ntok ´nto ´ntof ´ntos anon ´ntOt´n ´ntou 8

tern’. As the terminology suggests, verbal stems are derived from relatively abstract form-meaning representations (roots) by the superimposition of particular consonant–vowel patterns with a basic meaning. For example, a verbal root like /m-s / ‘BEAR , GENERATE’ may surface in at least four word formation patterns that are distinguished from one another by means of vowel change (Ablaut) and syllable structure. These are the absolute state mise ‘to give birth’, the nominal state mes, and the corresponding pronominal state mest ‘to deliver’, which combine with a nominal and pronominal object, respectively, and the Stative mose ‘to be bred’. The absolute state and the construct state (i.e. the nominal and the pronominal states together) describe events that change over time, while Statives refer to states or conditions that last for some time. The absolute state and the construct states, in turn, differ syntactically in the way they express the direct object relation. In the absolute state, the direct object is encoded as a prepositional phrase (e.g. mise n-uSe/ere n-shime ‘to give birth to a girl’ (Miracles of Apa Mêna 10b:33–4)). In the construct state, however, direct objecthood is indicated by the juxtaposition of the verb and the nominal or pronominal object (e.g. mes p-Sr-howt ‘to deliver the male child’ (Apocalypse 12:13) vs. mest-f ‘to deliver it’ (Miracles of Apa Mêna 10b:26)).

Conjugation system Coptic has more than 20 different verb conjugations (i.e. the pattern in which verb forms can appear) for the expression of tense (i.e. the location of events in time), aspect (i.e. the ongoing state, completion, or multiple occurrence of events), and mood (i.e. the commitment of the speaker toward the truth of the reported events). At the foundation of this richness of meaning distinctions is the subdivision of the four absolute tenses (the Present, the Habitual, the Future, and the Perfect) into two conjugation classes, traditionally known as first and second tenses. The second tenses are morphologically derived from the ‘basic’ first tenses by adding the relative markers e- or ´nt- in front of the verbal cluster (Table 2). First and second have exactly the same temporal and aspectual interpretation, but differ from one another with respect to their syntactic distribution. First TABLE 2

Absolute Tense System

First Tenses Present Habitual Future Perfect

f-sOt´m Sa-f-sOt´m f-na-sOt´m a-f-sOt´m

Second Tenses e-f-sOt´m e-Sa-f-sOt´m e-f-na-sOt´m ´nt-a-f-sOt´m

‘he listens’ ‘he (usually) listens’ ‘he is going to listen’ ‘he listened’

243

COPTIC EGYPTIAN tenses appear in pragmatically neutral declarative clauses (e.g. ei8s hE/Ete anon ne.k-h´mhal t´n-sOt´m ‘Look, we, your servants, are listening!’ (Eudoxia 62:03)). The derived second tenses are used in the context of relative clauses (e.g. u-hOB [ere p-nu…te moste ´mmo-f ]‘a thing [that God hates (it)]’ (Acts of Andrew & Paul 202:126–7), constituent questions (e.g. e-tet´n-Sine ´nsa nim‘Whom are you (woman) looking for?’ (John 18:4), as well as a range of declarative focus contexts (e.g. awo p-woei8n e-f-r woei8n h´m p´-kake ‘and (as for) the LIGHT, it is shining in the DARKNESS’ (John 1:5)). Apart from absolute tenses, Coptic has several syntactically dependent verbal tenses and moods. Relative tenses locate some event with respect to another event and express three types of temporal relations: precedence, subsequence, or simultaneity. In doing so, relative tenses often indicate a logical or causal connection between two events. In negative tenses, negative meaning (the equivalent of English not) and a particular time value are fused together in a single indivisible unit. Take for instance the Negative Future ´nne-fsOt´m ‘he shall not hear’, where the conjugation ´nne combines future tense with negative meaning, which are expressed by two separate elements (viz. shall and not) in the English equivalent. Moreover, Coptic has two different moods for the imperative (e.g. sOt´m ‘Listen!’), which expresses commands, and the optative, which expresses wishes (e.g. mare-f- sOt´m ‘may he listen’). See Table 3 for further illustration. The clause-initial or medial position of tenseaspect-mood markers, their morphological independence from the verb, as well as their agreement behavior (variant forms for nominal and pronominal subjects, e.g. (Habitual) Sare p-rome sot´m ‘the man listens (usually)’ vs. Sa-f-sot´m ‘he listens (usually)’

provides prima facie evidence for their categorical status as auxiliary verbs. Due to their semantic erosion, these auxiliaries have a fully grammaticalized meaning and function, which is typical of free functional morphemes.

Syntax The basic word order, from which other word-order patterns are derived, is subject–verb–object (SVO). Word-order alternations are, however, extremely common and motivated by pragmatic considerations. The topic status (i.e. the presupposed, familiar or known character) of a noun or pronoun is generally indicated by placing it into the left periphery of the clause, its grammatical function being resumed by a pronoun with identical person, number, and gender specification (e.g. p-aNgelos de m-p-tSoei8s a-fwonh-f e-p-arkhiepiskopos ‘the angel of the Lord, he revealed himself to the archbishop’ (Mêna, 4b:6–9)). Verb–subject order has a ‘presentative’ meaning and is used for the introduction of new discourse participants. Notice that the nominal subject that is removed from its sentence-initial position is supplied with the particle nkji (e.g. f-nEu nkji uaNgelos nte p-nu…te ‘(there) comes an angel of God’ (Budge, Coptic Martyrdoms 214:22)). To highlight the subject or object, a special construction type is used: the cleft sentence. It is called cleft sentence because it consists of two parts: a sentence-initial noun or pronoun and a relative clause (given in brackets) (e.g. awo m-pe.u8-kjBoi8 an pe [ ´nt-a-ftutSo-u88 ] ‘and (it) is not their arm that has saved them’(Psalm 43:4)). More research is needed to clarify the relation between pragmatic prominence and sentence form in Coptic Egyptian.

TABLE 3 Relative Tenses and Basic Moods Relative Tenses Temporal Terminative Conjunctive Finalis

Negative Tenses ´ntere-f-sOt´m ‘after he had listened’ Sant´-f-sOt´m ‘until he listens’ n´-f-sOt´m ‘and he listens’ tare-f-sOt´m ‘so that he will listen’

Negative Habitual Negative Future Negative Perfect Unexpected Negative Perfective

me-f-sOt´m ‘he does not listen’ ´nne-f-sOt´m ‘he shall not listen’ ´mpe-f-sOt´m ‘he did not listen’ ´mpate-f-sOt´m ‘he has not yet listened’

Moods Conditional

e-f-San-sot´m ‘if he listens’

Negative Conditional

Optative

mare-f-sOt´m ‘May he listen!’ sOt´m ‘Listen!’

Prohibitive

Imperative

244

e-f-San-t´m-sot´m e-f-t´m-sot´m ‘if he does not listen’ mp´r-sOt´m ‘Do not listen!’

CORPUS LINGUISTICS References Layton, Bentley. 2000. A Coptic grammar with Chrestomathy and Glossary. Porta Linguarum Orientalium N.S. 20. Wiesbaden: Harrassowitz Verlag. Loprieno, Antonio. 1995. Ancient Egyptian—a linguistic introduction. Cambridge and New York: Cambridge University Press. Polotsky, Hans Jakob. 1960. The Coptic conjugation system. Orientalia 27. 392–422. Reintges, Chris H. 2001. Code-mixing strategies in Coptic Egyptian. Lingua Aegyptia 9. 193–237.

Reintges, Chris H. 2003. Coptic Egyptian (Sahidic Dialect)—a learner’s grammar. Cologne: Rüdiger Köppe Verlag. Reintges, Chris H. 2004. Coptic Egyptian (Sahidic Dialect)—a learner’s grammar. Cologne: Rüdiger Köppe Verlag. Till, Walter C. 1961. Koptische Dialektgrammatik: mit Lesestücken und Wörterbuch. München: Verlag C.H. Beck.

CHRIS H. REINTGES See also Ancient Egyptian; Auxiliaries; Function Words

Corpus Linguistics During the last three decades of the twentieth century, computer technology has made it possible to conduct extensive and complex research on specific linguistic features — either lexical items or grammatical structures—and their systematic associations with other linguistic and nonlinguistic features. These nonlinguistic features include registers or specific varieties of language (e.g. religious, political, scientific) and dialects, which are regional or social varieties of a language. This new type of research is part of corpus linguistics, which is the empirical study of language using computer techniques and software to analyze large, carefully selected and compiled databases of naturally occurring language (Conrad 2000). Corpus linguistics represents a departure from the dominant mentalist approach to linguistic research, which emphasizes the processes taking place in the human mind (e.g. a Universal Grammar-centered approach to second language acquisition studies). This type of research is characterized by the empirical analyses of actual patterns of language use in large and principled databases or corpora. Corpus linguistics utilizes both quantitative and qualitative analytical techniques and relies on computers to perform complex analyses. (See Biber, Conard & Reppen (1998) and Kennedy (1998) for a detailed account of corpusbased investigations of language structure and use.)

Corpus-Based Investigations of Language Use Although language structure has traditionally been studied using nonempirical methods and relying on the researcher’s intuitions, extensive corpus-based studies describing various aspects of language use were carried out in the 1980s and 1990s (Aarts 1991, Aijmer

and Altenberg (eds.) 1991, Biber 1995, Leech 1991, Sinclair 1987, 1991, Stubbs 1995, Svartvik 1990). As Biber et al. (1996:115) point out, ‘it is in the area of language use that corpus-based techniques have had the most impact’. These studies complement descriptive language structure investigations and previously neglected aspects of English grammar (Crystal 1979). Early studies in corpus linguistics focused on the occurrence of linguistic items (e.g. noun, verb, and adjective frequencies), but the development of more powerful techniques has enabled researchers to identify and analyze complex association patterns. Corpusbased research has shown that these linguistic association patterns generally fall into two major categories: lexical associations and grammatical associations. In the first case, the goal is to investigate how a linguistic feature is systematically associated with particular words. In the second case, the researcher investigates how a linguistic feature is systematically associated with grammatical features in the immediate context. With respect to lexical associations, a concordancer shows which words collocate with each target word in a corpus that is representative of a specific register or dialect. For example, in a corpus of working-class northern New Jersey cyber discourse, we find that the mental verb know, which appears 15 times in a small sample thread (7,000 total words), collocates primarily with personal pronouns I, you, they, and we. This usage reflects the informal nature of cyber exchanges, which characterizes casual face-to-face interaction among peers or friends. The results of empirical, large-scale corpus-based research projects (data from a 5.7-million-word sample from the Longman—Lancaster Corpus) have

245

CORPUS LINGUISTICS

KWIC Display (TACT) know (15) (219) That’s how it works and they know it. At least now they’re ¦ (227) Are you a total moron? Do you know the number of lawsuits ¦ (314) and hazardous 33 years ago. I know of no environmental| ¦ (329) against Mar...and Cap...You know the condition of the ¦ (372) DPW does for this town. I don’t know what your job is, but ¦ (372) responsibility at all because they know you couldn’t handle ¦ (403) Just Curious about how…We all know that the building is a ¦ (418) he supervised his dog. I don’t know if your post even ¦ (419) electrician for years, you don’t know too much about his ¦ (432) you wouldn’t want anyone to know? Would you? ¦ (460) You are there aren’t you? I know it’s not election ¦ (467) IMPLANT Does Maru… know where the D.P.W. ¦ (473) Hard Worker #2: How do you know that Caputo visits Kiss ¦ (475) “does Maru…or the mayor know where the DPW is?” ¦ (479) to above post: how do you know who visits nortons ¦

as-a-Second Language (ESL) learners will incorporate the results of corpus-based research. Conrad (2000) and Granger (2002) explain that corpus research complements innovations in grammar pedagogy by encouraging instructors to design and implement consciousness-raising activities for second and foreign language learners. (See Conrad (2000) for an overview of the applications of Corpus Linguistics in grammar teaching in the twentyfirst century, and Granger (1998a; 1998b; 2002) for a discussion of learner corpora compilation processes.)

Corpus Analysis Tools and Corpora Used in Corpus Linguistics Since corpus linguistics utilizes large and representative collections of natural texts, there are several types of tools that can be used to conduct research: commercially available packages or concordancing programs (e.g. LEXA, MonoConc, MicroConcord, TACT, WordSmith, WordCruncher) and computer programs developed by researchers for specific types of analyses. These latter types of programs are used to investigate complex grammatical constructions or association patterns, such as Biber etal.’s (1998) study involving the omission of that from that-clauses. There are at least 31 commercially and publicly available corpora of written and spoken texts. These databases contain millions of words and are divided primarily into three categories: written, spoken, and historical. These corpora are further subdivided into American English, British English, and texts of other varieties of English. Corpora also exist for languages other than English. (See Appendix in Biber et al. 1998: 282–7.)

Potential Limitations of Corpus Linguistics shown that where certain words are nearly synonymous in isolation, careful analysis reveals that they tend to be used with very different kinds of words. For example, Biber et al. (1998, chapter 2) demonstrate that the word big commonly co-occurs with toe, while large commonly co-occurs with number. Linguistic features are also systematically associated with grammatical features in the immediate context as shown by corpus-based research. For example, one of the factors differentiating that-clauses and toclauses is their lexical associations. ‘The verbs such as suggest, conclude, guess, and argue can control a that-clause but not a to-clause; the verbs begin, start, and try can control a to-clause but not a that-clause’ (Biber et al. 1998, chapter 3). There are many promising applications of corpusbased research that have to be explored. For example, it is hoped that new grammar teaching materials for English246

It is important to understand that meticulous qualitative analyses of single texts (e.g. historical, spoken, written, and learner-centered) are usually undertaken before embarking on corpus-based research. A careful microanalysis of linguistic features in written or spoken discourse helps us frame new research questions and hypotheses. Thus, corpus-based analysis should be seen as a complementary approach to the more traditional approaches that have often focused on language structure. The strength of corpus linguistics lies in its investigations of language use, which necessitate empirical analysis of large databases of authentic texts. There are basically four potential limitations of corpus-based research. The first one is the time-consuming nature of compiling a corpus and tagging parts of speech, errors, or other features. The second limitation concerns the nature of the corpus data collected. This type of research often involves collecting and storing large corpora that may need to be continuously updated.

COURTROOM DISCOURSE The third limitation is that relying on computers forces linguists to concentrate primarily on written rather than spoken language. Lastly, there is a potential limitation of a philosophical nature, namely that a corpus is a finite sample of an infinite population. This means that researchers are often extrapolating from what is found in a corpus to what is true of the language or language variety it is supposed to represent (Leech 1998). Thus, one needs to be cautious in drawing general inferences from the results of corpus-based analyses. Despite its potential limitations, almost any aspect of linguistics can be studied from a use perspective. Corpus linguistics provides a variety of tools and methods that make large-scale research on complex linguistic phenomena an extremely productive and challenging undertaking. References Aarts, Jan. 1991. Intuition-based and observation-based grammars. English corpus linguistics, ed. by Karin Aijmer and Bengt Altenberg, 44–62. London: Longman. Aijmer, Karin, and Bengt Altenberg (eds.) 1991. English corpus linguistics. London: Longman. Biber, Douglas. 1995. Dimensions of register variation: a crosslinguistic comparison. Cambridge: Cambridge University Press. Biber, Douglas, Susan Conrad, and Randi Reppen. 1996. Corpus-based investigations of language use. Annual Review of Applied Linguistics 16. 115–36. Biber, Douglas, Susan Conrad, and Randi Reppen. 1998. Corpus linguistics: investigating language structure and use. Cambridge: Cambridge University Press.

Conrad, Susan. 2000. Will corpus linguistics revolutionize grammar teaching in the 21st century? Tesol Quarterly 34. 548–60. Crystal, David. 1979. Neglected grammatical factors in conversational English. Studies in English linguistics for Randolph Quirk, ed. by S. Greenbaum, Geoffrey Leech and Jan Svartvik, 153–67. London: Longman. Granger, Sylviane (ed.) 1998a. Learner English on Computer. New York: Addison Wesley Longman Limited. Granger, Sylviane 1998b. The computer learner corpus: a versatile new source of data for SLA research. In Granger, 3–18. Granger, Sylviane, Joseph Hung, and Stephanie Petch-Tyson (eds.) 2002. Computer learner corpora, second language acquisition and foreign language teaching. Amsterdam, Philadelphia: John Benjamins Publishing Company. Kennedy, Graeme. 1999. An introduction to corpus linguistics. London, New York: Longman. Leech, Geoffrey. 1991. The state of the art in corpus linguistics. English corpus linguistics, ed. by Karin Aijmer and Bengt Altenberg, 8–29. London: Longman. Leech, Geoffrey. 1998. Preface. In Granger. Sinclair, J. (ed.) 1987. Looking up. An account of the cobuild project in lexical computing. London: Collins ELT. Sinclair, John. 1991. Corpus, concordance and collocation. Oxford: Oxford University Press. Stubbs, Michael. 1996. Collocations and semantic profiles: on the cause of the trouble with quantitative studies. Functions of Language 2. 23–55. Svartvik, Jan. 1990. The London-Lund corpus of spoken English: description and research. Lund, Sweden: Lund University Press. TACT Database is under construction (cyber discourse of working-class northern New Jersey towns). Contact Susana Sotillo at [email protected].

SUSANA M. SOTILLO See also Universal Grammar

Courtroom Discourse Studies in courtroom discourse often focus on the significance of certain kinds of interactions between specific linguistic features and their pragmatic or sociointeractional functions in the courtroom. Research carried out in North Carolina beginning in the 1970s, for instance, resulted in a series of publications suggesting broadly that lay witnesses are persuasive or not depending upon the extent to which they use “powerful” or “powerless” language. The validity of the study’s methodology has been questioned, but the publications generated by the study have clearly served to advance our understanding of the role of language in judicial process, particularly where the questioning of lay witnesses is concerned.

Other research has explored the questioning tactics used by attorneys during direct and cross-examination of witnesses. The use of open-ended wh-questions (who, what, where, when, how, etc.) tends to elicit narrative responses from witnesses, and therefore more information, whereas yes/no questions tend and are usually intended to limit the amount and type of information that can be offered. A number of studies examine such patterns, and it is usual for trial practice courses in law schools to include material on such topics, often recommending that wh-questions be used during direct examination and that yes/no questions be used during cross-examination. For a recent example of such a study, see Sandra Harris’ (2001) examination

247

COURTROOM DISCOURSE of the O.J. Simpson, Louise Woodward, and Oklahoma federal building bombing cases. It is usual for defense attorneys to omit mention of the agent in cases of sexual assault or rape; the tactic is thought to take the focus off the accused rapist. The grammar and prosody of reported speech can be used to impeach or cast doubt on a witness’s testimony. Janet Cotterill examined the O.J. Simpson trial and explored how metaphors can be a rich source of indirect messages, noting that several were used in defense attorney Johnny Cochran’s closing arguments. In a more general sense, courtroom discourse includes attention not only to adversarial questioning techniques, particularly as used in direct and crossexamination, but also to all spoken language used during the trial process. In order to look meaningfully at the full range of courtroom discourse, one must recognize that a number of speech genres are used during the trial process. Anglo-American lawyers like to talk about the need for a theory of the case. Once that theory is developed, all portions of trial process are theoretically dedicated to developing that theory and persuading the judge and/or jury that it is the correct one. In the process of moving through trial process, lawyers make use of opening statements, voir dire, direct examination, crossexamination, possible redirect and recross, and then closing statements. They also usually participate in the wording of jury instructions, and they make and respond to objections during testimony. For each of these steps, discourse strategies are developed. Based on an examination of a single case, Gail Stygall (1994) suggests that analysis of legal language needs to move away from mere descriptive analysis to an examination of how the maintenance of legal language serves institutional power and dominance. More specifically, John Gibbons (2003) suggests that the very real power and dominance problems in the courtroom arise from the legal genres used in courtroom discourse and the fact that they are addressed to two very different audiences, lay persons, often jurors, on the one hand, and of course judges and lawyers on the other hand. Gibbons calls this the twoaudience dilemma and cites jury instructions as a particularly difficult genre, because “[j]urists persist in administering these instructions to jurors because they have survived on appeal to a higher court... The view that jurors remember and understand any instructions given them and that therefore it is more important to focus on an instruction’s survivability on legal appeal than on its comprehensibility, is ill-informed” (English and Sales 1997. 383). Where jury instructions are concerned, it is an established fact that since the early 1970s, findings by many social scientists confirm that lay persons are frequently bewildered by the wording of jury instructions. The

248

exact syntactic and semantic bars to juror comprehension of instructions are now well documented by linguists and psycholinguists. These scholars have demonstrated that instructions can be made more comprehensible by simplifying sentence structure and by giving additional information about the meanings of abstract terms in both civil and criminal cases. We know a great deal about the problems faced by lay jurors because of the extensive amount of research that has been completed. But lay witnesses face similar problems. They contend with very sophisticated question-and-answer tactics on direct and cross-examination, usually with little or no preparation. And, of course, court interpreters and translators face the same problems, even as they deal with the challenge of working with more than one language. All phases of pretrial, trial, and posttrial (appeal) procedure involve judicial procedures in which lay persons must deal with legal language and discourse patterns, including legal genres, with which they are quite unfamiliar. Jury instructions have been examined in great detail; research on other patterns and genres remains to be done. It seems likely that much additional research will be reported over the next few years. Increasing numbers of linguists have been focusing on language and law issues, and there are now at least two academic organizations that regularly focus on the kinds of topics involved in courtroom discourse. Since 1990, the Law and Society Association, an international association that meets outside the United States every third year, has scheduled sessions on Language and Law at its annual meetings. The International Association of Forensic Linguists (IAFL) came into existence as a result of a series of workshops and conferences in the early 1990s; it meets biannually, usually outside the United States The IAFL publishes a journal, until recently titled Forensic Linguistics: The International Journal of Speech, Language and the Law. Future issues will drop the original title and use the original subtitle only. References Coates, Linda, Janet Beavin Bavelas, and James Gibson. 1994. Anomalous language in sexual assault trial judgments. Discourse & Society 5(2 April).189–206. Conley, John M., William O’Barr, and E. Allen Lind. 1978. The power of language: presentational style in the courtroom. Duke Law Journal 78. 1375–99. Cotterill, Janet. 1998. ‘If it doesn’t fit, you must acquit’: metaphor and the O. J. Simpson criminal trial. Forensic Linguistics: The International Journal of Language and the Law 5(2). 141–58. Cotterill, Janet. 2002. ‘Just one more time…’ aspects of intertextuality in the trials of O.J. Simpson.” Language in the Legal Process, ed. by Janet Cotterill, Chapter 9, Basingstoke and New York: Palgrave. 147–61.

CRIOULO, GULF OF GUINEA Dumas, Bethany K. 2000a. Jury trials: lay jurors, pattern jury instructions, and comprehension issues. Tennessee Law Review 67(3 Spring). 701–42. Dumas, Bethany K. 2000b. U S pattern jury instructions: problems and proposals. Forensic Linguistics: The International Journal of Language and the Law 7(1). 76–98. English, P. W., and B.D. Sales. 1997. A ceiling or consistency effect for the comprehension of jury instructions. Psychology, Public Policy, and Law 3. 381–401. Gibbons, John. 2003. Forensic linguistics: an introduction to language in the justice system. Malden, MA and Oxford: Blackwell.

Harris, Sandra. 2001. Fragmented narratives and multiple tellers: witness and defendant accounts in trials. Discourse Studies 3 (1 February). 53–74. Matoesian, Greg. 1999. Intertextuality, affect, and ideology in legal discourse. Text 19(1). 73–109. O’Barr, William M. 1982. Linguistic evidence: language, power and strategy in the courtroom. New York: Academic Press. Stygall, Gail. 1994. Trial language: differential discourse processing and discursive formation. Amsterdam and Philadelphia: John Benjamins.

BETHANY K. DUMAS

Crioulo, Gulf of Guinea Gulf of Guinea Crioulo (GGC) is a cover term for four creoles lexically based on Portuguese. They are spoken on the three islands of São Tomé (around 100,000 inhabitants), Príncipe, and Annobón (about 5,000 each), of which the former two constitute an independent republic while the latter belongs to Equatorial Guinea. In the following, the languages will be referred to as Sãotomense, Principense, Annobonese, and Angolar, although they are also known by other names. Sãotomense and Angolar are both spoken on São Tomé, the latter by descendants of runaway plantation slaves in the south of the island. Previously uninhabited, toward the end of the fifteenth century the islands were settled by Portuguese and their African slaves, the latter of which were brought partly from what is today the coast of Nigeria, and partly from Bantu-speaking areas around the mouth of the Congo. It is in this early period that GGC is assumed to have emerged. GGC varieties are interesting for a number of reasons, not least because they probably are among the oldest creoles known. They are also typologically unusually distant from their lexifier. All four creoles coexist in a diglossic relationship with the official languages of the two countries (Spanish on Annobón, Portuguese in the case of the three others) and are used neither in media nor education. Although influenced by Spanish, GGC does not seem to be seriously threatened on Annobón, where it is the universal vernacular (although all speakers are proficient in Spanish). Angolar is believed to have about 9,000 speakers, but is giving way to both Sãotomense and Portuguese. It is not entirely clear

how many speakers use Sãotomense or Príncipense. According to a mid-1990s source, virtually everybody on São Tomé was capable of speaking Portuguese, while only half of the children were competent in Sãotomense, indicating language shift. For reasons yet to be fully understood, this process started earlier and has proceeded further on the smaller Príncipe, and in the 1987 census, less than 16% of the population were reported to speak the language, including virtually no children. The four GGCs share a number of features on all levels, suggesting a common origin. These include, among others things, a circumverbal sentence negation that can be reconstructed as /na … fa/, the second part of which appears clause-finally. This and other similarities can presumably be traced back to São Tomé, which was the first of the three islands to be settled. Subsequently, however, the languages have drifted apart to a degree where mutual comprehension is difficult, albeit possible. Like other creoles, GGCs are heavily analytic, and a participial suffix /-du ~ -ɾu/ seems to be the only attested bound morpheme. The Tense Mode Aspect system is mostly preverbal, and typically includes a past marker tava (< P estava), an imperfective element ka (< P ficar, cá or capaz?), indicating habitual aspect in isolation, and progressive when preceded by the copula sa), and completive particles za and kaba (< P já and acabar respectively). Features that set GGC apart from Portuguese, but make them similar to other creoles in the Atlantic area, include reflexive constructions involving the word for ‘body’, the use of 3pl as a nominal pluralizer, and use of verb serialization, for all of which African influence may plausibly be invoked.

249

CRIOULO, GULF OF GUINEA Although being SVO, the GGC languages display several rather exotic orderings from a European perspective, such as N NUM ɔmã dɔs ‘two hands’). Principense includes a coarticulated stop / /, corresponding to /kw/ elsewhere. It is not clear whether nasal+stop sequences should be regarded as cluster or as single phonemes. In all varieties except Principense, Portuguese /v/ turns up as // in some items. Principense and Angolar have a rhotic phoneme, but otherwise Portuguese rhotics appear as laterals or are deleted. All have depalatalized etymological /ʃ, / into /s, z/, but also sport a productive palatalization rule whereby /t, d, s, z/ are realized as [tʃ, d, ʃ, ] before high front vocoids (exceptions: Principense lacks [d], and Angolar has /θ, ð/ in positions where its sisters have /s, z/). Another segmental difference within the group is that in Annobonese the velar plosive /k/ is realized as [x] before back vowels (including /a/). The GCCs have been argued to be either tonal or pitch-accent languages. All varieties have a few words that are etymologically consonant-initial, but that have been equipped with a prothetic vowel, possibly deriving from the Portuguese definite article. These items all belong to the core lexicon, and are, interestingly, far more common in Principense (e.g. ufógo ‘fire’) than elsewhere (fógo). With the exception of Angolar (see below), the African lexical contribution is limited to a couple of hundred items at the most. A third of these are derived from languages of present-day Nigeria, most of the remainder being of Bantu origin. The latter group is

better represented in Sãotomense than elsewhere. Noteworthy African contributions include 3pl pronouns, and also a 2pl form in Principense. Not unexpectedly, the Annobonese lexicon nowadays contains a fair proportion of Spanish loans. Due to labor migration to Fernando Poo, where pidgin/creole English is the lingua franca, Annobonese has also assimilated some lexical items from this language. Lexically, Angolar is the odd man out in the GGC group. While being structurally very similar to Sãotomense, it contains an unusually high proportion of African lexicon, largely derived from Kimbundu. References There are five book-length treatments of GGC, listed below. In addition to these, some additional data can be found in scattered articles by Germán de Granda (mostly Annobonese), Marike Post (Annobonese), Tjerk Hagemeijer (Sãotomense) and Luís Ivens Ferraz, and in Valkhoff (1966). Barrena, Natalio. 1957. Gramática annobonensa. Madrid: Consejo Superior de Investigaciones Científicas. Ferraz, Luís Ivens. 1979. The creole of São Tomé. Johannesburg: Witwatersrand University Press. Günther, Wilfried. 1973. Das portugiesische kreolisch der Ihla do Príncipe. Marburg: Marburger Studien zur Afrika- und Asienkunde. Lorenzino, Gerardo. 1999. The Angolar Creole Portuguese of São Tomé. Munich: Lincom Europa. Maurer, Philippe. 1995. L’Angolar: Un créole afro-portugais parlé à São Tomé. Hamburg: Helmut Buske Verlag. Valkhoff, Maurius. 1966. Studies in Portuguese and Creole.Johannesburg: Witwatersrand University Press.

MIKAEL PARKVALL See also Crioulo, Upper Guinea

Crioulo, Upper Guinea Upper Guinea Crioulo is a continuum of Portuguesevocabulary creoles spoken mainly on the Cape Verde Islands and in Guinea-Bissau. A variety of GuineaBissau Creole also spills over into Casamance, the southernmost province of Senegal. The total number of native speakers is estimated at 400,000 in Cape Verde, 250,000 in Guinea-Bissau, and 40,000 in Senegal. In addition, there are half a million secondlanguage speakers in Guinea-Bissau and some 20,000 in Casamance. There is also an important diaspora of Cape Verdeans in both Europe and the United States.

250

Despite not having official status in either of the three countries (although since 1998 the ‘national language’ of Cape Verde), Portuguese Creole nevertheless has a strong position in Cape Verde by virtue of it being the native language of the entire population, and in Guinea-Bissau through being the only nationwide lingua franca. Judgments on how similar Cape Verde and GuineaBissau Crioulo really are differ. While some Cape Verdeans claim to be unable to understand GuineaBissau Crioulo, others treat all Upper Guinea

CRIOULO, UPPER GUINEA Portuguese Creoles as dialects of one language. In any case, the differences lie more in pronunciation and vocabulary than in structure. Also, Cape Verde Crioulo is in itself rather heterogeneous, and the varieties of the northern Sotavento islands are in most respects intermediate between those of the southerly Barlavento islands and Guinea-Bissau Portuguese Creole. These three main varieties thus constitute a continuum. Some authors insist that Cape Verde and GuineaBissau Portuguese Creole emerged independently, despite the great similarities. Most, however, seem to agree that they are somehow genetically related. Even so, there are two possibilities—did the ancestral Creole travel from the mainland to the islands or vice versa? Both hypotheses have been suggested, without any conclusive evidence being presented for either. Regardless of the direction of influences during the birth of the Upper Guinea Portuguese Creoles, it is clear that Cape Verde Crioulo has had some impact on the mainland varieties from at least the eighteenth century. First, the Portuguese often used Cape Verdean employees in Guinea-Bissau. Second, famine led to migration toward the mainland in the late 1800s, and third, islanders were prominent in the independence movement. Anyone accustomed to the better-known Caribbean creoles will notice the—as yet unexplained—absence of many of their characteristic features in the Upper Guinea Portuguese Creoles. Broadly speaking, the Sotavento varieties (and even more so Guinea Bissau Portuguese Creole) are typologically more distant from Portuguese, and present more features perceived as typical of creole languages. This is manifested e.g. in the use of (mostly) independent words to mark tense. Although usually referred to as a creole, it may be that Barlavento Cape Verde Crioulo should more properly be designated as a semi-creole. Although there is little evidence of this, many an observer has suggested that the Cape Verde dialects have in general been significantly more creole-like, and that decreolization has brought them closer to Portuguese with time. Today, Many Cape Verdeans are competent in Portuguese, which remains the official language. An obvious difference between the Cape Verde dialects and the mainland is that the latter creole varieties are still in close contact with African languages. At least some change in the direction of Portuguese has taken place during the twentieth century, and is reflected in variation between e.g. more Portuguese-like Guinea-Bissau /adivia/ vs. basilectal /dibia/ ‘to guess’ ( [voiced] / __ [voiced] then any representation that is [sonorant] will undergo the rule in the right context, but this set of representations need not be identified within the grammar as constituting a natural class, independently of this rule. References Chomsky, Noam, and Morris Halle. 1968. The sound pattern of English. New York: Harper & Row. Clements, G.N., and Elizabeth Hume. 1995. The internal organization of speech sounds. The handbook of phonological theory, ed. by John Goldsmith, 245–306. Oxford: Blackwell.

CHARLES REISS See also Feature Theory

Navajo and Athabaskan-Eyak Language The Navajo language is one of several Southern Athabaskan languages, more commonly known as Apachean languages, spoken in the Southwest. The Athabaskan languages, along with Tlingit and Eyak, make up the larger Na Dene language family. There are three major geographical groupings of Athabaskan languages: (1) Northern Athabaskan—including about 25 languages spoken in Alaska and Canada, including Koyukon, Holikachuk, Tanacross, Ahtna, Dena’ina, Deg Hit’an, Tanana, Upper Kuskokwim in Alaska, Han and Gwitch’in in Alaska and Canada, Tagish, Tutchone, and Tahltan at the Yukon headwater, Tahltahn, Sekani, Tsetsaut, Babine, Carrier, and Chilcotin in the far west of Canada, and Hare, Bearlake, Mountain, Kaska, Dogrib, Slave, Beaver, Chipewyan, and Sarcee across the rest of western Canada. It is important to note that, rather than discrete groups, the Alaskan and Canadian Athabaskan languages constitute a language and dialect

728

continuum, i.e. neighboring languages/dialects may be mutually intelligible. (2) Pacific Coast Athabaskan (and extinct Kwalhioqua-Tlatskanai)—including Hupa and Tolowa in California, and Tututni in Oregon. (3) Navajo and Apache—including seven Apachean languages divided into two main groups: (1) the Western group, including Navajo in one branch and a second branch divided into the San Carlos group (San Carlos, White Mountain, Cibecue, etc.) and the Mescalero-Chiricahua group, and (2) the Eastern group, divided into a branch containing Jicarilla and Lipan and a second branch containing Plains Apache. Of the Athabaskan languages, all but Navajo are moribund, with little if any intergenerational transmission. Even Navajo, with approximately 125,000 speakers, is threatened: while more than 95% of children entering bilingual education programs were fluent in Navajo in the mid-1970s, today fewer than 50% of

NAVAJO AND ATHABASKAN-EYAK LANGUAGE kindergartners are fluent. Of the Apachean languages, Navajo is by far the best documented, and indeed may be the best documented of any Native American language. This is due in large measure to the pioneering work of Robert Young and William Morgan. Their grammars and dictionaries of Navajo over the past half century (e.g. 1987, 1992) have set the standard for work both in Apachean and in Athabaskan more generally.

Phonology The distinctive sounds (phonemes) of Athabaskan include obstruents, sonorants, and vowels. The ancestor language—Proto Athabaskan—likely contained the following stop obstruents: T

TL

TS

CH

Ky

CHw Q

The modern languages show a range of shifts among the Proto-Athabaskan stops. The following list shows the modern reflexes of the Proto-Athabaskan word *-tsi’ ‘head’: Tanaina Tanana, Han Koyukon Gwitch’in (AK) Gwitch’in (Can.) Hare Bearlake Navajo

-tsi’ -tthì’ -tli’ -kì’ -chì’ -pfí’, fí’ -kwí’, kfí’ -tsi:’

The sonorants of Proto Athabaskan were w and y, m, n, and a palato-velar nasal. Proto Athabaskan vowels are reconstructed as i, e, a, u plus three reduced vowels (mid, low, and back). All of the modern Athabaskan languages have both oral and nasal vowels. Some of the languages have short and long vowels; others have full and reduced vowels. The phonemes of Navajo include plain, aspirated, and glottalized stops and affricates (d, dl, dz, j, and g are the symbols used for the plain stops), fricatives (s, z, sh, zh, h, gh), and liquids and sonorants (l, B , y, w, n, m). Navajo has plain and nasalized short and long vowels, and high vs. low tone (distinctive pitch).

Morphology and Verb Structure The Verb Complex The Athabaskan verb is a complex polysynthetic structure made up of a stem plus prefixes. The stem is

composed of a root plus suffixes (or other modification) indicating mode and aspect. Prefixes indicate subject and objects, mode and aspect, and adverbials. The prefix complex can be analyzed as a template comprised of basic positions or zones preceding the stem. The prefix chart in Figure 1 gives a general idea of the ordering of prefixes. Note that the stem is in position zero and subject prefixes are in positions 2 and 5, with tense, aspect, and mode in 3, and direct object in 6. Some of the prefixes noted above are obligatory in the sense that the particular position they occupy must be filled by one of the possible variants of that prefix class in every derivative of the verb. These obligatory prefixes are the person, tense, aspect, mode, and voice/valence (also called the ‘classifier’) prefixes. Other, derivational, prefixes, e.g. adverbial, iterative, are optional—he positions these prefixes occupy may or may not be filled depending on the meaning of the derivative. Sometimes, prefixes are thematic—they encode a situational participant—and they are lexicalized, which means that they are present in every derivation of a particular verb. These thematic prefixes in combination with the voice/valence prefix and the root of the verb are referred to as the VERB THEME. The Athabaskan theme is the underlying skeleton or verb construction to which prefixes and suffixal elements are added in producing an utterance. The theme itself has a meaning and is the basic unit of the Athabaskan verbal lexicon. Verb forms derived from themes have, in addition to the stem (i.e. the aspectually suffixed root) and theme prefixes, inflectional prefixes (person, number, etc.) and derivational prefixes (aspect, adverbials, etc.). These prefixes are arranged in relatively fixed positions preceding the stem as shown in Figure 1. In Navajo, the THEME consists of the valence prefix (position 1) plus the stem (position 0). The VERB BASE consists of the mode—conjugation prefixes (position 3) plus the subject pronoun (position 2) plus the verb theme. Examples of verb themes in Navajo, with verb words derived from them, are shown below: Theme: na + θ + né ‘play’ Example verb: naashné ‘I’m playing’ Analysis: na + θ + sh + né Thematic + imperf + 1st person subject + stem Theme:

ha # O + B + géésh

‘cut O (object) out’

8 9 3 5 0 10 # # 2 4 1 6 7 reflexive Subj. voice/valence stem derivational/ mode/ distrib./pl disjunct DO Subj. derivational/ post boundary (3rd) reversionary thematic conjugation (1st & 2nd positions thematic semiliterative sg & pl) iterative

Figure 1. Navajo prefix chart.

729

NAVAJO AND ATHABASKAN-EYAK LANGUAGE Example verb: hadeiilgéésh ‘we’re cutting it out’ Analysis: ha # da # θ + θ + sh + B + géésh Thematic # distributive plural # 3rd person direct obj. + imperf. + 1st person subj. + valence + stem

or paired objects, (3) slender, rigid, or stick-like objects, (4) animate objects, (5) plural separable objects, (6) plural objects in profusion, (7) noncompact matter, (8) objects in open containers, (9) flat, flexible, or cloth-like objects, (10) mushy wet objects, and (11) heavy, large, or bulky objects.

The Aspectual System Aspect in Athabaskan is expressed by a complex, multidimensional system. There are two major categories that combine in expressing the temporal contour of the state or activity described by the verb. The convention in Athabaskan linguistics is to call one of these two categories MODE and the other ASPECT (although some researchers refer to these categories as ‘aspect’ and ‘aktionsart’, respectively). Both of these categories are obligatory—every verb must be marked for both mode and aspect. This morphological marking is an intersecting one, i.e. mode and aspect are marked by prefixal and suffixal elements, which encode both categories. There are seven modes in Navajo, describing activity as incomplete, complete, ongoing, future, potential, customary, or recurrent. Examples of verbs in the imperfective and perfective modes are:

Most transitive Navajo sentences do not name both the subject and object in noun phrases. When only one is mentioned as a noun phrase, it is usually the direct object, that is, the new information. In this case, the direct object has an agreement prefix on the verb, which is yi-. However, when the topic, or old information, is mentioned in a noun phrase, this unusual case is marked by a different object agreement prefix, bi-. For example:

Imperfective: Perfective:

Mary hayííB tî Mary habííltî

yáshti’ yáB áti’

‘I’m talking’ ‘I talked

The second basic temporal category is aspect, which describes the manner in which an activity or event is carried out over time—whether it happens once, or repeatedly, or at length. There are 12 aspects in Navajo and they allow the expression of such distinctions in meaning as: ‘I am red’ vs. “I turned red’ and ‘I am chewing it’ vs. ‘I bit it’. For example: Durative: Momentaneous: Repetitive:

yáshti’ ‘ayániishtééh yádíshtih

‘I’m talking’ ‘I’m starting to talk” ‘I’m chattering’

Classificatory Verbs Athabaskan verbs in general show a wide range of classificatory functions. Some verb stems are specialized according to whether the subject of the action is singular (or dual) or plural. Others are specialized for singular vs. plural objects. Still other actions are referred to by different stems depending on the physical characteristics of the subject or object involved. These stems are the classificatory verb stems. The Athabaskan classificatory verbs categorize the material, shape, consistency, size, animacy, arrangement, quanta, and containment of the subjects of intransitive verbs of position and location, and the objects of transitive verbs of handling. The 11 Navajo classificatory verbs delineate (1) solid, round, or compact objects, (2) slender, flexible,

730

Word Order The word order of the Navajo sentence is Subject–Object–Verb (SOV). Aééchaa’í mósí yinooB chééB Dog cat it’s chasing it S O V

‘He carried Mary up’ ‘Mary carried him up’

The Future of Athabaskan Languages Work on language revitalization is under way in many Athabaskan-speaking communities, and efforts in the Navajo community are perhaps the most robust. There are language programs in many schools across the Reservation, and at the two branches of Diné College. The language is also taught at the state universities of New Mexico and Arizona. The Rock Point School on the Navajo reservation is a model bilingual education program. At Rock Point, children come expecting that they will succeed and they do—the Navajo program gives students pride in being Navajo, in their language, and in their culture. This effective program includes lessons in Navajo by community elders on Navajo cultural matters and a strong involvement of parents and so family both in conferences and in activities such as Language Fairs and book-making nights. References Axelrod, Melissa. 2003. The semantics of time. Lincoln: University of Nebraska Press. Cook, E-D., and Keren Rice (eds.) 1989. Athapaskan linguistics: current perspectives on a language family. Berlin: Mouton. Faltz, Leonard. 1998. The Navajo verb: a grammar for students and scholars. Albuquerque: University of New Mexico Press. Fernald, Ted, and Paul Platero (eds.) 2000. Athabaskan syntax and semantics. Oxford: Oxford University Press.

NEUROLINGUISTICS Jelinek, Eloise , Sally Midgette, Keren Rice, and Leslie Saxon (eds.) 1996. Athabaskan language studies: essays in honor of Robert W. Young. Albuquerque: University of New Mexico Press. Kari, James. 1979. Athabaskan verb theme categories. Fairbanks, AK: Alaska Native Language Center Research Papers, #2. Leer, Jeff. 1979. Proto-Athabaskan verb stem variation, part one: phonology. Fairbanks, AK: Alaska Native Language Center Research Papers, #1.

Rice, Keren. 2000. Morpheme order and semantic scope: word formation in the Athapaskan verb. Cambridge: Cambridge University Press. Sapir, Edward, and Harry Hoijer. 1967. The phonology and morphology of the Navaho language. Berkeley: University of California Press. Young, Robert W. 2000. The Navajo verb system: an overview. Albuquerque: University of New Mexico Press.

MELISSA AXELROD

Neurolinguistics Neurolinguistics is a relatively new discipline that comprises both clinical and basic research into the functional relationship between language and the brain. The research topics of neurolinguistics require a multidisciplinary approach and share parts of traditional disciplines such as linguistics, medicine, psychology, and computer science, and more precisely, parts of the fields of psycholinguistics, aphasiology, neurology, brain imaging, neuropsychology, and neuroinformatics. Due to the combination of theories and methods, all these fields contribute to a new discipline in life sciences, called cognitive neurosciences. As for linguistics, this leads to an innovative research area providing empirical and neurophysiological methods for testing aspects of linguistic theories and models. Looking back on the first emergence of common research methods in the history of linguistics, five main stages can be observed until now: (1) language philosophy, using introspection with an over 2000-year-old tradition; (2) first empirical studies on genealogy and roots of languages, using comparative approaches in typology for more than 200 years; (3) empirical behavioral studies of psycholinguistics, using e.g. reaction time experiments, with an almost 50-year-old tradition; (4) computer-aided simulation of language processes in computerlinguistics, using e.g. neural networks, for more than 30 years; (5) brain imaging of language processing and higher brain function in cognitive neurosciences during the last 20 years. Therefore, neurolinguistics is one of the newest, but also one of the most rapidly growing fields in contemporary linguistics. At present, neurolinguistics can be divided into three major branches, depending on the research domain (see Figure 1).

Clinical Neurolinguistics or Clinical Linguistics This field, formerly also called patholinguistics or aphasiology, mainly deals with clinical aspects of

brain-related language disorders (aphasia). Aphasia is an acquired language disorder caused by brain damage, which may induce severe impairment in both language comprehension and production, as well as in reading and writing. Reasons for brain damage are mainly strokes, due to an interruption of the blood flow that causes the death of nerve cells in the respective area. Possible reasons for the interruption may be a blocked vessel (arteriosclerosis), a blood clot (thrombosis), or a burst blood vessel accompanied by bleeding, e.g. caused by a ballooning expansion of a weak vessel wall (aneurysm). Moreover, traumatic head injuries and brain tumors in certain areas may cause such language deficits. Types of aphasic language disorders range from minor difficulties, which concern only naming, to major difficulties such as a complete loss of language. The strength of the impairment depends on the location and extent of the brain damage. Currently, aphasia affects more than 1,000,000 individuals (stroke survivors) in the United States. By combining linguistic theory and neurological knowledge on aphasia, clinical neurolinguists typically work in stroke units and rehabilitation clinics, providing mainly three occupations: (1) Diagnosis: In order to test aphasic patients on the extent and the special type of language deficit in standardized examinations, several neurolinguistic test batteries are used: e.g. the Boston Diagnostic Aphasia Examination

(Aphasiology, Medicine)

Clinical (Neuro)Linguistics

Experimental Neurolinguistics

Simulative Neurolinguistics

(Neurobiology, Medicine)

(Computerlinguistics, Informatics)

Figure 1

731

NEUROLINGUISTICS (BDAE), Porch Index of Communicative Ability (PICA), or the Western Aphasia Battery (WAB). (2) Speech therapy: This represents the main activity of the clinical neurolinguists, who have to perform adapted speech therapy programs with the patients for several weeks or months. (3) Clinical research: This part of their occupation covers the linguistic aspects of aphasia research in hospitals and universities and results in the optimization of treatment efficacy and effectiveness of therapy programs.

Experimental Neurolinguistics In contrast to clinical neurolinguistics, this field mainly covers basic research on the neurobiology of language of healthy probands. Experimental neurolinguists try to understand the nature of representation and physiological processes that contribute to normal language by investigating underlying neural mechanisms in general. Their main goal is to study the neurophysiological phenomena in the brain during language comprehension and production. Combining state-of-the-art neurophysiological noninvasive techniques with linguistic, neurobiological, and neuropsychological findings, experimental neurolinguists investigate certain aspects of language processing under laboratory conditions. Since this work requires high-tech equipment and a broad expertise in several disciplines, this kind of research involves typical teamwork, often represented by more than one laboratory. Common noninvasive techniques for the study of language processes in healthy participants are the following techniques, which represent the most frequently used empirical methods in neurolinguistics. Some invasive techniques, only used during preoperative diagnostics of patients, are described later. (1) Electroencephalography. By using scalp discelectrodes, which are attached on to the head skin by an electrolyte conductive gel, human electroencephalography (EEG) is a real noninvasive technique, first described in 1929 by Hans Berger. This technique gained its current importance in neurolinguistics after powerful computers became available. Recording the electrical activities of the underlying neural substrate (at least several 10,000 neurons), the EEG signal provides information on the time course of certain cognitive processes. Since the EEG measures the brain’s activity directly, it has a very high time resolution below 1 millisecond and this allows assessing ‘when’ a certain event is processed in the brain. Analyzing the brain waves elicited of a certain event (e.g. onset of a given word), the amplitude and time course of the wave can be interpreted (= eventrelated potential, ERP). Analyzing the coherence

732

changes within certain frequency bands, the correlated activity of different brain areas during a task can be measured (= coherence analysis). This latter analysis allows to monitor the cooperation of different neuronal networks during cognitive tasks such as language processing. (2) Magnetoencephalography. Like EEG, magnetoencephalography (MEG) is the second real noninvasive technique, and provides the same advantages of direct measurement of brain processes. In contrast to EEG, MEG measures the magnetic component of electrical activity. By using superconducting quantum interference devices (SQUIDs), which are contactlessly positioned around the head in a so-called Dewar, MEG records the very low magnetic fields accompanying the neural activity. Like the EEG technique, ERP- and coherence analysis of the MEG signal allows investigating language processes in the time and frequency domain. (3) Positron emission tomography. Even though the positron emission tomography (PET) technique is not based on invasion of the skull, it is a ‘small’-invasive technique, since radioisotopes are injected or inhaled. In clinical use since the late 1970s, PET scanners visualize differences in metabolic processes of the brains during a stimulus task (e.g. listening to words) and during a control task (e.g. listening to noise). Short-lived radioactive isotopes are intravenously injected and can be tracked from outside the head with a tomographic scintillation counter after their radioactive decay. While passing the brain tissue, the emitted positron of the former isotope collides after 2 to 8 mm with a body’s electron, producing two gamma rays. By detecting and analyzing such coincident gamma rays, the spatial position of the former isotope can be determined. The spatial distribution of the isotopes in the brain allows an insight into the extent of the metabolic processes, and thus the brain’s activity during a time window of several seconds. After the computational visualization of the data, false color pictures indicate the brain areas of higher or lower neural activity during a certain language task compared with a control task. (4) Magnetic resonance tomography (MRT) or magnetic resonance imaging (MRI) are synonymously used for a powerful technique, which allows threedimensional high-resolution brain imaging. After the head is exposed to the strong magnetic field of a cylindrical superconducting magnet, the spin axes of every hydrogen atom of the head brought into alignment. Then, short radio waves of the hydrogen resonance frequency (125 MHz) are sent to the tissue and by absorbing the energy the spin axes of the atoms are disordered for a short moment. After this radio wave stimulation, the hydrogen spin axes return to their

NEUROLINGUISTICS original, randomized state and emit energy in the form of a very weak radio wave. These remitted radio waves deliver the information on the head’s structure. After being analyzed and visualized by computer, cross-sectional images of the brain with a resolution of less than 1 millimeter can be made. Outside of clinical use, MRI scanners of up to 11 Tesla are used. This technique provides high-resolution information on the anatomy of the brain and can be combined with electrophysiological findings of EEG or MEG. In order to monitor cognitive processes in terms of metabolic changes, functional MRI must also be performed. Functional magnetic resonance imaging (fMRI) maps changes in oxygen concentration that correspond to nerve cell activity. By overlaying a map of oxygen concentration changes on the anatomical data, those brain areas that show increasing activity correlated to a given stimulus can be defined. Even though the time resolution is not excellent, the spatial resolution is in the range of several millimeters, which makes fMRI one of the most powerful techniques in neurolinguistic research. (5) Invasive techniques. During preoperative diagnostics, e.g. prior to epilepsy surgery, electrode arrays are implanted and monitored for several days to identify the focus of abnormal electrical activity. Such intracranial electrodes, placed directly onto the cortex, allow the recording of underlying cortical activity during cognitive tasks, e.g. object naming, speaking, reading, etc. Furthermore, the underlying cortex area can be electrically stimulated by short pulse trains of weak currents. If the stimulated cortex area contributes to language, speech arrest or another aphasiological symptom will occur. In other words, by electrical stimulation, temporary lesions of the brain can be made in order to test the possible role of a given cortex region in the language process. This test is required prior to the performance of a neurosurgical epilepsy operation to assess where certain cerebral functions are centered such as memory and speech. Another invasive technique applied for the same clinical reasons as mentioned above is the Wada test or Intracarotid Sodium Amytal Test (first described by Juhn A. Wada). After the injection of an anesthetic (e.g. sodium amytal) into the right or left internal carotid artery, the respective hemisphere is anesthetized for a few minutes. The type and extent of cerebral dominance of a patient can be assessed during the following behavioral test of language comprehension, production, object naming, and memory abilities. In a healthy, typical right-handed volunteer, suspension of the left (language-dominant) hemisphere leads to a Wernicke aphasia with a neologistic jargon, whereas suspension of the right hemisphere leads to a dysprosody.

Simulative Neurolinguistics In simulative neurolinguistics, the knowledge of theoretical linguistics, clinical linguistics, neuropsychological case studies, and results of experimental neurolinguistics are used as input for a computer simulation with special software environments. By using computational techniques of computerlinguistics and neuroinformatics (e.g. connectionist neural networks), previously observed real language processes can be used for a computer simulation. During simulation, each condition can be modified step by step, and the results can be obtained immediately. The time required depends only on the computing power of the hardware used. The advantage of this technique is the enormous flexibility of the simulated paradigm. Contrary to reallife processes, where investigators have to wait for patients with certain impairments, virtual impairment can be created and tested within seconds. On the other hand, the reliability of the predictions of such simulations depends on the implemented processes, models, and constraints. However, even though simulative neurolinguistics seem to have an enormous potential for future research, it did not have a remarkable influence on neurolinguistics or cognitive neuroscience until now. Based on the large progress in brain imaging and computerized analysis of brain data during the last few years, it can be expected that neurolinguistics will strongly contribute to linguistic theory and modeling during the next decades. For a few years now, brain processes during language can be investigated with a resolution in time of less than 1 millisecond and in space of less than 1 millimeter. Thus, a powerful new tool is available for linguists in the at least 2000-yearold attempt to understand human language. All methods developed and used in linguistics (see above) form an emergent instrument for the very different topics, questions, and approaches in studying language within the different fields of linguistics. For most of them, an integration of neurophysiological techniques would be valuable. For example, neurolinguistic studies indicate how and when language is learned as first or second language in a mono- or bilingual child, how different languages are learned by a juvenile or adolescent, and how it is represented in the brain as well as how learning can be facilitated or disturbed. In addition, the contemporary theories and models of language may be improved by neurolinguistic findings. The enormous impact of the behavioral data of psycholinguistics on linguistic modeling since the 1950s did show the necessity and usefulness of empirical evaluation. Neurolinguistic research offers an even stronger and more direct observation of cognitive phenomena and language processes in the brain.

733

NEUROLINGUISTICS References Benson, D. Frank, and Alfredo Ardila (eds.) 1996. Aphasia: a clinical perspective. Oxford: Oxford University Press. Brown, Colin M., and Peter Hagoort (eds.) 1999. The neurocognition of language. Oxford: Oxford University Press. Caplan, David. 1992. Language: structure, processing, and disorders. Cambridge: MIT Press. Eling, Paul (ed.) 1994. Reader in the history of aphasia: from Gall to Geschwind. Amsterdam: John Benjamins. Frackowiak, Richard S.J., Karl J. Friston, Christopher Frith, and Raymond Dolan (eds.) 1997. Human brain function. San Diego: Academic Press. Friederici, Angela D. (ed.) 1999. Language comprehension: a biological perspective. Berlin: Springer-Verlag. Gazzaniga, Michael S. (ed.) 2000. The new cognitive neurosciences. Cambridge: MIT Press.

Grodzinsky, Yosef, Lew Shapiro, and David Swinney (eds.) 2000. Language and the brain: representation and processing. San Diego: Academic Press. Kertesz, Andrew (ed.) 1997. Localization and neuroimaging in neuropsychology. San Diego: Academic Press. Rugg, Michael D. (ed.) 1997. Cognitive neuroscience. Cambridge: MIT Press. Stemmer, Brigitte, and Harry A. Whitaker (eds.) 1998. Handbook of neurolinguistics. San Diego: Academic Press. Zemlin, Willard R. 1998. Speech and hearing science: anatomy and physiology, 4th edition. Boston: Allyn & Bacon.

HORST M. MÜLLER See also Clinical Linguistics; Localization of Linguistic Information

New Guinea New Guinea, the second largest island in the world, lies roughly 100 miles north of Australia. It contains more than 1,000 languages, about one sixth of the world’s total (or one language in every 900 square kilometers), making it linguistically one of the most dense and complex regions of the world. The name was given to the island by the Spanish navigator Ortiz de Retes, possibly because he saw some resemblance between the indigenous population and the inhabitants of the Guinea coast of Africa. Politically, the island is divided between two countries: Papua New Guinea, occupying the eastern half of the island, and some 600 associated islands, the largest of which are New Britain, New Ireland, and Bougainville, and Irian Jaya (West Irian), a former Dutch colony until 1963, now the easternmost province of Indonesia. An independent nation since 1975, Papua New Guinea is historically an amalgamation of what were two separate colonies: Kaiserwilhelmsland, occupying the northeastern half of the mainland and islands under German rule from 1884 until 1914, and subsequently mandated by the United Nations to Australian administration under the name ‘New Guinea’; and Papua, the southern half of the mainland, formerly British New Guinea, which was proclaimed the Australian territory of Papua in 1906. Papua New Guinea has the highest population (5.2 million) and land mass of the Pacific island nations, with 64% of its population and 84% of its land mass. Irian Jaya has a population between 1.8 and 4 million and occupies a land mass approximately the size of California.

734

New Guinea was originally peopled by many different waves of migrants, whose prehistory is largely unknown. Although human habitation of the island extends back some 40,000 years, recorded history is very recent and in some cases goes back only a few decades. It was the last major land area in the world to be colonized by European powers; almost all regions have a history of contact of less than a century. The terrain is extremely rugged, with mountains reaching altitudes of more than 15,000 feet, dense tropical rain forests, and fast-flowing rivers, which have long cut the interior of the country off from outsiders. Many villages have no road or river links with other centers, and some can be reached only by walking for up to two weeks. The peoples of New Guinea are mainly settled villagers living in a subsistence economy. Their productive activities vary according to the zone they inhabit on the island’s extraordinary vertical ecology. The extreme highlands have an alpine climate, with widespread frost. There, an intensive agriculture based on the sweet potato has developed over the last few hundred years. This highly productive system has given rise to a local population boom, and the highlands support large, dense groups with large languages. The ten largest indigenous languages of Papua New Guinea belong to the large groups of the interior Highlands; they have from 30,000 to 100,000 speakers. Between them, they account for nearly one third of the population. Perhaps 80% of the languages have fewer than 5,000 speakers, and as many as one third

NEW GUINEA have fewer than 500. These languages are found mainly in the coastal lowlands and the intermediate areas known as the highland fringe, consisting of pockets of rain forest, swamps, and grassland, where the population is low and thinly spread. Linguists generally recognize two major language groupings in New Guinea: Austronesian and nonAustronesian (or Papuan). The Austronesian languages clearly constitute a family with a common ancestor, Proto-Austronesian, and comprise some 450 descendant languages in the Pacific basin and another 600 to 700 outside it, making it the largest language family. About one quarter of New Guinea languages are of Austronesian origin, and most of them (with the exception of the languages of western Irian Jaya) belong to the Oceanic subgroup established by the German linguist Otto Dempwolff. The relationships among the non-Austronesian (Papuan) languages are less clear, and the label is best seen as a cover term for a number, perhaps as many as 60, of distinct families. Figure 1 shows the distribution of the Austronesian and Papuan languages. Most linguists agree that the coastal distribution of most of the Austronesian languages indicates the later arrival of their speakers. The immediate ancestors of the Proto-Oceanic speakers migrated from eastern Indonesia through eastern Irian Jaya into the Bismarck Archipelago. The interior of New Guinea, where the majority of Papua languages are spoken, experienced no European contact until shortly before (and even in some cases some time after) World War II, and hence most of the languages were unknown to the outside world until quite recently. A number are classified as isolates, i.e. languages that seem to have no known relatives. The Indo-Pacific hypothesis, which attempts to link Papuan languages

with those of Tasmania (but not mainland Australia) and the Andaman Islands in the Indian Ocean, has not been generally accepted, nor has the suggestion of a link between Australian languages and the languages of the Eastern Highlands of Papua New Guinea. It is possible that Australia was settled from New Guinea because they were one continent until around 8,000 years ago. Only after the last Ice Age did sea levels rise to separate them. New Guinea languages are typologically diverse, displaying many interesting and unusual linguistic features. There are languages with SVO (subject– verb–object), SOV, VSO, VOS, and OSV word orders; OVS is the only unattested word order, and this order is quite rare across the world. Most Papuan languages, however, tend to be verb final. In addition, there are examples of noun classifier systems, such as in the Papuan language Nasioi, with a set of more than 100 suffixes added to nouns, adjectives, numerals, and derived nominals to classify the entity being referred to. Both Papuan and Oceanic languages, however, tend to have relatively simple sound systems, with sound inventories smaller than that of English. Consonant clusters are absent or rare in Austronesian, but they are present in many Papuan languages. Almost all Oceanic languages distinguish between inclusive (referring to the speaker and addressee or addressees, ‘I + you’) and exclusive first-person pronouns (referring to the speaker and some other person or persons, ‘I + he/she/it/they’), as well as a three-way distinction in number between singular (one), dual (only two), and plural, or paucal (more than two). Some languages have a trial, referring to only three. Very few Oceanic languages mark gender in pronouns, and most have a three-way distinction in their

NEW BRITAIN

IRIAN JAYA PAPUA NEW GUINEA

Non-Austronesian Austronesian Unclassified 0

20

40

60kms

Figure 1. Map of New Guinea showing New Britain and the distribution of Austronesian and non-Austronesian languages.

735

NEW GUINEA demonstratives between proximate (‘this’ near the speaker), intermediate (‘that’ near the addressee), and distant (‘that’ away from both speaker and addressee). Most Oceanic languages also distinguish two types of possessive marking, with all nouns being classified into alienable and inalienable. Alienable nouns refer to objects that belong to the speaker, where the speaker typically has control over the possession and may reject it (e.g. food, clothes, etc.). Inalienable nouns refer to items inseparably connected to the speaker, such as relatives or body parts. Inalienable nouns are marked with suffixes that vary in a minor way from language to language. Only a small number of Oceanic languages have a contrast between active and passive. Grammatical distinctions, such as tense, are mostly expressed by independent particles rather than affixes: Tuna yau ga ‘I’ PAST ‘I saw it’.

gire see

Papuan languages, however, often have a rich and strongly developed affixal word structure: Yimas ama-wa-t ‘I’ go PERFECT ‘I went’

Pronoun systems vary widely in Papuan languages but are generally not as complex as in the Oceanic languages. Very few Papuan languages distinguish inclusive and exclusive first-person pronouns, although a number of them distinguish gender in pronouns. Articles are virtually nonexistent, and possessive constructions are less complex than in Oceanic languages. A number of languages have more than one existential (‘be’) verb, such as Anggor in the Sepik region of Papua New Guinea, which has 18. The different verb forms depend on the shape of the object, its location, and posture. In Enga, ‘there are men’ is expressed as ‘the man stands’, whereas ‘there are women’ is expressed as ‘the woman sits’. Switch reference, although generally rare in the Pacific, is typical of complex sentences in Papuan languages. When two or more clauses join to form a complex sentence, the last verb in the clause retains the subject-tense marking of the first verb, but the other (medial) verbs do not. Instead, they incorporate a suffix indicating whether the subject of the verb is the same as or different from the subject of the following verb. Nasioi

kad-o-ma talk-‘I’-SAME ACTOR ‘While talking, I went.’

nan-ant-in go-‘I’-IMMEDIATE PAST

da? po-ko nan-amp-e-ain you come-DIFFERENT ACTOR go-‘we’-FUTURE ‘When you come, we two will go.’

736

Much less research has been done on the approximately 250 languages in Irian Jaya, only four of which are spoken by 40,000 people or more, and somewhat more on the roughly 860 languages in Papua New Guinea. In many cases, missionaries undertook the first studies of languages in New Guinea, and missionary linguist work continues today, particularly in Papua New Guinea, through the Summer Institute of Linguistics. Contacts between speakers of Papuan and Austronesian languages have led to rapid change and diversification to the point where languages such as Magori and Maisin in Papua New Guinea are difficult to classify as Austronesian or Papuan. The Motu language spoken along the southern coast of Papua New Guinea has the grammatical particle system typical of Austronesian languages, but it has the typical SOV word order of Papuan languages. tau ese au-na imea bogarai-na-I vada e hado man a tree-the garden middle-its-at PERFECT he/she/it plant ‘The man planted a tree in the middle of the garden.’

The small size of many Melanesian societies has also permitted change to spread more rapidly than in larger societies. Significant changes have affected the basic vocabulary and grammar of the Austronesian language Muyuw, spoken on Woodlark Island in Milne Bay Province of Papua New Guinea, in just a 50-year period. Local vernaculars are seen as a unique badge of identity and distinctiveness. Villagers in one community decided at a meeting that they would be different from other Selepet-speaking villages by adopting a new word (bunge) for ‘no’ to replace their usual word (bia) shared by all Selepet speakers. Multilingualism is widespread, and because people marry outside of their community, husbands and wives often speak different languages. Many people, especially men, know the languages of one or two neighboring communities, or perhaps a language with wider currency around their valley or coastline. Where language groups were large, as in the highlands, only those in the border areas would be multilingual. Where groups were small, everyone was effectively in a border area, and knowledge of multiple languages was universal. In the lowland village of Gapun, whose language Taiap is an isolate spoken by about 80 people, the average number of languages understood by men over 40 was five: the vernacular, a lingua franca, and three or so of the other local languages. In addition to the indigenous languages, there are a number of pidgins and creoles, as well as languages of the metropolitan powers, particularly English (and formerly also German) in Papua New Guinea and Bahasa Indonesia (and formerly also Dutch) in Irian Jaya. Among the pidgins and creoles are those based on

NIDA, EUGENE ALBERT indigenous languages such as Pidgin Yimas, based on the Papuan language Yimas and spoken in Papua New Guinea’s Sepik region, and Hiri Motu (‘trade Motu’), based on the Austronesian language Motu, spoken by a quarter of a million people and one of Papua New Guinea’s three national languages, along with English and Tok Pisin, an English-based pidgin and creole. Papua New Guinea’s most widespread language, Tok Pisin, with over 2 million speakers, is the largest pidgin/creole language in the Pacific. In many parts of Papua New Guinea, children grow up speaking Tok Pisin and no longer acquire their local village language. In Gapun, parents began speaking mainly Tok Pisin to their children, and now children over the age of ten no longer use Taiap. The vernacular languages have very little place in the national life of either Papua New Guinea or Irian Jaya. In the former, English is the main language of government and education, and in the latter, Bahasa Indonesia. References Capell, Arthur. 1969. A survey of New Guinea languages. Sydney: Sydney University Press.

Foley, William. 1986. The Papuan languages of New Guinea. Cambridge: Cambridge University Press. Kulick, Don. 1992. Language shift and cultural reproduction. Socialization, self and syncretism in a Papua New Guinean village. Cambridge: Cambridge University Press. Lynch, John. 1998. Pacific languages. Honolulu: University of Hawaii Press. Sankoff, Gillian. 1980. Multilingualism in Papua New Guinea. The social life of language. Philadelphia: University of Pennsylvania Press. Thurston, William. 1987. Processes of change in the languages of North-Western New Britain. Canberra: Pacific Linguistics Series B, No. 99. Wurm, Stephen (ed.) 1975. New Guinea area languages and language study, Vol I. Papuan languages and the New Guinea linguistic scene. Canberra: Pacific Linguistics Series C, No. 38. Wurm, Stephen (ed.) 1976. New Guinea area languages and language study, Vol II. Austronesian languages. Canberra: Pacific Linguistics Series C, No. 39. Wurm, Stephen (ed.) 1977. New Guinea area languages and language study, Vol III. Language, culture, society and the modern world. Canberra: Pacific Linguistics Series C, No. 40. Wurm, Stephen (ed.) 1980. New Guinea and neighbouring areas: a sociolinguistic laboratory. The Hague: Mouton. Wurm, Stephen (ed.) 1982. The Papuan languages of Oceania. Tübingen: Gunther Narr Verlag.

SUZANNE ROMAINE

Nida, Eugene Albert Eugene Nida’s research in linguistics and cultural anthropology spans a period of 60 years, during which he has continuously developed and refined his perception of languages and cultures. His initial studies of classical Greek and the New Testament as well as his early books on morphology and the structure of English were ideal qualifications for dealing with Bible translations, a task that he was asked to fulfill by the American Bible Society. A large part of Nida’s works and efforts were indeed devoted to showing translators how to better understand the Bible in order to make it understandable, in their turn, to receptors speaking highly different languages and belonging to most diverse cultures. This activity brought him into contact with over 200 languages and cultures in Asia, Australia, Africa, and the Americas. His views on the way in which languages function are thus based on extensive field surveys. In his description of the word and sentence structure of languages, Nida has resorted to the concepts of American structural linguistics, placing, however, the meaning at the center of his investigations and replac-

ing formal with referential classes. In his works, he has amply demonstrated that, despite formal similarities, words may develop different semantic relations between themselves and the use of strictly formal criteria of analysis can be misleading. Moreover, meaning cannot be located in words, which have ‘fuzzy boundaries’, but rather on a ‘molecular’ level, in their combinations with other words. Context thus becomes a fundamental factor in Nida’s analysis of meaning, and he follows Martin Joos (1972) in maximizing its role in lexical combinations. From word and syntax levels, Nida’s research has encompassed, in an increasingly detailed manner, the structures of discourse as well as textual organization, texts being regarded as the basic and ultimate carriers of meaning. Still, his investigation of contexts goes beyond strictly linguistic scrutiny. Like Edward Sapir and Bronislav Malinovski, Nida has constantly emphasized the crucial role of cultures in describing and explaining languages. Moreover, his adoption of a sociosemiotic stance has allowed him to explore the social aspects of language as well as the ways in which

737

NIDA, EUGENE ALBERT language signs relate to other sign systems and to the real world of referents. Eugene Nida has made a fundamental contribution to the development of translation theory and practice, and his impact on Bible translators, translation scholars, and translation schools has been outstanding. He has brought rigor and scientific objectivity to the subject by incorporating concepts, methods, and classifications from linguistics, pragmatics, semantics, and discourse analysis as well as necessary data from cultural anthropology, linking languages to cultures and broader contexts of communication. The concept of dynamic/functional equivalence that he has introduced in the translational discourse has shifted the emphasis in translation theory from the faithful reproduction of source messages to the creation of translated texts with a strong communicative impact, focusing on the receivers’ needs for clarity as well as on their linguistic and cultural expectations. This receptor orientation gives priority to the content of texts and may entail, in some cases, a more radical formal restructuring of the source text without, however, altering its meaning. In The theory and practice of translation (coauthored with Charles Taber, 1969) and in subsequent books, Nida provides a model for describing the translating process and analyzing meaning in more detail. In his view, the translating process involves analysis, transfer, and restructuring. Analysis consists in a reduction of surface structures to kernels (i.e. substructures), making use of the functional classes of objects, events, abstracts, and relations. At the kernel level, languages are found to ‘agree far more than on the level of more elaborate structures’ (1969:39), and it is at this level that the transfer into the receptor language occurs. The transformation into a new surface structure through restructuring takes place according to the deemed expectations of the receivers. The model draws on principles used in transformationalgenerative grammar, although the influence of this linguistic direction on Nida’s work has often been overemphasized. For one thing, Nida reverses Noam Chomsky’s model by starting from surface structures and moving to kernels and not the other way round. Secondly, Nida is interested in language-in-use (Ferdinand de Saussure’s parole or Chomsky’s performance), and not in the abstract level of Chomsky’s deep structures. Thirdly, Nida rejects the idea that languages are strictly rule-governed, as this would leave no space for their creative use. It is nevertheless true that transformational-generative grammar offered him a theoretical perspective for explaining processes of decoding and encoding texts. From a similar stance, he could also account for the similarities between languages on a conceptual and even on a formal level, in keeping with his conviction that effective interlingual

738

communication is always possible, although there is no such thing as absolute communication. In more recent publications in linguistics and translation theory, Nida undertakes detailed analyses at all language levels, broadening the area of interlingual investigations. The author includes a high number of case studies of Bible translations, of scientific and technical texts, of various literary genres, and of European Union texts, and a multitude of experiencebased examples from different languages and cultures illustrating his translation principles. Debates around Nida’s works were mainly generated by ideological and religious speculations on his receptor-oriented position in translation as well as by the principle of ‘equivalence of effect’, on which dynamic/functional equivalence is based. Although the American scholar has suggested a series of tests for checking and comparing the source to target text receivers’ understanding and response, “effect” has been regarded as too vague a notion to serve as a criterion for translation evaluation. Eugene Nida’s approach to linguistics and translation is avowedly eclectic. This encyclopedic perspective enables him to cover all major aspects of languages and their translation. He combines insights from structural linguistics, pragmatics, sociolinguistics, discourse analysis, functional grammar, transformational-generative grammar, semiotics, psycholinguistics, rhetoric, stylistics, information theory, etc. Nida openly declares his mistrust of the holistic systems that ‘can stifle creativity and lock minds shut to new evidence’ (2003:140). His books have aroused considerable interest among linguists, theologians, and translation scholars, and they have certainly been of great help to all those dealing with intralingual, interlingual, and intercultural communication, be it religious or secular.

Biography Eugene Albert Nida was born in Oklahoma City, Oklahoma on November 11, 1914. He received his B.A. in Greek and Linguistics from the University of California at Los Angeles an 1936, and an M.A. in Patristics from the University of Southern California in 1939. He taught at the Summer Institute of Linguistics in 1937–1952. He did his Ph.D. on English Syntactic Structures under the supervision of Charles C. Fries, Professor of English and Linguistics at the University of Michigan, in 1943. He was consultant for the American Bible Society and the United Bible Societies, 1943–1981; President of the Linguistic Society of America, 1968: and Translations Research Coordinator for the United Bible Societies, 1970–1980. He has studied languages and cultures, counseling on Bible

NIGER-CONGO translating in more than 90 countries, and has lectured in more than 100 universities, also actively participating in other scientific and academic reunions worldwide. He also received five honorary Ph.D.s. He is author and co-author of more than 40 books and 250 articles. Nida has lived in Brussels since 1995, consulting with Bible societies and translators from the European Union. References Black, Matthew, and William Smalley (eds.) 1974. On language, culture and religion: in honor of Eugene A. Nida. The Hague: Mouton. Fawcett, Peter. 1997. Translation and language. Manchester: St. Jerome. Joos, Martin. 1972. Semantic axiom number one. Language 48. 257–65. Larose, Robert.1989. Théories contemporaines de la traduction. Québec: Presses universitaires de Québec. Longacre, R.E. 1994. Eugene Nida. The encyclopedia of language and linguistics, Vol. 5, ed. by R.E. Ascher and J.M.Y. Simpson. Oxford: Pergamon Press.

Munday, Jeremy. 2001. Introducing translation studies. London, New York: Routledge. Nida, Eugene Albert. 1946. Morphology. The descriptive analysis of words. Ann Arbour, MI: University of Michigan Press. ––––––. 1964. Toward a science of translating. Leiden: E. J. Brill. ––––––. (co-authored with Charles Taber). 1969. The theory and practice of translation. Leiden: E. J. Brill. ———. 1975. Exploring semantic structures. Munich, Germany: Wilhelm Fink. ––––––. (co-authored with Johannes P. Louw). 1992. Lexical semantics of the Greek New Testament. Atlanta: Scholars Press. ––––––. 1996. The sociolinguistics of interlingual communication. Bruxelles, Belgium: Les Editions du Hazard. ———. 2001. Contexts in translating. Amsterdam and Philadelphia, PA: Benjamins. ———. 2003. Fascinated by languages. Amsterdam and Philadelphia, PA: Benjamins. Schmitz, John Robert. 1999. The contribution of Eugene Nida to the theory and practice of translation. Translation-transition. 15th World Congress of FIT, Mons, Vol. 1, ed. by Jean-Marie Vande Walle. Mons: Elma Edities.

RODICA DIMITRIU

Niger-Congo The Niger-Congo language family, previously labeled Nigritic, or western Nigritic, is one of the four main language families in Africa. The other language families are Khoisan, Nilo-Saharan, and Afro-Asiatic. Niger-Congo languages are spoken in the basins of the Niger and Congo rivers and in the Nuba Mountains in the Sudan. Joseph Greenberg’s (1955) original classification viewed the Niger-Congo family as a sister language of the Kordofanian family. He named the ancestor language from which the two emerged as Niger-Kordofanian. However, Kay Williamson (1989a,b) notes that the Kordofanian family falls within the Niger-Congo family; hence, she uses the label Niger-Congo in place of NigerKordofanian for the whole family. The Niger-Congo group of languages is spoken by well over 80% of the population of Africa. Geographically, they are spoken in areas stretching from Senegal (West Africa) to Kenya (East Africa). They also stretch from Sudan (North Africa) to the south in the Republic of South Africa. Niger-Congo has ten daughter languages, namely Adamawa-Ubangi (formerly Adamawa Eastern), Atlantic (formerly West Atlantic), BenueCongo, Dogon, Gur, Ijoid, Kordofanian, Kru, Kwa, and Mande.

Adamawa-Ubangi Adamawa-Ubanji, previously called AdamawaEastern, is spoken in parts of Central African Republic, eastern Nigeria, northern Cameroon, and southwestern Chad. It has languages like Gbaya, Banda, and Zande. The Adamawa-Ubangi family has two daughter languages: Adamawa and Ubangi. Adamawa has languages like Leko, Duru, Jen, Nimbari, Mbum, Bua, Kim, Day, Waja, Daka, and Fali. Ubangi languages are spoken in the area stretching from northern Cameroon across the Central Africa Republic to parts of southern Sudan and northern Democratic Republic of Congo (formerly Zaire). Ubangi languages include Banda, Gbaya, Ngbaka (Sere, Mba), Sango, and Zande. Sango is a lingua franca of the Central African Republic. While Adamawa allows words to end in consonants, none of the Ubanji languages, with the exception of Gbaya, allows final consonants. Whereas other language families may have nasal sounds, AdamawaUbanji has nasal morphemes (root words). Word order is Subject–Verb–Object, and verb reduplication (doubling of verbs in certain grammatical constructions) is common to all the languages in the subfamily. The

739

NIGER-CONGO languages have between two and four tonal phonemes, i.e. pitch patterns that may alone distinguish meaning.

Atlantic Atlantic languages, also called West Atlantic languages, are spoken in West Africa in the area stretching from the Senegal river down into Liberia. Most languages in this family are spoken in Senegal, The Gambia, Guinea, Guinea-Bissau, Liberia, and Sierra Leone. Atlantic has about 45 languages and over 30 million speakers. Languages in this family with large numbers of speakers include Fula, Wolof, Diola, Serer, Manjaku, Balanta, Basari, Limba, Kisi, Sua, and Temne. Important linguistic features include consonant mutation/alternation (specific consonants may change into another in particular sound patterns), noun class, and concord systems in which the choice of specific parts of speech (especially nouns) requires other elements (verbs, adjectives, adverbs, etc.) of the sentence to occur with particular affixes, depending on the grammatical construction.

(New) Benue-Congo The Benue-Congo language family is the largest subfamily of the Niger-Congo group. It is spoken in the areas in and around the Benue and Congo river basins. It is made up of the former Eastern Kwa––Yoruba, Edo, Nupe, Idoma, Igbo––and Bantoid (subclassified into non-Bantu and Wide Bantu). The New BenueCongo language family is subclassified into eight subfamilies: Defoid (Yoruba, Akoko), Edoid (Edo), Nupoid (Nupe, Ebira, Asu, and Gbagyi), Idomoid (Idoma), Igboid (Igbo), Kainji and Platoid (Kainji, Eloyi, Kagoma, Jukun), Cross River (Obolo, Ogoni, Legbo, Ogbia), and Bantoid (Tiv, Swahili, Kikuyu, Kirundi, Kinyarwanda, Shona, Zulu, and Xhosa). The Bantoid group is by far the largest group in the Benue Congo subfamily. Important linguistic characteristics of this large subfamily include cross-height harmony (where either only tense or only lax vowels occur in words) and vowel coalescence (where two or more vowels merge to become one). Syllable structure ranges from CV (consonant–vowel) to CCV and CVC. Consonant types include such uncommon sounds as breathy voiced plosives /bh dh/ and labio velar sounds /kp gb/. Other important features include noun class and concord systems and an SVO word order. Some of the Bantu languages have a fixed stress pattern––words in Swahili, for example, are always stressed on the penultimate (last but one) syllable.

Dogon Dogon, which is spoken on the mountains of Mali, was classified as a Gur language until 1989. However,

740

it may also be classified as an independent member of the Volta Congo group. Dogon has six major dialects: Donnɔ Sɔɔ, Tombɔ Sɔɔ, Torɔ Sɔɔ, Jamsay, Togo Kan, and TomO Kan. Important linguistic features include independent nasalized vowels (phonemic nasalization of vowels), vowel harmony (co-occurrence restrictions in the distribution of vowels), a two-tone system, and Subject–Object–Verb word order. In Dogon noun phrases, the noun always comes first.

Gur Geographically, the Gur group of languages is distributed over a wide area stretching from the Ivory Coast, through Ghana, Togo, Benin to Burkina Faso. Gur languages are also spoken on the fringes of Niger, Benin, and Mali. The Gur phylum subdivides into Oti-Volta, Bwamu, Kurumfe, Grusi, Kirma, Dyan, Gan, Dogoso, among others. There are nearly a hundred languages in the Gur family. Some of the languages in this group include Moore (spoken in Burkina Faso by nearly seven million people), Grusi, Gurenne, Wali (Dagaari), Dagbani, Buli, Kabre, Kotokoli, Kasem, Konni, Tayiri, Kusaal, Bassari, Ntrobo, Sisaala, Waali, Mampruli, and Nafaanra. Over twenty million people speak the Gur languages. Some of the phonological features found in Gur languages are tone, vowel harmony, uncommon consonants such as implosives, coarticulated sounds (e.g. /kp/ /gb/), and a syllable-timed rhythm.

Ijoid Ijo, the smallest branch of the Niger-Congo language family, is spoken by the Ijo of Nigeria and covers both Ijo and Defaka (Afakani). It is spoken in the Niger River delta region and adjacent riverine areas within the Rivers, Bendel, and Ondo states of Nigeria (Jenewari 1989). The Ijo language family is made up of seven languages: Biseni (Amegi), Okodia, Oruma, Nkoroo, Eastern Ijo (which comprises Kalabari, Okrika, and Ibani), Brass Ijo (Nember-Akassa), and Izon (Bumo, Kolokuma, Mein, and Arogbo). Speakers of Ijo are a little over one million. The Ijo were among the first West Africans to have contact with Europeans, and the Kalabari, an Ijo language, is believed to be one of the first Nigerian languages to be written. An important unique structural feature is a consonant harmony also called ‘implosive harmony’ (which requires that any given word either contains implosives or plosives, but never both), and a noun class system drawing grammatical distinctions based on animateness and biological gender. The basic word order is Subject– Object–Verb, although different word orders are possible (Object–Subject–Verb and Subject– Verb–Object) if object or subject need to be marked as the topic of conversation.

NIGER-CONGO

Kordofanian The Kordofanian language family is located in the Nuba Mountains of Sudan. Its status in the NigerCongo family has been widely debated. As mentioned earlier, it was first viewed as a sister family of NigerCongo in a larger grouping called Niger-Kordofanian, but later it came to be seen as a sub-branch of NigerCongo. Kordofanian has over half a million speakers and consists of about 20 languages. Moro, Mudo, Talla, Miri, Krongo, Talasa (Tumtum), Tiro, Utoro, Rere, Ngile, Tocho, Goy, Gom (Tegali), and Kalak are some of the languages of this subfamily. Kordofanian languages have dental /d, D, t/ (tongue touches teeth) and retroflex consonants /t ¢, d¢ / (tongue curls back). Plosives like /t/ occurring between vowels change to fricatives such as /s/, and the noun class system is reminiscent of those found in other Niger-Congo.

Kru The Kru language family is spoken in southwestern Ivory Coast and southern Liberia. There are 24 languages in this subfamily and the total number of speakers ranges between three and four million. Some of the Kru languages include Grebo, Klao, Dida, Godie, Bete, Nyabwa, Konobo, Bassa, Gbii, Bakwe, Kuwaa, Aizi, Wobe, Dewoin, SEmE, Guere (Krahn, WEE), Tepo, Chedepo, and Neyo. Most Kru languages have eight vowel phonemes and a vowel harmony where only vowels from a particular set may occur in any given morpheme. Kru syllables tend to end in a vowel. Central vowels such as /i, u, a,/ are found in some of the Kru languages. Uncommon consonant sounds such as implosives, double articulated sounds, and velar fricatives are found in the Kru languages. Kru languages have subject–verb–object word order with indirect objects preceding direct objects. When there is an auxiliary, then both the direct and indirect objects precede the verb. Sentences may be negated with the help of an auxiliary or a particle as well as changes in tone pattern.

Kwa Kwa has undergone tremendous reclassification with several languages and language groups such as Ijo, Kru, Yoruba, and others moved to other major language families. The New Kwa, as it is now called, is spoken in Côte d’Ivoire, Ghana, Togo, Benin, and Nigeria by about 30 million people. It subdivides into two main subfamilies, namely Nyo and Left bank. Important subfamilies under Nyo include AvikamAlladian, Agnéby, Potou-Tano, and Ga-Dangme. Gbe, Avatime-Nyangbo, Kposo, and Keby-Animere are identified as members of the Left Bank subfamily. Important New Kwa languages include Akan, Ewe

(and other Gbe languages), Ga, Dangme, Gwa, Avikam, Anyi, Baule, Chakosi, Nzema, Santrokofi, Likpe, Adele, Logba, and Kposo. Important linguistic features associated with some of the New Kwa languages are the occurrence of voiceless and voiced labial and velar fricatives, double articulated sounds (single sounds with two different places of articulations e.g. /kp, tp, gb/ ), vowel harmony, and tone terracing, where pitch is lowered toward the word end.

Mande The Mande group of languages is spoken in West African countries: Senegal, The Gambia, Mali, the Ivory Coast, Guinea, Sierra Leone, Liberia, Guinea Bissau, Burkina Faso, Mauritania, Benin, Niger, Nigeria, and Ghana. Bambara, Dyula, Susu, Mende, Kpelle, Vai, Lorma, Loko, Soninke, Kweni, Dan, Maninka, Kpelle, Busa, Bisa, Ligbi, Togo and Bobo are Mande languages. Over 40 languages are found in this family. The Mande languages are spoken by a little over 20 million people. Phonologically, many Mande languages have between seven and nine vowel phonemes, labiovelar stops, consonant mutation, two tones, and tone sandhi (adjacent tones influence one another). Unlike most African languages, Mande languages do not have serial verb constructions. Word order is Subject–Object–Verb. In noun phrase constructions, definite determiners, articles, and plurals tend to follow the noun. However, possessive pronouns precede nouns. Some of the languages such as Vai, Mende, Loma, and Kpelle in this group have combinations of Arabic script, Latin writing systems, and unique African writing systems. References Bendor-Samuel, John. 1989. Niger-Congo. The Niger-Congo languages. A classification and description of Africa’s largest language family (N-CL). New York: University Press of America. Bendor-Samuel, Olsen, and White. 1989. Dogon. In John Bendor-Samuel, pp. 169–77. Dwyer, David. 1989. Mande. In John Bendor-Samuel, pp. 47–66. Greenberg, Joseph. 1955. Studies in African linguistic classification. New Haven: Compass Publishing Company. Heine, Bernd, and Derek Nurse, 2000. Niger-Congo, N-CL. New York: University Press of America. Jenewari, Charles. 1989. Ijoid. In John Bendor-Samuel, pp. 105–18. Köhler, Oswin. 1975. Geschichte und Probleme der Gliederung der Sprachen Afrikas. Die Völker Afrikas und ihre traditionellen Kulturen, ed.by H. Baumann, 137-373. Wiesbaden: Franz Steiner. Marchese, Lynell. 1989. Kru. In John Bendor-Samuel, pp. 119–40. Mukarovsky. 1976. A study of Western Nigritic. Vienna: Institut für Aegyptologie und Afrikanistik. Naden, Tony. 1989. Gur. In John Bendor-Samuel, pp. 141–68. Samarin, William. 1971. Adamawa-Eastern. Samarin, William.

741

NIGER-CONGO 1971. Adamawa-Eastern. Current trends in linguistics, Vol. VII: Linguistics in Sub-Saharan Africa, ed. by T. A. Sebeok, 213–22. The Hague, Netherlands: Mouton. Sapir, John. 1971. West Atlantic: an inventory of the languages, their noun class systems and consonantal alternations. CTL 7. 45–112. Schadeberg, Thilo. 1989. Kordofanian. In John Bendor-Samuel, pp. 67–80.

Stewart, John. 1989. Kwa. In John Bendor-Samuel, pp. 217–46. Williamson, Kay. 1989a. Niger-Congo overview. In John Bendor-Samuel, pp. 3–46. Williamson, Kay. 1989b. Benue-Congo. In John BendorSamuel, pp. 247–74. Wilson, W. A. A. 1989. Atlantic. In John Bendor-Samuel, pp. 81–104.

SAMUEL GYASI OBENG

Nigeria The Federal Republic of Nigeria is the most populous country in Africa. This ‘giant of Africa’ gained independence from the British on October 1, 1960. Nigeria is located in West Africa. In the south, it is washed by the Atlantic Ocean. Its four neighbors are Nigér Republic in the north, Chad in the northeast, Cameroon in the east, and Benin Republic in the west. Presently, it has a presidential form of government. The centrally located city of Abuja is Nigeria’s new capital. The Niger river and its tributary, the Benue, divide Nigeria into three parts: the northern region or Hausaland, the southwestern region or Yorubaland, and the southeastern region or I(g)boland. The names of these regions mirror the linguistic basis of their nomenclature; Hausa is the primary language of Hausaland, Yoruba is the main language of Yorubaland, and Igbo (or Ibo) is the primary language of Igboland. The importance of River Niger is evident in the name of the country, which is exemplary of a creative linguistic process, which combines the words ‘Niger’ and ‘area’. ‘Nig- area’, spelled NIG-ER-IA, literally refers to the area around the River Niger. Nigeria has the largest number of languages found in any African nation, accounting for over one quarter of Africa’s languages. Nigeria’s linguistic landscape is as variable as its vegetation, which ranges from swampland on the southern coast, to lush tropical rainforests, and a continuum of savannah land––guinea savannah, through sudan savannah, sahel savannah, and finally desert, studded with the mighty baobab. The flora and fauna are distinctive to the region; many plants and trees are not known to the western world; hence, they have become part of the lexicon of Nigerian English. Prime examples are kola (nut), cam wood, the udala tree, bitter leaf (from which bitter leaf soup is made), and the acacia tree. In Things fall apart, a classic bestseller that has sold over two million copies and has been translated into over 50 lan-

742

guages, Achebe utilizes Nigerian English, which echoes the oral tradition the text celebrates. The text is rich in proverbs, metaphor, simile, and folklore, as exemplified by the following excerpts: a. ‘Okoye said the next half a dozen sentences in proverbs. Among the Ibo, the art of conversation is regarded very highly, and proverbs are the palmoil with which words are eaten’ (7). b. ‘A clan was like a lizard; if it lost its tail, it soon grew another’ (171)

Stories of Igboland are recounted and significant beliefs are celebrated, such as the Week of Peace, the New Yam Festival, and the precolonial conception of time in relation to market weeks and the position of the moon. In Nigeria, English serves multifarious functions––in administration, education, commerce, and more. The national anthem and the national pledge, the two most widely recited verses of patriotism, are both in English, and illustrate how Nigeria is conceptualized as both a fatherland and a motherland. As a result of the dynamics of English usage in different strata of Nigerian society, at least three varieties of English are discernable: the acrolect, the mesolect, and the basilect (see Bamiro 1991; Pandey 1997). Many Nigerians now use Americanisms. Multilingualism, with its accompanying problems as regards the selection of a national language, has ‘forced’ English––a neutral code––to function, in many ways, as the quasinational language (Emenyonu 1989:83). English is now a permanent resident in Nigeria, a primary Nigerian language in all senses of the term. In Achebe’s words, ‘It has been bent and twisted to bear the burden of the African experience’ (1975). It is a Nigerian brand of English (Ubahakwe 1979; Akere 1982; Bamgbose 1982; Odumuh 1987), with a distinctive local flavor. As Akere (1982) puts it,

NIGERIA What has happened here in Nigeria… is that the resources of a second language are superimposed on an intricate system of social and kinship relationships, and on a completely different pattern of cultural outlook and social expectations. …The English forms of address and greetings have become modified to suit local communicative needs. (97)

The educated Nigerian generally has access to the largest language repertoire and even initiates switches to the stigmatized Nigerian Pidgin (English) in conversations with peers and lesser-educated Nigerians. This is a marked difference between the ‘Outer Circle’ (see Kachru 1982) to which Nigeria belongs on account of the variety of English it has grown, and the ‘Inner circle’, where it is rather unusual to find college-educated individuals being the most multilingual and multidialectal citizens who readily and willingly switch to nonstandard varieties in exchanges with lesser-educated citizens (see Pandey 1999). Not surprisingly, code-mixing, code-switching, and speech accommodation are frequently used conversational strategies, even by Nigerian creative writers whose language use mirrors the conversational realities of Nigerian society (see Pandey 1997). Occupying approximately three fifths of Nigeria’s land area, the Niger-Benue river basin is a major source of sustenance. Its waters have also nourished and transported a variety of water-borne languages, namely, Pidgins and Creoles, and specifically Nigerian Pidgin English (NPE hereafter), the nation’s linguistic lifeline or primary lingua franca. NPE still bears linguistic evidence of the earliest European contact in Nigeria. The Portuguese are known to have arrived on the coast of West Africa in the fifteenth century. Exemplary Portuguese words that constitute part an parcel of the core vocabulary of this linguistic medley (NPE) include sabi (comprehend, understand), pickin (child or children), palava (trouble), and dash (a gift or bribe, including the act of giving). The excerpt that follows, an Efik Chief’s diary entry, provides some of the earliest evidence of the development of a pidginized interlanguage in Calabar: ‘about 6 am in aqua Landing with small Rain morning so I walk up to see Esim and Egbo Young so I see Jimmy Henshaw come to see wee and wee tell him for go on bord …’ (Forde 1956). Around Lokoja, a confluence town, different varieties of pidgin are spoken in a relatively small area. The same has been recorded for the Delta and Rivers regions, where several varieties of NPE have been documented (Marchese and Shnukal 1982; Faraclas 1985). Opinions differ regarding the conjectured expansion of NPE into a Creole in some parts of Nigeria (Donwa-Ifode 1983). Even though English is Nigeria’s official language, NPE is the preferred lingua franca; hence, it should

come as no surprise that NPE is the most widely spoken pidgin in the world, with over 40 million estimated speakers, not all of whom are Nigerians (Faraclas 1996:1). Some speakers are from neighboring Cameroon and Ghana––both of which witnessed a ‘brain drain’ (the Nigerian English term for intellectual loss) or an exodus of immigrants to Nigeria––during Nigeria’s oil boom. For the most part, these and other expatriates have acquired NPE. In the absence of a standard orthography, NPE is written the way it sounds (to the speaker). Faraclas (1996) observes that the name NPE is actually a misnomer, as it has developed trilectal varieties: a pidginized basilectal variant often referred to as ‘Pidgin Propa’ (i.e. ‘Pidgin Proper’ or PP) and frequently used by less educated or uneducated Nigerians; a creolized ‘mesolect’; and a decreolized variant that is often used by educated Nigerians. Eze (1980) refers to this variety as ‘Hyperanglised Pidgin’ (HPP) and notes that it differs extensively from ‘Pidgin Propa’ in its more-Standard-English-oriented use of prepositions, as opposed to the use of ‘for’ to represent all prepositional functions in ‘Pidgin Propa’. There are, of course, other differences, such as the more economical lexicon of ‘Pidgin Proper’. Examples include: (1) a. Im mother been come meet am in the office (HPP) b. Im moda been come see am for (im) office (PP) [Gloss: His/Her mother came to meet or see him/her in the office OR His/Her mother DID (indeed) come and meet him/her at the/his/her office]

(2) a. ’The thing pained me bad bad because I wanted to be big man like lawyer or doctor riding car and talking big big English’ (HPP, excerpted from Saro-Wiwa’s Sozaboy, p. 2) [Gloss: The thing/It hurt me very badly because I wanted to be an accomplished and successful person, like a doctor or a lawyer, driving a nice car and impressing people with my command of the English language]

b. E/De ting vex me bad bad. I wan be big man like lawya or docto wey get car and tok big big English (PP) (3) ‘Everyting scatta’ (from ’Zombie’) [Gloss: Everything is scattered/is in disarray] (4) ‘I look the D.O. well well. Then I tell am say no be tallness go fight the war…’ (from SaroWiwa’s Sozaboy, p. 27) [Gloss: I looked at/stared at the D.O. for a long time. Then I told him that when it came to fighting the war, it didn’t matter how tall you were]

743

NIGERIA (5) E get one pikin, I/Ai get four (pikin ) [Gloss: She/he has one child; I have four (children)] (6) ‘If you neva ready to carry the load, O’, why put am for another person head?… Dis kain life, na so so wahala’ (lyrics excerpted from ‘Wait for Me’ composed by King Sunny Ade and Shina Peters) [Gloss: If you aren’t ready to shoulder your responsibility, why burden others with it? That approach is full of troubles] (7) ‘No money, woman no go follow you, even if you fine fine pass everybody’ (from SaroWiwa’s Sozaboy, p. 21) [Gloss: If you have no money, women won’t come after you, even if you are the most handsome man in the world]

(8) Wettin you de talk? (What are you talking about? OR What thing are you referring to?) Nonetheless, some core features that different varieties of NPE share include an anglicized vocabulary: African grammatical and discourse structures, the nonuse of tense markers, as in (3) and (4), the zero copula and zero distinctions in gender and number (see (1b)), zero genitive forms (see (6)), reduplication––a feature of many African languages (see (4) and (7)), and multiple meanings associated with a single word (e.g. the word chop, which means consume(d), eat(en), ate, digest(ed), food, a snack, and a bribe). Other similarities include a mixed vocabulary that contains many words from Nigerian languages, reduced or zero articles, and demonstratives; communicative contractions such as ‘tori’ for ‘story’, ‘gree’ for ‘agree’, and ‘kain’ for kind”; communicatively expressive compounds like ‘sofahead’ (a creative term used to refer to one who worries too much), and coalesced creations, such as ‘likam’ (like him/her, likes him/her), ‘wey’ (who, whose, where), and ‘wettin’, a combination of ‘what’ and ‘thing’, as in (8). Like Nigerian English, NPE is particularly creative, and echoes the rich Nigerian oral tradition (over 2,000 years old). Proverbs are a vital ingredient and color this code in a refreshing manner. Examples include: Who go dash monkey banana? (Gloss: Who gives monkeys bananas out of sheer goodwill?/There are no free lunches) and Kaki no be leda (Gloss: Khaki is not leather; in short, it’s like comparing apples and oranges). Sayings in NPE enable their speakers to succinctly communicate a whole lot. In a culture where economy, metaphor, and wit are prized in conversation, this code is valuable. Because of its function as the language of the people, many Nigerian literary artists utilize NPE in their works (Lindfors 1974). In fact, for a long time, NPE served as a safe outlet for political expression, particularly during military regimes (Eka 1999). Exemplary works include Sozaboy: A novel in

744

rotten English written by the recently hanged Nigerian writer Ken Saro-Wiwa, who, although not a linguist, made the following sociolinguistic observation: Both ‘High Life’ and Sozaboy are the result of my fascination with the adaptability of the English Language and of my closely observing the speech and writings of a certain segment of Nigerian society. For, as Platt, Weber, and Ho accurately observe in their book The new Englishes (1984), ‘In some nations … the New Englishes have developed a noticeable range of different varieties. … ’ Sozaboy’s language is what I call ‘rotten English’, a mixture of Nigerian pidgin English, broken English, and occasional flashes of good, even idiomatic English. This language is disordered and disorderly. … It thrives on lawlessness, and is part of the dislocated and discordant society in which Sozaboy must live, move, and have not his being (‘Author’s Note’). Works by other Nigerian artists that are entirely or predominantly in NPE include ‘Zombie’, lyrics by Fela Kuti, one of Africa’s foremost musicians, and No longer at ease in which the characters’ use of NPE both mirrors and intensifies the neocolonial discord Achebe depicts through his artful use of varieties of Nigerian English. Some of Africa’s best-known writers, playwrights, and musicians come from Nigeria, and their works have been instrumental in the development of Nigerian English. Several are world-renowned and have contributed substantially to both African literature and world literature. In 1986, Wole Soyinka became the first African to receive the Nobel Prize in literature. Chinua Achebe was awarded the Nigerian National Merit Award, Nigeria’s highest award for intellectual achievement. Arrow of God won the New StatesmanJock Campbell Award, and Anthills of the Savannah was the finalist for the 1987 Booker Prize. Nigerian artists such as Chinua Achebe,Wole Soyinka, Ola Rotimi, Amos Tutuola, John Pepper Clark, Segun Oyewole, Tunde Fatunde, Christopher Okigbo, Gabriel Okara, and Femi Osofisan have been known to utilize NPE for literary effect––to color their characters or the plot (e.g. Amos Tutuola’s The PalmWine Drinkard), to add humor (e.g. Soyinka’s Jero’s Plays, Segun Oyewole’s Katakata for Sofahead ), sarcasm, local flavor, or to give the text a linguistic facelift (see Pandey 1997). NPE is also frequently used in Onitsha Market literature, including well-known plays like Ogali Ogali’s Veronica my daughter that has sold over 250,000 copies since 1956, and short stories written by a host of Nigerian pamphleteers (Sander 1980). Ogali uses NPE or ‘uncooked English’, as Veronica calls it, to portray her traditional father (Chief Jombo) and Chief Bassey, and Standard English to portray the other characters. Bomber Billy speaks a bombastic variety of English, and is ridiculed by the writer and by

NIGERIA one of the characters in his play, Paulina. She describes him as ‘a negligible pocket radio that utters useless words’. In Sander’s (1980) opinion, Ogali is making fun of those people in Nigerian society who, eager to demonstrate their mastery of the English language, go to any extremes in the use of polysyllabic words or the creation of new words. … Through the use of various levels of English, Ogali manipulates the response of the audience (x).

In short, NPE is frequently utilized in Nigerian literature, and its continued use contributes to its popularity. As far back as 1979, NPE was observed to have replaced Efik as the lingua franca of the Cross River area (Dunstan 1969:35). Being a West African pidgin, NPE is of vital importance to the study of creolistics, because of its historical role in the development of anglophone pidgins and creoles. In the words of Nicholas Faraclas (1996), ‘Nigerian Pidgin can be considered to be one link in a chain of English-lexifier pidgins and creoles spoken along the coast of West Africa and in African diaspora communities throughout the Atlantic Basin’ (1). Advocates of the Afrogenesis hypothesis, which traces the roots of pidgins and creoles to Africa, look to West African pidgins and creoles for answers (Romaine 1994; Dalphinis 1985). According to one view, ‘Created in West Africa, … work pidgins were transported across oceans where they took on new roles as lingua francas among the enslaved. …’ (McWhorter 1984:240). In the case of NPE, as well as Nigerian English, many of the ingredients are African. These include the names of food (akara balls, fried bean cakes; moimoi, steamed bean cakes; dodo, fried plantains; gari, powdered cassava; Jollof rice, rice flavored with Nigerian peppers; masa; millet cakes eaten in Hausaland; pounded yam, egusi soup, etc.); clothing (agbada, the flowing robes that Yoruba men wear; buba and wrapper, the women’s traditional outfit); local customs and festivals (naming ceremonies for newborns); regional celebrations (e.g. Masquerade Day, a tribute to the ancestors in the Ekiti region); and expressions of linguistic pride such as ajibotta, the well-known derogative Yoruba word for a butter eater or westernized Nigerian (the Nigerian English equivalent is been-to). Exemplary Yoruba words in NPE include Oga (a respectful term, roughly equivalent to ‘Sir’), katakata (major disturbance or problem), and the honorific O’. Some Hausa words that have entered NPE include wahala (trouble) and haba (Goodness! or Listen up!). Igbo words in NPE include sef (a discourse marker), yanga (boasting or boastfulness), and na (a particle with multiple meanings, such as ‘it is’, and the attention-getter Na wa, O’!, a popular expression of shock).

The vast majority of Nigeria’s languages––including Yoruba and Igbo, two of the three major languages––belong to the Niger-Congo family. Estimates vary regarding the exact number of indigenous languages thriving in Nigeria. While Bamgbose (1991:2) and McArthur (1992) put the number at 400, Katzner (1995) estimates that there are ‘about 250’ (352). While we cannot rule out the threat of language endangerment arising from the central place that English occupies in Nigeria’s linguistic horizon, the discrepancy in estimates has a lot to do with differences in opinions regarding what constitutes a language and what constitutes a dialect. For instance, Ibibio, Efik, and Annang, being mutually intelligible, are viewed by many Nigerians as one and the same language (Dunstan 1969:35), namely Efik/Calabar (or Ibibio to some Nigerians), which has acquired the status of a Standard dialect, on account of its use in broadcasts and textbooks. Some Nigerians do, however, regard them as separate languages. According to some, Ikwer(r)e/Ikwerri has 12 dialects. Many nonIbibio speakers regard Qua, Oron, Eket, Okobo, and Ibuno as dialects of Ibibio, while others view them as lesser-used languages. While some view Kolokuma and Ijo as dialects of the same language, others might not, as Nigerian codes of communication have distinct names. Most have several names. Isoko, for instance, is also called Biotu, Sobo, and Igaba; yet, some consider the last two names to be offensive. To most Nigerians, dialects tend to coincide with geographical boundaries or specific urban locations. Chief dialects identified for Eksako, for instance, are said to be those of Auchi, Aviele, Ekperi, South Ibie, Uzairue, and Weppe Wano––names of towns and cities, and of specific ‘clans’ (Dunstan 1969:47). The three main dialects of Fula/Fulani found in Nigeria also coincide with the names of major cities or geographical terrains, namely Sokoto, Adamawa, and the central dialect of Kano, Bauchi, Gombe, Katagum Division, and Jos (Dunstan 1969:57). Most educated Nigerians speak at least three languages, English (the official language), one or more of the major regional languages, a local language, and NPE, not to mention the other dialect(s) they might switch to in the middle of a conversation. For most Qua speakers, for instance, Efik is a second language, Igbo the third, NPE the fourth, and (Nigerian) English the fifth. In Nigeria, ethnicity and linguistic identities are inextricably intertwined. McArthur (1992) recognizes eight major ethnic groups, namely 21% Hausa, 21% Yoruba, 18% Igbo, 10% Fulani, 6% Tiv, 5% Kanuri, 5% Ibibio, 4% Edo, and 10% minorities. Because of the close ties between the Hausa and the Fulani, it is not unusual to find these groups conjoined under one

745

NIGERIA label, namely, the Hausa-Fulani. The minority ethnic groups would include, among others, the Ogoni, Isoko, Izon, Kwale, Degema, Itsekiri, Ibuno, Izi, Ewe, Andoni, Amo, Angas, Birom, Chip, Chella, Efik, Mada, Nabor, Nembe, Tera, Yergam, Gwari, and Gonja peoples of Nigeria. The names of their languages tend to coincide with their ethnic descriptors, further emphasizing the linguistic basis of ethnic identity in Nigeria. Each ethnic group has its own unique customs and language, making for a diverse group of Nigerians; yet, the minor languages tend not to be as carefully studied. Some (Fula, Tiv, Igala, Efik, and Edo) enjoy a more functional status than others, because of their use in the media. The nation’s linguistic and cultural diversity has a lot to do with its historical location at the heart of transcontinental migration routes. The greatest concentration of languages and ethnic groups is found in Central Nigeria. One has only to travel to Kwara State or Illorin to get a feel for the linguistic diversity. The most widely spoken indigenous language is Hausa (Katzner 1985; McArthur 1992). This is followed closely by Yoruba (18% of the population or some 20 million speakers), then Igbo (11% or some 15 million speakers), Fulani (approximately 8 million speakers), Kanuri (3 million), Efik and Ibibio (3 million), Tiv (spoken around Gboko by some 2 million), Ijo or Ijaw (2 million), Edo (one million), Urhobo (half a million), Nupe (in the Gur subfamily of NigerCongo languages, with roughly 500,000 speakers), Idoma (some 250,000 speakers), and Eksako (also spelled Esako and Etsako(r), an Edo language with some 120,000 speakers). These numbers are more estimates than up-to-date survey findings of ‘native speakers’ of these languages, as representative language surveys are hard to administer in this complex multilingual environment, and such surveys have rarely been administered (see Bamgbose 1995:34). Hausa, a Chadic language that contains many words borrowed from Arabic, is spoken by roughly 27% of the population or some 40 million Nigerians. It is the primary language or mother tongue of at least 25 million Nigerians (Katzner 1985:288). The Kano dialect is regarded as the Standard variety. The number of Hausa speakers has clearly grown; in 1979, an estimated 12 million Nigerians were native Hausa speakers (Dunstan 1969:73). Hausa is the lingua franca for the bulk of northern Nigeria and also plays the role of a regional language in West Africa. It is, in fact, the most widely spoken African language in West Africa, with speakers found in neighboring countries, including Nigér (as many as five million speakers), Ghana, Togo, Benin, Mali, and Senegal. The Hausa Language Board, established in 1955, has assisted in the standardization of the language. Prior to independence,

746

Hausa enjoyed official language status alongside English in what was then the Northern Region. At that time, Hausa was written in the Arabic script known as Ajami, but now the Roman script is used. Yoruba, a paradigm Kwa language and a subgroup of the Niger-Congo family, is spoken in southwestern Nigerian, from Illorin (Kwara State) down to Akure, Ado-Ekiti, and Igede-Ekiti (Ekiti State), Ibadan (Oyo State), and Lagos. It is Nigeria’s second major language. This language and the rich mythology (of Sango and Ogun) constitute seminal ingredients in many of Wole Soyinka’s works. The mythology is also resonant in the Afro-Caribbean beliefs transported to the Caribbean (see Burnett 1986). Some 20 dialects of Yoruba can be heard in Nigeria, and one, ‘based largely on the Oyo dialect’ (Dunstan 1969:80), is the Standard. Yoruba is also spoken in some other West African nations, including parts of Ghana, Togo, and Benin. Tonality is mirrored in its orthography. Three letters and their corresponding sounds are noteworthy: o, pronounced as in ‘sought’ (e.g. oba); é pronouced as in ‘let’ (e.g. éjé which means blood); and s, pronounced /sh/ as in ‘ekuse’ (i.e. Well done!). In orthography, the acute and grave accents indicate tone (not stress). Igbo, another Kwa language like Yoruba and Ewe, is Nigeria’s third major language. It is spoken in southeastern Nigeria, in Port Harcourt, Enugu, Calabar, Onitsha, Owerri, and surrounding areas. The three main dialects that are recognized are the Central, Owerri, and Umuahia dialects. Prior to 1961, its orthography was disputed (Dunstan 1969:85). The official orthography, adopted by the Onwu Committtee, is now compulsory in the School Certificate Igbo examination. Igbo features regularly in the works of Chinua Achebe. In Things fall apart, some 37 Ibo words and phrases are woven into the fabric of the text. Many have no translation equivalents, including chi (personal god), ogbanje (which Achebe describes as a ‘changeling’ that will keep dying until its ‘iyi-uwa’ is first dug up and destroyed), uri (a betrothal day, after the bride-price has been paid), Nne (mother, including one’s nonbiological mothers in a polygamous household), umunna (kinsmen), and ndichie (one’s elders). These indigenous words bring the richness of precolonial Igbo society to life. From a sociolinguistic standpoint, they constitute prime examples of the Sapir–Whorfian hypothesis, which illustrates how specific cultural concepts are embodied in words. Fula(ni) is generally termed Fulfulde in Nigeria. This Niger-Congo language (the West Atlantic branch) of the Fulani/Fulbe people is of great historical significance, as the Fulani are ‘a people of great antiquity’ (Katzner 1995:289) and have heavily influenced the sociolinguistics of northern Nigeria, where over eight

NIGERIA million of the estimated 15 million Fulani people reside. Many are cattle-herders, and follow the nomadic and seminomadic lifestyle they are accustomed to. Areas in which Fulani is spoken in Nigeria include Sokoto, Kano, Kaduna, Plateau, Bauchi, and Kwara States. In Mauritania, Gambia, Benin, Ghana, Mali, Burkina Faso, Chad, Nigér, and Cameroon, where it is also spoken, it is called Fula. In Senegal, it is called by its other name, Pulaar, and in GuineaBissau, it is termed Pulle. The root morpheme in all these names is {Ful-} or {Pul-} from which the Germanic name Fula(ni) and the Frenchified ‘Peul’ originated. Its regional spread makes it a frequently used lingua franca in West Africa. The national language question has received quite a bit of attention since the mid-1970s. In Language and the nation, Ayo Bamgbose observes that: The language question in Sub-Saharan Africa arises from the fact that not only are most of the countries multilingual, the colonial experience has led to the importation of foreign official languages which have taken on the roles of national communication, unification, administration and medium of education from early or late primary to university level. Thus, the existing multilingual situation is compounded by the addition of imported languages whose strength does not lie in numbers of speakers, but rather in the superior roles assigned to them (1).

During the colonial period, western education was restricted to only a small segment of the population that played a role in the ‘indirect rule’ in place; hence, even today, only the elite uses Standard English. British administrative policy in Nigeria relied on empowered Nigerians––primarily Chiefs and Emirs––to serve as intermediary rulers. English usage and English-medium western education were most widespread in parts of the country where Christian missions were established (Ajayi and Crowder 1985:68). Islam (roughly 47% of the population) and Christianity are the two primary religions, with an estimated 19% of the populace being ‘traditionalists’ (McArthur 1992). As Christianity was resisted in Islamic northern Nigeria, where the Jihad or Holy War led by Uthman dan Fodio in the early nineteenth century had left its religious mark, the discrepancy in English usage between northern and southern Nigeria has persisted to this day. So pronounced were the differences on the eve of independence that there was even a fear that those with an English-based Western education ‘would dominate the economic and political life’ (Ajayi and Crowder 1985:68) of the new nation. The establishment of major regional universities can be attributed, in part, to this regional-linguistic competition. As the University of Ibadan, Nigeria’s oldest institution, was garnering a lot of (inter) national attention, many Nigerians saw the need for a university in Igboland and another in Hausaland. At the wake of

independence, then, several universities were established in Nigeria, most notable among them The University of Nigeria in Nssuka, eastern Nigeria, Ahmadu Bello University, located at the heart of Hausaland, and the University of Ile-Ife, located in Ife, a very important city in Yorubaland, and home of the Oba of Ife, one of the most highly regarded Chiefs in the Yoruba chieftaincy. Ahmadu Bello University (better known as ABU) was named after (Sir) Ahmadu Bello who was northern Nigeria’s first Premier. A resident of Hausaland, and a well-known political figure, northerners (called ‘Gambaris’ by some southerners) named the University after him, as a tribute to his accomplishments. The University is located in Zaria, a famous northern city that played a pivotal role in the intricate emirate system that was in place prior to the colonial period. English and NPE usage declines steadily as one moves north of the town of Illorin. Instead, one tends to hear more Hausa and Fulani in everyday conversations and even in service-counter exchanges. Koranic schools and after-school Arabic classes are commonplace in the north, where Arabic is a language of great (religious) importance, although rarely used in conversations. In contrast, in the open markets in the Yoruba-speaking Ekiti area or in Onitsha, which is famous for its market, NPE is the preferred transactional code, ideal for bargaining. So important is NPE in the markets of Yorubaland and Igboland that mastery of NPE gives buyers a competitive advantage over non-NPE speakers, as haggling is a culturally acceptable and expected discourse practice. One who does not speak the language of the seller is usually at a disadvantage. Some vendors can be heard speaking what is called ‘Broken English’ in Nigeria, and mainly in their limited-but-meaningful exchanges (e.g. ‘Oga, buy tomato! Good Good!’) with the Oy˜ιbo(s), the Yoruba term used to identify outsiders or non-Nigerians. Some traders and children use the compound term ‘Oy˜ιbo man’. This term literally means ‘white man’ or ‘white person’, but its use is not limited to Europeans alone. Also, it is not intended as an insult, particularly when it is followed by the respect marker ‘O(h)’, although it could easily be misconstrued as such, particularly when one is surrounded by a crowd chorusing ‘Oy˜ιbo O’’. The term should be interpreted as more of a congregational announcement of the presence of a visitor whom one would like to honor by acknowledging. ‘O’’ is an honorific term in Yoruba, which means roughly ‘respectful Sir or Ma’am’, and has been adopted by NPE. The equivalent term for expatriates in Hausa is ‘Baturé’ (masculine) or ‘Baturiya’ (feminine). It would be a mistake to view Nigeria as an ESL environment, because it is an English-as-an-indigenized-language context (Bamgbose 1995). Even

747

NIGERIA though English is the medium of instruction from third grade onward, most Nigerians speak Nigerian English, which differs from Standard British or American English. In the past 10 years, a matter of grave concern to Nigerian educators has been the ‘mass failure syndrome’ (Bamgbose 1995:130–52) or the increasing number of failing grades in English obtained in the main college-entry examinations, namely, the WAEC or the West African School Certificate Examination, and the JAMB or the Joint Admissions and Matriculation Board Examination. These examinations are patterned along the colonial, British variety, yet the students hear and learn a Nigerian variety. Many of the idioms that students are tested on are, in fact, foreign to the culture and environment. Examples include the following italicized items: a finger in every pie, comparing apples and oranges, looking for a needle in a haystack, over the grape vine, and penny wise, pound foolish. Students are often tested on noncount nouns like information, even though most Nigerians pluralize these. The end result is a mismatch between the variety of English that is learned, spoken, heard, and taught, on the one hand, and what is tested. Admission into college is contingent upon a credit pass in English, and yet a curricular change is unlikely, as administrators do not recognize Nigerian English. In a report submitted to the Federal Ministry of Information, the Nigerian Public Service Review Commission drew attention to the need for a national language: An overriding problem, which affects the public service as it does all aspects of society, is that of language. Nigeria shares with many developing, and some developed nations, the lack of an indigenous lingua franca. What this means for efficiency in the conduct of government business is rarely even thought about perhaps because there seems to be no immediate answer. But it is perfectly clear to the careful observer that below the top-most levels in the various sectors of society, people are conducting their business in a language which, in varying degrees, they have not in fact mastered (6).

Since that time, some of the major motions that have been tabled include the idea of a trilectal national language policy, a rotational language policy (which also draws on the three regional languages, one by one), and the creation of a single conglomerate language, like Esperanto, which would be called WAZOBIA, and which would combine elements of Hausa, Yoruba, and Igbo. Another idea that has received considerable attention is the pan-Africanist idea of adopting the most widely spoken African language, namely, Swahili. This proposal has been led by Wole Soyinka. In the last few years, French has been added to the list of contested languages. In general, Nigerians in favor

748

of a national language have advanced the idea of an indigenous national language. Special programs have been designed to reduce tribalism and foster national unity. One such measure is the Nigerian Youth Corp Service, better known as the NYSC, which is a year-long commitment required of all graduates. It requires them to serve in a region that is linguistically different from their own, so that they learn to be more tolerant of other ethnic and linguistic groups. This year of service almost always ensures a process of assimilation. Most ‘youth corpers’ end up acquiring yet another language––usually a major regional language––during their NYSC year. The prop on which this program rests is essentially sociolinguistic. Another multilingual measure is the attempt to revive the local languages, particularly the three major languages that constitute the cornerstones of Nigeria’s linguistic ‘stool’ (a symbol of the seat of government in the precolonial Chieftaincy era). This includes making them instructionally viable (Rufa’i 1991; Arohunmolase 1998, 1999) through the development of textbooks, teacher preparation, and so on. The WAEC now offers certification examinations in some Nigerian languages. References Achebe, Chinua. 1958. Things fall apart. London: Heinemann. Achebe, Chinua. 1964. Arrow of God. London: Heinemann. Achebe, Chinua. 1987. Anthills of the Savannah. London: Heinemann; 2nd edition, 1989. Achebe, Chinua.1975. Morning yet on creation day. New York: Doubleday Press. Ajayi, Ade, and Michael Crowder (general eds.) (E. Dunstan, linguistics ed.) 1985. Historical atlas of Africa. Cambridge: Cambridge University Press. Akere, Funso. 1982. Sociocultural constraints and the emergence of Nigerian English. New Englishes, ed. by J.B. Pride, 85–9. Rowley, MA: Newbury House. Altback, Phillip, G. (eds.) 1999. Publishing in African languages: challenges and prospects. Asmara; Trenton, NJ: Africa World Press, Inc. Amkpa, Awam. 1998. Nigerian dramatic literature: the examples of Femi Osofisan and Tunde Fatunde. The growth of African literature: 25 years after Dakar and Fourah Bay, ed. by Edris Makward, et al. Asmara: Africa World Press, Inc. Arohunmolase, Oyewole (ed.) 1998. Nigerian languages for national development and unity. Trenton, NJ: Africa World Press, Inc. Avery-Coger, and Greta M.K. McCormick (eds.) 1988. Index of subjects, proverbs, and themes in the writing of Wole Soyinka. New York: Greenwood Press. Bamgbose, Ayo. 1982. Standard Nigerian English: issues of identification, In Kachru, pp. 148–61. Bamgbose, Ayo. 1991. Language and the nation: the language question in sub-Saharan Africa. Edinburgh: Edinburgh University Press. Bamgbose, Ayo, et al. (eds.) 1995. New Englishes: a West African perspective. Ibadan: Mosuro. Burnett, Paula. 1986. The Penguin book of Caribbean verse in English. Harmondsworth: Penguin Publishers.

NILO-SAHARAN LANGUAGES Dalphinis, Morgan. 1985. Caribbean and African languages: social history, language, literature, and education. London: Karia Press. Donwa-Ifode, S. 1983. Is Nigerian Pidgin English Creolizing? Journal of the Linguistic Society of Nigeria 2. 199–203. Dunstan, Elizabeth (ed.) 1969. Twelve Nigerian languages. New York: Africana Publishing Corporation. Eka, David. 2000. Aspects of language in Ken Saro-Wiwa’s Sozaboy: a novel in rotten English. Before I am hanged, ed. by O. Okome. Trenton, NJ: Africa World Press, Inc. Emenyonu, Ernest. 1989. National language policy in Nigeria: implications for English teaching. Language planning and English language teaching, ed. by C. Kennedy, 82–91. Oxford: Oxford University Press. Eze,Smart, N. 1980. Nigerian Pidgin sentence complexity, Band 8. Wien, Germany: Afro-Publishers. Faraclas, Nicholas. 1985. Rivers Pidgin English: tone, stress, or pitch-accent language? Language in African culture and society. Studies in the Linguistic Sciences 14(2), ed. by E. Bokamba. 67–76. ––––––. 1996. Nigerian Pidgin. New York: Routledge. Fatunde, Tunde. 1985. No food, no country. Benin City, Nigeria: Adena Publishers. Forde, C. D. 1956. Efik traders of old Calabar. London: International African Institute. Kachru, Braj (ed.) 1982. The other tongue: English across cultures. Urbana: University of Illinois Press; 2nd edition, 1992. Katzner, Kenneth. 1995. The languages of the World. New York: Routledge. Lindfors, Bernth (ed.) 1974. Dem-say: interviews with eight Nigerian writers. Austin, TX: University of Texas at Austin Press. ———. 1997. African textualities: texts, pre-texts, and contexts of African literature. Trenton, NJ: Africa World Press. Maja-Pearce, Adewale (ed.) 1994. Wole Soyinka: an appraisal. Ibadan: Heinemann. Marchese, L., and A. Shnukal. 1982. Nigerian Pidgin English of Warri. Journal of the Linguistic Society of Nigeria 1. 213–19.

McWhorter, John, H. 1984. The missing Spanish Creoles: recovering the birth of plantation contact. Berkeley: University of California Press; 2nd edition, 2000. Odumuh, E. 1987. Nigerian English. Zaria: Ahmadu Bello University Press. Osofisan, Femi. 1975. A Restless run of locusts. Ibadan: Onibonoje Press. Oyekunle, Segun. 1983. Katakata for Sofahead. London: Macmillan Education. Pandey, Anita. 1997. The Pragmatics of Code Alteration in Nigerian English. Studies in the Linguistic Sciences 25(1). 75–117. Pandey, Anita. 1999. Code alteration and Englishization across cultures: cyclic differences. Paper invited for presentation at the International Association for World Englishes (IAWE) Convention, Singapore, and invited for inclusion in The Three Circles of English by the Conference Organizer, Edwin Thumboo, University of Singapore. Romaine, Suzanne. 1994. Language in society: an introduction to sociolinguistics. Cambridge: Cambridge University Press. Rufa’i, Abba (ed.) 1991. Twentieth Year Commemorative Symposium of the Center for the Study of Nigerian Languages (1989: Bayero University), Nigerian Languages Yesterday, Today, and tomorrow: Proceedings of the Symposium. Kano, Nigeria: Bayero University Press. Sander, Reinhard. 1980. Veronica my daughter and other Onitsha plays and stories by Ogali Ogali. Port Harcourt: Three Continents Press, Inc. Saro-Wiwa, Ken. 1985. Sozaboy: a novel in rotten English. Port Harcourt: Saros Publishers. Ubahakwe, Ebo (ed.) 1979. Varieties and functions of English in Nigeria. Ibadan: African Universities Press.

ANITA PANDEY See also Bilingualism; Code-Switching; Hausa and Chadic Languages; Igbo and Igboid Languages; Niger-Congo; Official Language Selection; Pidgins and Creoles; Sapir–Whorf Hypothesis; Yoruba and Yoruboid Languages

Nilo-Saharan Languages The idea of a Nilo-Saharan (NiSa) phylum among African languages alongside Afro-Asiatic, NigerCongo, and Khoisan was brought into existence by Joseph H. Greenberg (1963). On the basis of 161 lexical and 29 grammatical sound–meaning similarities, he suggested genetic unity among 82 languages, which had been treated as 22 separate language units before by Tucker and Bryan (1956, 1966) and for some of which genetic unity had never been suggested (Table 1). On the level of language units as presented in Tucker and Bryan (1956), Greenberg proposed some

significant changes, most of which have been accepted since and partly also substantiated in detailed studies: (a) The inclusion of Mimi into the Maban group. (b) The combination of Tucker and Bryan’s ‘Nilotic’ and ‘Nilo-Hamitic’ into one group with three branches: Western, Eastern, and Southern Nilotic. (c) The combination of Moru-Mangbetu and Bongo-Bagirmi into a Central Sudanic group. Greenberg’s fusion of Gule and Koma into one group, however, has been revised. Bender (1996) suggests again two units although differently subdivided, i.e.

749

NILO-SAHARAN LANGUAGES TABLE 1

Language Groups of Tucker and Bryan (1956) and Greenberg (1963)

Tucker and Bryan (1956)

Greenberg (1963) 1. Songhai

2. Eastern Sudanic

2. Saharan: (a) Kanuri, Kanembu; (b) Teda, Daza; (c) Zaghawa [today also: Beria], Berti

10. Mimi 11. Maba group

3. Maban: Maba, Runga, Mimi (of Nachtigal), Mime (of Gaudefroy-Demombynes)

13. Fur

4. Fur

20. Nubian group 27. Didinga-Murle group 21.Barea 24. Tabi 15. Nyimang group 16. Temein group 12. Tama group 14. Daju group 30. Nilotic 31. Nilo-Hamitic 29. Teuso 1. Moru-Mangbetu 2. Bongo-Bagirimi 23. Berta 22. Kunama 25. ‘Gule’ [extinct] 26. Koma group

1. Nubian group 2. Murle, Logarim, etc. [today: Surmic] 3. Barea [today: Nara ~ Nera] 4. Ingassana (Tabi) [today also: Gaam/Jebel] 5. Nyima, Afitti 6. Temein, Teis-um-Danab 7. Merarit, Tama, Sungor 8. Daju group

5. Chari-Nile A. Eastern Sudanic

9. Nilotic 10. Nyangiya, Teuso [today: Kuliak] B. Central Sudanic C. Berta D. Kunama 6. Coman: Koma, Ganza, Uduk, Gule, Gumuz, Mao

I. Koman: T’wampa (Uduk), Gule, Komo, Kwama and Opo (Shita) and J. Gumuz (dialect cluster).

presented in these studies highlight significant features of the current state of discussion in the field of Nilo-Saharan genetic studies:

Several morphological sound–meaning similarities used by Greenberg (1963) to define Nilo-Saharan are still major arguments regarding Nilo-Saharan unity, among them:

—The idea of a Nilo-Saharan phylum remains widely accepted, although the external boundaries of Nilo-Saharan as well as significant parts of its internal subdivision remain highly controversial. —Greenberg’s proposal of a Chari-Nile group has been rejected as a valid genetic unit, while the genetic unity of Eastern Sudanic and of Central Sudanic remains unchallenged. —The lower-level units (Tucker and Bryan’s language units as revised by Greenberg) are also largely uncontroversial. —The most substantial progress since Greenberg has been made regarding language documentation, genetic classification, and reconstruction within these lower-level units.

—independent subject pronouns, especially in the singular, show the vowels a (1st person), i or u (2nd person) and e (3rd person), —ma and ko as relative and adjective formant, —n/k singular–plural alternation, —t/k singular–plural alternation, —nominal derivational prefix k- (‘movable’ k-), —verbal dative affix -kV-, —causative in t Several morphological sound–meaning correspondences proposed by Greenberg to support NiloSaharan unity, however, are also found in neighboring Afro-Asiatic (e.g. negation in b or m) or in NigerCongo languages, thus weakening the argument and pointing to the general difficulty of separating areal from genetic features. Most recently, only two comprehensive studies on the genetic classification of Nilo-Saharan since Greenberg’s have been published, i.e. Bender (1996) and Ehret (2001). The comparison of the results

750

External boundaries Discussions regarding the external boundaries of NiloSaharan relate particularly to the following topics: Songhay cluster: Its geographic distance from the other Nilo-Saharan languages and long-standing contact with Mande (Niger-Congo) and Berber (Afroasiatic) languages continue to cast doubt on the Nilo-Saharan affiliation suggested by Greenberg (1963) and supported

NILO-SAHARAN LANGUAGES by Bender (1996) and Ehret (2001). Nicolai (1990) reconstructs it as a post-Creole with a Berber base, while some other scholars follow an old argument by Delafosse and discuss Mande affiliation. Kadugli-Krongo (Kadu): This group appears in Greenberg (1963) under the name of Tumtum as one of five Kordofanian branches despite divergent properties. Following suggestions made by Dimmendaal (1987) among others and supported by the fact that this group shows several of the typical Nilo-Saharan features (1sg pronoun a'a, 2sg pronoun ’, movable k, n/k plural formation, etc.), Bender (1996) includes Kadugli-Krongo in his Core of Nilo- Saharan, while Ehret (2001) rejects any Nilo- Saharan affiliation. Shabo (formerly: Mikeyir): The genetic position of this rudimentarily documented language of SW Ethiopia remains unclear. Bender proposes a NiloSaharan affiliation, which Ehret rejects. Meroitic: The extinct language of the ancient Meroitic Empire (Sudan) preserved by a number of written records has been linked to various genetic groups and phyla, among them Afro-Asiatic, Tokharian, Saharan, and Eastern Sudanic. This latter affiliation was also supported by Greenberg, and has most recently again been substantiated by Claude Rilly, among others. Relation to Niger-Congo: Due to a number of lexical and morphological similarities between NiloSaharan and Niger-Congo languages, Gregersen (1972) proposed to combine these two phyla into a single one, i.e. Kongo-Saharan. His line of argument has been taken up by Roger Blench in recent years, who attempts to establish Niger-Congo as a branch of Nilo-Saharan.

Internal subgrouping General disagreement also characterizes higher-level units within Nilo-Saharan. Greenberg’s subgrouping (Table 1) has only partly been accepted. The reason for

this derives not so much from the still highly fragmentary documentation of numerous Nilo-Saharan languages, but from their generally great internal lexical and grammatical heterogeneity. This feature may point to a history of thousands of years for the whole phylum. Detailed studies of lower-level language units, however, also reveal continuous episodes of heavy language contact, thus complicating lexical and morphological reconstruction and hence higher-level grouping considerably. Based on morphological as well as on lexical comparison, Bender (1996) proposes four branches of NiloSaharan: (A) Songhay, (B) Saharan, (K) Kuliak, called ‘Outliers’, and the rest, called ‘Satellite–Core’, which he further subdivides as indicated in Table 2. Ehret’s (2001:88f, 70f) alternative genetic classification of Nilo-Saharan uses phonological and lexical isoglosses derived from an extensive comparative lexical database, but also considers grammatical and derivational morphemes as far as they are accessible through language descriptions. His result is presented in a mainly bilaterally branching tree, with one specific small language group or language branching off at each level and set against the rest, which receives novel geography-based labels. In the following condensed representation of his result, previously known group and language labels are underlined (see graph). The extreme lexical heterogeneity of Nilo-Saharan languages, even in the most basic vocabulary, is paralleled by a large grammatical diversity, and hardly any of these grammatical features is specific to the phylum. Nilo-Saharan languages are tone languages with two to four underlying level tones, which are used for grammatical as well as lexical purposes. The number of tones and their main functional domain correlate to a considerable degree with the morphological type of the respective language. Five-vowel systems with

TABLE 2 Bender’s (1996) Genetic Classification of Nilo-Saharan Outliers

Nilo-Saharan

A. Songhay B. Saharan K. Kuliak Satellites

Satellite-Core Core

C. Maban D. Fur F. Central Sudanic G. Berta H. Kunama E. Eastern Sudanic I. Koman J. Gumuz L. Kadugli-Krongo

Ek. Nubian, Nera, Nyima, Tama En. Surmic, Jebel (Gaam); Temein, Daju, Nilotic

751

NILO-SAHARAN LANGUAGES Nilo-Saharan Koman

Sudanic Central Sudanic

Northern Sudanic Saharo-Sahelian

Kunama

Saharan

Sahelian Songhay

Eastern Sahelian

For[Fur]

Eastern Sudanic

Maban

Nara

W-Astaboran Nubian

Kuliak

Kir-Abbaian

Astaboran

Tama

Kir

Jebel W-Jebel

Nuba-M.

Bertha

Temein

Daju

Nyimang

Surma-Nilotic Surmic

Nilotic

Graph. Ehret’s (2001) genetic classification of Nilo-Saharan.

advanced tongue root (ATR) vowel harmony, often with contrastive length, are frequently synchronically or diachronically reconstructable. The consonant systems show either four or five (e.g. Gaam, Berta, Koma, Western Nilotic) places of articulation, which partly also extends to the nasal. Ehret (2001:16) claims five places of articulation for Proto-Nilo-Saharan in the case of plosives and four in the case of nasals. He also suggests a complex set of plosives for Proto-NiloSaharan, with voiced implosives and explosives, voiceless aspirated and unaspirated as well as glottalized plosives. Although all of these sets do occur in Nilo-Saharan languages synchronically, only Twampa (Uduk) comes close to such a system. The morphological type of Nilo-Saharan languages ranges from largely isolating (e.g. Central Sudanic languages), to agglutinative (e.g. Southern Nilotic) and highly fusional (e.g. Kanuri, partly Western Nilotic). Languages of the same genetic unit often belong to different morphological types, which possibly has to be interpreted as an outcome of substantial language contact. Nonisolating Nilo-Saharan languages are mostly characterized by complex verbal derivational morphology, while isolating Nilo-Saharan languages tend to make use of serial verb construction for derivational and morphosyntactic purposes. Nominal morphology is equally heterogeneous, with some languages having gender (e.g. Eastern Nilotic), others not. A morphologically tripartite number system, in which either a singular/singulative or a plural form or both are morphologically derived from a nominal root, is widespread in Nilo-Saharan and may even be a potential isogloss. 752

TABLE 3 Nilo-Saharan Languages: Languages and Speakers Per Country Country

No. of NiSa Languages

Kenya Sudan Uganda Nigeria DR Congo Chad Niger Tanzania Mali Ethiopia CAR Eritrea Cameroon Burkina Faso Egypt Benin Libya

~6 ~ 64 ~ 20 6 22 ~ 33 ~7 ~4 1 16 13 2 2 2 1 2 2

No. of NiSa Speakers

% of Population

7.5 Mio. 7.2 Mio. 5.6 Mio. 3.5 Mio. 3.0 Mio. 2.6 Mio 2.7 Mio. 858 000 640 000 376 000 223 000 210 000 150 000 125 000 100 000 50 000 9 000

29.2 29 25.5 3.5 7 40.6 32.5 0.3 6 0.7 6.6 5.7 1.2 0.8 0.1 0.9 0.1

Syntactically, all the common basic word orders with its concomitant features occur, namely SOV (e.g. Kanuri, Nubian), VSO (e.g. East and South Nilotic, Kuliak, partly Kadugli-Krongo), and SVO. Morphological and partly also syntactic ergativity has been reported for Jur-Luwo, Anywa, Päri, Shilluk (Western Nilotic), and Toposa (Eastern Nilotic).

Demography Numbers given for Nilo-Saharan languages range from 80–90 (lumpers) up to 199 (splitters). The area in which they are spoken stretches from the river Nile in the

NOOTKA AND WAKASHAN LANGUAGES TABLE 4 Numerically Dominant Nilo-Saharan Languages Speakers

Language

Sub-division

Country

3–4 Mio.

Songhai Kanuri Dholuo Dinka Lwo (Acholi + Lango) Kalenjin Lugbara Alur Teso Maa Nuer Ngambay Lendu Mangbetu Fur Maba Teda-Daza Masalit Bari Karimojong Nile Nubian (Nobiin)

Songhay Saharan Western Nilotic Western Nilotic Western Nilotic Southern Nilotic Central Sudanic Western Nilotic Eastern Nilotic Eastern Nilotic Western Nilotic Central Sudanic Badha/Lendu Mangbetu-Balese Fur Maban Saharan Maban Eastern Nilotic Eastern Nilotic Nubian

Niger, Mali, Nigeria, Burkina, Faso, Benin Nigeria, Niger, Cameroon, Chad, Sudan Kenya, Tanzania Sudan Uganda, Sudan Kenya DR Congo, Uganda DR Congo, Uganda Uganda, Kenya Kenya, Tanzania Sudan, Ethiopia Chad, Cameroon, Nigeria DR Congo, Uganda DR Congo Sudan, Chad Chad Chad, Niger, Libyen Chad, Sudan Sudan, Uganda, DR Congo Uganda Sudan, Egypt

> 2 Mio. > 1 Mio.

> 500,000

> 250,000

northeast up to Lake Chad and––in the case of Kanuri and Songhay––further west to northeastern Nigeria and to the Mali-Niger-Burkina Faso border area, respectively. Nilo-Saharan languages are spoken as a first language in 17 African countries by over 30 Mio. speakers (Table 3). The highest number of speakers are found in Kenya (7.5 Mio.), Sudan (7.2 Mio.), Uganda (5.6 Mio.), Nigeria (3.5 Mio.), DR Kongo (3.0 Mio.), Niger (2.7 Mio.), Chad (2.6 Mio.), but these speakers represent significant portions of the population only in four countries, namely Chad (41%), Niger (32%), Sudan (29%), Kenya (29%), and Uganda (25%). With respect to the number of Nilo-Saharan languages per country, Sudan comes first (~64), followed by Chad, DR Congo, Uganda and––despite the overall low number of NiSa speakers––Ethiopia (16) and Central African Republic (13). Table 4 presents major Nilo-Saharan languages in terms of number of speakers and geographic distribution. As indicated, most of them are cross-border languages, thus making language-planning activities comparatively difficult.

References Bender, L.M. 1996. The Nilo-Saharan languages––a comparative essay. (LINCOM Handbooks in linguistics, Vol. 6.) München & Newcastle: Lincom Europe. Dimmendaal, Gerrit J. 1987. Krongo: between universal, areal and genetic norms. Journal of African Languages and Linguistics 9. 161–77. Ehret, C. 2001. A historical-comparative reconstruction of Nilo-Saharan. (Sprache und Geschichte in Afrika. Beiheft 12.) Köln: Rüdiger Köppe. Greenberg, J.H. 1963. The languages of Africa. International Journal of American Linguistics 29, 1 (Part 2). [Repr. 1966 by Bloomington: Indiana University; The Hague: Mouton.] Gregersen, E.A. 1972. Kongo-Saharan. Journal of African Languages 11. 69–89. Nicolai, R. 1990. Parentés linguistiques (à propos du songhay). Paris: Éditions du CNRS. Tucker, A.N., and M.A. Bryan. 1956. The non-Bantu languages of north-eastern Africa. London, New York, Cape Town: Oxford University Press for International African Institute. Tucker A.N., and M.A. Bryan. 1966. Linguistic analyses: the non-Bantu languages of north eastern Africa. London, New York, Cape Town: Oxford University Press for International African Institute.

MECHTHILD REH

Nootka and Wakashan Languages Wakashan is one of a dozen or so language families indigenous to North America’s Pacific Northwest, a lush and mountainous region of tremendous linguistic

diversity. The term is a variation of ‘Wakashian’, which Captain James Cook proposed to call the people he visited in Nootka Sound in 1778: 753

NOOTKA AND WAKASHAN LANGUAGES The word wakash ... was frequently in their mouths. It seemed to express applause, approbation, and friendship. For when they appeared to be satisfied, or well pleased with anything they saw, or any incident that happened, they would, with one voice, call out wakash! wakash!

The study of Wakashan languages has contributed significantly to the development of linguistics. For instance, the lasting and profound influence of such famous linguists as Franz Boas, Edward Sapir, Morris Swadesh, and Mary Haas is due in large part to their seminal work on these languages. Besides refuting the popular misconception that native languages are somehow primitive, their pioneering work spawned valuable research in linguistic affiliation relating to the ethnohistory of North America, and it also uncovered important structural and semantic phenomena that had not been found in more widely studied languages. Well-known contemporary linguists such as Emmon Bach and Stephen Anderson also credit their fieldwork on Wakashan languages for improving their understanding of language. Wakashan languages fall neatly into two branches, each comprising three languages. The southern branch (‘Nootkan’) includes Nootka (Nuu-chah-nulth) and Nitinat (Ditidaht) spoken along the west coast of Vancouver Island, as well as Makah spoken on Washington’s Olympic Peninsula. The northern branch (‘Kwakiutlan’) consists of Haisla-Henaksiala and Heiltsuk-Oowekyala spoken on the north and central coasts of British Columbia, and Kwakw’ala spoken on northern Vancouver Island and the adjacent mainland coast. The genetic relation between the two branches was discovered by Franz Boas in 1889. Their kinship is most clearly evidenced by the locational lexical suffixes (-i ‘indoors’, -as ‘outdoors on the ground’, -is ‘on the beach’, -a ‘on a rock’, etc.), which are of great frequency in all Wakashan languages. Edward Sapir and Morris Swadesh hypothesized that Wakashan belongs with Salish, Chimakuan, and Kutenai under a larger ‘stock’ called Mosan, and with Algic under a still larger ‘phylum’ called AlgonkinWakashan (or Almosan). At present, these primeval affiliations are not widely accepted by specialists, most of whom have chosen to recoil from such largescale classification until the histories of the smaller groupings are better understood. The lexical differences between Nootkan and Kwakiutlan alone are such that Swadesh estimated the two to have separated around three millenia ago, a time depth that is considered plausible by some (e.g. Jacobsen 1979) but too brief by others (e.g. Embleton 1985). For both branches, dialectal differentiation is greatest on Vancouver Island, which is therefore assumed to be the original home of the ancient Wakashans. The specialized

754

proto-Wakashan vocabulary on local maritime culture also suggests that the Wakashan Urheimat lies in its present area. Historically, the Wakashans were vigorous peoples. For instance, they used giant cedar trees to construct expansive houses for their extended families, to carve beautiful totem poles, and to build huge canoes which they used to fish for halibut and to hunt sea mammals ––including colossal humpback and gray whales. But they were almost decimated by the arrival of the Europeans. Epidemics (especially smallpox), alcohol, and firearms used in intertribal wars reduced their population from tens of thousands before contact (the 1700s) to a few thousand in 1929. Fortunately, Wakashan peoples are now recovering rapidly from these historical disasters; for instance, at present, the Nootka (Nuu-chah-nulth) and the descendants of Kwakw’ala-speaking tribes (Kwakwaka’wakw) number almost 8,000 and 6,000, respectively. Unfortunately, Wakashan languages remain in a grave state of decline. The Canadian government not only prohibited Wakashan peoples’ main social events (‘potlatches’) from 1884 to 1951 but also relocated their children to church-run residential schools from 1888 to 1983. These measures, which were taken deliberately to break down the transmission of culture and language, were all too effective. More passively, too, Wakashan languages have become obsolete under the influence of English, which has become the primary language at school and at home in every community thanks in part to the influence of the mass media. Thus, currently there are little more than a couple hundred speakers each of Nootka (Nuu-chah-nulth), Kwakw’ala, Haisla-Henaksiala, and HeiltsukOowekyala. Makah has just 20 to 30 speakers, while Nitinat (Ditidaht) has less than ten. Recently, however, communities have sought to counteract the loss of their ancestral languages by integrating them in schools. To this end, they have adopted standardized orthographies, compiled dictionaries, and developed multimedia materials on their languages. The long-term effect of these efforts on the survival of Wakashan languages is unknown but hopeful. Let us now focus on Nootka. This term is actually an error dating to Cook’s visit. On one popular account, the legendary captain mistook it as the name of the natives who told him nootka (‘circle about!’) to come to their village on the other side of an island. The term continues to be used widely in linguistics, but nowadays the people themselves prefer to be called Nuuchah-nulth (‘along the mountains’). There are a dozen distinct Nuu-chah-nulth tribes, with many different dialects. Grammars exist for three of these: the southern Tseshaht, the central Ahousaht, and the northern Ka:’yu:k’t’h’ (formerly Kyuquot). The following is a

NOOTKA AND WAKASHAN LANGUAGES brief survey of some structural characteristics of Nootka that the reader might find strikingly different and interesting. Nootka shares 12 consonant phonemes with English, and has at least 27 others. Fifteen of these involve various kinds of glottalization. For instance, the above-mentioned dialect name Tseshaht is pronounced with a postglottalized ‘ejective’ consonant and a glottal stop [ts ’iʃaʔat], while Nuu-chah-nulth is pronounced with a preglottalized ‘creaky’ consonant [nutʃanu]. Nootka consonants draw on all known places of articulation except dental. For example, the dialect names Ka:’yu:k’t’h’ and Ahousaht involve uvular, pharyngeal, and epiglottal consonants: [qajuk’at, ausat]. Nootka’s consonantal inventory would have been even larger had it not historically abandoned the lateral approximants [| |] and the voiced obstruents [b d dz d g g G G ], which are still in use in other Wakashan languages. In contrast to its manifold use of consonants, Nootka has just three basic vowels, in long [i a u] and short [i a u] varieties, plus two marginal vowels [ε, ɔ] found mainly in foreign borrowings. Nootka lacks schwa [ə], which is widely used in Kwakiutlan. Nootka phonemes are implicated in many active phonological processes, including lenition, glottalization, assimilation, shortening, and deletion. For instance, all these processes are involved in the following utterance (Tseshaht): /hi-as-uk-ap-ma/; [hijasuk ’ama] ‘I put mine in a container’ (LOC-‘in vessel’-POSS-CAUS-LS). Every Nootka syllable consists of a vowel obligatorily preceded by a single consonant, and optionally followed by up to three consonants, as in [.int.tin.ʔi.] ‘the one made of snot’. However, in the Ka:’yu:k’t’h dialect, vowels are regularly dropped inside and at the ends of words, such that any number of consonants is possible in sequence; for example, [t’ut’uʃink ʃtts’] (…-inuk ‘at hands’ + -ʃit mom. + -ts’i ‘at fire’) ‘he was drying his hands at the fire’. Incidentally, an extreme form of consonant sequencing occurs in Oowekyala: this Kwakiutlan dialect is remarkable in permitting consonant-only words such as [t’xt’k ’s] ‘fish hawk’ and even consonant-only utterances such as [t spstkts] ‘this (not visible) will be a nice thwart’. Words in Nootka show regular stress patterns (e.g. [haja akʃi at] ‘did not know now’), as in other Wakashan languages (but not all; Haisla has a pitch accent system and Heiltsuk is tonal). Nootka morphology is also very complex. On the one hand, it is polysynthetic in that hundreds of lexical suffixes can combine with well over a thousand roots to form large stems with composite meanings. For example, a verbal suffix like -nak ‘to have’ can be added to a nominal root as in tʃ ’apats-nak (canoehave) ‘to have a canoe’, to an adjectival root as in  i-

nak tʃ ’apats (big-have canoe) ‘to have a big canoe’, or to a ‘dummy’ root as in  u-nak  i tʃ ’apats (it-have big canoe) ‘to have a big canoe’. It can also combine with other lexical suffixes as in tupk-nak-maiqt (black-have-want) ‘to want to have a black one’. On the other hand, much of Nootka morphology is nonconcatenative in that many lexical and aspectual distinctions are expressed by reduplicating the root or by lenghtening its vowel(s). For instance, from the verb form mitx-ʃit ‘to make a turn’, we get the ITERATIVE mitxmitx-ʃit ‘to start in on turns at intervals’, the GRADUATIVE mitx-ʃit ‘making a turn’, the REPETITIVE mitxmitx-ʃit ‘to start turning around and around’, and (tenuously) the DISTRIBUTIVE REPETITIVE mimitxmitx-ʃit ‘to start turning around and around here and there’. In sum, the Nootka word consists of a root, which itself may be modified by reduplication or lengthening, and to which may be added a large number of suffixes with various lexical and grammatical meanings. Turning to syntax, it is often obscured in Nootka by the complex morphology that allows words to convey sentences, as in Ka:’yu:k’t’h’ ʔu-k i-tʃip-’itʃ-is-im (it-make-for-IMPV-you(PL)-me-will) ‘you’ll make it for me!’, ʔu-sup-intiʃ (it-kill-PST.IND) ‘he killed it’, and muk-sup-intiʃ (deer-kill-PST.IND) ‘he killed the deer’. When individual words are used instead of affixes, the following basic order is apparent: predicate–subject–object, as in qasap-intiʃ tʃ akup muwitʃ (killPST.IND man deer) ‘a man killed a deer’. Interestingly, any category of word can be a predicate in Nootka, including adverbs as in ʔati-intiʃ qasap muwitʃ (nightly-PST.IND kill deer) ‘it was at night he killed the deer’, nouns as in tʃ akup-intiʃ qasap muwitʃ (man-PST.IND kill deer) ‘it was a man that killed the deer’, and demonstratives as in ʔu-intiʃ tʃ akupi qasap muwitʃ (that-PST.IND man kill deer) ‘it was that man that killed the deer’. Of particular interest in this context is the oft-repeated claim that Nootka lacks category distinctions. This claim is controversial, but most linguists agree that this kind of distinction is weak in Nootkan syntax. For example, in Tseshaht quʔas ‘man’ and mamuk ‘work’ act as noun and verb, respectively, in mamuk-ma quʔas-ʔi ‘the man is working’ (work-he man-the), but these grammatical categories appear to be reversed in quʔas-ma mamukʔi ‘the working one is a man’ (man-he work-the).

References Boas, Franz. 1947. Kwakiutl grammar. Philadelphia: American Philosophical Society. Cook, Captain James. 1784. A voyage to the Pacific Ocean: undertaken by the Command of His Majesty, for making discoveries in the Northern Hemisphere; performed under the direction of Captain Cook, Clerke, and Gore, in His

755

NOOTKA AND WAKASHAN LANGUAGES Majesty’s Ships the Resolution and Discovery, in the years 1776, 1777, 1778 and 1780. London: W. and A. Strahan. Embleton, Sheila M. 1985. Lexicostatistics applied to the Germanic, Romance, and Wakashan families Word 36. 37–60. Goddard, Ives (ed.) 1996. Languages. Handbook of North American Indians, Vol. 17. Washington: Smithsonian Institution. Haas, Mary R. 1972. The structure of stems and roots in Nootka–Nitinat. International Journal of American Linguistics 38. 83–92. Jacobsen, William H. Jr. 1979. Wakashan comparative studies. The languages of North America: historical and comparative assessment, ed. by Lyle Campbell and Marianne Mithun. Austin, TX: University of Texas Press.

Lincoln, Neville J., and John C. Rath. 1980. North Wakashan comparative root list. Ottawa: National Museums of Canada. McMillan, Alan D. 1999. Since the time of the transformers: the ancient heritage of the Nuu-chah-nulth, Ditidaht, and Makah. Vancouver: University of British Columbia Press. Mithun, Marianne. 2001. The languages of Native North America. Cambridge, UK: Cambridge University Press. Nakayama, Toshihide. 1999. Nuu-chah-nulth (Nootka). Munich: LINCOM EUROPA. Sapir, Edward, and Morris Swadesh. 1939. Nootka texts: tales and ethnological narratives with grammatical notes and Lexical materials. Philadelphia: American Philosophical Society. Stonham, John T. 1999. Aspects of Tsishaath Nootka phonetics and phonology. Munich: LINCOM EUROPA.

DARIN HOWE

Northwest Caucasian Languages The languages of the Northwest Caucasian family––Abaza, Abkhaz, Adyghe, Kabardian, and Ubykh––were originally spoken in a contiguous region stretching from Abkhazia on the Black Sea in the south to the Kuban River in the north. As a result of the relocations following the Russo-Caucasian War (1817–1864), substantial communities of each now exist in Turkey, as well as smaller enclaves in Syria, Jordan, Israel, former Yugoslavia, Western Europe, and New Jersey. The languages of this family fall into three groups: Abkhaz/Abaza, Ubykh, and Circassian, the latter containing the languages Adyghe and Kabardian. In 1850, there were approximately one million speakers of Northwest Caucasian languages, most of them Circassians. The Russo-Caucasian war and the Russian colonization of the Caucasus drastically changed this situation: many Abkhaz and Circassians were forced to flee westward to the Ottoman Empire, and the Ubykh and Sadz Abkhaz were entirely displaced from their homelands near the Kwydypsta River just north of present-day Abkhazia. At present, some 800,000 Northwest Caucasians still reside in the Caucasus, most of whom retain their ancestral languages, as well as Russian. In addition, more than a million Northwest Caucasians now live in the diaspora, primarily in Turkey. Current estimates of the number of North Caucasians in Turkey range from one to six million, depending on the sources used. It is difficult to generate reliable figures, because many Northwest Caucasians now have complicated ethnic identities resulting from cultural and linguistic assimilation and intermarriage. In the Northwest Caucasus

756

proper, the different ethnic groups have generally preserved their own languages, although for some Russian has become the dominant language. Many Caucasian linguists believe that the Northwest Caucasian family is genetically related to the neighboring Northeast Caucasian family, but the linguistic arguments presented in support of this connection are not convincing, resting on typological similarity rather than the comparative method. All of the Northwest Caucasian languages have literary forms, with the exception of Ubykh, which is now extinct. Their current orthographic forms are based on the Cyrillic script, although there is currently movement to develop a new orthography based on the Roman script (most likely its Turkish manifestation). Abkhaz and Abaza are quite similar to one another and are considered by some scholars to be dialects of the same language; the same holds for Adyghe with respect to Kabardian. Ubykh is linguistically intermediate between these two groups. The Northwest Caucasian languages are most famous for pairing unusually rich consonant inventories (containing as many as 83 consonants, in Ubykh) with unusually small vowel inventories, typically consisting (in native words) of only a low vowel /a/ and a non-low vowel //. The richness of the consonant system derives for the most part from the addition of labial and palatal secondary articulations. It appears that the enormity and smallness of the consonantal and vocalic systems, respectively, are related: the ancestral (proto) language most likely had more conventional vowel and consonant systems, and subsequently reinterpreted the [back] and [round] features of the vowels

NORTHWEST CAUCASIAN LANGUAGES as resulting from secondary7 articulations on neighboring consonants. (Turkish, Arabic, and Persian but not Russian loans underwent this reanalysis as well: cf. Turkish dükkân ‘store’→Abkhaz /a-dwk’yan/ [ɑ´d buk’jɑ´n].) One can see in this example that what are˘now phonologically consonantal secondary articulations still surface on the vowels in most contexts. All of the Northwest Caucasian languages distinguish three types of laryngeal consonants, opposing plain voiced, voiceless aspirated, and ejective stops. Many also possess extremely complicated intonational stress systems, most famously Abkhaz. The system of agreement marking in Northwest Caucasian verbs is also uncommonly intricate. Abkhaz, for example, overtly marks verbal agreement with subjects, direct objects, indirect objects, and various other oblique arguments, as in the following Abzhuy example from George Hewitt (transcribed in the IPA): sɑɾɑ´ I

ɑ-pɥ ´s the-woman

ɑ-sɑp’´n the-soap

s-χɑɾph (∅-)ɑ-lɑ-l s´-ɾ-dɥ dɥɑ´-(∅-)jt’ my-shirt (it-)it-by her-I-cause-wash-(PAST-)FINITE ‘I got the woman to wash my shirt with (the) soap’ It is possible to string together as many as nine overt morphemes in a row in a sequence of verbal prefixes. The nominal morphology, on the other hand, is relatively simple, containing only two cases (direct object and oblique or adverbial). Possession is expressed via prefixes, as in Abkhaz s-nap’-kwɑ´ ‘my hands’ (I-hand-nonhuman.plural). Abkhaz (and Abaza) distinguish human vs. nonhuman in certain morphological classes, as can be seen in the plural in this example. All of the Northwest Caucasian languages distinguish two genders, two numbers, and three persons in their system of agreement markers. Finding lexical items shared by all of the Northwest Caucasian languages is difficult, but not impossible: ProtoNorthwest Caucasian

Common Circassian

Ubykh

Common Abkhaz

*pxja ‘back’ *bza ‘tongue’ *zwja ‘cook, boil’ *gwjə ‘heart’

pxa *bza *zwja *gwə

pšja bzja zwja gj ə

*(p)χja *bzə *zwjə *gwə

The Northwest Caucasian numeral system is interesting in that it preserves vestiges of an original vigesimal system in the numbers 30–99. For instance, Abkhaz (Cwyzhy dialect) eɥzɥejzɥɑbɑ ‘thirty’ is literally ‘twenty (eɥzɥɑ) and (-j) ten (zɥɑ-)’; ɥneɥzɥɑ ‘forty’ is ‘two twenties’, and so on.

Turning to the individual languages, Abaza had 34,800 speakers in the Karachay-Cherkess autonomous region of Russia at the time of the 1989 census; as of 1995, there were 10,000 more in Turkey, 80 (out of 150 members of the ethnic group) in Germany, and about 15 in the United States. There are three main dialects: Tapanta, Ashkhar, and Bezshagh. The literary language is based on the Tapanta dialect. Abaza is reported to be mutual with Abkhaz. Abkhaz had 101,000 speakers in Abkhazia in 1993, as well as 4,000 speakers out of approximately 15–30,000 ethnic Abkhaz in Turkey, and smaller numbers scattered in other countries. Abkhazia was under Georgian rule during the Soviet period, but seceded shortly after the dissolution of the Soviet Union in the mid-1990s. The continued refusal by foreign nations to recognize Abkhazia’s sovereign status, together with the forced immigration into Abkhazia of Slavs, Armenians, and Mingrelians during the 1930s, and Georgia’s continuing attempts to eliminate its ethnic minorities, have placed the Abkhaz language in danger of disappearing. The assignment of literary status to Abkhaz (including TV broadcasts since 1978) has helped counteract this tendency within Abkhazia, but few members of the younger generation are learning the language in the diaspora. There are three main dialect groups: Bzyp (north of Sukhumi), Abzhui, the literary dialect (south of Sukhumi), and Sadz (now spoken only by a handful of Abkhaz in Turkey). The Circassians live in the Adyghe, KabardinoBalkar, and Karachay-Cherkess republics in the Russian Federation. The Circassians have two main languages, Adyghe (West Circassian) and Kabardian (East Circassian), the former of which has four main dialects (Shapsugh, Bzhedug, Temirgoy (the literary dialect), and Abzakh) and the latter of which has six (Besleney, Malka, Bakhsan, Terek, Mozdok, and Lesser Kabardian). As of 1993, there were some 125,000 Adyghe and 46,000 Kabardians in Russia, 100,000 Adyghe and 202,000 Kabardians in Turkey, and a total of approximately 750,000 Circassians in the world as a whole. Both languages have literary status and are taught in the schools in their respective republics, but their survival in the diaspora is less secure following the closing of the Circassian school in Turkey by Atatürk in 1920 and the outlawing of publication in minority languages in Turkey in 1983. The Ubykh community originally lived on the Black Sea just north of Abkhazia, but all 50,000 of its members were deported to the Ottoman Empire following the Russian conquest of the Caucasus, ending up in Hac Osman, a town near Istanbul. Although the community retains a distinct identity, its language is now extinct. Isolated individuals can still be found who know pieces of the grammar and vocabulary, but

757

NORTHWEST CAUCASIAN LANGUAGES most of the remaining Ubykhs speak Adyghe and/or Turkish. References Allen, Sidney. 1956. Structure and system in the Abaza verbal complex. Transactions of the Philological Society. 127–76. Chirikba, Viacheslav. 1996. Common West Caucasian. Leiden: CNWS Publications. Colarusso, John. 1988. The Northwest Caucasian languages: a phonological survey. New York: Garland Press. Colarusso, John. 1992. A grammar of the Kabardian language. Calgary: University of Calgary Press. Dumézil, Georges. 1931. La langue des oubykhs. Paris: Champion. Dumézil, Georges. 1932. Études comparatives sur les langues caucasiennes du nord-ouest (Morphologie). Paris: AdrienMaisonneuve.

Hewitt, George. 1979. Abkhaz. Amsterdam: North-Holland. Hewitt, George (ed.) 1989. The indigenous languages of the Caucasus, Vol. 2: The North West Caucasian languages. Delmar, NY: Caravan Books. Kuipers, Aert. 1960. Phoneme and morpheme in Kabardian. The Hague: Mouton. Nikolaev, S., and S. Starostin. 1994. A North Caucasian etymological dictionary. Moscow: Asterisk Publishers. Paris, Catherine. 1974. Système phonologique et phénomènes phonétiques dans le parler besney de Zennen Köyü. Paris: Klincksieck. Smeets, Rieks. 1984. Studies in West Circassian phonology and morphology. Leiden: Hakuchi Press. Sommerfelt, Alf. 1934. Études comparatives sur le caucasique du Nord-ouest. Norsk Tidsskrift for Sprogvidenskap 7. 178–210. von Mészáros, Julius. 1934. Die Päkhy-Sprache. Chicago: University of Chicago Press.

BERT VAUX

Noun Incorporation The term Noun Incorporation refers to the linguistic phenomenon according to which a noun and a verb may combine into a single word. The two elements form one morphological unit that is often described as a complex verb (or as a compound). Usually, the incorporated noun is interpreted as the object of the verb. For example, in Onondaga, an American Indian language of the Iroquian language family, the noun (stem) hwist ‘money’, which is in bold below, may appear as part of the verb (stem) ahtu ‘lose’. In particular, the noun (stem) immediately precedes the verb (stem), as we can see in the following example: (a). Pet wa?-ha-hwist-ahtu-?t-a? Pat he-money-lost ‘Pat lost money’

In this case, the noun (stem) and the verb (stem) form a single word, namely the complex verb wa?-hahwist-ahtu-?t-a? ‘he-money-lost’. The incorporated noun (stem) hwist ‘money’ is interpreted as the object of the verb ahtu ‘lose’. In the same language, the object of the verb can be an independent word in the sentence. In this case, the object is one word: ne?o-hwist-a? ‘the money’, while the verb is another word, namely, wa?-ha-htu-?t-a? ‘he-lost’. The noun now follows the verb, as illustrated below: (b) Pet wa?-ha-htu-?t-a? ne?o-hwist-a? Pat he-lost the-money ‘Pat lost the money’

758

(Examples (a) and (b) are taken from Baker 1988:76.) Incorporating languages thus have two ways of expressing the verb and its object: either the two elements are realized as a single word after incorporation of the noun into the verb, as shown in example (a), or they appear as two separate words. In the latter case, the noun does not incorporate into the verb, but it follows the verb, as shown in example (b). The two constructions may differ slightly in their interpretation. For example, Incorporation sometimes has the function of narrowing the semantic range of the verb. This implies that the speakers of incorporating languages chose between the nonanalytic configuration (where the object is incorporated into the verb) and the analytic configuration (where the object is not incorporated into the verb) on the basis of the particulars of the information they want to communicate. Noun Incorporation is a productive process mainly in non-Indo-European languages. It is attested in some North American languages (Athapaskan, Caddoan, Chimariko, Iroquian, Siouan, Kiowa-Tanoan, Natchez, Takelma, Uto-Aztecan, Yana, Zuni), in Australian languages (Rembarnga), in Paleo-Siberian languages (Chukchee), in Oceanic languages, in Turkish, and in other languages. There has been a long debate in the literature of Generative Grammar regarding Noun Incorporation. The debate concerns the locus where these complex words are formed. On the one hand, it has been argued

NOUN INCORPORATION that Noun Incorporation is an instance of compounding in the Lexicon (Mithun 1984; Di Sciullo and Williams 1987; Rosen 1989). Compounding refers to the combination of two (or more) stems or words into one single word. For example, the English word bedroom is a compound consisting of two elements: bed and room. Hence, it has been argued that Noun Incorporation is a case of a noun and a verb that are put together in the same way that bed and room are combined into bedroom. According to this view, the noun–verb complex is not related structurally to its analytic counterpart. That is, there is no association between the noun–verb compound and the structure that involves the verb and its object, despite the fact that the two configurations may have very similar interpretation. The noun and the verb are put together morphologically and behave as one syntactic unit. This hypothesis is usually known as the lexicalist hypothesis. Baker (1988), on the other hand, was the first to argue that Noun Incorporation is a syntactic phenomenon. His main idea is that Noun Incorporation is a syntactic process that derives complex verbs from sentences where the verb takes a noun as its object. Specifically, Baker argues that the incorporated noun starts out as an independent word, in a position following the verb. It ends up in a position preceding the verb after an operation of syntactic movement. That is, syntactic movement changes the position of the noun and places it within the verbal complex. An empty slot is left behind, in the original postverbal position of the noun, which is called the trace of the moved element. Baker provides a straightforward explanation of the fact that the incorporated noun is interpreted as the object of the verb. It starts out as the complement of the verb, i.e. the two elements are adjacent and in a specific order: the noun follows the verb. This is considered as the structural position where the internal theta-role of the verb is assigned (following Baker’s Universal Thetarole Assignment Hypothesis––the UTAH––according to which each thematic role of the verb is assigned in a specific syntactic configuration). That is, any element that is base-generated in that particular syntactic position is interpreted as the direct object of the verb, namely as the theme (i.e. the argument that is affected by the action described by the verb). Hence, if the incorporated noun starts out as the complement of the verb, it is assigned the internal theta-role of the verb and is thus interpreted as the object of the verb, even after incorporating (by syntactic movement) into the verb. Incorporation is predicted to be possible with any noun that starts out from the verbal complement position. This is actually true: Noun Incorporation is a productive process in incorporating languages. In nonincorporating languages, like English, the object cannot form a single word with the verb. Words like truck drive are not in use. This means that the noun

truck, which is the object of the verb, cannot become part of the verb drive, but rather the two elements have to remain two separate words. The noun truck can be combined with the noun driver, giving rise to a complex noun: truck driver. However, it can be argued that these cases are different from genuine instances of Noun Incorporation in that the derived word is a noun rather than a verb. There are some complex words in English that resemble cases of Noun Incorporation. It could be argued that the formation of verbs like baby sit possibly involves an operation of syntactic movement. However, such examples lack a parallel analytic configuration. That is, to sit a baby cannot be considered as the parallel analytic configuration of the complex verb to baby sit. The latter has a very unexpected meaning. It expresses the notion ‘to look after a baby’. If baby sit is not an instance of Noun Incorporation, it could be viewed as an instance of compounding. This would mean that the verb baby sit is derived by a lexical or a morphological process and not by a syntactic process. Lastly, it has been argued that Incorporation is a widely spread phenomenon that is not necessarily restricted to nouns. It is a process that allows any two elements to combine into a single word. One of the two parts of the derived word is usually a verb. It has been proposed, for example, that certain kinds of pronouns incorporate into the verb. These are called clitic pronouns and are set apart from other pronouns by certain special characteristics. While ordinary pronouns do not appear as parts of the verb, clitic pronouns may do so. This process is often named cliticization. The process of cliticization can be distinguished, though, from Noun Incorporation in that the former possibly involves incorporation of a functional element (a clitic pronoun) whereas the latter involves incorporation of a lexical element (a noun). References Baker, Mark. 1988. Incorporation. Chicago: University of Chicago Press. Baker, Mark. 1996. The polysynthesis parameter. Oxford: Oxford University Press. Di Sciullo, Anna Maria, and Edwin Williams. 1987. On the definition of word. Cambridge, MA: MIT Press. Mithun, Marianne. 1999. The languages of native North America. Cambridge: Cambridge University Press. Mithun, Marianne. 1984. The evolution of Noun Incorporation. Language 60. 847–94. Rosen, Sara Thomas. 1989. Two types of Noun Incorporation: a lexical analysis. Language 65. 294–317. Spencer, Andrew. 1991. Morphological theory. Oxford: Blackwell. Spencer, Andrew, and Arnold Zwicky (eds.) 1998. The handbook of morphology. Oxford: Blackwell Publishers.

DIMITRA PAPANGELI See also Generative Grammar; Mohawk and Iroquoian Languages 759

NUMBER MARKING

Number Marking Number is the morphological category that expresses contrasts involving countable quantities. The most widespread number opposition is likely that between singular (one) and plural (more than one). Several languages, however, make further distinctions. Some of them (e.g. Greek or Sanskrit) exhibit dual forms that apply to two individuals, and a few languages (e.g. some Southwest Pacific languages) additionally distinguish a trial form that refers to three individuals. In some systems, there are also paucal forms for indicating a small number, as in Arabic. Although number implies a basic linguistic contrast, languages deal with this category in different ways. Most languages mark number in nouns; yet, there are languages (e.g. Maori, a Polynesian language) in which it is signaled by determiners, and others that lack number distinctions (e.g. Nancowry, spoken in India’s Nicobarese Islands). The explicit expression of number, on the other hand, may be obligatory, as in most English nouns; limited to certain nominal categories, as in most Algonquian languages (spoken by American Indians); or optional, as in Halkomelen (another American Indian language). Number marking differs greatly among languages, depending on the morphological system they display. In languages traditionally classified as analytic, such as Chinese, words consist of only one syllable and are invariable. Thus, plural nouns do not exhibit any formal marker and the notion of plurality must be inferred from the context. Only when this information is not contextually available or the idea of plurality is to be emphasized, separate words are added: ki ‘some’, šu ‘number’. Synthetic languages, on the contrary, rely on inflection, i.e. the modification of a word’s form, to express number distinctions. Modifications of words may equally result in derivation, which involves the creation of new words with new meanings. In traditional morphological theory, which has its origins in American structuralism, the analysis of word structure was based on the morpheme—the smallest meaningful unit of language. Morphemes are, however, abstract elements that are physically realized by morphs. The word houses, for example, consists of two morphs: house and -s. It is very frequent that a particular morpheme is represented by different morphs in different environments. For instance, the plural morpheme in the English noun oxen is not realized by the regular morph -s, but by the irregular -en. These alternants are called allomorphs.

760

Both inflection and derivation are traditionally explained exclusively in terms of affixation. As regards inflection, affixation consists in the addition of an inflectional affix (an obligatorily bound morph that contains grammatical information) to another morph with lexical content, as in house-s. This approach to inflection, however, is said to face important difficulties. One of them is illustrated by irregular English plurals such as men, in which plurality is marked not by the addition of an allomorph, but rather by means of a process that causes a vocalic change. The socalled portmanteau morphs (morphs that realize more than one morpheme, as the ending -us in the Latin noun annus ‘year’, which indicates simultaneously singular number and nominative case) are also problematic, since this approach assumes a one-to-one correspondence between form and meaning. In view of such shortcomings, recent morphological theories of inflection have abandoned this classical treatment and have adopted a model in which words are modified by means of different morphological processes that apply to the word itself, which is considered the base or stem. The most simple device to express number is, as in traditional accounts, affixation. Affixes can be divided into three main classes: prefixes, which attach to the front of the stem; suffixes, which come after the stem; and infixes, which occur within the stem. Most IndoEuropean languages make use of suffixes, such as the English or Spanish regular plural ending -s, to express number contrasts. Number marking by means of prefixation can be found in Bantu languages such as ι ru˜u˜thi), whereas the Kikuyu (mu˜-ru˜u˜thi ‘lion’, pl. m˜Uto-Aztecan language Oaxaca Chontal illustrates the use of infixes to distinguish number (kwepoʔ ‘lizard’, pl. kwe-B-poʔ ). Other processes by which number oppositions may be expressed involve the modification of the stem. This shift may affect the quantity or quality of the internal vowel, as in the English pair foot/feet. Vowel change is generally referred to as ‘ablaut’, although, because of its historical origin, the alternation just mentioned receives the distinct name ‘umlaut’. Not only vowels but also consonants may be modified to mark number. In the West African language, Fula, the plural of the word yiite ‘fire’, is formed by changing the initial consonant, giite. Such consonant alternations are frequently the result of a phonological change induced by an affix, which is then lost.

NUMBER MARKING In many languages, information about number is conveyed by means of modifications of the suprasegmental features tone and stress. In Somali (an East African language), for example, number is indicated by a change in the tone pattern: the singular form èy ‘dog’ takes a falling tone, whereas the plural éy takes a high tone. Russian illustrates the use of stress shift to distinguish singular from plural: in a particular group of neuter nouns, the singular is ending-stressed, while the plural is stem-stressed: oknó ‘window’ (nominative singular), ókna (nominative plural). Another common morphological process in certain languages is reduplication, which consists in the copying of the whole stem––total reduplication––or only of a part of it––partial reduplication, typically the leftmost portion. Reduplication is said to be frequently used iconally, i.e. the form of the word reflects its meaning. For this reason, reduplication is a common marker of plurality. Total reduplication to form the plural is found in Indonesian (babi ‘pig’, pl. babibabi), whereas partial reduplication is used in Motu, a language of Papua New Guinea (tau ‘man’, pl. tatau). The last morphological operation associated with number marking is suppletion. This process replaces one form by a phonologically unrelated form. A clear example of this phenomenon is the use of went as the past tense of go. Suppletion appears in many Western North American languages as a number marking

device: in Navajo, the form for ‘one is standing’ is different from that for ‘several are standing’. References Anderson, Stephen R. 1985. Inflectional morphology. Language typology and syntactic description, Vol. III, ed. by Timothy Shopen. Cambridge: Cambridge University Press. Anderson, Stephen R. 1988. Morphological theory. Linguistics: the Cambridge survey, Vol. I, ed. by Frederick J. Newmeyer. Cambridge: Cambridge University Press. Anderson, Stephen R. 1992. A-morphous morphology. Cambridge: Cambridge University Press. Bauer, Laurie. 1988. Introducing linguistic morphology. Edinburgh: Edinburgh University Press. Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart, and Winston; Chicago and London: The University of Chicago Press, 1984. Jespersen, Otto. 1922. Language. Its nature, development and origin. London: George Allen & Unwind Ltd. Matthews, P. H. 1972. Inflectional morphology. A theoretical study based on aspects of Latin verb conjugation. Cambridge: Cambridge University Press. Matthews, P. H. 1974. Morphology. An introduction to the theory of word-structure. Cambridge: Cambridge University Press. O’Grady, William, Michael Dobrovolsky, and Francis Katamba (eds.) 1987. Contemporary linguistics. An introduction. London and New York: Longman; 3rd edition, 1996. Spencer, Andrew, and Arnold M. Zwicky. 1998. The handbook of morphology. Oxford, Malden, MA: Blackwell.

LUISA GONZÁLEZ-ROMERO See also Affixation; Inflection and Derivation

761

O Official Language Selection Official language selection may be simply defined as the political choice of which language, or languages, to use in the legislative, executive, and judicial business of government. Given that all nations in the twenty-first century comprise populations speaking different languages, the selection of official languages is much more complex than the definition. This is because granting official status to a particular language will enhance its prestige, extend its use to educational and nonofficial domains, privilege its speakers, and impinge on the linguistic rights of speakers of other languages within the community. Both newly independent and well-established countries face difficulties when selecting official languages, although the contributing factors differ. Examples of each will now be considered. Countries that are colonized or subjected to imperialism in one form or another usually have the language of the colonial state imposed on them. Upon reaching independence, the selection of the official language is a matter of considerable importance––and controversy––because there are often several contending languages. One of them may be the first language of a majority of, or dominant group within, the population, and this would seem the obvious choice for nation building. This was the case with the nations that re emerged after the demise of the Soviet Union in the 1990s (Ukraine, Latvia, Belarus, and so on). However, in such cases, speakers of other indigenous languages may feel that they are regarded as second-class citizens for whom access to public services is restricted, if not denied completely, by the lack of official recognition of their mother tongue. This situation arose in

Pakistan, where Urdu was declared the official language on independence from Britain in 1947. However, because Urdu was not an indigenous language of East Pakistan, the overwhelmingly Bengalispeaking majority there felt marginalized; the unrequited demand for language rights led to civil war, after which Bangladesh achieved independence from Pakistan in 1971. Therefore, some newly independent nations select the language of the previous imperial power; this was the pattern among most ex-colonies of Britain in the 1960s and 1970s, such as Ghana, Nigeria, and Kenya. Although this choice means that a new nation may relatively easily build on the existing communications infrastructure, the choice may be inappropriate. There is often a residue of ill-feeling toward the culture and language of the erstwhile rulers; such a decision may also leave in positions of influence those who collaborated with the colonial power and who may be out of sympathy with a new regime. Because of this, some countries adopt another international language. The example of East Timor, which became independent in 2000, is a case in point. None of the indigenous languages was considered appropriate as the official language. The language of the erstwhile dominant power––Indonesia––was considered unacceptable on political and cultural grounds. Portuguese––the language of the earlier colonizer––was eventually chosen because it was widely understood and was used both within East Timor and internationally. Since 1994, South Africa has presented a very different example of language policy––one of national pluralism. At the national level, 11 indigenous languages have official status, and each provincial

763

OFFICIAL LANGUAGE SELECTION legislature can determine which of the languages will be used for internal official purposes. Longer established nations may also have difficulty in selecting official languages, as examples from major English-speaking countries illustrate. For many years, the francophone population of Canada struggled for legal recognition of their language until eventually the 1969 Official Languages Act declared that English and French had equality of status with equal rights and privileges. The case of the United Kingdom is different because although English is the national language, it has no official status: the only legally recognized official language in Great Britain is, since 1967, Welsh. Similarly, since 1987, the only official language throughout New Zealand is Maori. The reason for this apparent legal anomaly is that the indigenous speakers of those two minority ethnic groups campaigned to have their languages officially sanctioned for use in public services such as the legal system and education, health, and social services. With regard to English, there was no need to formalize such rights because they existed in fact, although not in law. Some English-speaking people in these countries opposed the granting of prestige and rights to these minority languages; they also opposed the allocation of public resources for their development and diffusion. None of these countries has an explicit overall national policy on languages, and there is considerable scope for uncertainty and even anxiety regarding the language rights of speakers of nonofficial languages:for example, immigrants and minority ethnic communities in all three countries and the indigenous First Nation peoples of Canada. By contrast, Australia has been developing a comprehensive National Languages Policy since 1987; although this now declares that English is the official language of Australia, it recognizes the aboriginal languages and also specifies the provision of services in languages other than English. The United States has never selected an official language, despite the large number of indigenous, colonial, and immigrant language communities. Recent attempts to change the federal constitution to adopt English as the official language throughout the

United States have been unsuccessful, although since the 1980s many state legislatures have passed laws proclaiming, with varying degrees of specificity, English as the official language within their own borders. Language is not merely a means of communication: it is the most important cultural symbol of a community. For many countries––Japan, Somalia, Malta, and many others––the national language is a core value that bonds the people together and indeed defines that very community. Increasingly, however, all nation states are host to diverse linguistic minority communities in their midst, whether this is caused by temporary business activity, permanent migration, urgent asylumseeking, or the revitalization of indigenous languages. It is, therefore, necessary for national policy-makers to consider the extent to which multilingualism (and multiculturalism) presents either a threat or a resource to their communities, and thereafter to select official languages and frame national language policies in the light of this understanding. References Bourdieu, P. 1991. Language and symbolic power. Cambridge: Harvard University Press. Eggington, William, and Helen Wren. 1997. Language policy: dominant English, pluralist challenges. Amsterdam, Philadelphia: John Benjamins Publishing Company. Fairclough, N. 1989. Language and power. London: Longman. Herriman, Michael, and Barbara Burnaby (eds.) 1996. Language policies in English-dominant countries. Clevedon, Philadelphia, Adelaide: Multilingual Matters Ltd. Phillipson, Robert. 1992. Linguistic imperialism. Oxford: Oxford University Press. Schiffman, H.F. 1995. Linguistic culture and language policy. London: Routledge. Tollefson, James, W. 1991. Planning language, planning inequality. London, New York. Longman. Varennes, Fernand de. 1996. Language, minorities and human rights. The Hague: M. Nijhoff.

ROGER BARNARD See also Australia; Canada; Language Planning; Nigeria; South Africa; Soviet Union: Successor States

Ojibwe and Algonquian Languages Ojibwe (sometimes Ojibwa, Chippewa) is a Native American language and belongs to the Algonquian family, one of the largest of precontact North America.

764

Although widely dispersed, these languages share many of the same grammatical features. Other members in this group include Blackfoot, Cheyenne and

OJIBWE AND ALGONQUIAN LANGUAGES Arapaho, Cree, Potawotami, Menominee, Fox, Illinois, Shawnee, Micmac, Passamaquoddy, and Abenaki. Algonquian speakers have inhabited North America for thousands of years. Their homeland was suggested to be somewhere in the region of Lake Huron or Georgian Bay, but given the ease of movement throughout the eastern woodlands area, most researchers are noncommittal in this regard. At the time of European contact, Algonquian languages were spoken along the eastern seaboard north of Virginia, as well as to the east of the lower Mississippi River. Today they can be found from the Atlantic Ocean to the Rocky Mountains along what is now the United States/Canadian border. Place names such as Connecticut, Manitoba, Massachusetts, Michigan, Minnesota, Mississippi, Ontario, Saskatchewan, and Wisconsin are all Algonquian in origin, as are the words moose, chipmunk, and toboggan. Linguists often speak of Algonquian in terms of regional subdivisions, but only Eastern Algonquian constitutes a family in the true sense; i.e. it can be shown that these languages have a common ancestor. Members of this group include Abenaki (spoken in Québec and Maine), Passamaquoddy-Maliseet (Maine, New Brunswick), and Micmac (Nova Scotia, New Brunswick). Arapaho, Cheyenne (Wyoming), and Blackfoot (Montana, Alberta) fall into the western group. ‘Central Algonquian’—of which Ojibwe is a member—is spoken around the area of the Great Lakes. Ottawa (Odawa), Algonquin, and Salteaux (all spoken in Canada) are essentially the same as Ojibwe; Cree, Fox, and Shawnee are close relatives. Divisions along geographical lines are not completely arbitrary: members of each region share certain sounds, vocabulary, and sentence patterns not found in others. The differences between one language and the next shift gradually from west to east, with western languages being the oldest. Algonquian was one of the first language families to be reconstructed by using methods developed in the nineteenth century for Indo-European. By comparing words in various languages, the American structuralist Leonard Bloomfield was able to determine the forms of ‘proto’ Algonquian, at the same time making significant contributions to our understanding of linguistic change itself. Since then, many other researchers have followed in Bloomfield’s footsteps, to the point where Algonquian is often looked on as a textbook case of structuralist analysis. Outside of taxonomy and analyses of sound change, not much systematic work has been done on Algonquian. Compared with other languages, our understanding of word and sentence structure is incomplete, and research on meaning and discourse is still in its infancy.

As a spoken language, Ojibwe is fairly robust, whereas several other Algonquian languages have declined or died out altogether. Most speakers live on reservations in Minnesota, Wisconsin, and Michigan (United States), as well as in the province of Ontario (Canada). One recent estimate puts the total number of Ojibwe speakers at 51,000, but most other estimates are lower. In any case, only a small percentage represents unilingual Ojibwe speakers. Many efforts are under way to preserve and/or revive the language. These usually take the form of bilingual education programs in the schools, language centers for adults, or the establishment of institutional or personal websites. Recently published word lists and dictionaries abound, and Ojibwe has received a relatively good deal of attention from linguists. Otherwise, the model of loss is similar to endangered languages worldwide: younger speakers are heavily influenced by the dominant culture (in this case, Anglo-America) and use their first language primarily in family or tribal situations. Borrowed speech habits continually erode native sound and sentence patterns until only the elders speak the language fluently. The sounds of Ojibwe are typical of other Algonquian languages. There are five vowels of different quality, and vowel length (indicated by a colon) distinguishes meaning. The vowels are [i, e, (schwa), i:, a:, and o:]. The consonants are [p, b, g, k, m, n, ch, j, w, y ,h, and ʔ]. Note the absence of [t, d]. Ojibwe (henceforth Algonquian) is a polysynthetic language, a term that carries with it many implications: single, complex words typically express the meaning of whole sentences, pronouns are ‘understood’ rather than overt, and expressions such as ‘all’ or ‘every’ may be separated from the nouns they modify by other words. In addition, polysynthetic languages are rich in agreement: subjects, direct objects, and even indirect objects cause the verb to inflect for person, number, and gender—and sometimes case. Noun phrases, meanwhile, do not appear in fixed positions within the sentence, as in English; instead they are considered ‘appositional’ (add-ons), only loosely linked to the verbs. Many meanings that are conveyed in full phrases in other languages are expressed via particles that attach to the verb. Verbs in Ojibwe (and all Algonquian languages) agree with subjects as well as direct objects, sometimes in very intricate ways. In the sentence [ni-ki:wa:pem-a:-n] kene:peko:n ‘[I saw] the snake’, the prefix ni- indicates that the subject is first person, whereas the absence of a plural marker ensures that it is singular. The third-person feature of kene:peko:n ‘snake’ is encoded by the suffix -a:, whereas -n signals that the (singular) object is different from the subject (cf. obviation below). Because of such rich agreement, there is no need for separate pronouns other than

765

OJIBWE AND ALGONQUIAN LANGUAGES emphatic ones. In the absence of an object, this sentence would simply mean ‘I saw him/her/it.’ When plural morphemes are added to the verb, agreement is ‘skewed’, yielding the order person prefix - person (verb stem) suffix SUBJECT OBJECT

number suffix SUBJECT

number suffix OBJECT

In other words, subject as well as object agreement is split between two nonadjacent positions. Various linguistic theories have attempted to deal with facts like these, but so far very few have succeeded in providing a natural account. Finally, nouns inflect for person, number, and obviation too, as in ni-gwis-ag [first person-son-plural] ‘my sons’. Note that often the same affixes involved in verbal agreement also show up in this context. A large class of elements called preverbs appear in fixed positions before the verb stem. These typically modify the state or action denoted by verb, much like modifier phrases do in English. Examples from Ojibwe are ojaanimi ‘busy, noisy’, wani ‘mistakenly’, and aano ‘in vain, without result’. Some preverbs are themselves verbal in character, such as giizhii ‘finish, get through doing’, gwiinawii ‘don’t know, not able to’, and goji ‘try, attempt’. Still others pertain to the way that a speaker perceives an event: bi ‘here, toward the speaker’, madwe ‘audible from a distance, can be heard’, ani/ni ‘going away from the speaker, approaching the time of doing something’. In short, preverbs have different functions that parallel phrasal expressions in nonpolysynthetic languages. Although it is possible for several preverbs to appear in a string, their order is not random. One of the characteristics of a polysynthetic language is the richness of word formation processes. In Ojibwe/Algonquian, many words are made up of multiple roots, not unlike English compounds, e.g. ‘killdeer’ (a kind of bird), ‘snow removal’, etc. Roots can be divided into initials, medials, and finals, depending on their position within the word. Initials carry most of the meaning, often describing processes or resulting states. Medials are usually nounlike, denoting body parts or instruments, and finals typically indicate simple states or actions. The complex Passamaquoddy stem napici koli-hqeh-mon ‘She or he sticks it onto something’ could be translated as ‘onto stuck-surface-make’ [preverb initial-medial-final]. Final roots in Ojibwe/Algonquian also specify the gender class of intransitive subjects and transitive objects (cf. the discussion of gender below). A common pitfall in understanding polysynthetic languages has been to ascribe variations in word form to the speaker’s viewpoint of the world, when in fact they probably result from regular grammatical processes.

766

The seemingly many names for snow in Eskimo (Yup’ik, Inuktitut) is a classic example. Similarly, Algonquian languages (like French) seem to contain an inordinate number of verbal paradigms. The difference, for example, between ‘if he comes’ and ‘when he comes’ depends on the change or deletion of a single vowel. The third major order in Algonquian is the imperative. The important thing to realize is that speakers do not memorize these myriad forms, but rather freely produce and understand them by internalizing a finite set of rules. Nouns in Ojibwe (Algonquian) are classified on the basis of animacy, roughly ‘being like a conscious living thing’. As one might expect, humans and animals are [animate], whereas rocks and similar types of objects are [animate]. The division is not always based on clear-cut properties, however. For example, trees are [animate], along with containers used for liquid (cups, spoons, pens, etc.). In Passamaquoddy (Eastern Algonquian), the words for rope, milk, and fallen snow are all [animate], as are those for fingernail and knee. Other body parts (heart, tongue) are [animate], however. The animacy of a given noun is overtly indicated by the verb-stem final, but only under certain grammatical conditions. Ojibwe speakers have a complex system of referring to things or people in a sentence or discourse. English normally uses pronouns like ‘he’ or phrases like ‘that guy’ to refer to someone after first introducing him by name. In Ojibwe, however, pronouns are not expressed overtly: one usually infers their presence through subject/object agreement on the verb or possessor agreement on the noun. The ‘obviation’ system guarantees that certain noun phrases (inaudible pronouns as well as audible noun phrases) are not confused with others. Within a complex noun phrase like niw wday-an ‘his dog’, for example, the possessive pronoun ‘his’ (which is only implied) would not be indicated by inflection, whereas niw ‘dog’ is marked as obviative (the suffix an). This ensures that ‘his’ and ‘dog’ do not refer to the same entity. The second domain of obviation is the clause. This can be seen in uki:-necci:we’a:-n [eniw kwi:wesse:ns-an] ‘He scolded [that boy]’, where the subject is unmarked and the object triggers obviation (underlined) on the verb as well as the noun phrase. As in possessed noun phrases, obviation within a clause is obligatory. In the third domain of obviation, subjects of main clauses are unmarked, but those in subordinate clauses are obviative (underlined): gii-boonii-w dash maa dVdibew [mitigoonsikaa-ini-g] ‘Then she landed on the shore [where there were bushes]’

Further restrictions determine whether subjects and objects are overtly expressed or implied. Although it is possible to say ‘I wrote them’ or ‘You hurt me’, for

OKANAGAN AND SALISHAN LANGUAGES instance, the opposite ‘They wrote me’ or ‘I hurt you’ cannot be expressed without changing the affixes of the verb. In the first two sentences, a ‘direct’ morpheme is attached to the verb stem, and in the latter two an ‘inverse’ one. The choice of direct or inverse is determined by a ‘participant hierarchy’ that ranks noun phrases according to person and obviation. In direct sentences, subjects outrank objects on the scale: 2 > 1 > 3 > 3—that is, second persons (‘you’) are ranked higher than first persons (‘I/we’), which in turn are higher than third persons (‘she or he’). Nonobviative third persons (3) are ranked higher than obviative (3) ones. In inverse sentences, exactly the opposite holds: objects must outrank subjects on the same scale. Strikingly, subjects in the inverse appear as objects in the direct, and vice versa: Direct ni-wa:pam-a:-k 1-see-third person/direct-pl ‘I see them’

Inverse ni-wa:pam-ik-o:k 1-see-third person/ inverse-pl ‘They see me’

Hierarchical effects such as those exhibited by Ojibwe/Algonquian pose a challenge to linguists attempting to explain them. References Aanii: an introduction to the central Ojibwe language. Video Series. Lessons in conversational Ojibwe as spoken at Wikwemikong, Ontario. Videotapes with Ojibwe and English subtitles, pauses for repetition, and additional vocabulary. Native Vision Productions, Box 23, Ohsweken, Ontario N0A 1M0, Canada (tel.: 519/445-0639; fax: 519/445-0639; e-mail: [email protected]; website: www.nativevision.homestead.com) Baker, Mark. 1996. The polysynthesis parameter. Oxford: Oxford University Press. Bloomfield, Leonard. 1946. Algonquian. Linguistic structures of native America, ed. by Hoijer Harry, 85–129. Viking

Fund Publications in Anthropology, Vol. 6. New York: Viking. ———.1957. Eastern Ojibwa. Ann Arbor: University of Michigan Press. Campana, Mark. 1996. The conjunct order in Algonquian. Canadian Journal of Linguistics 41(3). 201–34. *Campbell, Lyle. 1997. American Indian languages: the historical linguistics of Native America. New York: Oxford University Press. *Dahlstrom, Amy. 1991. Plains Cree morphosyntax. New York: Garland. Goddard, Ives. 1978. Central Algonquian languages. Handbook of North American Indians 15. 583–7. ———.1990. Primary and secondary stem derivation in Algonquian. International Journal of American Linguistics 56(4). 449–83. ———. 1994. The east-west cline in Algonquian dialectology. Papers of the 25th Algonquian Conference, ed. by W. Cowan, Carleton University, 187–211. Ottawa. Hockett, Charles. 1966. What Algonquian is really like. International Journal of American Linguistics 32. 59–73. Leavitt, Robert. 1992. Lexical exploration and educational insight. Papers of the 23rd Algonquian Conference, ed. by W. Cowan. Carleton University, Ottawa. ———. 1996. Passmaquoddy-Maliseet. Languages of the world/materials, Vol. 27. Newcastle: Lincom Europa. Lyovin, Anatole. 1997. An introduction to the languages of the world. Oxford: Oxford University Press. Mithun, Marianne. 1999. The languages of native North America. Cambridge: Cambridge University Press. Nichols, John. 1980. Ojibwe morphology. Doctoral Dissertation, Harvard University. *Rhodes, Richard. 1985. Eastern Ojibwa-Chippewa-Ottawa dictionary. Berlin: Mouton de Gruyter. ———. 1990. Obviation, inversion, and topic rank in Ojibwa. Berkeley Linguistic Society 16. 101–15. Siebert, Frank. 1967. The original home of the proto-Algonquian people. ed. by A.D. DeBlois, Contributions to anthropology: linguistics I (Algonquian), 13–47. Anthropological Series 78, National Museum of Canada Bulletin 214. www. ncs4.net/Ojibwe. The definitive link to Ojibwe linguistic and anthropological material. www.citilink.com. Nancy Vogt’s website, with brief descriptions of language and culture; bibliography of further references.

MARK CAMPANA

Okanagan and Salishan Languages Okanagan is one of four languages that belong to the Southern Interior branch of the Salishan linguistic family of North America. All but two languages of the family occupy contiguous territory that extends longitudinally from about 123°W to about 113°W, and latitudinally from about 52°N to about 45°N. In contrast

to early speculations that the Salishan languages had spread from an inland location (Boas 1905), recently scholars have proposed that speakers of the original language occupied a maritime area, and migrated outward from there, following routes to the interior along such major rivers as the Fraser and Thompson (Suttles

767

OKANAGAN AND SALISHAN LANGUAGES and Elmendorf 1963; Suttles 1987). The linguistic evidence adduced focuses on terms of flora and fauna (Kinkade 1991). Subgroupings: The family divides into five branches, the result of migrations. A single group is thought to have headed northwest and settled in the area where Bella Coola is now found, surrounded by non-Salishan languages, Wakashan seaside, and Athapascan inlandside. Another group is thought to have settled along the coast, and then spread further, one subgroup going southward to where Tillamook, now extinct, was spoken, and formed the third branch of the linguistic family. This language, too, is surrounded by non-Salishan languages, clockwise from the north: Chinookan, Athapascan, Takelman, and Maidu. A fourth group, forming the Tsamosan branch, also moved southward, and a fifth group, forming the Interior branch, moved eastward (see Figure 1). One now extinct Athapascan language, Nicola, was spoken in the approximate geographic center of the Salish area. The area where Okanagan is spoken by perhaps as many as 1,000 speakers of mostly mature age spans along the north–south expanse of the Okanagan valley from what is now Enderby to the south of Okanogan,

Washington, and westward in the Similkameen and Methow valleys; and along the north–south expanses of the Sanpoil and Kettle rivers, and the area west of the Columbia river as far as the bend around Wilbur, Washington. Dialectal differences are minor but remain to be described. (Figure 1 here) Phonology: In spite of the size of the family, and the number of different languages represented, it is possible to provide a generalized phonology of Salish (see Figure 2). The angled brackets notation should be read as either/or: with few exceptions, a Salishan language has one or the other series, the velar being the conservative one. Okanagan has the velar series. The northern dialects of Okanagan include voiced postpalatal resonants (γ', γ) that correspond to y' y in the other dialects. In the same northern dialects, the labialized pharyngeals have merged with their unrounded counterparts. Voiced stops are rare in Salishan languages, but occur in Coeur d’Alene, Twana, Lushootseed, and Comox, where they do not derive historically from proto-stops. Pharyngeals are found only in the Salish languages of the Interior; and no language, except Comox, has a phonemic nonejective lateral affricate.

Bella Coola Comox (Pentlatch) Sechelt Lillooet

Squamish SALISHAN

Halkomelem (Nooksack)

Shuswap Northern

N.Straits

Thompson Central

Coast

Interior Columbian

Straits

Okanagan

Klallam Southern Lushootseed

Kalispel (Twana) Tsamosan

Quinault

(L. Chehalis) (U. Chehalis)

Coeur d'Alene

(Cowlitz)

(Tillamook)

Figure 1

768

OKANAGAN AND SALISHAN LANGUAGES

Figure 2. Generalized Salish consonant inventory.

Okanagan has a symmetric three -vowel system, /i u a/, while the vowel systems of Salishan languages often include four vowels, sometimes five, with some asymmetries, notably the occasional absence of a high back vowel. Schwas are for the most part epenthetic. Salishan languages are wellknown for their propensity for consonant clusters; Bella Coola for several lexical items that consist entirely of voiceless obstruents. Morphosyntax: Salish languages have large numbers of affixes, predominantly suffixes. All Salish languages have developed lexical affixes, bound forms with lexical content, with a function that resembles that of incorporated nouns, and all Salish languages have several reduplicative patterns that mark primarily augmentative, diminutive, and inchoative forms. A nominalizer with the shape s- is nearly universally found, but homophonous aspectual prefixes also occur, notably in Coeur d’Alene and Okanagan. Person marking: Six Interior languages have different sets of person markers for transitive, intransitive, and subordinate predicates; the paradigms are less differentiated in Lillooet, closest to the languages of the coast, and in the noninterior languages. Possessive paradigms include prefixes and suffixes in all the languages. Okanagan has four main sets of person reference markers: the kn set (intransitive), the i(n)- set (possessive), the -(í)n set (transitive subject), and the (transitive) object set. The kn set consists of clitics (marked with the ligature) and a suffix: kn kw Ø

1sg 2sg 3sg

kwu p Ø ...-lx

1pl 2pl 3pl

These markers accompany stems that in English translate as intransitive verbs, nouns, and adjectives. kn ʔitx. I slept. kw sqilxw. You are an Indian/a person. ʔayχwt (axáʔ). This one is tired. A subset of these markers, identical in all persons except for 1sg kwu , co-occurs with the possessive set of person markers, and is reserved for double possessives and verb nominalizations. The possessive set, used with nouns, psych verbs, and verb nominalizations, consists of these markers

(prefixes and suffixes; parentheses abbreviate variants): i(n)a(n)-s/-c

1sg 2sg 3sg

-tt 1pl -mp 2pl -s-lx / -c-lx 3pl

which yield such forms as an-lʔíw in-χmínk

your father I like/want it

which, in turn, may combine with members of the kn set (kwu subset) to yield forms such as kwu an-1ʔíw I am your father. kw in-χmínk I like/want you (you are my wanting). kw i-ks-ʔam-t-ím an-lʔíw I am going to feed your father. The last is the nominalization of a future (ks-) possessor applicative (-t) verb form (root ʔam, feed), in which the suffix -(i)m, sometimes referred to as the antipassive, is required. The transitive subject set, often called the ergative set, consists of the following suffixes (parentheses abbreviate stressed and unstressed variants): -(í)n -(í)xw -(í)s

1sg 2sg 3sg

-(í)m /-t -(í)p -(í)s-lx

1pl 2pl 3pl

These markers follow the object markers, which, in turn, follow one of several obligatory transitive markers (see below). The (transitive) object set consists of the following markers (one proclitic and suffixes): kwu -s / -m -Ø

1sg 2sg 3sg

kwu ...-m -(úl)m -Ø... -lx

1pl 2pl 3pl

Because third -person object markers and thirdperson intransitive subject markers are Ø, Salishan languages are often characterized as split ergative systems. The allomorphy of the second singular object is transitivizer dependent. The disambiguation of number in the firstperson object is accomplished by the suffix -m and such forms are interpreted as 3rd indef subject - 1pl object: kwu sp’-nt-is He whipped me (-nt transitivizer). kwu sp’-nt-im They whipped us/We were whipped. 769

OKANAGAN AND SALISHAN LANGUAGES -(í)m occurs also with Ø, and the interpretation of these forms can be indefinite subject, or passive: sp’-nt-is sp’-nt-im

3rd person whipped 3rd person. 3rd person indef whipped 3rd person/3rd person was whipped.

Word classes: Aspectual criteria can be established to distinguish word classes, and, as expected, these may derive forms of other classes––nouns can derive verbs and verbs can be nominalized (for example, N. Mattina as reported in Kroeber 1999). A prototypical noun like kWilstn sweat lodge, culturally relevant and categorially marked (-tn instrumental), derives a verb with -m: kn kwilstn-m. I sweat bathed. Similarly, qwacqn hat derives qwacqn-m wear a hat (intransitive); ntχwχwin noon derives ntχwχwin-m do lunch (intransitive). Analogously, qiʔs dream (intransitive) derives s-qiʔs dream, and the latter form can be inflected with possessive markers and interpreted as a possessive noun form, or as a nominalized verb form. Most Okanagan stems can also be transitivized (see below). Nominal and pronominal arguments: Scholars have argued that Salishan languages are pronominal argument languages: a form like wik-nt-xw You saw it is a full sentence with a third-person object (Ø) and second-person subject (-xw). In this interpretation, any object expressed in nominal form is an adjunct, not a (nominal) argument. The claim is countered with the suggestion that in applicative sentences like kwu tq-t-is in-kílx. He touched my hand the noun phrase in-kílx my hand functions as one of the arguments of the possessor applicative verb form kwu tq-t-is He touched my ... and this argument is not, and cannot be, referenced in pronominal form on the verb. Intransitive forms are also analyzed as fully predicative. kn xwuy kW ilmíxwm kW χast

I went. You are a boss. You are fine.

In these sentences, the clitics kn and kw are the subjects, and the word to which the clitics are attached are the predicates. Third-person forms have Ø subject person marking, and forms like sql’tmixw have been analyzed as full predications that should be translated as something like ‘He is a man’ or ‘It’s a man.’ In the stream of discourse, such words can function as predicative elements. The normal way to express either of the isolated propositions ‘He’s a man’ and ‘It’s a man’ is with utterances like ixíʔ sql’tmixw That’s a man, or

770

sql’tmixw yaʔχís That one over there is a man; that is, by juxtaposing (in either order) the stem sql’tmixw and a deictic stem (ixíʔ, yaʔχís). In traditional terms, these sentences would be analyzed as exocentric equational constructions consisting of a subject and a predicate. The participant persons kn and kw are pronominal subjects; third-person forms can be analyzed as having a nominal subject of the classes mentioned, which, in context, can be deleted. Another complication for the interpretation of all full words as predicative is presented by the different markings for morphological and syntactic plurals: the morphological plural of citxw house is the reduplicated form ct-citxw houses, while the syntactic plural of the same form is citxw-lx (ixíʔ) (Those) are houses. In recent times, when scholars are preferring to view all constructions to have heads (or centers, in the old terminology), the question is raised as to what constitutes the head of such a sentence as kn sql’tmíxw. Most common is the hypothesis that the verb is the head of the sentence (here it would be the predicate nominal), but because the identification of head with lexical head can be dispensed with, just as abstract features within the Inflection or Agreement nodes have been proposed to head sentences, and just as the determiner has been proposed to head Determiner Phrases, so can kn be proposed to head the sentence kn sql’tmíxw. An utterance like xwuy He went, then, can be viewed as the abbreviation of xwuy ixíʔ That one went, and analyzed either as having a null subject, or as requiring a third-person nominal subject which undergoes deletion in the appropriate circumstances. Intransitive, possessive, and transitive paradigms: Beside the intransitive constructions already discussed, Okanagan uses kn inflection in a number of forms derived by means of prefixes, suffixes, and circumfixes. Among these forms are: To-Be nouns (kn k-noun). k-ilmíxwm snk’lip Coyote will be chief/is chief-to-be; k- have forms (kn k-noun). kn k-qwacqn I have a hat; inchoatives (kn verb+-ʔ- before stressed vowel). kn c’ʔax I got ashamed. (root c’ax); patient forms (kn verb+VC2). kn t’k’w-ak’w I fell. (cf. t’k’w-nt put something down); get patient forms (kn c+verb). u ilíʔ kn c-lak’ I was in jail a long time (cf. lk’-nt tie something); habitual/durative forms (kn c+verb). ilíʔkn c-wix I live there; imperfective forms (kn s-c+verb-(mi)x I have been X-ing); inceptive forms (kn ks+verb-(mí)xaʔx I am about to X); and past perfect forms (kn ksc+verb). kn ksc-nik’ I have cut some, I have some cut. Beside the possessive, double possessive, and psych forms discussed, other forms, some intransitive and others transitive (the latter always marked by the suffix -m), take possessive inflection: durative/intent forms (i+s+verb) s-q’sápiʔ-s ilíʔ i-s-ilíʔ I lived there a long

OLD CHINESE time (root ilíʔ there, lit. long-time there I-there); perfective forms (i+sc+verb). in-χást i-sc-ʔítx I slept well (my-good my-having-slept); future forms (i+ks+verb) lut a-ks-xwúy Don’t go; kw i-ks-(s)íw-m I’ll ask you; future imperative forms (i+kc+verb) lut a-kc-náq’w You will not steal. χast a-kc-k’wúl’-m You will work well; and future applicative forms (a-ks-verb-t-m) kw iks-may’-xít-m ... I am going to tell you ... Finally, all transitive forms take transitive person markers. Okanagan has two transitivizers, -nt and -st; a causative -st; three applicatives -t, -x(í)t, -tút; and three suffixes that prepare stems for transitivization: -nun, -min, -xix. Customary transitive forms are marked with the circumfix c-...-st, as in c-wik-st-n I always see it. References Boas, Franz. 1905. The Jesup North Pacific expedition Proceedings of the 13th International Congress of Americanists, 91–100. New York, Easton, PA: Eschenbach Printing Co. Czaykowska-Higgins, Ewa, and M. Dale Kindade (eds.) 1998. Salish languages and linguistics. Theoretical and descriptive perspectives. Berlin: Mouton De Gruyter. Kinkade, M. Dale. 1991. Prehistory of the native languages of the northwest coast. Proceedings of the Great Ocean

Conferences, Vol I: The North Pacific to 1600. Portland: Oregon Historical Society Press. Kinkade, M. Dale, William W. Elmendorf, Bruce Rigsby, and Haruo Aoki. 1998. Languages. Plateau Handbook of North American Indians, Vol. 12, ed. by Deward E. Walker (William C. Sturtevant, general editor), 49–72. Washington, DC: Smithonian Institution. Kroeber, Paul D. 1999. The Salish language family. Reconstructing syntax. Lincoln, NE: University of Nebraska Press. Suttles, Wayne. 1987. Northwest coast linguistic history. A view from the coast. Coast Salish Essays, ed. by Wayne, suttles. Vancouver: Talonbooks. Suttles, Wayne, and William W. Elmendorf. 1963. linguistic evidence for Salish prehistory. Proceedings of the 1962 Annual Spring Meeting of the American Ethnological Society, ed. by Viola E. Garfield, and Wallace L. Chafe. Seattle. Thompson, Laurence C. 1973. The Northwest. Current trends in linguistics, Vol. 10. Linguistics in North America, ed. by Thomas A. Sebeok, 979–1045. The Hague: Mouton. –––––– 1979 Salishan and the Northwest. The languages of Native North America: historical and comparative assessment, ed. by Lyle Campbell, and Marianne Mithun, 692–765. Austin: University of Texas Press. Thompson, Laurence C., and M. Dale Kinkade. 1998. Languages. Northwest Coast of Handbook of North American Indians, Vol. 17, ed. by Wayne Suttles. (William C. Sturtevant, general editor), 30–51. Washington, DC: Smithonian Institution.

ANTHONY MATTINA

Old Chinese Old Chinese may be subperiodized into Early Old Chinese, Middle Old Chinese, and Late Old Chinese. It is generally thought that most Early Old Chinese polysyllabic words became monosyllabic by or during the Middle Old Chinese period, so the Late Old Chinese language was overwhelmingly monosyllabic. Proto-Chinese, the ancestor of Old Chinese, is practically unknown. Although many scholars believe it to be related to the Tibeto-Burman family of languages in a ‘Sino-Tibetan’ genetic family, the theory remains controversial due to the lack of regular correspondences in phonology (sound structure), morphology (word structure), and syntax (sentence structure) between Chinese and Tibeto-Burman. Moreover, from Early Old Chinese times to the present, the Chinese language has spread into territory inhabited by peoples who originally spoke other languages, and for centuries it has been spoken by far more people than any other language in the world. Without a thorough comparative study of loanwords in languages known to have

bordered on the Chinese-speaking area in Antiquity, the affiliations of Chinese must be considered uncertain and the relationship with Tibeto-Burman likely to be due to convergence. Even if the Sino-Tibetan theory is correct, Proto-Chinese and the Tibeto-Burman languages must have diverged long before the Oracle Bone Inscriptions, the first documents written in Old Chinese. These texts appear in the mid-second millenium BCE and already contain distinctively Chinese phonological and syntactic characteristics that had developed in Proto-Chinese, before the language was first written down. For example, only one negative particle, *ma, is reconstructable for Proto-Tibeto-Burman, but the Oracle Bone Inscriptions have two negative roots, *pa- and *ma-. This distinction is preserved in all later forms of Chinese, including modern Mandarin, where it is still found in numerous bound forms and in the free negative forms bù [pu] and méi [mej]. The Old Chinese writing system contains important information for the reconstruction of the language.

771

OLD CHINESE Although it includes many characters with no phonetic elements, such as guî ‘tortoise, turtle’ (originally a pictograph), some characters are phonetically ‘borrowed’. (N.B.: Pronunciation of characters is given in Mandarin unless otherwise noted.) For example, the originally pictographic character used to write the word for ‘wheat’ was borrowed to write the then homonymous word lái ‘to come’. Most characters are actually constructed of two or more parts, of which one part is a ‘phonetic’ element and another is an often simplified semantic or ‘significant’ element usually called the ‘radical’, such as zhuî ‘short-tailed bird’ (originally a pictograph) and *zhuî (now usually pronounced, irregularly, huán) ‘grass used for making mats’, written with the ‘grass’ radical as the significant and zhuî as the phonetic. Although other examples––such as lu˘ ‘blunt, stupid; Lu, the home state of Confucius’ and its phonetic yú ‘fish’––have differences that are important for reconstruction (see below), the phonetic information contained in the script is not by itself sufficient to allow reconstruction of Old Chinese phonological structure. In fact, Chinese is not well attested phonetically until the seventh century CE, when Middle Chinese forms begin to be recorded in Old Tibetan alphabetic script. The phonology of Old Chinese can therefore only be recovered through reconstruction. Modern reconstructions of Old Chinese nearly all belong to a tradition that may be called Historic Sinological Reconstruction. It derives ultimately from the early Chinese grammarians, on whose work the pioneering Sinologist Bernhard Karlgren based his system of reconstruction. Although the appearance of most Old Chinese reconstructions has changed radically since the publication of Karlgren’s etymological dictionary Grammata Serica recensa (1957; A Chinese Grammar, Revised), Historic Sinological Reconstruction is still based on the method pioneered by Karlgren. The most important sources used in the traditional Historic Sinological Reconstruction approach to Old Chinese phonology are: the phonological information derivable from analysis of the characters themselves (mostly mid-second millenium BCE to second century BCE) and their variants or substitutions; the rhymes of the Shijing (late first millenium BCE; Book of Songs); Chinese transcriptions of known foreign words (c. first century CE onward); books by Chinese grammarians on dialects and rhymes (first century BCE onward), especially the lost Middle Chinese Qieyun (601 CE; Cut Rhymes); the rhyme books, particularly the Guangyun (Extensive Rhymes) and Jiyun (Collected Rhymes), both from the eleventh century CE; and, above all, the rhyme tables, especially the Yunjing (Mirror of Rhymes), from the twelfth century CE. The latter three works, although compiled much later, are based ulti-

772

mately on the Qieyun. However, unlike earlier works, the Yunjing organizes the material in tabular form. Historic Sinological Reconstruction of Old Chinese depends heavily on the projection back in time of the categories established by the tables in the Yunjing, adjusted according to the rhyme categories implied by the Shijing poems and to phonological information derivable from the characters themselves. Change in one part of the system necessarily entails change in many other parts in order to maintain consistent correspondences between the categories of the tables and the phonological categories of Middle Chinese and the modern Chinese dialects, hence the reconstructions of one scholar’s system often look radically different from those of another. Unfortunately, Historic Sinological Reconstruction does not allow subperiodization of Old Chinese––contemporary proponents of the method explicitly claim that they are not actually trying to reconstruct a real language spoken by real people in different places and at different times but only a theoretical construct from which all later forms of Chinese can be derived––so it is impossible to systematically use contemporaneous data to check the reconstructions themselves. This results in a tendency to ignore attested data in favor of the system. For example, Sergei Starostin (1989) reconstructs ‘woman; you (in the latter sense also written )’ as Old Chinese *nraʔ even though overwhelming Oracle Bone Inscription and other evidence indicates that the word initial must have been a bilabial in Early Old Chinese and apparently well into the Middle Old Chinese period. Another result is the production of many unlikely forms, such as Old Chinese * srjʔ(s), the reconstruction of ‘affair, matter’ by William Baxter (1992). Although scholars working in this tradition have made progress toward the reconstruction of Old Chinese, recently their method has been challenged. Reconstructions have been proposed that are not based on the categories of the rhyme tables but on contemporaneous data, emphasizing Old Chinese character variations, loanwords found in neighboring languages, and Middle Chinese reconstructions grounded in foreign transcriptions and the earliest attested text of the Qieyun (706 CE). Early Old Chinese (called by some scholars ‘preArchaic Chinese’), the language of the Yin, or late Shang dynasty period (fourteenth to eleventh centuries BCE), is preserved mainly in the Oracle Bone Inscriptions, which are divination texts inscribed on bones and turtle shells. These texts have been found in great numbers in the area of the lower Yellow River valley. The syntax of Early Old Chinese, although slightly different from that of the earliest classical texts from a millenium later in the late Middle Old Chinese and Late Old Chinese periods, is already SVO (Subject–Verb–Object) in basic sentence word-order,

OLD CHINESE as in modern Chinese. Several scholars have argued that Early Old Chinese must have had many disyllabic words or roots. Reconstructable examples include Early Old Chinese *Cwena (where ‘C’ stands for an unknown consonant) ‘woman; you’, and Early Old Chinese *kara ‘bitter’. However, Early Old Chinese phonology has yet to be studied in depth. Middle Old Chinese is essentially the language of the post-Shang bronze and stone inscriptions and the earliest transmitted Classical texts. The inscriptions on bronze vessels are often dated, and cover the period from the early Zhou dynasty (c. 1145 BCE onward) into the Warring States period (475–221 BCE). Some of them contain rhymed passages and variant characters from which phonological information can be derived. The Middle Old Chinese period may be divided into three subperiods: the early Middle Old Chinese of the Western Zhou dynasty, when the capital was located in the area of present-day Shaanxi Province and the prestige language was influenced by the local dialect; the middle Middle Old Chinese of the Eastern Zhou, when the capital moved eastward into another dialect area; and late Middle Old Chinese, the language of the earliest transmitted ‘Confucian’ classical texts, including the Shijing. These periods are marked by phonological changes, partly motivated by prestige-dialect shift when the capital moved to a new dialect region, and partly due to change over time. One example must suffice here. The usual first person singular pronoun (‘I’) in early Middle Old Chinese is (also written ) yú, but in middle Middle Old Chinese (beginning with the Stone Drum inscriptions, c. fifth to sixth centuries BCE) this word is replaced by wú, which by Late Old Chinese displaces yú except in deliberately archaizing texts. The character wú ‘I’ has as its phonetic element the ˘ ‘five’, while in other middle Middle character wu Old Chinese texts it is generally written instead with yú ‘fish’ as its phonetic. Because the Old Tibetan numerals are known to be cognate with (derived from the same historical source as) Chinese, Old Tibetan Iŋa ‘five’ is cognate with the Chinese word for ‘five’. (Whether the Tibetan numerals are borrowed from Chinese or inherited from a common ‘Sino-Tibetan’ ancestor is irrelevant for present purposes.) Since yú (from Middle Chinese *ŋɔ1) ‘fish’ is also the phonetic in the character lu˘ (from Middle Chinese *lɔ2) ‘Lu’, it is clear that not only must the Middle Old Chinese form of the word for ‘five’ have been close to *lŋâ, but it and the word for ‘fish’, the first person pronoun, ‘Lu’, and other words, such as yu˘ ‘speak, speech’, were pronounced the same in one or more of the Middle Old Chinese dialects. Because the character used to write the earlier first person pronoun yú is also used as a phonetic in other characters, it is reconstructable as *lâ, which derives from an earlier *laCa. It is clear that these

synonyms must be related. Since *lŋâ, one dialect form of the Middle Old Chinese first person pronoun, may be reconstructed as *laŋa, the other dialect form, *lâ, may be reconstructed as *laγa, both deriving from an Early Old Chinese *laga or *lege. This word is evidently related in turn to another early first person pronoun, Early Old Chinese *aga or *ege. During the Middle Old Chinese period, most previously disyllabic morphemes became monosyllabic. This change produced many homonyms by late Middle Old Chinese––such as *lŋâ ‘first person pronoun’ ~ ‘five’ ~ ‘fish’ ~ ‘speak’—and resulted in other drastic phonological changes, including widespread metathesis (segments changing place) in one or more dialects. For example, middle Middle Old Chinese *lŋâ ‘fish’ (from Proto-Chinese *laka) became late Middle Old Chinese *ŋlâ or *ŋrâ. The development of monosyllabism was also probably a major motivation for the eventual development of phonemic tone (intonational pitch distinguishing meanings). The shift to a monosyllabic morpheme structure was largely complete by the time the Shijing was recorded. Late Old Chinese is characterized by a number of phonological changes, including that of the Middle Old Chinese syllable-initial *l to either *y [j] or *d, of Middle Old Chinese medial *l or *r to *, and of the Middle Old Chinese syllable-final *r to *n or *y [j]. (The rules for these changes are still unclear.) The change of final *r to *n took place in the ancient central dialect and is found in all modern Chinese dialects, but it was not completed in some peripheral dialects, notably in the northeast, where ancient final *r was retained as a syllable coda at least into the Middle Chinese period. Another significant change includes the beginning of what has been called the Great Chinese Vowel Shift, in which the Middle Old Chinese vowel *â (or [a˘]) became Middle Chinese *ɔ and Mandarin u. Thus, late Middle Old Chinese *ŋlâ (or *ŋrâ) ‘fish’ became Late Old Chinese *ŋâ, Middle Chinese *ŋɔ1, Late Middle Chinese *ŋü1, and Mandarin yú [ü]. Some traces of Late Old Chinese forms are found in the modern Southern Min dialect. (Chinese is otherwise internally reconstructable only back to Middle Chinese.) The written Late Old Chinese of the Former Han dynasty (206 BCE to 9 CE) became the standard Classical Chinese literary language, which continued in use into the twentieth century and is still used for some purposes even today. References Baxter, William H. 1992. A handbook of Old Chinese phonology. Berlin: Mouton de Gruyter. Beckwith, Christopher I. 2002. The Sino-Tibetan problem. Medieval Tibeto-Burman Languages, ed. by Christopher I. Beckwith. Leiden: E. J. Brill.

773

OLD CHINESE Beckwith, Christopher I. 1999. Review of Laurent Sagart, The roots of Old Chinese. Amsterdam: John Benjamins; in Anthropological Linguistics 44.2 (2002). Benedict, Paul. 1972. Sino-Tibetan: a conspectus. Cambridge: Cambridge University Press. Karlgren, Bernhard. 1957. Grammata Serica recensa. Stockholm: Museum of Far Eastern Antiquities; reprinted 1972. Pulleyblank, Edwin G. 1962. The consonantal system of Old Chinese. Asia Major 9.

––––––. 1996. Early contacts between Indo-Europeans and Chinese. International Review of Chinese Linguistics 1.1. Sagart, Laurent. 1999. The roots of Old Chinese. Amsterdam: John Benjamins. Schuessler, Axel. 1987. A dictionary of early Zhou Chinese. Honolulu: University of Hawaii Press. Starostin, Sergei A. 1989. Rekonstrukcija drevnekitajskoj fonologieskoj sistemy (Reconstruction of the Old Chinese phonological system). Moscow: Nauka.

CHRISTOPHER I. BECKWITH

Old Church Slavonic Old Church Slav(on)ic (OCS) is the language of a group of Slavic texts containing copies of religious works, almost all originally translated from Greek. The corpus includes eight parchment manuscripts having 100 or more folia, and an approximately equal number of fragments. A half-dozen inscriptions survive from the same period and area. No manuscript bears an explicit date or locale, but paleographic and linguistic features indicate that most of the surviving manuscripts were copied in the eleventh century CE, with a very small number possibly from the late tenth. The original translations are generally assumed to go back to the second half of the ninth or early tenth centuries, and were connected with the mission of Cyril and Methodius to the Slavs and the activities of their pupils in Bulgaria following the death of Methodius in 885 CE. Every manuscript deviates in some ways from normalized Old Church Slavonic, but as a group they show an Eastern South Slavic dialect which has often been called Old Bulgarian, and it is in fact very close to the reconstructed Late Common Slavic (LCS), which is usually considered to have lasted until around the end of the first millennium CE. With the passage of time, this liturgical and literary language took on certain local characteristics in each area where it remained in use, producing Russian/Bulgarian/ Serbian, etc. recensions of Church Slav(on)ic, which remained mutually comprehensible and collectively served as a common literary language for the

high nonhigh

774

front

nonfront

¯ / ˘ e¯ / ˘e

u¯ / ˘u a¯ / ˘a

Orthodox Slavs and some other peoples, such as the Romanians. Old Church Slavonic was written in two alphabets: Glagolitic, invented (or, less likely, edited and brought into its final form) by St. Cyril, perhaps with the participation of his brother Methodius, for their mission to the Slavs in the 860s, and Cyrillic, invented later, probably in Eastern Bulgaria, and named in honor of Cyril. Cyrillic was based upon the Greek uncial letters; Glagolitic has no immediately obvious resemblance to any other alphabet, although some features compare suggestively to the Greek cursive of the time. Glagolitic was used in some parts of Croatia until the early twentieth century, but elsewhere (Kievan Rus’, Bulgaria, Serbia, etc.) was quickly displaced by Cyrillic, which has remained in use among the Orthodox Slavs until the present day. Late Common Slavic introduced far-reaching changes to the sound system of late Indo-European (IE). As in some other dialects of IE, *o and *a conflated to *å, with the long vowel giving Slavic a and the short vowel giving Slavic o. IE * e¯ became LCS * e (in OCS, a low front vowel [æ]), and IE * ˘e went to LCS *e. IE * ¯ gave LCS *i, IE * ˘ gave LCS *ь, a high short (lax) front vowel. IE *u¯ eventually became *y, a high unrounded nonfront vowel (*su¯nus > *synъ ‘son’), and IE * ˘u became LCS *ъ, a high short (lax) back vowel. This gave a vowel system with binary oppositions in height, fronting, and length (long/short):

realized as OCS

high nonhigh

front

nonfront

i/ь e / e

y/ъ a/o

OLD CHURCH SLAVONIC plus diphthongs. Those diphthongs in *w and *y were converted to monophthongs: *ow > *u; *ew > *ju; *oy > *e or *i; *ey > *i. Diphthongs in *r and *l were transposed (*gord- > *grad- ‘enclosed area,’ *xold- > xlad‘cold’, and diphthongs in *m or *n became nasal vowels (*ronka > *ro ka ‘hand,’ *se¯men > *se me ‘seed’). As a result, the vowel system of OCS contained i, ь, e , e, a, o, y, ъ, u, e , o ,with oppositions based upon front(i, ь, e , e, e )/back, high(i, ь, y, ъ, u)/low, rounded(ъ,u, o, o )/unrounded, tense/lax, and oral/nasal. Slavic is a satem language, as shown by such OCS words as sъto ‘hundred’ and zlato ‘gold.’ The consonant system shows only a two-way contrast, between voiced and voiceless, with a division into labial, dental, palatal (not palatalized), and velar. Stops were p b t d k g, continuants included s z š D x, and the affricates were c (ts),  (dz), J, and in some dialects,  . Sonorants were m n l r n¸ (=nj), ¸l(=lj), ¸r (=rj), glides contained w and j.

LCS syllable structure had undergone a number of significant changes. Syllables in LCS were always ‘open’ Syllable-closing consonants were dropped: *supnos > *sъnъ (su ˘ nu ˘ ). The basic syllable formula was strV, where s stands for s or z, t for almost any consonant, and r for a sonorant, although the details are quite complex. The consonant clusters that actually occur can be quite formidable: OCS umrъštvl¸o ‘I will kill.’ Syllables also showed a type of ‘synharmony’, where consonants and vowels were subject to mutual accommodation, such as fronting of vowels after j or palatal consonants, the ‘palatalization’ of the velars k g x to J D š or c  s before a front vowel, or ‘jotization’ of combinations of consonants with j: sj > š, kj > J , etc. This produced paired endings and consonant alternations in several instances. For examples of the changes, see the o/jo and a/ja declensions given below, including the forms of vlьkъ and ro ka given in the same tables.

TABLE 1 (a) Singular Gloss ‘city’ ‘man’ ‘wolf’ ‘village’ ‘field’ ‘woman’ ‘soul’ ‘hand’ ‘bone’ ‘son’

stem o jo o o jo a ja a ˘ ˘u

nom. gradъ mo Dь vlьkъ selo pol¸e Dena duša ro ka kostь synъ

voc. grade mo Du vlьe selo pol¸e Deno duše ro ko kosti synu

acc. gradъ mo Da vlьkъ selo pol¸e Deno dušo ro ko kostь synъ/-a

gen. grada mo Da vlьka sela pol¸a Deny duše ro ky kosti synu

loc. grade

mo Di vlьce

sele

pol¸i Dene

duši ro ce

kosti synu

dat. gradu mo Du vlьku selu pol¸u Dene

duši ro ce

kosti synovi

instr. gradomь mo Demь vlьkomь selomь pol¸emь Denojo dušejo ro kojo kostьjo synъmь

dat. gradomъ mo Demъ vlьkomъ selomъ pol¸emъ Denamъ dušamъ ro kamъ kostьmъ synъmъ

instr. grady mo Di vlьky sely pol¸i Denami dušami ro kami kostьmi synъmi

(b) Plural meaning ‘city’ ‘man’ ‘wolf’ ‘village’ ‘field’ ‘woman’ ‘soul’ ‘hand’ ‘bone’ ‘son’

nom./voc. gradi mo Di vlьci sela pol¸a z eny duše ro ky kosti synove

acc. grady mo De vlьky sela pol¸a Deny duše ro ky kosti syny

gen. gradъ mo Dь vlьkъ selъ pol¸ь Denъ dušь ro kъ kostьjь synovъ

loc. grade xъ mo Dixъ vlьce xъ sele xъ pol¸ixъ Denaxъ dušaxъ ro kaxъ kostьxъ synъxъ

(c) Dual meaning ‘city’ ‘man’ ‘wolf’ ‘village’ ‘field’ ‘woman’ ‘soul’ ‘hand’ ‘bone’ ‘son’

nom./voc./acc. grada mo Da vlьka sele

pol¸i Dene

duši r o ce

kosti syny

gen./loc. gradu mo Du vlьku selu pol¸u Denu dušu ro ku kostьju synovu

dat./instr. gradoma mo Dema vlьkoma seloma pol¸ema Denama dušama ro kama kostьma synъma

775

OLD CHURCH SLAVONIC TABLE 2 1 sg. 2 sg. 3 sg. 1 dual 2 dual 3 dual 1 pl. 2 pl. 3 pl.

present

root aorist

s-aorist

extended a.

imperfect

imperative

vedo vedeši vedetъ vedeve

vedeta vedete vedemъ vedete vedo tъ

vedъ vede vede vedove

vedeta vedete vedomъ vedete vedo

ve sъ vede vede ve sove

ve sta ve ste ve somъ ve ste ve se

vedoxъ vede vede vedoxove

vedosta vedoste vedoxomъ vedoste vedoše

vede axъ vede aše vede aše vede axove

vede ašeta vede ašete vede axomъ vede ašete vede axo

– vedi vedi vede ve

vede ta – vede mъ vede te –

Inflected words in OCS may be divided into nominal and verbal groups, with the nominals subdivided into nouns, adjectives, and pronouns. Numerals pattern with nouns or pronouns; ‘one’ through ‘four’ and their compounds also show gender agreement. Both verbs and nouns show number (singular, plural, and dual). Nouns distinguish up to seven grammatical cases marking subject (nominative), direct object (accusative), possession (genitive), location (locative), indirect object (dative), instruments/means (instrumental), and also a form of address(vocative). Nouns have inherent gender (masculine, feminine, and neuter). Verbs distinguish six tenses, and have indicative, imperative, and conditional forms; they can be active or passive and indicate completed vs. incompleted actions. Noun markings in OCS fall into five patterns (declensions); in Indo-European terms they are o/jo, a/ja, ˘ , u˘, and consonant stems. The difference between the o and jo stems and between the a and ja stems are due largely to the intrasyllabic changes referred to above as ‘synharmony’. The o and the u˘ stem endings have intermixed so much by the time the OCS texts were copied that one can no longer really set up u˘-stems as a separate declension. Consonant stems come in several varieties, e.g. rstems such as mati, gen. sg. matere ‘mother’; s-stems such as nebo, gen. sg. nebese ‘heaven’; en-stems such as vre me , gen. sg. vre˘mene ‘time’ and kamy, gen. sg. kamene ‘stone’. The productive declensions are the o/jo, a/ja, and ˘ -stems. Some of the u˘-stem endings occur frequently and even become productive in the o/jo declension.

between definite and indefinite; definite adjectives are made from indefinite adjectives by appending pronominal suffixes to the indefinite forms; thus, nova re ka means ‘a new river,’ but novaja re ka means ‘the new river’; novo selo means ‘a new village’ and novoje selo (o > e after j by ‘synharmony’) means ‘the new village.’ The comparative (more...) is formed by suffixation; the superlative (most ...) is made syntactically from the comparative. Although no single OCS verb is attested in all possible forms, the conjugation of ved- ‘lead’ must have been as shown in Table 2. Word order in OCS is most often S(ubject) V(erb) O(bject), but can be quite free. It is often quite difficult to know when OCS word order is following that of the Greek from which the text was originally translated and when it represents truly Slavic usage. All grammatical cases except the nominative occur with prepositions, but all, including the locative, occur without prepositions. Under the influence of Greek, OCS syntax often becomes rather convoluted. OCS vocabulary is primarily of Indo-European, or at least Balto-Slavic origin, although it also contains borrowings from several other sources, including Iranian (e.g. bogъ ‘god,’ rajь ‘paradise’) and Germanic (e.g. xle bъ ‘bread,’ kupiti ‘buy’). As one would expect, most of the religious terminology and an overwhelming majority of the abstract, legal, philosophical, administrative, and didactic vocabulary is either direct borrowings or loan translations (calques) from Greek. References

Sample Declensions See Table 1. Adjectives are inflected and show agreement in gender, number, and case. Masculine and neuter adjectives decline in principle like o/jo-stem nouns, while feminine adjectival forms are essentially like those of the a/ja-stem nouns. Adjectives also distinguish

776

Huntley, David. 1993. Old Church Slavonic. ed. by Bernard Comrie and Greville G. Corbett. The Slavonic Languages, London, New York: Routledge. Lunt, Horace G. 2001. Old Church Slavonic Grammar, 7th revised edition. Berlin, New York: Mouton de Gruyter. Schenker, Alexander M. 1995. The Dawn of Slavic. An Introduction to Slavic Philology. New Haven: Yale University Press.

CHARLES GRIBBLE

OLD ENGLISH

Old English Old English (OE) is the collective name given to the varieties of Germanic brought to Britain by the Jutes, Saxons, and Angles in the fifth and sixth centuries. The event separating Old from Middle English is the Norman Conquest of 1066. The root Engl- in the names Engl-ish and England < Engla-land ‘land of the Angles’ represents a form of the Latin and Common Germanic name of the tribe Angli. The OE language is also sometimes referred to as Anglo-Saxon (AS), although more frequently a distinction is made between the OE language and AS history, law, literature, culture, etc. The closest linguistic relative of OE is Old Frisian. OE and Old Frisian belong to the West Germanic subgroup of the Germanic branch of Indo-European. The date usually associated with the beginning of OE is 449 CE, when Germanic-speaking warriors, led by the legendary brothers Hengist and Horsa, sailed to Britain to fight against the Picts at the invitation of the Celtic king Vortigern. From the middle of the fifth century onward, Germanic settlers arrived in Britain in considerable numbers. The Jutes remained mostly in Kent, parts of Hampshire, and the Isle of Wight, and the Saxons occupied the lands south of the Thames, as well as Middlesex and Essex. The Angles spread westward and as far north as the Scottish Lowlands. The Germanic invasions and settlements resulted in the partial displacement of the indigenous Celticspeaking populations from central England to the more southern, western, and northern parts of the country. The most common periodization of OE is into early OE, from the beginnings to c. 800 CE, classical OE, c. 800–950, and late OE, c. 950–1100. Two notable historical events that had a profound effect on AS culture and the vocabulary of OE are the conversion to Christianity, initiated by the arrival of St. Augustine in Kent in 597, and the Viking invasions and settlements. The Vikings, who were Scandinavian seafaring warriors, began their raids on Northumbria in the late eighth century. In the ninth century, their attacks culminated in the establishment of a separate administrative unit, the Danelaw (< Dena lagu ‘law of the Danes’), which included large parts of the northern, central, and eastern regions of England, roughly northeast of a line linking London and Chester. The treaty which established the Danelaw, negotiated by the West Saxon king Alfred the Great, who ruled from 871 to 889, provided for a relatively peaceful period during the tenth century. At the beginning of the eleventh century, the balance of political power shifted again

toward Scandinavia: between 1016 and 1042, England was ruled by the Danish king Cnut and his sons.

Orthography The earliest OE written records are runic inscriptions dated c. 650–700 CE. They were either records of individual names, or brief decorative messages. The runic alphabet, originally linked to Germanic pagan rituals, was abandoned after the adoption and spread of Christianity during the sixth and seventh centuries. The alphabet used subsequently was a modified form of the Roman alphabet, known as the Insular hand in which the letter was written in its long form , and appeared as , known as ‘yogh’. The three specifically OE letters, not used in the Roman alphabet, are ‘ash’, for the vowel [æ], and ‘thorn’ and ‘eth’ (or ‘ðæt’), used interchangeably for the consonant sounds [ð], as in that, or [θ], as in thump. Consonant letter combinations specific to OE writing are: , which represented [sk] for most of the period, but probably also [ʃ] after c. 1000, and , which stood for [].

Dialects The three original groups of settlers, the Jutes, the Saxons, and the Angles, maintained their dialect differences throughout the AS period. The main dialects of Old English are Kentish, West Saxon, East Saxon, Mercian, and Northumbrian. Mercian and Northumbrian were two varieties of Anglian, spoken to the south and north of the river Humber. Prior to the Scandinavian invasions, Northumbrian was a dialect of great prestige due to the religious, artistic, and intellectual achievements of the Northumbrian kingdom. After the end of the ninth century, the kingdom of Wessex unified and dominated the rest of the country, and the new political and cultural situation led to the spread of the West Saxon literary norms to the neighboring dialects. West Saxon is the tenth century literary koiné of Anglo-Saxon England. Classical West Saxon is the variety of OE described in the standard reference works, and it is also the dialect on which the following descriptions are based.

Phonology The system of OE stressed vowels consisted of seven short vowels, seven long vowels, and two diphthongs. The colon sign in parentheses is the phonetic symbol for vowel length. Vowel length was not marked in the

777

OLD ENGLISH OE manuscripts, although many modern printings use the convention of a macron (-) over a vowel letter to indicate that it is long. Simple vowels

Diphthongs

i(:)

y(:)

u(:)

e(:)

eo or eə æa or æə

o(:) æ(:)

ɑ(:)

The most important phonological change, which occurred prior to the first OE written records and changed the quality of many vowels in some positions, was I-Mutation, also known as I-Umlaut. This is a right-to-left (regressive) vowel harmony phenomenon: back vowels became front and low front vowels were raised before /i, j/ in the same word. (The dashed line indicates the direction of the sound change from early OE to late OE and ME.) i(:) mal ‘bad’, ARBOREM > arbre ‘tree’, GUTTA > goute ‘drop’); introduced the affricates [t d ts dz] through palatalization, as in OF chalt ‘hot’, jorn ‘day’, cent ‘hundred’, onze ‘eleven’, respectively; began a long-term process of vowel nasalization culminating in forms such as MF un [œ] ‘one’ < UNUM, chien [j] ‘dog’ < CANEM, pont [p] ‘bridge’ < PONTEM; and deleted or weakened consonants between vowels, as in VITA > vie ‘life’, SAPONEM > savon ‘soap’. In word structure, the changes were no less dramatic. The loss of final vowels had profound effects on many grammatically relevant suffixes. Among the nouns and adjectives, the original CL system used suffixes to distinguish five separate classes of nouns, three genders (masculine, feminine, and neuter), two numbers (singular and plural), and six cases (nominative, genitive, dative, accusative, vocative, ablative, each with various syntactic roles indicating, among other things, subject [nominative case], direct object [accusative] and indirect object [dative], possession [genitive], terms of address [vocative], and so on). These were reduced in OF to three classes of nouns, two genders, two numbers, and two cases. With the disappearance of the final vowels, the suffixes disappeared in large part, so that the syntactic distinctions originally indicated by these suffixes were also directly affected. As a result, we see compensating developments in the increased use of prepositions (e.g. le livre de Cicéron ‘the book of Cicero = Cicero’s book’ rather than LIBER CICERONIS), and OF word order patterns become more constrained than in CL (where, because suffixes indicated the words’ function, words could appear in virtually any order depending on stylistic factors [CANEM [‘dog–accusative’] HOMO [man––nominative] VIDET > l’homme voit le chien ‘The man sees the dog’]). The determiner system also underwent radical change, with great increases in the use of demonstratives (e.g. OF cest ‘this’, cel ‘that’) and particularly articles (e.g. OF li, le, la, les ‘the (sg. and pl.)’, uns, un, une ‘a (masc. and fem.)’).

1 The symbol ‘>’ indicates the direction of development: Latin BREVIS becomes (>) OF brief. Conversely, ‘ OF chanterai ‘I will sing.’ Verb endings also eroded, so that person/number distinctions (I, you, he/she/it [sg.]; we, you, they [pl.]), formerly indicated by suffixes, were also threatened. As a result, pronoun use gradually increased, especially in the singular as a way of preserving the former differences. The form of OF words was significantly more variable than in the modern language. The effects of earlier sound change often resulted in different forms of noun and verb stems that were regularized subsequent to the OF period (OF truef––trovons > MF trouve–– trouvons ‘find’ [1 sg.––1 pl.]; serf––sers > MF serfs ––serfs ‘serf’ [sg.––pl.]; OF larc––large [masc.–– fem.] > MF large ‘wide’). Fusion of unstressed forms is also widespread: a le > au ‘to + the (sg.)’, en les > es ‘in + the (pl.)’, je le > jel ‘I + him/it’, ne les > nes ‘neg + them’, qui me > quim ‘who + me’, si me > sim ‘if + me’, and numerous others. In its sentence structure, OF still permitted relatively free word order compared to the modern language, but certain patterns, particularly those where the verb is in the second position in the sentence, began to predominate: subject–verb–object: Li vilains apele son fil. ‘The peasant calls his son.’ [Fabliaux 3, l. 39]; the frequent object–verb–subject (– X): Ses barons fist li rois venir. ‘The king had his barons come.’ [Le roman de Renart, l. 1807]; De venoison ont grant plente. ‘They

781

OLD FRENCH have a lot of venison.’ [Tristran, l. 1773]; and subject–object–verb: Li rois Tristran menace. ‘The king threatens Tristran.’ [Tristran, l. 770]. Given the disappearance of the CL passive verb forms, it is not surprising that replacements of these CL constructions appeared based on estre ‘to be’ plus the past participle (La pucele fut donc pendue. ‘They hanged the young girl’ [Vie de sainte Marguerite, l. 1]); on the increased use of pronouns such as on/l’en (On me desrobe en vostre terre. ‘I am being robbed in your land.’ [Fabliaux 11, l. 191–2]); or on the expansion of constructions with the reflexive pronoun se (Carles se dort. ‘Charles falls asleep.’ [Chanson de Roland, l. 724]). OF sentence patterns, in other words, remind us much more of the modern language than of CL. In vocabuary, the Latin origins of OF are clear ––the great majority of lexical items descend directly from Latin, although often from informal or popular rather than classical speech. In addition to popular items, which have become the norm (TESTA ‘jug’ > OF teste ‘head’; CABALLUS ‘nag, packhorse’ > OF cheval ‘horse’; GENUS ‘knee’ — GENUCULUM [vernacular diminutive form] > OF genoil [Modern French genou] ‘knee’), we see innovative use of many suffixes (e.g. OF -age used to indicate taxes of various types: abeillage, arivage, cheminage, melage [on bees, docking, roads and apples, respectively]; Germanic loans from the early Frankish conquerors (biere ‘coffin’, helme ‘helmet’, honte ‘shame’); learned words entering from religious or legal texts (avaricieux ‘avaricious’, crestiien ‘Christian’, testimonie ‘testimo-

ny’); and importations from a variety of other sources (Celtic chemin ‘road’, if ‘yew tree’; Arabic alchimie ‘alchemy’; Greek eglise ‘church’; Occitan amour ‘love’ [notably via the influence of the Troubadour poets], and many others). Nonetheless, the unmistakably major lexical source is Latin. But despite this parentage, OF is clearly French, no longer a ‘corrupted’ Latin, and has distanced itself from its origins much more than such well-known Romance relatives as Italian, Occitan, Catalan, or Spanish. While not easily comprehensible to modern readers, it leaves a stong impression of familiarity, and should inspire us as a source of great cultural and linguistic richness. References Buridant, Claude. 2000. Grammaire nouvelle de l’ancien français. Paris: SEDES. Einhorn, E. 1974. Old French. A concise handbook. Cambridge, New York: Cambridge University Press. Harris, Martin. 1978. The evolution of French syntax. A comparative approach. London, New York: Longman. Kibler, William. 1984. An introduction to Old French. New York: The Modern Language Association of America. Pope, Mildred. 1934. From Latin to Old French with especial consideration of Anglo-Norman. Manchester: Manchester University Press. Price, Glanville. 1971. The French language: present and past. London: Edward Arnold.

DOUGLAS C. WALKER See also France; Indo-European 4: Romance; Latin

Old High German The German language is estimated to be about 1,500 years old based on the earliest written sources, though as a spoken language it is almost surely older. Old High German is the traditional designation of the German language in its earliest stage, roughly between the years 750 and 1050 CE (some scholars use the period 600–1100). German is a member of the Germanic subfamily of the ‘centum’ branch of the Indo-European language family. The linguistic features that set Germanic off against other Indo-European languages are several, principal among which are the First Sound Shift (also called the Germanic Sound Shift, the First or Germanic Consonant Shift, or Grimm’s Law), Verner’s Law, the

782

fixing of main intonational stress on the initial syllable of the word, and a variety of vowel changes. The origins of the Germanic peoples––the speakers of the ancestor language Proto-Germanic––are cloaked in mystery, as indeed are the origins of the IndoEuropeans. We are relatively certain that sometime prior to 1000 BCE Germanic-speaking people occupied their primeval home (Urheimat) in what is today the area comprising southern Sweden and Norway, Denmark, and northern Germany. Between 1000 BCE and 500 BCE, at least some of the Germanic tribes began to move away from their original home, migrating farther north or striking out to the south and the east. The reason for this migration was doubtless due in

OLD HIGH GERMAN part to inundation of the Urheimat, much of which is covered today by the relatively shallow North Sea. Out of this Völkerwanderung (‘migration of peoples’) arose the traditional classification of the Germanic languages into East, North, and West Germanic. German belongs to the latter group, as do English, Dutch, Frisian, and Low German. We know nothing for certain about the people who occupied this territory before the Germanic tribes arrived. We know nothing about the early contacts between the Germanic intruders and the autochthonous inhabitants of the Urheimat, nor do we know what language(s) the latter spoke. The general rule of thumb in historical linguistics is that the greater the degree of contact between languages, the greater the amount of language change. The changes in pronunciation and vocabulary that Germanic languages underwent (the First Sound Shift and others) were extensive compared with more conservative Indo-European languages. Moreover, approximately one-third of the vocabulary of the Germanic languages is from other than Indo-European stock. Therefore, we assume that contacts with the autochthonous inhabitants were deep and extensive––and that the languages of these speakers had many fricative consonants (the source of English f, th, s, sh).

East Germanic

While there are competing representations for illustrating the relationships among the Germanic family depending on which criteria are given priority, the most widely used and traditional genetic classification of the Germanic languages is a tripartite one into East, North, and West. The family tree (stammbaum) of ProtoGermanic can be represented (and simplified) as follows. (The terms ‘Ingvaeonic’ and ‘Istvaeonic’ refer to the names of Germanic tribes; ‘North Sea Germanic’ and ‘Weser-Rhine Germanic’ are used almost interchangeably with ‘Ingvaeonic’ and ‘Istvaeonic’. Old Low Franconian gave us Dutch, and Old Saxon was the ancestral language of modern Low German, which is spoken today in northern Germany.) (See Figure 1.) The principal linguistic change that set Old High German apart from its other West Germanic sibling languages was the Old High German Consonant Shift, also called the Second Consonant Shift or the Second Sound Shift. This, along with the First Sound Shift differentiating Germanic from Indo-European, is one of the most illustrious sound changes in Germanic linguistics. It affected the voiceless stops /p t k/, changing them according to the phonetic environment into affricates /pf ts kx/ or fricatives /f s x/, the latter of which could occur as both long (geminated) /ff ss xx/

Gothic (†)

Icelandic West Norse Old Norwegian North Germanic Old Danish East Norse Old Swedish Proto-Germanic

Old English

Old Frisian Ingvaeonic Old Low Franconian

West Germamic

Old Saxon

Istvaeonic

Old High German

Figure 1

783

OLD HIGH GERMAN and short /f s x/. In general outlines, the Old High German Consonant Shift went as follows: Pre-Old High German ptk >

Old High German pf ts kx

>

ff ss xx

>

fsx

Phonetic Environment Word-initially and following m, n, l, r Medially following a short vowel Medially following a long vowel and word-finally

We can see the effects of the Old High German Consonant Shift from a comparison of modern German words with their English cognates: Modern German Pfund Pfeffer Wasser machen Zeit helfen auf kochen

English pound pepper water make tide help up cook

An additional sound change of Pre-Old High German d to Old High German t is considered by some linguists to be part of the Old High German Consonant Shift. Its effects are seen in cognates such as English do, day, ride vs. German tun, Tag, reiten. The designation ‘High’ has two meanings in the context of High vs. Low German. On the one hand, it is a purely linguistic designation for those dialects of West Germanic that underwent any part of the Old High German Consonant Shift. Its original meaning in this context, however, referred to altitude: those dialects of German that did not undergo the Old High German Consonant Shift are located in northern Germany on the flatlands bordering the North and Baltic Seas––‘low lying’in other words, hence ‘Low’ German––and the dialects of German that did undergo the Old High German Consonant Shift are located in the higher regions of southern and central Germany, Switzerland, and Austria, thus ‘High German’. One occasionally encounters the use of ‘High’ to mean ‘good, correct, elevated, superior, grammatical’ and ‘Low’ to mean the opposite, linguistically, of those qualities. One does not encourage this usage. Old High German is divided into Central German and Upper German dialects. Upper German consists of those dialects spoken in southern Germany, Switzerland, and Austria. The principal Upper German dialects of Old High German are Alemannic, spoken in present-day Switzerland, and Bavarian, spoken today in the German state of Bavaria and in Austria. Central German dialects

784

are spoken in an east–west band lying between Low German in the north and Upper German in the south. The principal Central German dialects are Franconian (Middle Franconian, Rhine Franconian, East Franconian) and East Central German, which overlaps with Thuringian, Upper Saxon, and Silesian. Upper German dialects are characterized by greater completeness in the effects of the Old High German Consonant Shift. It is assumed therefore that the Old High German Consonant Shift began in the far south and spread toward the north: the underlying idea is that sound change goes further toward completion in the area it has been around in longest. Most Upper German dialects have all of the changes stated above as comprising the Old High German consonant shift. Central German dialects show more mixed results. Central German dialects are most consistent in the shift of k to x and less consistent in the shifts of t to s and p to f. Thus, where Upper German dialects will consistently have fricatives in ich ‘I’, dorf ‘village’, and das ‘that’, Central German dialects can be found with ich, dorf, dat and ich, dorp, dat. The distinctive sounds of Old High German are: Vowels Long à u¯ e¯ o¯ a¯

Short i u e o a

Stops Fricatives Affricates Nasals Liquids Semivowels

vl. vd. vl. vl.

Diphthongs ei ou ie uo

Consonants Labial Dental Palatal Velar Glottal p t k b d g x f s s. ts kx pf m n l r w y h

([s] and [s.] are respectively dorsal––formed with the blade of the tongue against the alveolar ridge––and apical––formed with the tip of the tongue against the alveolar ridge. They have different historical sources: [s] is created by the shift of t to s in the Old High German Consonant Shift; [s.] continues Indo-European s). The back vowels [u o a u¯ o¯ a¯] had front variants [ü ö e æ ü¯ ö¯ æ¯ ] before syllables containing a high front vowel or semivowel [i à y], thus gesti ‘guests’, skôni ‘beautiful’, and ubil ‘evil’ were phonetically [gesti skoni übil]. These are the so-called ‘umlaut’ sounds of German (Gäste, schön, übel). The German language is traditionally divided into four periods: Old High German (750–1050); Middle High German (1050–1350); Early New High German

OLD HIGH GERMAN (1350–1650); and New High German (1650–present). The division between Old High German and Middle High German is dated from the reduction of the unstressed vowels [i e a o u] into a single sound schwa [ə]. Since this did not occur in a single day, the transition from Old High German to Middle High German is variously placed between 1050 and 1100, at which time the reduction of unstressed vowels to schwa was complete except in remote and isolated dialects. From the beginning of the Common Era, Old High German was in contact with Romans during the expansionist phase of the Roman Empire. Lexical borrowings were heavy, and their semantic domains reveal much about the nature of these early contacts: ‘wine’ (Old High German wîn, Latin vinum), ‘arrow’ (Old High German pfîl, Latin pilum). The earliest written records of the Old High German language are words and fragments written in runes. Somewhat later, after the Germanic tribes had been Christianized, we find Old High German glosses interspersed in Latin texts. In the Vocabularius Scti Galli, slightly later than 765 CE, we find sapiens wizzo (in normalized Old High German spelling), scitus wiser, fortis stark. Lengthier texts are found beginning in the ninth century, many of them translations into Old High German of religious writings originally in Latin: Isidor’s diatribe Contra Judaeos, the Weissenburg Catechism, the gospel harmonies Tatian and Otfrid, and the translation of Boethius made by Notker, an especially talented monk from the monastery of Saint Gall in Switzerland. Typical of the genre of most early Old High German writing is the Lord’s Prayer from St. Gall: Fater unsêr, dû bist in himile, wîhi namun dînan, kweme rîhhi dîn, werde willo dîn, sô in himile sôsa in erdu. brôd unsêr emezzihig gib uns hiutu, oblâz uns skuldi unsêro, sô wir oblâzêm uns skuldîgêm, enti ni unsih firleiti in khorunka, ûzzer lôsi unsih fona ubile.

Typical too were ‘How to say it in Old High German’ booklets such as the ‘Conversations from Paris’, written presumably for monks literate in Latin for traveling in Germany (the unnormalized spelling of Old High German points clearly and amusingly to the French background of the monk who wrote the thing): Gueliche lande cumen ger? (de qua patria?) ‘What land do you come from?’ Guer is tin erro? (ubi est senior tuus?) ‘Where is your master?’ Ne guez. (nescio.) ‘I don’t know.’ Gimer min ros. (da mihi meum equum.) ‘Give me my horse.’

Of particular cultural and historical interest are the ‘Oaths of Strassburg’, which date from 843. These are oaths of allegiance intended to end a fratricidal rivalry

between the kings Louis and Charles of the Franks, sons of Charlemagne. Because the eastern Franks spoke Old Low German and the western Franks Old French, the oaths were composed in both languages, thus giving us linguistic information as useful for a student of the history of the French language as for a student of the history of German. There is in Old High German a certain amount of original composition not derived from Latin translations and more substantial than ‘give me my horse’. The Muspilli is a religious poem of some length that is strikingly original, what we have of it, and in the later Old High German period there are Memento Mori and The song of Ezzo. It must be admitted that Old High German literature is inferior in comparison with the contemporaneous literature of Old English (Beowulf) or even its close relative, Old Saxon (Heliand). It is vastly inferior to the rich medieval literature of Middle High German with its courtly epics, its Nibelungenlied, its poetry. Old High German literature is not equal to the slightly later literature of Old Icelandic. There is very little original in Old High German literature, much that is derivative. However, its literary deficiencies should not detract from its linguistic importance. It was the language of the Second Sound Shift, the Old High German Consonant Shift, which along with the First Sound Shift (Grimm’s Law) and Verner’s Law is one of the great defining phonological events in the Germanic family of languages. The Old High German Consonant Shift—its causality, its spread, how it happened—has provided and continues to provide historical linguistics with grist for its theoretical mill. The importance of a language must never be judged by its literature alone. References Barber, C. Clyde. 1951. An Old High German reader. Oxford, Blackwell. Braune, Wilhelm. 1987. Althochdeutsche Grammatik. Tübingen: Max Niemeyer (numerous editions). Hawkins, John A. 1987. ‘German’ The world’s major languages, ed. by Bernard Comrie. New York: Oxford University Press. Lockwood, W.R. 1968. Historical German syntax. Oxford: Clarendon Press. Murdoch, Brian O. 1983. Old High German literature. Boston: Twayne Publishers. Priebsch, R., and W.E. Collinson. 1962. The German language. London: Faber and Faber. Waterman, John T. 1976. A history of the German language. Seattle: University of Washington Press. Wells, C.J. 1985. German: a linguistic history to 1945. Oxford: Clarendon Press. Wright, Joseph. 1906. An Old High German primer. Oxford: Clarendon Press.

ROBERT D. KING See also Germany; Grimm, Jacob; Indo-European 2: Germanic Languages

785

OLD IRISH

Old Irish Old Irish is a now extinct Indo-European language belonging to the Insular Celtic branch of the Celtic language family. Insular Celtic is that branch of Celtic that is (relatively speaking) indigenous to the British Isles. Old Irish is the direct ancestor of present-day Irish, as well as of Scots Gaelic and Manx. Within earlier Insular Celtic, it is possible to distinguish two major dialects, known respectively as QCeltic (sometimes called Goidelic or Gaelic) and P-Celtic (sometimes called Brythonic or Britannic). These labels have been retained by linguists to classify the later distinct Celtic languages. Q-Celtic is so called because it shows /k/ as the descendent of an original Indo-European /*kw/. (The name ‘Q-’, rather than ‘K-’, Celtic originates in Latin spelling.) In contrast, P-Celtic shows /p/ as the descendent of /*kw/. Irish, along with Scots Gaelic and Manx, belongs to the Q-Celtic group, while the P-Celtic group contains Welsh and Cornish, as well as Breton. We may contrast the development of the inherited word for ‘four’, which in Old Irish became cethair but in Old Welsh became petguar. The earliest surviving evidence for Q-Celtic comes from inscriptions written in the native Ogham alphabet, which date from the fourth to seventh centuries CE. These consist primarily of proper names, and so the amount of grammatical information that they can provide is limited; however, they do provide information about the sound system and word structure of QCeltic in the immediately pre-Old-Irish period. The Old Irish period proper is normally considered to fall between 600 and 900 CE, and it is from this period that the earliest literary evidence for Irish comes. Within the Old Irish period are sometimes further distinguished an Archaic Old Irish period (prior to the 700s CE) and a Later (or Classical) Old Irish period (the 700s and 800s CE). Ogham writing, incidentally, continued to be taught until around the seventeenth century, but its role after the inscriptional period of Q-Celtic was greatly restricted and most Old Irish material is thus written using a form of the Roman alphabet. The corpus of contemporary texts that represents Old Irish is relatively limited in both size and genre. The most important texts, dating from the Classical Old Irish period, consist of marginal notes (or ‘glosses’) on Latin manuscripts. Of these, the most important are the Würzburg glosses, dated mainly to around 750 CE, the Milan glosses, normally considered to date from sometime before 825 CE, and the St. Gall

786

glosses, dating in manuscript from around 845 CE but containing copies of earlier material. To these texts can be added a number of further sources consisting of additional glosses as well as a small number of continuous texts such as poems. In addition to the Classical Old Irish texts, there also exist a small number of older texts from the 600s CE. It is generally agreed that further Old Irish material is also preserved in later manuscripts, and, indeed, it has been argued that some of this material illustrates a period of Old Irish that pre-dates the earliest contemporary manuscripts. Certainly, some of these texts do show differences in comparison with other texts of the same date, notably in word order. However, this is a disputed area, and it has been argued convincingly that at least some of this variation may be due more to genre than to date. The texts may also have come under the influence of Latin models, and some of their distinctive characteristics may in part reflect interference from Latin. Thus, the contemporary Old Irish manuscripts still form the main, and safest, basis for the linguistic description of Old Irish, although scholars vary as to how far they will admit evidence from the later manuscripts. The Old Irish texts that have come down to us are linguistically very homogeneous and show rather little evidence of dialect variation. The classical view has been that what variation does exist is due to differences in date, although it has been tentatively suggested that these differences do, in fact, reflect dialectal variation. Certainly, the relative homogeneity of the Old Irish sources should not be taken to indicate that some degree of dialect variation did not exist within Old Irish; indeed, it would be unusual if this were the case, since the language was spoken in the whole of Ireland and in a large part of Scotland. We do not know how small or substantial these dialect differences may have been in the spoken language. As regards the homogeneity of the literary sources, it should be remembered that they were written by scholars in monastic communities, and some degree of dialect mixing, and possibly informal standardization, will undoubtedly have taken place. The vocabulary of Old Irish is predominantly of QCeltic origin, but also contains substantial loan material, primarily from P-Celtic and from Latin. Some of the Latin loans have undergone sound changes that suggest they were borrowed indirectly via P-Celtic, while others seem to have been borrowed directly.

OLD IRISH In structural terms, Old Irish has a number of features that are interesting from a historical and comparative perspective. On the one hand, the evolution of Old Irish to modern Irish shows trends toward the breakdown of complex systems of functional endings (inflexions), simplification, and the atomization of meaning-carrying elements that are paralleled in other language groups––for instance, in the development from Latin to modern Romance languages such as Italian and French. In the noun, for instance, Old Irish maintained a complex set of inflexions: for example, it had special dual forms for pairs of things, as well as a set of five case endings indicating grammatical relations such as subject, object, and possession. The same was also true of the adjective, which agreed with the noun that it modified. Although already largely obsolete in Old Irish, owing to the emergence of a fairly fixed word order, this system of endings was to be retained well into the modern Irish period (up to the 1600s CE). However, it has since broken down substantially, and must already have been breaking down at an earlier date in the spoken language. Some modern Irish dialects have practically abandoned noun inflexion, and even the more conservative standard written language uses a reduced set of endings. The Old Irish verb, too, was exceptionally complex, and has consequently received much attention from linguists. The basic distinction was between simple verbs and complex verbs, that is, those that were formed from simple verbs by the addition of various preverbal elements. For instance, from beirid (‘bear’) could be derived compound verbs such as do·beir (‘give’), for·beir (‘grow’), as·beir (‘say’), and tremi·beir (‘transfer’). In addition to these compounding preverbal elements, other preverbal affixes existed: for example, ro was used to distinguish, among other things, between the perfect and simple past (contrast English ‘I have run’ and ‘I ran’). Each of the two classes of verb had, in turn, two full sets of personal endings, which were used in parallel to one another, depending on the grammatical context: in the simple verb, they were known as absolute and conjunct endings, and in the compound verb as prototonic and deuterotonic endings. In addition, there were special endings in the simple verb to indicate a verb’s participation in a relative clause. Furthermore, as with other languages such as Latin, pronoun subjects were not expressed, and the meaning was carried by endings (contrast the unchanging bear in English ‘I bear’ and ‘you bear’ with Old Irish biru [‘I bear’] and biri [‘you bear’]). One of the main changes distinguishing Old Irish from later periods of Irish is the breakdown, and consequent simplification, of this complex verbal system, notably the disappearance of the two sets of

absolute and conjunct endings and the near disappearance of the prototonic and deuterotonic distinction. On the other hand, Old Irish also has features that are characteristic of itself and of the wider Celtic language family, in contrast with other Indo-European languages. In common with the other Celtic languages, Old Irish has Verb–Subject–Object as its basic word order. The Celtic languages are the only Indo-European languages to have this as their basic word order. There has been some discussion about the emergence of this word order in Old Irish. One suggestion has been that it is the consequence of generalizing the order that resulted from suffixed pronouns (see below). In Old Irish, as in other Celtic languages, the boundaries between the different language levels of sentence structure, word structure, and sound structure are not discrete. For instance, the initial sounds of Old Irish words can change in certain contexts (so-called ‘mutations’). These changes were originally triggered by the final sound in the preceding word, but a number of these sounds later disappeared and the pattern of mutation was also extended by analogy to other contexts. Some mutations thus came to be used to mark grammatical relationships, rather than being the result of local phonetic influence. This is the case, for example, in certain kinds of relative clause, where the main verb of the relative clause undergoes mutation. Thus, sentence function triggers a change in wordform, which was originally phonetically motivated. A further peculiarity of Old Irish is the frequency with which certain grammatically interrelated words within a sentence or clause are fused into a single word unit, where, in contrast, other languages (such as English) would normally use two or more free-standing words. One example of this is the situation where the object of a verb is a pronoun. In Old Irish, pronoun objects were often suffixed to the verb, rather than standing on their own as separate words–for example, beirthi (‘bears it’) or iurrus (‘will wound them’). Following one of the preverbal elements referred to earlier, a pronoun could actually become infixed within the verb—for example, ro-m·gab (‘has taken me’). In contrast to this Old Irish tendency, pronouns in modern Irish may also be free-standing, and the use of this option is becoming increasingly frequent, at the expense of the suffixed pronoun construction. A related phenomenon, which modern Irish has retained, is the case of the so-called ‘conjugated prepositions’, in which forms of the pronoun are also suffixed to prepositions, resulting in ‘one-word’ expressions such as lat (‘with you’, from la = with and -t = you). Aside from its importance as a crucial chapter in the history of modern Irish and the other Q-Celtic languages, Old Irish is linguistically important more

787

OLD IRISH generally for two main reasons. Firstly, from a historical perspective, it provides further evidence for the universality of certain principles of language change. Secondly, it is interesting typologically, in that it exhibits a number of features not normally found in Indo-European languages outside the Celtic language family. References Ahlqvist, Anders. 1988. Remarks on the question of dialects in Old Irish. Historical dialectology, regional and social, ed. by Jacek Fisiak. Berlin, New York: Mouton de Gruyter. Lewis, Henry, and Holger Pedersen. 1937. A concise comparative Celtic grammar. Göttingen: Vandenhoeck and Ruprecht.; 3rd edition with supplement, 1989.

Mac Eoin, Gearóid. 1993 Irish. The Celtic languages, ed. by Martin J. Ball and James Fife. London, New York: Routledge. McCone, Kim. 1996. Prehistoric Old and Middle Irish. Progress in Medieval Irish studies, ed. by Kim McCone and Katharine Simms. Maynooth: Department of Old Irish, St. Patrick’s College. ––––––. 1997. The Early Irish verb. Maynooth: An Sagart. Sims-Williams, Patrick. 1984. The double system of verbal inflexion in Old Irish. Transactions of the Philological Society 82. Strachan, John. 1904. Old Irish paradigms and selections from the Old Irish glosses. Dublin: School of Irish Learning ; 4th edition, revised by O. Bergin, Dublin: Royal Irish Academy, 1949. Thurneysen, Rudolf. 1946. A grammar of Old Irish. Dublin: Dublin Institute for Advanced Studies.

ANDREW WILSON See also Celtic Languages

Old Japanese Old Japanese is the oldest attested representative of the Japonic (Japanese-Ryukyuan) language family. There are two major varieties of Old Japanese: (1) Western Old Japanese (seventh to eighth century CE), a language based on the dialect of Asuka and Nara regions (roughly corresponding to modern Nara prefecture); and (2) Eastern Old Japanese (eighth century CE), a dialectal continuum located roughly in the southern part of the modern Chûbu and Kantô regions. Western Old Japanese was the basis of the literary language of the time, and consequently there are many more extant Western Old Japanese texts than Eastern Old Japanese texts. Furthermore, it is quite apparent that the westernmost dialects in the Eastern Old Japanese dialect continuum were considerably influenced by the Western Old Japanese standard, and this influence gradually diminishes from west to east. There is a widespread but mistaken opinion that Western Old Japanese is practically identical to Middle (Classical) Japanese. On the contrary, both languages are very different, with a number of important distinctions found in pronunciation, vocabulary, and grammar. Furthermore, Western Old Japanese is not even a direct predecessor of Middle Japanese: both are based on geographically close but not identical dialects. Sources: All Eastern Old Japanese and most Western Old Japanese texts are poetry, although Western Old Japanese also has two texts that are written in prose. The major noninscriptional texts are: poetry from the ‘Kojiki (Records of Ancient Matters)’

788

(712 CE), the ‘Nihonshoki (Annals of Japan)’ (720 CE), and the ‘Fudoki (Gazeteers)’ (713–737 CE); poetic anthology ‘Man’yôshû (Collection of myriad leaves [of words])’ (c. 759 CE), prose texts ‘Senmyô (Imperial Edicts)’ (seventh to eighth century), and ‘Norito (Liturgies)’ (seventh to ninth century). There are several important inscriptions as well, the largest of them being a poetic one: ‘Bussoku seki uta (Songs of the Buddha’s footprint)’ (c. 755 CE). Writing system: Western Old Japanese uses the writing system known as ‘man’yôgana’ (the ‘Man’yôshû’ writing). Man’yôgana is a syllabic system of writing, where Chinese characters are used phonetically (to represent syllables), although they can also be used logographically (to represent words or concepts). When Chinese characters are used both phonetically and logographically, there is a clear tendency to write roots or stems logographically and suffixes and particles phonetically; however, due to the syllabic nature of writing, the exact boundaries between stems (especially verbal stems) and suffixes may not be reflected in writing. Man’yôgana can be subdivided into two major types according to the historic stage of the underlying Chinese pronunciation of the characters on which man’yôgana is based. The later type is based on the Late Middle Chinese character readings of the eighth century, and it is used exclusively in the Nihonshoki. The earlier type is based on the Early Middle Chinese character readings of the sixth to seventh century; most likely it was not coined

OLD JAPANESE directly in Japan on the basis of these readings, but was borrowed from Korea. All other Western Old Japanese texts use this earlier type of the man’yôgana. Eastern Old Japanese also uses this type of Western Old Japanese writing system, and it is necessary to note that all Eastern Old Japanese texts are essentially written in Western Old Japanese orthography. Phonology: Western Old Japanese has the following consonants: p, b [mb], t, d [nd], k, g [ŋg], m, n, s, z [nz], w, y, r. No consonant is possible in the syllable-final position; thus, only syllables of C(onsonant)V(owel) and V structure are found in the language. Consonants b, d, g, z, and r occur only in word-medial position in native vocabulary. There are no sequences *wu and *yi in Western Old Japanese, but the rare exceptions to the general syllable template of [C]VCVCV…, like kai ‘oar’, suggest that at least the sequence *yi existed once at the pre-Western Old Japanese stage. The Eastern Old Japanese consonant system seems to be identical to Western Old Japanese. Western Old Japanese vowels include the following vowels (Yale notation is given with the most likely phonetic values in brackets): a [a], yi [i], u [u], iy [ï], ye [e], o [ə], wo [o], and one diphthong ey [əy], which was previously considered to be a unit vowel [ε]. The contrast between vowels yi/iy, ye/ey, and o/wo that later merged as /i/, /e/, and /o/ is not found in every possible position even in Western Old Japanese. Thus, there is no contrast in initial position for any of the three pairs; there is no contrast between yi/iy and ye/ey after coronal consonants (such as t, d, s, z), and a suggested contrast between po and pwo is questioned by some linguists, although it is likely that there is some supportive evidence for it in the earliest texts. In all cases where there is no contrast, it is customary to write just i, e, o. The set of Eastern Old Japanese vowels is markedly different: it includes only five vowels: a, i, u, e, o, although there is the possibility that one or two additional vowels can be hidden behind the Western Old Japanese spelling system. In any case, vocalic correspondences between Western Old Japanese and Eastern Old Japanese are not straightforward, indicating that a parent language of both used to have a system somewhat different from its descendants. Data on Western Old Japanese accent system are controversial: it has been suggested that in the ‘Nihonshoki’, Chinese characters with level tone were used to transcribe Western Old Japanese low pitch and Chinese characters with other tones to transcribe Western Old Japanese high pitch, but it appears to be true only statistically. Lexicon: Japanese poetic texts did not use words that, at the time, were obvious loanwords until quite late, so poetic texts contain only native vocabulary and

nativized loanwords, such as e.g. uma ‘horse’, umey ‘plum blossom’ (both from Old Chinese *mra and *mi), and tera ‘Buddhist temple’ (from Old Korean *tyerV). A few loanwords can be seen in prose texts, most from Chinese or Sanskrit (via Chinese intermediary). Morphology: The description of word structure and word formation in the present article is based on Western Old Japanese, because of its prestigious nature and due to the less complete data from Eastern Old Japanese. Both Western Old Japanese nominal and verbal morphology are significantly different from any later stages of the language, including Middle Japanese (morphological elements and their functions typical only for Western Old Japanese and not found in later stages of the language are indicated in bold typeface). Nominal morphology: Most nominals (with the exception of some pronouns) in Western Old Japanese have no formal markers, distinguishing them from other parts of speech, e.g. yama ‘mountain’, pyito ‘person’, yukyi ‘snow’, myidu ‘water’, piy ‘fire’, puta ‘2’, towo ‘10’. The following is the list of the most frequent WOJ affixes. Nominal prefixes: honorific myi-, intensive ma-, dimunitive wo- and kwo-, locative sa-. Plural suffixes: -ra (neutral plural marker), -domo, -tati, -na. Case markers: active-i, possessive -ga, genitive -no, genitive-locative -tu, dative-locatives -ni and -ra, accusative-absolutive -wo, comitative -to, ablative ywo/-ywori/-yu/-yuri, directives -gari and pye, comparative -no/-ni/nasu/-nosu, and terminative -made. Dimunitive suffixes: -ra, rø -kwo. Some Western Old Japanese pronouns have two stems: unextended, and extended with suffix -re, which may be treated as a formal marker of these pronouns, distinguishing them from other parts of speech, and also from other pronouns that do not have the special stem in -re. Personal pronouns: wa-/ware 1ps and 1pp (rare), a-/are 1ps, na 1ps, na-/nare 2ps and 2pp (rare), [myi]masi 2ps, si/so- 3ps. Reflexive pronoun: ono-/onore. It is necessary to note that Middle Japanese has a unified system of personal reflexive pronouns (...self), which does not exist in Western Old Japanese or Eastern Old Japanese. Demonstrative pronouns: proximal ko/kore ‘this’, mesial so-/sore ‘that’, distal ka/kare ‘that over there’. Demonstrative pronouns indicating place or direction: proximal: koko, koti, konata ‘here’; mesial: soko ‘there’; distal: kanata ‘over there’. The distal forms in Western Old Japanese and Eastern Old Japanese may be an innovation, as they are rather rare. In Western Old Japanese there is another distal demonstrative pronoun woti/woto/wote that is used more frequently than ka/kare, but its usage is predominantly limited to indicating distal place or

789

OLD JAPANESE time. Interrogative pronouns: ta-/tare ‘who’, nani ‘what’, ika ‘how’, idu/iduku ‘where’, iduti ‘where to’, idure ‘which’, itu ‘when’, iku/ikuda/ikura ‘how many’, nado/nazo ‘why’. Collective pronouns: miyna, moro ‘all’. The numerical system of Western Old Japanese sharply contrasts with later stages of the language; numerals of Chinese origin are not yet used, or at least they are not present in the texts. The main problem with attestation of numerals is that they are frequently written logographically, and the phonetic attestations are lacking. The following cardinal numerals are attested phonetically: pyito ‘1’, puta ‘2’, myi ‘3’, yo ‘4’, itu ‘5’, mu ‘6’, nana ‘7’, ya ‘8’, kokono ‘9, towo ‘10’. After towo ‘10’, attested tens include pata ‘20’, myi-swo ‘30’, yo-swo ‘40’, and ya-swo ‘80’. There is only one phonetic attestation of a numeral including both tens and digits: myiswo-ti amar-i puta-tu ‘32’ (30-class exceedinf 2-class) that, alongside later glosses to numerals written logographically in Western Old Japanese texts, provides us with the way in which the digits were added to tens: tens + classifier + infinitive of the verb amar- ‘to exceed’ + digit + classifier. Among higher numerals, only the following are attested phonetically: mwomwo ‘100’, i-pwo ‘500’, ya-pwo ‘800’, ti ‘1,000’, yorodu ‘10,000’. The system of classifiers is in its infancy, and only the following classifiers are attested: tu (objects used with digits), -ti (objects used with tens and hundreds), -ri (people), -moto (grassy plants), -pye (layers and folds), -ka (days). Verbal morphology: One of the greatest differences between Western Old Japanese and all later stages of the language is that quality verbs like modern Japanese siro-i ‘white’ or Middle Japanese siro-ki ‘id’ are still in the process of formation. This is because in Western Old Japanese stems of quality verbs can still behave as regular adjectives: e.g. sira namyi ‘white waves’, opo kapa ‘big river’, cf. Middle Japanese siroki nami ‘white waves’ and opo-ki kafa ‘big river’ with the obligatory attributive suffix -ki, which is not obligatory in Western Old Japanese. Therefore, Western Old Japanese adjectives are in transition from nominals to verbs. Western Old Japanese verbs are divided into several classes: consonantal verbs, with roots ending in a consonant (yuk- ‘go’, kyir- ‘cut’, omop- ‘love, think’, nokos- ‘leave’, etc.), vowel verbs, with roots ending in a vowel (kwopiy- ‘love’, tasukey- ‘save’, myi- ‘see’, miy- ‘turn’, etc.), and irregular verbs (ko- ‘come’, se‘do’, ar- ‘exist’, sin- ‘die’, in- ‘go [away]’). In addition, there are defective verbs—n-, to ‘be’, to ‘say’— that have only a limited number of forms and are irregular as well. Western Old Japanese verbs can take both prefixes and suffixes. The prefixal position is limited to one

790

slot, whereas suffixal positions are multiple. There are a sizeable number of verbal prefixes, most of which occur only in Old Japanese—i-, ka-, ari-, ta-, etc.— with mostly unclear or poorly investigated functions. Verbal suffixes can be divided into word-final and word-nonfinal, the main difference between these two groups being that the second group cannot conclude the verbal form and is always followed by another suffix. The rules governing the combination of verbal roots and suffixes are quite complex. However, a general rule of thumb is that suffixes with initial vowel keep the vowel after consonantal roots, and lose it after vowel roots, while consonant-initial suffixes keep initial consonant after vowel roots, but lose it after consonantal roots. Since some vowel-initial suffixes do not lose their initial vowels after vowel roots, but rather cause the last vowel of the verbal root to be dropped, it is convenient to subdivide vowel verbs into strong vowel verbs (that never lose their final vowels) and weak vowel verbs (that lose their final vowels before certain suffixes). Due to these complexities, it is necessary to list verbal affixes in two forms: after consonantal stems and after vowel stems (with the possible differentiation between forms found after weak and strong vowel verbs). Thus, for example the attributive suffix has the following forms: -u ( [e:] and [ou], [Ey] > [o:], e.g. Old Danish (ODa) enn ‘one’, dothaer ‘dead’ vs. Old Icelandic (OIc) einn, daudhr. Mutation by [i] and [u] was carried out more completely in WN than EN, and palatal mutation was also caused by -R and by [g]/[k] before high front vowels in WN, e.g. OIc nofnum ‘names (dat.)’, thaer ‘they (fem.)’ vs. Old Swedish (OSw) nafnum, thar. By contrast, breaking occurred more widely in EN. During the late tenth and early eleventh centuries, the Scandinavian countries were converted to Christianity, and Norse began to be written in a modified form of the Roman alphabet (borrowing the rune thorn and the letter eth from Old English usage in some regions to represent the dental fricatives, and adding diacritics to some vowels). The oldest surviving manuscripts in Norse are, however, from the late twelfth century. From this era until around 1350 (when the Middle periods of the Scandinavian languages might be said to begin), dialectal distinctions within EN and WN became increasingly marked, and one can refer to OIc, ONo, etc. EN, and in particular ODa, was generally more innovative than WN. ODa merged the unstressed vowels as schwa (written or ), e.g. ODa brothaer ‘brother’, dottaerson ‘daughter’s son’ vs. OIc brodhir, dotturson. Intervocalic consonants underwent lenition, with voiceless stops being voiced, voiced stops sometimes becoming fricatives, and voiced fricatives becoming semivowels or being lost, e.g. ODa nogaer ‘some, a certain’, sauthe ‘said’ vs. OIc nokkurr, sagdhi. The case system was substantially reduced, and there was leveling among the personal endings of verbs. The pitch accent distinguishing original monosyllables from bisyllables was changed into a reduced glottal stop, termed the stod, in stressed syllables with long vowels or final voiced consonants; this accent became lexically distinctive where old monosyllables became bisyllabic through the insertion of epenthetic vowels to break up consonant clusters, or through suffixation of the definite article. OSw underwent many of the same developments as ODa, but to a lesser degree, and later. A phenomenon of vowel balance affected unstressed vowels: after long syllables, they were pronounced more weakly, being written as in some manuscripts, and eventually merged and underwent apocope; after short syllables, however, they retained their qualities longer, being written as , and were more resistant to reduction and loss, e.g. mother ‘mother’ vs. fathir ‘father’, gangae ‘to go, walk’ vs. fara ‘to go, travel’.

OLD TIBETAN ONo, while generally similar to OIc, shared some developments with EN, e.g. initial clusters of [hl] and [hr] were simplified to [l] and [r]: ONo lut ‘share (acc.)’, ratt ‘pushed’ vs. OIc hlut, hratt. Some parts of ONo exhibited vowel harmony of unstressed vowels: following nonhigh vowels, [i u] were lowered to [e o], e.g. honom ‘him (dat.)’ vs. sinum ‘their (dat.)’, kononge ‘king (dat.)’ vs. bili ‘moment (dat.)’. OIc (i.e. ClON) was generally a very conservative variety; however, it innovated in some ways. Around the start of the literary OIc period, the language had a sevenvowel basic ‘triangle’: three front vowels [i e E] and four back vowels [u o o a], with two front rounded vowels [y o] that resulted from the CN mutations; all could be either long or short, and nasalized long vowels were also contrastive. There were three falling diphthongs: [Ei ou Ey]. This system is described in the remarkable twelfthcentury native work known as the First Grammatical Treatise. During the literary period, the low–mid vowels underwent mergers: [E] merged with [e], [E:] with [o:], [o] with [o] (later written ), and [o:] with [a:], e.g. OIc segidh ‘say!’, fell ‘fell’ vs. ONo saeghit, fell. Nasalization was also lost; [y] merged with [i]; [e:] diphthongized to [jE] and [o:] to [ai]. The consonantal system of ClON underwent limited change from CN. Before [l], [r] and most rounded vowels, [w] was lost; it survived as [v] in some environments, e.g. OIc reka ‘to drive’, ordhinn ‘having become’, vidh ‘against’ vs. OSw vraekae, wordhin, Old Gutnish withr. Velar stops palatalized before front vowels and [j] during the literary period, and consonant clusters ending in sonorants were broken up with epenthetic [u], e.g. Modern Icelandic madhur ‘man’ vs. OIc madhr. If a tonal accent existed in CN, it was lost in Icelandic. Grammatically, the morphology of CN was very well preserved in ClON. There were a number of loanwords from Latin, English, or German associated with the new Church, e.g. kirkja ‘church’, paskar ‘Easter’, biskup ‘bishop’. Word order was fairly free, although with a tendency to subject–verb–object, in prose at least. A notable feature of narrative style was the frequent use of the historical present, with tense often switching casually within sentences, e.g. Sidhan logdhu their saman flotann; teksk thar in grimmasta

orrosta ok fell mart af hvarumtveggjum… ‘Then they brought the fleets together; the fiercest battle begins there and many fell on both (sides)…’. ON is worthy of study not only for its intrinsic linguistic interest, but also for the access it gives to a wealth of literature, reflecting a time when pagan and Christian worlds met, when explorers and conquerors expanded out of their original homelands, and when the modern nations of Scandinavia were born. References Antonsen, Elmer. 1975. A concise grammar of the Older Runic inscriptions. Tubingen: Max Niemeyer. Benediktsson, Hreinn. 1982. Nordic umlaut and breaking: thirty years of research. Nordic Journal of Linguistics 5. Faarlund, Jan. 1994. Old and Middle Scandinavian. The Germanic languages, ed. by Ekkehard Konig and Johan van der Auwera. London, New York: Routledge. Gordon, Eric. 1927. An introduction to Old Norse, revised by Arnold Taylor. Oxford, New York: Oxford University Press; 2nd revised edition, 1957. Haugen, Einar. 1950. The first grammatical treatise: the earliest Germanic phonology. An edition, translation and commentary. Baltimore: Linguistic Society of America; 2nd revised edition, London: Longman, 1972. ———. 1976. The Scandinavian languages: an introduction to their history. London: Faber and Faber. Kristjansson, Jonas. 1988. Eddas and sagas. Reykjavik: Hidh islenska bokmenntafelag; 3rd edition, 1997. Nielsen, Hans. 1979. De germanske sprog: baggrund og gruppering. Odense: Odense universitetsforlag; as The Germanic languages: origins and early dialectal interrelations. Tuscaloosa, London: University of Alabama Press, 1989. Noreen, Adolf. 1884. Altislandische und altnorwegische Grammatik (Laut- und Flexionslehre) unter berucksichtigung des urnordischen (Old Icelandic and Old Norwegian grammar (phonology and morphology) in view of Old Norse). Halle (Saale): Max Niemeyer; 4th revised edition, 1923. ———. 1904. Altschwedische grammatik, mit einschluss des altgutnischen (Old Swedish grammar, with the inclusion of Old Gutnish). Halle: Max Niemeyer. Riad, Tomas. 1998. The origin of Scandinavian tone accents, Diachronica 15. Skautrup, Peter. 1944. Det danske sprogs historie (History of the Danish language), Vol. 1. Copenhagen: Gyldendal.

MARISA LOHR See also Gothic; Old High German; Old English; Swedish and Scandinavian Languages

Old Tibetan Old Tibetan is the ancestor of all modern Tibetan dialects. This language is attested from the seventh century onward. It was originally spoken in a relatively tiny

area (if we compare with historical Tibet’s huge size). The origin of the Tibetan people was most probably the Lhoka in southern Tibet, where remain the Yum-bung 793

OLD TIBETAN bla-sgang, the oldest fortress in Tibet, and the royal tombs in ‘Phyong-rgyas. From Emperor Srong-btsan sGam-po at the beginning of the seventh century up to the middle of the ninth century, the Tibetan Empire became one of the major powers in Central Asia. After this brief period of expansion, the Tibetan-speaking area has not changed much up to now. Old Tibetan should not be confused with Classical Tibetan, which is still the standard written language for most Tibetans. Classical Tibetan is an old dialect that is not the direct ancestor of all modern dialects, but it is the basis for the orthography of Modern Literary Tibetan, also called ‘Written Tibetan’. We consider only texts older than the tenth century as genuine ‘Old Tibetan’. The Buddhist conversion of Tibet entailed the coinage of new terms, and the creation of a dictionary, the Mahatvutpatti (bye brag rtog ched) at the beginning of the eighth century. Besides, the orthography underwent several reforms by the end of the Imperial Period. Few documents exist in genuine Old Tibetan. There are the famous Sino-Tibetan bilingual treaties inscribed on pillars (rdo ring), the extensive Dunhuang documents that have been partially translated into French, English, and Chinese, and only two transmitted texts that preserve the most archaic features: the bKa-chems kakhol-ma (testament of Srong-btsan sGam-po) and the sBa-bzhed (chronicle of bSam-yas). Many texts from the Buddhist canonical scriptures (bKa-‘gyur and bsTan-‘gyur) date from the Imperial Period, but they were rewritten so many times that most archaic features have been weeded out. No grammar of Old Tibetan exists yet and the extent of the difference between the Classical and the Old language is not fully understood.

Language Classification Tibetan is usually classified as a Bodic language. The Tamang-Gurung-Thakali languages of Nepal are generally considered its closest relatives. However, no reconstruction of the ancestor language Proto-Tibetan based on the comparison with these languages has been undertaken so far; hence, our knowledge of the prehistory of Tibetan is mostly based on internal reconstruction and on comparison with more remote cousins like Archaic Chinese and Burmese. Some scholars have proposed that Tibetan is especially close to Chinese within the Tibeto-Burman family, but further investigations are needed.

Vocabulary Tibetan has few loanwords from either Sanskrit or Chinese. Indic words are limited to animals and plants not found in Tibet (seng-ge ‘lion’, pad-ma ‘lotus’, ping kyur ma ‘a kind of bird’). Loanwords from

794

Middle Chinese are restricted to cultural vocabulary (lcog tse ‘table’, dong tse ‘coin’). The loanwords from Archaic Chinese have never been studied yet, as it is difficult to tell apart loanwords from words with the same ancestry (cognates) (one such example is lcags ‘iron’ : it must be a loanword, since iron appeared only in the fifth century BC in China). Besides, Tibetan absorbed many non-Tibetan languages during the expansion of the empire, and these have certainly influenced its vocabulary. This might be one cause for the existence of redundantly duplicated words with similar meaning, such as rmi and rmang ‘dream’. Tibetan has several striking lexical innovations, such as bdun ‘seven’ (Burmese khu hnac), khyed ‘you’ (Burmese nang), khrag ‘blood’ (Burmese sweh), not found in any other language of the family. Clear traces of the derivational morphology inherited from Proto-Sino-Tibetan are found in Tibetan. However, it probably was not in active use anymore even by the time of the first Buddhist translations, since new means were created to coin new terms: instead of using the -s suffix deriving an action noun from a verb (za ‘to eat’ zas ‘food’), translators chose to add the verb byed ‘to do’ in its future form bya after a verb to translate the Sanskrit past participle.

Writing System and Phonology The phonology of Old Tibetan is obscured by the irregularities in spellings found in most Dunhuang documents, which often make the translation quite difficult if not impossible. The exact pronunciation of Tibetan in the imperial period is mostly uncontroversial, except for a few letters. The prefixed ‘-, incorrectly called ‘a chung’ even by Tibetan specialists (‘a chung designates in fact the little ‘a used to indicate vowel length in Sanskrit words), has been the object of some debate, as it seems to have three distinct functions in Tibetan orthography: when used as an initial, it represents a voiced glottal fricative (as in ‘od ‘light’). This sound is still heard in some dialects. As a preinitial, it indicates prenasalization (as in ‘khor ‘to turn’, most likely pronounced as *ŋkhor). Finally, since the vowel a is not written by any special symbol in the indic alphabet, this letter also serves to disambiguate some combinations of letters (mda’ ‘arrow’, written M D ‘ as opposed to mad ‘under’ written M D). The final stops -b -d -g are written with the symbols for voiced consonants, although these were most probably pronounced as unreleased stops as in most languages in the area. Old Tibetan boasts of a great quantity of complex consonant clusters that are preserved in modern orthography but only partially in dialects. These clusters are extremely interesting typologically, since

OLD TIBETAN some of them are highly unusual: clusters such as rtare permitted, but not the cross-linguistically much more common tr-. Tibetan admits up to four initial consonants (brgyad ‘eight’), but no more than two finals (dmangs ‘people’). One of the most conspicuous differences between Old and Classical Tibetan is the da drag: while in Classical only -s can stand as a second final, in Old Tibetan some clusters exist with -d as a second final, such as in mnard ‘he oppressed’ (Classical mnar). In the Tibetan writing system, there are three series of initial consonants : voiced, aspirated, and plain unvoiced. The former can appear in all contexts, but the latter two seem to be variants of a single type in Classical Tibetan: except for some rare exceptions, only aspirated consonants can begin a word or stand after nasal preinitials (m- and ‘-), while when preceded by any other preinitial (stop, liquid or s-), only plain voiceless stops are permitted. This reminds one of English stops (p is aspirated in pit and nonaspirated in spit). In Old Tibetan, the distribution of these graphemes is much less predictable (gchig systematically for Classical gcig ‘one’, for instance), but it is not probable that these two series are truly distinctive in Old Tibetan. Finally, there is a vowel in Old Tibetan never found in the Classical language: the inverted I, transliterated as ï. This sign does not seem to represent a phoneme distinct from i, though. The rules governing their distribution are still unknown.

Verbal System Like modern Tibetan dialects, Old Tibetan was an ergative language, which means that the object of a transitive verb bears the same mark as the subject of an intransitive verb (in Old Tibetan, this case has no overt marking), while the subject of a transitive verb is marked with ergative case (variously marked as ‘is, kyis, gyis, or gis). klu-gong glo-ba nye’o (Zhol inscription, South:20) Klu-gong (no suffix)—intention––near—finite verb ending Klu-gong was loyal. gzhan ‘is bdag bsad sha zos dka’ myi dka’ (Pelliot Tibétain 126: 20) other—ergative suffix—yourself (no suffix)—kill––flesh (no suffix)—eat––hard—not—hard If others killed you and ate (your) flesh, wouldn’t it be hard? The ergative suffix, however, serves as an intrumental case and can mark emphasis. Unlike Kiranti or rGyalrongic languages, Tibetan has no trace of agreement between verb and e.g. the

subject. Transitive verbs have four forms (present, past, future, and imperative), and intransitive verbs have two forms. In most cases, one form is not sufficient to derive the four forms, so that the verbal conjugation is far from simple. Here are examples of verbal patterns found in Old Tibetan (for some forms unattested in Old Tibetan, we provide Classical Tibetan in italics): Present

Past

Future

Imperative

byed ‘dzin / zin ‘dogs za sbyor

byas bzung btags zos sbyard / spyard

bya gzung gdags bza sbyar

byos zung thogs zos sbyor

to take to tie to eat to join

Some elements of the verbal conjugation are quite old: the past tense -s suffix is also found in rGyalrong, Qiang, and Kachin, and the change of a to o in the imperative is related to the imperative -o suffix found in Tamang. However, many peculiarities of the verbal system (the change a to o in the present tense of some verbs, the past tense b- prefix) do not seem to be relatable to anything in other Tibeto-Burman languages. An interesting irregularity in conjugation that is common to Tibetan and Kiranti languages is the conjugation of the verb ‘to eat’. In Tibetan, it presents an exceptional vowel alternation in the past tense a à o unrelated to the more common alternations found in the present and imperative. In Hayu, ‘to eat’ is one of the four verbs with a -o stem alternation: dza-ng (I eat it), dzo (You eat it). The same irregularity is found in a dialect of Limbu. This must be an archaism going back far in the prehistory of the language. References Bacot, J.F.W. Thomas, and C. Toussaint. 1940–1946. Documents de Touen-houang relatifs à l’histoire du Tibet. Paris. Huang, Bufan, and De Ma. 2000. Translation and annotations on Dunhuang Tibetan manuscripts about the history of Tufan period. Beijing. Imaeda, Y. and T. Takeuchi. 1990. Choix de documents Tibétains, tome III: corpus syllabique. Paris. Imaeda, Y. et al. 2001. Choix de documents Tibétains, Tome IV: corpus syllabique. Tokyo. Li Fang Kuei, and W. South Coblin. 1987. A study in the Old Tibetan inscriptions. Taipei. McDonald, A. 1971. Une lecture des Pelliot Tibétain 128/6, 1287, 1038, 1047 et 1290. Etudes tibétains dédiées à la mémoire de Marcelle Lalou, 190–391. Paris. Richardson, H.E. 1985. A corpus of early Tibetan inscriptions. London.

JACQUES GUILLAUME See also Sino-Tibetan Languages

795

ONOMATOPOEIA

Onomatopoeia Onomatopoeia is better understood as a semantic phenomenon if language is likened as a diagram. A referent is codified through the phonetic-symbolic means of language into a word, i.e. the linguistic depiction of its representation. All languages show some kind of imitative linguistic codification known as onomatopoeia. Onomatopoeia represents a referent based on a diagrammatic sound–meaning identification. This direct sound–meaning relationship is nonarbitrary or isomorphic. Examples are easily found in many languages, especially the use of animal sounds like the Malay kokok ‘the crow of a rooster’, embek ‘the bleat of a goat’, the Mandarin gaga ‘the quack of a duck’, and miao miao ‘the meow of a cat’. Onomatopoeia as a universal language feature can be divided into three types: sound imitation, secondary sound symbolism, and echo. Onomatopoeia of the first type occurs when linguistic sound imitates sounds of the real world. Each linguistic imitation is taken as the diagrammatic sign of the signified referent. Many terms are derived from sound imitation, for example, cuckoo, cockatoo, boom, buzz, rumble, twang, chirp, slam, thump, clap, flap, and rattle. More common onomatopoeic words like ha ha ha that signifies laughter in English and Cantonese opera is used as a sarcastic tag as in ‘How funny, ha ha ha’ in specific contexts of communication. Onomatopoeic words are salient in Malay vocabulary as illustrated in Table 1. Sound symbolism occurs when there is a recurrence of sound–meaning correlation. An example would be the high front unrounded vowel taken to symbolize a sense of small, insignificant, slight, or weak. The vowel contrast in pinch vs. punch, gripe vs. grope, and sip vs. suck points to this sound–meaning symbolism in /i/. Research shows that the distribution of sounds in men’s and women’s names is not random. It is common for Caucasian women’s names to end with /i/ and men’s with consonants like fricative, i.e. /s, f/, and stop, i.e. /t, k/. The psychology experiment concludes that the gendered sound difference in personal names is best understood as reflecting emotional differences. Men’s names are associated with active, cheerful, nasty, and unpleasant emotions while women’s names are associated with passive, sad, soft, and pleasant emotions (see Whissell 2001). If language is diagrammatic, personal names are depictions of human beings whose respective names become the aspirations intended for them. Secondary sound symbolism in the onomatopoeia context is a linguistic sound taken to connote another

796

meaning based on phonetic similarity. Mandarin and Cantonese speakers believe that the sound of 8, i.e. ba, is sound symbolic to fa ‘prosper’. As such, car registration numbers that contain many 8’s are more expensive in Malaysia and Singapore. Since 4 sì is sound symbolically related to sì ‘die’ in these languages, karaokes in Singapore and Malaysia will add lucky phrases to tags with the number 4 before offering this queue number to customers. One can imagine how confident a typical Chinese needs to be in order to drive in a car with 4444 as the registration number. Secondary sound symbolism plays an important role in Chinese speech communities when one chooses a name for a child. Lóng ‘dragon’, a symbol of divinity and good fortune, is a favourite character used for boys. An excellent example would be Chen Long (Jackie Chan), who became the dragon of Chinese cinematography. Bruce Lee’s middle name is also the character of a dragon. Other favorite characters used as Chinese personal names are fa ‘prosperous’ as in Chow Yuen Fatt, an ‘safe’ as in Ang Lee, and kuang ‘bright’ the first name of the former Singapore Prime Minister. Onomatopoeia in the form of secondary sound symbolism is culturally significant as it influences one’s belief. TABLE 1 Onomatopoeic Lexicon

Meaning

Debuk Debum Debung Debur

The sound of slapping or boxing The sound of heavy things falling The sound of drumming The sound of waves breaking on the shore The sound of flapping of wings The sound of squeezing out of a small hole The sound of boxing each other The sound of heart pounding The sound of hitting with one’s palm The sound of blowing with a stick The sound of gulping water The sound of hearty laughter The sound of a little cough The wailing of a dog The sound of a plane Snoring The sound of hitting the iron The sound of falling coins The sound of thunder The ringing of the phone

Debus Debut Debak Debar Debik Degar Deguk Dekah Deham Dengkung Dengung Dengkur Dentang Denting Dentung Dering

OTO-MANGUEAN LANGUAGES TABLE 2 Semantic Criterion

Examples

Motion Tumult

huff-puff, flip-flap, flipperty-flopperty rumble-jumble, hubble-shubble, wringle-wrangle shally-wally, phoney-baloney, whimsy-whamsy hooky-crooky, hanky-panky, cuddlemuddle randy-dandy, shilly-shally, hazy-dazy equal-aqual, hob-nob, tiggle-taggle tip-top, wee-jee, itsy-bitsy mousey-pousey, tootsy-wootsy, loveydovey

Inferiority Trickery Mental instability Mutuality High degree Hypocoristic

Echo is a form of reduplication that occurs in the form of freezes, binomials, echo formation, and other forms of bipartites. Echoic morphology is language universal. Table 2 illustrates some English echos. Malay echoic reduplications like gunung-ganang ‘ranges of moutain’, batu-batan ‘rocks’, sayur-mayur ‘all kinds of vegetables’, bengkang-bengkok ‘very crooked’, and cucu-cicit ‘descendants’ are commonly used in daily conversation. Malay partial reduplication is important in word fomation, especially in naming things: tikus ‘mouse’ > tetikus ‘computer mouse’, kunci ‘key’ > papan kekunci ‘keyboard’, and pejal ‘solid (adjective)’ > pepejal ‘solid (noun)’. South Asian echos, on the other hand, are loaded with pragmatic functions. Bhojpuri echos like deslai-oslai ‘matchbox etc.’ is used as an indirect request for cigarettes by youngsters or subordinates because smoking in the presence of parents or elderly persons is bad manners. In Tamil, vendu ‘come’ is echoed as vendu gindu ‘come etc.’. The echo is used for the addressee to come (and run an errand for him). Cantonese echos like yùhn lu¯k lu¯k ‘rounded’, ch¯ lahp lahp ‘sticky/goey’, dung b¯ ng b¯ ng ‘freezing cold’, and baahk syu¯t syu¯t ‘snow-white’ come in tripartites. These echos illustrate more intense

meaning, i.e. quantity iconicity where more forms entail more meaning. Despite the skepticism about the relevance of onomatopoeia, it remains resilient in language. The fact that onomatopoeia exists through time and manifests creatively in various forms testifies to the significance of this component either as pragmatic tool, cultural routine, or semantic paradigm that prevails in language use. References Abbi, Anvita. 1992. Reduplication in South Asian languages: an areal, typological and historical study. New Delhi: Allied Publishers. Bolinger, Dwight. 1992. Sound symbolism. International Encyclopedia of Linguistics 4. Haiman, John. 1985. Natural syntax: iconicity and erosion. Cambridge: Cambridge University Press. Hinton, Leanne, Johanna Nichols, and John Ohala (ed.) 1994. Sound symbolism. Cambridge: Cambridge University Press. Jakobson, Roman, and Linda R.Waugh. 1979, 1987. The sound shape of language. Berlin: Mouton De Gruyter. Jespersen, Otto. 1922. Symbolic value of the vowel i. Philologica 1. Sew, Jyh Wee. 1995. Sound meanings in Malay. M.A. Thesis. Singapore: The National University of Singapore. ––––––. 1996. Symbolisation in Malay: evidence in genres and lexicon. Proceedings of the 4th International Pan-Asiatic Symposium of Language and Linguistics, Vol. I. Salaya: Mahidol University Press. Thun, Nils. 1963. Reduplicative words in English: a study of formations of the types tick-tick, hurly-burly and shillyshally. Sweden: Carl Bloms. Tiwary, K.M. 1968. The echo-word construction in Bhojpuri. Anthropological Linguistics 10. Waugh, Linda R. 1993. Against arbitrariness: imitation and motivation revived, with consequences for textual meaning. Diacritics 23. Whissell, Cynthia. 2001. Sound and emotion in given names, Names 49.

JYH WEE SEW See also Iconicity

Oto-Manguean Languages Oto-Manguean languages are distributed throughout Mesoamerica and form the language stock with the deepest time depth in the Americas. The emergence of its eight linguistic families patterns geographically with the ecological zones of the domestication of maize, chili, squash, and beans, which emerged as primary

agricultural products among the peoples of Archaic Mexico beginning about 4400 BCE. In this account, early Oto-Manguean diversified on a landscape where smaller, more mobile bands came to rely increasingly on domesticated plants and center their production in villages on the mesas and highland valleys of the

797

OTO-MANGUEAN LANGUAGES Central Altiplano, the Mixteca Alta, the Valley of Oaxaca, and the Tehuacán Valley. These regions help define four branches of Oto-Manguean, with Otopamean-Chinantecan and Tlapanecan-Manguean forming the Western division and Amuzgo-Mixtecan and Popolocan-Zapotecan the Eastern division. Most of the language families of West OtoManguean are centered in the highlands and watersheds from the Central Altiplano to the Pacific in Guerrero, and East Oto-Manguean families in highlands and watersheds from the Mixteca Alta to the Pacific in Oaxaca. Some consonant differences between the separate reconstructions of the ancestor languages for East and West suggest that the oldest accessible form of Oto-Manguean may already show dialect differentiation, possibly resulting from dispersal among these watersheds. The connections between the Chiapanecan and Manguean languages in the far east of Mesoamerica and Tlapaneco of West OtoManguean are accounted for as the result of population migrations in prehistory. West Oto-Manguean includes the two branches of Oto-Pamean-Chinantecan and Tlapanecan-Manguean. Northern Oto-Pamean consists of Chichimeco of Guanajuato and the Pame languages of San Luis Potosí. Southern Oto-Pamean consists of Matlatzinca and Ocuilteco around the Valley of Toluca, and an Otomian group comprising the Mazahua language and six branches of Otomí in the states of Mexico, Tlaxcala, and Michoacan. The Chinanteco language family represents the extension of the stock east of the Sierra Madre and contains at least six languages. The Tlapanecan-Manguean branch consists of TlapanecanSubtiaban and Chiapanecan-Manguean. Tlapaneco is spoken in several varieties in the state of Guerrero, and Subtiaba was historically spoken near the Pacific coast of Nicaragua. Chiapanecan-Manguean consisted of Chiapaneco in Chiapas and Mangue in coastal Honduras and Nicaragua. East Oto-Manguean includes the PopolocanZapotecan and Amuzgo-Mixtecan branches. Popolocan comprises four Mazateco languages of Oaxaca, Puebla, and Veracruz; four Chocho-Popolocan languages of Oaxaca and Puebla; and Ixcateco of Oaxaca. Zapotecan includes three Chatino languages in southwestern Oaxaca and five branches of Zapoteco in the mountains of southern, western, southwestern, and northern Oaxaca, and in the Valley of Oaxaca east to the Isthmus of Tehuantepec. Each branch of Zapoteco further divides into several distinct languages. Amuzgo-Mixtecan is made up of Amuzgo in Guerrero, two Trique languages in Oaxaca, and MixtecoCuicateco, which itself includes the Cuicateco language of Oaxaca and several divergent Mixtecan language areas in Guerrero, Puebla, and Oaxaca.

798

Over 100 years of Oto-Manguean scholarship provides us with evidence for several statements about the grammatical structure of Proto-Oto-Manguean, the assumed ancestor language. Simple words typically had only one syllable, and complex words could be formed by compounding simple word roots or by attaching prefixes. The language most likely had tone distinctions; i.e. melodic pitch differences between otherwise identical words could alone indicate a meaning difference. The vocalic system included minimally *i, *e, *a, *u, several diphthongs, the semivowels *w and *y, and the consonants *t, *ts, *k, *kw, *s, *x, *xw, *l, *n, *m (the asterisks indicate that these sounds have been reconstructed on the basis of documented languages). While parts of the Oto-Manguean family as presented here were separately grouped by researchers of the nineteenth and early twentieth centuries, it was Edward Sapir’s 1929 classification that was the first to include the earlier groupings of Mixtecan-Zapotecan, Otomian, Chiapanecan-Manguean, and Chinantecan in a single group. From the 1940s until the mid-1960s, Oto-Manguean research had a blossoming period during which several dedicated scholars worked out the relations of diversity within the Oto-Manguean branches. A student of Sapir, Morris Swadesh, compiled the first comparative study of the sound correspondences of the Zapotecan languages. In the 1950s, Sarah Gudschinsky worked on the internal diversity of Popolocan and supported a larger grouping that included the East Oto-Manguean families of Mixtecan and Popolocan, which she called ‘Proto-Popotecan’. Eric Hamp refined and revised both Gudschinsky’s earlier subgrouping of Proto-Popolocan and the analysis of the sound system of Chiapanecan-Manguean. María Teresa Fernández de Miranda also worked intensively with the Popolocan language family and the Manguean family and was working out a reconstruction of Proto-Zapotecan before her untimely death. Robert Longacre’s extensive work on the internal relationships of Amuzgo-Mixtecan also began with publications in the 1950s. In the same decade, our knowledge of West Oto-Manguean developed with Stanley Newman and Robert Weitlaner’s Proto-Otomí reconstructions. In the 1960s, Doris Bartholomew contributed work on Oto-Pamean and revised the Proto-Otomí reconstructions. Along with Michael Piper, she prepared Fernández de Miranda’s unfinished Proto-Zapoteco reconstructions for publication. A virtual dearth of publications spans over a decade before the publication of the most influential work on Proto-Oto-Manguean. Calvin Rensch began systematic work on Proto-Oto-Manguean in the early 1960s. A brief period of new interest developed after the 1976 publication of Rensch’s 1966 thesis, Comparative Oto-

OTO-MANGUEAN LANGUAGES Manguean phonology, in which he focuses on reconstructing the tone-carrying syllables of words in ProtoMixtecan. Based largely on Rensch (1976), J. Kathryn Josserand, Marcus Winter, and Nicholas Hopkins’s (1984) Essays in Oto-Manguean culture history treated the unique opportunity that Mesoamerica provides for integrating historical linguistic and archeological approaches to culture history. In the 1980s, a student of Eric Hamp, Terrence Kaufman, critically assessed the word and sound correspondences contained in the work of Rensch, Longacre, and other historical sources, and in 1994 presented the classification adopted here. The work of the handful of scholars mentioned above only begins to point to the interest of this language stock for linguistic and historical linguistic theory. The bulk of the above works focus mostly on the sound system and the vocabulary. However, the languages of Oto-Manguean also present challenges for research in intonation, sentence grammar, word structure, and meaning, for which the literature is almost entirely lacking. The Oto-Manguean sound system exhibits some unique characteristics. For example, many of the languages lack [p] and [b], and most consonants are voiceless, i.e. they are produced, like [t, k, s], without vocal cord vibration. Voiced consonants such as [d, g, z], on the other hand, are rare. Oto-Manguean is the world’s oldest language stock that provides sufficient evidence for tone distinctions in the ancestor language. The complicated tone systems distinguish between two and six pitch levels; i.e. in the most complex scenario, one could theoretically have six different one-syllable words differing only in the pitch or the pitch pattern of the vowel. ‘Whistled speech’ has been reported in Mazateco, Mixteco, and Zapoteco. This is a form of speech that is basically stripped of individual segments and consists almost exclusively of melody. It provides speakers with the ability to communicate a wide range of utterances across a distance. The grammar of Oto-Manguean languages makes heavy use of particles. In Oto-Manguean languages, they can carry meanings associated with intonational patterns in languages like English or the Romance languages. For example, yes/no questions, topicalization, exclamation, and emphasis are indicated by particles. However, at this time, very little work has been done on Oto-Manguean sentence structure. Oto-Manguean continues to have great import in the development of historical linguistic theory. It is a family with an internal complexity comparable to that of Indo-European but has diversified within a restricted cultural-ecological zone associated with highland agriculture. Thus, Oto-Manguean languages provide a natural testing ground for addressing the interaction of

language-internal changes and regionally defined patterns of linguistic diffusion. From Archaic through Post-Classic Mesoamerica, webs of political and economic interaction mediated the divergence of related languages as villages allied themselves to larger centers emerging with the origins of the states of PreColumbian Mexico. Oto-Manguean language history also involves contacts with languages outside the family but within a greater Mesoamerican culture area. Such relations include speakers of Huave, Totonac, Mixe-Zoque, Maya, and later arriving Uto-Aztecan speakers including the Aztecs. Loanwords and structural similarities shared among many of the languages of Mesoamerica support a view that networks of contact have been most important in shaping the patterns of linguistic features in this culture area. The Spanish Conquest brought a common language of the empire that must also be counted in the histories of the divergence of Oto-Manguean languages as well as in shaping our knowledge of those that were chronicled in Colonial and Post-Colonial Mesoamerica but are no longer spoken at the present time. Today the status of Oto-Manguean languages varies from village to village. Some families have well over 100,000 speakers, but the languages of that family may range from villages of several thousand, where all the children learn an Oto-Manguean language as a first language, to villages where less than a dozen elders have any memory of their language. In most communities, bilingualism is the norm, with Spanish being learned either at home or in school. Many villages that once spoke an Oto-Manguean language have entirely shifted to Spanish. This trend is occurring in some regions at an alarming rate. Several key languages and language families are already extinct. Chiapaneco and all of Manguean have long since passed. While the documents on Manguean are sufficient to demonstrate its membership in the Oto-Manguean stock, the data were not recorded with attention to tone and other complex issues of pronunciation. Most of the TlapanecanManguean languages have not been adequately documented at all. For Oto-Manguean overall, there are very few dictionaries and grammars, and few practical orthographies have been developed. The next couple of generations will decide whether Oto-Manguean voices will still be heard in Mexico’s future. References Bartholomew, Doris A. 1960. Some revisions of proto-Otomí consonants. International Journal of American Linguistics 26. 317–29. Campbell, Lyle, Terrence Kaufman, and Thomas C. SmithStark. 1986. Meso-America as a linguistic area. Language 62(3). 530–70.

799

OTO-MANGUEAN LANGUAGES de Miranda, Fernández, and María Teresa. 1995. El ProtoZapoteco, ed. by Michael J. Piper and Doris Bartholomew. Mexico D.F.: El Colegio de Mexico. 1995. Gudschinsky, Sarah C. 1995. Lexico-statistical skewing from dialect borrowing. International Journal of American Linguistics 21. 138–49. Hamp, Eric P. 1958. Protopopoloca Internal Relationships. International Journal of American Linguistics 24. 150–53. Josserand, J. Kathryn, Marcus Winter, and Nicholas Hopkins (eds.) 1984. Essays in Otomanguean culture history, Nashville: Vanderbilt University Publications in Anthropology No. 31. Kaufman, Terrence. 1994. The native languages of Mesoamerica. Atlas of the World’s languages, ed. by Moseley and Asher. London, New York: Routledge.

800

Longacre, Robert E. 1966. On the linguistic affinities of Amuzgo. International Journal of American Linguistics 32. 46–9. Newman, Stanley, and Robert Weitlaner, 1950. Central Otomian I: Proto-Otomi reconstructions. International Journal of American Linguistics 16. 1–19. Rensch, Calvin R. 1976. Comparative Oto-Manguean phonology. Bloomington: Indiana University Press. Swadesh, Morris. 1947. The phonemic structure of ProtoZapotec. International Journal of American Linguistics 13. 220–30. ———. 1960. The Oto-Manguean hypothesis and macro Mixtecan. International Journal of American Linguistics 26. 79–111.

MARK SICOLI See also Mexico; Tone Languages

P Pacific The area covered by the Pacific ocean occupies about a third of the Earth’s surface and is by far the world’s largest geographic region. It contains more than 1,000 indigenous languages, Asian mainland languages, half a dozen European languages, and a number of pidgins and creoles. Geographically, it includes the island nations and territories of Micronesia, Polynesia, and Melanesia, collectively known as Oceania. The islands occupy just less than 500,000 square miles of land area. New Guinea, the second largest island in the world after Greenland, represents 64% of this total, and New Zealand represents another 20%. The remaining 10% is divided among more than 10,000 scattered islands. Micronesia comprises the island groups east of the Philippines and north of the equator, including, for instance, the Mariana, Marshall, Caroline, and Gilbert islands. Melanesia encompasses the island groups in the southwestern Pacific extending southeastward from the Admiralty islands to Fiji, including such nations as Papua New Guinea, Solomon Islands, Vanuatu, and New Caledonia. Polynesia consists of a group of islands extending from New Zealand north to Hawaii and east to Easter Island. Politically, the status of the islands ranges from those that are independent (e.g. Papua New Guinea, Fiji, Kiribati, Palau, Samoa—formerly Western Samoa, Tuvalu, and Tonga), those that are quasicolonies or territories (e.g. New Caledonia, French Polynesia, and Wallis and Futuna—French overseas territories; Guam and American Samoa—United States territories; Norfolk Island—an Australian territory; Rapa Nui/Easter Island—a dependency of

Chile; and Northern Marianas—a Commonwealth of the United States), those that are self-governing in free association with other countries (e.g. Cook Islands, Tokelau, and Niue in relation to New Zealand, and the Federated States of Micronesia, Chuuk, Pohnpei, Yap, and Kosrae, in relation to the United States), and those that are colonies (e.g. Pitcairn Island—a British colony), to those that have been incorporated into other nations (e.g. Hawaii— formerly a territory of the United States, now the fiftieth state). The Pacific islands were peopled by influxes from both mainland and island Southeast Asia, with the earliest migrations to Melanesia. From Melanesia, generations of voyagers headed northward into eastern Micronesia and eastward into Polynesia. European exploration of the Pacific began with the voyages of Ferdinand Magellan in the 1520s, which inaugurated a period of Spanish and Portuguese influence, followed by Dutch, English, French, and German influences. Most of the major island groups were annexed by Britain, the United States, France, and Germany in the nineteenth century. Many Micronesian territories changed hands a number of times, beginning with Spanish colonial administration that was followed by German, Japanese, and American regimes. The historical partition of the islands among the various European colonizers has left the islands’ populations divided primarily into a Francophone and Anglophone set of nations, each with different political, economic, linguistic, and cultural orientations. More recently, however, notions of a Pacific islander identity

801

PACIFIC have begun to emerge, fueled partially by recognition of a common history of colonization, as well as by demands from nonindependent territories and nations for increased political autonomy. The indigenous languages of the Pacific belong to two language families: Austronesian and nonAustronesian (Papuan). The Austronesian language family is the largest and most widespread family in the world. The Oceanic or easternmost subgroup of the Austronesian family is the largest and best defined of all subgroups, containing all Polynesian languages, all Micronesian languages (except Palauan and Chamorro), and all Austronesian languages of Melanesia east of the Mamberamo River in Irian Jaya (the western half of the island of New Guinea, Indonesia’s easternmost province). The Polynesian languages are a subgroup of Oceanic containing approximately 30 languages organized into two major subgroupings: Tongic (including Tongan and Niuean) and Nuclear Polynesian (all the rest). The latter is divided into two further groups: Samoic Outlier (consisting of Samoan and Tuvaluan) and Eastern Polynesian (consisting of Hawaiian, Tahitian, Maori, Rapa Nui, and Marquesan). Speakers number fewer than a million people spread across a large area of the Pacific, ranging from Hawaii in the north to New Zealand in the south and Easter Island. This so-called Polynesian Triangle encloses an area approximately twice the size of the continental United States. Despite their great geographic spread, the Polynesian languages show a relative homogeneity, which indicates that they have dispersed only in the last 2,500 years from an original center in the Tonga-Samoa area. In addition, there are some 18 further Polynesian-speaking communities known as ‘Polynesian Outliers’ in Micronesia (e.g. Kapingamarangi and Nukuoro, two atolls in the southwestern portion of Pohnpei state) and Melanesia (e.g. Ontong Java atoll and Rennell island in the Solomon Islands). Although some Polynesian languages have a relatively large number of speakers, such as Samoan with about 200,000, many others have very few, and some, such as Maori and Hawaiian, face possible extinction as their speakers have shifted to English. Most of the Oceanic Micronesian languages (with the exception of Nauruan, which has not been sufficiently studied) form a closely related subgroup called Nuclear Micronesian. Palauan and Chamorro, however, appear to be the result of distinct migrations from Indonesia or the Philippines. Although Yapese is probably Oceanic, it does not appear closely related to any of the other Micronesian languages, because of its complex history of borrowing. It is also the only

802

Austronesian language containing glottalized consonants, (i.e. pronounced with simultaneous contraction of the vocal folds). Most of the Micronesian languages have relatively small numbers of speakers, with a few, such as Sonsorolese on Sonsoral island in Palau, having only several hundred speakers. Historical relationships among the Papuan languages, found primarily in the New Guinea interior, an area of immense linguistic complexity, are less clear than in the much better studied Austronesian language family, and the label is best seen as a cover term for many, perhaps as many as 60, distinct families. The Papuan-speaking peoples were the earliest people in the region, occupying the Sahul continent (which later partially submerged to become the island of New Guinea) at least 40,000 years ago. By 30,000 years ago, the Bismarck Archipelago east of New Guinea was occupied by speakers of Papuan languages. Speakers of Austronesian languages, who arrived much later, established coastal communities in New Guinea, the southeastern Solomons, Vanuatu, New Caledonia, and Fiji. Contacts between speakers of Papuan and Austronesian languages have made it difficult to classify some languages as Austronesian or Papuan, such as Maisin in Papua New Guinea, which displays features typical of both language families. The European languages spoken in the Pacific region are primarily those of the former colonizers and still have official status in most areas. Although French is important, no European language has had the impact that English has had. In addition to the standard variety of British or American English that serves as the official language in Anglophone territories, English exists in a number of locally distinct varieties in New Zealand and Hawaii, and it also exists in the form of numerous pidgins and creoles. French has left its legacy in the form of a number of local varieties of colonial French spoken, for example, in New Caledonia and Tahiti, as well as in a French creole, Tayo, in New Caledonia. Standard French still has official status in French territories of the Pacific, as well as in independent Vanuatu, formerly a condominium ruled jointly by Britain and France, where it shares co-official status with English. Because Germany lost most of its colonies after World War I, German never had a chance to spread as widely as other colonial languages. The language left some remnants, nonetheless, in the form of a German creole, Unserdeutsch, as well as borrowings, both into forms of pidgin English, such as Tok Pisin in Papua New Guinea, as well as into indigenous languages. Spanish has only a limited presence in the Pacific as the official language of Rapa Nui/Easter Island.

PACIFIC In addition to European languages, other languages were introduced as a result of colonization, such as Japanese, Korean, varieties of Chinese (for example, in Papua New Guinea, Nauru, Tahiti, and elsewhere), and Vietnamese in New Caledonia. Varieties of Southeast Asian languages, such as Hindi, were introduced into Fiji, and Javanese was introduced into New Caledonia. Philippine languages such as Tagalog, Ilocano, and Visayan were introduced into Hawaii through the importation of migrant labor and indentured plantation workers. The Pacific pidgins and creoles originated primarily in the late eighteenth and early nineteenth centuries. By the latter part of the nineteenth century, trade contacts between English speakers and Pacific islanders had led to the formation of English-based pidgins spoken in various forms and with differing degrees of stability in almost the entire Pacific basin from New Guinea to Pitcairn island, from the Marshall Islands and Hawaii to New Caledonia and New Zealand. Not all of these have survived. In Micronesia and most of Polynesia (with the exception of Hawaii), forms of pidgin English disappeared because their role in internal communication was quite limited. Chinese Pidgin English, the oldest form of Pidgin English in the Pacific, which developed around the port of Canton in 1700, has subsequently died out, although a mixed form of Melanesian-Chinese Pidgin English is still used on Nauru in a variety of commercial contexts, including Chinese trade stores and restaurants, and in the phosphate mining industry, which began in 1906. In addition, there are a number of pidgins and creoles based on indigenous languages such as Hiri Motu in Papua New Guinea, one of the three official languages of Papua New Guinea (along with English and Tok Pisin). A number of indigenous pidgins such as Pidgin Maori, Pidgin Fijian, and Pidgin Hawaiian are no longer in use. Pidgin and creole languages based on English are more numerous than those related to any other language, attesting to the greater spread of English than any other metropolitan language. Because the French and Germans were relative latecomers to the Pacific, they found varieties of Pidgin English well established in many areas. Melanesian Pidgin English emerged on plantations in Queensland, Australia, to which workers from all over Melanesia were recruited, and later evolved into a number of distinctly named varieties, including Tok Pisin, spoken in Papua New Guinea; Bislama, spoken in Vanuatu; and Pijin, spoken in the Solomon Islands. In these highly multilingual countries, these languages are the normal everyday medium of communication for millions of people. Tok Pisin is the largest language in the South Pacific today, with as

many as 2 million speakers. Most business in the House of Assembly, the country’s main legislative body, is conducted in Tok Pisin, which is the most widely shared language among the members. In Hawaii, a creole English developed on plantations and is still in use, primarily among working-class speakers. There is some dispute over the status of the variety of English spoken on Pitcairn and Norfolk Islands, spoken by descendants of the British mutineers from HMS Bounty. Only Tok Pisin and Bislama have received some official recognition. Tok Pisin is a de facto official language in Papua New Guinea and is spoken by more than half the population; however, English is the official medium of education. Although Bislama is recognized by the constitution of Vanuatu as the national language of the country, it is forbidden in the schools. English and French, the languages of the former colonial powers, are still the official languages of education. The only example of a German creole is Rabaul Creole German, called Unserdeutsch, ‘our German’, which arose at the turn of the century as a lingua franca of the Catholic mixed-race community in Vunapope near Rabaul in what is now East New Britain Province of Papua New Guinea. This was formerly part of the German colony Kaiserwilhelmsland from 1884 to 1914. This language was apparently used by children at a mission boarding school and creolized in one generation. It is now extinct. The only example of a French creole is Tayo, spoken in the southern part of New Caledonia around St. Louis, the site of a Catholic mission and plantation, where linguistic diversity among the surrounding tribes led to the emergence of a creolized French as a lingua franca.

References Bellwood, Peter. 1991. The Austronesian dispersal and the origin of languages. Scientific American July. 70. Foley, William. 1986. The Papuan languages of New Guinea. Cambridge: Cambridge University Press. Hollyman, K.L. 1971. French in the Pacific. Current trends in linguistics, Vol. 8, Linguistics in Oceania, ed. by Thomas A. Sebeok. The Hague: Mouton. Krupa, Viktor. 1982. The Polynesian languages: a guide. London: Routledge & Kegan Paul. Lynch, John. 1998. Pacific languages. Honolulu: University of Hawaii Press. Romaine, Suzanne. 1991. The Pacific. English around the World: sociolinguistic perspectives, ed. by Jenny Cheshire. Cambridge: Cambridge University Press.

SUZANNE ROMAINE See also Austronesian; New Guinea; Pidgins and Creoles

803

PAMA-NYUNGAN

Pama-Nyungan The term Pama-Nyungan, coined by Ken Hale in the early 1960s after the words for ‘man’ in the northeastern and southwestern extremities of Australia, refers to a group of 160 or so languages spoken over about seven eighths of the continent (shown in gray on the map). This group was initially proposed on the basis of lexicostatistical comparisons of 100 basic words, and was believed to represent a single family of genetically related languages. The remaining eighth of the continent is considerably more diverse, and the languages were assigned to some 28 additional families, referred to collectively as non-Pama-Nyungan. It was considered that almost all Australian languages, Pama-Nyungan and non-Pama-Nyungan, are ultimately genetically related. The lexicostatistical classification correlates well with a typological classification (the essence of which was proposed in 1940 by Arthur Capell). Most PamaNyungan languages are ‘suffixing’ (by definition, have only suffixes, and no prefixes) and dependent marking (grammatical relations signalled on noun phrases), while non-Pama-Nyungan languages are overwhelmingly ‘prefixing’ (by definition, have both prefixes and suffixes) and head marking (grammatical relations signaled on the verb, and sometimes also on noun

1 Kalibamic 2 Kukatyic? 3 Karantyic? 4 Walangama? 5 Waramungic 6 Kalkatungic 7 Yalarnngic 8 Yandic 9 Durubulic 10 Giabalic? 11 Bandjalanic

12 Gumbaynggir 13 Marawaric 14 Yotayotic 15 Yaithmathangic

phrases). There are a few exceptions: the Tangkic languages are suffixing and dependent marking but nonPama-Nyungan; Yanyuwa is Pama-Nyungan but prefixing and head marking. The classification has been the subject of considerable controversy in recent years. There is, however, no widely accepted alternative, and most Australianists accept it in broad outline.

Current State of the Languages Most Australian languages (Pama-Nyungan and nonPama-Nyungan) are endangered or moribund. Many languages of the east and southeast, where conflict with Europeans was particularly intense and violent, are no longer used as vehicles of everyday communication, and are effectively dead; some disappeared before adequate records could be made of them. On the other hand, some of the most viable languages, with the best chance of future survival, are PamaNyungan. These include the Western Desert language, with some 4,000–5,000 speakers; Arrernte and Warlpiri, with some 3,000 speakers each; Yolngu varieties with some 2,000; and Kala Lagaw Ya (Torres Strait Islands) with some 3,000–4,000 speakers.

Mabuiagic Kala Lagaw Ya Yolngic Arnhem Land Gulf of Daly Carpentaria River s ie il Warluwarric fam ngan -Nyu Tangkic ama

-P Non Kimberley

5

Garrwan

1 23

Lamalamic Yalanjic

Mbarbaram

Yidinyic Dyirbalic

4

Warluwarra Mayapic 6 7 8

Walmajarri Warlpiri Pilbara

South West Nyungic

Western Desert varieties

Arandic

Pama-Maric

Arrernte Waka-Kabic Karnic

Nganyaywana 9 10 11 12

13

ic

Wiradjuric

Yu in

-K ur

Darling Kaurna Narrinyeric Kulunic Boundary of Pama-Nyungan and non-Pama-Nyungan Group boundaries Questionable groups

804

14 15 Kurnic

PAMA-NYUNGAN Efforts are being made by speakers and their descendants to reintroduce some dead Pama-Nyungan languages (e.g. Kaurna in the Adelaide region), and to strengthen some weakening languages (e.g. Walmajarri in the northwest). But few such efforts have achieved appreciable results, and considerably more needs to be done before the process of attrition can be considered to have been arrested generally, or for any language.

Status of Pama-Nyungan and Relation to nonPama-Nyungan The notion of a Pama-Nyungan family is widely accepted by Australianists; most believe that a genetic family of roughly the size shown on the map will ultimately prove valid, although we are a long way from providing convincing arguments. It has not been established as a family by rigorous application of the comparative method: little of (ancestral) proto-PamaNyungan phonology, lexicon, or grammar has been reconstructed, and few subgroupings have been argued on the evidence of shared innovations. One reason for this situation is that investigators have assumed the genetic relatedness of Australian languages, and have focused on the search for shared innovations, rather than attempting to reconstruct proto-Pama-Nyungan. With improved descriptions of more languages, and use of the comparative method, the membership and scope of Pama-Nyungan have been refined somewhat since the 1960s, as have details of the subclassification. Thus, Mbabaram (Queensland rain forest region) and Nganyaywana (northeast New South Wales), originally excluded from the Australian superfamily, were shown by Robert Dixon and Terry Crowley, respectively, in the 1970s to be closely related to their neighbors; both had undergone radical sound changes that obscured cognates (related words). Perhaps the major revision was proposed in the late 1980s by Barry Blake (1988) and Nicholas Evans (1988). They reclassified the Tangkic languages, previously considered to be Pama-Nyungan (largely on typological grounds, since these languages show low vocabulary sharing with Pama-Nyungan) as nonPama-Nyungan, and Yanyuwa, previously taken to be a non-Pama-Nyungan isolate, as Pama-Nyungan. They assigned Yanyuwa to a subgroup with Warluwarra, spoken to the south and separated from Yanyuwa by the Garrwan languages (whose classification has also proved problematic). (These are indicated by vertical hatching on the map.) Their evidence mainly concerned forms of pronominals and case-markers. Blake (1988) reconstructed a pronominal system for proto-Pama-Nyungan, and a different one for

proto-non-Pama-Nyungan. Evans (1988) proposed further that Pama-Nyungan could be identified as a subgroup (within Australian) on the basis of specific phonetic and grammatical characteristics, for example, the wide occurrence of the variants -lu  -nggu as a marker for ergative case. Dixon has always been a virulent critic of the notion that Pama-Nyungan languages constitute a genetic family of their own, lambasting the idea on methodological and other grounds. Dixon has made some valid criticisms of lexicostatistical methodology. However, the lexicostatistical classification was always regarded as a tentative means of providing a preliminary genetic classification, one that must be refined and supported by the comparative method. Most Australianists who accept the classification do so tentatively (an imperfect or faulty classification is preferable to none, provided that its provisional status is borne in mind). Dixon also argues against Pama-Nyungan on the grounds that the putative innovations in the pronominal and case-marking systems—e.g. the first-person dual inclusive ngali ‘we two (including you)’ and the ergative variants -lu  -nggu—are not present in all Pama-Nyungan languages. Dixon avers that areal diffusion rather than common ancestry provides the simplest explanation for the distribution of these features. He suggests that the -nggu allomorph was independently innovated (via natural phonological conditioning) in four different places, and subsequently diffused areally. Others consider that the distribution of -nggu is better explained as a retention from a proto-language that was lost in some modern languages or groups. And as Evans (2003) pertinently observes, although instances of diffusion of specific morpheme variants are well attested in Australia, diffusion of variant sets (with complex conditioning factors) such as -lu  -nggu are not. In a similar way, many Australianists consider it to be probable that the absence of the first-person dual inclusive ngali in a few Pama-Nyungan languages represents loss of the form in these languages, rather than incomplete diffusion over the Pama-Nyungan region. Both Barry Alpher (e.g. 1990) and Harold Koch (e.g. 2003) argue the advantages, indeed necessity, of taking inflectional paradigms into account in reconstructing proto-Pama-Nyungan, and not merely the forms of roots and affixes separately. Both show that sets of paradigmatic forms—verb inflection paradigms by Alpher (1990), and case-paradigms for singular free pronominals and interrogative particles by Koch (2003)—can be identified in a diverse selection of Pama-Nyungan languages that permit reconstruction of inflectional paradigms, and thus support the genetic unity of Pama-Nyungan.

805

PAMA-NYUNGAN

Subgrouping The map gives an approximate idea of the groups within Pama-Nyungan. Some 30 groups are identified, which are further divided into almost seventy subgroups. The groups and subgroups differ considerably in distribution and size. One group, the Nyungic group, takes up nearly half of the Pama-Nyungan region, and is divided into about a dozen subgroups. On the other hand, Warumingic, Kalkatungic, and Yandic consist of single languages. The establishment of the groups and subgroups of Pama-Nyungan is in a very preliminary state, and there are doubts about the status of many. Various changes to the group and subgroup structure have been proposed since 1966. Attempts have also been made to motivate groups or subgroups by more reliable means than the original lexicostatistical criteria. In a number of cases, however, the genetic relatedness of a set of languages has been established, although not subgroup status by identification of shared innovations. In a careful study of the Pilbara situation, Alan Dench (2001) has pointed out a variety of difficulties in determining whether shared morphological and syntactic features are innovations indicative of common inheritance or the result of diffusion. He concludes that although the languages are probably genetically related, there is no evidence that they form a subgroup of any larger group; the languages form, in his view, a linguistic area within which much lexical and grammatical diffusion has occurred. Although Dench takes the extreme position that diffusional explanations should be privileged over genetic retention, he correctly observes the difficulties inherent in distinguishing genetically inherited features from areally diffused features. Dixon’s current position (2002) is that the putative Pama-Nyungan family is made up of some 30 lowlevel genetic ‘subgroups’, which are effectively small language families on their own (the prefix ‘sub-’, he explains, does not imply that there is necessarily any higher order containing group, only that there may be). Dixon further proposes that some of the accepted groups within Pama-Nyungan are areal groups, not genetic subgroups. These include the Arandic and adjacent Karnic groups of Central Australia. However, Koch (2004) and Claire Bowern (2001) provide carefully argued cases for their subgroup status on the basis of shared innovations, at least some of which (e.g. paradigmatic irregularities) are more likely to be retentions from an earlier proto-language than resulting from diffusion. Examples of such shared innovations peculiar to the Arandic languages are phonological changes (including complete loss of all initial consonants, centralization of vowels in unstressed syllables, and stress shift to the second syllable of words), and

806

innovations in personal pronouns, nominal inflections, and interrogatives. In Karnic languages, they include lexical and morphological innovations.

Origin and Dispersal of Pama-Nyungan Languages The uneven distribution of language families across Australia, and of groups within the Pama-Nyungan region is striking, and invites explanation. Some proposals have been put forward. These accept that protoPama-Nyungan was originally spoken in a much smaller homeland region, and that the continent as a whole once showed linguistic diversity comparable to that of the non-Pama-Nyungan region today. Around 3,000–5,000 years ago (the dating is a guess, based on the apparent diversity within the family), it began spreading out to ultimately cover the large region it occupies today. Stephen Wurm (1972) suggests a link with the emergence of an advanced technology (ground axes and hafted stone tools) that also appears on the archeological horizon some 5,000 years ago. He proposes that users of this technology, speakers of proto-PamaNyungan, expanded out into areas formerly occupied by speakers of different languages, their advanced technology presumably giving them an edge over the local inhabitants. A more elaborate model is suggested by Nicolas Evans and Rhys Jones (1997). These authors propose (on the basis of the diversity of subgroups of PamaNyungan and the number of potential sister families) a homeland region in the vicinity of the Gulf of Carpenteria. In contrast to Wurm, they do not link the spread of Pama-Nyungan to major migrations and conquests by the wielders of a new technology, but to a cluster of social changes, including the spread of ceremonial activity associated with the new highly valued technology, and the exploitation of new vegetable foods that could support large intergroup ceremonial activities. They suggest that initiation into the rituals associated with the new technology might attract payment in the form of wives for the sons of the initiators, thus giving rise to larger networks of exogamous social interactions than had existed previously. In such circumstances, proto-Pama-Nyungan could well have been a high-status language, and its usage a sign of the status of performers of the new ceremony. Thus, Pama-Nyungan could have spread out with relatively minor (though still discernible) demographic changes. Both models are of course speculative. Nevertheless they are noteworthy in that they attempt to forge links between linguistic, anthropological, and archeological evidence. They also make potentially testable predictions concerning the distribution of bio-genetic

PÀNINI " markers and substrate influences on modern PamaNyungan languages. References Alpher, Barry. 1990. Some proto-Pama-Nyungan paradigms: a verb in the hand is worth two in the phylum. Studies in comparative Pama-Nyungan, ed. by Geoffrey N. O’Grady and Darrell T. Tryon, 155–71. Canberra: Pacific Linguistics. Blake, Barry. 1988. Redefining Pama-Nyungan: towards the prehistory of Australian languages. Aboriginal Linguistics 1. 1–90. Bowern, Claire L. 2001. Karnic classification revisited. Forty years on: Ken Hale and Australian languages, ed. by Jane Simpson, David Nash, Mary Laughren, Peter Austin, and Barry Alpher, 245–61. Canberra: Pacific Linguistics. Bowern, Claire, and Harold Koch (eds.) 2004. Australian languages: Classification and the comparative method. Amsterdam: John Benjamins. Dench, Alan. 2001. Descent and diffusion: the complexity of the Pilbara situation. Areal diffusion and genetic inheritance: problems in comparative linguistics, ed. by Alexandra Y. Aikhenvald and Robert M.W. Dixon, 105–33. Oxford: Oxford University Press. Dixon, Robert M.W. 2002. Australian languages: their nature and development. Cambridge: Cambridge University Press.

Evans, Nicholas. 1988. Arguments for Pama-Nyungan as a genetic subgroup, with particular reference to initial laminalization. Aboriginal Linguistics 1. 91–110. ———. 2003 Introduction: comparative non-Pama-Nyungan and Australian historical linguistics. Comparative studies in non-Pama-Nyungan, ed. by Nicholas Evans, 3–25. Canberra: Pacific Linguistics. Evans, Nicholas and Rhys Jones. 1997. The cradle of the PamaNyungans: archaeological and linguistic speculations. Linguistics and archaeology: aboriginal Australia in global perspective, ed. by Patrick McConvell and Nicholas Evans, 385–417. Melbourne: Oxford University Press. Koch, Harold. The case for Pama-Nyungan: evidence from Remains forthcoming inflectional morphology. Paper presented to 17th International Congress of Linguists, Prague, Czech Republic, July 24–29, 2003, to appear in Proceedings of the congress. ———. The Arandic subgroup of Australian languages. In Bowern and Koch, pp. 127–50. Miceli, Luisa. Pama-Nyungan as a genetic entity. In Bowern and Koch, pp. 61–86. O’Grady, Geoffrey N., Carl F. Voegelin, and Frances M. Voegelin. 1966. Languages of the World: Indo-Pacific fascicle 6. Anthropological Linguistics 8. 1–197. Wurm, Stephen A. 1972. Languages of Australia and Tasmania. The Hague: Mouton.

WILLIAM B. MCGREGOR

PÁnini " Virtually nothing is known about the life of the great Indian grammarian PÁnini, except for the fact that he " was a native of S´alÁtura, in northwestern India (GandhÁra); from his matronymic, we learn that his mother’s name was DÁks¯i. None of the numerous commentators say anything" about the time in which he lived, but, based on the variety of Sanskrit he describes, he is now thought to have lived in the fourth (or maybe fifth) century BCE. His grammatical treatise, the AstÁ dhyÁ y¯i (‘Eight chapters’), is the most "" important grammatical description of Sanskrit, and a large part of the Indian grammatical tradition after PÁnini consists of commentaries on his work. " The style of the AstÁ dhyÁ y¯i appears quite striking if considered from the" "point of view of Western grammatical tradition: it is constituted by 3,996 su¯tra, or aphoristic rules, by which PÁnini not only describes " grammatical rules but also defines the way in which grammatical terms and constructions are used in his metalanguage. PÁnini’s terminology constitutes his " own achievement only in part: grammatical description already relied on a rather conspicuous tradition,

judging from the fact that PÁnini mentions the names " of ten of his predecessors. The language described by PÁnini is a variety of Sanskrit that does not correspond "exactly to the language of any known text. This fact has aroused suspicion, especially among Western scholars in the nineteenth century, leading some to blame PÁnini of " describing an invented variety. A further complication is given by the fact that even the earliest commentators are at variance with PÁnini on some points: most likely, different varieties of "Sanskrit underlie their descriptions. The philological tradition of PÁnini’s text is not without problems either, as some minor" parts probably belong to later commentaries (notably to Patañjali’s MahÁ bhÁ sya see below). " It is noteworthy that PÁnini occasionally mentions " differences between the variety he describes and earlier varieties, notably Vedic Sanskrit. This has led some scholars to state that PÁnini had a historical view of " of the existence of earlier grammar: however, in view grammarians, who had mostly described Vedic, it seems more likely that PÁnini only wanted to mention " 807

PÀNINI " variants found in other treatises, without historical intentions. Among the most important commentators of PÁnini are KÁtyÁyana, who, according to tradition, lived" in the third century BCE and annotated about a third of PÁn ini’s work, and Patañjali, author of the " Á bhÁ sya (‘Great commentary’, around 150 BCE). Mah Patañjali’s" commentary, which takes as a starting point KÁtyÁyana’s notes (which, in fact, we only know through the MahÁ bhÁ sya), and adds comments, partly " to PÁnini’s text, also contains to the notes, and partly " statements about the in the introduction some general Vedas and about grammar. PÁnini’s grammar became known to Western schol" end of the eighteenth century, with the beginars at the ning of Indological studies; since then, its fortune in the West has undergone quite dramatic changes. During the nineteenth century, the striking differences between the frameworks in which grammatical description was organized by PÁnini and by the Greco" increasing convincRoman tradition, together with the tion that insights into language could be gained only by means of historical investigation, progressively led to a negative evaluation of PÁnini’s work. Perhaps the " were held by the worst opinions about PÁn ini " American indologist Withney, who not only thought, as many of his contemporaries, that the Indian grammarian described an invented language but also dismissed his grammatical terminology as crude and philosophically not well founded. In spite of this and other severe judgments, Indian grammatical tradition influenced Indo-European linguistics in its early development, partly providing a basis for the terminology, partly because the Indian use of listing roots helped to develop the concept of morpheme. The twentieth century brought a complete reevaluation of PÁnini’s methods, in some cases with enthusi" astic appraisals. The systematic character of PÁnini’s " ¯tra treatment of grammar, and the formulation of the su as grammatical rule appealed to scholars working within various frameworks; PÁnini was hailed as a "

predecessor of both structuralism and generativism. It became apparent that PÁnini had a much more thor" ough understanding of phonology than his Greek contemporaries, and his description of vocalic sandhi was even said to presuppose a theory similar to (early) generative phonology. Perhaps the field in which PÁnini has enjoyed most admiration lately is case theory. In" the late 1960s, after the outburst of Fillmorean Case Grammar, it was pointed out that PÁnini’s notion of kÁ raka corresponded to " what Fillmore called ‘deep cases’, or, to state it in more modern terminology, semantic roles. Thus, there is kÁ raka of the agent, which can be expressed in the nominative or instrumental, the use of the latter ending being demonstrated by means of passive constructions. In fact, when describing the use of ‘affixes’ (meaning both verbal and nominal affixes), PÁnini I mostly has a semantically based approach, rather " than an approach based on forms, as was usual within the Greco-Roman tradition. Some later grammarians gave sematically finer definitions: for example, within the philosophical school of NyÁya the kÁ raka of the agent was conceived as implying intentionality, so that only animate entities could be regarded as agents. On the contrary, PÁnini and " his followers (PÁnin¯iyas) called agents all subjects of " action verbs, including inanimate entities (in examples such as ‘the cart goes to the village’). References Cardona, George. 1994. Indian linguistics. History of linguistics, Vol. 1, ed. by G. Lepschy. Oxford: Oxford University Press. Joshi, Venkateshshastri. 1980. Problems in Sanskrit grammar. Pune: 2 Dastane Ramchandra & Co. Rocher, Rosane. 1975. India. Current trends in linguistics, Vol. 13, ed. by T. Sebeok. The Hague, Paris: Mouton. Vasu, S´r¯is´a a Chandra. 1962. The AstÁdhyÁy¯i of PÁnini "" " Allahabad 1891, repr. Delhi: Motilal Banarsidass.

SILVIA LURAGHI See also India; Indian Traditional Grammar; Sanskrit

Papiamento Papiamento (Papiamentu) is a language spoken north of Venezuela on the Leeward Islands of the Netherlands, Antilles, Curaçao, and Bonaire, and on the island of Aruba, otherwise known as the ABC islands. It is estimated that Papiamento is spoken 808

natively by about a quarter of a million people, mostly in the Caribbean, and by some 40,000 people living in the Netherlands. Although the official language on the ABC islands is Dutch, Papiamento pervades all levels of society, including the educational system and

PAPIAMENTO government. Papiamento also has a rich literary tradition, an established presence in electronic and digital media, and several newspapers and magazines are also published in the language. The name is derived from the verb papia meaning ‘to speak’ and the suffix, -mento. Thus, Papiamento can be literally interpreted as ‘(the act of) speaking’. Papiamento is a creole language, which emerged during the second half of the seventeenth century around the period of European colonization of the Caribbean islands. Like other creole languages, Papiamento inherited structural characteristics from a mixture of West-African languages (the substrate languages). Approximately two thirds of its vocabulary comes from Iberian languages (the superstrate languages), about a quarter from Dutch, and the rest from other languages including French and English. The language also has traces of Amerindian and West African vocabulary as well. The history of the ABC islands provides some insight into the social and linguistic context in which Papiamento developed: originally populated by Amerindians speaking an Arawakan language, Curaçao was settled by the Spanish in 1527. After a century of Spanish occupation, the Dutch conquered Curaçao in 1634 at which time most of the Amerindians and the Spanish were driven from the island. Shortly thereafter, it became a depot for the slave-trading practices of the Dutch West-India Company. Not long after the fall of the Dutch Brazilian Empire in 1654, a substantial population of Portuguese Jews (and non-Jewish settlers as well) arrived on the island. Despite the diverse linguistic backgrounds of the island’s inhabitants, Papiamentu is believed to have stabilized as a creole on Curaçao around 1700 and spread to Bonaire and Aruba by the end of the century. Although this brief account of the islands’ history demonstrates the importance of Dutch and Portuguese in the incipience of the language, Papiamento clearly exhibits Spanish elements and in fact, a diary belonging to a Jesuit missionary on Curaçao dating from 1704, first referred to the language on the island as ‘broken Spanish’. This has caused some controversy surrounding the origins of the language with respect to whether it is actually a Portuguese- or Spanish-based creole. Some scholars argue that the Portuguese contribution to the language has been obscured by its continual contact with Spanish. Consequently, the earliest forms of Papiamento are thought to have exhibited the Portuguese contribution more so than the Spanish; however, the earliest written attestation of the language, a letter dating from 1776 written by a Portuguese Jew to his wife, shows a remarkable resemblance to the language spoken on the islands

today, replete with elements of the Spanish language thought to be less prominent in the earlier stages of Papiamento. Thus, questions concerning the precise origins of Papiamento continue and as it is not directly related to any other Portuguese- or Spanish-based creole, scholars can only continue to piece together linguistic and historical evidence to help shed light on the genesis and subsequent evolution of the language. Papiamento has existed in tandem with Dutch on the islands for centuries; however, the Dutch influence in the language is far less obvious than the Iberian influences, in both historical and contemporary contexts. In spite of the fact that Dutch is the official language of the regions where Papiamento is spoken natively, it is quickly becoming far less common even in the areas of government administration, education, and the judiciary system, realms of society in which it was traditionally used, testifying to the high level of social prestige affiliated with Papiamento—a fact that makes it somewhat unique among creole languages. Papiamento’s social currency is also demonstrated by the fact that it has an official orthographic system, differing slightly on the islands of Curaçao and Bonaire than on the island of Aruba. These spelling differences can be seen in the name of the language itself, while the former islands prefer Papiamentu as opposed to Papiamento, its Aruban counterpart. Additionally, after careful language planning efforts, in 1987, Papiamento was officially integrated into the educational system of the ABC islands. The phonemes (sounds) of Papiamento include almost all the consonants and vowels found in American Spanish, as well as some Dutch vowels. A linguistic trait of Papiamento not found in either the Iberian languages or Dutch, and which is rare in creole languages in general, is the use of high (´) and low (`) tones to distinguish between pairs of words (eg. párà ‘bird’, pàrá ‘to stand, to stop’; wárdà ‘guard service’, wàrdá ‘to wait’). This feature of the language is attributed to the West African, or substratal influences. West African structural influences typical of creole languages can be noticed in serial-verb constructions such as Ela kore bai su kas or ‘She or He ran (go) home,’ or focus particles found in sentence-initial positions like Ta e buki m’a dunabu or ‘(Focus) I gave you the book.’ Below are some Papiamento words that illustrate the Iberian and Dutch contributions to the language: Portuguese: bai ‘to go’; preto ‘black’; bringa ‘to fight’ (derived from brigar) Spanish: Dios ‘God’; salida ‘departure’; ruman ‘brother’ (derived from hermano) Dutch: wikent ‘weekend’; spiel ‘mirror’ (derived from spiegel), and wak ‘to see’ (derived from waken).

809

PAPIAMENTO References Agard, Frederick B. 1985. Papiamentu grammars and the structure of the language. For Gordon H. Fairbanks, ed. by Veneeta Z. Ascon and Richard L. Leed. Honolulu: University of Hawaii Press. Anderson, Roger W. 1990. Papiamentu tense-aspect, with special attention to discourse. Pidgin and Creole Tense-Mood-Aspect systems, ed. by Jon Singler. Amsterdam: John Benjamins. Arends, Jacques, Pieter Muysken, and Norval Smith (eds.) 1994. Pidgins and Creoles: an introduction. Amsterdam: John Benjamins.

Goilo, E.R. 1997. Papiamentu textbook. Aruba: VAD Printers. Kouwenberg, Sylvia, and Eric Murray. 1994. Papiamentu. Newcastle: Lincom Europa. Martinus, Frank. 1990. Papiamentu: the road to emancipation. Language reform: history and future, ed. by István Fodor and Claude Hagège. Hamburg: Helmut Buske Verlag. Wood, Richard E. 1972. New light on the origins of Papiamentu: an eighteenth-century letter. Neophilologus 56.

BETSY BARRY See also Caribbean; Pidgins and Creoles

Paralanguage The term paralanguage refers to a wide variety of nonverbal behaviors relevant to communication. Vocal paralanguage includes voice quality, intonation, and voice dynamics such as loudness, pitch range, and rate of speech. Bodily paralanguage incorporates facial expression, hand gestures, posture, touch, and physical proximity to others. These behaviors, all of which may co-occur with verbal communication, differ from language per se in that they do not form part of a hierarchically structured, coded system in which a discrete symbol is arbitrarily associated with a particular meaning. While paralanguage does not convey meaning in the same way as language, it plays important functions in communication. Paralinguistic behavior conveys the emotional state of speakers and listeners, expresses semantic nuances that could not be inferred from language alone, helps regulate the flow of interaction, and provides cues regarding the sex, age, culture, and socioeconomic status of speakers. The study of paralanguage has grown significantly over the past 30 years. Technological advances such as audio and videorecording, and more recently, the widespread availability of computers and specialized software analysis packages, have facilitated a more systematic approach to the study of paralanguage. The sections below describe some key paralinguistic channels: voice quality, intonation, and hand gestures (see Siegman and Feldstein (1987) and Feldman and Rime (1991) for an overview of these and other paralinguistic channels).

Voice Quality Voice quality may be defined as a quasipermanent setting of the vocal tract that characterizes the speech of

810

an individual, or a cultural, social, or linguistic group (see Laver (1980) for a thorough introduction to this subject; and Kent and Ball (2000) for an overview of recent research). Voice quality settings may be roughly divided into two categories: laryngeal and supralaryngeal. Laryngeal settings (also known as ‘phonation types’) involve different patterns of vocal fold vibration. These settings include modal voice, breathy voice, creaky voice, harsh voice, falsetto voice, whisper, and the various possible combinations of these phonation types (e.g. harsh whispery voice). Modal voice involves quasiperiodic vibration of the vocal folds, in that the vocal folds are open and closed for approximately equal periods of time, lending the voice a smooth, even quality. Other laryngeal configurations may be thought of as departures from this ‘neutral’ setting. For example, for a breathy voice, the vocal folds are open longer than they are closed, producing audible friction. By contrast, for a harsh voice, the vocal folds are closed longer than they are open, and the vibration is irregular, producing a rasping, uneven quality that is often enhanced by the vibration of the ventricular or ‘false folds’. For a whisper, the vocal folds do not vibrate at all: the vocal folds are closed along most of their length, and a triangular opening is left at the posterior part of the vocal folds, generating strong friction as air passes through. Some laryngeal settings are readily identifiable by their pitch, as, for instance, the very high pitch of falsetto voice, involving the vibration of a small portion of the vocal folds, and the very low pitch of creaky voice, in which the listener can hear each separate vibration of the vocal folds. Supralaryngeal settings involve different configurations of the larynx, pharynx, soft palate, tongue, lips, and jaw, all of which affect the resonance of the voice.

PARALANGUAGE The larynx may be raised or lowered, the pharynx widened or constricted, the soft palate raised or lowered, the tongue and lips protruded or retracted, and the jaw lowered or raised—all of which, separately or in combination, have distinct acoustic and auditory effects. For example, a speaker with a habitually lowered soft palate will have a nasal voice. A speaker with a lowered larynx will have a deeper sounding voice than he or she would with a neutral or raised larynx, by virtue of having produced a longer vocal tract with correspondingly lower formant frequencies. The combination of various laryngeal and supralaryngeal settings contributes to the characteristic ‘tone of voice’ or ‘accent’ that listeners come to associate with different individuals and regional or social groups. For example, North American females are more likely than their male counterparts to speak in a habitually breathy voice—a habit that some women may deliberately alter in order to sound more professional in a particular work or social setting. Certain regional accents of English may be distinguished by a particular voice quality setting, such as, for instance, the nasal quality characteristic of speakers in the southern United States of America. Some foreign accents are stereotypically associated with a particular voice quality setting, such as the quasi-permanent retraction of the tongue tip (or ‘retroflexion’) that often characterizes an East Indian accent. In addition to suggesting social, cultural, and linguistic characteristics of speakers, voice quality settings can be deliberately exploited by a single individual in different communicative settings to convey paralinguistic meaning—whether to signal the speaker’s actual emotional state or attitude toward what he or she is saying. For example, in Englishspeaking countries, a whisper is often used to signal secrecy, while a breathy voice may be used to convey intimacy or sexual attractiveness. A creaky voice may sometimes be used to suggest sadness or boredom, while lowered larynx voice may be used by speakers who wish to sound serious or authoritative.

Intonation Intonation refers to the variations in pitch or fundamental frequency that occur in speech, creating its characteristic melodies. In intonational analysis, these melodies are broken down into ‘tone units’ or ‘intonation groups’. Tone units usually consist of five or six words, often constituting a grammatical constituent such as a sentence, clause, or phrase, but sometimes consisting of a single word, or even a single syllable. The major characteristic of a tone unit, independent of its length or grammatical status, is that it contains one major pitch movement, realized primarily on the most

prominent, or ‘accented’ syllable in the tone group. Within the British system of intonation analysis, there are five major nuclear tones: falls, rises, levels, fall–rises, and rise–falls. Within the predominantly American autosegmental metrical system, (for example, the ToBI, or ‘Tone and Break Index’ system), these and other contours are expressed as different patterns of high and low tones (‘H’ and ‘L’). (See Cruttenden (1997) for a thorough overview of intonation in the British and American traditions.) Intonation may have linguistic or paralinguistic meaning. For instance, in English, an utterance produced with falling intonation is associated with a statement or declarative sentence, while a rising intonation is often used for questions. Such uses of intonation are often considered linguistic. However, there are many intonational variations that carry paralinguistic meaning. For instance, some speakers use rising intonation for declarative sentences as a way of checking whether or not the listener understands what they are saying, to suggest openness to an alternative point of view, or, possibly, to convey a lack of confidence in what they are saying. This particular usage is more common among young North American females than their male counterparts. The word ‘yes’, spoken slowly with a rising then falling pitch, may suggest doubt or hesitation on the speaker’s part, even though the word itself conveys agreement. The same word, spoken slightly more quickly, with a falling then rising pitch, may suggest impatience or irritation on the part of the speaker. These are but a few of many possible examples of the paralinguistic use of intonation. Intonation, in combination with voice quality, and other features of voice dynamics, such as variations in pitch range (wide vs. narrow), intensity (loud vs. soft), and rate of speech (fast vs. slow), provide the listener with significant cues regarding the speaker’s attitude and/or emotional state. For example, a wide pitch range is often characteristic of a happy mood, and a more restricted pitch range is characteristic of a more depressed state. Anger is often associated in a loud, high-pitched voice, while fear may be expressed with a quieter, high-pitched voice, sometimes combined with a whispery quality. In a state of excitement, people often speak more quickly than when they are calm. The qualities associated with a particular mood may be characteristic of a speaker’s habitual use of the voice, possibly suggesting more permanent personality traits.

Gestures While speaking, people can often be seen moving their hands up and down, pointing, or drawing pictures in the air, in ways that appear connected to what they are saying. These movements, as opposed to other hand

811

PARALANGUAGE movements that they might happen to make while speaking, such as smoothing their hair or removing lint from their pants (the latter are called “self-adaptors”), are considered communicative gestures. Adam Kendon—a pioneer in the research on gestures—has characterized hand gestures as existing along a continuum, ranging from completely nonlinguistic gestures to the fully linguistic hand movements that comprise actual sign languages (see McNeill (1992:37) for a fuller description of Kendon’s continuum). In the middle of this continuum are the semilinguistic hand gestures known as ‘emblems’. Emblems have a fixed form that is associated with a particular meaning, and may in this sense, be considered to be quasilexical. Emblems vary considerably from culture to culture. For example, the ‘thumbs-up’ gesture often used in English-speaking countries to signal a victory or a generally positive attitude toward a particular situation, has an obscene connotation in Iran. Some cultures—Neopolitan Italian culture, for example—have elaborate sets of emblems that are extensively used in day-to-day conversation. Because of their quasilexical nature, emblematic gestures can often be used instead of words in face-to-face interaction for communicative effect. However, the majority of communicative hand gestures are nonlinguistic, and co-occur with speech. Nonlinguistic gestures have been categorized in a variety of ways over the past 50 years; but the categories developed by McNeill (1992) are those most commonly used in the current literature on gestures. Nonlinguistic gestures may be classified as iconic, metaphoric, deictic, or beats. Iconic gestures illustrate concrete entities in the world, while metaphoric gestures depict abstract concepts. Iconic and metaphoric gestures have no fixed form, and are spontaneously invented by speakers to physically represent some aspect of what they are saying. For example, while saying ‘and the guy just grabbed her purse’, a speaker might physically depict this action in an iconic gesture—thereby often conveying information not fully expressed in words, such as the precise movement involved in the grabbing. Deictic gestures involve pointing, either to a location within the speaker’s environment, or to an area in space that symbolizes a location referred to in the accompanying speech. Beats are short, rhythmic movements of the hands, often physically oriented toward the listener (e.g. involving an open palm facing the listener) and do not represent anything in the accompanying speech. Rather, they are used to emphasize the delivery of information, help structure discourse, and help regulate aspects of interaction, for instance, turn-taking routines. As noted above, nonlinguistic gestures are closely synchronized with speech. Gestures may be divided into three phases: a preparatory or onset phase, in 812

which the hands rise from a position of rest; a stroke phase, the most ‘energetic’ portion of the gesture, or the point at which the gesture appears to have reached its ‘destination’; and a recovery or offset phase, the point at which the hands return to a position of rest. The stroke of a gesture almost always co-occurs with the word or phrase with which it is meaningfully connected. Moreover, the stroke often co-occurs with the most prominent, or accented syllable in a tone unit, suggesting that the planning of gesture is closely linked with the planning of speech. In recent years, there has been a surge of interest in the study of gesture and other paralinguistic phenomena, such as facial expression, voice quality, and intonation. Traditionally, these features have often been considered to reflect the emotional state of speakers. While this approach continues to form a significant part of research on paralanguage, recent research has also highlighted the important semantic information conveyed through paralinguistic channels (see, for example, McNeill 1992, 2000). Moreover, there is a growing interest in the complex ways in which linguistic and paralinguistic channels interact. This line of research is leading to an enriched understanding of communication as integrated, multimodal behavior— and perhaps challenging the traditional distinction between language and paralanguage. References Armstrong, David, William Stokoe, and Sherman Wilcox. 1995. Gesture and the nature of language. Cambridge, New York, and Melbourne: Cambridge University Press. Birdwhistell, Ray. 1970. Kinesics and context. Philadelphia: University of Pennsylvania Press. Cruttenden, Alan. 1986. Intonation. Cambridge, New York, and Melbourne: Cambridge University Press. 2nd edition, 1997. Feldman, Robert, and Bernard Rime (eds.) 1991. Fundamentals of nonverbal behavior. Cambridge, New York, Port Chester, Melbourne, and Sydney: Cambridge University Press, and Paris: Editions de la Maison des Sciences de l’Homme. Kent, Raymond, and Martin Ball (eds.) 2000. Voice quality measurement. San Diego, California: Singular Publishing Group. Laver, John. 1980. The phonetic description of voice quality. Cambridge: Cambridge University Press. McNeill, David. 1992. Hand and mind. Chicago, and London: University of Chicago Press. ———. (ed.) 2000. Language and gesture. Cambridge, New York, Melbourne, and Madrid: Cambridge University Press. Poyatos, Fernando. 1993. Paralanguage. Amsterdam, and Philadelphia: John Benjamins Publishing Company. Siegman, Aron, and Stanley Feldstein (eds.) 1987. Nonverbal behavior and communication. Hillsdale: Lawrence Erlbaum Associates.

ALLISON BENNER See also Communication Theory; Conversation Analysis; Emotion and Language; Signed Languages

PARSING

Parsing A recognizer is an algorithm that takes a string as input and either accepts or rejects it, depending on whether or not the string is a sentence of a particular grammar (Grosz et al. 1986). A parser is a recognizer that also shows all the ways in which the string can be derived, in the form of the invoked grammar rules arranged in trees or nested square brackets. The most common way to represent a grammar is as a set of production rules, which are also called rewrite rules. Phrase structure grammars are sets of production rules that specify the various ways in which a sentence can be decomposed into its constituent syntactic units, which themselves can be decomposed into smaller units, and finally into individual words. The production rule S → VP NP shows that a sentence can consist of a noun phrase followed by a verb phrase, and the rule ADJ → small|red|new shows that an adjective can consist of one of the following individual words: small, red, or new. Symbols that are themselves further expanded by rules are called nonterminal symbols, which generally correspond to syntactic classes. Symbols that correspond to individual word strings that must be found in the input sentence are called terminal symbols (Rich and Knight 1991). The nonterminal symbol S, which stands for the input sentence in its entirety, is called the root of the grammar. An example of a simple grammar which derives the sentence ‘John ate the cat’ is given by Allen (1995:42). (1) (2) (3) (4) (5) (6) (7) (8)

S → NP VP VP → V NP NP → NAME NP → ART N NAME → John V → ate ART → the N → cat

The parse tree corresponding to this parse is shown in Figure 1. S VP

NP NAME John

NP

V ate

ART

N

the

cat

Figure 1. A tree representation of John ate the cat.

Context-Free Parsers The term context free (CF) was introduced by Chomsky in 1956. CF grammars consist of rules with a single symbol on the left-hand side. Such rules are always applicable, since they are in no way dependent on the nature of constituents elsewhere in the input text. Most of the CF parsing algorithms were developed in the 1960s for use with computer language compilers, but pure CF grammars are not effective for describing natural languages. Grammars that are not CF are said to be context sensitive, where the rules have been augmented with additional conditions, meaning that they only apply if certain things are found elsewhere in the input text.

Top-Down and Bottom-Up Parsing A top-down parsing strategy starts with the start symbol S and then searches through different ways to rewrite the symbols until either the input sentence is reproduced or all different ways to rewrite the symbols have been tried. A top-down parse using the simple grammar above to derive the sentence ‘John ate the cat’ would proceed as follows: S        

NP VP (rewriting S) NAME VP (rewriting NP) John VP (rewriting NAME) John V NP (rewriting VP) John ate NP (rewriting V) John ate ART N (rewriting NP) John ate the N (rewriting ART) John ate the cat (rewriting N).

The bottom-up parsing strategy, on the other hand, starts with the actual words of the sentence and uses the rewrite rules in reverse until either the S symbol alone is generated or all combinations of the rewrite rules have been tried. A bottom-up parse of ‘John ate the cat’ might proceed as follows: John ate the cat        

NAME ate the cat (rewriting John) NAME V the cat (rewriting ate) NAME V ART cat (rewriting the) NAME V ART N (rewriting cat) NP V ART N (rewriting NAME) NP V NP (rewriting ART N) NP VP (rewriting V NP) S (rewriting NP VP) 813

PARSING LR parsers read their input from left to right and produce a rightmost derivation, i.e. expand the rightmost non-terminals of the production rules first.

Chart Parsers In a standard parser, if an early segment can be analyzed in more than one way, each step in the recognition of subsequent phrases in the sentence will be repeated once for each earlier analysis. Chart parsers (Winograd 1983) provide a means of eliminating such redundant computation. Kay described the chart as ‘a kind of well formed substring table’. It allows the parser to store the partial results of the matching it has done so far so that this work need not be duplicated. First, we will consider the bottom-up chart parser. Each word in the input sentence is considered in turn. The current input word is called the key. For each new key, we look for rules that match a word sequence involving the key—either a rule that begins with the key, or a rule that has already been started by earlier keys, which can either be extended or completed by the current key. Allen presents the following example of a grammar: (1) (2) (3) (4) (5) (6)

S → NP VP NP → ART ADJ N NP → ART N NP → ADJ N VP → AUX VP VP → V NP

If we are parsing a sentence that begins with ART, ART becomes the first key, and rules (2) and (3) are matched since their right-hand sides both begin with ART. The following notation is used to show that rules (2) and (3) can be continued from the point of the ART (denoted by ‘o’) onwards. The partially matched rules are stored in the chart, along with their locus (starting position in the sentence) and position of the last word to match so far: (2′) NP → ART o ADJ N (1,2) (3′) NP → ART o N (1,2) If the next key is ADJ, then rule 4 can be started, and rule 2′ can be extended: (2′) NP → ART ADJ o N (1,3) (4′) NP → ADJ o N (2,3) The chart keeps the record of all the rules found so far which have been matched partially by the previous keys. These partially matched rules are called active arcs. Active arcs are never removed from the chart; hence, their original and extended forms are simultaneously present. The chart also records the completely matched syntactic constituents (as defined by the grammar rules) found so far in the parse.

814

Completed constituents are placed on a list called the agenda. The algorithm proceeds by selecting a completed constituent from the agenda. For each matching rule in the grammar, an active arc is combined with the completed constituent in the agenda. Whenever the agenda is empty, all possible syntactic interpretations of the next word in the sentence are added. The parse ends when an active arc and a completed constituent combine to produce an entire sentence S. If all possible parses of a sentence are required, the process continues until the agenda is empty. The top-down chart parser is called the Earley parser (Earley, 1970).

Augmented Transition Networks An early model for natural language grammars was a transition graph called the finite state machine. This consists of a network of nodes (states) and directed arcs or transitions, marked by arrowheads pointing away from the previous state and toward the next state. Each arc has a symbol, representing terminal symbols (individual words such as ‘red’ or syntactic classes such as ‘adjective’) which may appear in an input sentence. There may be more than one arc arriving at each state or leaving each state. There is one distinguished state called the start state (with no inward arcs), and a number of distinguished states called final states (with no outward arcs). Starting at the start state, the words of the sentence to be parsed are read off in turn. If the first word read off matches the symbol on one of the arcs leaving the start state, the next state is the one found by following that arc. The second word is read in, and if that word matches an arc leaving the second state, the arc is followed to the third state. This process continues, and if we have arrived at one of the final states as a result of reading in the last word, the sentence has been accepted by the augmented transition network (ATN) grammar. Sometimes an arc can return to the very same state that it left, to show that optional additional words of the type specified by the arc’s symbol are permitted at that point in the sentence. ATN are finite state machines augmented by allowing not only terminal symbols on the arcs, but nonterminal symbols representing entire phrases (Woods 1970). These phrases are themselves represented by ATNs. For example, an ATN designed to recognize a prepositional phrase is shown in Figure 2.

PREP Start

NP S2

Finish

Figure 2. A simple augmented transition network.

PARSING To make the transition from S2 to finish, there must be a noun phrase (NP) in the input sentence. In an operation called a ‘push’, processing is suspended at the point the NP is requested, and a new ATN is entered, which accepts subsequent words from the input sentence. This is analogous to a subroutine call in a computer program. When the final state of the new ATN is reached (showing that the sequence of input words since entering the new ATN was indeed an NP), an operation called a ‘pop’ takes place and processing of the input sentence resumes at the point in the original ATN from where it left off. A called ATN may itself call other ATNs, and an ATN may call itself (recursion). The record of the sequence in which the ATNs have called each other is called a ‘stack’, analogous to a stack of plates, each plate representing an ATN, which may be ‘pushed’ onto or ‘popped’ off the stack. The most recently called ATN is at the top of the stack, but it cannot be ‘popped’ off until any ATNs it calls itself are completed. Advantages of ATNs are their generative power, efficiency of representation, and their ability to capture regularities in language.

Probabilistic Parsing A deterministic parser consists of a set of rules, each with equal a priori likelihood. It either generates just one parse according to the ordering of the rules, or all possible parses of a sentence, but does not mention whether any one of these parses is any likelier than the others. A probabilistic parser, on the other hand, may offer alternative rules for the decomposition of the constituents of a sentence, each with an associated probability value, e.g. VP → V NP 0.6, VP → V PP 0.4. The overall probability of a parse is found by multiplying together the probabilities of all the rules that were called upon in the derivation of that parse. The likeliest parse among many possible parses is found by the Viterbi algorithm. The strategy of exploring the paths with high probability constituents first is called Best First Parsing. The Penn Treebank contains over two million words of American English parsed manually. It has proved useful as a reference corpus for the comparison of different parsing algorithms, such as in the PARSEVAL evaluations (Black et al. 1991). In 1994, Magerman was able to derive rules automatically for a probabilistic parser through an analysis of the Penn Treebank. The simplest approach is to count the number of times each rule is used in the treebank and normalize this to estimate the probability of each rule. For certain grammars, the most probable parse cannot be found easily using the Viterbi algorithm. In such cases, a most probable parse may be found using Monte Carlo simulation techniques. Many parse trees

for the same sentence are randomly generated. The likelihood of each rule being used in a particular random parse depends on its probability. When the sample is large enough, the most frequently occurring output parse tree over all the random tests is assumed to be the most probable parse.

Tagging Tagging is marking items in a text with additional information, related to their linguistic properties. This information is attached to the text items in the form of codes called tags. Here, we will mainly discuss the automatic assignment of part of speech (POS) category tags to words in a text; but tagging can also involve the assignment of semantic categories as a by-product of word sense disambiguation. The Brill (1995) tagger operates as follows. An initial guess is made as to the POS of each word in the text by an initial state annotator. Simple methods of doing this include (a) labeling all words with their most frequent tag as found in the training corpus and (b) assuming all unknown words are nouns. The output of the initial state annotator is then compared with the manually annotated Penn Treebank, and a list of transformations is learned that can be applied to the output of the initial state annotator, according to the nature of words in the vicinity of each tagged word. The transformations have two components, a rewrite rule (such as ‘change the tag from modal to singular noun’) and a triggering environment (such as ‘if the preceding word is a determiner’). Using this transformation, if the word ‘can’ has been given an incorrect initial tag as in ‘the_DT can_MD rusted_VBD’, the modal (MD) tag is replaced by a singular noun (NN) tag, because the previous word has a determiner (DT) tag. Other transformations are ‘change an infinitive verb to a singular noun if one of the previous two tags is a determiner’, and ‘change a singular noun to a verb infinitive if the previous word is ‘to’. The CLAWS probabilistic tagger (Garside et al. 1987) takes an input sentence such as ‘Henry likes stews’. and compares it word by word against the CLAWS lexicon, which shows that ‘Henry’ must be tagged as a proper noun (NP); ‘likes’ and ‘stews’ could either be tagged as plural nouns or verbs (NNS or VBZ), and the full stop always has its own tag (.). For each of the four possible tag sequences spanning this region of ambiguity, a value is generated by calculating the product of the frequencies per thousand for successive tag adjacencies as found in the Brown corpus, as shown below. The number 17 shows that in the Brown corpus, the tag NP is followed by the tag NNS 17 times per 1,000, while 983 times out of 1,000 it is followed by something else.

815

PARSING Value(NP-NNS-NNS-.)  17  5  135  11,475 Value(NP-NNS-VBZ-.)  17  1  37  629 Value(NP-VBZ-NNS-.)  7  28  135  26,460 Value(NP-VBZ-VBZ-.)  7  0  37  0 Thus, the tag sequence ‘Henry_NP likes_VBZ stews_NNS ._.’ is found to be the likeliest. Since there are generally too many possible sequences to find a value for them all, the Viterbi algorithm (Forney 1973) is used, which means that only the most promising sequences are tested. The CLAWS tagger relies on bigram (pairs of adjacent tags) frequencies, while in 1988, Church described a tagger based on trigram (sequences of three adjacent tags) frequencies.

Affix Stripping Affix stripping and morphological parsing are parsing at the subword level. Instead of analyzing a sentence in terms of its constituent words, a word is analyzed in terms of its constituent morphemes. Klavans and Chodorow (1991) used an instructional morphological parser to teach morphological theory. Several sets of rules exist for the automatic removal and replacement of common suffixes (stemming), including those produced by Paice (1990), Lovins, and Porter. Recognizing suffixes automatically can help in POS identification, where, for example, any word ending in ‘ation’ must be a noun. The simplest form of stemming is the reduction of all regular nouns and verbs to their inflectional root, which is the singular form for nouns and the infinitive form for verbs. The rule sets for the removal and replacement of common suffixes are in the form of production rules, in the form ‘old suffix → new suffix’, meaning that if a word ends in the old suffix, this must be removed and replaced by the new suffix. The new suffix may be null, meaning that the old suffix is removed and is not replaced with anything. The rules must be presented in a fixed order. Four of the rules given by Paice are as follows: (1) (2) (3) (4)

ING → NULL E → NULL ION → NULL OLV → OLUT

Imagine that the word ‘resolving’ is given to each rule in turn. Rule (1) removes the suffix ‘ing’ and replaces it with nothing; hence, we are left with ‘resolv’. This does not end in ‘e’ or ‘ion’, hence, rules (2) and (3) have no effect. However, rule (4) states that the current suffix ‘olv’ must be removed and replaced by ‘olut’, leaving the string ‘resolut’. The words ‘resolve’ and ‘resolution’ will also be reduced to ‘resolut’, rendering all three words equivalent. Prefix removal may be less

816

useful than suffix removal, as the removal of a prefix may radically alter the meaning of a word: consider the group ‘bisect’, ‘dissect’, ‘insect’, ‘sect’.

Semantic Parsing A semantic grammar, like a syntactic grammar, consists of production rules, but instead of specifying only the syntactic constituents that may be present in a sentence, may also specify domain-specific semantic categories. Allen (1995) gives an example in the domain of airline bookings, of a semantic grammar that should cater for such input phrases as ‘flight 457 to Chicago’. This would include such rules as ‘FLIGHT-NP  FLIGHT-N NUMB’ (A noun phrase referring to a flight consists of a noun referring to a flight and a number) and ‘FLIGHT-NP  FLIGHT-N to CITY-NP’ (A noun phrase referring to a flight consists of a noun meaning ‘flight’, followed by ‘to’, then a noun phrase which is the name of a city). Different algorithms for semantic parsing vary according to whether it is necessary to produce a syntactic parse first, or whether one should proceed directly with the semantic parse. References Allen, James. 1995. Natural language understanding. Redwood City, CA: Benjamin Cummings. Black, Ezra et al. 1991. A procedure for quantitatively comparing the syntactic coverage of English grammars. Proceedings of the fourth DARPA Workshop on Speech and Natural Language. San Mateo, CA: Morgan Kaufman. Brill, Eric. 1995. Transformation-based error driven learning and natural language processing: a case study in part-ofspeech tagging. Computational Linguistics 21(4). 543–65. Earley, Jay. 1970. An efficient context-free parsing algorithm. Communications of the ACM 13(2). 94–102. Forney, G D Jr. 1973. The Viterbi Algorithm. Proceedings of the IEEE 61(3). 268. Garside, Roger, Geoffrey Leech, and Geoffrey Sampson (eds.) 1987. The computational analysis of English, a corpusbased approach. London: Longman. Grosz, Barbara J, Karen Spärck Jones, and Bonnie Lynn Webber (eds.) 1986. Readings in natural language processing. Los Altos, CA: Morgan Kaufmann. Klavans, Judith, and Martin S Chodorow. 1991. Using a morphological analyser to teach theoretical morphology. Computers and the Humanities 5. 281–87. Paice, Chris D. 1990. Method for evaluating stemming algorithms based on error counting. Journal of the American Society for Information Science (JASIS) 47(8). 632–49. Rich, Elaine, and Kevin Knight. 1991. Artificial intelligence. New York: McGraw-Hill. Winograd, Terry. 1983. Language as a cognitive process: syntax. Reading, MA: Addison-Wesley. Woods, W A. 1970. Transition network grammars for nnatural language analysis. Communications of the ACM 13(10): 591–606.

MICHAEL P. OAKES See also Phrase Structure

PAUL, HERMANN

Paul, Hermann Hermann Paul was one of the most influential linguists of the late nineteenth and early twentieth centuries. Paul was a neogrammarian (Germ. Junggrammatiker), a school of linguistics that flourished in his time and had an immense influence on linguistics thereafter, as discussed further below. His interest in languages and linguistics dates back to his schooldays, when he developed a taste for medieval German literature and Middle High German (superseding an earlier interest in mathematics). In a brief autobiographical essay published just after his death, Paul suggested that his work could be divided into two general categories: literary interpretations and textual criticism, with a focus on Middle High German, and the phonology and morphology of the Germanic languages. While Paul was certainly an important contributor to these fields, this self-classification is something of an oversimplification, as he made significant contributions to a number of other fields of linguistics and German studies. Paul was certainly an influential figure in literary studies and textual criticism. His dissertation and Habilitionsschrift both dealt with questions of Middle High German literature. He authored a variety of articles on literary topics, ranging from the life of Hartman von Aue, an important Middle High German poet, to the question of whether a Middle High German literary language ever really existed. As to textual criticism, he was the founding editor of the Altdeutsche Textbibliothek [Early German text library], and edited three of the volumes of the series, all Middle High German texts. His Mittelhochdeutsche Grammatik [Middle High German grammar], first published in 1881 and republished many times (11 editions appeared in Paul’s lifetime), remains a standard handbook in the field. Paul’s contributions to the study of Germanic phonology and morphology were also immense. He published a monograph-length article entitled ‘Zur geschichte des germanischen vocalismus’ (‘On the history of germanic vocalism’), as well as articles on the vowels of the inflectional and derivational syllables in the oldest Germanic dialects and the ablative case in Germanic, among others. His most important work in this area is the five-volume Deutsche Grammatik (German grammar), described in his obituary as ‘die einzige vollendete wissenschaftliche Grammatik der deutschen Sprache’ (‘the only completed scientific grammar of the German language’).

Other areas of linguistics Paul contributed to include lexicography, syntax, word formation, and language pedagogy. He compiled a German dictionary, the Deutsches Wörterbuch (German dictionary), which remains in print today. Paul was also interested in the theoretical aspects of lexicography, publishing various essays on lexicographic topics (although the titles of these essays indicate that his main concern was with his German dictionary). As for syntax, Paul recognized its importance early on; his Middle High German grammar was one of the first such handbooks to include a section on syntax (although it must be admitted that this section was not in the first edition of the book; it was added to the 2nd edition, published in 1884). His work in word formation was generally an offshoot of his lexicographical work; he published an important article on word formation, ‘Zur Wortbildungslehre’ (‘On the doctrine of word formation’), in 1896. Finally, a volume on language pedagogy, Über Sprachunterricht (On language teaching), was published posthumously in 1922. Probably Paul’s most significant contribution to linguistics was his work Prinzipien der Sprachgeschichte (Principles of the history of language), originally published in 1880. This volume rapidly became a standard handbook (five editions of the text were published during Paul’s lifetime, along with an English translation of the second German edition), and proved to have a lasting influence on the field. It is notable not only for its incisive statements of many Neogrammarian ideas but also for its innovativeness, and is rewarding reading even today. Finally, Paul’s activities as an editor cannot be underestimated. Together with his friend and former fellow student Wilhelm Braune, Paul founded a journal, the Beiträge zur Geschichte der deutschen Sprache und Literatur [Contributions on the history of the German language and [its] literature], which is still generally referred to as the ‘PBB’, for PaulBraune Beiträge. Paul served as coeditor of this journal for nearly 20 years, from 1874–1891, and was a consulting editor for 30 years after that. Paul was also the general editor of the Grundriss der germanischen Philologie (Outline of Germanic philology). Furthermore, as noted above, he was the founding editor of the Altdeutsche Textbibliothek (Early German text library), and edited three of the volumes of the series (listed above). Despite his many contributions to the field, Paul is too often dismissed as ‘merely’ a historical linguist. 817

PAUL, HERMANN This viewpoint is entirely natural, since most of his work was on historical linguistics, and because of statements like the following, found in his Prinzipien: ‘Es ist eingewendet, dass es noch eine wissenschaftliche Betrachtung der Sprache gäbe, als die geschichtliche. Ich muss das in Abrede stellen’. [‘It has been claimed that there is another way to view language than the historical. I must object to this.’] On the other hand, it has also been claimed that Paul was much more than ‘merely’ a historical linguist. It has been argued that Paul was a forerunner of Structuralism; more specifically, that certain concepts found in the work of Ferdinand de Saussure, the well-known Swiss linguist who is generally viewed as the intellectual father of Structuralism (especially the distinctions between synchrony and diachrony and between langue and parole), were inspired by Paul’s work (e.g. by his differentiation between individual linguistic acts and more general language use). The first of these viewpoints must be rejected; a closer reading of Paul’s Prinzipien readily indicates that he was also interested in synchronic topics. The second of these viewpoints is probably also an overstatement. It is clear that de Saussure codified a number of ideas that were current linguistic theory at his time, but to trace the critical distinction between synchrony and diachrony to Paul, as some have, for example, is probably an exaggeration. Regardless of these ideas, it is clear that Paul was an immensely important figure in linguistics. Paul played an extremely important role in the development of the Neogrammarian school of linguistics; he has been described by one historian of linguistics of the period as the Neogrammarians’ ‘most fertile and formidable theoretician’, and his Prinzipien codified many aspects of Neogrammarian thought. More specifically, this same historian of linguistics (Kurt Jankowsky; the relevant work is cited under ‘Further Reading’ below) has argued that Paul’s influence on his contemporaries can be traced by examining five major aspects of his work. First, Paul rejected the equation of linguistics to the natural sciences, suggesting that the term ‘sound law’ should not be understood as a ‘law’ is understood in physics and chemistry. Second, Paul distinguished between historical and comparative linguistics. Third, Paul was interested not only in how languages change, but why they do so. Fourth, Paul emphasized individual psychology over group psychology in his psychological approach to language. Fifth, Paul was deeply interested in linguistic geography, recognizing early on that a standard language is a convenient abstract fiction, made up of a variety of dialects and idiolects. A discussion of the place of the Neogrammarians in the history of linguistics is clearly beyond the scope of this entry, but their influence on linguistics, especially historical linguistics, was considerable. Numerous prominent linguists, in America and

818

abroad, ranging from Leonard Bloomfield, the dean of American linguistics in the 1930s and 1940s, to William Labov, a prominent sociolinguist of today, were deeply influenced by Neogrammarian thought. Given the importance of the Neogrammarians, and Paul’s importance within the group, his place in the history of linguistics seems assured.

Biography Paul Hermann was born in Salbkempierc bei Magdeburg, Germany on, August 7, 1846. He attended the University of Berlin 1866–1867. He received a Doctorate (1870) for his dissertation on Freidank, Habilition (1872) for work on Gottfried’s Tristan, University of Leipzig. He was Privatdozent in Leipzig in 1872–1874. He was also Professor of German, University of Freiburg im Breisgau, 1874–1893; Professor of German, University of Munich, 1893–1916; and Rektor, University of Munich, 1909–1910. He was a corresponding member of the Wissenschaftliche Akademie in Vienna in 1919–1921. Hermann died in Munich, Germany, on December 29, 1921. References Antal, Lázló. 1985. Some comments on the relationship between Paul and Saussure. Cahiers Ferdinand de Saussure 39. Henne, Helmut, and Jörg Kilian (eds.) 1998. Hermann Paul: Sprachtheorie, Sprachgeschichte, Philologie. Tübingen: Niemeyer. Jankowsky, Kurt R. 1972. The neogrammarians. A re-evaluation of their place in the development of linguistic science. The Hague: Mouton. Koerner, E.F.K. 1972. Hermann Paul and synchronic linguistics. Lingua 29. Labov, William. 1994. Principles of linguistic change. Internal factors. Cambridge, MA: Blackwell. Paul, Herman. 1879. Zur geschichte des germanischen vocalismus. Beiträge zur Geschichte der deutschen Sprache und Literatur 6. ———. 1880. Prinzipien der Sprachgeschichte; transl. into English by H.A. Strong as Principles of the history of language, 1889; reprint of 8th edition 1970. ———. 1881. Mittelhochdeutsche Grammatik. ———. (ed.) 1894. Grundriss der germanischen Philologie. ———. 1896. Zur Wortbildungslehre. Sitzungsberichte der philos.-philol. Klasse der bayerischen Akademie der Wissenschaften. ———. 1897. Deutsches Wörterbuch. ———. 1916–1920. Deutsche Grammatik. ———. 1922. Über Sprachunterricht. ———. 1992. Mein Leben. Beiträge zur Geschichte der deutschen Sprache und Literatur 46. Reis, Marga. 1978. Hermann Paul. Beiträge zur Geschichte der deutschen Sprache und Literatur 100.

MARC PIERCE See also Historical Linguistics; Leskien, August; Saussure, Ferdinand de

PEIRCE, CHARLES SANDERS

Peirce, Charles Sanders Charles Sanders Peirce was a most original and highly versatile American philosopher, logician, and scientist—by all standards, a veritable polymath. Son of Benjamin Peirce, a distinguished Harvard mathematician, Charles was born in 1839 and grew up in a family milieu that was most conducive to a promising career in academia. Yet, except for a brief period as a lecturer in logic at the Johns Hopkins University (1879–1884), Peirce never did any systematic teaching. He was a difficult person to get along with and the university refrained from renewing his contract on charges of personal misconduct. His most sustained employment was with the U.S. Coast Survey, where he worked for 31 years and made a successful career as a geodesist, until the termination of his appointment in 1891. Thereafter, he set up a private practice as a chemical engineer and, thanks to the intervention of his friend and admirer William James, was invited to give occasional series of lectures at Harvard. His last years were spent in penury and failing health. Peirce is widely known as the founding father of Pragmatism, a distinctively American philosophical movement whose central tenet is that belief guides action. Our ordinary beliefs as well as more sophisticated doctrines are nothing but dispositions to act in certain ways and are to be judged by taking into account their practical consequences. Peirce therefore rejected much of metaphysics—i.e. claims about the ultimate reality of what there is—since no practical consequences necessarily follow from alternative metaphysical conceptions. He was averse to the sort of intuitive rationalism as advocated by Descartes and foundationalism as encapsulated in the Cartesian quest for first principles, as well as his celebrated ‘Method of Doubt’. As he famously put it in an essay published in 1878 titled ‘How to make our ideas clear,’ truth is simply that opinion which is destined to be agreed to by everyone who is disposed to finding it and so-called reality is made up of the objects represented in that opinion. The idea of truth as resulting from a convergence of opinion was interpreted by many of his followers to mean that there is no more to it than a matter of mere consensus. William James for one regarded truth as ‘whatever proves to be good in the way of belief’. But it is clear from Peirce’s later writings that he subscribed to an objectivist conception of reality—the idea that there is a reality out there that is independent

of what anyone believed about it. As for his epistemology, Peirce rejected the Cartesian attempt to use self-knowledge as the doorway to knowledge about the world, arguing instead that all knowledge is fallible, although continuous inquiry makes it self-corrective. He also contended that the ultimate reality is mental, matter being nothing but ‘effete mind’, and that science provides us with the best method of arriving at it. Along with the Swiss linguist Ferdinand de Saussure, Peirce is also widely looked upon as a founding father of semiotics (or ‘semeiotic’, as he himself preferred to spell it). Unlike Saussure who claimed the sign relation to be dyadic—involving a sign (signifier) and a referent (signified) Peirce insisted that it was irreducibly triadic and involved, in addition to the object and the sign, an interpretant (the interpreting thought or meaning). The sign, in Peirce’s conception, mediates between its object and its interpretant, which are therefore its correlatives. Central to Peirce’s thinking is the idea that language does not simply and unproblematically refer to a reality external to it; it does so by representing it. This means, he claimed, all of cognition is necessarily mediated by the process of sign interpretation or semiosis. Peirce also viewed semiotics as fundamentally a classificatory science, like chemistry and biology. Furthermore, signs are not restricted to the activity of the human mind but are of the order of nature itself and, in this too, he differed from Saussure, for whom ‘semiology’ was part of social psychology. Unlike Saussure too, Peirce saw semiotics as teleological, tending toward the truth of things as they really are. Peirce was a meticulous system-builder and his mind worked typically by positing trichotomies. Building on his own triadic characterization of sign relation, Peirce initially distinguished three divisions of signs: based on the nature of the sign itself, the relation between the sign and its object, and the role of the interpretant in representing the object. In their turn, these three divisions harked back to three ontological categories: quality (firstness), relation (secondness), and representation (thirdness). The first division, the one based on the nature of the sign itself, yielded a three-way distinction among ‘qualisign’ (appearance), ‘sinsign’ (token), and ‘legisign’ (type). Under the second division, the one obtained by taking into account the relation of the sign to its object,

819

PEIRCE, CHARLES SANDERS Peirce distinguished icons (that bear a certain resemblance to the object), indices (that signify by direct causal interaction with the object in question), and symbols (that are associated with their object via convention). Finally, under the third division, the one where the interpretant enters the picture, yet another triad was postulated: rhemes (predicative signs), prepositional signs, and arguments. The arguments were further divided into the three categories—abduction (hypothesis/diagnosis), deduction (tautological reasoning), and induction (generalization)—the first two leading to qualitative or conceptual understanding and the last providing important quantitative details to it. Toward the end of his life, Peirce grew increasingly impatient with the way the word ‘pragmatism’ was being used (and, in his view, abused) by other writers including William James and John Dewey and, in 1905, in a paper called ‘What pragmatism is’ published The Monist, pleaded that the word ‘pragmaticism’ be used instead to refer to his own doctrine, and distinguished it from that of the others, adding that the neologism was ‘ugly enough to be safe from kidnappers’. There followed two other papers (forming a trilogy): ‘Issues of pragmaticism’ (1905) and ‘Prologomena to an apology for pragmaticism’ (1906). Peirce died in 1914, an almost forgotten man. Early commentators often described him as ‘a wasp in the bottle’, someone who constantly came up against formidable odds and did everything but avail himself of the only way out of his troubles. Starting with the last quarter of the twentieth century or so, however, there has been a resurgence of interest in Peirce’s work and a widespread recognition of the importance of his views to such areas as epistemology, linguistics, anthropology, psychology, and cognitive science.

Biography Charles Sanders Peirce was born on September 10, 1839 in Cambridge, Massachusetts (USA). In 1855, he entered Harvard College, graduating in 1859 and continuing as a ‘resident graduate’ for one year. The same year, he began his career as a field aid at the Coast Survey, specializing in geodesy and working his way up until being appointed an Assistant. He entered Lawrence Scientific School at Harvard in 1861 and two years later obtained a graduate degree in Chemistry. He delivered Harvard lectures on ‘The Logic of Science,’ in the spring of 1865 and Lowell Institute lectures on ‘The Logic of Science; or Induction and Hypothesis,’ October 24,–December 1, 1866. He was elected to the American Academy of

820

Arts and Sciences on January 30, 1867. He was Assistant at Harvard Observatory, October 1869– December 1872 and was elected to the National Academy of Sciences on April 20, 1877. He was appointed lecturer in Logic at Johns Hopkins in 1879 and was dismissed on charges of misconduct in 1884. He was elected to the American Association for the Advancement of Science in August 1880. He purchased a farm near Milford, Pennsylvania on May 10, 1888; over the next two years, he reconstructed the existing farmhouse and bought more land, and named the estate ‘Arisbe’. The manuscript of ‘How to reason’ was rejected by both Macmillan and Ginn Co (1894) and that of ‘New elements of mathematics’ was rejected by Open Court (1895). In 1902, a grant application for ‘Proposed memoirs on minute logic’ was rejected by Carnegie Institution. He gave lectures on ‘Pragmatism’ at Harvard March 26–May 17, 1903 and Lowell lectures on ‘Some topics of logic’ November 23–December, 17, 1903. He was invited to deliver the Harvard Philosophy Club lectures on ‘Logical methodeutic’ April 8–13, 1907. Peirce died at Arisbe on April 19, 1914, of cancer. References Apel, Karl-Otto. 1981. Charles S. Peirce: from pragmatism to pragmaticism. English translation by John Michael Krois. New Jersey: Humanities Press. Auspitz, Josiah Lee. 1994. The wasp leaves the bottle: Charles Sanders Peirce. The American Scholar 63. Brent, Joseph. 1993. Charles Sanders Peirce: a life. Bloomington, Indiana: Indiana University Press. Brock, J. E. 1974. Peirce’s conception of semiotic. Semiotica 14. Freeman, Eugene (ed.) 1983. The relevance of Charles Peirce. La Salle, IL: The Hegeler Institute. Hookway, Christopher. 1985. Peirce. London: Routledge & Kegan Paul. Ketner, Keneth Laine (editor) 1995. Peirce and contemporary thought. New York: Fordham Univerity Press. Moore, Edward C. 1993. Charles S. Peirce and the philosophy of science. Tuscaloosa, London: The University of Alabama Press. Murphey, Murray G. 1993. The development of Peirce’s philosophy. Indianapolis, Indiana: Hackett Publishing Company Inc. Peirce, Charles Sanders. 1931–1935. Collected papers of Charles Sanders Peirce, Vols. 1–6, ed. by Charles Hartshorne and Paul Weiss. Cambridge, MA: The Belknap Press of Harvard University Press; Vols. 7–8, ed. by Arthur Burkes. Cambridge, MA: The Belknap Press of Harvard University Press, 1958. ———. 1982. Writings of Charles S. Peirce: a chronological edition, 5 vols. Bloomington, IN: Indiana University Press. Rosenthal, Sandra. 1994. Charles Peirce’s pragmatic pluralism. New York: State University of New York Press.

KANAVILLIL RAJAGOPALAN See also Philosophy of Language; Sign Relationship

PERSONALITY AND LANGUAGE

Personality and Language Personality and language are two terms that suggest some promising interconnections, such as the likely effect of personality pathologies on speech, the effect of personality traits on the use of certain linguistic— especially discursive—patterns, or the influence of individual learner differences on the acquisition of a second language. The effect of language on the configuration of personality is apparent if we think of the lexical terms used in a given culture to refer to personality phenomena. The ways in which people describe their own or others’ personalities and the fundamental influence of a person’s speech style on others’ perception of his or her personality are examples of the interrelationship of the two. Additionally, personality may be seen as a key factor in the construction of language and of the speech style of each individual. Notwithstanding these areas of inquiry, we are still far from claiming any conclusive direct effect of one over the other. Not surprisingly, it is speech (i.e. the oral performance of language by a given speaker) that has been the primary focus of researchers interested in the links between language and personality. These researchers have generally avoided claiming strong connections between personality and language as a system and have instead claimed a relationship between speech style and personality features, either self-assigned by the speaker or attributed by listeners. Psychological research on personality has not been extremely consistent in providing a clear model against which to test any possible mutual influences of language and personality. Existing models range from viewing the structure of personality as composed of one or two dimensions, to the three or five factors that are currently the most widely accepted models. In the 1990s, the so-called Five Factor Model gained popularity within the research community. This model was first proposed in the 1960s, but it started receiving wide attention and empirical support three decades later. However, it still has to be confirmed in studies involving non-Western societies, where a different categorization of personality traits may be more accurate. In Western societies, the model seems to provide a good account of personality structure with the following five factors: extraversion, agreeableness, conscientiousness, emotional stability, and openness (also referred to as intellect). The existence of a predominant model may eventually clarify research questions on the interplay

between language and personality, and significant advances in this area may be seen in the near future. So far, the most intensive attempts to explore the common ground between the two areas of research have been conducted by social psychologists and—to a lesser extent—second-language researchers. Research has been primarily focused on two issues. The first issue is the effects of linguistic differences on the attribution of personality features by listeners, whose evaluation is based solely on speech. These studies started with the seminal work of Wallace Lambert and his associates. They found that listeners in Montreal rated the personality of a given speaker differently when he spoke in two distinct languages (i.e. French and English) and also that French-speaking listeners tended to assign personality traits to the speaker quite differently from English-speaking listeners. Other speech evaluation studies of evaluations assigned to Spanish and English speakers in the United States followed. Howard Giles and his associates explored the effect of different regional and social speech varieties on personality ratings. Their research uncovered stereotypes linked to speech differences, and it was very valuable precisely in demonstrating that laypeople’s connections between language and personality are the result of previously existing stereotypes. In all cases, two factors were identified: a competence factor, which had to do with efficiency and high status in society; and a solidarity factor, which involved a high affective empathy with the speaker. A common result of most studies conducted so far has been that language varieties that enjoy high prestige and can be regarded as the varieties of the most powerful people tend to be not so well valued in terms of personal empathy, and vice versa. More recently, some attempts have also been made to connect personality ratings and nonnative proficiency, with provisional outcomes that also point at the important role that language variation may play in listeners’ perceptions of a speaker’s personality, with all the major subsequent consequences for interpersonal relations. This method of study is less subjective than self-reports or selfaccounts of personality traits, which tend to be heavily mediated by the idea of what one would like to be like rather than what one really is or seems to be. Another positive aspect of attributions of personality by listeners is the fact that they are closely related to social stereotyping. Thus, the study of the influence of

821

PERSONALITY AND LANGUAGE speech style on personality evaluation leads to the uncovering of social stereotypes. The second area of research that links language and personality is second language acquisition, particularly the role of different personality traits in the ultimate achievement of second language proficiency. Surveys conducted among second language teachers and students show that personality is considered to be a primary factor in second language achievement. Empirical studies have mainly focused on a single personality factor: extraversion. Extroverted people seem to be more successful in acquiring basic communicative abilities because of their willingness to interact with other people and search for communication opportunities. Introverts are less engaged in communicative interactions but may benefit from an increased attention to the formal aspects of language. On the whole, research has not found one type of learner to be inherently better than another, but personality profiles seem to affect the preferred path to learning and, therefore, the final outcome of the process. Personality and language should be viewed as two uniquely human characteristics that are bound to interact; however, at this point, there is a need for much more research to investigate the relationship between the two.

References Ellis, Rod. 1994. The study of second language acquisition. Oxford: Oxford University Press. Furnham, Adrian. 1990. Language and personality. Handbook of language and social psychology, ed. by Howard Giles and W. Peter Robinson. Chichester: John Wiley & sons. Giles, Howard, and Peter F. Powesland. 1975. Speech style and social evaluation. London: Academic Press. Goldberg, Lewis R. 1981. Language and individual differences: the search for universals in personality lexicons. Review of personality and social psychology, Vol. 2, ed. by L. W. Wheeler. Beverly Hills, CA: Sage. Lambert, Wallace E., R.C. Hodgson, R.C. Gardner, and S. Fillenbaum. 1960. Evaluational reactions to spoken languages. Journal of Abnormal and Social Psychology 60. 44–51. Llurda, Enric. 2000. Effects of intelligibility and speaking rate on judgements of non-native speakers’ personalities. IRAL — International Review of Applied Linguistics 38. 289–99. Ryan, Ellen B., and Miguel A. Carranza. 1975. Evaluative reactions of adolescents toward speakers of standard English and Mexican American accented English. Journal of Personality and Social Psychology 31. 855–63. Saucier, Gerard, and Lewis R. Goldberg. 1996.The language of personality: lexical perspectives on the five-factor model. The five-factor model of personality; Theoretical perspectives, ed. by Jerry S. Wiggins. New York: Guilford. Skehan, Peter. 1989. Individual differences in second-language learning. London: Edward Arnold. Taylor, Harold C. 1934. Social agreement on personality traits as judged from speech. Journal of Social Psychology 5.

ENRIC LLURDA See also Discourse Analysis

Philippine Spanish Creoles The Spanish-based Creoles in the Philippines developed after the Spanish conquest in the sixteenth century and are commonly known as Chabacano. Despite its despicable original meaning (‘tasteless’, ‘vulgar’), this name is used by the speech-community as a self-designation. Alongside Papiamento and Palenquero in the Americas, Chabacano is one of the three Spanishbased Creoles worldwide. It shares interesting features with other European-based Creoles in Asia, a fact differentiating it from the Atlantic creoles. Further, unlike the latter, Chabacano has been acquiring a mixed character; especially from the twentieth century onward. The total number of Chabacano speakers can be estimated at no more than 500,000. Geolinguistically, Chabacano can be divided into two main varieties: Manila Bay Creole and Southern Mindanao Creole. Dialects of the former are found on the northern island

822

of Luzon and are spoken by relatively small communities in the towns of Cavite and Ternate, but the dialect of Ermita, a district of Manila, can now be regarded as extinct. Southern Mindanao Creole is spoken on the island Mindanao in the south of the archipelago. The highest degree of vitality of Southern Mindanao Creole, and of Chabacano generally, can be observed in the city of Zamboanga and the surrounding area in the extreme western part of Mindanao (approximately 300,000 speakers). Creole is also used here as a lingua franca by diverse ethnolinguistic groups. It is codified for public and private purposes with a hispanic orthography; there are bible translations, literary writings, and until recently press. Public oral use has recently been increasing, there is broadcasting in the language, and it is used in church and partially in school. A subvariety of Zamboangueño is

PHILIPPINE SPANISH CREOLES also spoken in the town of Cotabato; however, little is known about the vitality of the subvariety of Southern Mindanao Creole found in Davao. The genesis and history of Chabacano, especially Zamboangueño, is somewhat complex due to different contact situations in time and space. Keith Whinnom (1956) suggested that Chabacano grew out of the Portuguese creole of the Indonesian island of Ternate, which had been transported to the Philippines in the seventeenth century. He considered Zamboangueño a layer of Manila Bay Creole. This theory was contested principally by John Lipski (1992), who affirmed a typical creolization of a Spanish-based pidgin with Tagalog influences for Manila Bay Creole. For the special case of Zamboangueño, he proposes a development of different stages from the eighteenth century onward: it arose in the garrison, absorbing elements from Manila Bay Creole later on. During the nineteenth century, Hiligaynon elements were introduced through contact from ships coming from Manila having made a stop in Central Philippines. Afterward, it was re-hispanicized by native speakers from Europe and Mexico who came to Zamboanga. It was also influenced by Cebuano, like Hiligaynon, a Visayan language, which was brought by large-scale immigration. From the 1930s onward, the influence of English and, after World War II, that of the national language Pilipino have been increasing, especially affecting the lexicon. Consequently, Zamboangueño does not fit into the classification of plantation, fort, and maritime Creoles; nor can the dichotomy of exogene and endogene Creoles adequately describe the situation of Zamboangueño. In contrast to the Atlantic creoles, slavery played a marginal role. Whatever the different theories on the origin of Chabacano, first documented in Hugo Schuchardt’s Kreolische Studien (1883; Creole studies), in the beginning there was a pidginized form of Spanish used for intercommunication by different ethnic groups, namely, European Spaniards, Mexicans, West-Austronesian speaking groups, and Chinese immigrants. A pidgin created by the latter is still spoken by shopkeepers in Davao (‘Bamboo Spanish’). The current structure of the language, especially concerning the vocabulary and word structure, is heavily influenced by Philippine languages. The sound system shows Philippine features like the merging of [o]/[u] and [e]/[i] in nonaccented syllables and the occurrence of the glottal stop. Some of the productive morphemes for word formation in Zamboangueño originally come from Spanish, and many are from Hiligaynon or Cebuano (e.g. the verbalizing prefix man-: kwénto ‘story’ → man-kwénto ‘to tell’ or the adjectivizing prefix ma-: pwérte ‘strength’ → mapwérte ‘strong’). Their form is often identical to those

of the source languages, but in many cases they have different functions. The word-formation system seems to be unusually complex for a Creole. The plural of nouns is marked with the particle maga (el maga péhro ‘the dogs’). In Zamboangueño, first-person plural pronouns have a Philippine form encoding inclusion or exclusion of the addressee. The basic word order of V(erb)-S(ubject)-O(bject) stems from Austronesian contact, e.g. man-J ángge yo ‘I’m going shopping’. This is exceptional compared to other European-based creoles that are typically SVO. As in Philippine languages, human proper names in subject function are marked with si (ta-kantá si Maria ‘Maria is singing’). Human and definite objects are obligatorily marked with kon, perhaps a merger of an Austronesian element and Spanish con ‘with’, (konosé-le kon ese muhér ‘he knows that women’). There are many discourse particles of Philippine origin (e.g. daw for reported information). However, Chabacano also has features typical of Atlantic creoles, e.g. serial verb constructions (pwéde asé salé ‘to be able to make to go out’), no syntactic passive construction, lack of an equative copula (soltéro el hénte ‘the man is a bachelor’), identity of ‘have’ and ‘there is’ (both tyéne), and preverbal markers for tense, mood, and aspect (e.g. ya-andá yo na Samboanga ‘I went to Zamboanga’, ta-kusiná le ‘(s)he is cooking’ and ay-kantá silá ‘they will sing’). The lexicon principally consists of Spanish vocabulary (83%), with some Philippine (c. 15%) and English (2.5%) words. It can be assumed that, more recently, the influence of English has increased and that of Spanish has diminished. Very few words are of Portuguese origin (e.g. gumitá  port. gomitar ‘to vomit’). The importance of Mexican input is shown by words like J angge ‘market’  tiangue (a Mexican word of Nahuatl origin). Despite the close geographical contact with Zamboanga, the local languages, Tausug, Subanon, and Yakan have nearly no influence on the lexicon.

References Fernández, Mauro (ed.) 2001. Shedding light on the Chabacano language [Special Issue of Estudios de Sociolingüística. Linguas, sociedades e culturas 2, 27–56]. Vigo (Spain): University Press. Forman, Michael L. 1972. Zamboangueño texts with grammatical analysis: a study of Philippine Creole Spanish. Ph.D. Dissertation. Cornell University. Frake, Charles O. 1971. Lexical origins and semantic structure in Philippine Creole Spanish. Pidginization and creolization of languages, ed. by Dell Hymes, 223–42. Cambridge: Cambridge University Press. Lipski, John M. 1992. New thoughts on the origins of Zamboangueño (PCS). Language Sciences 14. 197–231. Papia. Revista de Crioulos de Base Ibérica 12 (2002) [Special Issue dedicated to Chabacano].

823

PHILIPPINE SPANISH CREOLES Riego de Dios, María. 1989. A composite dictionary of Philippine Creole Spanish (PCS). Studies in Philippine Linguistics 7. 1–210. Schuchardt, Hugo. 1883. Kreolische Studien IV: über das Malaiospanische der Philippinen [Creole Studies IV: on the Malayo-Spanish of the Philippines]. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften zu Wien 105. 111–50. Steinkrüger, Patrick O. 2003. Morphological processes of word formation in Philippine Spanish Creole (Zamboangueño).

Phonology and morphology of Creole languages, ed. by Ingo Plag, 253–63. Tübingen: Niemeyer. Whinnom, Keith. 1956. Spanish contact vernaculars in the Philippine Islands. Hong Kong: Hong Kong University Press.

PATRICK O. STEINKRÜGER See also Austronesian, Papiamento, Pidgins and Creoles, Spanish and Iberoromance Languages, Tagalog and Philippine Languages

Philology Etymologically, philology is the love of words (or, perhaps, the study of love). In the broadest sense, it covers everything having to do with the study of language— grammar, texts, history, civilization. Indeed, until a century or so ago, what is now called linguistics was simply part of philology, and even today, the term ‘comparative philology’ can still be found as a synonym for ‘historical linguistics’, especially within Indo-European studies. But here we should look at those branches of philology that have not been co-opted by (descriptive, historical, or theoretical) linguistics. Thus narrowed, philology can be considered to be the study of texts, with all that implies: first, determining exactly what the author of a text actually wrote; then, determining what the author said; then, determining what the author meant. Correspondingly, a typical edition—whether of a book of the Bible, a play by Shakespeare, a novel by Joyce, or an inscription from an Indian temple—will contain the text itself, notes on every detail, and an interpretative commentary. Traditionally, the interpretative work of philology relies on the prior accomplishments of four subdisciplines: epigraphy, paleography, diplomatics, and textual criticism.

Epigraphy Epigraphy is the study of inscriptions on hard material, such as stone and metal. People have been carving important texts into walls for as long as they have been writing—in Egypt, Mesopotamia, China, Anatolia, Greece, Arabia, Rome, Iran, India, Inner Asia, and Mesoamerica, in rough chronological order—and often such inscriptions are the only surviving written evidence of past civilizations. Sometimes they are official decrees, and by reading between the lines we can learn much about the societies that prompted them. Sometimes, as is almost exclusively the case with 824

Etruscan, they are funerary commemorations, from which we learn only of family relationships. The task of the epigrapher is to record inscriptions accurately (and photography can rarely substitute for the sharp eye and steady hand of the draftsman), to publish them first promptly and then systematically, and to provide such tools as charts of letter forms that will assist in the interpretation of newly discovered inscriptions. A separate branch of epigraphy is numismatics, the study of coins and their inscriptions.

Paleography Paleography is the study of manuscripts, written with pen or brush and ink on flexible material, namely papyrus, skin, or paper; textiles, wood, or leaves. These are perishable materials, so only rarely have they survived from more than a few centuries ago. The exceptions are due to extremely dry climates, as near the Nile and Jordan Rivers, areas where thousands of Egyptian papyri and the Dead Sea Scrolls have been found. In extremely rare cases, careful curatorial attention has preserved precious documents for a thousand years or more, but our knowledge of ancient texts results far more from a continual tradition of copying than from ancient attestations. (The tradition was fairly quickly superseded by printing beginning c. 1450.) The paleographer’s job is primarily to study the development of handwriting over generations of copyists, using dated manuscripts to anchor changes in stylistic trends and exercising judgment in assigning undated ones to this or that range of dates—although it must be recalled that any individual scribe’s hand will not change much over their professional career, so a date based strictly on paleographical grounds cannot be more precise than within a generation or so. There are as many fields of paleography as there are scripts with manuscript traditions—cuneiform, Egyptian, Chinese,

PHILOLOGY Greek, Latin, Indian, Southeast Asian, etc. (there are no general surveys of world paleography outside histories of writing like Diringer (1968) and Jensen (1969); Urry (1974) considers only Greek and Latin).

Diplomatics Diplomatics is the study of the form of documents. Its principal concern is detecting forgeries, so it deals principally with the physical properties of manuscripts and is the aspect of philology farthest removed from linguistics.

Textual Criticism Textual criticism is the study of the history of the content of texts. Whenever a scribe makes a copy of a manuscript by hand, however carefully, mistakes are introduced. If a work was popular, many copies would be produced, and copies of copies, and so on, every one of them differing very slightly from all the others. By noting the correspondences among the differences in different copies, the chain of transmission can be reconstructed, and even if the author’s original manuscript has not survived, it may be possible to reconstruct the author’s original text with considerable certainty, with successive changes being removed. Different scribal traditions handle the problem of scribal error in different ways. The rabbis who codified the text of the Hebrew Bible, known as Masoretes, did things like counting the number of words and even letters in each book. Scribes in the Islamic tradition included elaborate COLOPHONS in their manuscripts, naming not only the scribe and the date and circumstances of copying (as is often found elsewhere) but also a full list of all intervening scholar/copyists back to the original author. And Chinese scribes sometimes bypassed the problem by making copies of major works by printing them off woodblocks or even stone inscriptions of the texts, thus coming full circle in the fields of philology. An interesting offshoot of textual criticism is that it is often possible to determine that a given text has been translated from another language, and that other language identified with considerable certainty. Thus, R.H. Charles was able to say of the Book of Enoch, which was known in full only in Ge‘ez, with fragments surviving in Greek and Latin, that portions—chapters 1–5 and 37–104—were originally written in Hebrew, and chapters 6–36 were originally written in Aramaic (1912:lvii–lxx). Decades later, fragments of the Aramaic original were identified among the Dead Sea Scrolls.

Decipherment The philological task most relevant to linguistics is the decipherment of disused scripts. Usually, these are brought to light by archeological excavations. Rarely,

the unfamiliar inscription is accompanied by an inscription in a familiar script and language, and it can be assumed that the texts are at least roughly equivalent. This state of affairs is known as a BILINGUAL, and both the first decipherment, of the Aramaic language Palmyrene by Jean-Jacques Barthélemy in 1754, and the best known, of Egyptian hieroglyphics by JeanFrançois Champollion in 1822, involved bilinguals, although these two accomplishments were vastly different in difficulties. Pope (1999) includes detailed descriptions. When no bilingual is available, the ingenuity of the philologist is required to discover a VIRTUAL BILINGUAL. The most challenging, and for its impact on our understanding of human history the most important, decipherment was of CUNEIFORM. Late-eighteenth-century travelers brought back to Europe drawings of inscriptions they found at the ruins of the capital of the Persian Empire, Persepolis, which was destroyed by Alexander in 330 BCE. Each inscription was in three different scripts, which were assumed to represent the same content in three languages (hence, they are called TRILINGUALS), but none of the scripts was familiar. The characters in all of them were composed of incised wedges (cuneus in Latin), whence the name. The first success in interpreting cuneiform came in 1802. A German high school teacher, Georg Friedrich Grotefend, surmised that because the simplest of the three scripts always appeared in the most prominent position among the three, it probably wrote the ancient Persian language of the Empire. A few years earlier, the Frenchman Antoine Isaac Sylvestre de Sacy had deciphered royal inscriptions of a successor Iranian empire, the Sassanian, finding that they usually began with a genealogy in the form ‘X, great king, son of Y, great king ….’ Grotefend knew the names of the kings—Darius, Xerxes, and so on—from the Greek Histories of Herodotus and sought similar formulas at Persepolis. He soon found the expected recurring patterns in the inscriptions, and although he himself was not an Iranian philologist, his breakthrough discovery made it possible for specialists like the Danes Christian Lassen and Rasmus Rask to determine the basics of Old Persian. Edward Hincks, a Church of Ireland (Anglican) clergyman, showed that each of the 36 characters of Old Persian script represented a Consonant–Vowel (CV) syllable. The second most complicated of the three cuneiform scripts had about 100 different characters (none of them the same as in Old Persian), so it was not surprising when, by comparing what seemed to be the same names in corresponding Persepolitan inscriptions, it too was shown to record syllables—both CV and VC. Its language, which is today called Elamite, is still not well understood. Hincks was one of the scholars who did that work; but more importantly, he went 825

PHILOLOGY on to decipher the third, most complicated script, which used several hundred characters. By the time Hincks got involved, in the mid-1840s, numerous cuneiform inscriptions had been brought from the heartland, Mesopotamia, and they too were in the third script, so Hincks had a great deal of material to compare. This time, the second script did constitute a subset of the third script, but the most useful aspect of the inscriptions was their repetitiveness: what seemed to be the same formulas appeared in different texts, or even within one text, but they contained small divergences. (Some of the best examples came from an extensive inscription in what we now call Urartian, which was published in 1840.) A sequence like A B C D E F G might substitute for A B H E F G, and Hincks saw that this could mean that C D stood for C1V1–V1C2, while H stood for C1V1C2. He also discovered patterns like iprus–iparras, a sure sign that the language was Semitic; we now call it Akkadian (and the words were verbs). He found, rather incredibly, that characters could have more than one pronunciation—and he suggested that this was because the pronunciations found their origin in more than one language, what we now call Sumerian alongside Akkadian. Most importantly, he realized that some of the characters were being used not for the sounds they represented, but to stand for meanings. These characters—which he initially referred to as ‘ethnological boulders,’ since they were like alien rocks standing in a plowed field—were remnants of Sumerian routinely used in Akkadian texts. Thus, between 1846 and 1852, Edward Hincks provided the key to the Akkadian, Urartian, and Sumerian languages. One often hears of H.C. Rawlinson in this connection, but he was never afterward able to explain how he deciphered cuneiform; it is now known that, in his post in Baghdad, he was kept informed of Hincks’s work and used it in his analysis of the great inscription on a cliff at Behistun, Iran, which he copied with great difficulty—but which became available too late to be useful in the decipherment. Another celebrated decipherment was accomplished by Michael Ventris in 1952, when he realized that the Linear B tablets from Crete and Mycaenae on the Greek mainland were actually in an archaic form of Greek, not some unknown predecessor language. In this and several other cases, the virtual bilinguals were known place names that seemed to occur in texts found in those places; another is the decipherment of the Maya glyphs.

The Uses of Philology It is self-evident that without philology to make texts available for study, history could not be written and

826

historical linguistics would be limited to deductions that can be made by comparing attested languages. But philology has applications that are not so obvious: investigating the use and distribution of languages themselves reveals patterns in the human career that do not emerge when only the contents of historical documents are considered. On at least one occasion a philologist has used his professional skills in creating a fantasy universe; Tom Shippey (2003), besides introducing the work of philology to the committed reader, shows how J.R.R. Tolkien’s legion of imitators generally fail for lack of the kind of profound background Tolkien created through his many invented languages—and with example after example reveals that Tolkien’s deep learning in Germanic philology, as well as Celtic and Finnic, underlies feature after feature of Middle Earth. References Charles, R. H. 1912. The book of Enoch …. Oxford: Clarendon. Daniels, Peter T. 1994. Edward Hincks’s decipherment of Mesopotamian cuneiform. The Edward Hincks bicentenary lectures, ed. by Kevin J. Cathcart. Dublin: University College, Department of Near Eastern Languages. Daniels, Peter T. 1995. The decipherment of ancient Near Eastern scripts. Civilizations of the ancient Near East, ed. by Jack M. Sasson et al. New York: Scribner’s. Daniels, Peter T. 1996. Methods of decipherment. The World’s writing systems, ed. by Peter T. Daniels and William Bright. New York: Oxford University Press. Diringer, David. 1968. The alphabet: a key to the history of mankind, 3rd edition, 2 vols. New York: Funk & Wagnall’s. Herde, Peter. 1974. Diplomatics. Encyclopædia Britannica, 15th edition, Macropædia; since 1985 included under ‘History, The study of.’ Jensen, Hans. 1969. Sign, symbol and script, 3rd edition, translated by George Unwin. London: George Allen & Unwin. Kenney, E. J. 1974. Textual criticism. Encyclopedia Britannica, 15th edition, Macropædia; since 1985 included under ‘History, The study of.’ Larsen, Mogens Trolle. 1996. The conquest of Assyria: excavations in an antique land 1840–1860. London: Routledge. Ostler, Nicholas. 2005. Empire of the word: a language history of the World. London: HarperCollins. Pope, Maurice. 1999. The story of decipherment: from Egyptian hieroglyphs to Maya script, 2nd edition. New York: Thames & Hudson. Puhvel, Jaan. 1974. Epigraphy. Encyclopædia Britannica, 15th edition, Macropædia; since 1985 included under ‘History, The study of.’ Robinson, Andrew. 2002. Lost languages:the enigma of the World’s undeciphered scripts. New York: McGraw-Hill. Shippey, Tom. 2003. The road to Middle Earth: how J. R. R. Tolkien created a new mythology. Boston: Houghton Mifflin; 1st edition, 1982. Urry, William G. 1974. Paleography. Encyclopædia Britannica, 15th edition, Macropædia; since 1985 included under ‘History, The study of.

PETER T. DANIELS See also Aramaic; Semitic Languages; Sumerian

PHILOSOPHY OF LANGUAGE

Philosophy of Language Rather like the scholastic philosophers of the late Middle Ages, twentieth-century philosophers have seen philosophy as linguistic analysis, as the attempt to discern the logical structure of reality through discerning the formal structures—the superficial or deep grammar—of the language in which we report or think reality. Philosophers are particularly interested in certainty, in necessary truths as opposed to mere chance events. Nineteenth-century formalist mathematicians, similarly, honed a sense of mathematics as syntactically grounded in formal language or in a consistent set of linguistic conventions. Since philosophers conceived of their enterprise as a search for helpful logical truths and not as dependent on any experiential truths (philosophers do not do experiments), twentieth century philosophy became logicolinguistic analysis, and hence philosophy of language became, in short and for much of the past century, philosophy. Linguistic philosophy was not just seeing the features of language that revealed the world’s categorical structure but also seeing beyond language’s perhaps misleading surface features. To give a standard example, the verb to be in English plays three vastly different logical roles. Is can mean identity, as in 2 is (the same as) 2, Dubya is (the same as) President George W. Bush, or heat is (the same as) the average motion of molecular particles. Is, however, can also mean predication, as in The sky is (has the property of being) blue, Roses are (have the property of being) red, or The earth is (has the property of being) spherical. Finally, Is can mean existence, as in There is (exists) a Santa Claus or God is (exists) as opposed to God is (exists) not. Bertrand Russell (1919), while imprisoned as an antiwar demonstrator in World War I, wrote in Introduction to mathematical philosophy that it was important to keep these three senses wholly separate in linguistic analysis (so important that Russell added that he would declaim so even if he were ‘dead from the waist down and not merely in prison’). In the mathematical logical notation that Gottlob Frege and Russell created in the late nineteenth century, these notions were indeed wholly separate and specifically represented. With a notation that puts predicates in capitals and writes names and variables for individuals in lower case, ‘a  b’ means two names stand for the same individual, ‘Pa’ means predicating P of individual a, and ‘∃x[Dx]’ means, if D predicates divine, that there exists a God. Russell stressed that this notation commendably dissolved the

‘ontological argument for the existence of God’, which mostly simply runs, God is that being who possesses all possible perfections; existence is a perfection; therefore, God exists. This argument confounds the hypothetical predications ‘If something is Divine, then it has various Perfections’ and the existential ‘Something exists that is Divine’. Russell was particularly praised for his ‘paradigmatic’ analysis of ‘definite descriptions’, phrases of the form The so and so. Consider ‘The present King of France is bald.’ If you think this statement has a subject/predicate logical form, and believe that statements are either true or false, then you seem to have to say that either ‘The present King of France is bald’ is true or ‘The present King of France is not bald’ is true. The problem, of course, is that there is no present King of France (when Russell wrote and subsequently). It is no solution to say ‘The present King of France’ means nothing, because then ‘The present King of the United States’ and ‘The round square’ and countless other phrases would presumably also mean, or stand for, nothing; but it is evident that all such phrases differ in meaning. If a revolution occurred in the United States in 2084, there might then be a King of the United States, but that would not mean there would be a round square and a French King as well (or that nothing had changed to something). Russell insists that ‘The present King of France’ does not mean anything all by itself. Rather, ‘The present King of France is bald’ means ‘There exists an x that bears the predicate of being kingly of France, and if any y also bears that predicate, y  x, and x is bald.’ Following Russell’s lead, in the 1920s, the logical positivists of the Vienna Circle insisted that logicomathematical truths provide the formal structure within which the observational truths of experience array themselves. Theoretical terms, such as ‘vital spirit’, ‘electron’, or ‘virus’, are only acceptable if they can be specified completely in observational terms. The Vienna Circle philosopher W.V.O. Quine (1960) of Harvard University claims that by casting science into Russell’s austere predicate logic, you can most clearly determine what science says has to be real. Indeed, such a translation into predicate logic will strip off the possibly misleading superficial features of actual languages. Philosophical linguistic analysis, conceived in this fashion, cares nothing for the phonology of language, for literal physical sound streams and their transformation into the sharper, leaner, and deeper 827

PHILOSOPHY OF LANGUAGE structures that fluent speakers hear. Moreover, the exclusive concern is with the truth or falsity of sentences that describe the world, collectively feeding what scientific generalizations we can muster about the world. As Quine put it, startlingly, all of science is held up as one sentence to nature—but that is just to insist that we collectively assert and understand this scientific sentential consensus, not that anyone actually says this one sentence. Philosophy of language, as so understood, has little concern with how natural language is used in the acts and interactions of everyday life. Quine makes this clear when, at the beginning of Word and Object, he approvingly cites Otto Neurath’s metaphorical remark that we are like seamen who must make repairs in our boat while at sea. Here, ‘we’ means all scientific-minded humans, and ‘the boat’ means whatever natural and artificial symbol systems we use in our collective scientific description of the world, with Quine highlighting the austere notation of mathematical logic. By the mid-twentieth century, however, philosophers began to shift from a concern with the true/false logical relationship between sentences and the world to a more expansive, but also more narrowly linguistic, concern with sentences as speech actions of speakers, who carry individual responsibility to us for what they say, indeed for what they do in saying so. In the 1950s, the Oxford University philosopher J.L. Austin, among others, drew attention to performatives, sentences that, when uttered by the appropriate person in the appropriate circumstances, do something. If I say, ‘I promise to return your $20 tomorrow’, I am not describing, truly or falsely, some peculiar mental state; rather, by my speaking, I make a promise. If I say, ‘By the power vested in me by the State of New Jersey, I appoint you Port Commissioner,’ I make an appointment. ‘I promise X’ and ‘I appoint Z’ are explicit performatives, where the main verb of the sentence explicitly indicates what action is performed. However, ‘I certainly will return your $20 tomorrow’ will, given appropriate circumstances, constitute a promise. Similarly, if an officer says to his subordinate, ‘You will move your men to that bridge,’ he has given an order. Indeed, Austin maintained that every use of language, or every speech act, has a performative aspect. For example, if I say ‘There’s a bittern in your garden,’ I should know a little something about bitterns and have had some opportunity to identify the bird. If I have no idea of what bitterns look or sound like, and indeed, have not been in or near your garden, I have no right to say what I did. Given that I meet those minimal requirements, I may demur if you ask, ‘How do you know it’s a bittern?’ pleading perhaps ‘Well, I don’t know it’s a bittern but it is a large, white-feathered marsh bird’, or I may take a further plunge and say, ‘Oh, I know it’s a

828

bittern—I got a clear view and, growing up in the Fens, I’d know that booming anywhere’. Philosophical concern with semantic, performative, and pragmatic aspects of language has meant some fruitful interaction with linguistic science, particularly given the concern with syntax and logical form stressed by Noam Chomsky and other generative linguists since the 1960s. But still other philosophers have come to feel that philosophy is well rid of an exclusive emphasis on philosophy as linguistic analysis. In the middle decades of the twentieth century, philosophers concerned with values emphasized metaethics or the ‘logic of the language of morals’; more recently, philosophers have addressed specific normative issues. Similarly, many recent philosophers have constructed rational choice, and social contract theories. Thinkers as diverse as the logician Saul Kripke and the linguist Noam Chomsky have argued that necessary truths are more central to science than the real but often trivial analytic truths of language. Heat is the average motion of molecular particles is not, Kripke argues, an analytic or lexical truth, but rather a scientifically discovered necessary physical truth. In the same vein, water is H2O asserts a physical necessity, given the basic combinatory properties that form the fabric of atomic physics. Similarly, perhaps, the linguist’s claim that natural languages are generative and transformational follows from the structural character of the human linguistic faculty as a matter of biological necessity (it certainly does not follow from a lexical entry attributed to normal speakers, most of whom do not have ‘generative and transformational’ in their everyday vocabulary— or ‘average motion of molecular particles’, for that matter). In any case, few philosophers today would be wholly content to characterize philosophy as linguistic analysis. References Austin, J. L. 1961. Philosophical papers. Oxford: Clarendon Press. Ayer, A. J. 1936. Language, truth, and logic. London: Gollancz. Chomsky, Noam. 1986. Knowledge of language. New York: Praeger. Kripke, Saul. 1972. Naming and necessity, in Semantics of natural language, ed. by Gilbert Harman and Donald Davidson, Dordrecht, Holland: D. Reidel. Quine, W. V. O. 1960. Word and object. Cambridge, MA: MIT Press. Russell, Bertrand. 1919. Introduction to mathematical philosophy. London: Allen & Unwin.

JUSTIN LEIBER See also Austin, John Langshaw; Chomsky, Noam; Quine, Willard van Orman; Russell, Bertrand

PHONEME

Phoneme The phoneme [from Greek phonema—a sound] is recognized as the smallest abstract sound unit of a language, for example, map consists of three phonemes /m/, /æ/, and /p/. Each word in a given language must consist of a permitted sequence of phonemes. Phonemes may be defined as the class of sounds of a language that distinguish words from one another. The phoneme is a fundamental unit of phonology— the study of the sound systems of languages and the relationships between those sounds. Different languages have various numbers of phonemes; for instance, one Brazilian language—Piraha—is said to have only ten phonemes, while one of the African languages—!Xu—is known to have 141 phonemes. The term phoneme was introduced in the late 1870s, and the phoneme theory was first developed by Jan Baudouin de Courtenay and his pupils as a response to the need to systematize the sound patterns of the language. Since then, the phoneme and its features have been defined in various ways by different linguists. As Baudouin understood it, the term originally meant a ‘mental image’ of a real physical sound, and speech was the act of producing sounds as close to that mental image as possible. However, during the twentieth century, linguists came to a different view. Lev Scerba, who had initially followed Baudouin, shifted to the functional aspect of the phoneme, that is differentiating words. Daniel Jones also described a phoneme as a physical concept. If Nikolai Trubetskoy defined the phoneme as the smallest distinct unit and incapable of further subdivision, Roman Jakobson saw it as the sum of features that distinguish the phonemes from each other.

Contrastive Function It is traditionally accepted that the phoneme functions as the minimal contrastive unit of speech. In other words, the phoneme differentiates the meaning of words. The native English speaker easily discerns the difference in meaning between the words in minimal pairs, i.e. word pairs that differ in only one sound element: man and pan, where [m] and [p] differentiate between a human and a kitchen utensil, thus indicating that they possess different phonemes; bag and big, where [æ] and [I] make possible the distinction between an object and a size; den and then, where [d] and [ð] clearly indicate that one word means a ‘lair’ and the other one means ‘at that time’; or man and men, where [æ] and [e] identify one man as opposed to two or more of them.

Phonetic Context Every uttered word is a particular combination of sounds and is called the phonetic context of a sound. The English phonetician Daniel Jones gives the following definition of the phonetic context: …phonetic context of a sound is to be understood to mean the sounds next to it or near it in the sequence of which it is a part, together with its duration (length), stress and (if voiced) voice-pitch.

Minimal pairs provide the same phonetic context with only one different phoneme, as in bat and cat; where the sounds [b] and [k] occur in the same phonetic context, they differentiate the meaning of the words, and, therefore, are two different phonemes.

Allophones A phoneme may have several variants or allophones. As opposed to phonemes, allophones do not distinguish words, and they never occur in the same phonetic context. For example, cat and king possess a common sound [k]. However, the sound [k] is pronounced slightly differently: in the first instance, it is articulated—or pronounced—further back in the throat. In the second instance, the tongue—in an attempt to combine the sounds [k] and [i]—moves upward and forward producing a palatalized [k]. These two sounds vary only in the way they are articulated (pronounced); they cannot occur in the same phonetic environment, i.e. the palatalized [k] cannot occur before [æ] as in cat for the simple reason that our tongues are physically unable to make these two sounds together. The two sounds vary slightly in articulation; however, the difference is not significant, and may not even be picked up by a native English speaker. Thus, these two variants of /k/ are not distinctive in English; they are not different phonemes, but rather allophones of the same phoneme /k/. Daniel Jones gives the following examples of the allophones of the English phoneme /t/: The principal [t] is used before vowels in the strongly stressed position as in table or ten. A dental [t] is used before [th] as in eighth. A retracted [t] is used in [tr] as in train. An unaspirated or barely aspirated [t] is used in weakly stressed positions as in letter. A laterally exploded [t] is used in [tl] as in kettle. A nasally exploded [t] is used in [tn] as in mutton.

829

PHONEME For a native language speaker of a given language, it is not really important how carefully allophones are pronounced to understand what is being said. However, it is more important that phonemes are pronounced properly in order for the listener to get the correct message; instead of My brother bought you a big present, the speaker must not relay My brother bought you a pig present. Languages differ with respect to which sounds are perceived as distinctive (phonemic) and which sounds are mere variants or allophones of phonemes. English distinguishes between [l] and [r] as in lay and ray, which make them different English phonemes. Korean does not see any distinction between these two sounds, i.e. they are perceived by Koreans as allophones of one phoneme. Moreover, although English distinguishes between [d] and [ð] as in than, Malay does not distinguish them as two phonemes but as allophones of one phoneme. Of course, many languages make other distinctions that an English native speaker may not be able to hear.

Distribution Each language imposes limitations on the distribution of phonemes, i.e. each particular phoneme can occur only in certain phonetic contexts in that language. In English, there are no words that can start with the nasal [N] sound. There are no such words as ngull, ngreed, ngoose. In Chinese or Vietnamese, however, similar sound constellations are quite possible (e.g. names such as Ng or Nguyen). The distribution pattern of the phonemes in a language can be identified by analyzing which phonemes may occur in the following positions: syllable-initial word-initial

syllable-medial word-medial

syllable-final word-final

The English phoneme /p/ can be found in all of the above-mentioned positions, for example: post packet

apt upper

map ketchup

Other English phonemes can occur only in certain positions, for example, /œ/ as in rat, stamp, sachet, ant can only be found in syllable- or word-initial positions and syllable- or word-medial positions.

Phonological Opposition The phonological opposition, developed by the Prague School, illustrates the network of connections and relations between the phonemes. For example, proportional opposition features two or more pairs of phonemes, such as /p/ vs. /b/, /k/ vs. /g/, which are opposed to each other on the basis of voiced features, while /r/ and /l/ are in isolated opposition and are 830

opposed to each other on features that do not create a phonemic opposition in English. Bilateral opposition represents only pairs of phonemes, such as voiced–voiceless pairs /z/ - /s/, where the third element is not possible. Multilateral opposition, on the other hand, may have three or more elements, such as /p/, /t/, and /k/, which are all stop consonants, or /f/, /ð/, /θ/, /s/, which are all fricative consonants. Privative opposition can be illustrated by voiced–voiceless pairs, in which one member of each pair lacks a voiced feature (/d/-/t/, /b/-/p/). Equipollent opposition is where voiceless /p/, /t/, /k/ each have a different place of articulation. Gradual opposition involves gradation of a feature in the phonemes, for example, in vowels /i/, /e/, /a/, the difference is in aperture, or close–open continuum of tongue height. In many languages, the phonological opposition can be neutralized. For example, in English, when /s/ occurs in the word-final position, it is pronounced as [s] after a voiceless plosive: lips, rats, sacks, and as [z] after the voiced plosive: dogs, bags, ribs. In the Russian language, a voiced consonant can never occur at the end of the word. When the voiced consonant is found at the word-final position, it must be pronounced as voiceless. Neutralization in Russian can be illustrated by the following pair: pok spelled as [rok] (type of music), and pog spelled as [rog] (a horn); both are pronounced as [rok]. In the latter instance, the voiced [g] sound is neutralized and is pronounced as the voiceless [k].

Major Early Works on the Phoneme: Baudouin de Courtenay, Lev Scerba, Ferdinand de Saussure, Daniel Jones This section is a brief excursion into the major contributions of world linguists in relation to the notion of the phoneme only, and not phonology, which is beyond the scope of this paper. The term phoneme was introduced in 1875 to indicate the basic unit of human speech. Although it is said that the term was first used by French linguist Dufriche-Desgenettes in 1873, it was Jan Baudouin de Courtenay and Mikolaj Kruszewsky of the University of Kazan (Russia) who, between 1875 and 1895, conducted an extensive study of sounds, and established the phoneme theory. The Kazan School focused its research on the phoneme as ‘the psychological equivalent of the speech sound’. The phoneme, according to Baudouin, was impressed in the speaker’s mind. In the speech process, the speaker was aiming at uttering a sound similar to the psychological impression or the mental image; however, the pronunciation was always determined by the real phonetic context. The work of Baudouin influenced the Russian linguist Lev Scerba (1880–1944), who attended Baudouin’s

PHONETIC TRANSCRIPTION lectures in St. Petersburg. Scerba accepted the psychological definition of the phoneme, but later concentrated on the functional aspect of the phoneme, i.e. the differentiation of words. Scerba also introduced such phonological notions as complimentary distribution, which is a characteristic of a sound that never occurs in the same phonetic context. This can be illustrated by allophones of a phoneme, all of which occur in different phonetic contexts but are variants of the same phoneme. Since the allophones are different realizations of the same phoneme, it can be said that they are in complementary distribution to each other. Ferdinand de Saussure (1857–1913), a Swiss linguist, is considered the founder of modern linguistics. His main conception assumed that language is a system of mutually defining entities. He believed that language was a structured system. This view had a major impact on the later development of structuralism in linguistics, which studies the language as a system and not the individual elements of it. Structuralism was further developed in Europe by Trubetskoy and Jakobson, and in America by Sapir and Bloomfield. Daniel Jones (1881–1967), who was a lecturer of phonetics at University College, London, favored the idea of the phoneme as a practical tool for teaching purposes. His contribution to linguistics was in developing a more practical approach to phonology. He found phonology very useful for acquiring good pronunciation by people learning a foreign language, as a basis of nonphonetic branches of linguistics such as morphology and grammar, and as an important tool for creating alphabets as simply as possible. The Prague Linguistic Circle, which later came to be known as the Prague School, began its meetings in 1926. It consisted mainly of Russian and Czech linguists. Among them were the Russian linguists Nikolai Trubetskoy and Roman Jakobson. The contribution of Trubetskoy lies in the developing of the notion of phonological opposition, and relations between the members of an opposition. Jakobson’s view of the phoneme as a sum of distinctive features led to the creation of a new approach to phonological description. This approach of distinctive features of

the phoneme was based on the structuralist ideas of de Saussure. For example, a sound can be a vowel (a, i, o, e, u) or a consonant (b, d, g, z). If the sound is a consonant, it can be voiced (b, g, d) or voiceless (p, k, t), nasal (m, n) or nonnasal (all others). Vowel, voiced, and nasal are all examples of distinctive features. American structuralism was developed by American linguists Edward Sapir and Leonard Bloomfield. Sapir adhered to the idea that a native speaker knows the phonemic system of their language through intuition. For Sapir, the phonemic system was a mental reality that existed independently of the act of speech. Bloomfield saw the phoneme as reality, a bundle of distinctive features, but not as the intuition of a speaker. He believed that phonemic analysis was the process of isolation of its distinctive features through minimal pairs. According to Bloomfield, finding a distribution of redundant, nondistinctive features was also part of the phonemic analysis. References Abercrombie, David. 1965. What is a letter? Studies in phonetics and linguistics. London: Oxford University Press. ———. 1967. Elements of general phonetics. Edinburg: University Press. Bloomfield, Leonard. 1933. Language. New York: Holt. Hawkins, Peter. 1984. Introducing phonology. London: Routledge. Jakobson, Roman. 1962. Selected writings, Vol.1. The Hague: Mouton. Jones, Daniel. 1962. The phoneme: its nature and use, 30. Cambridge: W. Heffer & Sons Ltd. Kramsky, J. 1974. The phoneme. Munich: Wilhelm Fink Verlag. Trubetzkoy, Nikolai Sergeevich. 1958. Grundzuge der Phonogie, Gottingen: Vandenhoeck & Ruprecht; third edition, 1962, as Principles of phonology, translated by C.A.M. Baltaxe, Berkeley: University of California Press, 1969. Yallop, Colin. 1995. English phonology. Sydney: Macquarie University.

ALFIA ABAZOVA See also Baudouin de Courtenay, Jan Ignacy NiecisBaw; Bloomfield, Leonard; Feature Theory; Sapir, Edward; Saussure, Ferdinand de; Structuralism; Trubetzkoy, Nikolai Sergeyevich

Phonetic Transcription In speech, different letter(s) may represent the same sound (e.g. see, grief, key) or the same letter may represent different sounds (e.g. bad, many, want). In order to be able to express speech sounds of each

language known in the world in an unambiguous way, the International Phonetic Association developed a set of symbols known as The International Phonetic Alphabet (IPA). The use of sequences of phonetic 831

PHONETIC TRANSCRIPTION symbols to represent speech is known as phonetic transcription. The concept of a phonetic alphabet is to have one symbol for one sound. Over the years, different alphabets with modified IPA symbols have been developed for specific aims (as many symbols are not used), but the IPA is considered the standard.

In 1886, the International Phonetic Association was founded in Paris by language teachers who wanted phonetic notation to be used in schools, as a method of acquiring a realistic pronunciation of foreign languages. In 1897, it was named L’Association Phonétique des Professeurs de Langues Vivantes or the International Phonetic Association in English. The IPA is used, among other things, to indicate pronunciation in a dictionary, to record a language, and to transcribe speech for automatic speech recognition. The IPA is based on the Roman alphabet, but other symbols have been included to cover the wide variety of sounds found in the languages of the world. Phonetic transcription is placed between square brackets [ ] and there is no capitalization or punctuation. An overview of the different characters and symbols of the IPA is given in various charts elsewhere. The symbols will be discussed in more detail after highlighting some of the principles underlying the phonetic alphabet. Phonetic transcription does not only vary from language to language but also from speaker to speaker. Depending on dialect, fluency, etc., the same word can be transcribed differently. Speech segments can be divided into two major categories: consonants and vowels. They are described with reference to how they are produced and their auditory characteristics. It requires considerable skill to transcribe an unknown language. During speech production, the vocal tract varies continuously and the division between subsequent speech sounds may not be clear. Only linguistically relevant speech sounds are transcribed, not personal voice quality. A separate sign exists for each distinctive sound (e.g. /k/ for cat, track, kite, quick, monarch). The IPA also provides symbols for suprasegmental information (word boundaries, stress, intonation) and for refining the pronunciation of an utterance (e.g. aspiration).

can be distinguished with regard to voicing, manner, and place of articulation. Voicing refers to the presence or absence of vocal fold vibration (/b/ versus /p/, /d/ vs. /t/). Manner of articulation refers to the way in which speech sounds are produced (stop, fricative, nasal, trills, etc.), and place of articulation indicates the place of constriction in the vocal tract (lips, alveolar ridge, velum, pharynx, etc.). Some languages in the world contain nonpulmonic sounds, sounds that are not produced through air from the lungs: clicks and voiced implosives are indicated by special symbols, while ejectives are indicated by a plosive and an apostrophe. From an articulatory point of view, vowels can be classified according to the features ‘open’, ‘close’, ‘front’, and ‘back’. A vowel is considered ‘open’ when the space between the tongue and the roof of the mouth is large and ‘closed’ when the tongue is near the roof of the mouth. In front vowels, the tongue is fronted and raised toward the alveolar ridge; in back vowels, the tongue is near the back of the mouth (palate). Although the vowel space is continuous, most vowels are placed in relation to one of the eight cardinal (reference) vowels: /∑i, e, , a, , , o, u/. Several symbols are foreseen to designate suprasegmental information, i.e. intonation groups (), duration (), stress (), and to distinguish between words in tone languages (e.g. Chinese, Thai). Some of these symbols are iconic in the sense that the shape of the line indicates the height and possible movement of pitch. For example, [maÜ] with falling intonation means ‘to scold’ and [ma!] with fixed high intonation means ‘mother’ in standard Chinese. However, it is also possible to transcribe pitch height by adding diacritics to existing segments, e.g. [ó bà] meaning ‘it perched’ vs. [ó ba] ‘he or she hid’ in the West African language Yoruba. In this transcription, the symbols are not iconic, i.e. the accent does not indicate a rising or falling pitch, it means ‘high’ ‘mid’, or ‘low’. Another set of marks, called diacritics, has been included to further refine the transcription of phonation. This has been done to restrict the total number of characters. Diacritics are used to indicate aspiration, e.g. [p ha] ‘pie’, to indicate vowel centralization [ë] relative to the cardinal vowels, breathy and/or creaky voices, nasalization [fε˜] ‘fin’, meaning ‘end’ in French, and many other features.

IPA charts

Broad and Narrow Phonetic Transcription

The IPA charts give an overview of the different vowels, consonants, clicks, and other sounds occurring in the languages of the world. The charts are more than a list of symbols; they show how the different types of sounds can be classified. Consonants

Speech sounds can be transcribed ‘broadly’ or ‘narrowly’. A broad transcription is a phonemic transcription, while a narrow transcription captures more phonetic details of the speech sounds. For instance, the word ‘toast’ can be transcribed phonemically as

History

832

PHONETICS [tst] or phonetically as [th st]. Both types of transcription require some knowledge of the phonology of the language to solve ambiguities related to the segmentation of utterances. And even when the likeliness of occurrence of certain (combinations of) phonemes is known, the alignment between the physical speech sounds and the phonetic transcription can still be problematic. Phonetic transcription has proved to be a very useful skill, but it cannot be done automatically. A proper analysis of the speech signals requires considerable knowledge about the segmental and suprasegmental aspects of a language.

References Albright, R.W. 1958. The International Phonetic Alphabet: its background and development. International Journal of American Linguistics 24(1). 3. Handbook of the International Phonetic Association. 1999. A guide to the use of the International Phonetic Alphabet, 204pp. Cambridge: Cambridge University Press. Ladefoged, P. 1982. A course in phonetics, 2nd edition. New York: Harcourt Brace and Jovanovich. Ladefoged, P., and Maddieson, I. 1996. The sounds of the world’s languages, 426pp. Oxford: Blackwell Publishers. Roach, P. 1991. English phonetics and phonology, 262pp. Cambridge: Cambridge University Press.

ASTRID VAN WIERINGEN

Phonetics Phonetics is the scientific study of speech. It is concerned with all aspects of the production, acoustics, and perception of speech in the languages and dialects of the world. The starting point for almost any phonetic investigation is the identification of certain landmarks from a cross-sectional view of the vocal tract from which both consonants and vowels of the languages of the world can be classified (see the section on Basic Classificatory Principles). The classificatory system can be used to describe both the various types of speech sounds that occur in different languages and their dialects, as well as those aspects of pronunciation that are used contrastively, that is, to signal differences of meaning (see the section on Phoneme and Phonetic Variation). Phoneticians also analyze how languages make use of pitch, duration, and loudness to communicate meaning distinctions at both the level of the word and the utterance (see the section on Prosody). Phonetics is concerned with the way in which spoken communication is accomplished between speakers and hearers. This includes both the relationship between speech sounds and the neurological, physiological, and neuromuscular aspects of speech (see the section on Speech Motor Control and Physiology), as well as the resulting acoustic signal that is decoded by the listener (see the section on Speech Acoustics and Speech Perception). Phonetics is a highly interdisciplinary field and is informed by theory and methodology from Linguistics, Cognitive Science, Computer Science, Electronics, Signal Processing, Acoustics, Neurology, Anatomy, and Physiology. Within Linguistics,

Phonetics overlaps most directly with Phonology, Psycholinguistics, Sociolinguistics, and Dialectology and also with Syntax, Semantics, and Pragmatics in modeling the prosodic aspects of speech. Phonetics is relevant to disorders of speech communication and to speech pathology. In the last quarter of the twentieth century, phonetics has contributed in various ways to the field of speech technology, which is concerned with the development of computer systems for the automatic generation (synthesis) and recognition of speech.

Basic Classificatory Principles The range of possible sounds that can occur in the world’s languages and from which a language makes a selection can be described in terms of various interactions between the vocal organs (Figure 1). The great majority of speech sounds are produced when air is expelled from the lungs, passing through the larynx between two shelves of muscular tissue known as the vocal folds (Figure 2). When the vocal folds are apart, as they are in quiet breathing, they do not obstruct the passage of air from the lungs. Many speech sounds are produced with open vocal folds and they are known as voiceless. In many other speech sounds, including all vowels, the vocal folds can be drawn together and made to vibrate very rapidly, that is, they repeatedly alternate between being open and closed—such sounds are voiced. There is a direct relationship between the rate at which the vocal folds vibrate and the sound’s pitch. In most languages, there are pairs of

833

PHONETICS

A 5 4 1

6 B c

3

b

7 d

a 2

e

C

D E

Figure 1. A cross-sectional view of the vocal tract of an adult male taken from an X-ray image (adapted from Laver, J, 1994). Upper lip (1), lower lip (2), upper front teeth (3), alveolar ridge (4), hard palate (5), velum or soft-palate (6), uvula (7), tongue tip (a), tongue blade (b), tongue front (c), tongue back (d), tongue root (e), nasal cavity (A), oral cavity (B), pharyngeal cavity (C), larynx (D), and trachea (E). The places of articulation are obtained from the following combinations: bilabial (12), labiodental (23), dental (a3), alveolar (a4), apical postalveolar (a5), laminal postalveolar (b5), palatal (c5), velar (d6), uvular (d7) and pharyngeal (eC).

vocal fold

vocal fold

Figure 2. A photograph of the vocal folds taken from above (from Ladefoged, 2001b). The space between the vocal folds is called the glottis. Here the vocal folds are apart allowing the air to pass from the lungs through the trachea into the rest of the vocal tract.

sounds that differ only in whether or not the vocal folds vibrate, such as [f] (in ‘fan’, voiceless) and [v] (in ‘van’, voiced). Voiced sounds are usually accompanied by vibrations that can be felt by placing a hand at the level of the throat and then switching voicing on and off as in a repeated production of [fffvvvfffvvvffvvv]. The air from the lungs can be made to pass out of the vocal tract either through the mouth (the oral cavity) or the nose (the nasal cavity) depending on the

834

Figure 3. The configuration of the vocal organs for the production of [n]. The direction of the air from the lungs is shown by the lines and arrows. The velum is lowered allowing air to enter the nasal cavity and pass out through the nose.

actions of the velum or soft palate (Figure 3). When the velum is raised, the air passes out of the mouth or oral cavity and sounds produced in this way are said to be oral. When the velum is lowered, the air enters and passes through the nasal cavity and exits through the nose: such sounds are nasal. Many languages have pairs of sounds that are distinguished only by whether or not the velum is raised. For example, in English [b] in ‘bad’ and [m] in ‘mad’ are produced in the same way, except that [b] is oral and produced with a raised velum while [m] is nasal and produced with a lowered velum. Within the oral cavity, there are three other important landmarks: the upper front teeth, and behind them the alveolar ridge, which extends in an arch into the hard-palate (Figure 1). The tongue is a highly flexible and mobile vocal organ that has a very dense concentration of different muscles. The tongue is subdivided into the tip, blade, front, and back; in the rest position, these four landmarks lie roughly below the alveolar ridge, the back of the alveolar ridge, the hard palate, and velum, respectively. The tongue is attached to the epiglottis just above the vocal folds and includes a near-vertical section in the pharynx, the root. The upper and lower lips are involved in various ways in speech production: they may be protruded as in the vowel of ‘soon’ or completely closed as in [b] and [m] in ‘bad’ and ‘mad’. In producing any consonant, there is a point in the vocal tract at which there is the greatest degree of narrowing and therefore the greatest obstruction to the

PHONETICS flow of air: its location is the consonant’s place of articulation and it can vary anywhere between the lips and vocal folds (Figure 1). Three commonly occurring places of articulation are: bilabial (e.g. [p], [b], [m] in ‘pan’, ‘ban’, ‘man’), alveolar (e.g. [t], [d], [n], [l] in ‘tie’, ‘die’, ‘nine’, ‘lie), and velar ([k], ["], [ŋ] as in ‘cot’, got’, ‘sing’). For any place of articulation, there can be variations in stricture, which defines the extent to which the flow of air out of the mouth is obstructed. When the obstruction is complete, as in [b d " p t k m n ŋ] in the above examples, the consonant is phonetically a stop. A fricative is produced when the air is forced through a very narrow opening at high speed, often producing a hissing sound due to the air becoming turbulent: examples are [s] (‘sip’) and [ʃ] (‘ship’). If the obstruction is further reduced, the flow of air is no longer turbulent but laminar. A consonant produced in this way is an approximant; examples in English include [w] (‘we’), [j] (‘you’), [l] (‘led’) and in some English dialects [ɹ] (‘red’). [l] in ‘led’ is also a lateral, in which the air flows over the side of the tongue, and is distinguished from central sounds in which the air flows over the center of the tongue. The two principal features for describing vowel production are height and backness. Height is analogous to stricture and defines the degree of opening in the mouth. The smallest opening, which is never less than for approximants, occurs in high or close vowels as in ‘heed’ while the greatest opening is in low or open vowels like ‘had’. Backness is analogous to the place of articulation defined earlier: vowels can vary between front, in which the narrowing occurs at the hard palate, and back in which the greatest point of narrowing is often in the pharynx. Height and backness together form a two-dimensional space that defines the ranges within which vowel production (Figure 4) occurs. The edges of the vowel space are marked by the cardinal vowels, defined by the phonetician Daniel Jones in 1918 at the University

College of London. The Height  Backness space is used to define the phonetic differences between vowels of different languages and dialects. There is independent experimental evidence to show that the extent to which listeners judge two vowels to be similar depends to a large extent on the distance between them in this space. Various acoustic analyses in the last 40 years have shown that height and backness are correlated, respectively, with the first two resonances or formant frequencies of the vocal tract. Vowels can also vary in whether they are rounded (produced with protruded lips) or unrounded (the lips are unprotruded). There is a preference across languages for front vowels to be unrounded (e.g. ‘heed’, ‘head’, ‘had’ have unrounded front vowels) and back vowels to be rounded (e.g. ‘hoard’, ‘who’ have rounded back vowels). But there are many languages with front rounded (French, German) and back unrounded vowels (Japanese, Vietnamese) and in some languages, the presence or absence of lip-rounding may be the only feature that distinguishes between vowels (e.g. in French: high, front, unrounded [vi] (‘vie’, ‘life’) vs. high, front, rounded [vy] (‘vu’, ‘seen’)). The International Phonetic Alphabet provides a set of symbols for transcribing the sounds of the world’s languages based on classificatory labels or features of the kind discussed above. Each symbol therefore provides information both about the sound that is produced and its relationship with other sounds. For example, [p] is a voiceless bilabial oral stop, which means that like [f] and [t], it has open vocal folds (voiceless), like [m] and [b] it has a constriction at the lips (bilabial), like [b], [f], and [t] it has a raised velum (oral), and like [b], [m], [t], and [d] the passage of air through the mouth is completely obstructed at some stage during its production (stop). The set of features and their relationships with each other have been considerably informed by a detailed instrumental analysis of the sounds of the world’s languages in the last 40 years (e.g. Ladefoged and Maddieson 1996). Other research seeks to understand why certain feature combinations tend to be preferred by languages, whether babies are born with detectors for identifying features, and the extent to which listeners make use of features in perceiving speech.

Phoneme and Phonetic Variation

Figure 4. The vowels and their phonetic symbols of the international phonetic alphabet. The vertical dimension is Height, the horizonatal dimension is Backness. When the symbols appear in pairs, the one to the right represents a rounded vowel.

The consonants at the beginning of the words ‘pin’ and ‘bin’ are different from each other not only phonetically (the first is voiceless, the second voiced) but also in a more abstract phonological sense: they communicate a difference of meaning and are allophones of separate phonemes. The two ‘p’ sounds in ‘pin’ and ‘spin’ are also phonetically very different from each

835

PHONETICS other in most accents of English: the one in ‘pin’, transcribed as [ph], is aspirated and produced with a puff of air that can be felt by saying the word and holding a hand near the lips, whereas the one in ‘spin’, transcribed as [p], is unaspirated and is produced without any puff of air. But while [ph] and [p] are phonetically very different, they do not communicate a difference in meaning and are allophones of the same phoneme /p/. The fact that these sounds do not distinguish between meanings in English can be shown by saying the word ‘spin’ but with an aspirated [ph]—the pronunciation might be unusual, but native speakers of English would nevertheless still identify the word ‘spin’. Allophones of the same phoneme occur predictably in different contexts: whereas [p] occurs in the context after /s/ (‘spin’, ‘spoke’, ‘spring’, etc.), [ph] occurs in a different context at the beginning of stressed syllables (‘pin’, ‘pan’, ‘put’, etc.). A type of transcription that makes use of only the phonemes of a dialect or language is known as a broad transcription: in such a transcription, all the phonetic characteristics that are predictable from context, and, therefore, redundant for distinguishing between meaning, are filtered out. A narrow transcription is one that includes some phonetic detail, usually at least all the allophones that are common to the speakers of a particular dialect. Every language has a finite number of phonemes out of which it assembles its words to construct differences of meaning. An important goal in phonetics and phonology is to establish both the phoneme inventory of a language and the relationships between the phonemes and their allophones. All of these differ markedly between languages and dialects. For example, while [p] and [ph] are allophones of the same phoneme in English, in Korean they are allophones of separate phonemes: as a result, exchanging them can produce a meaning difference in Korean but not English. The same phonemic opposition is nevertheless likely to give rise to phonetic differences between two languages. For example, both English and French create differences of meaning by distinguishing between two phonemes /b/ and /p/ (‘bin’ vs. ‘pin’ in English; ‘beau’ (‘beautiful’) vs. ‘peau’ (‘skin’) in French). But in French the corresponding phonetic distinction is between a fully voiced [b] (the vocal folds vibrate while the lips are closed) and an unaspirated [p], whereas in English the distinction is between a devoiced [b#] (the vocal folds only start vibrating after the lips come apart) and an aspirated [ph]. Not only languages but dialects of the same language can differ in their phoneme inventory and how these phonemes are produced, or realized phonetically. Whereas most English dialects have a phonemic difference between /u:/ in ‘pool’ and /υ/ in ‘pull’, in

836

Scots English there is no such phonemic opposition and so these words are indistinguishable. And while most English dialects have two very different allophones of /l/ depending on its position in the syllable, Southern Irish English makes no such distinction. Therefore, /l/ in ‘leaf’ and ‘feel’ are phonetically very similar in Southern Irish English but quite different in most other dialects. Phonetic variation occurs at the level of the individual speaker. This variation is not simply due to the anatomical differences between speakers (which is responsible for considerable acoustic differences in the speech of men, women, and children) but also because speakers may have developed an idiosyncratic speaking style: a given speaker might have excessive nasalization or a tendency to protrude the lips during speech production. Developing a model that can represent speaker-specific aspects of pronunciation is part of voice quality research and it has important applications both in disorders of voice and speech as well as in the forensic analysis of speech that is increasingly used in criminal investigations.

Prosody The same utterance can be said in many different ways to communicate paralinguistic effects of emotion such as happiness, sadness, surprise, anger, and so on. These paralinguistic effects are usually brought about by manipulating the prosody or the timing, pitch, and loudness of an utterance. But these manipulations can be used linguistically to distinguish either between words (word-level prosody) or between utterances (sentence- or utterance-level prosody). Prosodic effects usually extend over, and modify, more than a single consonant or vowel; this is why the units of prosody are sometimes called suprasegmentals. Word-level prosody includes differences due to quantity that are largely communicated by timing differences (these are marked phonetically by a colon: thus, [t:] is a long version of [t]). Quantity differences can occur in the consonant (e.g. Italian: [fato], ‘fact’ vs. [fat:o], ‘done’) or in the vowel (German: [lam] (‘lamb’) vs.[la:m] (‘lame’)) or in both (Finnish: [mut:a] (‘but’) vs. [mu:ta] (‘other’) vs. [mu:t:a] (‘to change’)). Word-level prosody also includes tone differences in which word distinctions are based on pitch: for example, in Thai, the syllable [na] has different word meanings depending on whether it is produced with a high-falling pitch (when it means ‘face’), a lowfalling pitch (when it is a name), or a high-rising pitch (when it means ‘aunt’)—see Ladefoged (2001a) for examples and sound files. The syllable and word-stress are part of word-level prosody. Almost all languages organize their phonemes

PHONETICS into syllable units and there is psycholinguistic evidence to show that children can identify syllables from a very early age. Much of the phonetic variation (such as the two types of /l/ discussed above) can be attributed to the structural position of a sound in the syllable. In many languages, consonants at the end of a syllable are phonetically much weaker than at the beginning of a syllable. This is the source of many sound changes in which syllable-final consonants weaken or delete completely (e.g. Latin ‘septem’  French [set] (‘sept’, ‘seven’), processes that also occur synchronically (e.g. the /t/ of ‘fast’ is perhaps never pronounced in a context like ‘fast speech’, so that ‘fast’ rhymes with ‘pass’). Word-stress has to do with the relative prominence or salience and, to a certain extent, the relative loudness of syllables: for example, in the word ‘abracadabra’, the first and fourth syllables are more salient than the others and are said to be stressed, while the others are unstressed. In English, the position in the word of the syllable with the strongest word-stress, known as the primary stressed syllable, can vary (e.g. ‘pattern’ vs. ‘admit’), whereas in other languages like Icelandic and Polish it has a delimitative function and occurs at a fixed position in the word. In some languages, there are a handful of words that differ only in their word-stress pattern: German ‘übersetzen’ means ‘to transport’ (e.g. by ferry) when it has primary stress on the first syllable but ‘to translate’ when it has primary stress on the third syllable. In English, unstressed syllables (and the sounds within them) can be considerably shortened and often completely deleted (e.g. the second unstressed syllable in ‘Tower Bridge’ can be deleted resulting in a pronunciation close to ‘Tar Bridge’). This very clear phonetic difference between stressed and unstressed syllables that affects vowels and consonants in English is much less is evidence in languages such as French and this is one of the attributes that contributes to the very differently sounding rhythm of these two languages. Utterance-level prosody includes intonation and accentuation, both of which can be used linguistically to provide a range of different meanings. Intonation depends phonetically on the rise and fall in pitch and can mark syntactic differences in many languages such as distinguishing a statement from a question. Languages, and indeed dialects of a language, have quite different associations between meaning and intonation. For example, whereas in most accents of English, the intonation falls toward the end of a statement, in Belfast-English it usually reaches a plateau and may rise; and whereas English, in common with many languages, has a rising intonation in questions requiring a ‘yes—no’ answer (‘Did Marianna make the marmalade?’), the intonation of such ‘yes–no’ questions in Greek, Hungarian, and Romanian usually falls.

In the same way that word-stress concerns the relative prominence of syllables in a word, accentuation has to do with the relative prominence of words in an utterance. In many, but by no means all, languages, shifting the accentuation can evoke a different meaning that depends on the discourse structure of a dialogue. Consider that when said in isolation, the most prominent or accented word of the sentence ‘I don’t like classical music’ is usually ‘music’. But if the same sentence were said in reply to: ‘Have you ever listened to Beethoven?’ then the accent in the same reply would shift to ‘like’ to indicate that ‘classical music’ is part of the background or old information in the dialogue that is shared between the speaker and listener. Accented words tend to be produced with greater clarity precisely because they often carry new information and are therefore much more difficult to predict from the context in which they occur. For a related reason, function words, which rarely carry new information and which are therefore much less likely to be accented, are less clearly produced and may be entirely deleted: consider that ‘the man in the moon’ can be understood as such simply by lengthening the [n] of ‘man’ and deleting the vowel in ‘in’ entirely. Speaking style can have a considerable influence on the utterance’s prosody. At faster rates of speech, or in a more casual conversational setting, many of the types of weakening and deletion discussed in this section are likely to be in greater evidence.

Speech Motor Control and Physiology Speech motor control is concerned with the neurological and physiological bases of speech production. Investigations in this area are directly relevant to articulatory synthesis in which the aim is to be able to synthesize speech using a model of the speech production mechanism. Studies of speech motor control make extensive use of instrumentation for the analysis of the vocal organs including: laryngography and direct fiberoptic laryngoscopy for measuring and viewing the activity of the vocal folds; aerodynamic techniques for measuring airflow and air pressure at various points in the vocal tract; and electropalatography, which records the pattern of contact between the tongue and the roof of the mouth. Midsaggital electromagnetic articulometry (EMA) has been in use for about 15 years for recording the movement and velocity of the jaw, the lips and the tongue; this technique has to a certain extent replaced electromyography, which measures the electrical activity associated with muscle contraction. Considerable progress has recently been made with magnetic resonance imaging (MRI), which provides exceptionally clear cross-sectional images of the vocal

837

PHONETICS tract, and with ultrasonic techniques, which have been used for recording the movement of the vocal folds and the tongue based on measuring reflections of highfrequency sound waves between tissue and air. A major issue in speech motor control is understanding a type of phonetic variation known as coarticulation. This comes about because, in contrast to the way in which letters are written on a page, sounds in sequence overlap with, and therefore influence, each other. Consider for example the production of ‘pan’: the last two phonemes /an/ are oral–nasal meaning that the velum is raised for the /a/ vowel and lowered for /n/. But since the velum can not jump instantaneously between these two states, it already starts to lower during the /a/ making the vowel nasalized well before the tongue tip touches the alveolar ridge for /n/. The production of /a/ is therefore influenced by the following /n/: it is because of this type of coarticulation that the /a/ in ‘pan’ is phonetically very different from the /a/ in other contexts such as ‘pad’. Coarticulation is ubiquitous and, since speech sounds can be communicated at a faster rate if they overlap with each other in time, proponents of the motor theory of speech perception (see next section) have argued that coarticulation is necessary to ensure that speech is produced sufficiently rapidly. Coarticulation may be anticipatory as in the ‘pan’ example, in which a sound is influenced by a following sound, or perseverative in which the influence is from a preceding sound. Various experiments have shown that it is very difficult to tie coarticulation definitively to any category boundary: coarticulatory influences can spread across phoneme, syllable, and word boundaries; hence, it would be quite possible for the velum to start lowering during the vowel of ‘saw’ in anticipation of the following /n/ in ‘saw another’. Explaining how listeners manage to recover the phonemic and linguistic content of an utterance intended by the listener in the light of these ubiquitous coarticulatory effects is a major research undertaking in speech motor control and its relationship with speech perception. Another major issue is the extent to which speech motor control is regulated by feedback—that is, information that is relayed back to the higher centers of the brain from hearing oneself speaking (auditory feedback) and detecting the movement of, and contact between the vocal organs (tactile and proprioceptive feedback). Research in this area has been informed both by studying the way in which fine motor control in adults who have become deaf in later life deteriorates, and from so-called immediate compensation experiments, in which investigators seek to determine the extent to which a vocal organ like the jaw can compensate for the lack of movement in the lips, if these are artificially immobilized. Recent research

838

suggests that feedback may be used to develop an internalized model of feedback: this would allow speech production control to be a good deal more rapid because, instead of having to wait for actual feedback to determine whether speech has been produced as intended, speakers could calculate the feedback they would be expected to get and make any necessary corrections before speech was actually produced.

Speech Acoustics and Speech Perception This branch of phonetics is concerned with the analysis of acoustic speech signals. Major progress in this area has been possible as a result of at least three major technological developments: the invention of the spectrograph in 1948, allowing an acoustic signal to be represented in terms of its time, frequency, and amplitude components; progress in speech synthesis technology in the 1950s and 1960s, in which utterances could be generated synthetically by machine; and advances in digital speech processing in the 1960s, in which computer algorithms were developed for the rapid calculation of the frequency content of a signal and for representing acoustic speech signals in terms of a small number of parameters. When a speech sound is produced, an acoustic signal is created whose characteristics are entirely dependent on the actions of the vocal tract that give rise to it. In articulatory-to-acoustic relationships, the aim is to break down the acoustic signal into a number of components that can be related to individual vocal tract actions out of which the speech sounds are formed: we would like to know, for example, how the acoustic signal is changed by opening and closing the velum, protruding the lips for a rounded vowel like [u], opening the mouth for [a], and so on. A major breakthrough in our understanding of such relationships was in the development in the 1950s and 1960s of the acoustic theory of speech production. With such a model, it became possible to reduce the complex shape of the vocal tract to a relatively small number of parameters (for example, vowels can be modeled by the cross-sectional areas and lengths of four interconnecting cylinders) and then to use these parameters to predict the likely acoustic output. This type of model, which was developed primarily by Gunnar Fant in the 1950s and 1960s has been central to a range of research areas in phonetics, including synthesizing speech and predicting the types and distribution of vowels and consonants that are most likely to occur in the world’s languages. When an acoustic signal reaches the listener, it is transformed in various ways as it passes through the ear and is represented by electrical impulses in the

PHONETICS auditory nerve. Auditory phonetics seeks to model these types of transformations in order to obtain a more accurate representation of how listeners actually perceive an acoustic signal. Some researchers hold the view that the auditory transformations reduce the considerable acoustic differences between male and female speech. Speech perception is concerned with how listeners retrieve phonemes and linguistic units from the acoustic signal. The application of speech synthesis to speech perception—allowing the individual contribution of acoustic cues to the perception of speech to be investigated—was pioneered at the Haskins Laboratories in the United States of America in the early 1950s. Their experiments led to the influential motor theory of speech perception, in which it is proposed that listeners extract or decode phonemes from the acoustic signal by first reconstructing the speech production strategies that could have given rise to them: that is, listeners hear not the acoustic signal but the movements of the vocal tract. In Kenneth Stevens’ lexical-access from features model by contrast, which is partly based on the well-known quantal theory of speech perception, it is proposed that there are landmarks of acoustic stability that allow listeners to access the mental lexicon directly without recourse to speech production. In Björn Lindblom’s hyper and hypo (H&H) model of speech, the notion that listeners make use of either invariant acoustic or articulatory landmarks is rejected. Instead, speech is said to vary along a continuum from hypo- to hyperarticulation, depending on the speaker’s desire to produce speech with a minimum of effort (hypoarticulation) and the need to produce speech clearly (hyperarticulation) at e.g. points of accent and information focus (see above). Speech perception is further complicated by the way in which decoding linguistic information from the speech signal interacts with ‘top-down’ processing, that is, with the knowledge of the language that the listener brings to bear in recognizing and understanding speech. It is quite clear that listeners do not recognize speech by first decoding the speech signal into phonemes and subsequently transforming these into words: a listener instead often recognizes a word well before the acoustic signal for that word has occurred in its entirety. Listeners can identify words that may be present only in a substantially impoverished form in the acoustic signal, or perhaps not even physically present at all (see the examples above). A further complicating factor is that, in contrast to the presence of white spaces between printed words on the page, there is no direct information in the acoustic signal about where one word ends and the next begins. Various experiments have suggested that listeners may make

use of prosody (in particular, syllable-structure and word-stress) to divide the stream of speech into separate words. Developing a model of how the available information from the acoustic signal interacts with the listener’s knowledge of the language is a major challenge both to understanding how speech is perceived by humans and to developing computers to accomplish an analogous task.

Conclusions Phoneticians have made considerable progress in understanding how abstract phonemic and prosodic units are related to the physical characteristics of speech sounds in the production and perception of speech in the languages of the world. The acoustic theory of speech production and the synthesis of intelligible speech from text are two examples of significant scientific discoveries in the last 50 years that have very clear practical applications in technology. The dramatic improvement in the storage and analysis of speech data by computer in the last decade has provided the tools for analyzing a far greater range of speaking styles of naturally occurring spontaneous speech. However, most of the phonetic knowledge is still derived from stylized laboratory speech of English and there is a great lack of comprehensively annotated databases from other languages and in particular from endangered languages. New experimental approaches to measuring the vocal organs have advanced our understanding of speech production control. There is now recorded material spanning several decades, which can be used to provide an experimental basis to sound change. Careful analyses of all these new types of data will lead to more sophisticated models and new discoveries in phonetics in the future. References Clark, J., and C. Yallop. 1995. An introduction to phonetics and phonology, 2nd edition. Oxford: Blackwell. Fant, G. 1973. Speech sounds and features. Cambridge, MA: MIT Press. Johnson, K. 1997. Acoustic and auditory phonetics. Cambridge, MA: Blackwells. Ladefoged, P. 2001a. A course in phonetics, 4th edition. New York: Harcourt Brace College Publishers. Ladefoged, P. 2001b. Vowels and consonants: an introduction to the sounds of language. Oxford: Blackwell. Ladefoged, P. 2002. Phonetic data analysis: an introduction to fieldwork and instrumental techniques. Oxford: Blackwell.

Ladefoged, P., and I. Maddieson. 1996. The sounds of the World’s languages. Oxford: Blackwell. Ladd, D. 1996. Intonational phonology. Cambridge: Cambridge University Press. Laver, J. 1994. Principles of phonetics. Cambridge: Cambridge University Press.

839

PHONETICS Laver, J., and W.J. Hardcastle. 1997. The handbook of phonetic sciences. Oxford: Blackwell.

Pickett, J. 1998. The acoustics of speech communication. Boston: Allyn and Bacon.

Lehiste, I. 1967. Readings in acoustic phonetics. Cambridge, Ma: MIT Press. ———. 1970. Suprasegmentals. Cambridge, MA: MIT Press.

JONATHAN HARRINGTON See also Spectral Analysis

Phonology The aim of phonology is to examine the way sounds are organized in languages and to explain the variations that occur. While it is physically possible to produce a wide range of sounds, only a relatively small number of these are used in a language. For example, a nasal preceding a plosive, such as in mbeke does not occur in English at the beginning of a word, but it does in several African languages. This is not because English speakers cannot produce the specific sequence of sounds, but because of the way speech sounds in a particular language are organized. In phonology, the most characteristic properties of different sounds, known as features, are compared to develop rules underlying the use of sounds in groups of languages. Systematic surveys of a representative number of languages are necessary to be able to generalize about sound systems, and to relate the findings to other areas of language (e.g. syntax and morphology). The UPSID database (UCLA Phonological Segment Inventory Database), developed by Ian Maddieson and colleagues in the 1980s, contains inventories of over 450 languages of the world. This database, as well as the Stanford Phonology Archive, has become a popular tool in teaching and in phonological and phonetic research. A detailed description of a language is not only of interest for the sake of having a survey of different languages of the world but is also essential to be able to teach foreign learners a language, or to treat children who are not acquiring the sound system of their language properly, or people who have lost the capacity to speak through injury or illness. In addition, speech technology communication systems require knowledge of the structure of a language.

Phonemes and Allophones A classical approach in phonology is to begin by establishing the phonemic system of a language, i.e. to determine which sounds are phonemes and which are allophones. Substitution of one phoneme for another will result in a word with a different meaning;

840

substitution of one allophone for another one only results in a different pronunciation of the same word. Phonemes are the ‘contrastive’ sounds of a language: /c/ and /s/ are different phonemes, because ‘cat’ and ‘sat’ contrast in meaning. Allophones are the phonetic variations of distinctively used sounds. Sounds that are not distinctive are also known as ‘redundant’. In English, aspiration is a redundant feature. However, in Thai, aspiration is distinctive, as the use of aspiration affects the meaning of the utterance. By substituting one sound for the other (the ‘commutation test’), it is possible to determine the phonemes of a language. Pairs of words that differ by just one sound, such as ‘cat’ and ‘sat’ or ‘robe’ and ‘roam’, are called minimal pairs. Other criteria also help to determine whether speech sounds belong to the same phoneme or not. One is complementary distribution, which refers to the situation where two sounds should not occur in the same environment. For instance, in English, aspirated plosives occur at the beginning of words, but not in consonant clusters. It can be predicted which allophone is pronounced in which context. Where we find the one, we do not find the other: they are mutually exclusive, never occurring in the same phonetic environment. Another criterion is free variation: sounds that occur in the same place in a word can belong to the same phoneme only if they do not change the meaning of the word. For instance, substitution of glottal stop [] for [t] as in ‘butter’ does not change the meaning of the word. Therefore, /buer/ and /butter/ are not minimal pairs. Also, sounds ought to display a reasonable amount of physical similarity to belong to the same phoneme. In summary, when two sounds are in complementary distribution or free variation, they are allophones of the same phoneme.

Distinctive Features A phoneme can be described by several articulatory and acoustical features. One of the main aims of

PHONOLOGY phonology is to identify the set of distinctive features required to describe the sounds of a language. Distinctive features serve to distinguish one phoneme from another. It is important to describe these detailed aspects of speech sounds, in order to understand how sets of sounds are related. Tables 1 and 2 list some distinctive features (not exhaustive) of some English vowels and consonants. If a feature is present, it is marked with a () sign; if it is not present, it is marked with a minus () sign. The features ‘front’, ‘back’, ‘high’, ‘low’, and ‘round’ refer to the position of the tongue in the vocal tract. For front sounds, the body of the tongue is fronted re its neutral position (for the /ə /, a schwa), for back sounds it is retracted re its neutral position. For high sounds, the body of the tongue is raised re the / ə /, while for low sounds it is lowered re the / ə /. Rounded sounds are produced with protruding lips. Voicing refers to the presence or absence of vocal fold vibration. The feature ‘continuant’ distinguishes between stops (nasal stops included) and other sounds. Anterior sounds are produced with the tongue tip at or before the alveolar ridge in the vocal tract, and coronal sounds are produced with the tongue tip or blade raised (it includes some palatal consonants). Strident sounds are fricatives or affricates (stop  fricative) with high-frequency noise (a hissing sound). Nasal sounds are produced with the velum lowered, resulting in airflow through the oral and nasal cavities. The notion of distinctive features is not only relevant for analysis purposes but also for the description of phonological processes. Some phonological processes, such as for instance assimilation, affect certain combinations of features more than others,

and this should be reflected in the phonological representation. This is possible if it is assumed that each distinctive feature is free to act independent of the other features it may be associated with. This approach consisting of ‘interconnecting levels’ (tiers) is known as autosegmental phonology: each feature can influence a neighboring sound segment, irrespective of whether other associated features do so. In particular, feature geometry describes the (nonlinear) action of one feature on another by representing their relationship as hierarchical (tree) structures.

Rules Phonological rules make statements about which allophones of a phoneme will occur in a specific context. The change of one item to another item in a certain environment can be stated by a rule. The validity of the rule can then be tested against other examples to determine exceptions between sounds within and across languages. Rules are important for discovering the universal principles governing the use of sounds in languages. It is necessary to specify the item(s) affected, the change(s) that takes place (indicated by ‘→’), and the environment in which the change occurs (indicated by ‘/’). In linear rule writing, the original state of affairs is given on the left of the arrow and the structural change and environment on the right of the arrow. For example, in English, vowels are nasalized before a nasal stop. This can be stated as follows: ~ ]/ ___ /n/ e.g. /æ/→ [æ It is also possible to use brace notation to describe one process in two different environments. The following example states that /t/ can be pronounced as a glottal

TABLE 1 Possible Distinctive Features of Some English Vowels (Not Exhaustive) Front Back High Low Round

u

ɔ

o

ɑ



%

e

ε

i

i

ə

    

    

    

    

    

    

    

    

    

    

    

TABLE 2 Possible Distinctive Features of Some English Consonants (Not Exhaustive) Voice Continuant Strident Anterior Coronal Nasal

p

t

k

b

d

g

m

n

f

s

ɵ

ʃ

v

z

ð

h

l

r

w

j

     

     

     

     

     

     

     



     

     

     

     

     

     

     

     

     

     

     

    

   

841

PHONOLOGY stop before a consonant or the end of a word (such as in ‘button’).



C /t / → []/ —– #

By using parentheses and braces it is possible to formulate rules of greater complexity. Moreover, the Greek alphabet, or alpha notation, can be used to match features in different places in the rule. While some rules state how features are affected by the context of other features (feature-changing rules, such as nasalization, glottalization, flapping), other rules can affect the entire segment. For example, in English a schwa is inserted between a final liquid and a nasal, as in [filəm] (instead of /film/). This example of insertion can be described as: ∅ → ə/



cons son -nas



#  cons nas 

—–

where ‘cons’ refers to consonantal sounds, ‘son’ to sonorant sounds, i.e. sounds that contain considerable energy (such as /m,l,j/), and ‘nas’ to nasal sounds. Given the presence of certain features, one can predict the value of other features in that segment. For example, there are no rounded front vowels in English. If a vowel is specified as [-back] it is also [-round]. It is therefore not necessary to specify [-round]. A blank indicates that the feature is predictable by a phonological rule of the language. Similarly, if a phoneme is [nasal] it is also [voiced] in English. Sometimes, nasal phonemes are ‘devoiced’ after an initial /s/ [snu' p]. However, this does not affect the meaning of the word, as is the case in Burmese where /ma/ means ‘health’ and /ma' / means ‘order’ (therefore, they are different phonemes).

Suprasegmentals Until now, we have considered the individual phonemes of a sound system. However, several phonological processes affect units that are larger than phonemes, such as syllables, words, phrases, and sentences. In some languages, words contain (with some exceptions) only front or back vowels (e.g. Turkish). The analysis of phonological features in terms of units larger than a phoneme is dealt with by prosodic phonology, and autosegmental phonology. Another branch of phonology, suprasegmental phonology, deals with aspects of pitch, loudness, tempo, rhythm, and tone. A phonological theory concerned with organizing phonemes into groups of relative prominence that emphasizes the relationship between phonemes and rhythm, and intonational stress is known as metrical phonology.

842

Optimality Theory Until now, phonological processes have been addressed by means of rules and derivations (derivational Generative Phonology). Application of one rule often affects the subsequent application of some other rule. However, during the last decade, Optimality theory has obtained a dominant role in phonology. This approach has introduced an alternative way of modeling the relationship between words and sounds. Rather than trying to define how a language′s words may be derived by combining given phonemes according to certain rules, Optimality theory approaches the problem from the opposite direction: in principle, any combination of features/sounds is assumed to be possible, but each language imposes certain constraints in a language-specific way on the endless possibilities. The constraints are assumed to be universal while they may be ranked differently, which allows for language-specific grammars. While phonology attempts to model knowledge of a language, it should also reflect the fact that languages are learnable. Proponents of Optimality theory argue that apart from dealing with phonological phenomena, it also includes a better learning theory than traditional (derivational) approaches. In addition, Optimality theory has already proven to be a productive new tool in other domains of linguistic analysis, such as morphology, syntax, and even semantics. References Anderson, Stephen. 1974. The organization of phonology. New York: Academic Press. Chomsky, Noam, and Morris Halle. 1968. The sound pattern of English. New York: Harper & Row; Paperback edition 1991, Cambridge, MA: MIT Press. Clements, George N. 1992. Phonological primes: features or gestures?. Phonetica 49. 181–93. Durand, Jacques, and Francis Katamba. 1995. Frontier of phonology. London: Longman. Goldsmith, John. 1990. Autosegmental and metrical phonology. Oxford: Blackwell. Goldsmith, John. 1995. The handbook of phonological theory. Oxford: Blackwell. Jakobson, Roman. 1990. On language. Cambridge, MA: Harvard University Press. Kager, R. 1999. Optimality theory. Cambridge: Cambridge University Press. Ladefoged, Peter, and Ian Maddieson. 1996. The sounds of the World’s languages. Oxford: Blackwell. Maddieson, Ian. 1984. Patterns of sounds. Cambridge: Cambridge University Press. Roca, Iggy. 1994. Generative phonology. London: Routledge. Spencer, Andrew. 1996. Phonology: theory and description. Oxford: Blackwell. Vihman, Marilyn. 1996. Phonological developments: the origins of language in the child. Oxford: Blackwell.

ASTRID VAN WIERINGEN

PHRASE STRUCTURE

Phrase Structure Sentences are not just linear strings of words. There is ample evidence that words in a sentence can form a unit that excludes other words in the same sentence. Such units are referred to as constituents. Constituents are hierarchically organized, forming larger phrases. For example, a sentence will often consist of a subject and a predicate. The constituent that forms the predicate can in turn consist of several smaller constituents, such as a verb, a direct object, an adverbial, and so on. These constituents, too, may be complex and contain even smaller constituents. The internal structure of sentences and smaller constituents is generally referred to as ‘phrase structure’. Evidence for the assumption that the internal organization of a sentence consists of hierarchically ordered constituents is manifold. One piece of evidence concerns the fact that the words in a sentence can be redistributed in certain ways such that the result is again a grammatical sentence. In such a redistribution process, certain words stay together. This indicates that they form a unit—a constituent. For instance, in the sentence in (1a), the phrase the films by Pierre Paolo Pasolini functions as the direct object of the verb likes: it is the thing that is liked. It thus forms a semantic unit. That it also forms a syntactic unit is indicated by (1b): the phrase can be placed at the beginning of the sentence as a whole (a process technically known as ‘topicalization’) and the result is still grammatical. This cannot be done with just any string of words that happen to stand next to each other. This is shown by (1c), where the attempt to topicalize likes the films, to the exclusion of the rest of the direct object, results in ungrammaticality. This shows that this string does not form a syntactic unit on which syntactic rules such as topicalization can operate. Thus, likes the films is not a constituent of the sentence in (1a). (Ungrammatical examples are marked by an asterisk, following the common convention in the literature). (1a) John likes the films by Pierre Paolo Pasolini. (1b) The films by Pierre Paolo Pasolini, John likes. (1c) *Likes the films, John by Pierre Paolo Pasolini. Another indication that phrases are structured and consist of constituents is that certain strings of words can be replaced by one word, whereas others cannot. In (1a), the films by Pierre Paolo Pasolini can be replaced by the question word what (which in English has to be placed at the beginning of the sentence), as

in (2a). Again, this is not possible for likes the films, as shown by (2b). Replacement thus gives the same indication of constituency as does redistribution. (2a) What does John like? (Answer: The films by P.P. Pasolini). (2b) *What does John by Pierre Paolo Pasolini? (Answer: Like the films). The smallest constituents of a sentence are words, which are further organized into phrases, which in turn are further grouped into larger phrases, and so on. Phrasal constituents are built around a word that functions as the head of the phrase. The head of a phrase is that word in the phrase which in a sense is its most important part: it cannot be omitted, and it determines most of the syntactic properties of the phrase as a whole. For example, subjects of sentences are very often phrases that are built around a noun. The phrase can consist of just this noun, as in (3a), but more material can be added, as in (3b), where the head music is modified by an adjective preceding it, and (3c) where a relative clause following the noun is added. In this way, arbitrarily large phrases can be formed. The head cannot be left out, however, as illustrated in (3d). (3a) [Music]NP is his favorite pastime. (3b) [Classical music]NP is his favorite pastime. (3c) [Classical music written in the twentieth century]NP is his favorite pastime. (3d) *[Classical written in the twentieth century]NP is his favorite pastime. If a phrase is built around a noun, the resulting structure is a noun phrase, abbreviated as NP. Any phrase headed by a noun has a syntactic distribution that is determined by its head. For instance, phrases headed by a noun may appear in subject, direct object, indirect object positions, etc. In contrast, phrases headed by an adjective (APs) cannot normally function as such. (4a) [The girl]NP gave [the boy]NP [a new cd]NP. (4b) *[Pretty]AP gave the man a new cd. Adjectival phrases, on the other hand, can act as modifiers to nouns (e.g. the AP classical modifies the noun music within the NP classical music in (3b), whereas NPs do not ordinarily appear in this position (bar exceptional cases like a London bus). The distribution of phrases headed by a verb (VPs such as cook pasta or like music), phrases headed by prepositions

843

PHRASE STRUCTURE (PPs such as in the cupboard and under the stairs), and phrases headed by an adverb (AdvPs such as very quickly or yesterday) is yet different. Depending on the properties of the head itself, other material within a phrase besides the head can also be obligatory. Certain lexical items require additional elements to be present in a phrase that they are the head of. This can be due, at least in part, to their meaning. For instance, the verb like has to combine with an object (5a). Omission of this so-called complement to the verbal head leads to ungrammaticality (5b) in this case. However, not all complements are obligatory. The verb eat can combine with an object (5c), but need not necessarily do so (5d). In the latter case, an object is implicitly understood. (5a) (5b) (5c) (5d)

John likes music. *John likes. John is eating plums. John is eating.

Apart from heads and complements, phrases may contain additional, purely optional, material. Such optional elements are referred to as adjuncts. In VPs, they are most often adverbials such as time, place, and manner adverbials; in NPs, modifying relative clauses, for example, can act as adjuncts. Adjuncts differ from heads and complements in not being unique in a phrase. Whereas a phrase can only have one head, and (barring some problematic cases) also only one complement, adjuncts can be freely stacked. This is illustrated in (6). (6)

John often walks his dog in the park, in the spring, on a nice Sunday afternoon, while whistling The blue Danube.

An early formalism of expressing the possible constituent structures of sentences made use of so-called rewrite rules (also known as phrase structure rules; see Chomsky 1957, 1965). They have the general form of (7), where XP stands for a phrase of any category, and Y, Z, and W for its constituent parts. The rule in (7) states that the phrase XP consists of Y, Z, and W (in technical parlance, XP is said to be rewritten as Y Z W). Y Z W can be words, but can also be phrases themselves. (7)

XP → Y Z W

Simplifying somewhat, the structure of NPs can be expressed by the rewrite rule (8a). Rule (8a) states that NPs consist of a determiner, optionally an adjective, the head noun, and optionally a PP. The structure of a VP can be given by the rule in (8b), which states that a VP consists of a verb, an NP (functioning as an object), and (optional) adverbs. The structure of a PP can be represented as given by the phrase structure

844

rule in (8c), which combines a preposition and its complement. (8a) NP → Det (Adj) N (PP) (8b) VP → V NP (Adv) (8c) PP → P NP It is worth noting that the NP mentioned in (8c) can be rewritten in accordance with (8a), meaning that it can contain a PP. This PP can be rewritten again in accordance with (8c), so that it contains an NP, which can be rewritten yet again in accordance with (8a), and so on, ad infinitum. This results in sequences like in (9). In principle, there is no limit to such sequences, although for practical purposes they will stop at some point. This shows a pervasive property of the phrase structure of natural languages: it is recursive. (9) I saw the mouse in the hole in the wall of the house next to the river beside the meadow in the county next to … Sentences are formed by combinations of several phrases, in the simplest of cases an NP that functions as the subject of the sentence and a verb phrase that functions as the predicate (as for instance in [[John]NP [likes Mary]VP]S). This is expressed by the rewrite rule in (10). (10) S → NP VP Rewrite rules were replaced in favor of a theory of phrase structure, which proved to be very influential. This theory is known as X’-theory (pronounced as X bar theory). It was first proposed in Chomsky (1970), and further elaborated in such works as Jackendoff (1977), Fukui (1986), and Speas (1990). It originated from the idea that all phrases, regardless of the categorial status of their head (i.e. whether they are NPs, VPs, PPs, AdjPs, or AdvPs), are built along the same structural schema, given by the tree diagram in Figure 1. X is a variable standing for a head of any category (N, V, P, A, Adv), YP is the complement of the head, while ZP is called its specifier. X’ and X” are the projections of the head X. When the head X combines with the complement YP, it projects up to the X’ (pronounced X bar) level. The X’ constituent further combines with a specifier leading to the projection of the X” (pronounced X double bar) level. In the classical X" (=XP)

X'

ZP

X

Figure 1

YP

PHRASE STRUCTURE version of the theory, the X” level is considered to be the highest level up to which the head can project. It is therefore also referred to as the XP level, designating a full phrase. Examples of how phrases of various types fit into the general scheme in Figure 1 are given in Figure 2. Like the complement and the head, the specifier is unique. For example, while a noun can take a possessive phrase as specifier, and can take a determiner as specifier, it cannot take both: (11a) Mary’s copy of Dracula (11b) a copy of Dracula (11c) *a Mary’s / Mary’s a copy of Dracula There are several structural relations that can be defined over the tree diagram in Figure 1. Any two constituents that are immediately dominated by the same node are called sisters. X and YP are sisters, being immediately dominated by the X’ node, as are X’ and ZP, the latter two being immediately dominated by the XP node. The XP node is said to be the mother of X’ and ZP, while the X’ node is similarly the mother of the X and YP nodes. The specifier of the head X can therefore be defined as the immediate VP NP

V′ V

NP

loves

ice-skating

His brother (a)

PP

daughter of X” and sister of X’, and the complement can be defined as the immediate daughter of X’ and sister of X. The schema in Figure 1 cannot accommodate all the possible constituent parts of phrases. In particular, it cannot accommodate the adjuncts in a phrase, of which there can be arbitrarily many (see above). To accommodate adjuncts, it is assumed that bar levels can be reiterated. When an adjunct is added to a phrase, the head does not project to a higher bar level; instead, the bar level is just repeated. In this way, an adjunct that is added at the double-bar level can be formally distinguished from the (unique) specifier of the head: whereas both the adjunct and the specifier are daughters of an X”-node, only the specifier is the sister of an X’-node; the adjunct is the sister of an X”node (Figure 3). Because the specifier can also be distinguished from adjuncts by the very fact that adjuncts can be stacked, whereas the specifier is unique, it is not clear that it is essential to distinguish between the two formally as well (see below). An assumption built into most modern phrase structure theories is that all nodes are maximally binary branching. This means that no node can have more than two daughters. If this assumption is adhered to, ‘flat’ structures of the type in Figure 4, for a VP that contains both a direct object and an indirect object, are impossible. The consequence is that there must be a more hierarchical structure inside VP. Evidence for this is said to come from examples like (12a, b). These show that the indirect object can act as antecedent for a reflexive element (for instance, himself) inside the direct object,

P′

AdvP

X"

right (b)

P

NP

across

the bridge

WP

X"

NP YP AP N

PP

pictures

of Amsterdam

beautiful (c)

X

Figure 2

VP A′

AdvP

(d)

ZP

Figure 3

AP

very

X'

N′

A

PP

fond

of music

V

NP

NP

give

mary

a book

Figure 4

845

PHRASE STRUCTURE but not the other way around. Evidence from other types of sentences shows that the antecedent for the reflexive has to be higher in the structure than the reflexive itself. The data in (12a, b) therefore indicate that the indirect object and the direct object are not on the same level in the structure. Instead, the indirect object must be higher than the direct object.

DP

D′

(12a) Mary showed Bill a double of himself. (12b) *Mary showed a double of himself Bill. The scheme in Figure 1 is not intended to express a linear order of constituents. The theory of phrase structure in the first instance is about the hierarchical relationship between phrases. Different principles must then determine whether a complement precedes or follows the head, as well as whether the specifier precedes or follows the head. Syntactic theory distinguishes between two classes of words: lexical and function words. The position of function words in phrase structure theory has always been rather unclear. Recently, such elements have come to be regarded by some as heads of phrases, on par with lexical categories. In line with this development, items like determiners (such as the and a in English), complementizers (such as that, if, whether), auxiliary verbs and modals (such as to be, can, will), and even inflectional affixes for tense and agreement (such as the English –ed suffix in past tense forms) are regarded as the heads of phrases. This view has various consequences for how the internal structure of phrases is supposed to be built up. For instance, a phrase like the house is no longer seen as an NP, but rather as a determiner phrase (DP), headed by the determiner D, which takes the NP house as its complement, as represented in Figure 5. Similarly, sentences are now regarded as projections of the inflection we see on the verb, and hence as inflectional phrases or IPs (Figure 6). Subjects are in the specifier position of IP. In the same vein, sentences introduced by a complementizer are regarded as complementizer phrases or CPs. The complementizer head in the CP takes an IP as its complement, which in turn takes a VP as its complement (Figure 7). The specifier position of CP can also host elements in some sentences, such as question words. The proliferation of functional categories in the theoretical inventory also has consequences for the notion of adjunction. With every additional functional projection, an additional specifier position becomes available. This means that it is at least possible to discard the assumption that bar levels have to be repeated in order to accommodate adjuncts. Instead, every adjunct may be regarded as a specifier of a particular functional projection, as extensively argued in Cinque (1999).

846

D

NP

the

house

Figure 5 IP DP

I'

John

VP

I

V'

is V

DP

singing

a song

Figure 6 CP C' C

if

IP DP

john

I' I

VP

is

V' V

DP

playing the cello

...

Figure 7

Classical X’-theory not only needs the possibility of repeating bar levels to accommodate all the possible constituents of a phrase, but it suffers from the opposite problem as well: it sometimes forces more syntactic positions to be present in the structure than are actually necessary. For instance, a head can take a specifier while not taking a complement. Whereas (13b) is as well formed as (13a) is, in classical X’-theory there nevertheless is a complement position present in (13b) as well. (13a) John’s collection of mushrooms (13b) John’s collection It is even possible that a phrase contains just its head, without a specifier or a complement, as for instance in (14b) (vs. (14a), with specifier and complement present). If phrases always have the form of

PIDGINS AND CREOLES Figure 1, then in such cases there is both a specifier position and a complement position, neither of them containing any lexical material. (14a) (I expect) [Mary to win the race] (14b) (I expect) [to win] Partly because of this inelegant aspect of X’-theory, a more flexible view of phrase structure has been adopted (see Speas 1986; Chomsky 1994), usually referred to as Bare Phrase Structure theory. Its basic assumption is that a head projects as often as it combines with another phrase, no more and no less. One consequence of this is that it is no longer possible to distinguish between specifiers and adjuncts in structural terms (see also the previous point on functional structure). Instead, the specifier can be distinguished relationally, since it establishes a unique relation with a head. References Abney, Stephen. 1987. The English noun phrase in its sentential aspect. Ph.D. Dissertation. Cambridge, MA: MIT.

Cinque, Guglielmo. 1999. Adverbs and functional categories: a cross-linguistic perspective. New York, Oxford: Oxford University Press. Chomsky, Noam. 1957. Syntactic structures. The Hague: Mouton de Gruyter. ———. 1965. Aspects of the theory of syntax. Cambridge, MA: MIT Press. ———. 1970. Remarks on nominalization. Readings in English transformational grammar, ed. by Roderick A. Jacobs and Peter S. Rosenbaum. Waltham, MA: Ginn. ———. 1994. Bare phrase structure. Cambridge, MA: Distributed by MIT Working Papers in Linguistics. Fukui, Naoki. 1986. Theory of projection in syntax. Stanford, CA: CSLI Publications, Center for the Study of Language and Information, and Tokyo: Kurosio Publishers. Grimshaw, Jane. 2003. Words and structure. Chicago: University of Chicago Press. Jackendoff, Ray. 1977. X’ Syntax: a study of phrase structure. Cambridge, MA: MIT Press. Kayne, Richard S. 1984. Connectedness and binary branching. Dordrecht: Foris. Pollock, Jean-Yves. 1989. Verb movement, universal grammar, and the structure of IP. Linguistic Inquiry 20. 365–424. Speas, Margaret. 1986. Adjunctions and projections in syntax. Ph.D. Dissertation. Cambridge, MA: MIT.

AMELA CAMDZIC AND PETER ACKEMA

Pidgins and Creoles Although until recently creolists were more or less agreed on the definition of ‘pidgin’ and ‘creole’, this is no longer so. Also, the adage that ‘a creole is a nativized pidgin’ (Hall 1966) is no longer universally accepted. It has been observed, for example, that at least some of the French-lexicon creoles do not have a pidgin ancestry (Chaudenson 2001). Further, the more we learn about the transition from pidgin to creole, the less certain we are about where to draw the line between the two (Baker 1995). At the same time, creolists are becoming aware that it may not be justified to group pidgins and creoles together, even though in the past they have often been treated as if they belong to one category. In spite of all these difficulties, some definitions will be presented here, if only to give the reader an idea of the distinctions. A pidgin is a nonprimary language that is the result of language contact. A creole is a primary language that is the result of language contact. Before explaining these definitions, a brief survey of the pidgins and creoles of the world is useful. According to the most extensive survey that has been made until now (Smith 1995), the total number of pidgins and creoles is around 350, including both the ones that are extinct as well as varieties that are intermediate

between a ‘true’ creole and its related ‘lexical donor language’ (also called ‘lexifier language’, ‘superstrate language’, or simply ‘superstrate’). The total number of speakers, including those who only use a creole as a second language (a pidgin is a second language by definition), may be estimated at around 100 million. By far, the largest group, especially among the creoles, is formed by those that have derived the bulk of their lexicon from an Indo-European language such as Portuguese, French, English, Spanish, and Dutch. It is no coincidence, of course, that these are the languages spoken by the European nations that played the leading role in the European expansion. It was the contact between the European languages of the explorers and colonizers, on the one hand, and the non-European languages of the people with whom they came into contact, on the other (referred to as ‘substrate language(s)’, or simply ‘substrate’), that gave rise to the emergence of these pidgins and creoles. Other Indo-European languages, apart from the ones mentioned above, that have played a role in the emergence of pidgins and creoles include German, Russian, Italian, and Hindustani. Although these European-lexicon pidgins and creoles have attracted the lion’s share of the attention of

847

PIDGINS AND CREOLES linguists, there is a significant and linguistically just as important group of pidgins and creoles whose lexicons are based on non-Indo-European languages. The lexifier languages in this group include Arabic, Japanese, Chinese, and a number of Bantu, Austronesian, Australian, Papuan, and Amerindian languages. An important difference with the group of European-lexicon pidgins and creoles is that in many cases no IndoEuropean language was involved in their formation. Because of this, the study of non-European pidgins and creoles may serve to correct any biases that may have resulted from the focus on European pidgins and creoles that has characterized the field of pidgin and creole linguistics for a long time (see Thomason 1997, 2001). An important question in this regard is whether the features of pidgins and creoles that are sometimes regarded as universal might not be an artifact, resulting from the fact that all the major lexifier languages are structurally similar, belonging as they do to only two branches (Germanic, Romance) within one and the same language family, Indo-European. To get an idea of the variety as well as the geographical distribution of the world’s pidgins and creoles, the maps in Holm (2000) are a good starting point. As far as the Pacific is concerned, much more detailed information can be gleaned from the splendid three-volume language atlas edited by Wurm et al. (1996). Unfortunately, a similar work for the Atlantic region is not (yet) available. As mentioned earlier, the definitions given above require some comments. First of all, what is meant by ‘primary’ and ‘nonprimary’ language? If you are multilingual, your primary language is the one with which you feel most at home, the one you use most often, even though it may not be your native language. When we say that a pidgin is a nonprimary language, it means that pidgin-speakers have at least one other language in their repertoire—their primary language— and that the pidgin is used only as an auxiliary language, when communicating with speakers with whom they have no other language in common. Creoles, on the other hand, are primary languages by definition, although, again, this does not mean that they are necessarily native languages. Second, if pidgins are nonprimary languages, is it justified to view them as ‘true’ languages in the first place? Although this is an interesting question in itself (when is something a language?), it is too complex to be dealt with here. For the purpose of this essay, however, pidgins are referred to as languages. As for creoles, there is no question about it: they are languages just like any other. The one thing that pidgins and creoles have in common, according to the definitions given above, is that they are a product of language contact. Language contact, therefore, is the key concept in these definitions: it is a necessary, although not sufficient, condition for pidginization and creolization (the process by which a 848

pidgin or a creole comes into being). This means, first, that there will be no pidginization or creolization without contact, and, second, that contact does not automatically lead to pidginization or creolization. When we look at pidginization and creolization as language contact phenomena, the following questions present themselves: (1) (2)

(3)

What other types of language contact are there, apart from pidginization and creolization? In what ways do pidginization and creolization differ from other types of language contact? What are the factors that decide whether language contact will lead to pidginization or creolization?

These three questions will guide the remainder of the discussion. What other types of language contact are there, apart from pidginization and creolization? The linguistic effects of language contact are dependent on a number of factors, the most important of which is whether or not imperfect language learning plays a role in the contact situation. On the basis of this, two types of language contact can be distinguished (Thomason 2001): (a) Shift-induced interference. When people give up their native language while adopting another one, the new language often displays features from their native language. In this case, the changes in the affected language are a result of imperfect learning. (b) Borrowing. When people maintain their native language, it may still be influenced by another language. This happens when native speakers incorporate features from another language into their native language. In this case, the changes in the affected language are not a result of imperfect learning. The crucial difference between the two is that in the case of shift-induced interference, speakers introduce changes in a language that is not their native language, while in the case of borrowing they introduce changes in a language that is their native language. Although the end-results of these two types of language contact are sometimes hard to distinguish, the processes by which these results come about are very different, for example, in the order in which the different components (lexicon, phonology, etc) of the language system are affected. While borrowing always starts with words and only involves the grammatical system later, it is the other way around in shift-induced interference. The results of the two processes may range from minor changes in the lexicon, such as adoption of a loanword, to major structural changes, such as changes in basic word order.

PIDGINS AND CREOLES In what ways do pidginization and creolization differ from other language contact phenomena? Pidginization and creolization differ from other language contact phenomena (except, perhaps, so-called ‘mixed languages’; see relevant lemma) in that neither pidgins nor creoles can be justifiably viewed as changed versions of the languages in contact, while this is the case with borrowing and shift-induced interference. In other words, pidgins and creoles are autonomous languages. This entails that they have a grammar of their own, one that is not derivative of or dependent on a preexisting language. The autonomy of the grammar is the result of a process whereby grammatical ‘components’ (rules, patterns, processes) that were initially related to one or more of the contributing languages, have become independent. The transition from dependency to independency occurs when these components start interacting with one another to produce a new grammar, rather than forming an incoherent collection of elements derived from other languages. The idea of pidgins and, especially, creoles as autonomous languages is shared by most creolists, including those who view creolization in terms of either borrowing or shift-induced interference. The view of creolization as a case of (extreme) borrowing is held by Lefebvre (1998), who sees Haitian Creole as the result of a process of relexification, whereby words from the slaves’ native language(s) were replaced by similar words from French, while the grammatical structure of the native language(s) was maintained. A very different view is represented by Chaudenson (2001), who believes that—especially the French-lexicon—creoles are the result of a process of language shift (from the slaves’ native languages to French), whereby some features may have been introduced as a result of interference. No matter how far apart, both views agree that creoles are autonomous languages, not modified versions of one or more of the contributing languages. The idea of creoles as autonomous languages is also inherent in Thomason’s (2001) view of pidginization and creolization as cases of ‘abnormal’ language transmission. While in ‘normal’ transmission both the lexicon and the grammatical apparatus are transmitted from one and the same ‘parent language’ to one or more ‘daughter languages,’ in cases of ‘abnormal’ transmission, there are at least two ‘parents’, one for the lexicon and one for the grammatical system. As a result of this, a pidgin or a creole cannot be a variety of either one of the parent languages; therefore, it is an autonomous system. Finally, in generative approaches to pidginization and creolization (e.g. DeGraff 1999), the question of autonomy hardly arises, as every language—whether pidgin, creole, or something else—is seen as an instantiation of Universal Grammar. Although this brief discussion does not exhaust the list of theories of creole genesis,

it seems fair to say that they all share the concept of autonomicity as a basic feature of pidgins and creoles. What are the factors that decide whether language contact will lead to pidginization or creolization? This is one of the fundamental questions regarding pidginization and creolization. Since little attention has been devoted to it, especially from a historical point of view, the discussion here is necessarily tentative. It should also be noted that what follows is restricted to European-lexicon pidgins and creoles (by far the largest group), and that it is not necessarily valid for pidgins and creoles that are lexically based on other languages. It is hoped, however, that insights into the conditions for the genesis of the European-lexicon pidgins and creoles will also shed some light on the other cases. The European pidgins and creoles both arose in the historical context of the European expansion. This particular phase in history began around 1430 when Prince Henry (‘the Navigator’) started sending out Portuguese ships to explore the northwestern coast of Africa. Although the exact date when the contacts between Europeans and Africans for the first time led to the emergence of a pidgin or creole language in not known, it is clear that the genesis of the European-lexicon pidgins and creoles took place in the context of the European expansion, whether in Africa, America, Asia, or the Pacific. The age of European expansion was a very special, perhaps even unique, phase in the history of the world, which may be characterized by the following features: (a) The introduction of trade and production on a truly global scale. (b) The presence of contact situations characterized by unusually wide social, cultural, and psychological gaps between the parties involved (Europeans on the one hand and Africans, Asians, Native Americans, and Polynesians on the other). (c) The use of forced labor and forced displacement of large numbers of people (slave trade, indentured labor). (d) The creation of a new type of society known as ‘the plantation complex’ (Curtin 1998). With regard to the last feature, it should be added that, although plantations in the strict sense were not present in every single situation where pidgins and creoles developed, the concept of ‘plantation complex’ may be interpreted in a sufficiently wide sense to encompass the latter situations as well. Although the exact relationship between the presence of particular external conditions on the one hand and the emergence of pidgins and creoles on the other is not very clear, a few things may still be said about it. For example, there is the issue of access to the target language. With the increasing disproportion between the 849

PIDGINS AND CREOLES European and non-European segments of the population and the concomitant change from the small-scale ‘homestead society’ to the large-scale ‘plantation society’ (Chaudenson 2001), access of the non-Europeans to the European language decreased. However, this cannot be the whole story because there are a number of cases, especially in Spanish colonies, where such disproportion did not lead to pidginization or creolization (McWhorter 2000). Other factors, such as lack of motivation on the part of the enslaved to learn the language of their masters (Baker 1990), may have been involved as well. However, there is another, more elusive, factor that has received little attention until now. If we look at the period of the European expansion from a contemporaneous rather than from a modern perspective, we cannot help but recognize that this unique phase in history confronted all those who were involved in it with an entirely ‘new world’. Never before had so many different parts of the world been in contact on such a large and intensive scale. Although the fact that this new contact situation was accompanied by a true ‘explosion’ of new languages does not necessarily mean that the one was caused by the other, the synchronicity of the two ‘events’ is striking. It is hoped that future research, especially into the historical dimension of pidginization and creolization, will shed more light on this issue. References Arends, Jacques, Pieter Muysken, and Norval Smith (eds.) 1995. Pidgins and creoles: an introduction. Amsterdam, Philadelphia, PA: Benjamins. Baker, Philip. 1990. “Off target?” Journal of Pidgin and Creole Languages 5. 107–19

———. 1995. Some developmental inferences from historical studies of pidgins and creoles. The early stages of creolization, ed. by Jacques Arend, 1–24. Amsterdam, Philadelphia, PA: Benjamins. Chaudenson, Robert. 2001. Creolization of language and culture. Revised in collaboration with Salikoko Mufwene [translation of Des îles, des hommes, des langues. Paris: L’Harmattan. 1992]. London and New York: Routledge. Curtin, Philip. 1998. The rise and fall of the plantation complex, 2nd edition. Cambridge: Cambridge University Press. DeGraff, Michel, (ed.) 1999. Language creation and language change: Creolization, diachrony, and development. Cambridge MA: MIT Press. Hall, Robert. 1966. Pidgin and creole languages. Ithaca, NY: Cornell University Press. Holm, John. 2000. An introduction to pidgins and creoles. Cambridge: Cambridge University Press. Lefebvre, Claire. 1998. Creole genesis and the acquisition of grammar: the case of Haitian Creole. Cambridge: Cambridge University Press. McWhorter, John. 2000. The missing Spanish creoles: Recovering the birth of plantation contact languages. Berkeley CA: University of California Press. Smith, Norval. 1995. An annotated list of creoles, pidgins, and mixed languages. ed. by Jacques Arends, Pieter Muysken, and Norval Smith, Pidgins and creoles: an introduction, 331–74. Amsterdam, and Philadelphia, PA: Benjamins. Thomason 1997. Thomason, Sarah, (ed.). Contact languages: A wider perspective. Amsterdam/Philadelphia PA: Benjamins. Thomason, Sarah. 2001. Language contact: an introduction. Philadelphia: Georgetown University Press. Wurm, Stephen, Darrell Tryon, and Peter Mühlhäusler (eds.) 1996. Atlas of languages of intercultural communication in the Pacific, Asia and the Americas, 3 vols. Berlin: Mouton de Gruyter.

JACQUES ARENDS See also Language: Contact-overview; Lexical Borrowing

Pike, Kenneth Lee Kenneth Pike, like his predecessors, based his theoretical contribution firmly on linguistic field work. He followed their lead in attempting to explain the ways in which the languages of North America structured words, phrases, and clauses. Unlike his teachers, he added a needed emphasis on meaning and context to earlier analytical methods. Pike’s interest in linguistics began when he attended the second session of linguistics training led by William Cameron Townsend in 1935 at a remote Arkansas farm. After this rudimentary training, Pike applied his new knowledge to the study of the complex tone patterns, morphology, and syntax of the Mixtec language in

850

the highlands of Oaxaca, Mexico. The following year (1936), he taught the phonetics component of the Townsend’s course, using his personal field experience to explain phonetic concepts. His introduction to the initial writings of Edward Sapir (1921) and Bloomfield (1933) motivated him to attend the 1937 Linguistics Institute of the Linguistics Society of America, where he also met the leading American linguists of the day. He studied under Sapir from 1937 to 1942 at the University of Michigan, focusing on phonetics and descriptive linguistics. Pike was later hired by the university and taught linguistics there from 1948 to 1979.

PIKE, KENNETH LEE Pike spent much of his early efforts (1935–1948) on phonetics and phonology (phonemics) and published textbooks in these areas that were used extensively for many years. His personal field studies and his experience as a consultant to other researchers in Latin America led Pike to investigate phonetics and phonology (then called phonemics). His Phonetics (1943) was the most complete book of its kind. Pike’s phonetic studies provided insights into what he called contrast, variation, and distribution. Contrast refers to the ways in which sounds are different from one another in a particular language. Variation describes the fact that individual sounds are not always pronounced the same way, although native speakers perceive them as the same sound. Distribution deals with the locations in which a sound appears. These three themes continued throughout Pike’s writing and formed the basis of his later work on morphology and syntax as well. Pike pioneered the study of tone languages and helped many other linguists analyze the diverse tone systems of the world. One important contribution was using a constant tone pattern or frame to analyze similar tones. By keeping the frame (such as ‘This is a…’) constant, linguists could better hear the contrastive tones of the words being studied. Pike’s studies in intonation were likewise an essential tool for many scholars and people involved in teaching English to speakers of other languages. Pike applied intonation studies to poetry as well and published examples of poems (his own included) marked for intonation patterns reflecting how the poem was read by different speakers. In 1936, Pike developed a ‘monolingual demonstration’ in which he showed students how to learn a language without relying on English translation. Using a variety of sticks, leaves, and other common objects, Pike would elicit language data from a speaker of a foreign language without using an intermediary language. Within an hour, Pike would be able to explain major features of the sound system, morphology, and syntax of the language. The lecture was very popular, both on campus and in nonacademic settings. It provided a clear illustration of the role of a hypothesis and experimentation in linguistic studies. In 1977, the University of Michigan distributed a video version of one such demonstration of Pike working with a Javanese speaker. Pike is best known for his work in tagmemics, a theoretical approach to morphology and syntax that integrates language forms and grammatical functions. Based on the concepts of sounds (phonetics) and sound patterns (phonemics), Pike proposed that there were similar features in morphology (morphemes) and syntax (syntagmemes). He defined the tagmeme as a combination of particular grammatical ‘slots’ or positions

with the elements that could ‘fill’ that position. In English, for example, the subject position (slot) of a sentence can have (be filled by) a Noun Phrase, a pronoun, or proper noun, but not by an adverb or a conjunction. Eventually, Pike would add semantic information to his tagmemic model. In the previous example, Pike noted that the role or function of a Subject in English can be either the person who does the action (an Agent) or the person who is affected by the action (a Patient or Undergoer), as in a passive clause. Using principles of contrast, variation, and distribution from phonemics (phonology) and phonetics, Pike coined the terms ‘emic’ and ‘etic’ in 1954 to discuss word and clause structures. Pike’s seminal work Language in relation to a unified theory of the structure of human behavior presents his philosophy of language and culture that formed the basis of the emic/etic distinction later used in anthropology and other fields. With his wife and coteacher, Evelyn, Pike developed a unified approach for studies of morphology, syntax, and discourse at a time when other linguists limited syntax to the clause or sentence level. Their Grammatical analysis (1982) has been widely used as a textbook in descriptive linguistic methodology. Pike’s early students at the University of Michigan include Alton Becker, Ruth Brend, Charles Fillmore, John Gumperz, Ilse Lehiste, and Velma Pickett. His pioneering work in tagmemics has also been developed by Walter Cook, Soedarjanto, Linda K. Jones, and Robert Longacre, among others. Other linguistic theories overshadowed Pike’s work at times, but his teaching and writing influenced hundreds of field linguists and helped them to document indigenous languages around the world. After his retirement from the University of Michigan in 1979, Pike served as an adjunct professor of linguistics at the University of Texas–Arlington. Pike’s contribution to linguistics was overshadowed when tranformational approaches to syntax became popular. Although tagmemics handled a larger portion of the overall linguistic analysis, it did not handle such areas as active vs. passive transformations or the ways in which certain patterns can be derived from other, more basic, patterns. Pike was one of the first people involved in the Summer Institute of Linguistics (SIL), a Christian organization that sponsors linguistic and cultural research as well as literacy and translation programs in indigenous language communities. When Pike was not teaching at the University of Michigan, he split his time between his long-term research among the Mixtec people and helping his SIL colleagues. In 1951, Pike and his coworkers published a translation of the New Testament in San Miguel Mixtec, the first such translation completed by SIL. From 1936 through 2000, he was regularly involved with SIL schools, first in

851

PIKE, KENNETH LEE Townsend’s training program in Arkansas, then at the University of Oklahoma (1941–1987), and later at other locations. As President of the Summer Institute of Linguistics from its incorporation in 1942 until 1978, Pike oversaw the development of SIL’s worldwide linguistic research and training in over fifty countries. Pike personally supervised the beginnings of the SIL schools at the University of Oklahoma (1941), in Australia (1950), and in Great Britain (1951). He was also active in the Linguistic Society of America and the Linguistic Association of Canada and the United States. Pike was the author of more than 20 books and 200 articles, including articles on the interaction of his personal spiritual beliefs and his academic work. He was also a prolific poet and frequently incorporated his poems in his linguistic writing and teaching. He was an outstanding consultant who assisted other linguists to see the patterns in the languages they studied. He frequently coauthored articles with younger colleagues and was coeditor of linguistic volumes in such diverse settings as Indonesia, Nepal, and Papua New Guinea. Pike lectured in 43 countries covering five continents and received eight honorary degrees and numerous other honors including the Presidential Merit Medal from the Philippines (1974). His work with SIL led to his being nominated for the Nobel Peace Prize each year from 1982 to 1996, based on the positive impact of the linguistic, literacy, and translation efforts of SIL in over 1,000 languages.

Biography Kenneth Lee Pike was born in Woodstock, Connecticut on June 9, 1912. He received his Th.B. (1933) and Ph.D. (1942) for studies in phonetics, mentored by Edward Sapir, University of Michigan. He was a member of the Summer Institute of Linguistics, 1935–2000; President, 1942–1979; and President Emeritus, 1979– 2000. He was also Associate Professor of Linguistics, University of Michigan, 1948–1953; Professor of Linguistics, 1954–1979; Chair of Linguistics Department, 1974–1979; and Professor Emeritus 1979–2000. He was Permanent Council Member,

International Phonetic Association, 1945; member, Linguistics Society of America; and President, 1961. Pike was a member of the Linguistics Association of Canada and the United States and its President in 1977. He was also a member, American Academy of Arts and Sciences, 1973 and member of the National Academy of Science, 1985. He was nominated for the Nobel Peace Prize in 1982–1996. He received Fulbright lectureship in the USSR in 1988. Kenneth Lee Pike died in Dallas, Texas on December 31, 2000. References Brend, Ruth M. (ed.) 1972. Selected writings to commemorate the 60th birthday of Kenneth Lee Pike. The Hague: Mouton. Brend, Ruth M. (compiler) 1987. Kenneth Lee Pike bibliography. Bloomington, Indiana: Eurasian Linguistic Associa-tion. Brend, Ruth M., and Kenneth L. Pike (eds.) 1977. The Summer Institute of Linguistics: its works and contributions. The Hague: Mouton. Headland, Paul, Kenneth L. Pike, and Marvin Harris (eds.) 1990. Emics and etics: the insider/outsider debate. Newbury Park, CA: Sage. Pike, Kenneth Lee. 1943. Phonetics: a critical analysis of phonetic theory. Ann Arbor, MI: University of Michigan Press. ———. 1945. The intonation of American English. Ann Arbor, MI: University of Michigan Press. ———. 1947. Phonemics: a technique for reducing languages to writing. Ann Arbor, MI: University of Michigan Press. ———. 1948. Tone languages. Ann Arbor, MI: University of Michigan Press. ———. 1967. Language in relation to a unified theory of the structure of human behavior. The Hague: Mouton. ———. 1982. Linguistic concepts. Lincoln, NB: University of Nebraska Press. Pike, K.L., and Evelyn Pike. 1982. Grammatical analysis. Dallas, TX: Summer Institute of Linguistics. ———. 1993. Talk, thought, and thing: the epic road to conscious knowledge. Dallas, TX: Summer Institute of Linguistics. ———. 1977. Seasons of life: a complete collection of Kenneth L. Pike’s poetry, compiled and ed. by Sharon Heimbach. Huntington Beach, CA: Summer Institute of Linguistics. Pike, Eunice V. 1981. Ken Pike: Scholar and Christian. Dallas, TX: Summer Institute of Linguistics. The Summer Institute of Linguistics provides additional information about Kenneth Pike at http://www.sil.org/klp/.

PETER J. SILZER See also Mexico; Tone Languages

Pintupi and Pama-Nyungan Languages The Pintupi language is an Australian indigenous language spoken in the Western Desert, eastern Central Australia, which is situated approximately 400 km

852

northwest of Alice Springs in the Northern Territory. It is mainly spoken in and around the Papunya Indigenous Community Settlement. Pintupi belongs to a group of

PINTUPI AND PAMA-NYUNGAN LANGUAGES languages which includes Ngaanyatjarra, Pitjantjatjara/ Yankunytjatjara, Luritja/Pintupi, with the last two language groups being very similar in formation (Simpson 1993:136, 142). These languages form part of the Pama-Nyungan grammar group of Australian languages and exclude other indigenous languages from the north of Western Australia and the north and northwest of the Northern Territory (Yallop, C., p. 16). According to an Inquiry by the Australian Government in 1992, 90 of the original 250 Australian indigenous languages are still living languages and approximately 3,000 people still speak the eastern Western Desert group of languages. The Western Desert, western Central Australia group of languages include Manjiljarra, Yulparija, Kukatja, Ngaanyatjara, and Ngaatjatjarra languages, with approximately 1,000 people speaking these languages. This brings the total number of people speaking the Western Desert languages of Central Australia to approximately 5,000, making it the largest group of speakers of traditional indigenous languages in their natural environment in Australia. In 1995, Peter Trudgill wrote in his book Sociolinguistics: In Australia, for instance, there used to be about 200 aboriginal languages. Of these, 50 are already dead, and another 100 are very close to extinction. Perhaps as few as 30 will make it to the year 2000. (p. 177)

It has therefore become an important task for linguists today to do as much as they can to work toward the preservation of these endangered languages. Australian indigenous people have been either discouraged or forbidden to speak their own languages since colonization, and many Aboriginal languages have been lost to the world. These languages are ‘oral’ only and have no written form apart from drawings, such as is found in caves, in their natural settings, and it is only in recent times that linguists have been able to do more extensive fieldwork to give written form to the surviving few indigenous languages. Pintupi is one of these languages, and translators, such as Aboriginal Bible translator Ken Hansen, who lived in a Pintupi community for several years, have done significant work in this area. In his article on ‘Translating for the Pintupi’, Ken Hansen says that when he first began to learn to speak Pintupi his progress was very slow because the sounds of the indigenous language were more complex than the sounds of English. It was therefore very difficult to transcribe into a written form. For example, Hansen says the Pintupi language has words such as: (1) (2)

ngarrinpa and ngarinpa wangka and wanka

lying standing talk alive

While these words seem very similar in structure, just the addition of a longer ‘r’ sound gives a completely different meaning to the word. The sound ‘ng’ is also often used at the beginning, or in the middle, of words, making the language difficult to pronounce for English speakers because we are used to using this sound at the end of words, such as in the word ‘sing’ (Walsh and Yallop 1993:xi). Hansen also found that it was very important to use the correct words in different social situations, and because of the strange combination of sounds, this was difficult for him to do. His goals were based on the use of the Bible in traditional indigenous language and culture, and he was obviously very committed to his work. He had estimated that it would take him at least 15 years to work out how to write the language by just listening and transcribing the sounds into an alphabet, a dictionary, and a grammar, let alone have enough information about the culture to translate the Bible. There is now a Pintupi/Luritja to English Dictionary available that has been written by him. Another enormous problem in doing this type of translation is that the Aboriginal people had never heard of things such as a synagogue, and similar concepts such as ‘meeting place’ had to be used instead. The indigenous word for ‘ceremonial offering’, ‘kunatinpa’, could possibly have been used or incorporated here. Some other examples of Hansen’s Pintupi/English translations are (with literal translations and/or explanations): synagogue king angel pharisee Holy Spirit God Christ Christ

tjuwuku tjaatji the Jew’s church mayutju pulka big boss nganka ngurrara one who belongs in the sky tjuwuku luwuku mikunytju a lover of the Jew’s law katutjaka kurrunpa God’s personal spirit katutja the one who pertains above kirritja Christ pronounced in the Pintupi way katutjalu tjamatatjunkula wantirriyantja the one whom God promised to send long ago

As can be seen from these translations, the effect of British colonization has been that many words that originally come from English origins would not have been in existence in indigenous language prior to this event. They have since been added to the Australian indigenous language and culture. This is particularly

853

PINTUPI AND PAMA-NYUNGAN LANGUAGES the case with words such as ‘big boss’, because many indigenous men became stockmen on cattle properties, and the owner was called the ‘big boss’, ‘mayutja pulka’, which was then transcribed into Traditional indigenous language. This is a concept that is important to modern social use of indigenous language and has meant the creation of new Aboriginal word usage (Simpson 1993:137–78). One of these words is the Western Desert word for ‘spirit’, ‘kardiya’, which is now also used to refer to ‘white people’. As can be seen in modern word usage in indigenous language is the current, and extremely significant, social event taking place in Australia today, the Reconciliation Process, and the use of the word ‘sorry’. The Indigenous people are asking the Australian Government and the Australian people as a whole to say ‘sorry’ for the past injustices that have been done to the Australian Indigenous people (Haebich 2000:569). These injustices include taking the indigenous children from their natural parents and communities and placing them with white families ‘for their own good’ (Haebich 2000:567–8). Although the present Australian multicultural population was not specifically involved in perpetrating these injustices, it is seen as a token of respect toward the indigenous people, and an acknowledgement of the shame that the people who were responsible for these injustices should feel, for committing these acts. The associated feelings for this can be incorporated into words such as ‘kunta’, or ‘shame’, by saying ‘sorry’ to what is now left of the indigenous community. Part of the meaning of the Pintupi word ‘kunta’ relates to the English words and concepts for ‘shame’ and ‘respect’. When used to mean ‘respect’, it is usually symbolic of ‘shyness’ in not wanting to show disrespect, such as refusing a person to his face without excuses. Someone who is being respectful of others shows embarrassment by saying whatever they have is less significant than the other persons’ possessions, and this type of shame is called ‘kuntarrinpana tjanampa’. Children are socialized by shaming, and a child is said to be deaf or unheeding, ‘patjarru’, when young, but needs to be respectful when older and to be aware of ‘kunta’ as they learn of the need to show respect within the community (Mawr, B:126–157). These emotional concepts have a great significance in the Pintupi understanding of social behavior and it could be further said that this concept is of importance to the indigenous people as a whole. It can then be an integral part of the Reconciliation Process of saying ‘sorry’, therefore incorporating an important indigenous concept of being sorry for the bad things that someone has done within the indigenous concept of their own culture.

854

The following are Pintupi concepts with their English meanings relating to feelings: kunta rama ngulu rarru pukulpa ngarru

shame mad fear anger happy happy

These concepts have been researchers, such as Meyers, as:

interpreted

by

(1) ‘Mad, crazy persons are said to be “deaf”, “rama”, because they take no notice of what is being said to them. “Rama”, or “not hearing”, therefore relates to people who do not take notice of advice and are therefore not thinking.’ (2) ‘The English and Pintupi concepts of “fear” or “ngulu” are the same and generally mean to be frightened or afraid of the consequences of one’s actions. Men’s sacred objects and ritual paraphenalia are often described as “ngulu”, or “frightening and dangerous”.’ (3) ‘“Anger” or “rarru” means to “get wild” or “very angry” because he does not feel sorry for his actions. Pintupi descriptions of “rarru” means that their “ears are closed” and they are not thinking of the consequences of their actions.’ (4) ‘When a relative comes to visit the Pintupi say one becomes happy, or rejoices, “pukularrinpa”. Or men bringing back meat make people happy. It is the opposite of anger, fighting, and sorrow, and is tied to a major Pintupi image of sociality and cooperation, says Meyers. Ceremonies that include song and dance make people happy, and “pukulpa” was traditionally used to stop fighting with an approaching enemy.’ Another way in which indigenous language use is being preserved in Australia today is in bilingual education programs in schools. Pintupi/Luritja is being taught bilingually in Walungurra School, Northern Territory, since 1983, with Luritja language being taught at Papunya (Black 1983:207–20). Another place of bilingual education is of the Diwurruwurrujarru Community, where the Walpiri language, (also a Pama-Ngungan language) is being taught at Lajamanu, Yuendumu, Nyirripi, and Willowra schools, near Katherine in the north of the Northern Territory (Bavin 1993:85). Kriol is being taught at Barunga School, also in this area (Rhydwen 1993:155–8). In some of these communities, linguist/teachers work alongside indigenous women while the children are taught traditional ways of food gathering. The names

PITCAIRNESE of ‘bush tucker’ (food), including plant and animal names, and names for ‘body parts, are taught in Kriol and Walpiri Language by the Indigenous women in communities near Katherine, and the linguist/teacher teaches the same names in English. In other bilingual teaching programs, everyday classes are taught in both traditional indigenous language, and English. A significant problem that arises because indigenous language is an oral not written language, is that new words have to be invented to explain how to write things such as commas and full stops, and also in the teaching of mathematics. This is how and where linguists and teachers are working together with indigenous people to preserve Australia’s original languages. It is especially important that this work be done as quickly as possible so that indigenous languages are not only lost to their original users, but to the world as a whole as globalization and the common use of English as the predominant language take over. References Amery, Rob. 1986. Languages in contact: the case of Kintore and Papunya (with particular reference to the Health Domain). Language in Aboriginal Australia 1. Bavin, Edith. 1993. Language and culture: socialisation in a Walpiri community. In Walsh and Yallop. Black, Paul. 1983. Aboriginal languages of the Northern Territory. Batchelor, NT: School of Australian Linguistics, Darwin Community College. Blake, Barry. 1991. The handbook of Australian languages. Cambridge: Oxford University Press. Devlin, Bruce. 1990. Some issues related to vernacular language maintenance: a Northern Territory view. ed. by, C. Walton and W. Eggington. Language maintenance, power and education in Australian Aboriginal contexts, Darwin: Northern Territory University Press.

Dixon, R.M.W. 1980. The languages of Australia. Cambridge: Cambridge University Press. Eades, Diana. 1993. Language and the law: White Australia v Nancy. In Walsh and Yallop. Haebich, Anna. 2000. Broken circles: fragmenting indigenous families 1800–2000. Fremantle, Western Australia: Fremantle Arts Centre Press. Hansen, Kenneth. 1977. Pintupi/Luritja dictionary, 2nd edition. Alice Springs, Northern Territory: Summer Institute of Linguistics, Australian Aboriginals Branch. Hansen, Kenneth C. 1980. Translating for the Pintupi. Nungalinya Occasional Bulletin No 22. Darwin, Northern Territory: Nungalinya Aboriginal and Torres Strait Islander Church Training and Research Centre, University of the Northern Territory. House of Representatives Standing Committee on Aboriginal and Torres Strait Islander Affairs. June 1992. Language and culture — a matter of survival. Report of the Inquiry into Aboriginal and Torres Strait Islander Language Maintenance. Australian Government Publishing Service. Meyers, Fred R. 1976. Cultural concepts of the emotions. From: to have and to hold: A study of persistence and change in Pintupi social life. Ph.D. thesis. Bryn Mawr. Myers, Fred R., 1986. Pintupi country, Pintupi self: sentiment, place and politics among Western Desert Aborigines. Washington, DC: Smithsonian. Rhydwen, Mari. 1993. Kriol:the creation of a written language and a tool of colonization. In Walsh and Yallop. Schmidt, Annette. 1990. The loss of Australia’s Aboriginal language heritage. Canberra: Aboriginal Studies Press. Simpson, Jane. 1993. Making dictionaries. In Walsh and Yallop. Tindale, Norman. 1974. Aboriginal tribes of Australia. Los Angeles: University of California Press. Trudgill, Peter. 1995. Sociolinguistics : an introduction to language and society. England: Penguin Books. Walsh, Michael, and Yallop, Colin, (eds.) 1993. Language and culture in Aboriginal Australia. Canberra, Australia: Aboriginal Studies Press. Wierzbicka, Anna. 1997. Understanding cultures through their key words. New York: Oxford University Press.

JOANNE MARIE BAUMGARTNER

Pitcairnese Pitcairnese is the name of one of two closely related varieties of a language initially formed on Pitcairn island and subsequently transplanted to Norfolk island by descendants of the nine English-speaking sailors who mutinied on HMS Bounty in 1789 and their Polynesian companions from Tahiti and Tubuai, one of the Austral Islands of the Tahiti group. Linguists use the name ‘Pitcairn-Norfolk’ (PN) to refer to both, but speakers call the variety spoken on Pitcairn ‘Pitkern’ and that on Norfolk Island ‘Norfolk.’ In 1996, Pitcairn Island Council, prompted by the concern that Pitkern

might die out because none of the children were speaking it spontaneously, declared it an official language of the island. Although there is some disagreement about whether PN is a creole rather than a variety of English, the conditions under which it developed were clearly exceptional because we know the place and time of its formation, as well as details of the origins of the primary individuals involved. PN is one of the few cases in which it is possible to directly trace the survival of some British dialectal features to some of the sailors.

855

PITCAIRNESE The Bounty, sent to Tahiti to transport breadfruit trees to the West Indies as a source of food for slaves, contained no pressganged men and was probably the first British ship to sail with an all-volunteer, handpicked crew of 45. The cliché ‘motley crew’ well describes the mutineers, for they covered not only a wide social spectrum ranging from their leader, Fletcher Christian, master’s mate, born of an aristocratic family originally from the Isle of Man, to common seaman John Adams, a Cockney and orphaned son of a Thames waterman, brought up in a poor house, but they also covered a wide range of regional varieties of English. Four of the nine mutineers were English, two Scottish, one American (from Philadelphia), one partWest Indian from St. Kitts in the Caribbean, a welleducated son of a navy captain (Edward ‘Ned’Young), and one, John Williams, assistant armorer, was a native of Guernsey and spoke French, as Captain William Bligh noted in his list of identifying traits of his men. Of Peter Heywood, he remarked that he spoke with the ‘Manx, or Isle of Man’ accent. In their search for a place to hide from the British navy, the mutineers initially sailed back to Tahiti and to Tubuai, where they picked up six men and 12 women, one with a daughter, before settling in 1790 on a remote Pitcairn island more than 2,000 kilometers from Tahiti. This small, two- by one-mile, harborless island with its steep cliffs and lack of reefs or anchorage was at the time uninhabited and incorrectly charted on contemporary maps. Pitcairners’ first contact with the outside world occurred 18 years after settlement, in 1808, when Captain Folger of the Topaz from Boston was greeted in English by Thursday October Christian, Fletcher Christian’s son, and was asked whether he knew Captain Bligh. Apart from this brief encounter, the settlers lived there for the first 33 years in almost complete isolation, thus permitting no outside influences during the formative period of the language. It was not, however, a tranquil time by any means, with nearly 90% of the founding male population murdered by 1800. Disputes over land and the women led to the violent deaths of all the men except four Polynesians and four mutineers, one Englishman (John Adams), a Scot (William McCoy), one Cornishman (Matthew Quintal), and one West Indian from St. Kitts (Edward Young). Young, McCoy, and Quintal died by 1800, leaving only John Adams, then 33 years old, who remained with nine surviving women and the 23 children born on the island. Outsiders did not enter the colony until the 1820s. The first separation in a community that had lived together as a family for more than 60 years occurred in 1856, when the 194 Pitcairners were resettled on Norfolk Island, originally a penal colony. Eighteen months after their arrival, 17 members of the

856

Young family sailed the 3,700 miles back to Pitcairn. Five years later, a second party returned. Today, fewer than 50 speakers remain on Pitcairn Island. Presentday Norfolk is spoken by about 700 people; the majority of the island’s population is not descended from the original Pitcairners. Some 700 other descendants of the original population live in Australia, New Zealand, and elsewhere. Because the mutineers were outnumbered by two to one, the odds were that Pitcairn would become predominantly Polynesian in language and culture, particularly because all the women were Polynesian (most of them from Tahiti) and would have been the primary caregivers to the children. There are also hints in both boatswain’s mate James Morrison’s and midshipman Peter Heywood’s accounts of the mutiny that the sailors had begun to intersperse Tahitian words in their speech to one another. Midshipman Roger Byam recalls that by the time the Bounty sailed with the by now heavily tattooed sailors after a stay of more than five months in Tahiti, all the men knew some Tahitian words. Some were said to be quite fluent and could be heard carrying on conversations with one another in which barely a word of English was spoken. Yet, the language of power on Pitcairn island was clearly English, and many linguists argue that PN was not a fully developed pidgin or creole because there was no point in its history in which English was not the primary language of the community. Stable pidgins rarely develop where contact involves only two languages. Moreover, the descendants seem to identify more strongly with their English rather than Tahitian lineage, and the predominant influence on the vocabulary, pronunciation, and sentence structure of PN is English. The main influence of Tahitian survives in the form of a large element of Tahitian vocabulary, particularly for flora, fauna, and foodstuffs, e.g. buhe, ‘sea eel’ (Tahitian puhi), and miti, ‘coconut cream sauce’ (Tahitian miti). Tahitian influence can also be seen in certain semantic (meaning) shifts that have affected English words, such as hand, meaning both ‘hand’ and ‘arm’ in PN, paralleling Tahitian rima. Otherwise, most of the vocabulary is ultimately of English origin, including a number of dialects. Words can sometimes be traced to the dialects of individual mutineers, e.g. the negative imperative (‘do not’) dune from the Scots probably spoken by John Mills, gunner’s mate from Aberdeen, or William McCoy of Rosshire, or moge, ‘thin’ (and possibly also the negative form kaa/kannt, ‘cannot’), from Edward Young’s West Indian English. Other Scots forms include bole, ‘to make a small hole in anything’, devil’s needle, ‘dragonfly’, gaggle, ‘to cackle’, and possibly tayte, ‘potato’. There are also archaisms such as dub, ‘to square and smooth (timber)’, tardy, ‘late’, paunch, ‘stomach’, and a few words of American origin, such as corn, ‘maize’,

PITCAIRNESE and candy, ‘sweets’, possibly because of the presence of Isaac Martin, a mutineer from Philadelphia, or because of the influence of American whalers who visited the island. Other dialect forms include dunnekin, ‘lavatory’, and grub, ‘food’. A variety of terms with original nautical reference, such as all hands, ‘everyone’, deck, ‘floor’, and flog, ‘spank’, have extended their meanings in PN. Similarly, heave ho, a cry of sailors in raising the anchor up, survives in PN heway me (heave me away) ‘to lift up/ throw/take away/ remove’, as does heway (‘heave away’), now obsolete in English. However, this once nautical usage with the meaning ‘to fling, throw or cast, to haul up or raise by means of a rope’ does exist in dialectal usage in Scots, where heave away means ‘to throw away’, as well as in some English dialects in which heave means ‘to lift/raise’. Although PN was formed independently of Pacific Jargon English, an English-based foreigner talk used in encounters between Europeans and Pacific islanders and found on almost all islands in Polynesia and Micronesia by the 1830s because of the whaling industry, it shares some features in common with other Pacific pidgins and creoles descended from it, e.g. bimorphemic (i.e. two-part) question words such as whatawe, ‘how’ (English ‘what way’), verbs generally unmarked for tense (past, present) and aspect (whether the action is completed or not), use of dem (English ‘them’) as a plural marker, e.g. dem ai, ‘the eyes’, and absence of the copula ‘be’, e.g. whatawe hem, ‘how is he?’ The sailors probably knew some rudiments of Pacific Jargon English with adaptations made during their stay at Tahiti. These features could also have been acquired through contact with visiting whalers. Between 1813 and 1852, 400 visits of ships are recorded, three quarters of which were American, mostly whaling ships. For a time, there was a steady market for the islanders’ vegetables. Norfolk islanders also had contact with the whaling trade in the late nineteenth and early twentieth centuries. Many worked as crewmen on whaling ships. They were also in contact with the varieties of Melanesian pidgin spoken by Solomon Islanders and New Hebrideans. In 1867, the Melanesian Mission established a training school for Solomon Islanders on Norfolk. From the 1890s until World War II, Norfolk Islanders worked as plantation managers and cadets in the Solomons. Otherwise, however, PN has far more in common with the Caribbean creoles, no doubt attributable to the influence of Edward Young from St. Kitts. The possessive construction makes use of English ‘for’ as in dem ai fo yoen, ‘your eyes’. A few grammatical features, particularly the personal pronouns, reflect Tahitian influence, e.g. the first-person inclusive dual hemi (‘you and I’), a literal

translation of Tahitian taua, composed of English ‘thou’  ‘me’, with the initial sound of ‘thee’ changed to /h/. This pronoun contrasts with uklun (‘we/us plural’), whose etymology is unclear. PN, like Tahitian, has dual marking in forms such as yu tu, ‘you two’, contrasting with yoli, ‘you plural’, i.e. three or more. This may also reflect archaic English ‘all ye’ or ‘you all’. The sound system of PN also shows some influence from Tahitian. Lack of a significant distinction between /r/ and /l/, a feature common to many Polynesian languages, including Tahitian, is found in PN forms such as stolly, ‘lie’ (English ‘story’), morla, ‘tomorrow’, and tomolla, ‘day after tomorrow’ (both from English ‘(to)morrow’). This feature (along with variation between /v/ and /w/, also similar to Tahitian) is a feature typical of St. Kitts too, which may suggest additional influence from the speech of Edward Young. Unlike many pidgins and creoles, PN has preserved consonant clusters under some circumstances, whereas Tahitian has none. However, final t/d are often absent in PN, a common feature of colloquial and dialectal English. PN pronunciations such as /flaid/ for ‘Floyd’ and /kloi/ for ‘cry’ may reflect a sound shift that probably originated in London and was well under way by the mid-nineteenth century. It is possible that John Adams is the origin of this feature in PN, although it could also have arisen later by independent innovation or contact with other varieties, such as the one spoken by Young. The vowels of words such as price and choice still rhyme in St. Kitts today in many people’s usage. The /ie/ diphthong (double vowel sound) of gate could reflect West Indian or English dialectal influence. Despite the presence of two Scots among the founding English speakers, PN has no trace of /r/ after vowels in words such as cart, barn, etc. References Buffett, Alice, and Donald C. Laycock. 1988. Speak Norfolk today. Norfolk Island: Himii. Dening, Greg. 1992. Mr. Bligh’s bad language. Passion, power and theatre on the bounty. Cambridge: Cambridge University Press. Harrison, Shirley. 1972. The language of Norfolk Island. B.A. Thesis. Sydney: Macquarie University. Harrison, Shirley. 1985. The social setting of Norfolk speech. English World-Wide 6(1). 131–53. Laycock, Donald C. 1990. The status of Pitcairn-Norfolk: creole, dialect, or cant. Status and function of languages and varieties, ed. by Ulrich Ammon. Berlin: de Gruyter. Mackaness, George. 1938. A book of the ‘Bounty’ and selections from Bligh’s writings. London: J.M. Dent & Sons, Ltd. Ross, A.S.C., and Moverley, A.W. 1964. The Pitcairnese language. New York: Oxford University Press.

SUZANNE ROMAINE See also Pidgins and Creoles

857

PLACE OF ARTICULATION

Place of Articulation Consonants are formed by creating a constriction in the vocal tract; in producing a consonant, the place of articulation is the part of the vocal tract having the greatest constriction. The symbols used for the description of sounds are taken from the IPA chart, which lists the International Phonetic Alphabet. The major constriction, which may be a complete or partial closure, is made by the articulators coming together, generally by a lower articulator moving toward an upper articulator. The lower articulators are elements of the lower jaw—the lower lip, the lower teeth, and the tongue. The upper articulators are the upper lip, the upper teeth, the palate, the velum, the uvula, and the rear wall of the pharynx. Places of articulation usually have a compound name giving the lower and upper articulators, with the name of the lower articulator first. Thus, apico-dental indicates that

the lower articulator is the apex of the tongue and that the upper articulator is the upper teeth. Occasionally, when the lower articulator is obvious or unimportant, only the upper articulator is named: e.g. velar, used alone, is interpreted as meaning dorso-velar. The major places of articulation are shown in Figure 1. Places of articulation can be divided into four general categories: labial, coronal, dorsal, and guttural. At each place of articulation, sounds with different manner of articulation can be produced: stops have complete closure at the indicated place, whereas fricatives are pronounced with partial closure, which produces a hissing sound. Labial sounds are made with one or both lips and include bilabials, labio-dentals, and linguo-labials. The lower lip articulates with the upper lip to form a bilabial consonant. The term bilabial is used rather

Palatal Velar

Bilabial

Postalveolar Alveolar Dental

Uvular

Labio-dental Pharyngeal

Figure 1

858

PLACE OF ARTICULATION than labio-labial. Bilabial stops [p, b] and the nasal [m] are extremely common in languages; indeed, a language without them (e.g. some Oto-Manuean languages) is noteworthy. The bilabial fricatives [ϕ, β], however, are rather uncommon; languages tend to have bilabial stops and labio-dental fricatives. The lower lip articulates against the upper teeth to form labio-dental consonants. Labio-dental fricatives [f, v] are very common. Labio-dental stops do not occur as distinctive sounds in any language. The labio-dental nasal [(] occurs as a variation of [m] in English words ) such as symphony [ si(fəni], in which it takes the place of articulation of the following labio-dental fricative. Linguo-labial sounds are very rare sounds made with the tip or blade of the tongue articulating with the upper lip. The IPA uses a diacritic [ ] underneath a labial symbol for this place of articulation: . Coronal sounds are made with the apex, or lamina, of the tongue. They include dentals, alveolars, alveolo-palatals, postalveolars, and retroflex sounds. Dental sounds can be made with either the tip of the tongue (apico-dental) or with the blade (lamino-dental). The dental fricatives are voiceless [θ] and voiced [ð]; these sounds both occur in English as in the words thin and then, respectively. Other dental sounds are shown by using the alveolar symbol with the diacritic [' ]; e.g. [t' d' n' 'l r' ]. Lamino-dental fricatives [s' z' ] can be made with the blade of the tongue near the back of the upper teeth, in addition to the apico-dental fricatives [θ ð]. Danish has a dental approximant [ð +]. Interdentals are made by thrusting the tongue slightly forward so that the tip protrudes between the teeth; they can be symbolized as [θ, ð,]. In English, apico-dentals are normal, although some people occasionally use interdentals in emphatic or exceptionally careful speech. Like the dentals, alveolar sounds may be made with either the tip or blade of the tongue, known accordingly as apico-alveolar or lamino-alveolar. The symbols for either are [t d n s z]; all of these sounds occur in English. Alveolar and dental sounds are both extremely common in languages; it is rare, however, for a language to have both, except for fricatives. Postalveolar sounds are made with the blade of the tongue articulating with the area at the back of the alveolar ridge. The fricatives [ʃ ] and affricates [tʃ d] are quite common in the languages of the world. English has these sounds in shin [ʃ], leisure [], chin [tʃ], and gin [d]. The alveolar fricatives [s z] and the postalveolar fricatives [ʃ ] involve some difficulty in describing their point of articulation accurately. Different speakers hold their tongues in somewhat different positions. With the alveolars [s z], the airstream hits the back of the upper teeth. This causes a turbulence that gives them a distinctive sound. For [ʃ ], the air stream hits

the back of the lower teeth and produces a slightly different sound. Two fricatives are given in the International Phonetic Alphabet as alveolo-palatal; these are different from the postalveolars. They are produced with the tip of the tongue near the lower teeth and with the blade quite close to the back of upper teeth. The symbols are [-] for the voiceless fricative and [.] for the voiced fricative. These sounds occur in Mandarin Chinese and in Polish. For retroflex consonants, the tip of the tongue curls back, and the underside of the tip articulates with the area at the back of the alveolar ridge. The body of the tongue is quite concave. Retroflex consonants are symbolized as [/01 2 34]; they use the alveolar symbols modified by a lower hook. These sounds are common in languages of India. The English r sound is usually called retroflex; however, in practice, the place of articulation may vary from alveolar [ɹ] to retroflex [5]. Further, many speakers of English use a bunched r, in which the tip of the tongue is near the lower teeth and the body of the tongue is pulled up and back toward the palate. Although the articulations of the retroflex and bunched r are very different, the acoustic effect is quite similar. Dorsal sounds include those made with the front or back of the tongue: palatal, velar, and uvular. With palatal sounds, the front of the tongue articulates with the palate. The tip of the tongue points down, often touching the lower teeth. Italian has a palatal lateral [ʎ], as in figlio, and a palatal nasal [1], as in signore; German has a voiceless palatal fricative [ç], as in ich. The palatal stops [c 7] and the voiced fricative [ 8] are not common. Dorso-velar sounds are made with the dorsum of the tongue articulating with the velum. The dorso-velar stops are [k "] as in English could, good; the dorsovelar nasal is [ŋ] as in English sing. The fricatives are [x], as in German Bach, and [γ], the voiced equivalent. Frequently, these sounds are simply called dorsals. The uvular consonants are made with the dorsum of the tongue articulating with the uvula. They are like velar sounds made very far back in the mouth. The symbols are [q] and [g] for the stops, [n] for the nasal, and [χ] and [ʁ] for the fricatives. Uvular sounds are found in Arabic and Inuktitut. Some French accents have an approximant [ʁ +] or fricative [ʁ] for the r sound. The guttural sounds include the pharyngeal, epiglottal, and laryngeal places of articulation. Pharyngeal consonants are made by moving the root of the tongue backward toward the pharyngeal wall. The root of the tongue is the vertical part, forming the forward wall of the pharyngeal cavity. Many people cannot make a complete pharyngeal stop. A pharyngeal nasal stop is physiologically impossible

859

PLACE OF ARTICULATION because, if no air flows through the pharynx, then no air can flow through the nasal passage. For the fricatives, the symbols are [©] and ; these sounds are found in Arabic and other Semitic languages. Epiglottal sounds are extremely rare, but they occur in some Semitic and Caucasian languages. For them, the epiglottis moves backward to articulate with the pharyngeal wall. The symbols are for the stop and for the voiceless and voiced fricatives. Laryngeal sounds involve the vocal folds; they include the glottal stop and the fricatives [h] and [ ]. From a phonological point of view, these sounds can all be thought of as consonants with a laryngeal place of articulation, with the glottal stop as a laryngeal stop and the sounds [h] and [ ] as laryngeal fricatives. From a purely phonetic point of view, however, each of these could be categorized differently: as a state of the glottis involving complete closure of the vocal folds, [h] as a voiceless vowel, and [ ] as a breathy voiced vowel. Glottal stop and [h] are very common; [ ] is somewhat rare, although it occurs between vowels in English, as in ahead. In general, the major categories of place of articulation (labial, coronal, dorsal, and guttural) are independent of each other. Thus, it is possible to make two stops at the same time, e.g. a dorsal [k] and a labial [p]. Such a sound is said to have a double articulation; for example, with [kp], the place of articulation is labialvelar. It is symbolized as [kp], with the tie-bar showing that the [k] and the [p] are made simultaneously. English [w] is a doubly articulated labial-velar approximant: the lips are rounded and the dorsum of the tongue is partially raised toward the velum. Many English speakers pronounce a double articulation in kitten . Swedish has a doubly articulated voiceless postalveolar-velar fricative [], velarization [t?], and pharyngealization . The IPA has alternative symbols for either velarization or pharyngealization . Labialized sounds involve lip rounding; they are quite common. Palatalized consonants have the front of the tongue more toward the palatal region than usual. Palatalized consonants are common in some Slavic languages. Velarized consonants have the tongue toward a position that is more velar than usual, i.e. a high back unrounded vocalic position. The English dark [@], occurring at the ends of words such as ball, is an example of a velarized lateral; here, the dorsum of the tongue is raised slightly toward the velum. Pharyngealized sounds are found in Arabic; they involve a retraction of the root, thus effecting a narrowing of the pharynx. References Catford, J.C. 1977. Fundamental problems in phonetics. Bloomington, IN: Indiana University Press. Catford, J.C. 1988. A practical introduction to phonetics. Oxford: Oxford University Press. Gimson, A.C. 1980. An introduction to the pronunciation of English. London: Edward Arnold. Hardcastle, William J., and John Laver (eds.) 1997. The handbook of phonetic sciences. Oxford: Blackwell. International Phonetic Association. 1999. Handbook of the International Phonetic Association. Cambridge: Cambridge University Press. Ladefoged, Peter. 1993. A course in phonetics. Ft. Worth, TX: Harcourt Brace and Jovanovich. Ladefoged, Peter. 2000. Vowels and consonants: an introduction to the sounds of languages. Oxford: Blackwell. Ladefoged, Peter, and Ian Maddieson. 1996. Sounds of the world’s languages. Oxford: Blackwell. Laver, John. 1994. Principles of phonetics. Cambridge: Cambridge University Press. Minifie, Fred, Thomas J. Hixon, and Frederick Hixon. 1973. Normal aspects of speech, hearing, and language. Englewood Cliffs, NJ: Prentice-Hall. Pullum, G.K., and W.A. Ladusaw. 1996. Phonetic symbol guide. Chicago: University of Chicago Press. Rogers, Henry. 2000. The sounds of language: an introduction to phonetics. Harlow, England: Pearson.

HENRY ROGERS See also Anatomy of the Articulatory System

Plurality The majority of languages have singular and plural forms of words to indicate a single object or a set of more than one object. Some languages (Philippine lan-

860

guages, some languages of Southern Asia, Baltic languages) have a dual form of words to indicate a set of two objects. Dual nouns are used in archaic dialects of

PLURALITY Lithuanian, and some relics are still known in Russian, Ukrainian, and some other Slavic languages. For example, in Russian the word eye has the form: gπaA [glaz] (singular), gπaAa [glaza] (plural). The plural form is actually a relic of the dual. The origin of the dual in these languages can be explained by the fact that some things exist only in a pair: like ears, eyes, knees, legs, and sides. In most of the modern languages, the dual form has disappeared. Only two categories, i.e. the singular and plural (more than one object) are used; for example, the singular word leg in English takes the plural form of legs; in Russian it would be HOg a [noga]—HOg u [nogi]; and in Turkish the same word would be bacak in singular and bacaklar in plural. There are also languages that do not have a specific form to indicate plural objects. This can be illustrated by examples from Malay, where a noun mata (eye) would normally imply a plural meaning (there are usually two eyes on a face). The same principle would apply to kaki (legs). When necessary, the plural form in Malay is created by repeating the noun—kamuskamus (dictionaries), kanak-kanak (children). The plurality of nouns and pronouns as objects in the sentence affects subject–verb agreement. For example, the singular English pronoun I requires a singular verb—I have a beautiful vase for these flowers. While the plural form of I–we requires a plural verb—We have a beautiful vase for these flowers. The singular noun sister requires a singular verb—My sister goes to the movies every Sunday. The plural sisters requires a plural verb—My sisters go to the movies every Sunday. The following English nouns are grouped according to objects they refer to: (1) Sister, house, table, book, mother, minister, teacher, doctor; (2) Milk, coal, chocolate, mud, water, gold, information, sugar, cream; (3) Love, hatred, friendship, hope, charity, justice; (4) Family, group, committee, class, parliament. The first group refers to concrete, tangible objects, or persons. They are count nouns. The second and third groups of nouns indicate substance (2) or abstract (3) nouns that are not discrete, and therefore cannot be counted. They are mass nouns. The last group represents entities or bodies that consist of a number of members. These nouns can be counted. They are collective nouns.

Count Nouns Count nouns can be preceded by the word some or the articles a, an, and the. For example: A tree has fallen

during the storm. The tree has fallen during the storm. Some trees have fallen during the storm. They can have both singular and plural forms. Most of the count nouns form their plurals by adding -(e)s. However, there are a number of irregular words, and words borrowed from other languages, from Greek and Latin in particular, that create their plurals in specific ways (see examples below). There are also some English words that can only be either singular or plural. The plural form of count nouns is formed by adding -(e)s to the noun: flower lady balloon lunch bush fairy potato

flowers ladies balloons lunches bushes fairies potatoes

A number of irregular nouns in English form their plurals differently, i.e. by changing their vowel or by adding -en. man woman foot goose mouse louse child ox

men women feet geese mice lice children oxen

Those count nouns, which end with -f, form plurals by changing -f to -ves, for example, wolf wife scarf

wolves wives scarves

Words borrowed from other languages, Latin and Greek in particular, create their plurals according to the foreign (original) rule, for example: datum stratum larva criterion

data strata larvae criteria

Some English nouns are used only in their singular form, such as music, chemistry, physics, ceramics, mathematics, etc. At the same time, there are nouns that are only used in the plural: jeans, trousers, scissors, spectacles, measles, etc. However, the language allows for the following phrases if there is a need to indicate the number of some of the above-mentioned words. For example—Tanya knew she desperately needed another pair of jeans. The office assistant has ordered two pairs of scissors for the department.

861

PLURALITY There are cases of nouns that do not change for plural, e.g. sheep, deer, fish, salmon, etc. Also, there are words that have two different plural forms that convey different meanings, such as: craft:

craft (boats) crafts (handicrafts) indexes (alphabetical lists of names, subjects, etc.) indices (algebraic exponents) mediums (at spiritual séances) media (newspapers, television)

index:

medium:

Collective Nouns Collective nouns are the nouns that indicate a group, a body, or an organization that consists of a number of members, for example family, class, parliament, committee. They form their plurals in the same way the count nouns do by adding (e)s: families, classes, parliaments, committees. The fact that a collective noun refers to a number of people does not affect the subject–verb agreement in a sentence. Such words in their singular form are used with singular verb, e.g.—My family enjoys skiing. The Social Club Committee usually plans the recreational activities of the employees.

Mass Nouns Mass nouns are not usually used with the definite article; for instance—Loyalty is his most valuable characteristic. Sugar adds sweetness to any drink. However, in the following examples the mass nouns are modified by further information and can be

used with the definite articles—The love John and Matt shared for their mother finally ended the war between the brothers. The chocolate made in this town is delicious. Since the mass nouns refer to abstract, not discrete objects, they cannot be counted. Therefore, they usually do not allow for plural morphology. However, they allow for other means of measurement: A bottle of milk; Three cups of tea; Bars of chocolate; Patches of mud; A useful piece of information. The appearance of such words as sugars or lights is possible only if they refer to concrete, discrete entities; for example: Switch off the lights please. Can I have two sugars in my coffee please? References Dummet, Michael. 1993. Grammar and style for examination candidates and others. London: Duckworth. Haegeman, Liliane, and Jacqueline Gueron. 1999. English grammar and generative perspective. Oxford, UK: Blackwell Publishers. Horan, Ronald S. 1996. Exploring English: a complete analysis of modern grammar. Leichhaardt, Australia: Honeysett Press. Hudson, Richard. 1984. Word grammar. Oxford, UK: Blackwell. Johansson, Stig. 1980. Plural attributive nouns in present-day English. Lund: CWK Gleerup. Bach, Emmon, et al. (ed.) 1995. Quantification in natural language, Vol. II. Dordrecht: Kluwer Academic Publishing. Reid, Wallis. 1991. Verb and noun number in English: a functional explanation. London: Longman.

ALFIA ABAZOVA

Polish and West Slavic Languages The present-day West Slavic language area is made up primarily of the national territories of Poland, the Czech Republic and Slovakia, as well as a small part of south-eastern Germany. By the end of the sixth century CE, migrations of Slavic-speaking people had settled in a wide area of northwestern Europe, and these communities extended as far west as the River Elbe. By the ninth century CE, distinctly West Slavic elements had become noticeable within the Common

862

Slavic used by these settlers, and, by the end of the tenth century, features distinctive of the three West Slavic subdivisions of Lechithic, Czecho-Slovak, and Sorbian had emerged. Apart from one or two isolated remaining pockets of Slavic, the more westerly Slavic-speaking areas were gradually resettled by German speakers in the middle ages, resulting in approximately the West Slavic language area as it exists today.

POLISH AND WEST SLAVIC LANGUAGES The modern West Slavic languages are Polish, Czech, Slovak, and Upper and Lower Sorbian. To these is also often added a further member, Cassubian, whose status is disputed between its being a separate language or a dialect of Polish. Until the mid-eighteenth century, another West Slavic language, Polabian, also existed. Polish, Cassubian, and Polabian make up the Lechithic subgroup of West Slavic, Czech, and Slovak make up the Czecho-Slovak subgroup, and Upper and Lower Sorbian make up the Sorbian group. Other divisions have been suggested in the past, involving the merging of Sorbian with either Lechithic or Czecho-Slovak, but this three-group model is the most widely accepted. West Slavic is distinguished from South and East Slavic primarily on the basis of phonetic differences, such as the retention of /tl/ and /dl/ clusters, which have been reduced to /l/ in South and East Slavic languages. West Slavic languages also have fixed word stress, while most other Slavic languages have moveable stress. All the West Slavic languages use a Roman-based alphabet, with each language having a set of accented and/or special characters. Polish has the largest set of special characters, including ‘hooked’ vowel characters that represent nasalized vowels. Two features of the West Slavic languages have been of particular general interest to linguists. Both are shared with other Slavic languages, but one is especially closely associated with work on West Slavic. The first feature is the category of aspect, which stands alongside tense to indicate whether an action or event is completed or ongoing, as opposed to whether its occurrence is in the past, present, or future. This is evident in the existence of pairs of related verbs, such as the Lower Sorbian cytas´ and docytas´, where cytas´ focuses on the general act of reading, while docytas´ is used for a completed reading session (meaning roughly ‘read through to the end’). The other feature is word order within the sentence. All the West Slavic languages mark grammatical functions by the use of special word endings, rather than by word order as is the case in English. They thus permit considerable flexibility in word order, and they use this to give prominence to particular pieces of information within the sentence. This phenomenon has sometimes gone by the name of ‘functional sentence perspective’ and its theoretical status within general linguistics originated with work on Czech by the Prague School in the 1930s.

Polish and Cassubian Polish has the largest number of speakers of the West Slavic languages. It is the official language of the Republic of Poland and is thus spoken by around 38.5

million people in that country. It has been estimated that there are several million more speakers outside Poland, including immigrant communities in the United States of America and elsewhere. The earliest evidence that we have for Polish dates from the twelfth century and consists solely of proper names within an otherwise Latin document. The earliest surviving complete sentence in Polish dates from the thirteenth century and the earliest complete text dates from the fourteenth century. Within present-day Polish, five main dialect areas are usually recognized: Wielkopolska is an area toward the northwest of Poland, centered around Poznan, Mazowsze is in the northeast and includes Warsaw, MaAopolska is in the southeast, S`la sk is a small area toward the southwest, and Kaszuby is another small area in the north around the port of Gdansk. In addition to these, there are also mixed dialect areas, both in the extreme northeast and in the west along the border with Germany. These latter areas were part of Germany until the end of World War II and were then settled by Polish speakers of various dialects. An increasingly standard form of literary Polish, drawing mostly on the dialects associated with the main cultural centers of Poznan and Gniezno (in the Wielkopolska area) and Cracow (in the MaAopolska area) can be identified from the sixteenth century, when there was a flourishing ‘golden age’ of Polish literature. This literary language was also influenced to a certain extent by Czech, as well as by borrowings from Latin. Following a period of decline for written Polish in the early eighteenth century, a revivalist movement emerged in the late eighteenth and early nineteenth centuries, which aimed to codify the standard language. Although members of this movement took note of the earlier literary language of the sixteenth and seventeenth centuries, it was not primarily an archaizing movement (in contrast to the codification of Czech, outlined below) and consequently the differences between the codified written and contemporary spoken languages were not all that great. Throughout this period, and also thereafter, there was a strong opposition to the adoption of distinct regionalisms into standard Polish. A version of Polish based on the standard written language has become adopted as the standard spoken variety. Today, most Polish speakers who speak a dialect are also able to speak this standard language, and they use it in more formal and public settings. Distinctive features of Polish include the nasalized vowels mentioned earlier, the marking of gender as well as number and person in certain forms of the verb, and the use of third-person (‘he’ or ‘she’ or ‘it’) verb forms instead of second-person (‘you’) forms in respectful address.

863

POLISH AND WEST SLAVIC LANGUAGES The Kaszuby dialect (Cassubian in English) is, as stated earlier, often recognized as a separate language within the West Slavic group, although within Poland there is a greater tendency to see it as a dialect of Polish. It certainly possesses greater differences from standard Polish than do the other dialects, and Poles from outside the Cassubian region have difficulty in understanding it. Its status as a separate language is strengthened by a strong Cassubian ethnic identity. It also has its own literary tradition and is used regularly in the local media. Despite these outward attributes of a relatively cohesive language, Cassubian has a high degree of internal dialect variation. There have been various attempts to codify the language, but, to date, there is no formally recognized standard. Arguments against the status of Cassubian as a language are its close relation to Polish and the fact that the Cassubian area is politically a part of the Republic of Poland. References will sometimes be found to a further West Slavic language, Slovincian, which is now extinct. This was located geographically adjacent to Cassubian, to the northwest of the main Cassubianspeaking area. Although the residents of that area referred to their language as Slovincian, it is almost universally recognized by linguists as having been a subdialect of Cassubian. A combination of earlier Germanization and postwar population movements led to the extinction of Slovincian by the 1950s.

Czech Czech is the official language of the Czech Republic, spoken by between 9 and 10 million inhabitants, as well as by around 60,000 inhabitants of neighboring Slovakia and by immigrant communities elsewhere. The earliest surviving example of Czech dates from 1057 CE and consists of a set of glosses on a Latin manuscript. From the thirteenth century onward, a prolific Czech literary tradition exists, which flowered especially in the sixteenth century. Czech is unusual in that there is a larger than average difference between the standard written and spoken languages. The standard spoken language—known as Common Czech—represents the natural evolution of the Czech language over time. In contrast, standard written Czech stems from essentially artificial codification by linguists in the early to mid-nineteenth century, the most influential of whom were Josef Dobrovský and Josef Jungmann. Dobrovský, who was mainly responsible for codifying the grammar of Czech, based his grammar on substantially earlier literary usage from the golden age of the sixteenth and seventeenth centuries. Thus, in comparison with Common Czech and other West Slavic languages, standard written Czech shows a number of archaic features. For instance, the

864

six sets of noun endings from earlier Slavic are better preserved in Czech than in any of the other West Slavic languages. Jungmann was more concerned with widening the use of Czech, although his attitude was also essentially purist and he was responsible for introducing a number of distinctively Czech new words in preference to borrowing from other languages. Besides standard spoken Common Czech, Czech divides into three main dialect areas: Bohemian, Hanák (or Central Moravian), and Lach (or Silesian). Additionally, there are two transitional dialect areas: a southeast Moravian area, whose dialects are closely related to Slovak, and a mixed Czech-Polish dialect area. Standard Common Czech is linked most closely to the central Bohemian dialects, and particularly to the dialect of the main urban center, Prague. There is, however, a degree of regional variation even within Common Czech, and some now suggest that these are of sufficient magnitude to argue that two distinct regional versions of Common Czech exist—Bohemian Common Czech and Moravian Common Czech. As with Polish, there are also mixed dialect regions in areas along the border with Germany, which had been German-speaking prior to the end of World War II.

Slovak Slovak is the official language of Slovakia, spoken by around 4.6 million people there. A further 300,000 or so speakers are resident in the neighboring Czech Republic. Slovak is closely related to Czech and the two languages are, broadly speaking, mutually intelligible. This is particularly the case with the southeast Moravian dialects of Czech. However, there are also some differences between the two languages in both sound and word structure. This is especially true of the written languages, owing to the different language varieties that were drawn on during the period of their codification. As with Czech, the earliest surviving evidence for Slovak comes from a set of eleventh-century glosses. However, in contrast, no substantial literary tradition existed for Slovak prior to around 1780: instead, Czech took on the role of the written language, although Slovak writers did introduce some distinctively Slovak features into their written Czech. A further similarity with Czech is in the date that, and the process by which, the eventual standard written language did develop. Written Slovak was not codified until the 1840s, when a standardized form was produced by L’udevít Štúr in association with a group of other Slovak intellectuals. However, unlike Czech, where much earlier written usage exerted a strong influence on the newly codified language, the Slovak written language was based primarily on a synthesis of the contemporary spoken Central Slovak dialects. This

POLISH AND WEST SLAVIC LANGUAGES must at least in part have been a necessity, owing to the paucity of a written Slovak tradition, but it also mirrors the codification of Polish, where contemporary usage had more influence than it did with Czech. Slovak has three main dialect areas—West, Central, and East—with the West Slovak dialects being closest to their Czech neighbours.

Sorbian (Upper and Lower) The Sorbs are an indigenous Slavic minority who live in the region of eastern Germany known as Lusatia. Until 1990, Lusatia was part of the former German Democratic Republic. It now belongs to the federal states of Saxony and Brandenburg. In earlier times, Sorbian dialects extended eastward beyond the River Neisse into what is present-day Poland, as well as westward and northward into further regions of Germany. However, Sorbian is now wholly contained within the Lusatian region. Today, there have been estimated to be around 60,000–70,000 speakers of Sorbian, although all Sorbian speakers are now bilingual in Sorbian and German, and a number of these speakers use the German language for most everyday purposes. Sorbian has survived most strongly in small, closely knit rural communities, and has been noted as being especially strong in a group of villages to the north of Bautzen in Saxony. However, although its strongest base is among the rural community, Sorbian is not a low-prestige rural dialect and has a lively literary and scientific culture. Apart from proper names, the earliest surviving evidence for Sorbian vocabulary consists of glosses in Latin manuscripts dating from the twelfth century. The earliest continuous Sorbian texts date from the sixteenth century. These earlier Sorbian texts were written in local dialects with distinct differences from one another. In the eighteenth century, three standard written varieties emerged: Lower Sorbian, Catholic Upper Sorbian, and Protestant Upper Sorbian, with Upper Sorbian thus being split on denominational lines. (Since the Lower Sorbians were Protestants, the issue did not arise for that variety.) At that time, only religious texts were published in Sorbian. From the midnineteenth century onward, alongside an increase in nonreligious writing, the denominational split in Upper Sorbian was gradually diluted, although it was not until after World War II that the present twofold division of Sorbian was firmly established. There are two present-day standard varieties of Sorbian. Upper Sorbian is the language of the Sorbs in Upper Lusatia (centered around Bautzen in Saxony) and is the larger of the two varieties with roughly 40,000–50,000 speakers. Lower Sorbian is the language of the Sorbs in Lower Lusatia (centered

around Cottbus in Brandenburg), spoken by approximately 20,000 speakers. Upper and Lower Sorbian have a number of differences from one another, and these occur at all language levels, including sound structure (e.g. Lower Sorbian has góra in place of Upper Sorbian hora), vocabulary (e.g. Lower Sorbian has jo for ‘yes’, whereas Upper Sorbian has haj), and grammar (e.g. Upper Sorbian makes more use of possessive adjectives than Lower Sorbian). In contrast to the two standardized varieties, the features that distinguish the several Sorbian dialects show a rather complex distribution, given the small size of the language area. Nevertheless, broadly speaking, it is possible to make a primary distinction between Upper and Lower Sorbian dialect areas, with a transitional area, showing features of both, located in between. Having coexisted for such a long time alongside German, Sorbian has been affected by a number of Germanic influences, several of which are not shared by the other West Slavic languages. These influences have occurred not only at the vocabulary level but have also affected the grammar of Sorbian. For example, most Slavic languages do not have articles (such as a, an, and the in English). However, under German influence, Sorbian has made use of its demonstrative pronouns and the number ‘one’ as, respectively, definite (the) and indefinite (a/an) articles, although this practice is discouraged in the standard written languages, which have been subject to Slavic purist movements. Lower Sorbian probably retains more German influences than Upper Sorbian, which was subject to a more intensive push toward purism. German influence on Lower Sorbian was formerly also evident at the orthographic level, with much Lower Sorbian publishing using the German Fraktur type until around the time of World War II. As a small minority language within a Germanspeaking context, the number of speakers of Sorbian has declined substantially over the past century. However, there are lively movements to preserve its use and the future is not necessarily bleak. The rights of the Sorbs have been legally guaranteed within a bilingual framework since the end of World War II, firstly by the former German Democratic Republic and subsequently by the federal states of Brandenburg and Saxony. To promote and maintain the language, a number of initiatives exist, including the provision of bilingual schooling from an early age (the WITAJ initiative).

Polabian Polabian, sometimes known as Dravaeno-Polabian, is an extinct West Slavic language. The last native speaker is reported to have died in 1756, although there are records of someone able to recite the Lord’s Prayer in Polabian as late as 1798. Polabian was formerly spoken in a small

865

POLISH AND WEST SLAVIC LANGUAGES part of the Lüneburg area, which is situated in north Germany between Hamburg and Hannover. It was thus surrounded by German-speaking territory from an early point in its history and was geographically isolated from the other Slavic languages. (Contrast Sorbian, which has retained a border with both Polish and Czech.) Not surprisingly, therefore, it shows evidence of a very strong German influence, especially from the regional Low German varieties, both in vocabulary and grammar. Very little Polabian material has survived, and much of what has is in the form of isolated words and phrases compiled into dictionaries in the last century of the language’s existence. Only a small number of complete texts, such as the Lord’s Prayer and a folk song, have been preserved. A figure of around 6,000 recorded vocabulary items has been suggested. This has nevertheless provided enough evidence to suggest that at least three dialects existed. Despite this lack of material, and in view of its unusual isolation from related languages, Polabian has received continued and substantial attention from linguists, including notable figures such as Trubetzkoy.

References Auty, Robert. 1960–1961. Dialect, Koιv` and tradition in the formation of literary Slovak. Slavonic and East European Review 39. Auty, Robert. 1976. Problems in the formation and development of the Czech literary language. Slavic linguistics and language teaching, ed. by Thomas F. Magner. Cambridge, MA: Slavica. Comrie, Bernard, and Greville G Corbett (eds.) 1993. The Slavonic languages. New York: Routledge. De Bray, R.G.A. 1980. Guide to the West Slavonic languages. Columbus, OH: Slavica. Olesch, Reinhold. 1989. Gesammelte Aufsätze. 1. Dravaeno polabica. Cologne and Vienna: Böhlau. Polan´ski, Kazimierz. 1993. Polabian. In Comrie and Corbett. Rothstein, Robert A. 1993. Polish. In Comrie and Corbett. Short, David. 1993. Czech. In Comrie and Corbett. Short, David. 1993. Slovak. In Comrie and Corbett. Stone, Gerald. 1972. The smallest Slavonic Nation: the Sorbs of Lusatia. London: The Athlone Press. Stone, Gerald. 1993 Cassubian. In Comrie and Corbett. Stone, Gerald. 1993. Sorbian. In Comrie and Corbett. Townsend, Charles E. 1990. A description of spoken Prague Czech. Columbus, OH: Slavica.

ANDREW WILSON See also Indo-European 5: Slavic

Politeness Politeness refers to how a speaker chooses to phrase an utterance in a particular context, a choice that can reflect the speaker’s view of the interpersonal context. There are a number of different approaches to politeness. In a social normative view, to be polite is to speak formally and to behave in accordance with situationspecific rules regarding appropriate linguistic behavior. A variant of this approach has been to regard politeness as one component of pragmatic competence (Leech 1983). In this view, there are conversational maxims (e.g. tact and generosity) that stipulate which linguistic form is preferred over another in any particular context. Currently, the most popular approach to politeness, based on the work of Brown and Levinson (1987), views politeness as the linguistic means by which people manage each other’s ‘face’. The concept of face (Goffman 1967) refers to the identity that people present during the course of their interactions with others. Face is assumed to be universal and of two types: negative face, or the desire to have autonomy of action, and positive face, or the desire for closeness with others. One’s face is fragile and subject to continued threat during social interaction. For example, criticisms threaten the recipient’s positive face, and requests threaten the recipient’s negative face. Social 866

interaction thus presents a dilemma for interactants; they want to maintain each other’s face, but they are also motivated to perform actions that are face-threatening. This conflict is solved by engaging in facework, or, more precisely, by being polite. According to Brown and Levinson (1987), politeness is conveyed by deviating from maximally efficient communication; politeness is thus roughly equated with indirectness. There are many ways this can be accomplished, and Brown and Levinson organized politeness into five superstrategies. They provided evidence for these strategies from three different languages and argued that the linguistic strategies for conveying politeness are universal. The least polite strategy is to perform an act without any politeness, or to perform the act directly. The most polite strategy is to perform the act with an off-record form. Off-record forms are deliberately ambiguous and hence perform a face-threatening act indirectly (e.g. It’s warm in here as a request to have the window opened). Falling between these two extremes are on-record (i.e. relatively unambiguous) acts with redress emphasizing either positive face or negative face. The former, termed ‘positive politeness’, functions via an exaggerated emphasis on closeness or solidarity with the hearer. Examples

POSSESSIVES include the use of slang and familiar address terms, jokes, and presumptuous optimism (You’ll loan me your car, won’t you?). The latter, termed ‘negative politeness’, functions via attention to the recipient’s autonomy. This is accomplished by symbolically lessening any imposition on the hearer. A common method is to use conventionally indirect forms (e.g. Could you shut the door?). Research has provided partial support for this typology. It appears, however, that politeness strategies are very diverse and may not be completely captured with this scheme. Also, it is unclear whether politeness should be equated with indirectness; very indirect strategies may sometimes backfire and may be perceived as manipulative. One of the most important features of a face-management view of politeness is the attempt to specify the relationship between politeness and the interpersonal context. Brown and Levinson (1987) proposed that politeness will vary as a function of the weightiness (or degree of face-threat) of the act to be performed. Weightiness is contextually determined and assumed to be an additive weighting of the following three variables: the intrinsic (and culturally bound) degree of imposition of the act itself (e.g. asking for a loan of $5 is less imposing than asking for a loan of $500), the power of the hearer relative to the speaker, and the degree of social distance between the interlocutors. Thus, increased weightiness (hence, in general, greater politeness) occurs as a function of increasing imposition, hearer power, and relationship distance. Note that in this way the politeness of an utterance cannot be determined in isolation, but rather only in a specific interpersonal context. For example, an utterance perceived as polite when used with an underling may not be polite when directed toward a superior. Research has provided partial support for these ideas. Many studies have demonstrated that greater speaker power is associated with decreased politeness. Fairly consistent support has also been found for the imposition variable, with increasing imposition associated with increasing politeness. The effects of relationship distance have been the most problematic; whether greater politeness is associated with increasing relationship distance is not clear.

One of the most important issues surrounding politeness is its status as a cultural universal. Although few would argue the claim that politeness exists in all cultures, the claim that positive and negative faces are universal desires motivating the form that politeness takes has been questioned. For example, it has been argued that negative face is relevant only in Western cultures, or cultures in which there is an emphasis on individual autonomy. Although Brown and Levinson (1987) marshaled impressive evidence for the cross-cultural validity of their politeness strategies, empirical support for their claim that they are universal has been rare. Clearly, there is great cultural variability in politeness. The crucial question is whether this variability is a result of differing cultural conceptions of face or whether these differences can be explained at a lower level of abstraction. Face wants may be universal, but cultures will vary in terms of what threatens face, who has power over whom, how much distance is typically assumed, and so on. Recently, there have been some new developments in politeness theory. Some researchers have used it as a framework for explaining how people interpret indirect utterances. Others have begun to examine nonverbal politeness. Finally, many researchers are now using politeness theory in applied contexts as a means of examining and explaining linguistic behavior in specific contexts, such as courtroom trials, classroom interaction, and political campaigns. References Brown, P., and Levinson, S. 1987. Politeness: some universals in language usage. Cambridge, UK: Cambridge University Press. Fraser, B. 1990. Perspectives on politeness. Journal of Pragmatics 14. 219–36. Goffman, E. 1967. Interaction ritual: essays on face to face behavior. Garden City, NY: Anchor. Holtgraves, T. M. 2001. Politeness. Handbook of language and social psychology, 2nd edition, ed. by H. Giles and W.P Robinson. London: Wiley. Lakoff, R. 1973. The logic of politeness: or, minding your p’s and q’s. Papers from the ninth regional meeting of the Chicago Linguistic Society, 292–305. Leech, G. 1983. Principle of pragmatics. London: Longman. Watts, R. J. 2003. Politeness. Cambridge, UK: Cambridge University Press.

THOMAS M. HOLTGRAVES

Possessives Possessive marking is used to indicate that two nouns or noun phrases stand in a modifying relationship, namely that of possession. One noun (phrase) is the

modifier, the possessor, and the other noun (phrase) is the modified, the possessed. According to the position of the modifier, two kinds of possessives are available 867

POSSESSIVES in English: one is marked by the genitive case (my father’s books) and the other by a preposition (the books of my father). According to traditional grammars, the genitive possessive refers to a specific entity or individual. It is more frequent with modifiers referring to animate entities, i.e. living things, but certain inanimate modifiers can also take the genitive, such as geographical names (Canada’s farms), temporal nouns (yesterday’s paper), and the names of institutions (the university’s council). The use of inanimate modifiers also depends on the modified noun and is generally possible with modified nouns referring to places (the earth’s surface, the water’s edge, at the road’s end) and measures (arm’s length). The preposition of is used in partitive constructions (a piece of wood, a kind of letter) and expressions denoting a place of origin (Jesus of Nazareth), quality (a man of affairs), and material (a sword of steel). The genitive modifier can function as a determiner and thus excludes the use of other determiners (such as the or a). For example, it is not possible to say the my father’s book. To use a determiner, the so-called ‘double genitive’ construction must be used—that is, a book of my father’s. The double genitive can be used with the indefinite article a and with demonstrative pronouns such as that or this (a/that/this book of my father’s). The definite article can generally not be used in double genitive constructions: the book of my father’s is not a possible English phrase. However, if there is a modifying relative clause, use of the definite article becomes possible: the book of my father’s that is on the table. Genitive and prepositional markings can reflect the subject/object distinction in constructions like the enemy’s destruction of the city. Here, the nouns enemy and city receive the same interpretation as in the sentence the enemy destroys the city. In both cases, the enemy has a subject-like function, while the city is the object. This parallelism was an important piece of evidence in a heated linguistic debate concerning nominalization, the process of deriving a noun such as destruction from the corresponding verb, here destroy. In grammatical theories of the 1960s, it was assumed that a noun phrase like John’s presentation of the project is derived from the synonymous sentence John has presented the project. This view changed with Noam Chomsky’s widely quoted article ‘Remarks on nominalization’ (1970). Chomsky argued—convincingly to many linguists— that any noun can ‘project’ structures containing a subject and an object. Thus, for phrases like the enemy’s destruction of the city, the base structure would be [enemy destruction city]. For comparison, the base structure for The enemy destroyed the city would be [enemy destroy city]. Nouns and verbs thus ‘generate’ or ‘project’ very similar structures at an abstract level.

868

The only difference between noun phrases and verb phrases (or sentences, for that matter) lies in the elements that are used to mark the exact relationships between the individual items. In the case of noun phrases, the possessive is the relevant grammatical marker for the subject. The markers ’s and of thus were considered to be ‘semantically null’, i.e. they do not contribute any particular meaning to the sentence. Chomsky (1986) assumed that both are simply overt realizations of the genitive (case). In fact, possessive markers have no semantic content in cases of so-called ‘inalienable’ possession. For example, the expression John’s picture may have two ‘inalienable’ readings: (1) ‘John has created the picture’ vs. (2) ‘John is in the picture’. Both are cases of inalienable possession, because nobody can take away the fact that ‘John created the picture’, or nobody can take John out of the picture (without destroying the picture). In other words, the entities John and picture are inseparable and presuppose each other (i.e. an inherent relationship). If John is construed as a creator of the picture, this noun is inserted in the subject position [John picture], which automatically leads to the genitive expression John’s picture. If, on the other hand, John is the object of picture, the base structure would be [picture John], and the default phrasing would be the picture of John. However, if no creator is specified, i.e. if the subject position is left empty in the base structure, English allows ‘promotion’ of the object, which again yields the genitive John’s picture. Hence the potential ambiguity of English genitive expressions. However, Mona Anderson (1983) argued that the genitive ’s does contribute independent meaning in cases of ‘alienable’ or ‘true’ possession. Thus, the expression John’s picture can also be interpreted as ‘The picture that John owns’. This means that the relation between John and picture is not inherent but is established by means of the ‘word’ or ‘lexical unit’ ’s. Note that ’s can be labeled a word, since it fulfills a function similar to ‘full’ words, namely possess or own. In other words, in this case the phrase John’s picture is projected not from the noun picture but rather from the possessive ’s, and the base structure would be [John ‘s picture], similar to [John possess picture] or [John own picture]. The idea of lexical ’s has been commonly accepted since the second half of the 1980s, and it has been extended to other structures. For instance, his book has been analyzed as deriving from the base structure [he’s book], where he is the subject and book is the object of the word ’s. In a crosslinguistic perspective, the distinction between genitive and prepositional possessives is not universal. In French, for example, a genitive marker analogous to the English ’s does not exist. Thus, the

PRAGMATICS English expression Robert’s friends is translated in French as les amis de Robert (literally the friends of Robert). Russian, on the other hand, has no preposition corresponding to the English of. For instance, two constructions are available to translate the English expression Robert’s friends: (1) Robertovy druzja, where -ov- is a possessive marker, and (2) druzja Roberta, where -a is a morphological manifestation of genitive case. Historically, the marker ’s originates from the genitive case singular marker -es used in Old English. For example, in the sentence Him sceamode þæs mannes ‘He felt ashamed for the man’, this ending is observed on the noun stem mann ‘man’ preceded by the genitive form of the demonstrative se ‘the, that’, i.e. the literal translation would be something like ‘To-him therewas-shame the-of man-of.’ Similarly, þurh geswicenysse yfeles ‘by cessation from evil’ is literally ‘through cessation evil-of’, and cyninges botl ‘a king’s palace’ is ‘king-of palace.’ Since the genitive ending already sufficiently performed the function of expressing possession, the preposition of did not mark possessive phrases in Old English. Prepositional marking becomes much more frequent in Middle English. The latter is characterized by a common use of the socalled ‘group genitive’ constructions like the king of France’s son corresponding to þæs cyninges sune Frances ‘the-of king-of son France-of’ in Old English.

This change was caused by the increasing loss of case inflections and the resulting increased importance of function words. In conclusion, the possessive markings ’s and of are usually realizations of genitive case in English and establish a grammatical relation between two nouns, but English ’s also has a lexical use. The manifestation of genitive case varies across languages and historically also within a given language. References Allen, Cynthia L. 1997. The origins of the ‘Group Genitive’ in English. Transactions of the Philological Society 95.1. 111–31. Anderson, Mona. 1983. Prenominal Genitive NPs. The Linguistic Review 3. 1–24. Chomsky, Noam. 1970. Remarks on nominalization. Readings in English transformational grammar, ed. by Roderick A. Jacobs and Peter S. Rosenbaum. Waltman, MA: Ginn and Company. Chomsky, Noam. 1986. Knowledge of language: its nature, origin, and use. New York: Praeger. McCawley, James D. 1988. The structure of noun phrases. The syntactic phenomena of English, 2 vols. Chicago and London: The University of Chicago Press. Taylor, John R. 2001. Possessives in English: an exploration in cognitive grammar. New York: Oxford University Press.

EGOR TSEDRYK See also Case

Pragmatics Pragmatics is the study of language use. Whereas the study of grammar focuses on the language system, pragmatics offers a complementary perspective on language, providing an insight into the linguistic choices that users make in social situations. Pragmatics is, for instance, interested in how people pay compliments, engage in small talk, or write e-mails. The communicative functions of utterances or texts, and the speakers’ or writers’ intentions behind them, are of particular interest. Historically, the emphasis was on spoken language. The term ‘pragmatics’ was first used in semiotics, the general theory of signs. In this theory, pragmatics pertains to the relationship between signs and their users. In linguistics, pragmatics deals with verbal signs (words, utterances, texts) and how they are used by humans in communication. The term ‘pragmatics’

is derived from Greek pragma, which means ‘action’. Action is defined as intentional behavior. Pragmatics studies verbal communication as a complex form of intentional behavior, which is interactive, i.e. partner oriented, and symbolic, i.e. conventionalized and culture specific. In the English-speaking world, pragmatics was originally considered part of sociolinguistics. Pragmatics is a relatively young field of study. So far, there is no coherent pragmatic theory. Pragmatics developed from linguistic, philosophical, and sociological approaches to language use. The so-called pragmatic turn in linguistics was brought about by the writings of a group of philosophers known as speech act theorists in the late 1960s. Until then, mainstream linguistics had focused on linguistic forms and structures, neglecting meaning and ignoring communicative functions and language users.

869

PRAGMATICS Initially, pragmatics was dominated by the speech act theory. The basic insight of this tradition is that speech is action. Whenever we say something, we do not only produce sound waves, utter words, and produce sentences but we also perform an action. If a mayor says, I declare this bridge open, then this bridge is open. In this sense, the mayor has performed an action. Speech act theory aims at establishing how many fundamentally different ways there are of doing things with words. Bridge-opening belongs to ‘declarations’, a class of speech acts that all require an institutional context. Four further classes have been identified, which are more likely to occur in everyday communication. These are ‘directives’, ‘commissives’, ‘expressives’, and ‘assertives’. Typical examples for these classes are requests, promises, apologies, and statements, respectively. In linguistics, the study of language use focuses on the individual speech acts that belong to these fundamental classes, e.g. requests, promises, and apologies, and on the linguistic resources used to perform them. Empirical pragmatics aims at systematically determining the different strategies that people actually use to realize a particular speech act. A classic case is requesting. For instance, if you want a person to take out the garbage, you may say: Take out the garbage!, You ought to take out the garbage, How about taking out the garbage?, Can you take out the garbage?, Garbage day tomorrow, etc. These strategies differ in their degree of directness. The most direct strategy is to use an imperative construction, and the least direct strategy is to only hint at the desired action. Directness must not be confused with politeness. The degree of directness is a feature of any strategy independent of context, whereas the politeness value of an utterance depends entirely on the social situation. Theories of verbal politeness have also contributed to the development of pragmatics. According to one particular politeness theory, any speech act may, at least potentially, interfere with the wants and needs of hearers. This is obvious if a speaker wants a hearer to do something. It is less obvious in the case of social niceties, but a compliment, for example, may also be interpreted as harassment. Given the threatening potential of all speech acts, speakers are thought to calculate the imposition involved in the respective situation and then select from a range of direct and indirect strategies available for realizing the act in question. Also, any of these strategies can be modified in a number of ways. For instance, speakers can say, Could you perhaps take out the garbage, darling?, where perhaps and darling serve to modify the basic strategy, Could you take out the garbage? Modifications are motivated by polite considerations and aim at reducing the social risks involved in performing the act in question.

870

A distinction can be made between pragmalinguistics and sociopragmatics. ‘Pragmalinguistics’ is concerned with the verbal resources available for realizing any given speech act. By contrast, ‘sociopragmatics’ focuses on the polite norms governing the selection of resources relative to social situations. For example, pragmalinguistics identifies the word choices, meaning patterns, and sentence constructions that are used to pay a compliment, whereas sociopragmatics determines who may compliment whom on what in which situations. Pragmalinguistics is language specific, and sociopragmatics culture specific. In empirical pragmatics, five types of inquiry can be distinguished. Studies may focus on speech act realization in one language alone, or they may compare speech acts across languages. The second type is known as contrastive pragmatics. It aims at answering questions such as these: how do speakers of English and speakers of Japanese formulate complaints? By contrast, cross-cultural pragmatics examines differences and similarities in social norms and cultural values governing polite choices. A typical question is this: how do members of US American culture and members of British culture refuse offers? A fourth branch of empirical pragmatics is interlanguage pragmatics, which is the study of language use in a second or foreign language. The fundamental question here is the following: how do language learners perform and interpret speech acts in the target language? Finally, applied pragmatics focuses on problems of communication in practical contexts, such as courtrooms, medical interviews, or international business encounters. Both natural and elicited data are analyzed in empirical pragmatics. Natural data, which occur in everyday situations, are recorded by using field notes, audiotape, or video. Further sources are large corpora of spoken language, such as the Santa Barbara Corpus of Spoken American English. Alternatively (or additionally), a wide range of research instruments is used to elicit data. Tools include role plays, questionnaires, and interviews. Fictional material is also used, e.g. dialogue in narrative prose or drama. All approaches to language use discussed so far focus on speakers and their intentions, rather than on hearers and their interpretations. How hearers understand utterances is dealt with in a theory of conversational maxims, which also developed in speech act philosophy. According to this theory, hearers assume that speakers behave cooperatively at all times (cooperative principle). More specifically, hearers expect speakers to observe a set of conversational maxims, involving ‘Don’t lie’ (maxim of quality), ‘Say as much as is necessary, no more and no less’ (maxim of quantity), ‘Be relevant’ (maxim of relation), and ‘Be brief’ (maxim of manner). If an utterance does not, for example, seem to relate to

PRAGMATICS what was said before, or if maybe more or less is said than is necessary, then the hearer tries to work out an interpretation that makes sense in the given context. This process is called ‘inferencing’. A pragmatic theory of communication and cognition that focuses on the process of understanding utterances is relevance theory. It is based on the conversational maxim of relation. In everyday communication, people do not always observe the maxims of conversation. For example, people lie, talk too much, or make obscure statements. Such behavior can be explained by assuming that a politeness principle may conflict with the cooperative principle. In a particular situation, conversationalists may rate politeness maxims higher than the conversational maxims. For instance, a ‘white lie’, which violates the maxim of quality, can sometimes be more considerate than telling the truth. Generally speaking, maxims of politeness minimize unfavorable effects for the hearer or increase favorable effects. For example, positive evaluations expressed in compliments are often maximized, whereas impositions involved in requests are usually reduced. Most theories of language use and most empirical approaches are speech act-based, i.e. they focus on utterances in isolation. Speech act-based pragmatics is referred to as micropragmatics. Its shortcomings are remedied in macropragmatics, which provides a fuller picture of language use by accounting for the dialogical and interactive nature of human communication. Macropragmatic approaches do not focus on the speaker or hearer alone, but rather on participants in communication who alternately adopt the role of speaker or hearer. Also, they challenge the view that the speech act is the central communicative unit and claim that the ‘exchange’ is more important. An exchange minimally consists of two speech acts. The first act functions as an initiating ‘move’, and the second act as a responding ‘move’. Moves indicate the interactive function of an act relative to the immediate verbal context. Typical exchanges are question– answer, offer–refusal, and greeting–greeting. Between initiation and response, countering moves can occur, as in Can you take out the garbage? (initiating move) Why don’t you take it out yourself? (countering move) Okay, okay. Calm down, I will (responding move). An exchange is closed when an outcome has been reached. In macropragmatics, complete communicative events such as conversations, interviews, debates, etc. are analyzed. Communicative events are considered hierarchical structures. The hierarchy, starting with the smallest unit, involves acts, moves, exchanges, sequences, and phases. At each of these levels, there are obligatory units (heads) and optional units (supportives). Thus, we distinguish head moves and sup-

portive moves, head exchanges and supportive exchanges, and so on. For example, the utterance Can you lend me ten dollars? I’ve left my purse at home consists of a head move, which is the request, and a supportive move, which is an explanation. The type of macropragmatics sketched here is based on British discourse analysis, which was developed in linguistics. An alternative approach is conversation analysis, which originates in American sociology. In conversation analysis, the term ‘adjacency pair’ is used for the exchange unit. Another unit identified in conversation analysis is the ‘turn’. Conversationalists take turns at talk. A turn is defined as a speaker’s contribution at a particular point in a conversation. It is surrounded by speaker switches. Turns may differ considerably in length. Prototypically, a turn consists of a second pair-part to match the previous speaker’s first pair-part and a first pair-part that must be matched by the next speaker. In conversation analysis, the mechanisms of turn-taking and turn-allocation for encounters involving two or more speakers have been identified, and simultaneous speech and interruptions have also been investigated. Conversation analysis differs from discourse analysis in using only naturally occurring conversation, whereas discourse analysis also works with other data types, including fictional dialogue. Most research in pragmatics deals with present-day language. Recently, however, an interest has also developed in earlier periods. Two branches have evolved, namely, historical pragmatics and diachronic pragmatics. Historical pragmatics deals with previous stages of language, e.g. Old English or Middle English. By contrast, diachronic pragmatics studies development across time. In both types, two alternative perspectives can be adopted. On the one hand, the way in which a particular communicative function, e.g. a speech act, was realized at different times may be examined. On the other the researcher may start with a particular expression, e.g. a routine formula, and describe its uses in certain periods or in the course of history. It is not always possible to keep these two perspectives apart. Typical research topics in historical pragmatics include greetings in the fifteenth century and politeness strategies in Shakespeare’s plays. Topics for diachronic pragmatics include, for instance, insults in Old, Middle, and Modern English or changes in apologizing through the ages. Although all historical sources are in the written mode, some include a considerable amount of information on spoken language. For example, court reports may incorporate detailed accounts of what defendants or witnesses actually said. Finally, recent and ongoing pragmatic change may be studied in e-mails and chat-room conversations, because netiquette is only just emerging.

871

PRAGMATICS Another recent development in pragmatics is the study of synchronic variation in contemporary language use. Sex differences have been examined to some extent, but regional, social, and ethnic variation has not received much attention in pragmatics. Sociolinguistic research suggests that we can expect speech act realizations, politeness markers, discourse strategies, and turn-taking, etc., to vary across regions, urban and rural communities, social classes, and ethnic groups. This is an area for future investigation. In short, pragmatics is the study of language use. Pragmatics is not a component of the language system, but it offers a different perspective on verbal phenomena. It examines how linguistic resources are used in communication and investigates a speaker’s (or writer’s) intentions and a hearer’s (or reader’s) interpretations. Language use is considered a complex form of social action. The general question addressed in pragmatics is how language functions in the lives of human beings. The focus can be on utterances (micropragmatics) or on longer stretches of discourse (macropragmatics). Pragmalinguistics is concerned with the structural resources that a language provides for conveying particular intentions, whereas sociopragmatics considers language use relative to social situations. There is no coherent pragmatic theory to date. Theoretical pragmatics draws on a number of disciplines, including philosophy, sociology, and anthropology. Speech act theory, politeness theory, discourse analysis, and conversation analysis provide major contributions. Empirical pragmatics involves the investigation of speech act realizations,

politeness phenomena, and discourse strategies on the basis of natural, elicited, and also fictional data. The comparison of language use across languages and cultures is also an important empirical research component. Applied pragmatics focuses on problems of communication in practical contexts. Most pragmatic research investigates present-day language, but recent developments also include historical and diachronic perspectives. References Blakemore, Diane. 1992. Understanding utterances: an introduction to pragmatics. Oxford, Cambridge, MA: Blackwell. Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper, (eds.) 1989. Cross-cultural pragmatics. Norwood, NJ: Ablex. Jucker, Andreas, (ed.) 1995. Historical pragmatics. Amsterdam: Benjamins. Leech, Geoffrey. 1983. Principles of pragmatics. London and New York: Longman. Levinson, Stephen. 1983. Pragmatics. Cambridge: Cambridge University Press. Mey, Jacob (ed.) 1998. Concise encyclopedia of pragmatics. Oxford: Elsevier. Mey, Jacob. 2001. Pragmatics: an introduction, 2nd edition. Oxford and Cambridge, MA: Blackwell. Trosborg, Anna. 1995. Interlanguage pragmatics. Berlin and New York: Mouton de Gruyter. Verschueren, Jef. 1999. Understanding pragmatics. London and New York: Arnold. Wierzbicka, Anna. 1991. Cross-cultural pragmatics. Berlin and New York: Mouton de Gruyter.

KLAUS P. SCHNEIDER See also Speech Acts

Predication Although predication is extremely important in contemporary linguistic research, it is a very old concept tracing back to Aristotelian logic, where it emerged to indicate the semantic relationship existing between the event described in a proposition and the individuals participating in such an event. Due to its origin in the Western philosophic tradition, the linguistic analysis of predication has been approached for a long time from a semantic perspective. From the 1980s onward, however, its study took a different course due to the advent of several transformational theories mostly devoted to examine the syntactic behavior of predication. Being traditionally accounted for in semantic terms as the connection between a predicate—the event

872

expressed in the clause—and its arguments—the entities involved in the event—the semantic analysis of predication has been mainly interested in how this linking takes place. The most common answer to this question is found in Theta Theory—that is, the lexicosemantic module of the transformational theoretical framework known as Government and Binding. Since it is chiefly concerned with the predicate’s assignment of theta roles to its different arguments, it specifies the exact number of arguments a particular predicate requires and their semantic nature; that is, if they take part in the predication as agents, patients, experiencers, themes, etc. It must be noted, however, that, being unlimited in number, there is no agreement as regards

PREDICATION the name and catalogue of these detailed semantic traits alloted by a predicate to its arguments, generally called theta roles. Regardless, the main principles underlying Theta Theory are formulated in the well-known ThetaCriterion, which postulates the following restrictions on the assignment of theta roles: (a) every argument in a predication must receive one theta role from its predicate; (b) the same theta role cannot be assigned to more than one participant in the predication. According to Theta Theory, then, and as illustrated in Mary gave some money to her little brother, a predicate like give needs three arguments bearing the following theta roles to be completed: an agent (Mary), a theme (some money), and a recipient (her little brother). From the syntactic viewpoint, predication is defined via saturation: a predicate is a maximal projection— that is, a full phrasal category—and, consequently, an open function, which must be syntactically closed by the appropriate arguments to form a grammatically acceptable chain. Despite being the maximal projection with the strongest predicative capacity, the verbal phrase is not the only one able to function as predicate: The girl read her cousin a book. The adjectival (Bill is fond of Jan), the nominal (This ratty piece of leather is a wallet), the prepositional (Those sisters are on drugs), and the adverbial phrases (She feels a bit down today) can also perform, as illustrated between brackets, the predicative function. Although different predicates demand a different number of arguments, as the previous examples show, they all share the property of having, at least, one compulsory argument: their subject. This argument (The girl, Bill, This ratty piece of leather, Those sisters and She in the previous examples) differs from the rest of possible arguments that a predicate may have both semantically and syntactically: semantically, because it constitutes the starting point for the predication; that is, it is the participant the predication is about; and syntactically, because it is the predicate’s only external argument. Although the subject and predicate constituents are unanimously said to be in a relation of mutual c-command—that is, they are syntactic sisters dominated by exactly the same maximal projections—the two main transformational approaches to the study of predication explain the subject–predicate syntactic relationship in different ways: the Predication Theory, for instance, considers that subject and predicate are independent structural components, whose syntactic relation has to be marked by indexing at surface structure, and the Clausal Theory defends, in turn, that they both form a single constituent having clausal status either with or without verb. If this is the case, the clause containing the predication relationship is catalogued as small. An example of a small clause is, thus, represented by Jill a nice person in Rachel considers Jill a nice person.

Although the previous syntactico-semantic characterization applies to predication in general, some refinements must be made because not all instances of predication exhibit the same grammatical behavior. Predication cannot be considered, therefore, a homogeneous linguistic phenomenon, but rather a heterogenous one involving two distinct classes of predication: primary and secondary predication. The former corresponds to the prototypical notion of predication because, being syntactic and semantically independent, it embodies the basic and matrix predicative relationship occurring at the level of the clause: John painted the house, John eats carrots. The latter lacks, in turn, such syntactico-semantic autonomy and must always appear, consequently, complementing an instance of primary predication: John painted the house red, John eats the carrots raw. The aforementioned contrast is, furthermore, extremely important because it determines the status of the subject of each type of predication: whereas the subject of a primary predication is systematically, like John in the previous examples, the predicate’s theta-marked external argument, the subject of a secondary predication displays a dual function: it is the secondary predicate’s theta-marked external argument at the same time as a theta-marked argument of the primary predication; notice, for instance, that in the examples above, the house and the carrots are, on the one hand, the subjects of the secondary predicates red and raw, and on the other, the direct objects of the primary predicates painted and eats. Besides these contrasts, two semantic types of secondary predication must be distinguished: resultative and depictive secondary predication. The former, as illustrated by the house red in the previous example, signals a potential result of the action denoted by the verb and can only be predicated by the direct object argument of the primary predication; the latter, in turn, does not maintain any semantic relation with the verbal predicate it complements, because, as shown by raw in the aforementioned example, depictive secondary predicates denote intrinsic and temporary properties of the subject or direct object arguments of the primary predication, which must exist at the time of the verbal action. References Aarts, Bas. 1995. Secondary predicates in English. The verb in contemporary English. Theory and practice, ed. by Bas Aarts and Charles F. Meyer. Cambridge: University Press. Dik, Simon C. 1981. Predication and expression: the problem and the theoretical framework. Predication and expression in functional grammar, ed. by A. Machtelt Bolkestein, Henk A. Combé, Simon C. Dik, Casper de Groot, Jadranka Gvozdanovi_, Albert Rijksbaron, and Co Vet. London: Academic Press. Frawley, William. 1992. Linguistic semantics. New Jersey: Lawrence Erlbaum Associates.

873

PREDICATION Nakajima, Heizo. 1990. Secondary predication. The Linguistic Review 7. 225–309. Napoli, Donna Jo. 1989. Predication theory. A case study for indexing theory. Cambridge: Cambridge University Press. Rapoport, Tova R. 1990. Secondary predication and the lexical representation of verbs. Machine Translation 5. 31–55. Rothstein, Susan D. 1985. The syntactic forms of predication. Bloomington: Indiana University Linguistics Club.

Stowell, Tim. 1983. Subject across categories. The Linguistic Review 2. 285–312. Williams, Edwin S. 1980. Predication. Linguistic Inquiry 11. 1. 203–38. Williams, Edwin S. 1983. Against small clauses. Linguistic Inquiry 14. 2. 237–308.

BEATRIZ RODRÍGUEZ-ARRIZABALAGA See also Thematic Structure

Professions for Linguists Linguists study aspects of language, languages, and language use. With this diversity of focus, linguists may be found in many different professions. Linguists are used in both private and public sectors, in jobs related directly to linguistics, as well as in professions where a knowledge of language or language use is highly relevant, although not directly required. The nature of language has interested scholars for over 2,000 years. The earliest works of a linguistic nature date to some of the first writings known to humankind. Such early linguistic writings include discourse on the nature of language and humanity (e.g. Socrates), the reconstruction of the spoken language of the Vedas (e.g. Panini in India), and the translation of texts and codification of languages (e.g. the trilingual Rosetta Stone, Varro’s grammar of Latin). The description of child language acquisition has also fascinated scholars through the ages, including scientists (e.g. Charles Darwin), psychologists (e.g. Piaget), philosophers (e.g. Rousseau), and others. In fact, the disciplines within which linguists have worked have included philosopher (e.g. Plato, Locke, Hume), grammarian (e.g. Dionysius Thrax, Cicero), philologist (e.g. Saussure), anthropologist and field linguist (e.g. Boas, Sapir), psychologist (e.g. Skinner, Piaget), scientist (e.g. Bloomfield), and ‘linguist’ within an independent field of research (e.g. Chomsky). In their professions, linguists may focus on the sounds of the language (e.g. phonetics, phonology), the words of the language (e.g. semantics), how words combine (e.g. syntax, discourse), how speakers acquire and/or use language (e.g. sociolinguistics, psycholinguistics, applied linguistics, dialectology), how the human brain processes language (e.g. psycholinguistics, neurolinguistics), how to program machines to carry out voice recognition/voice synthesis (e.g. computational linguistics), or even how animal communication systems, or languages, work (e.g. zoosemiotics).

874

The sounds of language are the focus of phoneticians in the fields of acoustic phonetics (e.g. the study of the physical properties of speech sounds using instrumental techniques of investigation to provide an objective account of speech patterns, which can be related to the way sounds are produced and heard), auditory phonetics (e.g. the study of the way people perceive sound, as mediated by the ear, auditory nerve, and brain), or articulatory phonetics (e.g. the study of the use of the vocal organs (articulators) to produce the sounds of speech). The sound systems of language are of interest to phonologists, who study the sound systems of languages and the general or universal properties displayed by these systems. Phonetics and phonology are also relevant to the teaching of pronunciation, speech pathology and speech therapy, accent reduction, voice and drama coaching, elocution and public speaking, forensic linguistics (e.g. voice recognition, voice printing), sound engineering, and the growing area of speech synthesis and voice activation in the computer industry. Syntactians study the structure (or grammar) of languages, including the way words are combined to form sentences, relationships between the elements of sentence structure, and the rules governing the arrangement of sentences in sequences. Meaning in language, in terms of relations such as synonymy and antonymy, is the domain of semanticians. Issues such as to what extent semantic concepts are universal is a current area of focus. In addition to academic research, syntacticians and semanticists work on dictionaries and grammars of different languages, develop and study computer languages and computer translation, preserve endangered and dying languages through field research, provide forensic analysis for the legal system, develop language tests, develop language policy and other educational curriculum, work in language reconstruction

PROFICIENCY TESTING (with archeologists), etc. Linguists also work for intelligence agencies as spies, translators, interpreters, code-breakers, and computer programmers. Professions for linguists related to the acquisition of first/child language include developmental linguistics and developmental psychology, psycholinguistics, early childhood education, deaf education, teaching, speech therapy, speech pathology, textbook and curriculum development, and software development. Similarly, professions for linguists related to the acquisition of second or additional languages include applied linguistic research, ESL/EFL teaching, foreign and modern language teaching, bilingual/bicultural education, teacher training, translation, proofreading, marketing, textbook writing and publishing, editing, speech therapy, speech pathology, software development, website design and marketing, on-line teaching, recruiting students for language schools (e.g. agents), owning language schools and other related businesses, developing and implementing multicultural policies, facilitating minority language rights, language policy, language planning in business and government services, and as freelance consultants (e.g. providing program development/analysis, giving seminars on gender issues in the workplace, cross-cultural training, etc.). Some additional areas within linguistics in which linguists are employed include anthropological linguistics (the study of languages in relation to social and cultural patterns and beliefs); areal linguistics (the study of geographical areas that are characterized by shared linguistic properties); audiology (the study of hearing and hearing disorders) biolinguistics (the study of the biological preconditions for language development and use); dialectology (the study of dialects, dialect geography, or linguistic geography); historical linguistics (the study of language variation and/or change); neurolinguistics (the study of the basis in the human nervous system for language development and use); philology (the study of language history, including historical study of literary texts); philosophical linguistics (the study of the role of language in relation to philosophical concepts, as well as the philosophical status of linguistic theories, meth-

ods, and observations); pragmatics (the study of the conditions on language use deriving from the social situation); semiology (the study of signs and their use, focusing on the mechanisms and patterns of human communication and on the nature and acquisition of knowledge); semiotics (the study of language as one type of sign system, along with bodily gestures, clothing, and the arts); sociolinguistics (the study of how language is integrated with human society with reference to race, ethnicity, class, sex, and social institutions); and zoosemiotics (the study of animal communication systems). In the new millennium, with globalization and the explosion of new communication technologies, the knowledge and expertise of linguists are increasingly in demand in the fields of education, business, technology, health, law, and communications. The field that began with philosophical explorations of language and humanity is now recognized as being in the forefront of human/machine interfaces in the twenty-first century. References Brand, Amy. A short road from academic research to academic publishing: one linguist’s trip. Linguistic Enterprises Advice. http://web.gc.cuny.edu/linguistics/enter/elgin.htm. Crystal, David. 1987. The Cambridge encyclopedia of language. Cambridge, UK: Cambridge University Press. Haden Elgin, Suzette. Another item in the freelance linguist’s survival kit: Seminars, Linguistic Enterprises Advice, http:// web.gc.cuny.edu/linguistics/enter/elgin.htm Levi, Judith, Questions and answers about language consulting with lawyers, Linguistic Enterprises Advice, http://web.gc. cuny.edu/linguistics/enter/elgin.htm O’Grady, William, and Michael Dobrovolsky. 1996. Contemporary linguistic analysis. Toronto, CA: Copp Clark Ltd. Wylie, Dovie. But Mom, who is going to pay you to know these things? Linguistic Enterprises Advice. http://web.gc.cuny. edu/linguistics/enter/elgin.htm Other Work-Related Websites Dave’s ESL Café: http://www.eslcafe.com/search/jobs/ Linguist List: http://www.emich.edu/linguist/jobsindex.html American Association of Applied Linguistics: http://www. aaal.org Linguistic Enterprises: http://web.gc.cuny.edu/linguistics/enter/

KAREN WOODMAN

Proficiency Testing The term proficiency testing embodies two main viewpoints: an earlier, traditional one referring to tests of

ability to use a language for some extralinguistic purpose, and the more recent ‘Proficiency Movement’

875

PROFICIENCY TESTING in foreign language assessment associated with the American Council on the Teaching of Foreign Languages (ACTFL). Traditionally, proficiency testing has been contrasted mainly with achievement testing in which the test content and method are based on a language teaching syllabus. The purpose of an achievement test is to find out how much students have learned of what has been taught. By contrast, proficiency tests are not linked to any syllabus but rather attempt to measure how well a learner can use the language regardless of the circumstances in which it was acquired. Lado (1961) was the first to make this contrast explicit: [Achievement] tests...attempt to measure how much of a language a student knows. When the...tests are thought of independently of the learning experience, they may then be referred to as proficiency tests. Proficiency tests measure how much of a foreign language a person (not necessarily a student) knows. (p. 369)

Note that Lado speaks of the purpose of the proficiency test as measuring how much of the language a person knows, reflecting his ‘discrete point’ approach to test development wherein specific grammatical points are tested more or less in isolation. At about the same time, other scholars were calling for proficiency tests that required language performance in contexts of use. For example, Carroll (1961 [1972]) argued that an ideal English language proficiency test should make it possible to differentiate, to the greatest possible extent, levels of performance in those dimensions of performance which are relevant to the kinds of situations in which the examinees will find themselves after being selected on the basis of the test. (p. 319)

Thus, Carroll explicitly links proficiency testing to performances relevant to future situations, and this has become the standard view of proficiency: ability to use a language for some extralinguistic purpose (Davies et al. 1999). The extralinguistic purpose that proficiency testing is most often associated with is admission to academic programs. For example, the two best-known proficiency tests, the Test of English as a Foreign Language (TOEFL) and the International English Language Testing System (IELTS), are both used primarily for admission to university programs in the United States, the United Kingdom, and Australia. These tests were developed to answer the question of whether a candidate for admission to an English-speaking academic institution possesses sufficient English ability to cope with academic work. However, proficiency tests need not be limited to academic situations only. The focus on future uses of the language naturally leads to tests of languages for specific purposes (LSP). For example, the Occupational

876

English Test (OET) was developed in Australia to ascertain whether migrant candidates for licensure in the health professions are able to use English with sufficient ability to work with English-speaking patients and clients; the Japanese Language Test for Tour Guides was also developed in Australia; the Test of English for International Communication (TOEIC) is a US test aimed primarily at international business people; and the Proficiency Test in English for Air Traffic Controllers is a European test for trainee air traffic control officers. These and other specific-purpose language tests were developed without reference to any particular course of instruction but rather on the basis of an analysis of a particular target language use situation (Douglas 2000). The distinction between achievement and proficiency testing is not entirely rigid, either. For example, Bachman (1990) argues as follows: Whether or not the specific abilities measured by a given proficiency test actually differ from those measured by a given achievement test will depend, of course, on the extent to which the theory upon which the proficiency test is based differs from that upon which the syllabus is based (p. 71).

Note that Bachman defines proficiency tests as theory-based, and indeed, it is generally the case that proficiency tests are related to some theory of what it means to know and use a language. Since Hymes’s (1972) formulation of communicative competence, proficiency test developers have used theories about the components of language knowledge to guide them in deciding what to test. The current, most well-known framework is that proposed by Bachman (1990) and Bachman and Palmer (1996) of communicative language ability, consisting of grammatical, textual, functional, and sociolinguistic language knowledge, plus strategic competence, which directs the use of one’s language knowledge in actual communication. The so-called Proficiency Movement, although more recent than proficiency testing itself, has its origins in the early days of the assessment of the oral language skills of US Government employees, most notably the Foreign Service Institute (FSI), the Peace Corps, the Central Intelligence Agency (CIA), and the military (Lowe 1988). The FSI (now the Interagency Language Roundtable [ILR]) Oral Proficiency Interview became the standard for the testing of spoken language ability, and the American Council on the Teaching of Foreign Languages referred to it in responding to a mandate from the 1979 President’s Commission on Foreign Languages and International Studies to establish a series of nationally recognized descriptors that would facilitate assessment of the proficiency of both students and teachers in foreign language programs. The ACTFL Proficiency Guidelines

PRO-FORMS (1986, 1999) thus ‘represent the combined efforts of groups of educators to provide an operational definition for [proficiency], to represent it as phenomena observable and evolving’ (Galloway 1987: 26). The Guidelines are directed primarily at assessing the foreign language proficiency of college and university students in the United States, and consist of descriptors of performance in speaking, listening, reading, and writing at Superior, Advanced, Intermediate, and Novice levels, with the latter three levels divided into high, mid, and low categories. The focus of the Proficiency Movement, as embodied in the ACTFL Guidelines, is the assessment of language production or performance rather than internalized language knowledge or ability for use (Galloway 1987; Lowe 1988), thus distinguishing the Movement from the more traditional understanding of the concept of proficiency described above. References American Council on the Teaching of Foreign Languages. 1986. ACTFL proficiency guidelines. Hastings-on-Hudson, NY: American Council on the Teaching of Foreign Languages. American Council on the Teaching of Foreign Languages. 1999. ACTFL proficiency guidelines—Speaking. Hastings-onHudson, NY: American Council on the Teaching of Foreign Languages. Bachman, Lyle. 1990. Fundamental considerations in language testing. Oxford: Oxford University Press.

Bachman, Lyle, and Adrian Palmer. 1996. Language testing in practice. Oxford: Oxford University Press. Carroll, John B. 1972. Fundamental considerations in testing for English language proficiency of foreign students. Teaching English as a second language: a book of readings, ed. by H.B. Allen and R.N. Campbell. New York: McGrawHill. (Originally published in Testing, Washington, DC: Center for Applied Linguistics, 1961.) Davies, Alan, Annie Brown, Cathie Elder, Kathryn Hill, Tom Lumley, and Tim McNamara. 1999. Dictionary of language testing. Studies in language testing, Vol. 7. Cambridge: Cambridge University Press. Douglas, Dan. 2000. Assessing languages for specific purposes. Cambridge: Cambridge University Press. Galloway, Vickie. 1987. From defining to developing proficiency: a look at the decisions. Defining and developing proficiency: guidelines, implications, and concepts, ed. by Heidi Byrnes and Michael Canale. Lincolnwood, IL: National Textbook Company. Hymes, Dell. 1972. On communicative competence. Sociolinguistics, ed. by J.B. Pride and J. Holmes. Harmondsworth, UK: Penguin. Lado, Robert. 1961. Language testing: the construction and use of foreign language tests. London: Longman, Green, and Company. Lowe, Pardee. 1988. The unassimilated history. Second language proficiency assessment: current issues, ed. by Pardee Lowe and Charles Stansfield. Englewood Cliffs, NJ: Prentice-Hall Regents. McNamara, Tim. 1996. Measuring second language performance. London, New York: Longman.

DAN DOUGLAS See also Second Language: Learning; Second Language: Teaching

Pro-forms In linguistics, a pro-form is generally taken as an element used in place of other linguistic element(s). Pronouns are the most common pro-forms used to substitute for a noun or a noun phrase. ‘She’, for example, is a third-person singular pronoun used in place of a singular human female animate noun such as a woman as in, ‘A woman is coming to see you. She has called you earlier.’ Although pronouns have often been used as examples of pro-forms, there are other linguistic elements that have comparable properties but do not substitute for a noun or a noun phrase. ‘So’, as in ‘He thinks it will rain tonight but I don’t think so,’ substitutes for a whole clause, i.e. ‘it will rain’. Here and there are pro-forms for locative expressions; bare noun phrases may be substituted with what

and one/ones, and adjectives by such and verb phrases by do. The term ‘pro-form’ was probably first used by Jerrold Katz and Paul Postal (1964) as a mechanism to explain both syntactic and semantic aspects of the substitutions in the above examples. Syntactically, the pro-constituent guarantees the recoverability of a substitution or deletion. Semantically, the pro-form calls for interpretation by retrieving its equivalents. The term pro-form, since its introduction, has often been used alternately with pronoun, and now it seems to replace pronoun. However, some linguistic elements seem to have comparable properties to pronouns but they are not substitutes for nouns. In fact, there are many other classes of words than nouns that get a different form in the following mention in a text.

877

PRO-FORMS Thus, the term pro-form seems appropriate to be used collectively for any kind of substitution. In the literature of generative grammars, a pro-form is often found as an element that assumes the process of substitution. In these theories, pro-forms can be used as one of the key tests for constituency in syntax: if a string of words can be replaced by a pro-form, these words form a constituent, i.e. a structural unit within the sentence. To illustrate, in the sentence ‘The student read a book in the library,’ the pronoun she can substitute for the noun phrase ‘the student’, a there can stand in for the prepositional phrase ‘in the library’, and a did can substitute for the verb phrase ‘read a book in the library’. Consequently, we may conclude that these three types of phrases are constituents. Moreover, there are some other terms that are loosely used in place of pro-forms. One of these is ellipsis. Andrew Radford (1997) considers ellipsis a process by which redundant information in a sentence is omitted. Pro-forms, however, are not omissions but

substitutions. However, some recent work in natural language processing (e.g. by Daniel Hardt 1993) includes pro-form as one category of elliptical forms. In terms of semantics, a pro-form has no meaning in itself, rather it requires a retrieval of meaning from a previously mentioned element, or antecedent, i.e. the element for which it substitutes. In other words, proforms are semantically bound by other elements. References Hardt, Daniel. 1993. Verb phrase ellipsis: form, meaning and processing. Ph.D. dissertation. University of Pensylvania. Katz, Jerrold, and Paul Postal. 1964. An integrated theory of linguistic descriptions. Cambridge: Cambridge University Press. Radford, Andrew 1997. Syntactic theory and the structure of English: a minimalist approach. Cambridge: Cambridge University Press. Van Valin, Robert D. Jr. 2001. Introduction to syntax. Cambridge: Cambridge University Press.

PRANEE KULLAVANIJAYA

Proper Nouns Proper nouns (usually known as ‘proper names’ by philosophers) are basically names of specific people, places, organizations, months, festivals, and so forth. Typical examples are John, London, Oxfam, January, and Christmas. The investigation of the semantic meaning of proper nouns goes back to the Greeks. In 200 BCE, the Greek grammarian Dionysius Thrax gave one of the most often used definitions: a proper noun signifies an individual being, whereas a common noun signifies a general substance. As time has gone by, philosophers (e.g. John Stuart Mill, Gottlob Frege, Bertrand Russell, and Saul Kripke) and linguists (e.g. Otto Jespersen, Alan Gardiner, and John Algeo) have put forward a number of criteria for the semantic description of proper nouns with reference to the traditional assumption of a proper–common dichotomy. One of the earliest and most famous criteria was introduced by the English philosopher John Stuart Mill in A system of logic (1843). He suggests that a pure proper noun is a denotative sign without connotation. In contrast to a common noun that both denotes and implies attributes, a proper noun is used to denote a specific or unique individual entity regardless of what properties it has. For example, a proper noun like Victoria simply denotes the individuals who are called

878

by Victoria. It does not indicate any attributes as belonging to these individuals. Sometimes, we do not even know whether it is a woman’s name or a place name. With reference to the semantic information about the entities, pure proper nouns are lexically opaque and the entities they designate are normally unpredictable. On the other hand, a common noun is used to subsume similar entities under a generic concept. For example, various books, irrespective of material, size, or purpose, are subsumed under the generic concept book. Thus, the use of a common noun for an entity is completely determined by the meaning of the noun. Common nouns such as chair are semantically transparent because they have lexical meaning and carry information. It has been argued that while proper nouns have no meaning in isolation, they carry descriptive meaning when applied in a specific context to a particular person or place. For example, the proper noun John is used in English-speaking countries as a name for a male. However, this culturally specific information should not be identified with the meaning carried by a common noun. The information associated with a proper noun is usually changeable, whereas the meaning of a common noun is basically constant.

PROPER NOUNS Within a given context of utterance, proper nouns together with pronouns are two distinct kinds of singular definite referring expressions that enable the hearer to denote the actual referent from the class of potential referents. Unlike pronouns that have some descriptive meaning, proper nouns have reference but not sense. They identify their referents by making use of the unique and arbitrary association that holds between a proper noun and its referent. Moreover, unlike pronouns that are used by different speakers to denote different entities in different contexts, proper nouns are always claimed in any particular context to refer to one and only one entity, although a proper noun may name many individuals. It follows that a proper noun merely serves as an identification mark by singling out an entity from among similar entities. Its function is precisely that of labeling. In the 1970s, the American philosopher Saul Kripke (1972, 1980) espoused the view that proper nouns are ‘rigid designators’ that designate the same individual in all possible states of the world whatever their characteristics happen to be. According to Kripke, the proper name John is a rigid designator because it picks out particular individuals who happen to be John without any defining characteristics in whatever state of affairs we entertain. However, the common noun chair refers to the same thing in the world that happens to fit the defining characteristics. In some languages such as English, it has long been held that the definition for typical proper nouns should also be laid down on the basis of three additional criteria: orthographic, morphological, and syntactic. Orthographically, it is a graphic convention that English proper nouns are marked with initial capitalization. By the morphological criterion (i.e. the relationships between the parts that make up the proper noun itself when it is more than a single morpheme), a typical English proper noun has no plural suffix in the strictest sense e.g. Britain—*Britains. By the syntactic criteria (i.e. the relationships of the proper noun with other items in the same grammatical construction), typical English proper nouns are not preceded by the article or other determiners since the references that proper nouns have is by their very nature self-determining or inherently definite. Unlike common nouns, proper nouns normally lack the contrastive definiteness (e.g. *an America, *some Americas; an orange, some oranges). Another major syntactic criterion is that proper nouns lack modification. When they have the normal unique denotation, they can only accept nonrestrictive modifiers (e.g. Mary, who is my sister, is studying abroad—*Mary who is my sister is studying abroad.). Proper nouns require the when they are modified by a restrictive clause (e.g. The Mary I know is my friend’s wife), but not when they are unrestricted.

While numerous philosophers and linguists have considered proper nouns and common nouns as two discrete categories and the boundary between them as unambiguous, Jespersen (1924) holds that no sharp line can be drawn between proper and common nouns, the difference being one of degree rather than of kind. Later, Quirk et al. (1985) explicitly stated that the class of proper nouns has unclear boundaries and the degree of membership involves the notion of gradience. There are some circumstances in which proper nouns (e.g. the Avon, the Crimea, the Himalayas) behave like common nouns by taking the definite article and the plural form. On the other hand, a number of common nouns with unique denotation are sometimes capitalized and thus enjoy a similar semantic function as proper nouns e.g. Fortune, Wealth, Fate. References Algeo, John. 1973. On defining the proper name. Gainesville, FL: University of Florida Press. Frege, Gottlob. 1892. Über Sinn und Bedeutung. Zeitschrift für Philosophie und philosophische Kritik, Vol. 100; as On sense and reference, translated by Max Black, in Translations from the Philosophical Writings of Gottlob Frege, ed. by Peter Geach and Max Black. Oxford: Basil Blackwell, 1960. Gardiner, Alan. 1940. The theory of proper names: a controversy essay. London: Oxford University Press; 2nd edition, 1954. Jespersen, Otto. 1924. The philosophy of grammar, London: Allen and Unwin; New York: Norton, 1965; Chicago and London: The University of Chicago Press, 1992. Kripke, Saul A. 1972. Naming and necessity. Semantics of natural language, ed. by G. Harman and D. Davidson. Dordrecht, Netherlands and Boston, USA: D. Reidel Publishing Co.; as Naming and necessity. Oxford: Basil Blackwell, revised edition, 1980. Lyons, John. 1977. Semantics, 2 vols. Cambridge: Cambridge University Press. Marmaridou, Sophia A.S. 1991. What’s so proper about names? A study in categorisation and cognitive semantics. Athens: University of Athens. Mill, John Stuart. 1843. A system of logic, ratiocinative and inductive, being a connected view of the principles of evidence, and the methods of scientific investigation. London: John W. Parker; in Collected works of John Stuart Mill, Vol. VII, ed. by J.M. Robson. Toronto and Buffalo: University of Toronto Press, and London: Routledge and Kegan Paul, 1973. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985. A comprehensive grammar of the English language. London and New York: Longman. Russell, Bertrand. 1950. An inquiry into meaning and truth. London: George Allen and Unwin; London and New York: Routledge, revised edition, 1995. Ullmann, Stephen. 1962. Semantics: an introduction to the science of meaning. Oxford: Basil Blackwell. Zabeeh, Farhang. 1968. What is in a name? An inquiry into the semantics and pragmatics of proper names. The Hague: Marinus Nijhoff.

GRACE YUEN WAH TSE See also Naming; Reference

879

PROSODY

Prosody Prosody refers to the (variations in) rhythm, intonation, and stress patterns in speech. Its role is to emphasize words, to segment a sentence into smaller units, to change the meaning of an utterance (not of a speech sound!), and to resolve syntactic disambiguity (e.g. ‘They fed her dog biscuits’ Who ate the biscuits? She or the dog?). Apart from these language-dependent aspects, prosody is also used in a more universal manner, to provide cues to the state of the speaker and to convey the speaker’s mood (happiness, impatience, boredom, etc.). Prosodic patterns are very important during speech development. They appear in the infant’s speech long before the first word: babies often speak nonsense with the correct intonation! Moreover, as boundaries between words are not clearly marked in speaking, infants use prosodic information to ‘segment out’ words in the continuous stream of speech sounds. This process is termed ‘prosodic boot-strapping’. And the importance of prosody also becomes clear when listening to foreign speakers: even if they articulate well, they are not easily understood if their use of stress and timing is incorrect. The same can be said of deaf speakers: their speech is often unintelligible because they cannot master the prosody of their language well. In noisy environments prosody helps the listener to understand the speaker. Moreover, spectrally distorted speech with the original prosody is easier to understand than the original speech signal with distorted prosody. Prosody does not only contribute significantly to the intelligibility of speech but also to its naturalness. Although present-day speech synthesis is highly intelligible, it still sounds artificial. More research on speech prosody is needed to make it sound natural or pleasant enough to listen to for long stretches of time. The most important prosodic features are intonation, stress, and quantity. The acoustic correlates of these features are fundamental frequency (perceived as pitch), intensity, and duration (although there is no one-to-one correspondence).

Intonation Variations in tone (word level) and intonation (sentence level) are used to change the meaning of words, to change the function of the utterance (declarative vs. interrogative), to signify attitude and emotion, etc. These changes in fundamental frequency are related to the rate of vibration of the vocal folds. An increase in

880

subglottal pressure and an increase in the tension of the laryngeal muscles (thereby increasing the stiffness of the folds and reducing their local mass) result in an increase in the fundamental frequency. Prosody can be used grammatically, by producing a question with a rising pitch, and a statement with a falling/flat pitch. Intonation also helps to group words, especially in syntactically ambiguous phrases (e.g. a light house keeper). However, there is quite a complex interaction between stress, intonation, and duration in the case of syntactic ambiguities. At some word boundaries, such as between ‘gray tie’ vs. ‘great eye’, duration is the primary cue, although the ambiguity can also be resolved through differentiation in intonation.

Stress Stress usually refers to relative prominence and increased intensity due to increased physical effort. Note, however, that stress is not perceived as loudness, and that intensity is not the only acoustic correlate of greater effort. An increase in stress often cooccurs with an increase in fundamental frequency. After all, both stress and intonation are brought about by an increase in respiratory effort and subglottal pressure. The minimum size of the unit of stress placement is the syllable. At the word level, stress on the first or second syllable is used to differentiate meaning (cf: ‘DIgest’ and ‘to diGEST’ or RECord and to reCORD). This is termed lexial stress. At the sentence level, stress is used to draw attention to certain words (emphatic stress) or to differentiate between two words (contrastive stress). Languages are spoken with a certain rhythm. English is a stressed-timed language due to the fact that stressed syllables occur at regular intervals. French and Japanese are referred to as syllable-timed languages because the duration of their syllables remains relatively constant. Stressed syllables have a higher intensity and fundamental frequency than their unstressed counterparts. Apart from these voice source factors, the timing of the articulatory movements of the vowels and consonants produces differences in duration, depending on the position of the utterance in a word or phrase. In many languages, stressed syllables are longer than unstressed ones. The acoustical correlates of stress (especially duration) are, however, language dependent: in Czech, there is hardly an increase in duration on stressed syllables.

PSYCHOLINGUISTICS

Quantity Quantity is the linguistic term for ‘constrastive duration’. Each individual speech sound has an ‘intrinsic duration’ to specify its identity, but the duration of speech segments or larger units can yield contrastive information, depending on context or the phonological system of the language. For example, the duration of vowels in Finnish can be constrastively short and long. In Norwegian, a long vowel is followed by a short consonant and vice versa. In English, the duration of a vowel preceding a voiced consonant is 1.5 times greater than that of the same vowel preceding a voiceless consonant (sight197 ms, side297 ms; Peterson and Lehiste 1960). Vowel duration, and consonant duration to a lesser extent, is also affected by the number of syllables in a word, by syllable stress (where it interacts with fundamental frequency), and by the type of word (function words such as ‘in’ or ‘on’ are relatively short). Very often, the last syllable of a word or the last word of an utterance is longer than the same syllable or word in other positions. This temporal effect, called prepausal lengthening, results from a slowing down in the speaking rate and it signals phrase and sentence boundaries. At word level, duration can signal the syntactic structure of a sentence, e.g. ‘a gray tie’ vs. ‘a great eye’. At a sentence level, relative duration does not affect the meaning of individual words.

However, significant changes in tempo may convey information about the mood of the speaker or the circumstances under which the utterance is made. References Cooper, W., S. Eady, and P. Mueller. 1985. Acoustic aspects of contrastive stress in question–answer context. Journal of the Acoustical Society of America 77. 2124–56. Cutler, A., and D.R. Ladd. 1983. Prosody: models and measurement. New York: Springer-Verlag. Cutler, A. 1991. Linguistic rhythm and speech segmentation. Music, language, speech, and brain, ed. by J. Sundberg, L. Nord, and R. Carlson. Houndmills and London: MacMillan. Lehiste, I. 1970. Suprasegmentals. Cambridge, MA: MIT Press. Lehiste, I. 1972. The timing of utterances and linguistic boundaries. Journal of the Acoustical Society of America 51. 2018–24. Lehiste, I. 1996. Suprasegmental features of speech. Principles of experimental phonetics. Mosby-Year Book, 226–44. St Louis, MO: Mosby. Maeda, S. 1976. A characterization of American English intonation. Cambridge, MA: MIT Press. Peterson, G.E., and I. Lehiste. (1960). Duration of syllable nuclei in English. Journal of the Acoustical Society of America 32. 693–703. Terken, J. 1991. Fundamental frequency and perceived prominence of accented syllables. Journal of the Acoustical Society of America 89. 1768–76.

ASTRID VAN WIERINGEN See also Soundwave Analysis

Psycholinguistics Psycholinguistics is the discipline of psychology that studies the mental mechanisms of language processing—speaking, listening, reading, and writing—in both a native and a second tongue. Psycholinguistics also studies the processes underlying the acquisition of language, how language processes break down in language pathologies such as dyslexia and aphasia, and how these processes relate to brain function. Psycholinguistics borrows many of its theoretical constructs from linguistics. Levels of processing, distinguished in theories of language comprehension or language production, correspond to linguistic levels, such as semantics (meaning), the lexicon (vocabulary), syntax (sentence structure), morphology (units that make up words), phonology, and phonetics (sound systems). Furthermore, the processing units that these theories assume correspond to linguistic units, such as the phoneme, the syllable, the mor-

pheme, and the clause. Psycholinguistic experiments provide information about the psychological reality of linguistic units and the way linguistic information is represented and processed in the mind of the language user. A crucial difference between linguistics and psycholinguistics is the latter’s focus on the mental process. For instance, in natural speech we can easily produce two words per second. How do speakers find these words so quickly in their (vast) mental lexicon, which contains at least 10,000 words? What happens to these processes when we are unable to say a word, but it is on the tip of our tongue? As another example, when listening to sentences, we can sometimes be led up the ‘garden path’. A classic example is ‘the horse raced past the barn fell’. Here, one builds up a structural representation of the sentence, but upon reading the last word (‘fell’), it turns out to be wrong. How do

881

PSYCHOLINGUISTICS readers build up a structural representation? What went wrong with this process in the example? The scientific inquiry into language processing began in the second half of the nineteenth century. Rudolf Meringer (1859–1931) started to collect speech errors. In contrast to his contemporary Sigmund Freud (1856–1939), he studied the linguistic, rather than psychological, properties of such incidents. Wilhelm Wundt (1832–1920) proposed the first theory of language production, and Gustav Aschaffenburg (1866–1944), by using the technique of word association, investigated the representation of word meaning. The Dutch ophthalmologist Fransiscus Donders (1818–1889) introduced the method of mental chronometry (measuring the time mental processes take). Reaction time studies are still the most common methodology in psycholinguistics. Around the same time, the French neurologist Paul Broca (1824–1880) tested a patient who had lost his ability to speak. Monsieur Tan, as the patient was known, could produce only a single syllable (‘tan’). After Mr. Tan’s death, Broca examined the brain of this patient and discovered massive damage to an area in the front part of the brain, which is now known as Broca’s area. Broca’s discovery, and subsequent studies by the German neurologist Carl Wernicke (1848–1905), marked the beginning of the inquiry into the relationship between language and brain. In the first half of the twentieth century, psychology was dominated by behaviorism. This approach rejected the notion of ‘mental representation’ and ‘mental process’ but focused on what is directly observable: behavior. An exponent of behaviorism was Burrhus F. Skinner (1904–1990), who published his Verbal behavior in 1957. In Skinner’s approach, the language-acquiring child begins without any linguistic knowledge. It learns language by experience only. Sentences would be produced as associative chains, where one word would be the stimulus that triggered the next word. A few years later, a young linguist from the Massachusetts Institute of Technology, Noam Chomsky (1928–), published an extremely sharp criticism of Skinner’s book. Chomsky argued that Skinner’s theory could not work—for instance, a child is exposed to far too few examples of sentences to learn the language by experience only. Instead, Chomsky argued that the child is born with a tacit knowledge of ‘universal’ grammar, and learning means tuning this grammar to its own language. Around the same time, psychologists were abandoning behaviorism and became interested in linguistics again. In 1951, a conference was organized at which leading psychologists and linguists discussed the possibilities of ‘merging’ the fields. At this conference, the term ‘psycholinguistics’ was launched.

882

Research in the 1960s and 1970s embraced Chomsky’s theory of transformational grammar. This theory provided representations of sentences (called deep structure and surface structure), and procedures that transformed these representations. However, little convincing evidence was found for a correspondence between linguistic transformations and psychological processes, and this approach was largely abandoned in the early 1980s. Research in the 1980s and the 1990s was characterized by three major developments. First, Jerry Fodor published his The modularity of mind (1983), in which he presented the theory that many cognitive processes are performed by dedicated processing systems (modules) that operate independent of other processing systems. A large number of studies in modern psycholinguistics have been and still are centered on this theory. Second, psycholinguists designed computer models of language processing. Seminal work in this area was done by James McClelland and David Rummelhart (1989), in particular, on the modeling of word perception. Third, technical advances made it possible to measure the activity of the brain while it was engaged in verbal tasks. Although still in its infancy, the use of such techniques can be considered very promising in studying the representation and processing of linguistic information in the brain. What are the mental processes involved in using language? In the following paragraphs, a brief overview will be presented of theoretical issues in pivotal language processes: comprehending, speaking, and ‘control’ processes. The first challenge in comprehending is to identify words. This is no easy task: the speech signal has to be distinguished from noise, which may distort it. The speech signal is also very variable. One way of dealing with this variation is ‘categorical perception’. That is, tokens of a speech sound can be acoustically quite different, but they tend to be perceived as the same sound (e.g. as the speech sounds /b/ or /p/, but never something in between). Another challenge for the listener is that in continuous speech, there are often no pauses between words. How then do listeners know where one word ends and the next one begins? Psycholinguistic research shows that this depends on the language: different languages provide different cues to word onset (e.g. the pattern of stress), and listeners use these cues. Visual word recognition—reading words—poses its own problems. English is an alphabetic language (as opposed to, e.g. Chinese). However, the correspondence between printed letters or groups of letters and sounds is often irregular (e.g. in the word ‘yacht’). This has led psycholinguists to distinguish two routes for reading: a lexical route, in which letters or groups of letters directly access a word, and a sublexical

PSYCHOLINGUISTICS route, in which letters are transformed into a phonological code that guides word recognition. The lexical route would be used for reading irregular words; the sublexical route would be used for reading ‘nonwords’ (e.g. ‘climp’). Support for this theory comes not only from distinct patterns of reading disorder (dyslexia) but also from experiments in which the speed of reading irregular and regular words is measured. Once words are identified, we construct a structural representation of the sentence, so that we can infer who did what to whom. This process is called parsing. Many studies of parsing have analyzed structural ambiguities. In a sentence such as ‘the man hit the girl with the guitar’, the guitar can be in the possession of the man (and used for hitting the girl), or it can be in the possession of the girl being hit. These interpretations correspond to different sentence structures. Studies with ambiguous sentences have revealed some of the processes involved in parsing. A very important debate centers on the question of what kind of information is available to parsing. In some views, only syntactic information is used. According to other views, parsing is influenced by a variety of factors (e.g. plausibility, or the frequency with which a word occurs in a certain structure). Sentence understanding does not stop with parsing. We have to integrate the word meanings and the sentence structure into a sentence meaning, and we have to relate this meaning to prior discourse. In one study, brain responses were measured to sentences such as ‘I spread my bread with socks’, and it was discovered that such stimuli evoke distinct patterns in the electroencephalogram signal. This finding has proved useful for studying these integration processes. Speaking starts with the decision to engage in a communicative act. This is based on the speaker’s intentions, background knowledge, and a mental model of previous dialogue. On the basis of these sources of information, the first step in speaking is the generation of a message. This is the interface between thought and language: the message is not yet language, but it is specific to language. In most theories of language production, the message contains concepts for the things and actions the sentence is about, perspective (is ‘in front of’ relative to you or me?), and focus (which information is given, and which is new?). The next step consists of formulation of this message: retrieval of the words and construction of a sentence. There is much evidence that word retrieval consists of two steps: a process that is based on meaning and a process that is based on form. This distinction is based on research that analyzed collections of speech errors and research that measured reaction times in word production. In these tasks, so-called distracter words with either a form or a meaning relation to the target word have different effects. A major issue

is whether these data also imply that there are two separate lexical representations. Another task of formulation is the construction of a sentence structure. A key finding is the observation that sentence structure is ‘persistent’. Speakers have a tendency to reuse a structure that has been recently used before, even if the two sentences do not share any meaning, words, or prosody. This finding strongly suggests that structured representations of sentences are psychologically real. It also provides a technique for determining how sentences are mentally represented. A final task of formulation is to retrieve pronunciations. A major problem here is that words in real speech do not correspond in a one-to-one fashion to words in their citation form: in connected speech, sounds from one word are often ‘merged’ with those of another (e.g. ‘give it’ will often be spoken as ‘gi-vit’). Another issue is whether certain representations that are distinguished in theoretical phonology are psychologically real. The earlier discussion focused mainly on contentdirected processes in language use. Yet, there is increasing interest in control processes, which determine when the content-directed processes should engage, and in mental resources, such as memory and attention, which are the ‘fuel’ on which the contentdirected processes run. First, language is predominantly used in interaction with others. There is a growing body of evidence that partners in a dialogue coordinate with each other: that is, they converge on using the same linguistic structures, such as words, syntax, and description schemes, to denote the same things in the external world. How and why we accomplish such coordination is an important item on the agenda for study. Second, the language processes may contain errors. Fortunately, there are self-monitoring processes dedicated to intercepting errors and ‘repairing’ them. How these monitoring processes operate—and how it is possible that they do the job as fast as they do—is, at present, a matter of debate. Third, the language user is constrained by factors such as attention and memory. In recent years, there has been an increasing interest in how limitations in such ‘resources’ affect processing. Debate here centers on the nature of such resources: whether they are comparable to the same short-term memory in which one briefly stores the digits of a phone number, or whether there are specialized resources for language processing. References Carrol, David. W. 1999. Psychology of language. Pacific Grove, CA: Brooks/Cole. Harley, Trevor A. 1995, 2001. The psychology of language: from data to theory, 2nd edition. Hove, UK: Psychology Press.

883

PSYCHOLINGUISTICS Fodor, Jerry A. 1983. The modularity of mind. London, England and Cambridge, MA: MIT Press. Gernsbacher, Morton-Ann (ed.) 1994. Handbook of psycholinguistics. San Diego, CA: Academic Press. Levelt, Willem J.M. 1989. Speaking: from intention to articulation. London, England and Cambridge, MA: MIT Press. Brown, Colin M., and Peter Hagoort (eds.) 1999. The neurocognition of language. Oxford, UK: Oxford University Press.

884

Dijkstra, Ton, and Koenraad De Smedt (eds.) 1996. Computational psycholinguistics. London, UK: Taylor & Francis. McClelland, James, L, and David E. Rummelhart. 1989. Parallel distributed processing: explorations in the microstructure of cognition, Vol. 2: Psychological and biological models. London, England and Cambridge, MA: MIT Press.

ROBERT J. HARTSUIKER

Q Quantification A sentence such as John is happy consists of a subject (John) and a predicate (is happy). The predicate attributes the property it describes (‘being happy’) to the individual the subject refers to (‘John’). When the subject noun phrase contains a determiner (a, the, some, every, etc.), this element plays a crucial role in the predication process. The contrast in meaning between the sentences A student is happy and Every student is happy is caused by the different nature of the determiner that expresses the number or quantity of individuals to which the relevant property applies. We say, then, that these sentences differ in quantificational force. The first sentence states that there is (at least) one student who is happy in the situation under consideration. The second sentence states that every individual under consideration is happy. This semantic contrast can be expressed as a contrast in logical form, as defended by the philosophers Gottlob Frege and Bertrand Russell. In first-order logic, the determiner a/some is treated as an existential quantifier (meaning ‘there is at least one individual for which it is true that…’), and the determiner every as the universal quantifier (meaning ‘it is true for all individuals that…’). Quantifiers are assumed to bind a variable (x) in a logical statement or proposition. Going back to the above sentences, the paraphrases of their respective logical forms would be the following: There is at least one individual x such that x is a student and x is happy and For every individual x if x is a student then x is happy. The first-order logic analysis of quantification has been criticized for positing that the constituents of a

noun phrase (i.e. the determiner and the noun) contribute independently to the meaning of a sentence: the determiner quantifies over the whole proposition of which the noun is a part. This runs counter to the common linguistic analysis of sentential expressions. Another limitation is that the first-order logic analysis is not designed to capture the content of the extensive variety of natural language determiners. For example, it has been shown that the English determiner most is not expressible in first-order logic. The sentence Most students are happy does not mean that most individuals under consideration are students and they are happy (first-order logical meaning). Rather, it means that most of those individuals who are students are also happy. Generalized Quantifier Theory developed during the 1980s, following the initial contribution of the philosopher Richard Montague, as an attempt to improve the shortcomings of traditional logical analyses of quantification in natural language. The functional nature of the subject—predicate relation that we described at the beginning is reversed. A predicate still expresses a property, but it is not treated as a function that applies directly to the subject (argument). Rather, a subject is viewed as a higher-order expression that expresses a property of the predicate or, equivalently, a set of properties. In general, noun phrases express generalized quantifiers (sets of properties). We can represent the sentences in the previous paragraph compositionally as some student (happy) and every student (happy), where the generalized quantifiers associated with some student and every student are respectively

885

QUANTIFICATION interpreted as the set of properties that some student or every student has, and what the sentence asserts is that happiness is one of those properties. The sentence will be true or false depending on whether this is the case or not. The meaning of a noun phrase such as some student can also be determined in a compositional fashion. The noun student expresses or describes, more technically ‘denotes’, a set of individuals and the determiner some relates it with another set of individuals, namely, that denoted by happy. This yields the following logical representation: a/some (student) (happy). Given the particular meaning of the determiner some, this expression is true if the intersection of the sets denoted by students and happy is not empty, i.e. if there is at least one individual who is a student and who is happy. Other determiner meanings can be characterized in a similar fashion: every (student) (happy) is true if the set denoted by student is a subset of the set denoted by happy; no (student) (happy) is true if the intersection of the sets denoted by students and happy is empty; more than three (students) (happy) is true if the number of individuals in the intersection of the sets denoted by students and happy is greater than three; most (students) (happy) is true if the number of individuals in the intersection of those two sets is greater than the number of the individuals in the intersection of the sets denoted by students and not happy, i.e. the number of those students who are happy is greater than the number of those students who are not happy; etc. In sum, a natural language determiner is a function with two arguments. The first argument corresponds to the noun denotation and the second argument to the content of the sentence verb phrase. Natural language determiners also satisfy a series of constraints that set them apart from their logical counterparts. For example, all natural language determiners are conservative or live on their first argument—motivating the intuition that the determiner has a close bond with the noun. The sentence Some sailors left is equivalent to the sentence Some sailors are sailors who left. If we substitute any other determiner for some in this sentence, the equivalence still holds. The effect of this constraint is to make natural language quantification inherently restricted to the first argument of the determiner. Thus, in checking whether some (sailors) (left) is true, we do not have to consider those individuals who are not sailors. In other words, in processing the sentence Some sailors left, we do not first check who left and then determine whether any of these individuals are sailors; rather, we look at relevant sailors and determine whether some of them left.

886

Determiners are also characterized by the type of inferences that they license, namely, set-to-subset inferences or set-to-superset inferences. For example, the sentence No students smoked entails the sentence No students smoked cigars. The property denoted by smoked cigars is a subset of the property denoted by smoked, so the determiner no licenses set-to-subset inferences for its second argument. The same is true for its first argument: the sentence No students smoked entails the sentence No female students smoked, given that the property denoted by female students is a subset of the property denoted by students. The determiner function some has the opposite pattern and licenses set-to-superset inferences for its two arguments. Thus, we predict that the sentence Some female students smoke cigars entails both Some female students smoke and Some students smoke cigars. Quantifiers can also be classified according to whether they can occur in an existential construction, i.e. a sentence that asserts the existence of something (There is…). Consider the structure There is/are Q student(s) in the garden. The determiners some, three, no, fewer than five, and many can be substituted for Q and occur in this construction, whereas the occurrence of determiners such as every, most, and all but three would make the sentence ungrammatical. The determiners that can occur in an existential construction are intersective, i.e. they express a relation of intersection of their two arguments. On the other hand, those determiners that cannot occur in an existential construction are characterized by expressing a relation of inclusion, such as every, or proportionality, such as most. Definite determiners, whether simple (the) or complex (the ten…), demonstrative (this) determiners, and possessive determiners (my) do not occur in existential sentences either. These determiners are all inherently context dependent and thus presuppose or do not assert existence. When more than one quantifier occur in a sentence, a form of semantic interaction called scope arises. Scopal relations are determined by the different order of quantifiers in the semantic representation of a clause. For example, the sentence Every student read a book is ambiguous because it contains two generalized quantifiers: every student and a book. Under one interpretation, every student read a different book. Under the second, there is a unique book such that every student read it. This ambiguity is a genuine scope ambiguity. In the first reading, the scope order of the quantifiers is the one that respects the linear order of the noun phrases; i.e. the subject noun phrase is more dominant or has scope over the object: every

QUANTIFICATION student  a book. The universal generalized quantifier every student takes scope over the existential generalized quantifier expressed by a book. The second reading is an inverse scope reading in that the scopal order of the quantifiers differs from the surface linear order. In this reading, the existential quantifier takes scope over the universal quantifier: a book  every student. So far, we have considered only nominal quantification, i.e. quantification corresponding to the meaning of noun phrases. However, other elements may also contribute to the quantificational force of a sentence. Adverbs such as always, sometimes, and often are not merely temporal adverbs. In the majority of contexts, they behave as adverbs of quantification. For example, the sentence Peter always drinks coffee does not mean that Peter drinks coffee at every moment, but rather that whenever Peter drinks, he drinks coffee. Adverbs of quantification can thus be taken to express quantification over situations. Interrogative generalized quantifiers are wh-phrases such as who, what, where, and which. Interrogative determiners are also conservative. The interrogative sentence Which students are rich? is equivalent to Which students are students who are rich? Interrogative determiners are also uniformly intersective, because the meaning of an interrogative determiner depends on the intersection of its two arguments. For example, a complete answer to which (students) (rich) specifies, in a situation, the intersection of the set of students and the set of rich individuals in that situation. Interrogative determiners and existential determiners such as some have a majority of their properties in common. From this point of view, it is not surprising that a significant number of the world’s languages use the same lexical expression for interrogatives and existential determiners. Sometimes, the quantificational force of a sentence cannot be associated with any overt element in the clause. The sentence Lions are fierce is usually interpreted as stating that most lions are fierce or that lions are normally fierce. To characterize the semantic properties of this sentence, the existence of hidden elements of diverse quantificational force has to be postulated. In the example under consideration, there is a hidden element of generic force, Gen, so the logical representation of this sentence would be Gen (lions) (fierce). Nominal and adverbial quantification interact in apparently unexpected ways. This interaction arises mostly when an indefinite occurs in the scope of an adverb of quantification in relative constructions, conditional sentences, etc. In the sentence Every

farmer who has a donkey beats it, the indefinite a lacks its typical existential force. The sentence does not mean that every farmer beats one donkey or other (at least one) that he has. Rather, the correct interpretation is that for every pair consisting of a farmer and a donkey owned by that farmer, it is also the case that the farmer beats the donkey. The indefinite seems to have universal force here; so we may conclude that the universal quantifier headed by every behaves as a binary quantifier here and associates both with the restriction of the universal noun phrase and with the indefinite one. The same pattern can be observed in the sentence Always, if a farmer owns a donkey, he beats it, where the indefinite noun phrases a farmer and a donkey seem to inherit the universal force from the adverb always. Some theories have concluded that indefinites lack quantificational force and inherit their apparent force from other surrounding elements in the clause. In general, adverbs of quantification are taken to transmit their quantificational force to all the indefinites that appear in their scope. This phenomenon is known as unselective quantification. Quantification thus covers a wide range of interesting phenomena and has triggered many debates among linguists and philosophers. References Benthem, Johan van, and Alice ter Meulen (eds.) 1985. Generalized quantifiers in natural language. Dordrecht: Foris. Gärdenfors, Peter (ed.) 1987. Generalized quantifiers: linguistic and logical approaches. Dordrecht: Reidel. Gutiérrez-Rexach, Javier. 1998. Semántica lógica y cuantificación nominal. Introducción a la teoría de cuantificadores generalizados. New Orleans: University Press of the South. ——— (ed.) 2003. Semantics: critical concepts. London: Routledge. Hawkins, John. 1978. Definiteness and indefiniteness. London: Croom Helm. Heim, Irene. 1982. The semantics od definite and indefinite noun phrases. Ph.D. dissertation. University of Massachusetts, Amherst. Kamp, Hans, and Uwe Reyle. 1993. From discourse to logic. Dordrecht: Kluwer. Keenan, Edward, and Leonard Faltz. 1985. Boolean semantics of natural language. Dordrecht: Reidel. Lappin, Shalom. 1996. Handbook of contemporary semantic theory. Oxford: Basil Blackwell. May, Robert. 1987. Logical form. Cambridge, MA: MIT Press. Reuland, Eric, and Alice ter Meulen (eds.) 1987. The representatio of (in) definiteness. Cambridge, MA: MIT Press. Szabolcsi, Anna. 1997. Ways of scope taking. Dordrecht: Kluwer.

JAVIER GUTIÉRREZ-REXACH

887

QUESTIONS

Questions The term ‘questions’ in everyday usage refers to utterances inviting a response and that, when spoken, may have specific intonation patterns. Questions can be analyzed for their grammatical forms and also for their functions in discourse. Discussion of questions generally involves both these aspects. The analysis of questions can be used to relate linguistic findings to patterns of social behavior. In English, questions can be identified and categorized on the basis of syntactic and phonological forms. Four types of questions in English can be distinguished: (1)

Wh-questions

These are forms beginning with an interrogative element (who, which, what, why, when, how, and so on): Where does Antony work? What can Antony do? The subject (Antony) and either an auxiliary (does) or modal verb (can) are inverted in English. Wh- questions generally require more elaborate answers than those elicited by yes/no questions. (2)

Yes/no questions

This broad category refers to questions to which the hearer is expected to respond either with yes or no. The following forms can be identified: a. Inversion of the subject and an auxiliary or modal verb Does the race start here? Can we finish this now? b. Statements with a question intonation In these questions, the speaker presents an utterance in the form of a statement, but marks it as question by a rising and falling tone. You will be sure to bring it back soon? (3)

Alternative forms

In these forms, the speaker presents two or more alternative answers within the question. Will the new administrator work here or on the first floor? (4)

Tag forms

These forms consist of a declarative utterance followed by a question. The computer is connected to the internet, isn’t it? Yes/no questions are said to either be neutral or biased toward a positive or negative orientation. For

888

example, the use of the nonassertive form ‘anyone’ in Has anyone left? can be argued to render this question neutral in that it leaves open whether the response will be affirmative or not. However, the use of the more assertive form ‘someone’ in Has someone left? is argued to indicate speaker expectation of a positive response, whereas the use of a negative form hasn’t in Hasn’t anyone left? is argued to indicate a bias toward a negative response. Generally, questions are understood to be based on presuppositions, or underlying assumptions. For example, the question When does your brother finish college? is based on the presuppositions that (1) you have a brother and (2) your brother is at college. Questions can be examined in relation to their pragmatic functions, namely, their functions to request specific types of linguistic responses. A discourse-based analysis of questions involves a consideration of contextual factors, such as the situation, who knows what, and how the discourse unfolds. Identification of the functions is the primary interest, but identification of the forms that typically realize these functions may follow from this. From a pragmatic viewpoint, a question can be identified as a question if it fulfills certain conditions: the act of questioning must be genuine, and the speaker must believe in the presuppositions of the question and desire to know the response. For example, a speaker (sitting in the driver’s seat in a car with some passengers at the back) produces the following utterance: Are you ready? Without a response, the speaker then sets forth on the journey. This indicates that the driver does not desire to know the response. Also, the absence of any response from the passengers indicates that the hearers do not interpret the driver’s utterance as seeking a response. Thus, despite the question-like form, the utterance would not be termed a question. In discourse-based analyses, questions are often referred to as ‘elicits’ or ‘requests’, and they are categorized in relation to the specific functions they require of the hearer. One such category is elicits of, or requests for, information. In the following example (data taken from Schegloff 1972:107), A requires information about the location of an address from B. A I don’t know where the—uh—this address is. B Well, where do—-which part of the town do you live in?

QUINE, WILLARD VAN ORMAN Another category in a discourse-based analysis is that of elicits of, or requests for, confirmation. These function to invite the hearer to verify or disconfirm the speaker’s idea or proposition, as the following example taken from Coulthard and Brazil (1981:84) illustrates. A So the meeting’s on Friday B Thanks A No I’m asking you This example also illustrates that identification of the exact function may be ambiguous and can only be resolved as the discourse unfolds. A’s first utterance asks B to confirm that the meeting is on Friday. However, the function of this utterance is unclear, and B understands it as information being provided by A. B thus thanks A for the information. In utterance 3, A then makes explicit the function of the first utterance as an elicit of confirmation. Other discourse functions of questions have been identified. Elicits of commitment require the hearer to promise further action (e.g. You will come on Wednesday?). Elicits of agreement require the hearer to verify that the idea proffered by the speaker is selfevidently true. The following example taken from Tsui (1992:107) illustrates: (On a sunny day) A Lovely day, isn’t it? B Yes, beautiful! The analysis of questions according to function has been applied to the study of language use in specific situations. For example, analysis of teacher talk has been shown to reveal various didactic functions. These didactic functions include higher cognitive level questions requiring the hearer to manipulate bits of previously learned knowledge to create a response, lower cognitive level questions requiring the hearer simply

to recall or recognize factual information, display questions functioning to check or test the student on information known to the teacher, and referential questions seeking information unknown to the teachers themselves. Conversation analysts have analyzed questioning to identify rules underlying interaction. Questions are identified as the first part of a two-part sequence: question (Q) and answer (A). The person who asks a question has the option of speaking again to ask another question. Thus, there is a chaining rule providing for the sequence Q-A-Q-A-Q-A…. This approach to the analysis of questions has been used to relate linguistic findings to power structures in society. For example, the analysis of interaction may reveal one speaker asking more questions, thus directing the discourse and indicating asymmetrical power relationships. References Aijmer, K. 1996. Conversational routines in English: convention and creativity. New York: Longman. Coulthard, R.M., and D.C. Brazil. 1981. Exchange structure. Studies in discourse analysis, ed. by R.M. Coulthard, and M.M. Montgomery, 82–106. London: Routledge and Kegan Paul. Dillon, J.Y. 1990. The practice of questioning. London: Routledge. Greenbaum, S., G. Leech, and J. Svartik. 1972. A grammar of contemporary English. London: Longman. Schegloff, E.A. 1972. Notes on conversational practice: formulating place. Studies in social interaction, ed. by D. Sudnow, 75–119. New York: Free Press. Tsui, A.B.M. 1992. A functional description of questions. Advances in spoken discourse, ed. by M. Coulthard, London: Routledge. ——— 1994. English conversation. Oxford: Oxford University Press. van Lier, L. 1988. The classroom and the language learner. Harlow Essex: Longman.

HELEN BASTURKMEN

Quine, Willard van Orman Quine is one of the most important analytic philosophers of the twentieth century, a fact eloquently attested to by the fact that in 1987 a supplement to the Oxford English dictionary recognized his importance as an eponym by including the entry ‘Quinean’ with the following definition: ‘Of, pertaining to, or characteristic of Willard van. Quine or his theories’. At the

time of his death in December 2000, Quine was the Edgar Pierce Professor Emeritus at Harvard, where he had pursued an active career spanning over half a century. His early work was heavily influenced by Rudolf Carnap and other logical positivists. But, as an empiricist and (from the 1960s on) an advocate of a naturalistic epistemology, he distrusted all talk of meaning

889

QUINE, WILLARD VAN ORMAN equivalence, which was a mainstay of Carnap’s reductionist program. A prolific writer, Quine is widely perceived as a philosophical iconoclast. His early essay ‘Two dogmas of empiricism’ (published in the 1953 collection) had great impact on the philosophical community. In it, Quine challenged the distinction between ‘analytic’ and ‘synthetic’ statements, one of the most fundamental and well-entrenched distinctions of modern philosophy—originally introduced by Kant but already foreshadowed in the work of Locke and Leibniz. According to Kant, an analytic statement was one whose predicate was contained in the concept of its subject. In the case of a synthetic statement, there was no such relation of inclusion between the subject and the predicate. Post-Kantian interpretation had resulted in an extension of the distinction to cover cases other than the simple subject–predicate statements originally envisioned by Kant and, with it, the key criterion for analyticity had become the idea of a sentence being true by virtue of its meaning. The idea had been embraced by the logical positivists of the Vienna Circle who had also extended its scope to include, among other things, all propositions of mathematics. Quine’s attack on the ‘analytic/synthetic’ distinction was based on his rejection of the notion of ‘synonymy’, which he thought was a philosophical mare’s nest. Synonymy was suspect because it rested on the notion of meaning that too was, Quine argued, incapable of surviving critical scrutiny, notwithstanding the widespread use of the term in philosophical circles. Whatever heuristic value there was to the distinction was due to the fact that understanding a sentence was largely a matter of knowing what experiences would make a given sentence true, a claim long made by empiricists—the so-called analytic statements simply being, under this interpretation, those that have been confirmed on all known occasions. Unlike the early Wittgenstein and the positivists, Quine saw philosophy as a natural ally of science. In his view, a scientific theory was an intricate web of ideas whose borders touched on concrete experience. No part of the web was immune to revisions in the light of new experience. Furthermore, revisions are never strictly localized but rather affect several parts of the web at once. This meant that individual sentences, considered in isolation, have no meaning, strictly speaking. In addition, Quine held that in talking about the world one characteristically moves from talking in certain terms to talking about them, that is, from so-called material mode to formal mode—a process that he famously referred to as ‘semantic ascent’. On Quine’s holistic view of meaning and

890

evidence, an entire network of beliefs and theoretical predilections is involved every time a scientific hypothesis is put to test. This means, he went on to argue, that no scientific hypothesis can be tested in isolation (this claim is often referred to as ‘Quine–Duhem thesis’), just as no single sentence can be said to lend itself to an analysis of its putative meaning. Skepticism concerning meaning and with it a whole array of assorted concepts such as property, proposition, and necessity as well as the very concept of a concept led Quine to unveil one of his most influential and controversial doctrines, namely the thesis of the ‘indeterminacy of translation’. Quine argued that there can be no such thing as radical translation, i.e. translating, as it were, from scratch between two languages totally foreign to each other. A radical translator or interpreter will only have access to certain external facts, but these will be of no help when it comes to deciding between alternative candidates for the translation of an expression from one language to a radically different one. The impossibility of radical translation stems from what Quine called the ‘inscrutability of reference’ (later renamed ‘indeterminacy of reference’). The native word ‘gavagai’ uttered by the informant as he points at a rabbit jumping past does not authorize one to hypothesize that the two words ‘gavagai’ and ‘rabbit’ are translation equivalents; a number of alternative hypotheses, such as that the informant might have been referring to specific rabbit parts or determinate rabbit movements, cannot be ruled out as equally plausible in light of the available empirical evidence. At a broader level of application, Quine’s position meant that, when different theories all seem equally adequate in explaining a given set of behavioral patterns, there can be no choosing among them on grounds of greater psychological reality or whatever. In an important paper titled ‘Methodological reflections on current linguistic theory’, Quine challenged Noam Chomsky’s contention that the best theory is the one that is not only observationally and descriptively adequate, but explanatorily so as well. Quine insisted that, given a native speaker who presumably has a consummate command of his or her language and a hypothetical language learner whose linguistic performance is indistinguishable from that of the former, there is no sense in which the former may be claimed to know the language in any deeper sense than the latter. Quine’s point was that, insofar as they are meant to be explanations for certain verifiable behavioral phenomena, the ‘theory’ that the language learner presumably brings to bear on the task of speaking the foreign language is just as good as the

QUINE, WILLARD VAN ORMAN one the native speaker is being claimed to have internalized. Accordingly, Quine dismissed Chomsky’s celebrated notion of ‘linguistic competence’ as scientifically untenable. In the final analysis, Quine’s objection to such notions had to do with his firm conviction in the irreducibly public nature of language. Toward the end of his career, he became more and more convinced that one can only learn language from experience with the world at large and that, therefore, there is no way linguistics can avoid embracing behaviorism. Quine will be remembered as one of the most challenging and dynamic philosophers of the second half of the twentieth century. His hobbies included traveling and collecting stamps. He could never bring himself to use a computer, contenting himself instead with a 1927 Remington typewriter whose keyboard had to be modified to suit his special requirements of characters and type faces.

Biography Quine was born in Akron, Ohio on June 25, 1908. He graduated from Oberlin College (Oberlin, Ohio) in 1930, and in 1932, barely two years later (an all-time Harvard record), received his doctorate from Harvard, with the famous English philosopher and mathematician Alfred North Whitehead as adviser. He traveled to Europe on a Sheldon Traveling Fellowship and spent the next year in Vienna, Prague, and Warsaw, studying, lecturing and meeting various members of the Vienna Circle, among them Philip Frank, Moritz Schlick, Alfred Tarski, A.J. Ayer, and Rudolf Carnap as well as Kurt Gödel. He started his teaching career at Harvard in 1936 as an Instructor in Philosophy and taught there ever since, except for four years in the US Navy during World War II, serving in Washington and working with cryptoanalysts trying to break the German submarine code. He was promoted to Associate Professor (1941), Professor (1948), and Edgar Pierce Professor (1956). He paid two visits to Oxford: in 1953–1954 as Eastman Visiting Professor, and in 1973–1974 as Savile Fellow of Merton College and Wolfson Lecturer. He was awarded 18 honorary degrees by international institutions, including University of Lille, Oxford University, Cambridge University, Uppsala University, University of Bern, and Harvard University. He received innumerable honorary fellowships and awards, including: Society of Fellows, Harvard University (Junior Fellow, 1933–1936; Senior Fellow, 1949–1978), American Academy of Arts and Sciences (fellow 1949 –), Harvard University (Chairman, Philosophy, 1952–1953), Association for Symbolic Logic (President, 1953–1955), Institute for

Advanced Studies (Princeton, NJ, 1956–1957), American Philosophical Association (President 1957), American Philosophical Society, member (1957 –), Centre for Advanced Studies in the Behavioral Sciences (Palo Alto, CA, 1958–1959), British Academy corresponding fellow (1959–), Instituto Brasileiro de Filosofia corresponding member (1963–), Centre for Advanced Studies (Wesleyan University, Middletown, CT), Nicholas Murray Butler gold medal (1965), Columbia University (New York, 1970), National Academy of Sciences fellow (Washington, DC, 1977), Institut de France (1978), Norwegian Academy of Sciences (1979), F. Polacky gold medal (Prague, 1991), Charles University gold medal (Prague, 1993), Rolf Schock Prize (Sweden, 1993), and the Kyoto Prize (Japan, 1996). He retired as Edgar Pierce Professor Emeritus in 1978 from a teaching career in which his pupils had included not only influential philosophers like Donald Davidson but also the satirical songwriter Tom Lehrer and Theodore J. Kaczynski, the so-called ‘Unabomber’. He died in Boston on December 25, 2000, aged 92.

References Barrett, R., and R. Gibson, (eds.) 1993. Perspectives on Quine (Philosophers and their critics). London: Blackwell. Davidson, D., and J. Hintikka (eds.) 1969. Words and objection: essays on the work of W.V. Quine. Dordrecht, Holland: D. Reidel. Gibson, R. 1982. The philosophy of W.V. Quine: an expository essay. Florida. Guttenplan, S. (ed.) 1975. Mind and language. Oxford: Oxford University Press. Hahn, L. and P. Schilpp, (eds.) 1986. The philosophy of W.V. Quine. New York: Open Court. Orenstein, A. 1977. Willard van Orman Quine. New York: Twayne. Quine, Willard van Orman. 1934. A system of logistic. Cambridge, MA: Harvard University Press. ———. 1940. Mathematical logic. New York: Norton. ———. 1950. Methods of logic. New York: Holt. ———. 1953. From a logical point of view. Cambridge, MA: Harvard University Press. ———. 1960. Word and object. Cambridge, MA: MIT Press. ———. 1963. Set theory and its logic. Cambridge, MA: Harvard University Press, Belknap Press. ———. 1970. Philosophy of logic, Foundations of Philosophy Series, Englewood Cliffs, NJ: Prentice-Hall. ———. 1972. Methodological reflections on current linguistic theory. Semantics of natural language, ed. by D. Davidson and G. Harman. Doredrecht, Holland: D. Reidel Publishing Co. ———. 1974. The roots of reference. The Paul Carus Lectures, Vol. 14. LaSalle, IL: Open Court. ———. 1976. The ways of paradox, and other essays. Cambridge, MA: Harvard University Press. ———. 1977. Ontological relativity. New York: Columbia University Press.

891

QUINE, WILLARD VAN ORMAN ———. 1981. Theories and things. Cambridge, MA. and London: Harvard University Press. ———. 1989. Quiddities: an intermittently philosophical dictionary. Cambridge, MA: Harvard University Press. ———. 1992. Pursuit of truth. Cambridge, MA: Harvard University Press. ———. 1995. From stimulus to science. Cambridge, MA: Harvard University Press.

892

Romanos, G. 1983. Quine and analytic philosophy. Harvard: MIT Press. Shahan, R. and C. Swoyer, (eds.) 1979. Essays on the philosophy of W.V. Quine. Oklahoma: University of Oklahoma Press.

KANAVILLIL RAJAGOPALAN See also Philosophy of Language

R Reading Over the past several centuries, reading has gone from being a relatively obscure intellectual endeavor limited to a small elite segment of the population to a relatively common one. You are, in fact, reading at this very moment. Despite the fact that reading has become so common, and is something you probably do every day, picking a useful definition of reading turns out to be surprisingly difficult.

whether comprehension is necessary for something to be considered reading. Studies of decoding printed letter strings sometimes use nonwords rather than real words to eliminate the effects of individual differences in vocabulary knowledge. If you can read, you will have little trouble working through a list of nonwords such as HEZ, RAF, POTE, and HINKER, and the processing that is done overlaps with that done when reading real words. But without comprehension, is this reading? A pragmatic answer to the question of what is reading is to define reading as extracting meaning from print, and to acknowledge that it can be useful to study parts of the reading process such as aspects of decoding that are best viewed when people are asked to pronounce nonwords. In addition, it is important to acknowledge that successful reading requires the coordinated execution of decoding processes that are unique to processing print and comprehension processes that may be common to listening as well as reading.

What Is Reading? Reading a newspaper article surely counts as reading. But does scanning the stock market results to see how a favorite stock fared this week amount to reading? How about ‘reading’ a map? Is working through a mathematical proof, or a computer program, reading? Is it reading when you decode and pronounce a nonword such as BOPE? Definitions of reading run the gamut from ‘decoding printed symbols’ to ‘thinking with a book in front of you’. Differences in opinion about what constitutes a useful definition of reading arise from different answers to two fundamental questions. First, for something to be defined as reading, must it be unique to reading or can it be general to oral language comprehension? Reading almost always is done for the purpose of comprehending meaning, but most comprehension appears to be the same regardless of whether the message was initially processed through the eyes or the ears. Should a book on reading be mostly about language comprehension or should it be mostly about what is unique to reading as opposed to listening? The second fundamental question is

What Is Being Read? Humans have used spoken language for at least 100,000 years and perhaps quite a bit longer. In contrast, the earliest artifacts of writing, and hence reading, are cave paintings in southern France and northern Spain, that are approximately 20,000 years old. The paintings depict animals and occasional humans, and it is not clear whether they tell a story or are just pictures that serve some religious or magical purposes. A major advance in writing occurred around 5,000 years

893

READING ago in Sumeria. When goods were sent to market, they were accompanied by clay balls. Sealed inside the clay balls were small pieces of clay that represented the goods that were being transported. When the shipment reached the market, the clay ball would be broken open and the shipment would be inspected to make sure nothing had been diverted. Eventually, it became apparent that marks could be pressed into wet clay to represent the goods, saving the trouble of forming the balls filled with small pieces. This, of course, marks the birth of a real writing system. Today, printed languages vary in terms of what they represent, with three major categories of writing. The Chinese writing system is called a logography. Each character represents a morpheme or minimal unit of meaning. Beijing means ‘north capital”, and is represented in writing by two symbols, one for north and one for capital. The Kana system of Japanese is a syllabary. Each symbol represents a spoken syllable. Finally, English, modern Korean, and many other written languages are alphabets, in which a small set of characters are put together to make a much larger number of words. Dividing the world’s languages into three categories is an oversimplification. Thus, some sounds as well as meaning are represented in Chinese writing, and English is best described as a morphophonemic representation system as English spellings represent both sounds (phonemes) and (units of) meaning (morphemes). Consequently, the printed word HEALTH has a spelling that makes apparent its relation to the morpheme ‘heal’ as opposed to a strictly phonetic spelling such as HELTH.

How Do Children Learn to Read? The obvious answer is that children are taught to read by teachers and parents. But a closer investigation of what is involved suggests it is immodest to assert that teachers or parents actually teach children to read.

Reading involves the coordinated orchestration of just about every process ever studied by cognitive psychologists and linguists. At best, teachers put print in front of children, and point out associations between print and sound or meaning, but the real work of learning to read is done by cognitive and linguistic machinery that is not readily accessible to teachers or learners. It is clear that learning to read an alphabetic system of writing is easier for children who have awareness and access to the sound structure of their oral language (i.e. phonological awareness) that is represented in writing. Thus, the spoken words ‘cat’, ‘rat’, and ‘sat’, have different initial sounds and identical medial and final sounds. To a child who is aware of these similarities and differences, the English system of writing will seem sensible in that these relations in sounds are represented in spelling. Cat, rat, and sat have different initial letters and identical medial and final letters. To a child lacking such an awareness, an alphabetic writing system will seem to be arbitrary and complex. References Adams, Marilyn. 1994. Beginning to read: thinking and learning about print. Cambridge, MA: MIT Press. Crowder, Robert, and Richard Wagner. 1991. The psychology of reading. New York: Oxford University Press. Perfetti, Charles. 1985. Reading ability. New York: Oxford University Press. Rayner, Keith, and Alexander Pollatsek. 1989. The psychology of reading. Englewood Cliffs, NJ: Prentice-Hall. Wagner, Richard, and Catherine McBride-Chang. 1996. The development of reading-related phonological processes. Annals of Child Development 12. 177–206. Wagner, Richard, and Joseph Torgesen. 1987. The nature of phonological processing and its causal relations with reading. Psychological Bulletin 101. 192–212.

RICHARD WAGNER See also Handwriting; Literacy; Reading Impairment; Writing Systems

Reading Impairment Reading impairment, reading disability, and dyslexia all refer to a level of reading that is below expectations. Expectations for reading can be based on normative data from age-matched peers, or based on an individual child’s oral language ability or cognitive ability.

894

Nature of Reading Impairment Dyslexia is perhaps the most commonly used term to describe reading impairment, and surely the most misunderstood. The common view of dyslexia is that of a

READING IMPAIRMENT visual perceptual problem that results in seeing mirror images of words or letters. Thus, individuals with dyslexia are reported to read WAS as SAW, or to confuse the letters ‘b’ and ‘d’. The origin of this view is easily established. Children with reading problems typically became obvious to teachers and parents at the second-grade level, and indeed they can be observed reading WAS for SAW and confusing the letters ‘b’ and ‘d’. But it turns out that normal beginning readers also make similar errors. The fact that words in English are to be read from left to right as opposed to right to left is arbitrary and must be learned. In addition, the letters ‘b’ and ‘d’ are both visually confusable and similar in sound (i.e. both stop consonants). Careful analysis of reading errors has shown that second-grade readers with dyslexia make no more reversal errors than do younger normal readers who are at the same level of reading. Thus, secondgrade teachers observed only poor readers making reversal errors, whereas kindergarten teachers would know that such errors are quite common. Another mistaken idea about reading impairment is that it results from erratic or inefficient eye movements. Reading requires highly sophisticated and coordinated eye movements characterized by ballistic movements called saccades and pauses called fixations. During the movements or saccades, little information is available to the eyes beyond a blur. Nearly all information is acquired during fixations. As you read the words on this page, your perception is that your eyes are moving smoothly across the page. This is an example of a situation where perception does not reflect reality. If you observe normal readers reading text by having them read directly across from you while holding a book low enough so you can observe their eyes, you will indeed see that the eyes move across the page in a series of small, but observable jerky movements. Perform the same experiment on individuals with reading impairment and it will be apparent that their eyes move much more erratically, even moving in the wrong direction at times. Observations like these resulted in the belief that faulty eye movements were the origin of reading impairments, and also resulted in interventions based on eye-movement training. It turns out that this view has it backwards. Indeed, the eye movements of individuals with reading impairments are more erratic than those of normal readers, but the erratic eye movements are the by-product, not the cause, of the impaired reading. The eye movements of individuals with reading impairments do not move across the page as smoothly as do those of normal readers because they are having trouble reading the words. This also explains their greater frequency of

backward eye movements or regressions. Conclusive evidence was provided by careful studies in which normal readers were given material that was as difficult for them to read as is grade-level reading material for individuals with reading impairment, and individuals with reading impairment were given very easy reading material that they could read as well as normal readers could read grade-level material. Under these conditions, the eye movements of normal readers deteriorated to match the previously reported erratic eye movements of individuals with reading impairment, and the eye movements of individuals with reading impairment now looked normal. Additional confirmation came from the results of eye-movement training studies. Although eye-movement training did result in gains performance on eye-movement tasks outside the context of reading, reading performance did not improve. For the vast majority of individuals with reading impairment, the problem appears to be based in language as opposed to the visual system, and is commonly compounded by ineffective instruction. Compared to reading-level matched controls, most individuals with reading impairment perform poorly on measures of phonological awareness and phonological decoding, and have fewer words that can be decoded by sight. Phonological awareness refers to an individual’s awareness and access to the sound structure of an oral language. Phonological decoding refers to decoding words by sounding them out, as when one is asked to decode nonwords such as TANE. The underlying language problem for individuals with reading impairment is likely to be a subtle and not well-understood problem in forming accurate phonological representations, which in turn leads to poor phonological awareness and phonological decoding. Once beginning readers fall behind, they are exposed to reading instruction designed for more advanced readers, which provides little assistance, until they are finally identified as having a reading problem and more appropriate instruction is provided.

Distribution of Reading Impairment Two important facts about the distribution of reading impairment are counterintuitive. First, despite the fact that boys outnumber girls by roughly four to one in classes and clinics that serve individuals with reading impairments, epidemiological studies in which all children are tested reveal that reading impairment rates for boys and girls are roughly comparable. For every boy with a reading impairment, there is likely to be a girl with an equally severe reading impairment. The overrepresentation of boys in classes and clinics appears to

895

READING IMPAIRMENT result from a referral bias that arises because boys are more likely to be disruptive than girls, and hence more likely to be referred for evaluation. Second, reading impairment is not something that one has or does not have. It is not an all or none phenomenon. The distribution of reading performance is continuous, with no obvious breaks or bumps in the tails of the distribution. Where one draws the line is arbitrary. References Blackman, Benita (ed.) 1997. Foundations of reading acquisition and dyslexia. Mahwah, NJ: Lawrence Erlbaum Associates. Crowder, Robert, and Richard Wagner. 1991. The psychology of reading. New York: Oxford University Press.

Foorman, Barbara (ed.) 2003. Preventing and remediating reading difficulties: bringing science to scale. New York: York Press. Shankweiler, Donald, and Isabelle Liberman. 1989. Phonology and reading disability. Ann Arbor, MI: University of Michigan Press. Spear-Swerling, Louise, and Robert Sternberg. 1996. Off track: when poor readers become learning disabled. Boulder, CO: Westview Press. Stanovich, Keith, and Linda Siegel. 1994. Phenotypic performance profile of children with reading disabilities: a regression-based test of the phonological-core variable-difference model. Journal of Educational Psychology 86. 24–53. Wagner, Richard, and Tamara Garon. 1999. Learning disabilities in perspective. Perspectives on learning disabilities, ed. by Robert, Sternberg and Louise Spear-Swerling. Boulder, CO: Westview Press.

RICHARD WAGNER See also Reading

Reference The term ‘reference’ is ambiguously used to designate (a) the relation between a referring expression and that to which it refers, (b) that to which a referring expression refers (also known as the ‘referent’), and (c) the act of referring to some extralinguistic entity or state of affairs, etc. by using a referring expression. Sometimes the terms ‘denotation’ and ‘designation’ are used interchangeably with reference in its sense (a)—‘denotatum’ and ‘designatum’ being the corresponding variants for the word in its sense (b). Sense (c) introduces, over and above the referring expression and the referent, a third element, namely the speaker, into the picture. This in turn introduces the possibility that a speaker may succeed in referring to an object, although the referring expression she or he uses is, strictly speaking, inaccurate as a description of it—as when someone says ‘That horse is agitated’, pointing at a certain restless quadruped that, on closer inspection, turns out to be a mule. A speaker may also successfully refer to an individual whose identity is otherwise unknown to everyone including himself or herself as when someone says ‘Smith’s murderer is insane’, meaning by the referring expression ‘Smith’s murderer’ something like ‘whosoever murdered Smith’ rather than, say, ‘that man in the dock over there, about to be cross-examined’. These cases show that speaker reference and denotation may not always coincide, raising the thorny issue as to which of them should be regarded as basic.

896

It is also important to distinguish among different kinds of referring expressions. ‘That man over there’, ‘Smith’, ‘The President of the United States’ are examples of singular definite expressions whereas ‘horse’, ‘gold’, etc. exemplify general expressions. Scholarly discussion about reference has tended to concentrate on singular definite expressions. Broadly speaking, theories of reference may be grouped under two conflicting approaches: descriptive vs. direct reference approaches. On the descriptive view, a referring expression does its job by describing features or properties of the referent. The descriptive view hinges crucially on a distinction between Sinn (sense) and Bedeutung (reference) proposed by the German logician and philosopher Gottlob Frege in order to explain the difference in cognitive value between identity statements of the form ‘x  y’ and ‘x  x’, where x and y stand for referring expressions. Frege argued that two referring expressions may have different senses but nonetheless have the same reference (i.e. referent), thus making the former equation more informative than the latter. He also insisted that a referring expression can only have a referent provided it has a sense, although having a sense was not a sure guarantee that the term will have a referent (as in the case of ‘unicorn’ or ‘square circle’). Bertrand Russell, who went on to spell out the description theory more articulately, took a narrower view of reference, maintaining that only ‘logically

REFERENCE proper names’ do the referring and they do it infallibly by naming objects known through ‘acquaintance’ that included sense-data, certain universals, and possibly ourselves. As for descriptions, he argued—contra Frege—that they are ‘incomplete symbols’, meaningful only in the sentential contexts in which they occur. Ordinary proper names like ‘Bucephalus’ and ‘Mount Everest’ are for Russell, as indeed they were for Frege too, disguised descriptions. So a descriptive phrase such as ‘The present King of France’ in the sentence ‘The present King of France is bald’ does not qualify as a genuinely referring expression, because a proper analysis of its logical form will show that what the sentence really says is that France currently has a (an only) king and that he is bald. Thus paraphrased, the singular definite description no longer has a referring function and its presence is explicated in terms of the existential quantifier and the variable within its scope as in (∃x) (Kx & ( (∀y) (Ky → y  x) ) & Bx) where ‘K’ and ‘B’ stand for the predicates ‘The present King of France’ and ‘Bald’, respectively. Russell’s ‘theory of definite descriptions’ came under attack from Peter F. Strawson, who took up a line of reasoning initiated by Frege and, departing from it in significant ways, insisted that anyone who utters the sentence in question in fact presupposes that there is presently a king of France, rather than asserting it, as Russell’s analysis implied. For Strawson, whereas referring expressions indeed have their senses, it is the actual uses of those expressions that do the referring. What specific horse is being referred to by the expression ‘that horse’ can only be resolved by examining an actual use (utterance) of the expression as part of, say, the making of a statement (and not merely the issuance of a sentence), along with the attendant set of actual circumstances. In other words, Strawson foregrounded the pragmatics of reference and claimed speaker reference to be paradigmatic, in opposition to Russell who was primarily interested in sense (a) or the semantics of the term ‘reference’. In contrast to the descriptive approach, direct reference approach harks back to a view taken by J.S. Mill in the nineteenth century. Unlike Frege and Russell, who argued that proper names had senses and that these coincided with the sense of corresponding definite descriptions, Mill had contended that they had no ‘connotations’ but only their ‘denotations’ (Mill’s terms for sense and reference, respectively). In other words, they simply referred, period, by functioning as a label attached, as it were, directly to the object. Just how this direct contact with the referent is secured is an open question. One suggested answer is that there is a causal link between the use of a term and an original ‘baptismal’ episode of attaching the name to its

bearer. Not all direct theorists, however, subscribe to the thesis of causality. Among modern defenders of the direct reference approach are Keith Donnellan, Saul Kripke, and Hilary Putnam. Highlighting the pragmatic aspect of reference, Donnellan argued that there are two distinct uses of definite descriptions, namely ‘attributive’ and ‘referential’ uses. In its referential use, a referring expression such as ‘Smith’s murderer’ can succeed in picking up the intended referent, namely, say, the man in the dock awaiting the sentence, even though it may eventually turn out that the man in question is innocent of the charges. In other words, in its referential use a referring expression can refer in spite of its sense, contrary to what the descriptivist view predicts. In Donnellan’s view, the Russell–Strawson debate was on a nonissue, since the former was thinking exclusively of the attributive use of definite descriptions while the latter was interested only in the referential use, but wrongly believing that the accuracy of the description was a necessary condition for the success of a referential use thereof. An interesting counterargument to Donnellan’s thesis was offered by John Searle, who claimed that the two uses, far from being mutually exclusive, can be explicated in terms of the number of ‘aspects’ under which a speaker is in a position to refer to the referent in question. Kripke held that proper names are ‘rigid designators’, meaning that they have the same referents in all possible worlds. So it is wrong to think that a proper name such as ‘George W. Bush’ has the same sense as ‘The President of the United States of America’ because, although the two may indeed refer to the same individual in the actual world, they may not do so in some other possible world (say, the one in which the Supreme Court decision had favored Al Gore instead). In the 1970s, Hilary Putnam developed the thesis of ‘semantic externalism’ whose central tenet is that, in many cases such as words referring to natural kinds like ‘gold’ and ‘water’, the reference of an expression cannot be thought of as a function of descriptions associated with it in the mind of the user. Thus, in Putnam’s view, ‘water’ refers to the chemical compound H2O in all possible worlds, irrespective of whether or not one is aware of that scientific truism. Direct reference theories run into difficulties when called upon to explain the behavior of referring expressions in what are known as intensional or referentially opaque contexts as in ‘Juan believes that the Malvinas rightfully belong to Argentina’, which has—on one interpretation—a different truth value than the sentence ‘Juan believes that the Falklands rightfully belong to Argentina’ although ‘Falklands’ and ‘Malvinas’ designate the same group of islands (a fact that, as it happens,

897

REFERENCE Juan is not aware of). The usual answer that whatever difference there may be between the two sentences is not of interest to semantics has not satisfied critics. Causal theorists are also hard put to it to explain changes in the reference of a term over the years (as in the case of ‘Yugoslavia’ whose referent has significantly changed after successive wars) as well as natural kind terms that are posited rather than effectively attested (as in the case of a missing link to account for the development of human beings from apes). References Donnellan, Keith. 1966. Reference and definite descriptions. Philosophical Review 75.

Frege, Gottlob. 1949. Über Sinn und Bedeutung (1892). Readings in philosophical analysis, English translation by Herbert Feigl, ed. by Herbert Feigl and Wilfrid Sellars. New York: Appleton-Century Crofts, Inc. Haack, Susan. 1978. Philosophy of logics. Cambridge: Cambridge University Press. Kripke, Saul. 1972. Naming and necessity. Semantics of natural language, ed. by Donald Davidson and Gilbert Harman. Dordrecht, Holland: D. Reidel Publishing Company. Lyons, John. 1977. Semantics, Vol. 1. Cambridge: Cambridge University Press. Russell, Betrand. 1905. On denoting. Mind 14. Searle, John. 1979. Referential and attributive. Expression and meaning. Cambridge, UK: Cambridge University Press. Strawson, Peter F. 1971. On referring. Logico-linguistic papers. London: Methuen & Co.

KANAVILLIL RAJAGOPALAN

Register Broadly defined, ‘register’ refers to the way people use language in different situations. The term is often used as a full or partial synonym for style, speech variety or variation, field of discourse, and text type. It is the subject of sociolinguistic, linguistic, and applied linguistic analysis. Most often, register is used to mean style. In that sense, it refers to the ‘stylistic variety’, or ‘stylistic variation’ that occurs in a person’s speech or writing in different social contexts. Typically, different registers vary according to their degree of formality. For example, expressions like ‘sure’, ‘no problem’, and ‘let’s have a blast’ are associated with informal conversational styles or registers. Utterances like ‘in spite of our deliberate efforts to condemn fraud…’, on the other hand, are used in formal written texts. In her Grammar book, D. Larsen-Freeman offers numerous examples of appropriate and inappropriate use of both formal and informal registers. In the case of phrasal verbs, ‘put off’, ‘call off’, and ‘show up’ are common in informal registers, whereas their one-word synonyms ‘postpone’, ‘cancel’, and ‘arrive’, reflect higher levels of formality. The same applies to relative adverbs. For example, ‘1950 is the year that I was born in’ would occur in more formal texts than ‘1950 is when I was born’. A connector, such as ‘moreover’, normally occurs in formal written texts. It would be inappropriate in an informal conversational setting like this, ‘Let’s go to the beach. *Moreover, let’s rent a boat’. It appears from these and other examples that there is an abundance of structures that reflect the subtle nuances of formality, both within a separate linguistic category and across different categories. 898

Linking register solely to levels of formality in speech and writing, however, would be an oversimplification. Along with levels of formality, scholars also study changes in register based on the speaker’s awareness of a broader network of differences in situation, topic, addressee(s), or location. John Lyons emphasizes the complex sociolinguistic nature of register. He discusses it in the framework of context, style, and culture, and more specifically as a feature of ‘communicative competence’. He relates it to the appropriate choice of language with regard to ‘domain’, i.e. a cluster of social situations constrained by a common set of behavioral rules. Within domain, Crystal and Davy identify ‘province’, i.e. occupational or professional activity with no reference to the individuals engaged in it. For example, the language used in advertising, science, law, and sports can be identified as being typical of the corresponding province, hence the use of ‘love’ in tennis and ‘sentence’ in law. Fishman relates domain to ‘subject-matter’ on the one hand, and to ‘locale’ and ‘role-relations’ on the other. Examples of these would be ‘family—home’ and ‘religion—church’. Within these subject-locales, he identifies certain role-relations, such as ‘mother-toson’ and ‘priest-to-parishioner’. Thus, a mother talks to her son at a different level of formality if the conversation is held at home or not. Similarly, her choice of language will be different if the topic was ‘her son’s drunk driving’ and ‘her son’s recent award’. She would also choose a different register if she were to discuss these subjects with other people, at other places. Similar to Crystal and Davy’s use of ‘province’, Quirk defines

REICHENBACH, HANS register in relation to the ‘field of discourse’, or the activity in which a speaker is engaged. He categorizes speech varieties as dependent and interdependent according to ‘field of discourse’. Speakers who retain regional features in their use of Standard English are examples of the former. Interdependent speech varieties, on the other hand, in a slightly broader use of Fishman’s ‘locale’, are exemplified by topics that are associated with specific regions. Thus, discussions of baseball are in American English. Similarly, coaching is handled by speaking rather than by writing. Linguistically, register could also be examined on the phonetic, morphological, syntactic, and semantic levels. Yet, it can be most fully analyzed on the ‘text’ level, e.g. it is suprasentential or above and beyond the sentence. In his An introduction to functional grammar, M.A.K. Halliday discusses register in relation to ‘texture’, or ‘text structure’, which comprises theme and focus, lexical cohesion, reference, ellipsis and substitution, and conjunction, all of which have a semantic, not grammatical structure. He points out some of the differences in ‘theme and rheme’ in different text types, such as narratives and instructions. In the following example of a narrative, the participant (‘he’) remains the topical theme: ‘He went to many colleges but he didn’t complete any. He always fell into some difficulty. He never hit upon the right thing’. In instructions that have stepwide structure, the theme, or given subject of one clause, becomes the rheme, or new object of the clause that follows it, e.g. ‘Turn the stove on. Put a pot on the stove. Pour water into it. Add all ingredients to the water’. Thus, it would be inappropriate to use this kind of stepwide-structured register when writing an expository text, e.g. ‘I have twin sisters. My parents love them. They buy them many toys. One such toy is a huge stuffed cat. It now occupies the couch’. Halliday also defines register by different types of ‘lexical cohesion and reference’. He notes that they are built through interlocking referential chain complexes that vary from one register to another and produce a certain global effect, e.g. ‘A boy called John. . . John . . . he . . . the lad . . . him . . .’ The global effect created by such overlapping referential chains is the source of the dynamic flow of discourse in narratives. According

to him, ellipsis and substitution, which are mostly found in dialogue, have a more local effect because of their shorter lexical reach. Typically, their reach extends over no more than a few consecutive moves, ‘A: Would you like to come here? B: Could I bring a friend too? A: Sure. Which one? B: I’ll know tomorrow. A: Is it still raining outside?’ Such texts are related not so much by ideational as by interpersonal meaning. Halliday also defines the features of oral and written texts or registers. Generally characterized by lexical simplicity, oral texts become complex by being grammatically intricate. They use clauses to express relationships and to achieve dynamic complexity. On the other hand, written and technical texts, in particular, are lexically dense and statically complex because they use nominalization, or a lot of noun phrases, to express relationships. While they have simple clausal patterns, they pack ideas in nominal constructions, i.e. ‘advances in technology’ instead of ‘technology is getting better’. Thus, nominalization, which helps to construct technical terms and develop step-by-step arguments, is often considered the most prominent feature of texts or registers of expert knowledge, prestige, and power. References Cheshire, Jenny. 1992. Register and style. International encyclopedia of linguistics, ed. by W. Bright. New York: Oxford University Press. Crystal, David, and D. Davy. 1968. Investigating English style. London: Longman. Downes, W. 1994. Register in literature. The encyclopedia of language and linguistics, ed. by R.E. Asher and J.M.Y. Simpson. Oxford: Pergamon Press. Fishman, J.A. 1972. The sociology of language. Rowley, MA: Newbury House. Halliday, M.A.K. 1994. Functional grammar. London, Melbourne and Auckland: Edward Arnold. Larsen-Freeman, Diane, and Marianne Celce-Murcia. 1999. The grammar book, 2nd edition. New York: Heinle & Heinle. Lyons, John. 1977. Semantics. London, New York and Melbourne: Cambridge University Press. Quirk, Randolph, et al. 1985. A comprehensive grammar of the English language. London and New York: Longman. Richards, Jack, et al. 1992. Longman dictionary of language teaching and applied linguistics. Harlow, Essex: Longman.

LILIA SAVOVA

Reichenbach, Hans The philosopher Hans Reichenbach was a central figure of logical empirism in the 1920s and 1930s. The starting point of his research were the philosophical dimensions

and implications of relativity theory (Relativitätstheorie und Erkenntnis a priori, 1920; relativity theory and cognition a priori), but also more general issues in the 899

REICHENBACH, HANS philosophy of space and time (Axiomatik der relativistischen Raum-Zeit-Lehre (1924; Axioms of relativistic space–time), Philosophie der Raum-Zeit-Lehre (1928; Philosophy of space– time)). In 1930, he founded the journal Erkenntnis (Cognition, later called Journal of Unified Science), together with Rudolf Carnap from the Vienna Circle. Furthermore, he developed a threedimensional logic for the interpretation of quantum mechanics (Philosophic foundations of quantum mechanics, 1944) and worked on probability. But linguists usually do not appreciate Reichenbach as a philosopher. In linguistics, he became famous for not more than 12 pages of his complete works, namely for §51 ‘The tenses of verbs’ of his book Elements of symbolic logic, 1947. According to Reichenbach, the tenses of verbs display token reflexivity. The tenses determine time with reference to the time point of the act of speech, i.e. of the token uttered. The time point of the token is called the point of speech (S), and the time point at which the event being talked about took place is called the point of the event (E). Using these two points means assuming three tenses, namely E  S (the simple past in English), E  S (the simple present), and E  S (the simple future). But obviously, there are more than these three tenses in English (and in other languages as well). This is the reason why Reichenbach introduced a third parameter, namely the point of reference (R). What this parameter does becomes most easily clear in narratives. Compare the following passage from W. Somerset Maugham’s Of human bondage: But Philip ceased to think of her a moment after he had settled down in his carriage. He thought only of the future. He had written to Mrs. Otter, the massière to whom Hayward had given him an introduction, and had in his pocket an invitation to tea on the following day.

The events related in the simple past perfect (had settled down..., had written to Mrs. Otter..., had given him an introduction...) are viewed as being completed from the perspective of a specific point of reference, which is constituted by those events that are related in the simple past (ceased to think of her..., thought only of the future..., had in his pocket an invitation...). Thus, the point of reference is a kind of perspective point. With some tenses, two of the three parameters E, R, and S coincide. The simple past is such a case, with E and R being simultaneous, and both E and R being before S. With other tenses, the three parameters denote distinct times, e.g. in the case of the simple past perfect: E and R are both before S, but E is also before R. The different usages of the simple past and the simple past perfect in the quotation above show this quite clearly.

900

As for the simple tenses, Reichenbach proposes the following tables, in which the direction of time is represented as the direction of the line from left to right: Past Perfect I had seen John –X–X–X–> ERS Present I see John –X–X–X–> S,R,E

Simple Past I saw John –X–X–X–> R,E S

Simple Future I will see John –X–X–X–> S,R E

Present Perfect I have seen John –X–X–X–> E S,R Future Perfect I shall have seen John –X–X–X–> SER

If there is a progressive form, Reichenbach assumes that the event E covers a certain stretch of time. He calls these extended tenses. Here are two examples: Past Perfect, Extended I had been seeing John –XXXX–X––X–> ERS

Simple Past, Extended I was seeing John –XXXX––X–> R,E S

As for adverbial modification, Reichenbach states that in the presence of positional adverbs like now or yesterday, only modification of R is possible. This can be illustrated with sentences like I had met him yesterday, where yesterday is the point of reference, but the meeting (E) has occurred at some time before yesterday. Here, R is the carrier of the time position. Therefore, Reichenbach speaks of the positional use of the reference point here. On the whole, there are 13 possibilities of ordering the three time points E, R, and S. But for English, there are only six tenses in the traditional grammars. The following table summarizes the possibilities, Reichenbach’s terms for them, and the traditional terms: Possibility E_R_S E,R_S R_E_S R_S,E R_S_E E_S,R S,R,E S,R_E S_E_R S,E_R E_S_R S_R,E S_R_E

Reichenbach’s Term Anterior past Simple past Posterior past Posterior past Posterior past Anterior present Simple present Posterior present Anterior future Anterior future Anterior future Simple future Posterior future

Traditional Term Past perfect Simple past — — — Present perfect Present Simple future Future perfect Future perfect Future perfect Simple future —

RELEVANCE IN DISCOURSE With Reichenbach’s terms, the position of R relative to S is indicated by past, present, future, and the position of E relative to R is indicated by anterior, simple, posterior. As for the posterior future, it is used for instance in I shall be going to see him. Sentences like I did not expect that he would win the race express the posterior past—like posterior future this is commonly not classified as a tense. Reichenbach has been criticized for assuming the point of reference because this often coincides with the point of the event or with the point of speech. Nevertheless, his work on tense is considered classic. Within the modern theory of tense, there are many reconstructions and adaptations of Reichenbach, e.g. Klein (1994), Ehrich (1992), Musan (2000), Comrie (1985), Thieroff (1992), and Fabricius-Hansen (1986), to mention just a few.

Biography Hans Reichenbach was born in Hamburg, Germany, on September 26, 1891. He studied civil engineering, mathematics, physics, and philosophy at Technische Hochschule Stuttgart and Universities of Berlin, Göttingen, and Munich; Ph.D. (1915, Erlangen) for mathematical work about probability; and Habilitation (1920, Technische Hochschule Stuttgart) for philosophical work about relativity theory. He was assistant Professor, Technische Hochschule Stuttgart, 1920–1926; and Professor for Philosophy of Nature and Physics, University of Berlin, 1926–1933; he lost his teaching

appointment when the Nazis seized power and emigrated to Turkey; he was Professor for Philosophy, University of Istanbul, 1933-1938. He moved to the United States, and was Professor for Theory of Science, University of California at Los Angeles, 1938–1953. He died in Los Angeles on April 9, 1953. References Comrie, B. 1985. Tense. Cambridge: Cambridge University Press. Ehrich, V. 1992. Hier und jetzt. Studien zur temporalen und lokalen Deixis im Deutschen. Tübingen: Niemeyer. Fabricius-Hansen, C. 1986. Tempus fugit: Über die Interpretation temporaler Strukturen im Deutschen. Düsseldorf: Schwann. Klein, W. 1994. Time in language. London and New York: Routledge. Musan, R. 2000. The semantics of perfect constructions and temporal adverbials in German. Habilitationsschrift, University of Berlin. Reichenbach, Hans. 1920. Relativitätstheorie und Erkenntnis a priori. ———. 1924. Axiomatik der relativistischen Raum-ZeitLehre. Braunschweig: Vieweg. ———. 1928. Philosophie der Raum-Zeit-Lehre. Braunschweig: Vieweg. ———. 1935. Wahrscheinlichkeitslehre. Leyden: A.W. Stthoff. ———. 1944. Philosophic foundations of quantum mechanics. Berkeley: University of California Press. ———. 1947. Elements of symbolic logic. New York: Macmillan. ———. 1951. The rise of scientific philosophy. Berkeley: University of California Press. Thieroff, R. 1992. Das finite Verb im Deutschen. Modus— Tempus—Distanz. Tübingen: Narr.

MONIKA RATHERT

Relevance in Discourse Since discourse analysis is interested in stretches of text longer than a sentence or an utterance, it has to address the question of coherence—how the different parts of texts hang together. One crucial aspect of coherence will be relevance, how the meanings conveyed by one sentence can be related to or interact with information already in the reader’s/hearer’s mind, especially information already given in the preceding text. The first philosopher to emphasize the importance of relevance to discoursal pragmatics was Grice (1975 [1967], 1989). His cooperative principle proposes four maxims that govern conversation in particular, namely quantity (give as much information as is required, no

more), quality (only say things that are true for which you have evidence), manner (be clear, orderly, concise, and avoid obscure language), and relation (be relevant). Sperber and Wilson (1995) and Wilson and Sperber (1987) developed Gricean theory by subsuming all these maxims under relevance. For them, communicative utterances carry the presumption of optimal relevance. Utterances are relevant if the information they convey interacts with the hearer’s existing assumptions. These interactions, called contextual effects, can be either the strengthening of an existing assumption, the weakening/cancelation of an existing assumption, or the production of a contextual implication. In the

901

RELEVANCE IN DISCOURSE case of contextual implication, the information given by an utterance interacts with an existing assumption by deductive logic, to create a new assumption. One of Sperber and Wilson’s examples is as follows: A: Would you drive a Rolls Royce? B: I wouldn’t drive any expensive car. B does not obviously answer A’s question. However, if A already knows that a Rolls Royce is an expensive car, this assumption interacts with B’s reply to produce the contextual implication ‘B would not drive a Rolls Royce’. So, B’s answer is still relevant to A’s question. The assumption supplied by the hearer ‘a Rolls Royce is an expensive car’ is called an implicated premise. That different contextual implications arise from different implicated premises is particularly important to the intertextuality of discourse. Intertextuality, put simply, means the way in which one text impinges on others (Kristeva 1974:59–60). One aspect, viewed from a relevance perspective, is that the information derived from one text can provide implicated premises and effect the contextual implications during the processing of a later text. For example, take the following text: It is difficult to see how any organism could remain healthy if feedback messages to the decision-making center were systematically stifled for long periods.

It will be processed and interpreted quite differently, more metaphorically, if the preceding intertext is a) a report showing that many politicians are devoting less and less time to their constituency surgeries/feedback sessions with the electorate rather than b) a feature article on the advances in the effectiveness of the pain-killing sprays used by football-players/athletes. The relevance of utterances can be computed as a fraction: contextual effects processing effort. All other things being equal, the greater the contextual effects the greater the relevance, and the greater the processing effort the less the relevance. We can see that contextual effects correspond to Gricean maxims of quality and relation; and processing effort corresponds to Grice’s maxim of manner (and perhaps quantity). According to this formula, B’s reply might not be seen as optimally relevant. B could have reduced the processing effort by simply replying ‘No’. However, B’s actual answer is likely to produce far more contextual effects than the answer ‘No’. For example, if A knows that a Cadillac and a Lexus are also expensive cars, then B’s actual reply will generate the contextual implications ‘B would not drive a Cadillac or a Lexus’. Like Gricean pragmatics, relevance theory opposes a code-based theory of communication. A successfully decoded message may be intelligible, as B’s answer

902

above is to A, but to function as relevant in discourse it has to be interpretable, which may involve implication or inferencing. So Grice and Sperber/Wilson prefer an inferential account of communication. In this, coding and decoding using conventional signs are never sufficient for communication, and sometimes even unnecessary. If you ask whether I enjoyed my skiing trip, and I raise my broken leg covered in plaster, this might communicate the answer ‘no’. But such an action has no conventional coded meaning. In inferential theories, utterances give evidence for a hypothesis about an intended meaning, often an implied one, as in the examples above. However, relevance theory goes further than Grice, in maintaining that pragmatic inferential processes are important in making explicit the full propositional content of the utterance, even before implications are made. For example, imagine John has cooked dinner, Mary is sitting in the dining room reading the newspaper, and John comes in and puts the two plates of food on the table. He then says It’ll get cold.

In order to make the prepositional content complete, we have to infer what ‘it’ refers to (for example, Mary’s dinner rather than the newspaper), disambiguate ‘cold’ (to mean ‘low in temperature’ rather than ‘experiencing low temperatures’), and narrow down the time reference of ‘ll’ (to mean ‘will soon’). This process, an example of what is termed explicature, will give us the relatively complete proposition Mary’s dinner will soon become low in temperature

This is now intelligible as a complete message. But in order for it to be interpretable, it would have to generate a relevant contextual implication. If it interacts with Mary’s existing assumption John wants Mary to eat her dinner while it is still hot.

the resulting contextual implication would be John wants Mary to come and eat her dinner very soon

(cf. Sperber and Wilson 1995:176–93). Relevance theory was seen by Wilson and Sperber as a complement to Chomskyan linguistic theory, which deals with semantics but fails to account for the pragmatic meanings generated in discourse. Chomsky is famously uninterested in sociolinguistics. But, as early critics of relevance theory pointed out (Clark 1987), without some theory of social purposes the notion of relevance is rather vacuous. Goatly (1994, 1997) argues that in order to develop as a tool for discourse analysis, relevance theory needs developed notions of social contexts such as the Hallidayan

RHETORIC AND LINGUISTICS concepts of register and genre. However, Blass (1990) has argued that relevance theory gives a more satisfactory account of cohesion and coherence than the model of Halliday and Hasan (Cohesion in English 1976). Relevance theory has been interestingly applied to different discourse genres, including work by Tanaka (1994) and Forceville (1994) analyzing advertisements, Campbell (1992) and Moeschler (1989) on argumentation, Wilson (1990) on political language, and Mayher (1990) for language in education. References Blakemore, Diane. 1992. Understanding utterances: an introduction to pragmatics. Oxford: Blackwell. Blass, Regina. 1990. Relevance relations in discourse: a study with special reference to Sissala. Cambridge: Cambridge University Press. Campbell, J. 1992. An applied relevance theory of the making and understanding of rhetorical arguments. Language and Communication 12(2). Clark, H. 1987. Relevance to what? Behavioural and Brain Sciences 10(4). Forceville, Charles. 1994. Pictorial metaphor in advertising. London: Routledge.

Goatly, Andrew. 1994. Register and the redemption of relevance Theory: the case of metaphor. Pragmatics 4(2). ———. 1997. The language of metaphors. London: Routledge. Grice, Paul. 1975 [1967]. Logic and conversation. Syntax and semantics, Vol. 3: speech acts, ed. by Peter Cole and Jerry Morgan. New York: Academic Press. ———. 1989. Studies in the way of words. Cambridge, MA: Harvard University Press. Halliday, M.A.K., and R. Hasan. 1976. Cohesion in English. London: Longman. Kristeva, Julia. 1974. La revolution de langage poetique. Paris: Seuil. Mayher, J. 1990. Uncommon sense: theoretical practice in language education. London: Heinemann. Moeschler, J. 1989. Pragmatic connectives, argumentative coherence and relevance. Argumentation 3.3. Dordrecht: Kluwer. Sperber, Dan, and Deirdre Wilson. 1995. Relevance: communication and cognition, 2nd edition. Oxford: Blackwell. Tanaka, Keiko. 1994. Advertising language: a pragmatic approach to advertisements in Britain and Japan. London: Routledge. Wilson, Deirdre, and Dan Sperber. 1987. An outline of relevance theory. Notes on Linguistics 39. Wilson, J. 1990. Politically speaking: the pragmatic analysis of political language. London: Blackwell.

ANDREW GOATLY

Rhetoric and Linguistics No discussion regarding the interrelationships between these two disciplines can proceed without a thorough understanding of what their content encompasses, the key to the alliance. The roots of rhetoric date back to the work of Aristotle (384–322 BCE) and the fifth-century sophists (460–380 BCE) such as Protagoras and Gorgias, who in their dialogues about the relationship of truth and reality, and the nature and purpose of spoken discourse consistently demonstrated that, inevitably, for these scholars there was a ‘philosophical and pedagogical concern with language’ (Gillam 1998:15). Like linguistics, rhetoric has exhibited organic changes in focus, orientation, and scope over time, evident in the meaning of the word rhetoric, which has acquired meanings that are ‘irreducibly multivalent’ (Fleming 1998:174). Often associated in modern lay thought with the political rhetoric of ‘deceit’, rhetoric in academia has moved from being a ‘pejorative to an honorific term’ (Fleming 1998:169), a return to the classical art of persuasion. Modern uses of the term have broadened the scope of rhetoric to include ‘virtually all humanly created symbols from which audi-

ences derive meanings’ (Foss et al. 1999:6), an influence triggered by the modern rhetorician Kenneth Burke, who envisioned rhetoric to be ‘symbolic inducement’ (Burke 1966:296) for the purposes of either cooperation or competition. The site of location has shifted from public spheres of debate such as law, politics, and religion to more private constructions of meaning, particularly the manner in which individuals make sense of their world, a shift that has resulted in the synonymous connotation that rhetoric has acquired in modern times with the term ‘composition’, the private creation of meaning via the most common symbol of persuasion, writing, in modern universities. In this entry, these terms are sometimes used synonymously. Although the focus in classical accounts is with the construction of truth, modern conceptualizations see rhetoric as providing an arena to question the privileging of some truths over others, in short, ‘a tool of critique’ to demonstrate that ‘reality, belief and language are not lined up as unproblematically as people would like’ (Fleming 1998:170), a shift prompting an additional return to antiquarian roots espousing the sophist tradition of Isocrates which stresses civic development

903

RHETORIC AND LINGUISTICS or Arete (virtue) on the part of the individual. In such a conceptualization, ‘rhetoric is a study that in addition to imparting an art and guiding good practice, encourages critical and substantive reflection about the situated relations of discourse to reason, character, and community in human action’ (Fleming 1998:184), a definition stressing the link with linguistics, particularly sociolinguistics. The link between the two fields is often conceptualized as the applications of linguistic theory to rhetoric, an alliance that can be broadly divided into two periods: optimistic applications of linguistic theory to rhetoric from the 1950s to the mid-1980s— a period of ‘a good deal of excitement’ (Smitherman 1999:351)— to pessimistic speculations as to the actual applicability of linguistic theory in the early 1990s to the present. It is important to stress that even though the literature seems to insist that the influence was unidirectional, linguistic research has also been influenced by studies in rhetoric. This might be attributed to the parallel development of composition-rhetoric and linguistics in the 1950s and 1960s as ‘both fields sought to reinvent themselves and stake intellectual claim to distinct identities among the established disciplines of the academy’ (Smitherman 1999:351). Consequently, the antiquarian interest in audience-influenced discourse styles saw linguistic applications in the work of Martin Joos (1961), who outlined five major audienceinfluenced styles of discourse. Since language has and continues to be an integral part of rhetoric, especially in approaches labeled epistemic rhetoric, whose primary focus is on the private sphere of semantic generation, particularly the extent to which language simultaneously embodies and generates knowledge, linguists have constantly influenced and been influenced by the field. For instance, Kenneth Pike (1970) developed his theory of tagmemics in response to the argument that since the basis of language was the embodiment of knowledge, such a process could only proceed via access to reoccurring, repeatable units, the kind that language already offers. The dual relationship between these disciplines emerges in the unique conceptualization of Rogerian rhetoric, a persuasive technique based on conciliatory rather than adversarial argumentation. Similar links can be made with the work of Kenneth Burke (1966) whose model of language as action, ‘an equipment for living’ and ‘a strategy for coping’ (Warnock 1998:11) had linguistic parallels in the performative model of J.L. Austin’s speech act theory. The influence in the 1950s through the 1970s came from theoretical linguistics, particularly in the early phases, structuralism and, later, generative grammar, as well as applied linguistics, particularly the work of sociolinguists. One of the most overt influences in

904

rhetoric-composition was in the form of the 1974 National Council of Teachers of English (NCTE) statement on ‘Student’s rights to their own language’, a document first drafted by the Conference on College Composition and Communication (CCCC). The 1970s split in linguistics between Cartesian/theoretical linguistics and applied linguistics prompted a shift in influence, with most of the influence coming from applied linguistics. Some of the most commonly cited linguists at this time were Charles C. Fries, Kenneth Pike, Paul Roberts, and Noam Chomsky. These linguistic influences prompted a shift from prescriptive accounts of written discourse to more descriptive orientations to student writing, as well as an overt recognition that the cultural, linguistic, and rhetorical backgrounds of writers shaped their discursive practices. Other areas of interest in this period came from studies in contrastive rhetoric, whose propounder, Robert Kaplan (1966), was heavily influenced by the work in contrastive analysis in applied linguistics. The aim of such early as well as later theorists such as Ulla Connor was to demonstrate first language sociocognitive influences of speakers engaged in second language rhetorical practices, work resonating the cultural determinism inherent in the linguistic theories of Benjamin Lee Whorf. Recognizing the sociocultural basis of writing, linguists such as Muriel Saville-Troike and Donna M. Johnson have synthesized ethnography of communication approaches with such conceptualizations and devised linguistically oriented rhetorical approaches whose basic premise is ‘text as praxis’ (Cai 1998:123). Applications of contrastive rhetoric in first language writing are manifest in the works of Mina Shaughnessy and David Bartholomae whose research stresses that student errors in writing may be reflections of systematic sociocultural patterns rather than random performance slips, an approach heavily adopted by theorists in basic writing paradigms such as that developed by Patricia Bizell and Bruce Herzberg who urge for the teaching of effective rhetorical strategies that permit both inter- and intracultural communication. Research in the 1980s saw a deep alliance between the fields and ‘scholarship influenced by Pragmatics, Discourse Analysis and Functionalist Linguistics appeared with some frequency in composition journals and monographs’ (Kimball 2000:180). Shirley Brice Heath’s (1983) Ways with words, an ethnography of speaking approach, influenced rhetoricians interested in issues of literacy and written composition then and even today. Other connections to linguistics came in the alliance with stylistics. This marriage of disciplines saw a culmination in the publication of two books, one by Raskin and Weiser (1987) of Language and writing: applications of linguistics to rhetoric and

RHETORIC AND LINGUISTICS composition, a book demonstrating pragmatic applications of linguistics in the teaching of rhetoric, and the other by Beale (1987) of A pragmatic theory of rhetoric, a rhetorical model based on J.L. Austin’s Speech Act Theory (1962) since ‘classical concepts of rhetoric square nicely with speech act theory, making them mutually illuminating’ (Winterowd 1987:282). One can say that this was the last happy alliance. By the early 1990s, institutional changes that separated English and Linguistics departments at a number of universities triggered pessimistic feelings on the part of rhetoricians concerning the utility of linguistic analysis, a sentiment often voiced in articulations that the scientific method of linguistics, particularly ‘the atomistic nature of most linguistic analysis’ (Roy 1991:580), was in contradiction to the humanistic method of rhetoric, a difficulty accentuated even further by the metalanguage adopted by linguistics that prompted Alice Roy (1991), in a classic review of four studies on the alliance, to describe the link as ‘problematic’ (580). Recent rhetoricians have commented that this pessimism is unwarranted, a consequence of a misapplication of linguistic theory, a view based on ‘limited and inaccurate views of linguistics’ (Kimball 2000:181). Utilizing research from Gary Olson and Lester Faigley, who review Chomskian applications to rhetoric concluding that ‘It’s very rare that you ever get a free ride from some other field’ (34), prompts Kimball (2000) to call for a renewed interest in the applicability of linguistic theories to rhetoric today. Obvious examples of such renewed interest come in discourse theories in rhetoric propounded by Kinneavy et al. (1990), as well as social constructivist research in the field by theorists such as Susan Jarratt and Lynn Worsham, Lester Faigley, and James Berlin, all of whom view language to be a complex, pluralistic signifying practice that simultaneously constructs realities as it presents it—a rhetorical paradigm rooted in Roland Barthes’s (1968) theory of social semiotics, an approach emphasizing the social and ideological underpinnings of symbol use. These conceptualizations show heavy influences from linguistic research in critical discourse analysis, particularly social semiotics. Renewed interest has occurred in the area of generative rhetoric, a paradigm originally devised by Christensen and Christensen (1976), who, influenced by generative linguistics, argued that texture in discourse proceeded via syntactic elaboration, particularly via modification and addition. There has been a recent pedagogical push for more overt sentence combining instruction in writing. An exponent, Kimball (2000) argues for this approach to be utilized in the revision stage of writing. Other current applications of linguistics are in genre theory, a theoretical paradigm strongly influenced by the works of M.A.K. Halliday’s

systemic functional linguistics. Researchers such as Susan Jarratt and Lynn Worsham, Lester Faigley, and James Berlin push for a view of written rhetoric as situated social action, arguing that the only means to empower neophyte writers from being silenced is to educate them about the genres of powered discourse. Recent reapplications of early psycholinguistic and sociolinguistic models have manifested themselves in whole language teaching approaches to rhetoric, which emphasize linguistic socialization within and not apart from reading, writing, listening, and speaking. Other perspectives in genre theory have influenced rhetorical theories such as the literacy-orality theories of Walter Ong, an extension of Whorfian linguistics situating the source of writer difference in oral vs. analytical cultures, and/or individuals, an outcome of which may be a difference in cognitive complexity. Although this approach has been heavily critiqued in rhetorical and linguistic circles, applications of genre differences in spoken and written discourse have been recently cited by Kimball (2000) as an arena for research. Using the linguistic analyses of linguists such as M.A.K. Halliday, Wallace Chafe and Danielwicz, and Douglas Biber, she argues that some of the difficulties that novice rhetoricians face in this era may be an inability to distinguish between the features of this genre, a knowledge of which is presumed by the academy, and oral discourse. Returning to the unity of the disciplines in classical antiquity has caused some researchers in linguistics to recreate an alliance that benefits both pragmatics and rhetoric. With the renewed interest in applied linguistic theories, the alliance between rhetoric and linguistics looks promising, a possibility accentuated by the fact that more and more students and faculty in composition-rhetoric and linguistics departments have some exposure to each other. What makes the synthesis even more viable is that the area of interest in both rhetoric and linguistics from antiquity to the present has been questions of language. In both fields therefore, we have to continue to see the product of language as a ‘subject of study rather than an object of intervention’ (Kimball 2000:188). With this said, one can only predict more of an alliance in this century.

References Austin, J.L. 1962. How to do things with words. Oxford: Oxford University Press. Barthes, Roland. 1968. Elements of semiotics, transl. by Anetto Lavers and Colin Smith. New York: Hill and Wang. Beale, Walter H. 1987. A pragmatic theory of rhetoric. Carbondale: Southern Illinois University Press. Burke, Kenneth. 1966. Rhetoric and poetics. Language as symbolic action: essays on life, literature and method, 295–307. Berkeley: University of California Press.

905

RHETORIC AND LINGUISTICS Cai, Guanjun. 1998. Contrastive rhetoric. Theorizing composition: a critical sourcebook of theory and scholarship in contemporary composition studies, ed. by Mary Lynch Kennedy, 52–6. Westport, CT: Greenwood Press. Christensen, Francis, and Bonnijean Christensen. 1976. A new rhetoric. New York: Harper Press. Fleming, David. 1998. Rhetoric as a course of study. College English 61(2). 169–91. Foss, Karen, A., Sonja K. Foss, and Cindy L. Griffin. 1999. Feminist rhetorical theories. Thousand Oaks, CA: Sage Publications. Gillam, Alice. 1998. Classical rhetoric. Theorizing composition: a critical sourcebook of theory and scholarship in contemporary composition studies, ed. by Mary Lynch Kennedy, 15–26. Westport, CT: Greenwood Press. Heath, Shirley Brice. 1982. Protean shapes in literacy events: ever-shifting oral and literate traditions. Spoken and written language, ed. by Deborah Tannen, 91–117. Norwood, NJ: Ablex Publishing Company. ———. 1983. Ways with words: language, life, and work in communities and classrooms. New York: Cambridge University Press. Joos, Martin. 1961. The five clocks. New York: Harcourt, Brace and World. Kaplan, Robert B. 1966. Cultural thought patterns in intercultural education. Language Learning: A Journal of Applied Linguistics 17. 1–20.

Kimball, Sara. 2000. Linguistics and composition. Reforming college composition: writing the wrongs, ed. by Ray Wallace, Alan Jackson and Susan Lewis Wallace, 179–90. Westport, CT: Greenwood Press. Kinneavy, James L., William J. McCleary, and Neil Nakadate. 1990. Writing in the liberal arts tradition, 2nd edition. New York: Harper. Pike, Kenneth L. 1970. Tagmemics and matrix linguistics applied to selected African languages. Norman: Summer Institute of Linguistics of the University of Oklahoma. Raskin, Victor, and Irwin Weiser. 1987. Language and writing: applications of linguistics to rhetoric and composition. Norwood, NJ: Ablex Publishing Corporation. Roy, Alice. 1991. Four studies of linguistics and composition. College English 53(5). 580–6. Smitherman, Geneva. 1999. CCCC’s role in the struggle for language rights. College Composition and Communication 50(3). 349–75. Warnock, Tilly. 1998. Burkean theories of rhetoric. Theorizing composition: a critical sourcebook of theory and scholarship in contemporary composition studies, ed. by Mary Lynch Kennedy, 9–13. Westport, CT: Greenwood Press. Winterowd, Ross W. 1987. Literacy, linguistics and rhetoric. Teaching composition: twelve bibliographic essays, ed. by Gary Tate, 265–90. Fort Worth: Texas Christian University Press.

ANJALI PANDEY

Romanian The Romanian language (Limba româna ˘) is a Romance language that has been continuously spoken in the eastern part of the Roman Empire, including the Romanized provinces of Dacia, Southern Pannonia, Dardania, and Moesia, from the moment Latin was brought there until the present. The Romanian language has four main dialects: Daco-Romanian (or Romanian proper), Macedo-Romanian or Aromanian, Megleno-Romanian, and Istro-Romanian. The Aromanian dialect is spoken in Albania, Macedonia, Greece, and Bulgaria by scattered minorities, and the main area where MeglenoRomanian is spoken is in Greece. The Istro-Romanian dialect was spoken on the Istrian Peninsula of Croatia. The Daco-Romanian dialect (‘Romanian’) is spoken by more than 25 million people in Romania, Serbia, Croatia, Slovenia, Bosnia and Herzegovina, Bulgaria, Hungary, and Moldova, as well as in Romanian communities in the USA, Australia, Italy, Germany, etc. The variety of Romanian spoken in the Republic of Moldova (east of the Prut river) is officially called ‘the Moldavian language’. The earliest known written document in Romanian is a letter dating from 1521. The present spelling system was settled around 1860–1880. The basis of the standard

906

Romanian language is a combination of the two main geographical varieties, the Moldavian (west of the Prut river) and the Wallachian (Muntenian), the latter playing the more important role in setting the standard. Given the language’s relative isolation from the other Romance languages and its close contact with Hungarian and the Slavic languages surrounding Romania, the development of Romanian followed a different path from those of most other Romance languages. One of the most important foreign influences came from the Slavic languages. As a liturgical language, Old Slav(on)ic (or Old Bulgarian) provided a relatively large number of words to Romanian. This situation, combined with the fact that until the nineteenth century the Cyrillic alphabet was used to write Romanian and with the geopolitical position of Romania, has misled many people into thinking that Romanian is a Slavic language and that Romania is a successor state of the Soviet Union. Another major influence in the development of the Romanian vocabulary came from Greek, especially during the Phanariot period (1711–1821). However, very few of the Greek borrowings are still in use today. Other foreign influences came from Turkish, Hungarian, and German. Turkish influence was prominent in Moldavia

RULES VS. CONSTRAINTS and Wallachia during the Phanariot period, and Hungarian influence was prominent in Transylvania while it was part of the Austro-Hungarian Empire. The German elements used by the Romanian minorities of German origin also played a role in the diversification of the Romanian vocabulary. More recent foreign influences came from French in the first part of the twentieth century, while today massive English influence on the Romanian vocabulary can be easily observed. Within the geographical area of Romania, there are three main varieties of the language: Moldavian, Wallachian (Muntenian), and Transylvanian. Additionally, in the Republic of Moldavia the inhabitants speak the Bessarabian variety of Romanian, which is characterized by the massive influence of Russian. All these varieties, however, are mutually intelligible. The standard pronunciation of Romanian is based on orthographic norms. In this respect, the principle generally promoted is that of identity between writing and pronunciation. As far as the vocabulary is concerned, Romanian inherited c. 2,000 original Latin words, but Latin elements make up 60% of the Romanian vocabulary. Most of the original Latin words refer to parts of the human body and have several meanings, including metaphorical ones. In contrast to the vocabulary, Romanian grammatical structure is almost entirely Latin. There are Latin prepositions (e.g. cu ‘with’) and conjunctions (e.g. s,i ‘and’), Latin interrogative pronouns (e.g. cine ‘who’), Latin endings for marking the plural, Latin verbal endings for marking person and tense, Latin pronouns, and Latin numerals from one to ten. Romanian grammar involves ten parts of speech: six inflected (noun, article, adjective, pronoun, numeral, and verb) and four noninflected (adverb, preposition, conjunction, and interjection). The following grammatical categories all stem from Latin: gender, number, case, person, comparison, voice, mood, and tense. However, the development from Latin to Modern Standard Romanian generated several differences. Thus, for example, Latin had six cases, while Romanian preserves only five—nominative, genitive, dative, accusative, and vocative.

Gender in Romanian distinguishes masculine, feminine, and neuter (masculine and feminine genders are inherited from Latin, while neuter is a creation within Romanian). The changes from Latin are typically simplifications (as in all Romance languages), which led to a reorganization of the vocabulary by reducing the complicated forms of the noun and adjective, by simplifying and rearranging the verbal system, and by simplifying the case system. While the Romance languages usually have reduced nominal declensions, Romanian has three forms of noun inflection inherited from Latin. The declension of adjectives is identical to that of nouns, varying according to gender, number, and case. The pronominal declension is the richest and the closest to the complex Latin system of all the Romance languages. The Romanian verbal system is largely based on Latin. The most noticeable differences from the other Romance languages are the analytic future (voi citi ‘I shall read’) and the supine (de citit ‘for reading’). Romanian is the only Romance language with a free sequence of tenses and a relatively free word order. References Constantinescu-Dobridor, Gheorghe. 1994. Sintaxa limbii române (The syntax of the Romanian language). Bucharest: Editura Stiintifica. B Coteanu, Ion. 1993. Gramatica de baza a limbii române (The fundamental grammar of the Romanian language). Bucharest: Garamond. Daniliuc, Radu, and Laura Daniliuc. 2000. Descriptive Romanian grammar. An outline. Munich: Lincom Europa. Doca, Gheorghe. 1996. Limba Româna ˘. Probleme de sinonimie s, i de cultivare a limbii (Romanian language. Synonymy and style issues). Bucharest: TUB. Harris, Martin, and Nigel Vincent (eds.) 1990. The romance languages. London: Routledge. Pop, Liana. 1993. Romanian with or without a teacher. Cluj: Echinox. Posner, Rebecca. 1996. The romance languages. Cambridge: Cambridge University Press. Serban, Domnica. 1986. Syntactic functions in universal grammar. A contrastive study of the subject in English and Romanian. Bucharest: TUB.

RADU DANILIUC

Rules vs. Constraints The question of whether grammar should be modeled using rules or constraints presents a morass of terminological confusion in addition to some legitimate

theoretical issues. In some work, e.g. Karttunen (1993), the terms ‘(declarative) rules’ and ‘constraints’ are used interchangeably. There are even positions

907

RULES VS. CONSTRAINTS such as Mohanan’s (2000:146), who argues that the distinction is incoherent once we recognize that both rules and constraints express propositions. This article proposes a definition of ‘rule’ vs. ‘constraint’ that will surely be inconsistent with other usages, but which provides the basis for discussing competing models of grammar, focussing here on phonology. The use of both rules and constraints requires a representational matching procedure (RMP) to determine whether a given rule or constraint P applies to a given representation q. Part of the statement of P must be a structural description, SDp. The RMP must evaluate q to determine whether it satisfies SDp. The RMP outputs two possible results: YES, q satisfies the structural description of P; or NO, q does not satisfy the structural description of P. In other words, the RMP defines the domain of the function P. In many theories, the RMP relies on subsumption—q is in the domain of P if SDp subsumes q—although there are other possibilities. If the RMP determines that q is in the domain of P, we can now distinguish between a rule and a constraint. Consider a rule of the form R(I)  O, that is, R maps input I to output O. If q is in the domain of R (i.e. if the RMP determines that q matches I), then that part of q that is identical to I is rewritten as O yielding q: R(q)  q. Thus, we define rules as functions from representations to representations, with both input and output built from the same representational primitives. In a constraint-based system, if the RMP determines that q is in the domain of a constraint C, then an operation of constraint evaluation must determine whether q satisfies the constraint. If q satisfies C, constraint evaluation outputs the value NOVIOLATION; if q does not satisfy C, constraint evaluation outputs the value VIOLATION. These values then serve as input to another component of the grammar. In some models, if there is any constraint that q violates, then q is ungrammatical. In optimality theory (OT), in contrast, the evaluations of q for each constraint are fed to another component of the grammar, which compares q to competing representations. The output of the grammar typically does violate some constraints in OT, but they must be lower ranked than constraints violated by competing representations. To reiterate: a rule is defined as a function from representations to representations; a constraint is defined as a function from representations to the set {VIOLATION, NOVIOLATION}. Note that a constraintbased system requires two evaluations of a given representation q—one corresponding to the RMP to determine whether a constraint C is relevant to q and one to the evaluation of whether q violates C. Equipped with this working distinction between rule and constraint, we can now offer arguments to

908

favor rule-based models of grammar, by pointing out a number of conceptual and empirical problems with constraints. Our first criticism of constraint-based models is related to issues of learnability, acquisition, and the nature of universal grammar (UG), i.e. the human language faculty. Many linguists, especially phonologists, have assumed that both UG and particular grammars contain constraints—either positive or negative conditions that, respectively, must or must not be satisfied by grammatical representations. However, such conditions cannot be learned by positive evidence. This is because a generative grammar generates an unbounded number of well-formed structures, but only a finite sample can be encountered by any point in acquisition. The learner may find a supposed illformed structure in the next sentence. For example, a putative constraint that all syllables must have onsets may be consistent with the data received up to a given point; however, there is no guarantee that the inductive conclusion consistent with the constraint will turn out to be correct. Therefore, such conditions could only be learned via negative evidence. However, it is generally accepted (see Marcus (1993) for convincing arguments) that negative evidence is neither supplied to the child with sufficient regularity, nor attended to by the child when supplied, to play a significant role in language learning. Thus, the constraints cannot be learned via positive evidence (for reasons of logic), nor through negative evidence (according to the empirical data from language acquisition). So they must be innate. This conclusion follows from the premises, and it has led to implausibly rich versions of UG. In phonology, for example, innate constraints have been posited to account for the vast variety of phonological alternations seen in the languages of the world. The OT literature provides examples of constraints to account for voicing of initial obstruents in the second member of Japanese compounds, flapping in English, and so on, thus trivializing the notion of UG. Fortunately, the conclusion of innate constraints can be rejected: the problem lies with the assumption that UG, and also particular grammars, consist of constraints. By positing a set of representational and computational primitives, such as syllables and feet, precedence relations, identity evaluations, etc., and a learning mechanism that combines these primitives into rules that capture alternations, a learner can build a language-specific grammar based solely on positive evidence. By positing highly specific rules that conform to the data, the learning path is constrained without the use of constraints. A cluster of arguments against constraint-based models relate to the fact that they typically depend on adopting a notion of (absolute or relative) ill-formedness

RULES VS. CONSTRAINTS of linguistic representations. Ill-formed representations violate constraints, and this result is passed on to other modules of the grammar. An obvious objection to this approach is that there is a much simpler alternative to the idea that grammars generate ill-formed representations and then reject them as possible outputs—why not assume that the grammar just generates those representations that it outputs? The grammar itself does not need the notion of ill-formedness. Of course, a linguist may refer to a representation as illformed, in the sense of ‘not generatable by a particular grammar L’, but this description by a linguist should not be interpreted as a description of an actual property of L. Instead of accounting for the infinite set of representations that L does not generate, normal scientific practice seems to dictate that we account for what L does generate. Ill-formedness, or rather nonformedness, is thus a derivative notion. Various theories of grammar, including OT and some versions of minimalism and its predecessors posit a mechanism that allows unconstrained generation of linguistic representations. In OT, this device is GEN which, given an input, generates the universal candidate set of possible outputs. In various syntactic theories, an analog to GEN is the ‘free’ concatenation of morphemes, or the ‘free’ application of operations such as Move α. A derivation which is thus generated will either satisfy certain constraints at PF and LF, the grammar’s interface levels, and thus converge; or it will not satisfy those conditions and it will crash. Both the OT approach and the free-generation-with-interface-conditions approach in syntax are flawed in the following (related) ways. First, it is easy to proclaim something like ‘GEN generates any possible linguistic representation’ or ‘The syntactic component allows Move α to apply freely’. However, it is not clear what such statements mean. One could argue that the theory of grammar need not be computationally tractable, since grammar models knowledge and does not necessarily provide an algorithm for humans to produce speech output. However, there must be some mechanism that does generate particular linguistic representations, and it is not unreasonable to ask a theory of grammar to characterize the formal properties of this mechanism. It seems that any implementation of GEN or the syntactic component that incorporates Move α will have to be explicit about what it does. One way to achieve this is to be explicit about what the abstract grammar generates. Second, the free generation-cum-filters model stinks somewhat of antimentalism. It basically says ‘We don’t care how the candidate forms are generated, as long as they are generated. One way is as good as the next, as long as they are extensionally (empirically) equivalent.’ In defining I-language, a matter of

‘individual psychology’ as the domain of inquiry for linguistics, Chomsky (1986) argued convincingly that the fact that knowledge of language is instantiated in individual minds/brains means that there is necessarily a ‘correct’ characterization of a speaker’s grammar. Rejecting the view that the language faculty consists of constraints has methodological implications. Just as we do not want the language faculty to contain constraints against grammatical structure, models of the language faculty should not be described via constraints. If our current hypothesis concerning UG is stated only in positive terms, as statements of what grammars have access to or consist of, without prohibitions or constraints, we can achieve a more economical model. The positive terms are just those entities and operations (features, syllables, deletions, insertions (in phonology); and Merge, Move and featurechecking (in syntax)) which have been observed empirically or inferred in the course of model construction. When faced with a phenomenon that is not immediately amenable to modeling using existing elements of the vocabulary, scientific methodology (basically Occam’s Razor) guides us. We must first try to reduce the new phenomenon to a description in terms of the vocabulary we already have. If this can be shown to be impossible, only then can we justify expanding the vocabulary. Once one accepts that modules or processes of grammar must have a certain set of properties and that these properties ultimately must be describable in terms of a set of positive statements (a vocabulary) and can be incorporated into the structural descriptions of rules, it appears to be the case that a procedural, or rule-based approach to grammar that generates a sequence of representations constituting a derivation is to be preferred to a constraint-based, nonderivational theory. In other words, grammars can be understood as complex functions mapping inputs to outputs. A rule-based model just breaks the complex function into simpler components in order to understand the whole. A theory that incorporates GEN or Move α avoids the problem of characterizing the function that is the grammar. A further objection to the notion of ill-formedness is the fact that the justification for evaluating representations as ill-formed is typically derived from some version of markedness theory. Depending on the formulation of a given constraint, either matching or failing to fulfill the condition specified by the constraint signals VIOLATION. For example, a constraint formulated as ‘Don’t have a coda’ leads to an evaluation of illformedness for a syllable that has a coda. A constraint formulated as ‘Have an onset’ leads to an evaluation of ill-formedness for a syllable that does not have an onset. Relative and absolute ill-formedness or markedness

909

RULES VS. CONSTRAINTS evaluations of linguistic representations are ascribed by linguists for grammar-external reasons, such as ease of articulation or perception, avoidance of ambiguity, etc., and thus have no place in a formal model of grammar— humans know about ambiguity, and can try to avoid it, but grammars do not know anything. Marked or ill-formed structures typically are claimed to have at least one of the following properties: ● ●









Relative rarity in the languages of the world. Late ‘acquisition’ by children (typically referring to the recognizability of a form in child speech). Loss in aphasia (typically referring to the recognizability of a form in aphasic speech). Relative difficulty of perception (not always experimentally validated). Relative difficulty of articulation (based on impressions of what is hard to say). Tendency to be lost in language change and to not arise in language change.

All of these criteria have been criticized by Hale and Reiss (2000a,b; see references therein). These works conclude that the best way to gain an understanding of the computational system of phonology is to assume that the phonetic substance (say, the spectral properties of sound waves, or the physiology of articulation) that leads to the construction of phonological entities (say, feature matrices) never directly determines how the phonological entities are treated by the computational system. The computational system treats features as arbitrary symbols. An additional consideration in evaluating markedness and the notion of grammatical ill-formedness, pointed out by John Ohala over the years, is that building into a theory of universal grammar the articulatory or perceptual basis for a recurring sound pattern leads to an unacceptable duplication of explanatory mechanisms. For example, if we can show in a laboratory that consonant place distinctions are less salient in codas than in onsets, we do not need to build into

phonological theory innate constraints that lead to place feature neutralization in codas. The facts of perception demonstrated in the laboratory, combined with a theory of language acquisition and change, will account for the attested patterns. Offering an additional cause, within the phonology, is just bad science. Based on learnability considerations, the weakness of markedness considerations, and various conceptual arguments, we conclude that a rule-based derivational model of grammar is superior, since it can be stated in purely positive terms, without prohibitions. The problem of inductive uncertainty is avoided, by the learner and the linguist, if we adopt a rule-based model of grammar, as defined here. References Chomsky, Noam. 1986. Knowledge of language. Westport, CT: Praeger. Epstein, S., E. Groat, R. Kawashima, and H. Kitahara. 1998. A derivational approach to syntactic relations. Oxford: Oxford University Press. Hale, M., and C. Reiss. 2000a. Substance abuse and dysfunctionalism: current trends in phonology. Linguistic Inquiry 31. 157–69. Hale, M., and C. Reiss. 2000b. Phonology as cognition. Phonological knowledge, ed. by N. Burton-Roberts, Philip Carr, and Gerry Docherty. Oxford: Oxford University Press. Karttunen, Lauri. 1993. Finite state constraints. The last phonological rule, ed. by John Goldsmith. Chicago, IL: University of Chicago Press. Mohanan, K. P. 2000. The theoretical substance of optimality formalism. The Linguistic Review 17. 143–66. Ohala, J. 1990. The phonetics and phonology of aspects of assimilation. Papers in laboratory phonology I: between the grammar and physics of speech, ed. by J. Kingston and M. Beckman. Cambridge: Cambridge University Press. Reiss, C. 2003. Deriving the feature-filling/feature-changing contrast: an application to Hungarian vowel harmony. Linguistic Inquiry 34. 199–224. ———. Constraining acquisition without constraints. Rules and constraints in phonology, ed. by A. Nevins and B. Vaux. Oxford: Oxford University Press, to appear.

CHARLES REISS

Russell, Bertrand Arthur William A leading British philosopher, logician, essayist, pacifist, and political activist, Bertrand Russell was born in 1872 into a liberal and aristocratic family but was sadly orphaned at the age of 4. He grew up as an enfant terrible and later a tireless critic of the establishment,

910

only to be invested with the Order of Merit in 1949 and, a year later, awarded the Nobel Prize for Literature. By the time of his death in 1970 at the age of 97, he had authored some 70 books, and around 4,300 articles, book chapters, reviews, pamphlets etc.,

RUSSELL, BERTRAND ARTHUR WILLIAM besides more than 61,000 letters and other documents not originally intended for publication. Along with G.E. Moore, he is widely acclaimed as one of the founders of analytic philosophy and, with Kurt Gödel, he shares the reputation of having made the most significant contribution to the study of logic in the twentieth century. An unflinching pacifist, he was imprisoned for involvement in antiwar protests and dismissed from his job at Cambridge University in 1916, and in 1940 from his job at the City College, New York; years later, in 1961, he was again arrested and imprisoned for a week, this time for engaging in antinuclear protests. As an essayist and pamphleteer, Russell addressed a number of issues of contemporary social, political, and moral concern and never shied away from expressing his views boldly and openly as in his book Why I am not a Christian. Russell was a strong advocate of logicism—the view that all of mathematics is ultimately reducible to logic. To him goes the credit for having directed the attention of fellow philosophers and logicians to the utmost importance of detecting and rooting out contradictions in set theory and elsewhere. In 1901, Russell discovered what is famously known in the literature as ‘Russell’s paradox’, later fully elaborated in his work Principles of mathematics (1903). The paradox, or antinomy as it is sometimes called, was identified by imagining a set of all sets that were not members of themselves and asking, apropos of that resultant set, if it was a member of itself, the answer being that it can only be a member of itself by not being a member of itself—which is a blatant contradiction. The discovery of this paradox stimulated intense research in set theory and led to several suggested solutions. Russell’s own solution was elaborated in his theory of types, which basically consisted in postulating a hierarchy of sentences with strict prohibition on mixing levels and thus preempting the very question that led to the discovery of the paradox in the first place. Among Russell’s most significant contributions to the philosophy of language is his famous ‘theory of definite descriptions’ put forward in response to the German logician Gottlob Frege’s theory of ‘sense’ (Sinn) and ‘reference’ (Bedeutung). Frege had argued that two expressions, say, ‘the morning star’ and ‘the evening star’, although clearly different in their senses, can nevertheless have the same referent, namely, in this particular instance, the planet Venus. Despite its intuitive appeal, Russell found the distinction unsatisfactory and was appalled by what Frege had shown to be an important consequence of the distinction for the very enterprise of logic as it had been traditionally conceived. For, in Frege’s view, the sentence ‘The present king of France is bald’, uttered at a time when France was no longer a monarchy (or, equivalently,

when there was no referent to the descriptive phrase ‘the present king of France’), was bereft of any definite truth value, because the existence of the referent is a precondition (or, technically, presupposition) for what is asserted by the sentence to be capable of being either true or false. When the presupposition is not true, what is asserted by the sentence will be neither true nor false. This meant that the time-honored principle of logic known as the law of the excluded middle according to which, given a sentence and its (contradictory) negation, one of them has to be necessarily true no matter what the world is like would be, from now on, subject to the additional constraint that stipulated ‘provided the sentence’s presuppositions are true’. Russell realized that such a consequence was simply devastating for logic’s reputation as a ‘science of pure reason’. Russell’s alternative analysis began by showing that the logical form of the sentence under discussion was much more complex than Frege had thought: it was, he argued, actually a conjunction of three separate propositions, namely: (1) There is a present King of France. (2) There is only one king of France. (3) Whosoever is the King of France is bald. Now, it is commonplace in logic that the negation of a compound sentence could be shown to be true if any one (or more) of its constituents is known to be false. So, the nonexistence of a King of France was as strong a reason for a sentence affirming his baldness to be false as the fact of the king being hirsute. What Frege took to be a presupposition of the sentence was thus shown to be capable of being treated as one of its straightforward entailments. Whereas the Fregean analysis demanded a thorough revision of one of the fundamental claims on behalf of logic (in addition to the familiar ‘true’ and ‘false’, there was the need to admit of ‘neither true nor false’ as a third possibility), Russell’s solution required no new apparatus other than the standard bivalent logic. From the 1970s onward, many linguists such as Deirdre Wilson (1975) and Ruth Kempson (1975) have followed Russell’s lead and urged that many of the phenomena treated by earlier researchers as presuppositions should be better treated as semantic entailments, and any residual problems such as favored readings (including the intuitive feeling that certain presuppositions do hold good in determinate circumstances) should be accounted for by a separate pragmatic component that will have recourse to such theoretical apparatuses as Grice’s theory of conversational maxims. The decision to consign part of the problematic to pragmatics may be seen as having as its ultimate inspirational source a key element in P.F. Strawson’s objections to Russell’s analysis on the grounds that it did not do full justice to ordinary language. (In a protracted polemic with Russell, Strawson

911

RUSSELL, BERTRAND ARTHUR WILLIAM insisted that there was a need to distinguish a sentence or an expression from a use and an utterance thereof.) Russell’s theory of descriptions provided the logical basis for his epistemology centered on a distinction between ‘knowledge by acquaintance’ and ‘knowledge by description’. Among the objects that could be known through acquaintance Russell included what he referred to as ‘logically proper names’, whose function was to logically refer to certain objects such as sense data and universals. Ordinary names, by contrast, were regarded by Russell as functioning as definite descriptions in disguise. Russell thus pioneered the program known as ‘logical atomism’, the locus classicus of which is Wittgenstein’s Tratactus logico-philosophicus (translated into English under the same title in 1922), originally prepared under Russell’s own supervision. In Russell’s view, the world is wholly made up of atomic facts and every meaningful proposition must be composed of constituents with which we are acquainted. Among Russell’s major publications is the monumental work Principia mathematica (1910–1913), written in collaboration with his Cambridge tutor Alfred North Whitehead. His 1912 work, The problems of philosophy, has long served as one of the popular introductions to the subject for undergraduates.

Biography Bertrand Russell was born on May 18, 1872 at Ravenscroft, Wales. He was orphaned at the age of 4, with the death of his mother and sister in 1874, and his father in 1876. In 1876, Russell’s grandfather, Lord John Russell (a former Prime Minister of Great Britain), and grandmother succeeded in overturning his father’s will to win custody of Russell and his brother. With the death of his grandfather in 1878, Russell’s grandmother began supervising his upbringing. In 1890, he went to Trinity College, Cambridge, and three years later was awarded a first class B.A. in Mathematics. In 1901, he discovered Russell’s para-

dox. From 1907 to 1910 he worked on Principia mathematica in collaboration with A.N. Whitehead. He was elected Fellow of the Royal Society in 1908. He was dismissed from Trinity College for antiwar protests (1916), and was imprisoned for six months for involvement in antiwar protests (1918). He was awarded the Order of Merit (1949) and Nobel Prize for Literature (1950). In 1955, he released the Russell–Einstein Manifesto condemning the nuclear arms race. In 1958, he became founding President of the Campaign for Nuclear Disarmament and was imprisoned for one week in 1961 for leading antinuclear protests. He died in Penrhyndeudraeth, Wales, on February 2, 1970. References Ayer, A.J. 1972. Russell. London: Fontana/Collins. Clark, Ronald. 1975. The life of Bertrand Russell. London: J. Cape. Kempson, Ruth. 1975. Presupposition and the delimitation of semantics. Cambridge: Cambridge University Press. Irvine, A.D. (ed.) 1998. Bertrand Russell: critical assessments, 4 vols. London: Routledge. Monk, Ray. 1996. Bertrand Russell: the spirit of solitude. London: J. Cape. Russell, Bertrand Arthur William, and Alfred North Whitehead. 1910, 1912, 1913. Principia mathematica, 3 vols. Cambridge: Cambridge University Press. Russell, Bertrand Arthur William. 1912. The problems of philosophy. London: Williams and Norgate; New York: Henry Holt and Company. ———. 1914. Our knowledge of the external world. Chicago and London: The Open Court Publishing Company. ———. 1927. Why i am not a Christian, London: Watts; New York: The Truth Seeker Company. ———. 1929. Marriage and morals. London: George Allen & Unwin; New York: Horace Liveright. ———. 1945. A History of western philosophy. New York: Simon and Schuster; London: George Allen & Unwin, 1946. ———. 1959. My philosophical development. London: George Allen & Unwin; New York: Simon and Schuster. Sainsbury, R.M. 1979. Russell. London: Routledge. Wilson, Deirdre. 1975. Presuppositions and non-truth conditional semantics. London: Academic Press.

KANAVILLIL RAJAGOPALAN

Russian and East Slavic Languages The East Slavic languages—Russian (R), Ukrainian (U), Belarusian (Br)—constitute one of the three branches of the Slavic language family. All three East Slavic languages are written in the Cyrillic alphabet, but Belarusian has a parallel Latin orthographical tradition dating from the mid-sixteenth century.

912

The East Slavic languages developed from the eastern dialects of Late Common Slavic, when speakers migrated northward and eastward from the presumed Slavic homeland north of the Carpathians around 500 CE into territory sparsely populated by Finno-Ugrian and Baltic peoples. By the eighth century, a lively

RUSSIAN AND EAST SLAVIC LANGUAGES trade in furs, amber, wax, honey, slaves, and silver across this vast territory attracted Scandinavian warrior-traders, the Rus’, who sought to control the routes that linked the markets of Europe and Asia. The Rus’ dominated the East Slavic tribes in the ninth and tenth centuries, consolidating power first in the lands of the north around Lake Il’men’ and at Novgorod, and then in the south at Kiev. They gradually assimilated into the larger Slavic population, leaving traces in the form of a few dozen loanwords (e.g. R jákor, U jákir, Br jákar ‘anchor’; cf. Old Swedish ankari), but most importantly in the very name Rus’, ascribed first to the East Slavs around Kiev and derivatively to the Kievan lands. The name ultimately came to refer to East Slavic lands collectively and by extension to all the early East Slavs and to their language, which for the early historical period (up to the fourteenth century) can be called Rusian. Our direct knowledge of Rusian is complicated by the fact that Rusian Church Slavonic, and not Rusian, was the first written language, gradually introduced after the conversion of the East Slavs to Christianity in 988. Rusian Church Slavonic (hereafter Slavonic) was a local variant of Old Church Slavonic, the liturgical language devised by Saints Cyril and Methodius on the basis of their ninth-century South Slavic (Macedonian) dialect for use in spreading Byzantine Christianity among the Slavs. Slavonic was basically comprehensible to the East Slavs but contained a number of superficially South Slavic features, primarily of a phonological and morphological nature. For example, the Common Slavic roots *gard- ‘fortified town, wall’ and *berg- ‘bank, shore’ were realized as grad-, brêg- in Slavonic, whereas the vernacular Rusian had the regular East Slavic pleophonic counterparts gorod-, bereg-, respectively. Similarly, the Slavonic and Rusian reflexes of the Common Slavic sequence *tj were šJ (shch) and J (ch), respectively, and thus Slavonic svêšJ a and Rusian svêJ a ‘lamp; candle’, both derived from the Common Slavic root *svêt- ‘light’ plus the suffix -j-. Despite such formal differences, the basic grammar and the majority of the most common lexical items in Slavonic and Rusian were identical. Some scholars prefer to treat them as variants or registers of a single language, whereas others view them as two closely related but separate languages. The latter view is adopted here. In addition to the superficial distinctions noted above, written Slavonic differed from vernacular Rusian in its richer lexicon, and in its potentially more complex syntax, both inspired by elements of the Bulgarian Slavonic, Serbian Slavonic, or Greek texts that served as sources for translation. Although the primary written record of early Rus’ appeared in Slavonic rather than Rusian, lapses in the copying and editing of

texts provide evidence of the actual speech habits of the Rusian scribe producing them. Since the 1950s, archeological exploration has revealed direct examples of written Rusian in documents uncovered in predominately northwestern locales, especially Novgorod, Pskov, and Staraja Russa. Unlike their Slavonic counterparts preserved on costly parchment, these legal and financial documents or personal correspondence were scratched onto disposable birchbark. They provide valuable examples of the Rusian vernacular in a geographically circumscribed area from the eleventh through the fifteenth centuries. From the very beginning, strictly Rusian features mingled with those of Slavonic, a situation that became more common as new texts, created and borrowed, provided fresh opportunities for linguistic choice. Generic and thematic factors certainly influenced the predilection for Slavonic in religious and other solemn texts and Russian in mundane ones, but they were not overriding. At times, for example, a scribe might prefer Slavonic grad ¯b over Rusian gorod ¯b simply to save the space of one letter on the expensive parchment at the end of a line. Although there was dialectal divergence among the East Slavs by the time of Christianization, virtually no evidence exists to suggest that it caused any difficulty in communication between residents of the Novgorodian north and the Kievan south. For this reason, the vernacular language of the early East Slavs is best considered as Rusian alone, with dialectal variants distributed geographically. Nonetheless, linguistic evolution and historical circumstance helped to extend and deepen these distinctions to a point that it is more appropriate to speak of separate languages rather than dialects. The effects of the phonological change known as the ‘jer shift’ produced dramatic differences north and south, alternative choices that affected consonant and vowel inventory as well as morphophonemic and morphological alternation (see below). Historical and political factors also served to create distinct communities of speakers whose speech patterns became increasingly divergent from one another and subject to influence from other languages. For example, the growing importance of such northeastern centers as Rostov, Suzdal’, and Vladimir (twelfth century) in the face of older Novgorod and Kiev; the Mongol invasions (mid-thirteenth century); the Novgorodian colonization of the north (thirteenth to mid-fifteenth centuries); the incorporation of southwestern Rus’ into the Grand Duchy of Lithuania (thirteenth to fourteenth centuries) and Poland (fourteenth century); and the ascendancy of Moscow (fourteenth to fifteenth centuries) all contributed to the enhancement of regional linguistic difference.

913

RUSSIAN AND EAST SLAVIC LANGUAGES By the fourteenth century, two language areas of Rus’ had developed: a northern, Russian one destined to come under the hegemony of Moscow, and a southern, Ruthenian one, at first found largely within the Lithuanian political orbit, and then from 1569, under the rule of the Polish–Lithuanian Commonwealth. In Ruthenian territory, the prominent distinctions between Belarusian and Ukrainian were apparent by the fifteenth to sixteenth centuries, although their beginnings are discernable three to four centuries earlier. From the mid-seventeenth century, parts of the Ukrainian and Belarusian territories gradually fell under the control of Muscovy, then the Russian Empire, and finally the Soviet Union, a process that extended well into the twentieth century. Consideration of the structure and dialectal distribution of Rusian provides a common basis for comparing and contrasting the development of the East Slavic languages with respect to each other. Rusian had a rich system of vowels (i,y,u,ê,e,o,ä,a,b,¯b), including two high, lax vowels, the jers (symbolized as front b and back ¯b, respectively), a higher-mid tense front vowel (symbolized as ê), and a low tense front vowel (symbolized as ä). Over the course of the eleventh and twelfth centuries, the jers were either reidentified as mid-vowels e and o, or were lost (e.g. Russian pbnb ‘stump’, s¯bn¯b ‘sleep; dream’  R,Br p’en’, son, U pen’, son). The jer shift reduced the number of distinctive vowels by a third and nearly doubled the inventory of consonants by creating sets of distinctively paired plain (hard) labial and dental consonants (p,b,m,v,t,d,s,z,n,l,r) and palatalized (soft) counterparts (p’,b’,m’,v’,t’,d’,s’,z’,n’,l’,r’). Russian has largely preserved this later Rusian phonemic inheritance, whereas Ukrainian, apart from the dispalatalization (hardening) of consonants before Rusian i and e, preserved most soft dentals, but hardened soft labials. Belarussian preserved soft labials before vowels, hardened r’ to r, and affricated t’ and d’ to c’ (ts’) and ’ (dz’), the result of assibilation, so-called cekan’e and dzekan’e, respectively. After the jer shift, most new sequences of dentals or palatals plus j yielded soft geminate (doubled) dentals and palatals (r’ and j are the exceptions) in Ukrainian and Belarusian, a reflex not typical of Russian, e.g. Rusian Dit b je ‘life’, U Dyt’t’a, Br Dic’c’o, R Dit’jo. Many Rusian dialects had a seven-vowel system after the jer shift: i,e,a,o,u plus the tense vowels ê and ô. The latter two were distributed differently, north and south, and were ultimately the most unstable. At the present time, standard Russian and Belarusian each have five-vowel systems (i,e,a,o,u), whereas standard Ukrainian has a six-vowel system (i,y,e,a,o,u). The higher-mid front y is a characteristically Ukrainian distinctive vowel, derived from merged Rusian i and y,

914

with ê (and typically ô as well) subsequently raising and fronting to i (so-called ikavism), e.g. Rusian lês¯b ‘forest’, stol¯b ‘table’  U l’is, st’il. The oldest layers of dialectal differentiation in Rusian follow a path from southwest to north and northeast, reflecting the early direction of Slavic development from the original Carpathian center to the northern and eastern periphery. Later innovation is stimulated by the rise of new centers of cultural diffusion, as noted above. Thus, among the oldest isoglosses is one that distinguishes dialects that preserve Common Slavic *g as a stop (the more ‘peripheral’ North Russian) from those that have lost closure, producing a fricative like γ or h (the more ‘central’ South Russian, Belarusian, Ukrainian). Another isogloss distinguishes the development of the tense jers (in the environment before j) as mid-vowels e and o (the more ‘peripheral’ Russian) or vowels i or y (the more ‘central’ Ruthenian), e.g. Rusian m¯bju, R moju, U myju, Br miju [mýju] ‘I wash’. A third, involving the hardening of palatalized labial and dental consonants before Rusian e, separates the dialects of Russian and Belarusian (presence of palatalization) from those of Ukrainian (absence of palatalization). Some archaic Carpathian dialects of Ukrainian have preserved the Rusian phonemic distinction between i and y, unlike the rest of East Slavic. Rusian, like Slavonic, had a rich system of grammatical categories. Nouns, pronouns, and adjectives were inflected by gender (masculine, feminine, neuter), number (singular, dual, plural), case (nominative, accusative, genitive, locative, dative, instrumental), animacy, and a special vocative form. Verbs were capable of distinguishing mood, tense, voice, and aspect. Participles and adjectives were inflected by gender, number, and case according to two declensions, a ‘short’ nominal declension and a ‘long’ adjectival declension, the latter providing expression for definiteness or generic status. All three East Slavic languages have retained the dynamic, free stress characteristic of Rusian. By the fourteenth century, a number of Rusian grammatical categories had been lost, no longer productive in any of the East Slavic languages. These include dual number, the aorist and imperfect tenses, and case in short adjectives and participles. The perfect took over the function of past tense in all of East Slavic. Ukrainian has preserved a separate vocative form. Ukrainian and Belarusian both have a rarely encountered pluperfect tense. The morphophonemic velar ~ dental sibilant alternation found in most Rusian dialects was retained on a more limited basis in Ukrainian and Belarusian, but lost in Russian (e.g. Rusian ruka/rucê ‘hand, arm’ nom.sg./dat.-loc. sg., U ruka/ruc’i, Br ruka/ruce, but R ruka/ruk’e). Of syntactic interest are

RUSSIAN AND EAST SLAVIC LANGUAGES the many East Slavic constructions without overt nominative subjects (e.g. R mn’e ujt’i ‘I have to go’ [lit. to me to go], U cerkvu bulo zbudovano ‘The church was built’ [lit. churchacc. wasneut.sg. builtneut.sg.], Br m’an’e n’e bilo [byló] doma ‘I wasn’t home’ [lit. of me not wasneut.sg. at home]).

Russian Spoken by approximately 275 million people, 165 million of them natively, Russian is by far the largest of the Slavic languages. The vast majority of Russian speakers reside in the Russian Federation or constitute substantial minorities in several of the republics of the former Soviet Union, including Ukraine, Belarus, Latvia, Kyrgyzstan, and Kazakhstan. It is the major language of the Russian Federation and one of the official languages of Belarus and Kazakhstan. There are significant Russian-speaking communities in Western Europe, the United States, and Canada. Russian is one of the five official languages of the United Nations. The history of written language in Russian territory is one of continuing tension between Slavonic and vernacular Russian from the fourteenth to the eighteenth centuries. The church-supported emergence of Moscow as the center of the evolving Russian state saw the development of Slavonic into an ever more ornamental, complicated medium with intentional archaization from the late fourteenth century onward, a trend that resonated with the late fourteenth-century reforms in Bulgaria. These changes had the effect of distancing Slavonic from the spoken language by eliminating vernacular forms and introducing artificial elements and grammatical patterns. At the same time, during the fifteenth and sixteenth centuries, the language of the growing Muscovite bureaucracy—the formulaic, vernacular-based chancery language—displayed a normative tendency that elevated the speech patterns of Moscow and avoided the dialectal features of other areas of Russia. The union of Left-Bank Ukraine and Muscovite Rus’ in the mid-seventeenth century provided Ukrainian bookmen with the opportunity to bring to Muscovy their better developed grammatical tradition based on Latin and Greek models. They codified and enriched the Slavonic tradition, particularly with foreign vocabulary from the West, including Latin, primarily through Polish and Ukrainian intermediaries. In their turn, Slavonic and the chancery language influenced the development of vernacular Russian. It was only in the mid-eighteenth century, especially through the grammatical and lexical reforms of Adodurov and especially Lomonosov, that there was a reasoned attempt to control the distribution of foreign, Slavonic, and Russian elements in the Russian written language.

Karamzin in the late eighteenth to early nineteenth century and then Pushkin helped to refine the Russian language by example, setting the stage for it to become a vehicle capable of artistic and scientific expression on the highest levels of European culture. Russian is traditionally divided into three major dialect zones: north, south, and central. South Russian is distinguished by the presence of two innovations: (1) the fricative γ instead of original stop g and (2) akan’e. Akan’e refers to the pronunciation of unstressed o and a as [a] initially or after a hard consonant in the first pretonic syllable (the syllable immediately preceding the stressed one) and as central lax schwa elsewhere (symbolized [ə]): compare molodój [məladój] ‘young’ and nagradnój [nəgradnój] ‘pertaining to reward’. The large transitional zone between North and South Russian is composed of the central dialects (including Moscow), which show their intermediate status in favoring northern consonantism with g and southern vocalism with akan’e. It is the Muscovite dialect that ultimately served as the basis of the standard language: compare NoR mogú [mogú], SoR moγú [maγú], and standard Russian mogú [magú]. In addition to akan’e, standard Russian is also characterized by ikan’e, the pronunciation of unstressed o,a,e as [i] in the first pretonic syllable after a soft consonant, e.g. n’os’í [n’is’í] ‘carry’ (imperative 2sg.), zan’alá [zən’ilá] ‘occupy’ (past fem.sg.), n’emój [n’imój] ‘dumb’ (adjective nom.sg.masc.)

Ukrainian Ukrainian is spoken by approximately 49 million people, making it the second largest Slavic language. Most of the speakers of Ukrainian reside in Ukraine, but there are substantial communities in Russia, Poland, Slovakia, Canada, and the United States. Ukrainian is the state language of Ukraine. In addition to Ruthenian (rus’ka mova) and plain talk (prosta mova), it was also called Little Russian from the midseventeenth century to the early twentieth, a derivative from Little Russia, itself a calque of a Byzantine Greek administrative term used to distinguish the more proximate Ukrainian territory from the more distant Muscovite (called Great Russia). From the late fourteenth century, Ruthenian Slavonic in Ukrainian and Belarusian territories underwent archaization analogous to that of Slavonic in Russian territory. By this time, Galicia had been absorbed into the Kingdom of Poland and the rest of the Ukrainian and Belarusian territories into Lithuania. This turn of events had far-reaching linguistic implications for both Ruthenian and Slavonic. In Polish Galicia the vernacular-based Ruthenian chancery language was ultimately replaced by Latin in accordance with a 1433 decree. In

915

RUSSIAN AND EAST SLAVIC LANGUAGES Lithuania, on the contrary, Ruthenian was elevated in status to become the official language of government administration. Based on the East Slavic vernacular spoken around Vilna (Vilnius), the chancery language developed as a kind of koine, which favored specific Belarusian elements over Ukrainian ones, but avoided the more remarkable characteristics of either, such as Belarusian cekan’e/dzekan’e and akan’e, and Ukrainian ikavism, in its various developmental stages. By the middle of the sixteenth century, spoken and written Ruthenian was making its influence felt in Slavonic in the presence of greater numbers of vernacular forms. The 1569 Union of Lublin resulted in the merger of Poland and Lithuania as the Polish–Lithuanian Commonwealth. With Poland as the dominant partner, Polish and, to a lesser extent, Latin were introduced as the official languages of government administration throughout the land. Without administrative function, Ruthenian became subject to local adaptation. Features from the western and northern Ukrainian dialects began to dislodge their Belarusian counterparts in the Ukrainian territories to the south, whereas more prominent Belarusian features were developed in the Belarusian territories to the north. Both were subject to heavy Polonization and influence from Slavonic as well. After the 1596 Union of Brest brought together Catholics and Orthodox Uniates, Polish and Latin began to challenge Slavonic as well in the spiritual sphere. A Slavonic revival begun in the western Ukrainian town of Ostrih resulted in the publication of a complete translation of the Bible in 1581. The revival spread later to L’viv and Kyiv and resulted in new Slavonic dictionaries and grammars. Ukrainian had become the administrative language in Ukrainian territory east of the Dnieper after an uprising in 1648 resulted in the establishment of a Cossack state, the Hetmanate. It retained that role even after the union with Muscovy in 1654. Mazepa’s unsuccessful attempt in 1709 to wrest the Hetmanate from Russian control led to the replacement of Ruthenian Slavonic by Russian Slavonic. Late in the century, Russian replaced Ukrainian as the language of administration, both in the territories of the old Hetmanate, and in the Right-Bank Ukrainian territory acquired in the 1793 and 1795 partitions of Poland. The emergence of the modern Ukrainian standard language begins with Kotljarevs’kyj’s translationreworking of the Aeneid (1798, 1809, 1842), in which he showed a preference for southeastern dialect features (Poltava, Kharkiv, Kyiv). He prepared the way for a flowering of literary Ukrainian during the period of Ukrainian Romanticism (1820s–1840s), chiefly found in the writings of Kvitka, ŠevJenko, and Kuliš, all of whom sought to realize a synthesis of Ukrainian

916

and Slavonic, although not to the degree seen in Russian. Harsh restrictions were imposed by the Russian government on the use of Ukrainian in 1863, leading to the ban on its public use in 1876, a prohibition lifted only in 1905. Meanwhile, the Ukrainian literary tradition that had continued in Polish Galicia (ceded to Austria in 1772, 1795) had an impact on the future development of Ukrainian in eastern Ukraine. The two centers of Ukrainian culture were joined only after World War II. The Soviet period was marked by an intense government campaign of Russification. Modern Ukrainian competes with Russian and a Russianized Ukrainian hybrid called surzhyk, especially in the capital of Kyiv and in eastern Ukraine. Since independence (1991), Ukrainian, the sole state language, has gained considerable ground in the official sphere, especially in government, communication, and education. Ukrainian is traditionally divided into three major dialect zones: northern, southwestern (including the Carpathians, eastern Poland, and eastern Slovakia), and southeastern. A typical feature of the northern dialects is the presence of diphthongs from older ê and ô under stress that correspond to monophthongal vowels in the standard language in either the southwest or southeastern variant. The southeastern dialects are characterized by ikavism, whereas this process has been only partly realized in many southwest dialects.

Belarusian Belarusian has approximately 10 million speakers, most of whom reside in Belarus. There are small communities of Belarusian speakers in Poland, the United States, and Canada. One of the official languages of Belarus, Belarusian (also rarely Belarusan) has been called Belorussian or Byelorussian, a name influenced by the incorporation of Belarusian territory into the Russian Empire. The English rendering of Belarusian as ‘White Russian’ produces an unfortunate ambiguity with the political term referring to the enemies of the Bolshevik Revolution in Russia. Following the Union of Brest in 1596, Belarusian (plain talk) increasingly yielded to Latin and finally to Polish as the authoritative written language, and in 1697 it was banned from administrative use completely. During the eighteenth century, Belarusian was largely limited to the burlesque interludes for religious school dramas, themselves written in Slavonic, Latin, or Polish. After Belarusian territory was incorporated into Russia during the partitions of Poland in 1772 and 1795, the Russian government banned the use of Belarusian in schools and in publications, a proscription that lasted until 1905. The publication of a

RUSSIAN AND EAST SLAVIC LANGUAGES normative grammar by Branislaw TaraškeviJ (1918 and later editions) was a major step toward the establishment of a standard literary language. Belarusian was subjected to heavy Russification during the Soviet period. Its fate remains precarious in contemporary Belarus: even in the capital of Minsk, one is more likely to hear Russian or a Russianized Belarusian hybrid called triasanka than Belarusian itself. There are three major dialect areas of Belarusian: northeast, southwest, and a broad central one (including Minsk), transitional to both. The northeastern dialects, themselves phonologically and morphologically transitional to Russian, feature dissimilative akan’e and jakan’e. Jakan’e refers to the neutralization of o,a,e in unstressed environments after soft consonants in which one of the pronounced variants is [a]. The dissimilative principle yields two basic variants of neutralization in the first pretonic syllable: (1) the low vowel [a] and (2) a nonlow vowel, commonly [ə] in akan’e and [i] in jakan’e. The first pretonic vowel in each case dissimilates the stressed (tonic) vowel in height: if the tonic vowel is high, the first pretonic variant is low, and vice versa; thus vodí [vadý] ‘water’ (gen.sg.), but vodá [vədá] ‘water’ (nom.sg.), n’osú [n’asú] ‘carry’ (nonpast, 1st sg.), but n’os’éš [n’is’éš] ‘carry’ (nonpast, 2nd sg.). The southwestern dialects are phonologically and morphologically transitional to Ukrainian. In all but the southernmost regions, which are devoid of akan’e and jakan’e, strong (nondissimilative) akan’e and jakan’e are predominant: the first pretonic neutralized variant of o,a,e is always realized as [a], regardless of the nature of the tonic vowel; thus [vadý], [vadá], [n’asú], [n’as’éš]. Standard Belarusian,

based on the central dialects (including Minsk), has strong akan’e and jakan’e as well. References Avanesaw, R.I. 1964. Narysy pa belaruskaj dyjalektalohii. Minsk: Navuka i tèxnika. Avanesov, R.I., and V.G. Orlova. 1965. Russkaja dialektologija. Moscow: Nauka. Comrie, Bernard, and Greville G. Corbett (eds.) 1993. The Slavonic languages. London and New York: Routledge. Issatschenko, Alexander V. 1980. Russian. In Schenker and Stankiewicz. Mayo, Peter. 1993. Belorussian. In Comrie and Corbett. McMillin, Arnold. 1980. Belorussian. In Schenker and Stankiewicz. Schenker, Alexander M., and Edward Stankiewicz (eds.) 1980. The Slavic literary languages: formation and development. New Haven: Yale University Press. Shevelov, George Y. 1979. A historical phonology of the Ukrainian language. Heidelberg: Carl Winter Universitätsverlag. ———. 1980. Ukrainian. In Schenker and Stankiewicz. ———. 1993. Ukrainian. In Comrie and Corbett. Strumins’kyj, Bohdan. 1984. The language question in the Ukrainian lands before the nineteenth century. Aspects of the Slavic language question, Vol. 2: East Slavic, ed. by Riccardo Picchio and Harvey Goldblatt. New Haven: Yale Concilium on International and Area Studies. Timberlake, Alan. 1993. Russian. In Comrie and Corbett. Vlasto, A.P. 1986. A linguistic history of Russia to the end of the eighteenth century. Oxford: Clarendon Press. Wexler, Paul. 1977. A historical phonology of the Belorussian language. Heidelberg: Carl Winter Universitätsverlag. Worth, Dean S. 1998. Language. Modern Russian culture, ed. by Nicholas Rzhevsky. Cambridge: Cambridge University Press. Cylko, F.T. 1966. Narysy z dialektolohiji ukrajins’koji movy, 2nd edition. Kyiv: Radjans’ka Škola.

MICHAEL S. FLIER

917

S Sandhi Sandhi rules are phonological alternations that are triggered at junctures of words or morphemes. The name sandhi comes from Sanskrit and means ‘juncture’. It does not designate one particular phonological process. Instead, it is a nonspecific cover term for any kind of sound mutation that occurs at the edges of words and morphemes, and that is triggered in environments created by morphological or syntactic concatenation operations. Sandhi phenomena are very common across the languages of the world. Some of the best-known examples are tone sandhi, which is found in Chinese for example, Welsh and Irish consonant mutations, and a set of Sanskrit sandhi rules. Two types of sandhi can be distinguished: internal and external sandhi. Internal sandhi rules are triggered word internally, between two morphemes. External sandhi rules operate at word edges, when two words become adjacent as a result of some syntactic process in the language. An example of internal sandhi can be found in English words such as symbol and symbiosis. Both words are derived by addition of the prefix syn to the stem. In its basic form, this prefix can be observed in words like synthesis, synopsis, syntax, synchronicity, synonym, etc. When combined with a stem beginning with /b/, as in the words symbol and symbiosis, the last consonant of the prefix changes to /m/. This internal sandhi rule is an instance of the phonological process known as place assimilation, in which the place of articulation of some consonant becomes dependent on the place of articulation of an adjacent consonant. In the case at hand, the place of articulation of the consonant /n/ becomes identical to the place of

articulation of the consonant that follows it. The latter consonant can therefore be said to act as the trigger for this phonological change. One of the internal sandhi rules of Sanskrit is a phonological process known as voice assimilation. The phonological process in question leads to a change of voiceless consonants, such as /p/, /t/, and /d/, turning them into their voiced counterparts /b/, /d/, and /g/. The effect of this rule is observable in the compound sadaha ‘good day’ derived from sat ‘good’ and aha ‘day’. Compound formation leads to a configuration in which the two segments––the devoiced word final /t/ and the voiced, word initial /a/––become adjacent, and thus creates the conditions for the application of voice assimilation. Interestingly, the same rule that operates word internally in Sanskrit, between two portions of a compound, also applies as external sandhi between two words of a phrase. For instance, the phrase samyag uktam ‘correctly spoken’ consists of the word samyak ‘correctly’ and uktam ‘spoken’. Again, voice assimilation of the word-final consonant, in this case /k./, can be observed under the influence of the following voiced segment. In English, external sandhi may occur in phrases such as miss you, kiss you, caught you, etc, where the word final consonant of the verb changes to [sh] or [ch] to give mi[sh] you, ki[sh] you, and cough[ch] you. This phonological process is known as palatalization. Another example of external sandhi is known by the name of intrusive-R. It exists in some, but not all, dialects of English. It is a feature of the nonrhotic variants (i.e. those in which the word final /r/ as in bar, star, Zanzibar, etc. is not pronounced). Intrusive-R is an

919

SANDHI insertion of [r] between two words of which the first word ends in a vowel, while the second begins with one, e.g. vanilla-r-ice-cream, the idea-r-of it, I saw-r-it, etc. An external sandhi rule that is present in some dialects of Italian is known as raddoppiamento sintattico. It is triggered only when the two words between which the sandhi operates are contained in the same local domain. The phonological process involved is the lengthening (or doubling) of the word-initial consonant. It is triggered when the word is preceded by another word ending with a stressed vowel, eg. città ppulita ‘clean city’ from città ‘city’ and pulita ‘clean’, or metà llibro ‘half book’ from metà ‘half’ and libro ‘book’. However, for consonant doubling to take place, the words have to be contained within the same phonological phrase. When this requirement is met, as in (1a), and the two words are found within the relevant domain (i.e. the phonological phrase here marked by the subscript Φ), raddoppiamento sintattico takes place. On the other hand, if the two words are separated by a phonological boundary (marked as //), raddoppiamento sintattico fails. This is illustrated by (1b), where the word-initial consonant in molto ‘old’ fails to undergo lengthening. (1a) ( Avra lletò )Φ ( il libro )Φ. ‘He will have read the book.’ (1b) (Visita la città )Φ // ( molto vechie )Φ. ‘He visits very old cities.’ An important feature of external sandhi is that an application of the phonological process in question is

determined by the phrasal groupings of the words involved, as shown by the data in (1a) and (1b). Since the phrasal organization of the sentence is determined by syntax, the conditions under which external sandhi rules apply are to some extent influenced by this component of the grammar. However, this influence only goes some way. The constituents into which words are syntactically organized such as syntactic phrases are not isomorphic to the groupings of words into phonological constituents such as phonological phrases, although there are certain connections between the two. The relationship of syntax and phonology (the syntax–phonology interface) and the field of prosodic phonology in general is a large and fertile field of research (Nespor and Vogel 1986; Selkirk 1984). References Andersen, Henning (ed.) 1986. Sandhi phenomena in the languages of Europe. Berlin, New York and Amsterdam: Mouton de Gruyter. Chen, Matthew Y. 2000. Tone Sandhi: patterns across Chinese dialects. Cambridge: Cambridge University Press. Inkelas, Sharon, and Draga Zec. The phonology syntax connection. Chicago and London: University of Chicago Press. Nespor, Marina, and Irene Vogel. 1982. ’Prosodic domains of external Sandhi rules’ The structure of phonological representations, Part I, ed. by Harry van der Hulst and Norval Smith, 225–55. Dordrecht: Foris Publications. ———. 1986. Prosodic phonology. Dordrecht: Foris Publications. Selkirk, Elisabeth O. 1984. Phonology and syntax: the relation between sound and structure. Cambridge, MA: MIT Press.

AMELA CAMDZIC

Sanskrit Sanskrit belongs to the Indo-Iranian branch of IndoEuropean, whose subgroupings are Iranian and IndoAryan. Indo-Aryan includes Vedic Sanskrit, Classical Sanskrit, and the Prakrits, or ‘vulgar’ varieties, which have as their offshoots the modern Indo-Aryan languages, such as Hindi-Urdu, Bengali, Gujerati, Marathi, Panjabi, Sinhalese, and Romani. Sanskrit is one of the most anciently attested IndoEuropean languages and has played a crucial role in the reconstruction of Proto-Indo-European. Apart from some isolated forms (see below), the most ancient literary sources are the poems collected in the books of Veda (hence the name ‘Vedic Sanskrit’), which record an oral tradition, going back to the sec-

920

ond millennium BCE. Classical Sanskrit is a highly standardized language, as such used without interruption from its first written attestations, and virtually unchanged until the present; however, in the early phases dialectal variation is visible, even between Vedic and Classical Sanskrit. The former is not the direct ancestor of Sanskrit: its peculiarities include both archaisms and innovations with respect to the classical literary language, pointing toward some kind of diatopic variation. Furthermore, the Prakrits of the Middle Indian period (300 BCE–200 CE) do not derive directly from Classical Sanskrit, but display a variety of different features, some directly connectable to Vedic. Among the Prakrits, the most important as a

SANSKRIT literary language is Pa¯li, the language of the Buddhist Canon. The coexistence of Sanskrit with Prakrits is ancient, and is mostly a matter of social variation: in classical drama, socially inferior characters (including women) usually speak Prakrit. The Aryan populations reached India toward the end of the second millennium BCE, and established themselves in the upper Indus valley. Before that time, the Aryans had already built a linguistic group distinct from the Iranians, as shown by the traces left during their journey through the Middle East. In the first half of the second millennium BCE, an Aryan aristocracy dominated the Hurrian kingdom of Mitanni, located on the upper course of Tigris and Euphrates. Names of Indic deities are mentioned in a treaty with a Hittite king, and some Indic words are found as loanwords in a Hurrian text of the period, among which are some numerals, which attest a pre-Sanskrit variety (e.g. aika, ‘one’, in which the monophthongization /ai/ > /e/ had not yet taken place, cf. Sanskrit eka).

Grammatical Sketch Phonology and Script Sanskrit is written in devana¯gari script, still used for the modern Aryan varieties, which derives from an earlier script, called bra¯hmÃ. The devana¯gari is a peculiar writing system, halfway between a syllabary and an alphabet, consisting of 13 signs for vowels and 35 for consonants, plus a number of diacritics and several compound signs used for representing consonant clusters. Consonant signs not followed by any diacritic are understood as containing the consonant and the vowel /a/. The devana¯gari script is very well suited for the representation of Sanskrit phonemes, each sign corresponding to a single phoneme (except for diphthong signs) and each phoneme being represented by a sign. Thus, Table 1 can also serve as an inventory of Sanskrit phonemes. An important feature of Sanskrit phonology, especially in comparison with the other ancient IndoEuropean languages, lies in the phonologization of both voiced and voiceless aspirates, the existence of a series of retroflex (also called ‘cerebral’) stops, borrowed from the non-Indo-European substrate languages of India, and the retention of vocalic liquids, partly with length opposition. (Length opposition is only relevant for /r# / ~ /r# /; besides, /l#/ is a very rare phoneme.) From a diachronic point of view, the most significant innovations are found in the vowel system. At an early stage in Proto-Indo-Iranian, the three vowels of Proto-Indo-European, /a/, /o/, and /e/, merged in /a/ (this change involved both short and long vowels). Later, after the separation of Iranian and Indo-Aryan,

short diphthongs underwent monophthongization: /ai/ > /e/, /au/ > /o/. Hence, already in Vedic there is again a distinction, at least for long vowels, but the /e/’s and /o/’s of Sanskrit do not reflect the same vowels of Proto-Indo-European. Sanskrit belongs to the so-called satə m group within Indo-European, i.e. it is one of those languages, along with the remaining Indo-Iranian languages, Slavic, and Armenian, in which velars changed into spirants or affricates, and labiovelars changed into velars. A major role in Sanskrit phonology is played by sandhi, i.e. a set of phonological rules determined by the clustering of vowels or consonants. Morphology Sanskrit morphology is extremely rich. Inflectional classes of nouns are based on the Indo-European distinction between thematic and athematic declension, with subclasses of the latter for stems in nasal, liquid, /i/, /u/, and long vowel. Nouns have eight cases (nominative, vocative, accusative, instrumental, dative, ablative, genitive, and locative) and three numbers (singular, dual, and plural); gender distinctions include masculine, feminine, and neuter. Adjectives, which also inflect for degrees of comparison (comparative and superlative), follow the inflection of nouns for number and case, while pronouns follow special patterns of inflection. Verbs are traditionally divided into ten classes (eight in Vedic), of which the first, fourth, sixth, and tenth follow the thematic conjugation, while the remaining follow the athematic conjugation. In the thematic conjugation, the thematic vowel -a- (< PIE * -o-) is inserted between the stem and the ending; note that, contrary to the other Indo-European languages, which have-o¯ as ending of the first person singular of the present indicative, Sanskrit has reintroduced the ending -mi of the athematic declension, so that we find a cumulation of both endings: bhara¯mi, ‘I bring’, first class, thematic, built with the present stem bhar-, ending-a¯ plus ending -mi (cp. with Latin fer-o– , same meaning). Verbs distinguish three persons in the singular, dual, and plural. Tenses include the present, the imperfect, the aorist, the perfect, and the future; in addition, Vedic also had a compound pluperfect. In origin, the verb system was based on an aspectual opposition between present and perfect. The present denoted actions or processes, while the perfect denoted states. The Indo-European opposition between perfective/imperfective is continued only formally, in the opposition between aorist and present, but not semantically (see below). Verbs have a special stem for the present, the perfect, and the aorist; the differences among stems are usually based on root gradation, which opposes a reduced

921

SANSKRIT

Devana¯gar¯ alphabet Primary vowels Short

Long

Unrounded low central Unrounded high front

Initial  a  i

Diacritic  pa  pi

Initial Á  Ã

Diacritic 

pÁ  pÃ

Rounded high back



u



pu



Æ





Syllabic variant

~

r9



pr9



r9¯



pr9¯

Diphthongs Initial Diacritic

Secondary vowels Unrounded front



e



pe



ai



pai

Rounded back

=

o



po

>

au



pau

Other symbols  

W R

an· anusvÁ ra - nasalises vowel ah¢ visarga - adds voiceless breath after vowel

am¢ anunÁsika/candrabindu - nasalises vowel p virÁma - mutes vowel

Consonants Occlusives Voiceless plosives unaspirated aspirated  ka % kha < ca E cha ? ²a @ t ¢ha 8 ta 5 tha  pa  pha

Velar Palatal Retroflex Dental Labial

Voiced plosives unaspirated aspirated

ga & gha ; ja D jha 4 °a 6 °ha K da C dha ! ba F bha

Nasals G A B 3 

n˙a ña n¢a na ma

Sonorants and fricatives Palatal ' (

Sonorants Sibilants

Retroflex 7 

ya s´ a

Dental # H

ra s¢a

Other letters 

ha

la sa

va

Variant letters used in Mumbai n "

l ¢a

G

Labial

jha

c

n¢a

A selection of conjunct consonants |

ks¢a

S

jña

_

tta

Z

W'

tra

pya

O

tka

? B

t k¢ a

e

hya

A

ttva

Numerals )

*

+

/

,

-

.

0

1

2

*)

0

1

2

3

4

5

6

7

8

9

10

(Source: http://www.omniglot.com/writing/devanagari.htm)

922

SANSKRIT grade (e.g. kr #-, ‘make’) to full grade, also called gun a #ddhi (ka¯r-). (kar-), to lengthened grade, also called vr (Note that there is no one-to-one correspondence between vowel grade and a specific tense.) The perfect is usually made on a reduplicated stem (e.g. ca-kar-a, ‘I have made’). Already in Vedic Sanskrit, the aspectual value of the aorist was given up: the aorist is used in reference to recent past, and never without past reference in contrast to the present (as was the Ancient Greek aorist). The perfect added to its original stative meaning a more recent resultative meaning, which also had its temporal implication of remote past. Furthermore, the so-called imperfect, derived from the present tense, was used in narration, where it did not display any particular tendency toward expressing the imperfective aspect. In Classical Sanskrit, moods include the indicative, the optative, and the imperative for the present tense; the aorist has an indicative and a precative, and the future has a conditional. Vedic had a more developed mood system, which included an indicative, a subjunctive, an optative, and an injunctive, all made on all three stems (present, aorist, and perfect). From a typological point of view, the most interesting feature of this mood system is constituted by the injunctive. In a way, it stands outside mood oppositions, and is used for general statements. Voice includes an active (parasmaipada) and a middle (a¯ tmanepada); not all verbs have both voices, but for those that do, the middle mostly has reflexive value. Furthermore, the middle can have passive meaning, except for the present stem, which builds a special passive by adding the suffix -ya-. The Sanskrit verb has a variety of nominal forms: active and middle present, future and perfect participles, a past passive participle, from which a past active participle is also made, a future passive participle, also called gerundive, an indeclinable participle, also called gerund, and an infinitive. A peculiarity of Sanskrit is the large extent to which composition is used for building nouns. Following the Indian classification, types of compound include dvand¯h, ‘lions and va (copulative compounds, e.g. sim  hagaja ha-, ‘lion’, plus gaja-, ‘elephant’), elephants’, from sim  tatpurus a (determinative: tat-purus a ‘that - man’= ‘the man of that one’, ‘his man’), and bahuvrà hi (possessive, or exocentric: bahu-vrà hi ‘much - rice’ = ‘having much rice’). Syntax Sanskrit has free word order, but in unmarked statements displays a tendency toward SOV order; unaccented pronouns and other clitics are consistently placed after the first accented word in the sentence in which they occur (Wackernagel’s position), modifying

adjectives, genitives, and restrictive relative clauses precede their head, and adpositions mostly follow the noun. OV features are all present in Vedic, too, but Classical Sanskrit is more consistent with respect to them. Subordination includes relative clauses and a variety of adverbial clauses, with conjunctions derived from the stem of the relative pronoun. Besides, participles and gerunds also contribute to the building of hypotaxis. Participles (especially the past passive participle in -ta) and gerunds are often used without the verb ‘be’, as if they were finite verb forms. Sanskrit displays a peculiar preference for passive, rather than active construction. The would-be-subject of the active is thus expressed through an agent phrase in the instrumental: tenedam uktam, ‘he said this’, lit.: ‘this (was) said by him’ (tena, ‘by him’ + idam, ‘this’, uktam, past passive participle of vac-, ‘say’). Impersonal passives with instrumental agent phrases are also found in intransitive verbs; the past participle, however, occurs more frequently with an agent–subject phrase in the nominative, in which case it has active meaning (so sah gra¯mam gatah , ‘he went to town’, sah, ‘he’, gra¯mam, ‘to-town’, gatah , past participle of gam-, ‘go’, rather than tena gra¯mam gatam, lit.: ‘by him (tena, instr.) (it was) gone to town’).

Sanskrit and Indo-European The importance of Sanskrit for Indo-European linguistics cannot be overstated: it was only after European scholars learned about it that they realized that their own European languages presented striking similarities among each other, and set out to reconstruct a proto-language. Actually, for some time Sanskrit was even thought to be the proto-language: but also later, after it lost this status, Sanskrit was regarded as the most conservative Indo-European language, especially with respect to its numerous inflectional categories, fully developed in rich and complex paradigms. The discovery of Anatolian, at the beginning of the twentieth century, gave a serious stroke to the primacy of Sanskrit. Some scholars started to reconstruct a Proto-Indo-European with a smaller case system, with less verbal moods, and in which some basic Greek– Sanskrit isoglosses (augmented past tenses, aorist, dual number, etc.) played a less relevant role. Another challenge to the possible archaic character of Sanskrit has come from the so-called ‘ejective model’, according to which the stops of Proto-Indo-European include voiceless/glottalized/aspirate, rather than voiceless/ voiced/voiced aspirate of the traditional reconstruction (see Hopper 1973). A full discussion of new reconstruction trends can be found in Gamkrelidze and Ivanov (1995).

923

SANSKRIT References Allen, Sidney. 1962. Sandhi. The Hague: Mouton. Bloch, J. 1965. Indo-Aryan from the Vedas to modern times. Paris: Adrien-Maisonneuve. Burrow, T. 1965. The Sanskrit language. London: Faber & Faber. Cardona, George. 1990. Sanskrit. The world major languages, ed. by B. Comrie. Oxford: Oxford University Press. Gamkrelidze, Th. V., and V. V. Ivanov. 1995. Indo-European and the Indo-Europeans. Berlin, New York: Mouton de Guyter.

Gonda, Jan. 1951. Remarks on the Sanskrit passive. Leiden: Brill. Hoffman, Karl. 1967. Der Iniunktiv im Veda. Heidelberg: Winter. Hopper, Paul. 1973. Glottalized and murmured occlusives in Indo-European. Glossa 7(2). 141–66. McDonell, A.A. 1916. A Vedic grammar for students. Oxford: Oxford University Press. McDonell, Arthur A. 1926. A Sanskrit grammar for students. Oxford: Oxford University Press. Speijer, J. S. 1980. Sanskrit syntax. Delhi: Motilal Banarsidass (reprinted from 1886 edition).

SILVIA LURAGHI

Sapir, Edward As a senior in the undergraduate program in German language and literature at Columbia University in 1903–1904, Edward Sapir enrolled in a graduate seminar in American languages taught by the distinguished anthropologist Franz Boas. Sapir had already studied Latin, Greek, French, Spanish, and his major subject German, and now he was exposed to the diversity of the native languages of North America. The effect was not immediate, for Sapir continued on to the master of arts degree, still concentrating on Germanic philology and literature. But his 1905 master’s thesis, on traditional German linguistics, did include data from several American languages to demonstrate the grammatical complexity of all languages, those without writing systems as well as those with long established literary traditions. Soon Sapir was engaged in fieldwork, traveling to the West Coast of the United States to study Wishram in Washington in the summer of 1905, and Takelma in Oregon in 1906. Although his Germanic studies had been largely philological and historical, his grammar of Takelma was entirely a synchronic analysis, describing the language without recourse to historical information, a characteristic of most of the linguistic work done under the direction of his mentor, Franz Boas. Sapir published his master’s thesis, in the journal Modern Philology, but virtually all of his other early publications were on native American languages. His doctorate was awarded in 1909, with the Takelma grammar serving as his dissertation, and by the end of that year he had published nine articles, reviews, and monographs on the Kwakiutl, Upper Chinook, Yana, Yuchi, and Wishram languages, as well as Takelma. His reputation as a specialist in native American languages was established early, and continued to grow throughout his lifetime, reaching nearly mythic pro-

924

portions as some admirers attributed to him extensive knowledge, and even fluent command, of languages which he had investigated for just a few days in consultation with a single speaker. After receiving the doctorate, he took up the post as chief of the Division of Anthropology in the Geological Survey of the Canadian National Museum. This was both a hindrance and an asset to his career as a linguist. On the one hand, the position isolated him from his colleagues in the United States and at times directed his energies toward administration. But the advantages were many, not least of which was control of a research staff and a budget, always too small, with the authority to determine and direct fieldwork. It was Sapir’s intent to create a systematic survey of the languages and cultures of all of Canada. Since he was nearly always the only staff member with formal training in linguistics, it was he who did most of the linguistic work, concentrating his attention on the languages of Western Canada. He made many field trips in the years 1910–1915, and thereafter worked frequently with speakers at his office in Ottawa. The use of trained native assistants for the recording of text was a common practice with Boas and his students, but Sapir was exceptional in recognizing that native speakers’ insights were relevant to the development of linguistic theory, especially in the area of phonology. In the early years of Sapir’s career, linguists were struggling with the new concept of the phoneme, which Sapir later came to define as ‘a functionally significant unit in the rigidly defined pattern or configuration of sounds peculiar to a language . . ., as distinct from . . .the ‘sound’ or ‘phonetic element’ as such (an objectively definable entity in the articulated and perceived totality of speech)’. (Sapir, ‘The psychological reality of phonemes’ 1985[1933]: 46)

SAPIR, EDWARD One of Sapir’s most important and long-lasting contributions to the development of the phoneme concept was his view that the phonemic structure of a language was part of the unconscious knowledge that speakers of a language possessed, part of their ‘phonological intuitions.’ Drawing on his years of work with native interpreters of American Indian languages, he argued that ordinary speakers hear phonemes, not phonetic details, and he offered evidence in the way that interpreters would transcribe their language, always recording distinctive contrasts in sounds while not noting phonetic differences that were irrelevant to their phonological system. The paper in which these concepts were most clearly expressed was written in French and published in Paris, France in 1933. At that time, it was little noted by American linguists, but when it appeared in an English version in 1949, ten years after Sapir’s death, it sparked something of a revival of interest in Sapir’s work. His notions of the psychological reality of the phoneme, of native speaker intuition, and of unconscious linguistic knowledge were all to find their way into the development of generative phonology in the late 1950s and 1960s, as did his description of phonological systems as ‘sound patterns’, explicated in his widely read 1925 article ‘Sound patterns in language’ published in the first issue of Language, journal of the Linguistic Society of America. Also growing out of Sapir’s work in Canada was his genetic classification system for American Indian languages. Drawing on his background in Germanic historical linguistics, Sapir began to apply the techniques of the Indo-European comparative method to the original languages of America. Beginning with the 1891 classification system developed by John Wesley Powell and the Bureau of American Ethnology, Sapir took Powell’s 55 ‘stocks’ (or language families) and eventually reduced them to six groups: Eskimo-Aleut, Algonquian-Ritwan, Na-Dene, Penutian, HokanSiouan, and Aztec-Tanoan. There was a good deal of controversy at the time over this plan, in part because Sapir provided little evidence to support it. Sapir himself considered the scheme to be probable but not at all proven, given that ‘scientific comparative work on these difficult languages is still in its infancy’ (‘Central and North American languages’, 1985[1929]: 169), but his system remains the fundamental reference point for those working on the genetic classification of American Indian languages. Sapir’s most widely read work is the small book Language, published in 1921 while he was still in Canada. Aimed at both students of linguistics and the public, the volume predated his work on phonology. Its chapters focus on words and grammar, with a particular emphasis on what is today called linguistic typology, the classification of languages and of features of

languages into types based on the structural properties of their units (as opposed to genetic classification based on shared history). Sapir’s classification system drew heavily on his background in Germanic linguistics and can be viewed as the twentieth-century apex of nineteenth-century European typology. The system provides categories for the morphology (the processes of word formation) of languages along two dimensions: the extent to which words employ affixes and the ease with which the root and affixes of a word can be segmented. For example, an isolating language would be one in which most words consist solely of roots, with little affixing. Sapir gives Chinese as an example. A language with many affixes intertwined and difficult to segment would be labeled fusional; Sapir cites Salinan, a language of southwest California. This part of Sapir’s typology, based as it was on the well–known categories of earlier linguists, has been modified somewhat in terminology, but it remains useful in the classification of morphological types. Sapir, however, also attempted to overlay this system with another, a conceptual classification that distinguished concrete concepts and relational concepts, with further bifurcation into simple and complex. This aspect of his typological classification has not been pursued, but Sapir himself recognized the preliminary nature of his system and the need to explore a greater number of more diverse languages than he was able to consider at the time. This is exactly what modern typologists have done. Sapir’s interest in conceptual categories was manifested in another area of research in which he engaged in the later 1920s and early 1930s––the issue of an international auxiliary language. Beginning in 1925, at about the time he accepted a position at the University of Chicago, Sapir became involved with the International Auxiliary Language Association, Inc. (IALA), headquartered in New York City. Working in cooperation with the co-founder and honorary secretary of the association, Alice Vanderbilt Morris, Sapir wrote a number of scholarly and popular pieces supporting research on and development of an international auxiliary language. He was not alone among linguists in this interest, which was shared most enthusiastically by the Danish linguist Otto Jespersen, a number of other European linguists, as well as by many of Sapir’s American colleagues. Sapir served briefly as IALA’s linguistic research director and under IALA sponsorship produced three studies of conceptual structure, which comprise Sapir’s total output on semantics. He made a number of programmatic statements but did no further work in this area and has had no lasting influence. Linguists’ interest in an international auxiliary language dissipated during the 1930s, a period of professionalization within the discipline and of increasing American isolationism on the national scene.

925

SAPIR, EDWARD In 1931, Sapir moved from the University of Chicago to Yale University. He never again conducted fieldwork, and he largely withdrew from linguistic study to concentrate on culture and personality. But he did supervise dissertations and postdoctoral work on American languages by a number of Yale students who went on to distinguished careers in linguistics, including Mary R. Haas, Stanley Newman, Morris Swadesh, and Charles F. Voegelin. At Yale, Sapir also came to know Benjamin Lee Whorf, a largely self-educated man interested primarily in philosophical issues and the study of meaning, who associated himself with Sapir and his circle of graduate students. Whorf’s linguistic relativity hypothesis (that grammatical structure of a language influences its users’ perceptions of the world) became well known beyond the field of linguistics. It is sometimes referred to, incorrectly, as ‘the Sapir–Whorf hypothesis’, but it was developed fully only after Sapir’s death in 1939. Most contemporary linguists found it too speculative, and the hypothesis never became part of mainstream linguistics. Edward Sapir was a key figure of American linguistics in the first half of the twentieth century, equaled in stature for that period only by his contemporary Leonard Bloomfield. Sapir’s students protected and promoted his work and ideas and he remains ‘a symbol of scope and insight in the study of language’ (Dell Hymes 1985: 598).

Biography Edward sapir was born on January 26, 1884 in Lauenberg, Germany, now part of Poland, the son of Lithuanian parents who soon moved to England and then on to the United States. He spent his childhood in Richmond, Virginia up to the age of ten, and then in New York City. He began the study of German, French, Spanish, Latin, and Greek at Peter Stuyvesant High School, matriculating with a German major at Columbia University in 1901. He received his B.A. in 1904; M.A. (1905) in Germanic philology and literature; and Ph.D. (1909) in anthropology, specializing in linguistics with a dissertation on the Takelma language of Oregon supervised by Franz Boas. He was Research Associate, Department of Anthropology, University of California, Berkeley, 1907–1908, and Fellow and then instructor at the University of Pennsylvania, 1908–1910. He was Chief, Division of Anthropology, Geological Survey of the Canadian National Museum, Ottawa, 1910–1925. He went to the University of Chicago in 1925, and was promoted to Professor of Anthropology and General Linguistics in 1927. Concurrently, he was also Director of Linguistic Research, International Auxiliary Language Association, New York City, from October 1930 to July 1, 1931. He went to Yale University as

926

Sterling Professor of Anthropology and Linguistics in 1931. He was President, Linguistic Society of America (LSA) in 1933; he also taught Introduction to Linguistic Science and Field Methods at the 1937 summer LSA Linguistic Institute at University of Michigan. Sapir died in New Haven, Connecticut on February 4, 1939. References Anderson, Stephen R. 1985. Phonology in the twentieth century: theories of rules and theories of representations. Chicago and London: University of Chicago Press. Campbell, Lyle. 1997. American Indian languages: the historical linguistics of native America. New York and Oxford: Oxford University Press. Cowan, William, Michael K. Foster, and Konrad Koerner (eds.) 1986. New perspectives in language, culture, and personality: proceedings of the Edward Sapir centenary conference, Ottawa, 1–3 October 1984. Amsterdam and Philadelphia: John Benjamins. Darnell, Regna. 1990. Edward Sapir: linguist, anthropologist, humanist. Berkeley and Los Angeles: University of California Press. Falk, Julia S. 1995. Words without grammar: linguists and the international auxiliary language movement in the United States. Language & Communication 15(3). 241–59. Hymes, Dell. 1985. Epilogue. Selected writings of Edward Sapir in language, culture and personality, ed. by David G. Mandelbaum, paperback edition, 598–600. Berkeley and Los Angeles: University of California Press. Koerner, Konrad (ed.) 1984. Edward Sapir: appraisals of his life and work. Amsterdam and Philadelphia: John Benjamins. Lee, Penny.1996. The Whorf theory complex: a critical reconstruction. Amsterdam and Philadelphia: John Benjamins. Sapir, Edward. 1989. The collected works of Edward Sapir, 16 vols. Berlin and New York: Mouton de Gruyter. Sapir, Edward. 1921. Language: an introduction to the study of speech. New York: Harcourt, Brace. Whaley, Lindsay J. 1997. Introduction to typology: the unity and diversity of language. Thousand Oaks, CA: Sage Publications. All the following works are reprinted in Selected writings of Edward Sapir in language, culture and personality, ed. by David G. Mandelbaum. Berkeley and Los Angeles: University of California Press, 1949. Page references in the text are to the paperback edition of 1985. Sapir, Edward. 1925. Sound patterns in language. Language 1. 37–51. ———. 1929. Central and North American languages. Encyclopaedia Britannica, 14th edition, Vol. 5, 138–41. London and New York: Encyclopaedia Britannica Co. ———. 1929. The status of linguistics as a science. Language 5. 207–14. ———. 1931. Function of an international auxiliary language. Psyche 11. 4–15. ———. 1933. Language. Encyclopaedia of the social sciences, Vol. 9, 155–69. New York: Macmillan. ———. 1933. La Réalité Psychologique des Phonèmes. Journal de Psychologie Normale et Pathologique 30. 247–65. ———. 1944. Grading: a study in semantics. Philosophy of Science 11. 93–116.

JULIA S. FALK See also Boas, Franz; Sapir–Whorf Hypothesis; Whorf, Benjamin Lee

SAPIR–WHORF HYPOTHESIS

Sapir–Whorf Hypothesis The idea that one’s native language ‘colors’ his Weltanschauung has been around at least since the days of Johann Gottfried Herder (1744–1803) and Wilhelm von Humboldt (1762–1835). Charles Sanders Peirce (1839–1914) postulates that man’s symbolic universe could only make sense via language. Ferdinand de Saussure (1857–1913) (Cours de linguistique générale [1916:155]) states that: ‘No ideas are established in advance, and nothing is distinct, before the introduction of linguistic structure.’ However, the principle of linguistic relativity has largely become associated with Benjamin Lee Whorf (1897–1941), who along with Edward Sapir (1884–1939), his linguistic mentor at Yale University, used linguistics to advocate their position that language influences the way in which a speech community perceives and conceives of its reality. Part of the groundwork for the Sapir–Whorf hypothesis was laid by Whorf’s work as a fire insurance investigator. During his career, he had the opportunity to analyze many reports as to why fires broke out. He found that workers would use extreme caution when around ‘full’ drums of gasoline. Just as one might expect, workers were careful not to smoke around ‘full’ drums. Yet, these same workers, when around ‘empty’ drums of gasoline, would sometimes toss lit cigarettes nearby. This caused a violent explosion because an empty drum still contained volatile gasoline vapor. Thus, an ‘empty’ drum was really much more of a threat than a ‘full’ one. Using these data, Whorf concluded that the meanings of certain words had an effect on a person’s behavior. It was the extensive research of both Sapir and Whorf into the grammatical systems of American Indian languages, however, that proved to have the greatest impact on the hypothesis. By predicating their insights into the interrelationships of language and culture on what they had learned from the structures of these languages, the basic idea of language shaping the perceptions of its speakers and providing for them a vehicle so that their experiences and emotions could be placed into significant cognitive categories was given its scientific underpinnings. Generally, Sapir is credited as giving the problem of establishing the link between language and culture its initial formulation, while Whorf is honored as the one who took this idea and developed it further to include grammar in addition to lexis, thereby making it a bona fide hypothesis. Hence, the resultant supposition is commonly given

the designation the ‘Whorfian hypothesis’. Some, pointing to Sapir’s preeminent stature as a linguist, prefer the appellation the ‘Sapir-Whorf hypothesis’. When viewed in terms of output, however, one could counter that a more appropriate label would be the ‘Whorf–Sapir hypothesis’. A rather interesting development in this debate over giving credit where credit is due has been the attempt to disassociate Sapir from the hypothesis entirely. Desirous of preventing the image of the great maestro Sapir from being tarnished by the taint of controversy, some, most notably Alfred L. Kroeber, have claimed that Sapir’s views were not really that (pro)Whorfian. This viewpoint is not borne out by an examination of Sapir’s own writings. For example, as one can see in a (1929) passage (from the journal Language, p. 209), he equated language and thinking in terms of the speech community’s overall culture: Language is a guide to ‘social reality’…it powerfully conditions all our thinking about social problems and processes. Human beings do not live in the objective world alone, nor alone in the world of social activity as ordinarily understood, but are very much at the mercy of the particular language which has become the medium of expression for their society.

There are actually two different versions of the hypothesis. This is understandable when one considers that Whorf did all of his writing from 1925 to 1941, and his ideas were continuously evolving. The strong version of the hypothesis, which is called linguistic determinism, holds that language determines thinking. This position is most difficult to defend primarily because translation between one language and another is possible, and ‘thinking’ can take place without language; e.g. an artist or sculptor can and often does think with his fingers. Mirroring Sapir’s thoughts, Whorf notes in his (1940) ‘Science and linguistics’: We dissect nature along lines laid down by our native languages…We cut nature up, organize it into concepts, and ascribe significances as we do, largely because we are parties to an agreement to organize it in this way –– an agreement that holds throughout our speech community and is codified in the patterns of our language. (Carroll 1956:213).

The milder version of the hypothesis is labeled linguistic relativity, coined by Whorf, since he always tried

927

SAPIR–WHORF HYPOTHESIS to qualify his assertions. This claims that the native language influences one’s thoughts and perceptions. Linguistic relativity can be illustrated using one of Whorf’s favorite sources of data, (Uto-Aztecan) Hopi. With the exception of birds, there is, according to Whorf, only one word in Hopi for everything that flies. One also notes that Bedouin Arabic dialects have more than 6,000 different lexemes for various types of camels and camel paraphernalia. Perhaps the most incontrovertible piece of evidence in its favor comes from the realm of numbers. Some languages (e.g. Hottentot [Nama]) only have words for the numerals ‘one’ and ‘two’ and a word roughly translatable as ‘many’ for ‘three or more’ (and a few languages have no numbers at all). Thus, such concepts as googolplex are beyond everyday verbal expression in these languages. Mathematics (as we know it) is thus not possible, i.e. people are indeed ‘at the mercy of’ their native tongues. In ‘An American Indian Model of the Universe’ (International Journal of American Linguistics [1950]), Whorf argues that since there is neither an explicit nor an implicit reference to time in Hopi and thus no tenses for its verbs, time and space cannot be the same concepts for Hopis as they are for English speakers. The Hopi Weltanschauung is different from that of an SAE (Standard Average European, a Whorfian term) speaker. Many of his other essays present data for his basic contention that Hopi metaphysics, which underlies its cognition, is different from that of an English speaker; i.e. the Hopis calibrate the world differently because their language defines experience differently for them. As more information has surfaced on Hopi, some of Whorf’s specific grammatical points have not held up well. This explains why most linguists today believe the Sapir–Whorf hypothesis to be invalid. However, the

truth of the matter is that while no one has proved it wrong, neither has anyone proved it right. Whichever end of the continuum one considers in the relationship between language and culture, it is important to realize the interpenetration of the two. In areas like bilingualism, is it really possible to learn a foreign language without also simultaneously learning the Weltanschauung of its speakers? Not only is language part of culture, it allows for its acquisition. Sapir maintains that ‘the worlds in which different societies live are distinct worlds, not merely the same world with different labels attached’ (Mandelbaum 1949:162). Whorf took that one step further, asserting that grammar not only allows for the voicing of ideas but it also ‘is the shaper of ideas’ (Carroll 1956:221). References Carroll, John B. (ed.) 1956. Language, thought, and reality: selected writings of Benjamin Lee Whorf. Cambridge, MA: MIT Press. Darnell, Regna. 1990. Edward Sapir: linguist, anthropologist, humanist. Berkeley: University of California Press. Gumperz, John J., and Stephen C. Levinson (eds.) 1997. Rethinking linguistic relativity. Cambridge: Cambridge University Press. Hoijer, Harry (ed.) 1954. Language in culture. Chicago, IL: University of Chicago Press. Koerner, Konrad (ed.) 1984. Edward Sapir: appraisals of his life and work. Amsterdam and Philadelphia: John Benjamins. Lee, Penny. 1996. The Whorf theory complex. Amsterdam and Philadelphia: John Benjamins. Mandelbaum, David G. (ed.) 1949. Selected writings of Edward Sapir in language, culture, and personality. Berkeley: University of California Press.

ALAN S. KAYE See also Humboldt, Wilhelm von; Peirce, Charles; Sapir, Edward; Saussure, Ferdinand de; Whorf, Benjamin Lee

Saramaccan Saramaccan (or Saamáka) is the name of the creole language spoken by the Saramaka people, who live along the Suriname River in central Suriname (in the northeast of South America). The names ‘Saramaka’ and ‘Saamáka’ derive from the fact that the first settlements of these people were located along the Saramacca River, also in central Suriname. The c. 50,000 Saramaccans living today are the descendants of African slaves who escaped from the plantations to create their own communities in the Suriname rain forest in the seventeenth 928

and eighteenth centuries. These runaway slaves and their descendants are often referred to as ‘Maroons’, a word derived from Spanish cimarron, meaning ‘stray animal’. The creole languages spoken by Maroon communities may be referred to as ‘maroon creoles’, to distinguish them from (former) ‘plantation creoles’, such as Sranan. There are two Maroon creoles (or rather: groups of Maroon creoles) in Suriname: Saramaccan (including three different dialects) and Ndyuka (used here as a cover term for four different dialects). Maroon

SARAMACCAN creoles outside of Suriname that are still spoken today include Palenquero (Colombia) and Angolar (São Tomé, an island off the coast of West Central Africa). The main reason for distinguishing Maroon creoles as a separate category is the fact that, due to their relative isolation from outside influence, they are assumed to be more ‘radical’ than (former) plantation creoles, which have remained in contact with the language from which they have derived their lexicon (the ‘lexifier language’). The term ‘radicality’ refers to the typological distance between a creole and its lexifier language. Although until now very little comparative research regarding the degree of radicality of different creoles has been done, it seems clear that the typological distance between, say, Saramaccan and its (main) lexifier, English, is larger than that between, say, Cape Verdean Creole and Portuguese. Therefore, Maroon creoles are assumed to be closer to creoles at the time of their formation (say, 300 years ago) than (former) plantation creoles. Among the Maroon creoles, Saramaccan has acquired a special status, especially in research associated with Bioprogram Theory (see e.g. Bickerton 1984), to the extent that it has come to be considered as the most radical creole. According to this view, Saramaccan provides a rare opportunity to get a closer look at the process of creolization. While there are still important gaps regarding the history of the Saramaka and their language, the following seems to be clear (Smith 2002). The origins of the Saramaka people and their language go back to escapes from plantations that took place before 1700, at a time when the formation of a plantation creole must have been well on its way. Assuming that most of the runaway slaves had been on the plantations for some time before making their escape, they took at least some knowledge of this evolving plantation creole with them. This explains the structural similarities between Saramaccan and Sranan, both of which are descendants of the earlier plantation creole. In spite of these similarities, however, there are also a number of differences between the two languages. One of these is the proportion of Portuguese-derived words, which is much larger in Saramaccan than it is in Sranan. In the former, one third of the basic vocabulary (basic words such as ‘water’, ‘go’, ‘big’) is derived from Portuguese, while this is much smaller in Sranan. The remainder of the basic vocabulary is largely derived from English (Suriname was an English colony before it became Dutch), while there are also a few items from West African languages. The presence of these Portuguese-based words is explained by the fact that, as has been established by historical research (e.g. Price 1976, 1983), many of the first Saramaka came from plantations owned by Sephardic Jews, who were Portuguese- (and to some extent Spanish-) speaking. Although the presence of many Portuguese-derived

words could lead one to view Saramaccan as a creole with two lexifier languages—English and Portuguese–– the fact that most function words are from English suggests that the Portuguese element was added later. For this reason, Saramaccan is generally categorized as an English-lexicon creole, albeit one with a strong Portuguese element. A second difference between Saramaccan and Sranan is related to the fact that the former has a higher percentage of words derived from African languages. This is probably due to the fact that there was much less contact with other languages in the case of Saramaccan than in the case of Sranan. As to the origins of these words, three African languages (or rather: language clusters) have been especially important: Kikongo (a Bantu language), Gbe, and Akan (both of them Kwa languages; both Bantu and Kwa belong to the family of Niger-Congo languages). Although nothing is known about the specific African origins of the individual runaway slaves who formed the ‘founder population’ of the Saramaka people, we do have reliable information about the origins of the African slaves in general, who were brought to Suriname in the 1675–1700 period (Arends 1995a:243). In this period, roughly speaking, half of all Suriname slaves came from an area where Bantu languages such as Kikongo, were spoken, while the other half came from an area where Kwa languages, such as Gbe and Akan, were spoken. The connection between the ethnolinguistic origins of the Suriname slaves and the traces that were left in the Suriname creoles by their native languages receives further support from the fact that Saramaccan exhibits some rather marked phonological features, such as lexical tone and nasal and complex stops, which are characteristic of these three languages (Bruyn 2002). Like its sister language Sranan, Saramaccan is well documented in its early stages. In the case of Saramaccan, however, the early documentation is limited to a very short period, roughly 1780–1820. This has to do with the fact that the Moravian Brethren, to whom we owe these early writings, stopped their missionary activities among the Saramaka around 1820. These early documents, which together number well over 2,000 manuscript pages, consist mainly of religious texts, such as Bible translations, although some linguistic descriptive works, such as dictionaries, are included as well (see Arends (1995b) for further information). Unfortunately, however, up to now only a few of these documents have been made available for linguistic research (Arends and Perl 1995). As to the major structural features of Saramaccan, many of these are also found in Sranan. Here, we will only mention and illustrate some of the features in which the two languages differ. (This section draws heavily on Bruyn 2002; see also Bakker et al. (1995.) 929

SARAMACCAN Lexicon. Some examples of Portuguese-derived basic vocabulary items are búka (