Contact Languages: A Wider Perspective (Creole Language Library)

Citation preview

CONTACT LANGUAGES

CREOLE LANGUAGE LIBRARY (CLL) A companion series to the "JOURNAL OF PIDGIN & CREOLE LANGUAGES"

Editors Pieter Muysken (Amsterdam) John Victor Singler (New York)

Editorial Advisory Board Mervyn Alleyne (Kingston, Jamaica) Norbert Boretzky (Bochum) Lawrence Carrington (Trinidad) Chris Corne (Auckland) Glenn Gilbert (Carbondale, Illinois) John Holm (New York)

George Huttar (Dallas) Salikoko Mufwene (Chicago) Peter Mühlhäusler (Adelaide) Pieter Seuren (Nijmegen) Norval Smith (Amsterdam)

Volume 17

Sarah G. Thomason (ed.) Contact Languages: A wider perspective

CONTACT LANGUAGES A WIDER PERSPECTIVE

Edited by

SARAH G. THOMASON

JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences — Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data Contact languages : a wider perspective / edited by Sarah G. Thomason. p. cm. -- (Creole language library, ISSN 0920-9026 ; v. 17) Includes bibliographical references and indexes. 1. Languages, Mixed. I. Thomason, Sarah Grey. II. Series. PM7802.C66 1996 417'.22-dc20 ISBN 90 272 5239 4 (Eur.) / 1-55619-172-3 (US) (alk. paper)

96-43552 CIP

© Copyright 1997 - John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. • P.O.Box 75577 • 1070 AN Amsterdam • The Netherlands John Benjamins North America • P.O.Box 27519 • Philadelphia PA 19118-0519 • USA

Table of Contents Acknowledgments Contributors List of maps Introduction Hiri Motu Tom Dutton Pidgin Delaware Ives Goddard Ndyuka-Trio Pidgin George L. Huttar & Frank J. Velantie Arabic-based Pidgins and Creoles Jonathan Owens Kitúba Salikoko S. Mußvene Sango Helma Pasch Prior Pidginization and Creolization in Swahili? Derek Nurse Michif: A Mixed Language Based on Cree and French Peter Bakker & Robert A. Papen Media Lengua Pieter Muysken Callahuaya Pieter Muysken Mednyj Aleut Sarah G. Thomason Ma'a (Mbugu) Sarah G. Thomason Language Index Names Index Subject Index

vii ix xi 1 9 43 99 125 173 209 271 295 365 427 449 469 489 495 503

Acknowledgments The contributors to this volume have been so patient, during years of delay in the editing process, that I owe them an unusually emphatic expres­ sion of gratitude — not only for their lack of overt impatience, but for their amiable reaction to repeated editorial queries and requests for revision. The series editors, Pieter Muysken and John Singler, have provided valuable advice and helpful nudges throughout this period. My most enthusiastic thanks got to Anna Fenyvesi, whose extreme diligence and word-processing skill produced the final manuscript and whose moral support heartened me throughout the last stages of the process.

Contributors Peter Bakker Institute for General Linguistics University of Amsterdam Spuistraat 210 1012 VT Amsterdam The Netherlands Tom Dutton Department of Linguistics Research School of Pacific Studies The Australian National University P.O.Box 4 Canberra Australia Ives Goddard NHB Rm. 85, MRC 100 Smithsonian Institution Washington, D. C. 20560 USA George L. Huttar International Linguistics Center 7500 W. Camp Wisdom Rd. Dallas, TX 75236 USA Salikoko Mufwene Department of Linguistics University of Chicago 1010 East 59th Street Chicago, IL 60637

x

Pieter Muysken Instituut voor Algemene Taalwetenschap Universiteit can Amsterdam Spuistraat 210 1012 VT Amsterdam The Netherlands Derek Nurse Department of Linguistics Memorial University of Newfoundland St. John's Newfoundland Canada A1C 3X9 Jonathan Owens Universität Bayreuth Sprach- und Literaturwissenschaftliche Fakultät Arabistik Postfach 101251 D-8580 Bayreuth Germany Robert A. Papen Département de Linguistique Université du Québec à Montréal . .8888 Succursale A Montréal, Québec H3C 3P8 Canada Helma Pasch Institut für Afrikanistik der Universität zu Köln Meister-Ekkehart-Strasse 7 5000-Köln 14 Germany

Contributors

Contributors Sarah G. Thomason Department of Linguistics University of Pittsburgh Pittsburgh, PA 15260 USA Frank J. Velantie Afd. Alfabetisering (Volwassenen Edukatie) MINOV Commewijnestraat 31 Paramaribo Suriname

XI

List of Maps Map of Hiri Motu area Approximate current residential areas of Trios, Wayanas, and Bushnegro groups of Suriname Map of the Sudan Languages of the Southern Sudan and adjoining regions Position of the Kongo Kingdom in the 17th-18th centuries Map of Sango area Map of Media Lengua area

10 100 128 137 174 210 370

Introduction Sarah G. Thomason University of Pittsburgh

The best-known contact languages are pidgins and creoles with European lexicons, scattered along the major routes that were followed by European powers engaged in trade and colonization, starting in the Age of Exploration. These languages arose as a direct result of contact with Europeans, as re­ flected in the fact that their vocabularies are drawn primarily from the European languages of visiting traders, colonizers, or resident slavemasters. But it is becoming increasingly obvious that languages such as the Caribbean and Indian Ocean creoles, the various varieties of Pidgin English, and the several Portuguese-based creoles in South Asia and elsewhere owe their fame (and their prominence in the linguistic literature) to their links with European history rather than to any special linguistic status among contact languages. Pidgins and creoles with non-European lexicons are now being studied in many places around the world, and a third type of contact language — bilingual mixtures that (unlike pidgins and creoles) must have been created by bilinguals — has been reported reliably from widely separated locations. The purpose of this book is to present linguistic and historical sketches of some lesser-known contact languages, in an effort to provide some balance in the worldwide picture of these most dramatic results of language contact. An added goal is to facilitate the investigation of linguistic and historical similari­ ties and differences among contact languages of all three major types. The twelve case studies collected here range over all three types of contact languages, and several authors also deal explicitly with the problem of finding the boundary between contact language and dialect of the lexifier language. Three chapters focus on particular pidgins (Hiri Motu, Pidgin Delaware, and Ndyuka-Trio Pidgin), two on creoles (Kitúba, Sango), one on a

2

Sarah G. Thomason

specific set of pidgins and creoles (Arabic-based languages), one on the question of early pidginization and/or creolization in Swahili, and five on bilingual mixtures (Michif, Media Lengua, Callahuaya, Mednyj Aleut, and Ma'a). The chapters are arranged according to the type of language: pidgins first, then the chapter covering both pidgins and creoles, then creoles, then the question-mark chapter on Swahili, and finally the two-language mixtures. Geographically, the Americas and Africa are best represented (five chapters each); there is one chapter each from the Pacific and northern Eurasia. This geographical skewing was not a deliberate choice. To some extent it resulted from a historical accident — several scholars who were asked to contribute were unable to do so in the end — but the main reason for the imbalance is that more contact languages that meet the basic criterion for inclusion are known in Africa and the Americas than in other parts of the world. This picture may change dramatically soon, as more information becomes available about contact languages elsewhere, notably among nonAustronesian languages of New Guinea (Foley 1986:30-31, Williams 1993). All the authors were asked to address a series of questions designed to make the case studies easy to compare with each other and with descriptions of other contact languages. The first set of questions concerned the circum­ stances of the language's use: location, number of speakers, domains of usage, other languages spoken by the users. The next set focused on demon­ strating the "languageness" of the language under consideration: is it crystal­ lized, so that it must be learned as a language rather than simply being produced by ad-hoc simplification (or other distortion) from any given speaker's native language? And is it mutually intelligible with any other language, specifically the main lexical-source language? The third set of questions was historical. In discussing the historical circumstances under which the contact language arose, authors were asked to distinguish carefully between documented social conditions and inferences drawn from inevitably incomplete documentation, and to lay out the bases for inferences drawn from linguistic facts. (For instance, if the language is a creole, is there direct evidence of a fully crystallized pidgin stage, or could creolization have been abrupt in a new contact situation?) Finally, authors were asked to describe the language briefly and to consider which of the three obvious candidates for sources of contact-language lexicon and grammar — the lexifier language(s), other language(s) spoken by the creators of the contact language, and univer­ sal structural tendencies — might be responsible, alone or in combination, for particular features found in the language.

Introduction

3

These questions could not be answered fully for all the languages, because of limitations in the available information or, as in the Swahili chapter, a scope that included non-contact varieties as well as contact lan­ guages. The chapters also differ in depth of coverage. In some cases the contact languages, and often their lexical and grammatical source languages as well, are still too little known to permit a detailed analysis; in other cases, such as Hiri Motu, more detailed discussions of certain topics are available elsewhere, so that readers are referred to other sources for details. The authors approach their topics in different ways as a result of the kinds of information that are available. So, for instance, Goddard's discussion of Pidgin Delaware is based on philological analysis of the documentation on this long-dead language, while Dutton and Muysken, to name just two of several authors, draw on their own fieldwork data in describing Hiri Motu and Media Lengua. An introduction to a book about contact languages must define the object of study: what is a contact language? The book's title presupposes that contact languages are a well-defined linguistic phenomenon, but — as with so much else in language-contact studies — the point is controversial. Probably no answer will satisfy everyone, but the following definition is at least consistent with the usage of all the contributors to this volume: a contact language is a language that arises as a direct result of language contact and that comprises linguistic material which cannot be traced back primarily to a single source language.1 Because the historical linguist's technical concept of genetic rela­ tionship requires that members of a language family descend primarily, as whole systems, from a single parent language, contact languages do not belong to any language family: by definition, their genesis was not a matter of descent with modification from a single parent. This definition is thus funda­ mentally historical; it is based on diversity in the sources of the linguistic structures rather than on (say) typological characteristics of the language. The reason for insisting on a historical definition is that synchronic definitions don't work; there is, for instance, no such thing as a master list of linguistic features that are universally shared by and exclusive to contact languages, or even pidgins and creoles as a set (leaving bilingual mixtures aside). There are, in my view, just three types of contact languages — namely, the types represented in this book. There are prototypical contact languages in all three categories, and there are also various kinds and degrees of deviations from the prototype. All these deviations (e.g. "semi-creoles" and "koinés") are best analyzed in relation to the prototype, as on a continuum, and not as

4

Sarah G. Thomason

separate types of contact languages. Treating pidgins, creoles, and bilingual mixtures as discrete language types of course raises problems of identifica­ tion. There are many borderline cases: languages that have some pidgin characteristics but are analyzable as simplified versions of their lexifiers; speech forms that have some systematicity but are not clearly stable enough to be classed as languages; and so forth. But there are also many clear cases of contact languages. And since wrestling with fuzzy boundaries is a standard part of the historical linguist's job — consider, for instance, the impossibility of drawing a sharp dividing line between dialects of the same language and separate languages, during a process of language split — the discovery that fuzzy boundaries also exist in the study of contact languages is no surprise. Some languages are clearly contact languages; for other languages there is conflicting evidence, so that no firm identification is possible; and still other languages clearly are not contact languages. Of the languages covered in this book, most are definite contact lan­ guages by my definition: Hiri Motu, Pidgin Delaware, Ndyuka-Trio Pidgin, various Arabic-based pidgins and creoles, Kitúba, Sango, Michif, Media Lengua, Callahuaya, and Mednyj Aleut. Nurse argues that early Swahili was not a contact language, however, and Owens discusses (critically) the view that Arabic went through a pidgin stage. Ma'a is a contact language if it is a separate language, but its status as an independent language is debatable. Contact languages are not necessarily lingua francas,2 though many of them are. Of the ones discussed here, Hiri Motu, Pidgin Delaware, NdyukaTrio Pidgin, the Arabic-based pidgins and creoles, Kitúba, and Sango are (or were) also lingua francas; Swahili too has long served as a lingua franca, but it did not arise through pidginization or creolization. In sharp contrast, none of the bilingual mixtures fulfills (or fulfilled) such a function. This split is typical: bilingual mixtures usually or always serve as salient markers of ethnic-group identity (in fact, that's why they come into being), while pidgins and creoles arise as lingua francas. As noted above, a major purpose of this book is to make information about lesser-known contact languages more readily available. This in turn should enable specialists who are interested in generalizing over the whole range of contact languages to avoid the European-lexifier bias (especially for pidgins and creoles) of the majority of theoretical proposals in this area. The chapters provide two particularly striking illustrations of the usefulness of a broader data-base.

Introduction

5

The first concerns the issue of how many languages must be present in a new contact situation in order for a stable pidgin to develop. Virtually all generalizations about pidgin and creole languages cite the presence of more than two languages as a typical, or in some cases an absolute, condition for the emergence of a pidgin or creole. Probably the most widely cited formulation of this view is Whinnom's argument that "no pidgin has ever consolidated itself in other than a multilingual situation" (1971:104); numerous other authors (e.g. Mühlhäusler 1986:147) have accepted Whinnom's view. But two authors in this volume describe pidgins that apparently arose in twolanguage contact situations. Goddard suggests that Pidgin Delaware origi­ nated in communication between Delawares and Dutchmen and then spread to other European groups; Huttar & Velantie say that Ndyuka-Trio pidgin developed for communication between Ndyuka and Trio speakers only. To­ gether with other cases that have been described recently — most notably Kouwenberg's analysis of Berbice Dutch Creole as a two-language creole that arose out of contact between Dutch and the West African language Eastern Ịjọ (1994) — these examples show that the standard view about numbers of languages in pidgin genesis is simply mistaken. The second striking example is, in effect, the explanation for the first. Whinnom's common-sense reasoning behind his claim that more than two languages are needed for pidgin genesis is that, in a two-language situation, "there cannot be any rèally effective withdrawal of the target language" (1971:104). But here common sense is misleading. All three of the pidgins described in this book — Hiri Motu, Pidgin Delaware, and Ndyuka-Trio Pidgin — emerged in the first instance because the lexifier-language speakers deliberately withheld their full language from outsiders, instead using a simplified foreigner-talk variety that formed the basis for the eventual pidgin. Comparable cases are reported from other parts of the world, e.g. Halb­ deutsch (Lehiste 1965), but few of these turned into stable pidgins. The three examples side by side in this book provide solid evidence for the crucial importance of speakers' attitudes in determining the linguistic results of language contact. Another important feature of these papers is the thoroughness of the authors' descriptions of the historical circumstances in which the contact languages arose. These descriptions reveal a number of shared historical features, but there are also significant differences. Among the pidgins, Pidgin Delaware, Ndyuka-Trio Pidgin, and some Arabic-based pidgins arose as

6

Sarah G. Thomason

trade pidgins, as did many other pidgins around the world. The setting for the development of Hiri Motu was more complex: in Dutton's account, the Motu used their foreigner-talk variety with all foreigners, but a stable pidgin crys­ tallized only when many foreigners arrived — including the first police force, whose members came from other islands and were not Motu speakers — to provide a multilingual setting in which a single lingua franca was needed. The three creoles described in the book followed three different routes of development. East African Nubi, though connected with Arabic-based trade pidgins, was creolized when ethnically mixed soldiers established their own permanent settlements. Kitúba arose out of contact between West Africans and Bantu speakers, on the one hand, and among the Bantu speakers them­ selves in new multilingual settings, on the other; and Sango emerged as a work-group language. The five bilingual mixed languages fall into three general categories, historically speaking. Michif, Media Lengua, Mednyj Aleut, and Callahuaya all arose abruptly, by sudden creation, rather than gradually; but there were apparently two quite different motives for their development. The first three serve as the special languages of new ethnic groups or subgroups — the mixed-blood Métis (French, Cree, Ojibwe), acculturated Quechuas who belong neither to the Quechua world nor to the Spanish world, and mixed-blood residents of Mednyj Island (Russian, Aleut), respectively. The fourth, Callahuaya, functions as a secret language, used by male healers during rituals. The fifth bilingual mixture, Ma'a (Mbugu), is not the result of sudden creation but rather of long-term linguistic persistence in the face of intense cultural pressure from Bantu; far from being a new ethnic group, Ma'a speakers are an old community that has stubbornly resisted total cultural assimilation. At first glance, then, these five languages seem quite different in nature. But they share a characteristic social feature: all of them are in-group languages, used within the community as a sign of community solidarity and not understood by outsiders. They also share a salient linguistic feature: they are all comprised of two components, one from each of two source languages, and the linguistic material from each source language is adopted wholesale, without the kind of distortion that would occur in the absence of bilingualism. Finally, the structural descriptions in the first six case studies provide a strong antidote to the still common view that all pidgins and creoles have similar and simple structures. Features like the systematic OSV and SOV word order patterns of Hiri Motu, the noun class system of Kitúba, and the

Introduction

1

/kp/ and /gb/ phonemes of Sango will surely help to eradicate the idea that pidgins and creoles have maximally simple and more or less identical gram­ matical structures. The set of case studies in this book is a very small sample of the rich variety of contact languages around the world. We hope, in presenting these studies, to encourage other work along the same lines. Only with the accumulation of many solid case studies can our data-base become substantial enough to support robust generalizations about the nature and development of all types of contact language.

Notes 1.

At least one hedge must be added immediately: non-contact languages may, and often do, contain a great many loanwords, but few loanwords appear in the BASIC vocabulary. English is the most famous example of a loanword-heavy language, with up to 75% of its lexicon borrowed from French and/or Latin. But it is nevertheless easy to prove that English is a Germanic language, because the vast majority of the items in its basic vocabulary, together with most of its grammar, are of Germanic origin. An added complication is that "basic vocabulary" is not a precise concept. The idea is that basic vocabulary items are those that occur in every language, and are therefore (among other things) less likely to be borrowed than culture-specific words. In spite of its vagueness, this notion is of considerable methodological importance in historical linguistics.

2.

And of course not all lingua francas are contact languages. English, for instance, is the world's most-used lingua franca, and English itself is not a contact language.

References Foley, William A. 1986. The Papuan languages of New Guinea. Cambridge: Cambridge University Press. Kouwenberg, Silvia. 1994. A grammar of Berbice Dutch Creole. Berlin & New York: Mouton de Gruyter. Lehiste, Ilse. 1965. "A poem in Halbdeutsch and some questions concerning substratum". Word 21.55-69. Mühlhäusler, Peter. 1986. Pidgin & creole linguistics. Oxford: Basil Blackwell. Whinnom, Keith. 1971. "Linguistic hybridization and the 'special case' of pidgins and creoles". Pidginization and creolization of languages, ed. by Dell Hymes, 91-115. Cambridge: Cambridge University Press. Williams, Jeffrey P. 1993. "Documenting the Papuan-based pidgins of insular New Guinea". Atlantic meets Pacific: A global view of pidginization and creolization, ed. by Francis Byrne & John Holm, 355-367. Amsterdam: John Benjamins.

Hiri Motu Tom Dutton The Australian National

1.

University

Introduction1

Hiri Motu is the name now used officially to refer to the language that used to be known as Police Motu. This is a pidginized form of Motu, the native language spoken around Port Moresby, capital of Papua New Guinea, and for some fifty kilometers or so east and west of it along the coast (Dutton 1969) — see map.2 Until independence in 1975 Hiri Motu (as Police Motu) was the principal lingua franca between peoples speaking mutually unintelligible languages (including Europeans) and the unofficial language of administra­ tion of much of the southern half of the country that was then known as the Territory of Papua. In the years immediately preceding and following inde­ pendence Hiri Motu assumed an important role in national politics. Today not so much is heard of it in this role although it still has the potential to become so should social conditions change to favor it. It is still, however, an important language in Papua New Guinea and one of two recognized unofficial national languages of the country, even though it is not being learned and used by the younger generation to the same extent as previously, as Tok Pisin and/or English become more widely known.3 Historically the origin of this language is uncertain as there are no relevant written records. Until recently the popular view was that (and this is the view that underlies the name change from Police Motu to Hiri Motu) it is a continuation of a trade language used by the Motu on annual trading voyages, or hiri, to the Gulf of Papua where the Motu traded with linguistically unrelated peoples.4 However, recent research shows that this cannot be the case. Instead the language most probably developed out of a special variety or register of Motu used by the Motu in talking to those who came to visit or trade with them

Tom Dutton

10

Hiri Motu

11

in their own area, whether from linguistically unrelated areas or not. This variety was a simplified form of Motu, a kind of Motu Foreigner Talk, that was, it will be claimed, taken up, used, and spread in a revised form by members of the first police force and others in British New Guinea, as Papua was called then, from the late 1880s onwards. As a result it became associated with the police force and soon became known as Police Motu.

2.

Distribution and varieties

A survey of the distribution of Police Motu in 1961 (Brett et al. 1962) showed that it was at that time spoken throughout most of Papua except for those areas which had had little contact with the Administration (as, for example, in the Southern Highlands District and in distant parts of the Western, Gulf, and Milne Bay Districts) or where there were competing church languages (as, for example, in the Milne Bay and Western Districts). The Central District contained the highest number of speakers and the number of speakers dimin­ ished roughly in relation to the distance traveled away from Port Moresby in any direction. The only exception to this was in the coastal area of the Purari River delta where the language was adopted as the language of the Tommy Kabu Movement5 in the 1950s and was spoken "as a conscious preference" to the mother tongue of the members (Hitchcock & Oram 1967:11). In 1962 the number of speakers was estimated to be approximately 65,000, although this did not include 12,000 Motu and Koita villagers around Port Moresby who spoke Motu as their first or second language and who generally also know some kind of Hiri Motu. In the 1966 census, however, something like 110,000 persons over the age of ten years claimed to be able to speak "simple Police Motu", or at least could answer census questions in it, and in the 1971 census upwards of 150,000 persons living in Papua New Guinea are said to have spoken it. Not all of these were Papua New Guineans but the majority of those that were came from the six administrative districts that used to make up the old political unit of Papua. The present situation is unknown as there are no more recent census statistics available. There are some indications also, as noted above, that younger speakers are not learning the language but are learning and using Tok Pisin and/or English instead. Because of its wide distribution it comes as no surprise to find that there is no such thing as standard Hiri Motu. Instead it exists as a series of varieties

12

Tom Dutton

(distinguishable chiefly by their sound systems and vocabulary) representing varying degrees of difference within two dialects — a Central one and a nonCentral one. Thus, for example, whereas speakers of the Central dialect generally maintain phonological distinctions made by the Motu (because their languages are closely related to Motu), speakers of the non-Central dialect do not. Consequently speakers of the non-Central dialect will be heard to say lau for both "I" and "go", whereas Central dialect speakers will say lau for "I" and lao for "go". Similarly, speakers of the Central dialect use more Motu words than speakers of the non-Central dialect do, the latter using words taken from different varieties of English or other languages of Papua. Thus, for example, speakers of the Central dialect will be heard to say ginigunana for "first" while those of the non-Central dialect will be heard to say namba wan, a form borrowed from the variety of pidgin English formerly spoken in Papua and the adjacent Torres Straits.6 Or again, while Central dialect speak­ ers will use rata for both "breast" and "milk", non-Central dialect speakers will use rata for "breast" and susu for "milk"; this latter word comes from some form of pidgin English or from one or more of the languages of Milne Bay, most probably Suau. Finally, whereas Central dialect speakers will say tamagu for "my father" and tamamu for "your father", non-Central dialect speakers will say lauegu tamana and oiemu tamana respectively. The Central dialect of Hiri Motu is that used mostly by the Motu and speakers of other closely related languages in the neighboring area who generally know some Motu. These languages are those generally referred to in the literature as Austronesian, as distinct from most others in Papua New Guinea, which are non-Austronesian or Papuan — see map. The other dialect is that used by speakers of all other languages, which are mostly not related to Motu, except in the Milne Bay Province and adjacent areas. These are all Papuan languages. This latter dialect is much more widespread than the former and several commentators have advocated that it, or selected parts of it, should be taken as the standard dialect for mass communication purposes. In fact the Central dialect is the one most used for these purposes.7

3.

The pidgin nature of Hiri Motu

As already indicated, Hiri Motu is a pidgin language. That is, it shares a number of features with pidgins observed elsewhere. Thus, for example, it is

Hiri Motu

13

reduced in structure compared with its major lexical-source language, Motu. It lacks, for example, most of the inflectional and derivational morphology of Motu, has no irregular verbs, makes no distinction between alienably and inalienably possessed nouns, has restricted adjective agreement, and has generalized dual forms and a universal postposition dekena or dekenai which is very much like long in English-based pidgins of the neighboring region. Its vocabulary is, moreover, restricted and not based solely on Motu. Hiri Motu can also be said to be like other pidgins in being nonnative to most speakers — there are reputed to be a number of speakers who speak it as their first language, but these have never been surveyed or studied — and in being used for communication between speakers of mutually unintelligible languages across Papua. Thus, even though, as will be argued here, Hiri Motu is most probably a continuation of a simplified form of Motu that was used for communicating with visitors to the Motu area, it has developed its own structure which has to be learned just as does that of other well-recognized pidgins. Thus Motu speakers cannot produce Hiri Motu by ad-hoc simplifica­ tion, just as English speakers cannot produce Tok Pisin, for example, by adhoc simplification of English (however much most nonspeakers think they can).8 It is, moreover, not immediately intelligible to Motu speakers who have not been previously exposed to it. In other words, Hiri Motu is a language in its own right and is not some debased or broken form of Motu.

4.

The origin and development of Hiri Motu

At the time of first European settlement in 1874 the coastal area around Port Moresby was occupied by two completely different and linguistically unre­ lated groups of people, the Motu and the Koita (or Koitabu as the Motu call them). The Motu lived in maritime villages between Kapakapa in the east and Galley Reach in the west and were divided into two groups or tribes, the Eastern Motu and the Western Motu, who lived east and west of Bootless Bay respectively. Of these two groups the Western Motu played the most impor­ tant role in the origin and development of Hiri Motu (as Police Motu), for it was they who were drawn into closest and increasing contact with foreigners after the arrival of the first Europeans. They were, however, distinguished culturally from their eastern counterparts by the fact that at the time of European contact (and for an unknown number of years before) they were

14

Tom Dutton

involved in a complex network of trading relationships with linguistically related and unrelated groups east, west, and inland of their present position. The most spectacular and important part of this trade was the hiri, or annual trading voyage to the Gulf of Papua several hundred kilometers away to the west. During these voyages the Motu visited such groups as the Elema (the general name given to a group of coastal peoples speaking eight closely related languages of the Eleman Family between Cape Possession in the east and the Purari River delta in the west) and their neighbors immediately westwards, the Kikori, who inhabit the delta of the Purari River. These groups speak languages unrelated to Motu and only very distantly related to each other (if they are related at all). On these visits and on return visits made by some of these groups, the Motu and their trading partners communicated with each other in one or more of at least two different trade languages which were pidginized forms of the languages spoken by the Gulf traders. One of these languages, the Hiri Trading Language, Eleman variety, hereafter HTL(E), was based largely on the component languages of the Eleman Family, and the other, the Hiri Trading Language, Koriki variety, or HTL(K), was based on the single language, Koriki. Both of these languages had a number of features in common, although they were not mutually intelligible. The examples in Table 1 show what these languages were like, and most importantly, that they were quite different from, and should not be confused with, Hiri Motu, as has been the case in the past, although they were obviously similar in structure to it in many respects.

Table 1. The Hiri trading languages and Hiri Motu compared. HTL(K)

HTL(E)

Hiri Motu

English

Koa(nu) vapeo? who(his) canoe

A, neia enane eh, who his vevara? canoe Ara enane vevara. I my canoe

(Inai be ) daika (this focus) who ena lakatoi? his canoe (Inai be ) lau(this focus ) I egu lakatoi. my canoe

Whose canoe is this?

Na vapeo. my canoe

(It's) mine.

Hiri Motu

15 Eme enane rare you your name neia? who Ara enane rare I my name maro Moi. focus Moi

Oi- emu ladana you-your name (be ) daika? (focus) who Lau-egu ladana I- my name be Moi. focus Moi

E, uapekai. yes, stay Moi, anene pei Moi, come food navai. eat Na okuai! me give

Ene enane pamora you your friend rare maro neia? name focus who (Ara enane pamora (I my friend rare maro) Elamo. name focus) Elamo Elamo abuviti? Elamo stay E, abuv iti. yes, stay Moi, abusi ma Moi, come water siahu abulari. hot eat Ara porohalaia! me give

Enane pu miai go sago get anea! come

Abuari pai go sago avaia abusi! get come

Pu peo. sago not

Pai penepene sago some lasi. not

Oi- emu pamora you- your friend (be ) daika? (focus) who (Lau-egu pamora (I- my friend be ) Elamo. focus) Elamo Elamo ia noho? Elamo he stay Io, ia noho. yes, he stay Moi, oi mai Moi, you come aniani oi ania. food you eat Lau oi henia/ oi me you give/ you henia lau dekena! give me to Oi lao rabia you go sago oi mailaia! you bring Rabia (ia noho) sago (it stay) lasi. not

Ni noe koana? you name who

Na noe Moi. my name Moi

Ni vake(nu) your friend noe koana? name who (Na vake(nu) (my friend(his) noe) Elamo. name) Elamo Elamo uapekai? Elamo stay

What's your name?

My name is Moi.

What's your friend (or trade partner)'s name? (My friend (or trade partner)'s name is) Elamo. Is Elamo here? Yes, he's here. Moi, come and eat (some food).

Give it to me!

Go and bring some sago!

There's no sago.

16

Tom Dutton

But the hiri was only part of the complex network of trade and exchange that the (Western) Motu were engaged in at the time of European contact, albeit the most spectacular part. Other parts of this network involved contact with speakers of other languages that lived in their immediate neighborhood and a little farther afield. Some of these trading groups spoke (and still speak) languages that are closely related to Motu, while the remainder spoke (and most still speak) languages which, as already indicated, are not related to Motu and are only distantly related to each other in groups or families, if they are related at all. The closest to the (Western) Motu were the Koita, who, at the time of first European contact, lived amongst the Motu as minority sections in Motu villages or in separate villages close by. Inland of the Koita and closely related to them linguistically lived the Koiari and the Mountain Koiari (also spelt Koiali). Because of their position, however, these two groups had little direct contact with the Western Motu, although there was some trade between them. They apparently knew very little if any Motu. Beyond the Koita the Motu also traded directly with linguistically related groups on either side. In the west they visited the Doura and Gabadi of the Galley Reach area and to the east they had developed rather special relation­ ships with the Vulaa, a tribe inhabiting Hula and associated villages around Hood Point. Motu villagers who did not go on the hiri depended on these people for fish while the hiri traders were away. In return the Vulaa received sago and other gifts from the returning hiri canoes. It is not known for certain what language was used by the Motu and their trade partners in these more local contacts, but it can be assumed on the basis of the available evidence that this was a simplified form of Motu, hereafter referred to as Simplified Motu. Thus, for example, when the Rev. W. G. Lawes of the London Missionary Society settled in Port Moresby in 1874 as the first European missionary amongst the Motu, he attempted to learn Motu. However, it was not until some time later that his son, Frank, who played with the boys in the village and learned Motu from them, drew his father's attention to the fact that he did not speak "true" Motu but only a simplified version of it,9 which he also used in making his first translations into Motu (Taylor 1978). Later Lawes noted that this simplified version of the language that he had been taught was characterized by "a good many colloquialisms" or "instances of pidgin Motuan" in use amongst the Motu which were "not correct grammatically" but were "sanctioned by usage" (Lawes 1896:30).

Hiri Motu

17

This Simplified Motu was, moreover, used by the Motu "in speaking to foreigners", although they themselves would "never do so amongst them­ selves". Although Lawes did not ever describe this foreigner talk in more detail we have been able to get some further insight into its nature and use in various ways. As a result it is possible to say that this Simplified Motu was not a separate language from Motu (in the sense that it was unintelligible to native Motu speakers and had to be learned by them as a second language), but was merely a special variety or register of Motu used for communicating with a particular class of people, notably those seeking entry to the Motu world. This variety was furthermore most probably not a stable one but probably varied from speaker to speaker over time, and depending on whether the person being spoken to was a complete stranger to the Motu or not. As a variety it was distinguishable from "true" Motu by a number of grammatical features of the sort already noted and to be discussed in more detail below, some of which are not found in Motu. Thereafter nothing is heard of this Simplified Motu until more than twenty years later when a distinct language with many of the same features appears that was to become known as Police Motu and later Hiri Motu. Although there is again no linguistic evidence on the point it is presumed, on the basis of other evidence to be presented, that the two are connected and that Police Motu is a continuation of this Motu Foreigner Talk or Simplified Motu (in a modified form) and not a separate development. There are two reasons for this. The first has to do with the fact that the received oral tradition claims that Police Motu is a continuation of a hiri trade language (as already indicated) and the second with social developments in the Port Moresby area following the arrival of London Missionary Society missionary Lawes in 1874. With respect to the received oral tradition it is to be noted that the principal reason for suggesting the name change Police Motu to Hiri Motu was that this language was supposed to be a "lineal descendant of the language of the Hiri" (Chatterton 1971:2). But it has already been shown that this could not have been the case, because not only did the Motu use at least two such languages (the HTLs) in hiri trading, but also these were markedly different from Police Motu, especially in vocabulary. So either the tradition has to be rejected as nonsense, or it has to be seen as having been misunder­ stood or misinterpreted at some point. Alternatively Police Motu has to be seen as a "lineal descendant" in a sense different from that normally accepted in linguistics. Given, however, that the Motu had another contact "language",

18

Tom Dutton

Simplified Motu, which shares most grammatical features and vocabulary with Police Motu, a point which is demonstrated and discussed further below, the simplest explanation would appear to be that the tradition has indeed been misunderstood or misinterpreted at some time. This position is attractive, moreover, because it saves both the tradition, in a modified form, and the linguistic definition of continuation. Thus we may suggest that what the tradition really said was simply that Police Motu was the continuation of a trade language, one of several used, and not THE hiri trade language. Thus, both the received oral tradition, once reconstructed, and the linguistic evi­ dence are consistent with an hypothesis that Police Motu is a lineal descen­ dant of the contact "language" Simplified Motu. However, for this to be true particular social forces must have been active in and around Port Moresby to keep this simplified form of Motu alive and to extend its use into a general lingua franca before 1890, the year that the first police force was established in Port Moresby. One of these forces, notably the need to trade or converse with non-Motu coming into the Motu area, has already been referred to. A second force, probably not unrelated to the first, was that the Motu were apparently not keen for non-Motu to learn Motu; other forces have to do with social developments in the Port Moresby area following the arrival of mis­ sionary Lawes in 1874. Thus when news of the discovery of Port Moresby began to filter down through the Australian colonies and overseas, people of all sorts began making their way towards this new terra incognita as if drawn to it by some strange magnet. The first to do so were missionaries, who were soon followed by "scientists", entrepreneurs, explorers, adventurers, traders, and others, who in turn were followed by Government officials. Each of these groups had their effect on the use and transmission of different languages in the area, but it would appear to be the unofficial "visitors" who were the most important as far as the development of Police Motu was concerned. These "visitors" began coming to the area soon after Lawes arrived there in 1874 and continued to do so unchecked for the next sixteen years before the governor of the newly proclaimed colony of British New Guinea was in a position to control them. They included a wide range of ethnic types such as Chinese, "Malays", "South Sea Islanders", persons of mixed race, Maltese, Ceylonese, Indians, Filipinos, Europeans, Americans, and a large number of British-Australian origin. Many of these stayed and "married" Papuan women from the Port Moresby area and their descendants are still to be found there

Hiri Motu

19

today. Apart from a few extraneous cases these unofficial visitors generally fall into two classes: "South Sea Islanders" (who came mainly from what used to be called the New Hebrides but is now called Vanuatu) and "Malays" (who came from around the Singapore-West Indonesian area). Because these "visitors" were foreigners and were dependent on the Motu for food and other services, they would presumably have been in a situation similar to that that missionary Lawes was in when he first arrived. In other words they would have been treated like other visitors and addressed in Motu Foreigner Talk or Simplified Motu and not "true" Motu. And because these foreigners came from so many different sources (although collectively many of them may be labeled "South Sea Islanders" or "Malays"), and because they presumably had no language in common if they did not know some form of "broken" (or pidgin) English,10 the lingua franca of the Torres Straits and Queensland from where most, if not all, are known (or presumed) to have come to Port Moresby, there would have been increased pressure on them to learn and use some form of Motu as a lingua franca. Besides, they were in the minority and therefore in a weak social position. Consequently they would have been forced to accommodate to the Motu and not vice versa. Finally, because many of these foreigners were traders and traveled from one language area to another where some Motu was already known (and no English was) at the time of European contact,11 some form of Motu was bound to become the natural lingua franca between foreigner and non-Motu. In short, the whole sociolinguistic context of the Port Moresby area in the period preceding government intervention predetermined Motu in some form or other to becoming the established lingua franca of the area. But this was not just a change in frequency of use of the language; rather, it was a fundamental one in kind where its function was widened, or changed, from being a unidirectional, or vertical, one (that is, Motu-to-foreigner and vice versa) to a multidirectional, or horizontal, one (that is, foreigner-to-foreigner — includ­ ing non-Motu Papuans — and vice versa). Just what form of Motu developed out of this contact is difficult to say without written records. However, given (a) that present-day descendants say that some of their forefathers spoke "pidgin Motu", and (b) that, as will be indicated below, a subset of these foreigners was employed by the incoming Government from 1884 onwards and that thereafter Police Motu appears amongst those who were under their charge, it seems reasonable to assume that something similar was the lingua franca of this foreigner community.

20

Tom Dutton

That is, it seems reasonable to assume that a form of Simplified Motu similar to that taught to Lawes, but not necessarily identical with it — in fact, most likely not identical to it, given the different origins, social standing, and presumably different language learning abilities of these foreigners12 — was the lingua franca of this community. In any case, it would seem to be highly likely that this community would play a significant role in establishing a form of simplified Motu as the general lingua franca of the Port Moresby area. But there were other important social changes in and around Port Moresby at the time which also had their effect on the language. These changes were intro­ duced by the establishment of a Government presence in Port Moresby. When the four-year-old Protectorate of British New Guinea was con­ verted into the colony of British New Guinea in 1888, Dr. (later Sir) William MacGregor was appointed first Governor. He arrived in September, 1888, proclaimed the colony, and immediately set about the two main tasks of getting to know as much of the country as possible and of building up a suitable administrative structure with the limited resources available. At that time the colony was divided into three divisions — Eastern, Central, and Western — with headquarters at Port Moresby and Samarai and outstations at Rigo and in the Louisiade Archipelago where gold mining was developing. Port Moresby had been surveyed as a town in 1886 but had not yet developed into anything resembling one. At this time much of the country was still unknown and not yet under government influence or control. Those areas that were under some sort of control were those more or less coextensive with mission influence or areas such as the following where commercial activity had been going on for some time: the Port Moresby-Aroma coast and immediate hinterland in the Central Division, the Kiwai area in the Western Division, and the coastal parts of the mainland tip around Milne Bay and islands in the Louisiade Archipelago in the Eastern Division. With MacGregor's arrival conditions changed dramatically for the na­ tive population. Hitherto, the Protectorate Government had been merely conducting a holding operation, attempting as far as possible to avoid distur­ bances and to protect the local population from undesirable outside influences and from itself. It had few legal powers to make laws and to enforce them. But a colony was a different matter and it was the change in the nature and methods of the colonial Government that had such an impact on the local population and, in turn, on the linguistic situation. Indeed, the impact was so

Hiri Motu

21

great that by the time that MacGregor left British New Guinea in 1898, what was later to become known as Police Motu had become the principal, al­ though not the sole, unofficial language of administration in many areas and the scene was set for its further expansion into other areas as similar policies and methods continued bringing new areas under control. The principal agents in this development were the three instrumentalities of law and order: the police force, the village constable system, and the prison system. These three systems were closely related and integrated and fed on, and into, each other in the following way: the prisons provided recruits for the police force and the village constable system, which in turn provided recruits for each other as well as fresh prisoners. When Sir William MacGregor arrived to proclaim the new colony of British New Guinea, one of the most pressing requirements was for a suitable police force with which to help extend government influence over an increas­ ing area and to enforce law and order over those areas. Hitherto administra­ tors of the Protectorate had to depend on "commodore justice" and on an unofficial police force composed of a heterogeneous collection of foreigners and local Papuans who acted in various capacities but who had never been trained in any way for the police work they could be called on to do. Consequently MacGregor soon set about organizing an official force which he called the Armed Native Constabulary. But as no trained personnel were available in British New Guinea to form the nucleus of this unit, MacGregor appealed to the Governor of Fiji for assistance, with the result that two Fijians and twelve Solomon Islanders from the island of Malaita were recruited to go to British New Guinea for periods between one and three years — the two Fijians, a sergeant and a corporal, for one year and the remainder, constables, for three years. The members of this force all spoke mother tongues which are Austronesian languages distantly related to Motu. In addition the Solomon Islanders probably spoke some form of Fijian as a result of having spent between six and twelve years working on plantations and/or for the Govern­ ment in Fiji before being recruited, and all are presumed to have also spoken some form of English as a necessary prerequisite to being chosen to serve under English-speaking commandants in British New Guinea. This nuclear force arrived in British New Guinea in 1890 and was gradually expanded by the addition of Papuan recruits mostly from the Kiwai area of the Western Division, until, by the end of MacGregor's term as Governor in 1898, the force consisted of one hundred and ten Papuan non-commissioned officers and constables drawn from most parts of the country then under control.

22

Tom Dutton

The way in which the police force was formed and developed has a number of implications for the history of Police Motu, whose name is so closely associated with it. The most important of these are: (i) Because the founding members of the force were drawn from areas outside the Central Division, where "true" Motu was (and still is) spoken natively, they did not know any Motu on arrival. Consequently they must have been forced to communicate with one another and with their superiors in some form of English, which the majority, if not everyone, must have known. For the founding Fijians and Solomon Islanders this English was most prob­ ably what is generally referred to as "broken" English but was in reality a form of pidgin English similar to that found in Vanuatu and the Solomon Islands today. Similarly the founding Papuan members from the Kiwai area of western Papua spoke a form of "broken" or pidgin English, a creolized version of which is the primary language in the Torres Straits today and whose official name is Broken (Shnukal 1988). Thus even though Police Motu eventually became associated with the force as THE language of the force, it was not the sole language of the force; "broken" or pidgin English was equally important in the formative years. In fact both languages were required for a long time as each was used in different parts of the country and a member of the force could be transferred from one part of it to another on duty at any time. But to understand why only Police Motu became associated with the force (as the name indicates) and what it was like, we have to turn to the second main implication of the way the force was formed mentioned above. (ii) Because the force was formed in the way it was and was housed and trained in Port Moresby, and because the language situation in that area was the way it was (viz. Motu was the most widely known language while at the same time very little English of any kind was known), the police could not avoid learning some kind of Motu if they were to carry out their duties in the Port Moresby area. Thus, it will be remembered that the initial nucleus of the force consisted of imported Fijian and Solomon Islanders and that Papuan members were only added "gradually" (to use MacGregor's words). When these imported men arrived they would immediately have recognized a number of what would today be called wantoks (or persons from the same area and/or cultural background) in Papua New Guinea amongst the foreigner population of Port Moresby. As these wantoks had been in Port Moresby for some time and were living in or near Motu villages, it is presumed that they

Hiri Motu

23

could speak some sort of Motu. In addition it is presumed that they would have very soon introduced their newly arrived "friends" into the Port Moresby scene, in particular to its local politics and the utility and necessity of knowing some kind of Motu for survival purposes in the Port Moresby area. Even if this had not taken place, the new arrivals would surely have quickly come to the same conclusion themselves, for they were very soon put to work after their arrival in the Rigo and Mekeo areas east and west of Port Moresby. Here they would have been working with Government officials and other employees, some of whom were the wantoks already referred to and all of whom it is presumed (for reasons given above) spoke some kind of Motu. Consequently, they would have been exposed to Motu being put to use in the field. Presumably they would also have been told (even if they had not needed to know) that in the Port Moresby area "Motu" was the most widespread language and that one had to be able to speak it if one wanted to communicate with the local people. At the same time, they were probably given some elementary advice about its nature and possibly even how easy it was to learn. However, they would hardly have needed much of that kind of instruction since all of these men, as already noted, spoke languages related to Motu. Indeed, they would themselves have easily recognized many of its basic words and structural elements as being similar to their own. This can be seen by looking at Table 2, in which Motu words are compared with some of the most common words from some of those languages that are known or are suspected as having been spoken by various members of the force. For similar reasons, the form of Motu adopted by these first police is likely to have been similar to that used by the Government officers, other employees, and unofficial policemen with whom they worked and/or whom they were replacing — in other words, some form of Simplified Motu. The formation of the police force was thus a new and important element in the development and spread of this variety of language. It provided a particular social environment in which men from different parts of the coun­ try (and overseas initially) were brought together to work. In this situation, which is akin to those on plantations elsewhere, some common language of communication was soon needed. Initially, for reasons already given, this language must have been "broken" (or pidgin) English. However, the use of "broken" (or pidgin) English in the Port Moresby area was limited by the fact that "Motu" was the established lingua franca there. Consequently, although "broken" English never died out, its use in the Central Division was more

24

Tom Dutton

Table 2. Some basic vocabulary and structural elements of Motu, Fijian, and some Solomon Islands languages compared. Fijian13 English belly bird bone butterfly come die eye five fly (v.) foot go he his hornbill house I in life lobster louse mother my name paddle road salt, sea sugarcane taro two what your (sg.)

Solomon Islands Languages

Motu boga manu turia bebe mai mase mata ima roho ae lao, laka ia -na bina ruma lau lalonai mauri ura tumu sina -gu lada hode dala tasi tohu taro rua dahaka -mu

Bau

Kadavu

Lau

Kwara'ae Kwai Is

kete manumanu sui beebee lako mai mate mata lima vuka yava lako koya -na

kete manumanu tua bebe lako mai mate mata lima vuka laga lako kia -na

oga manu sili bebe mai mae maa lima lofo 'ae lea nia -na bina luma ñau lalo mouri ura 'uu tee -gu sata fote tala asi ofu alo rua ta -mu

sira hai'no' sulia beb mai' mae maa lima loh a'e leka nia -n(a) bina lum naua sae maori deng 'uu tea' -ku sata fote tal asi uuh alo rua tae -mu

14

vale yau e lomani bula urau kutu tina -qu yava voce sala maasima dovu dalo rua cava -mu

vale yau i lomani bula urau kutu tina -qu ila voce saa levu maasima tovu suli rua yava -mu

oga no suli bebe mai mae maa nima lofo 'ae leka nia ? bina luma nau laona mauo uragou 'uu tee ? rata fotee tala asi ofu alo rua taa ?

Hiri Motu

25

restricted than elsewhere. The fact that recruits were taken to Port Moresby for initial training before being distributed around the various Government stations meant that the language traditions, once established, were perpetu­ ated and, moreover, that because, as MacGregor noted in his Annual Report for 1892/3 (p. xxviii), "thirty or forty men leave the force each year for their own villages", Police Motu went with them to distant parts at a constant rate. This situation was reinforced by the two other ancillary law enforcement agencies, the village constable system and the prisons, both of which were considered by Governor MacGregor as important agents of change. These systems contributed recruits to the police force from an increasingly wide area and returned Police Motu speakers with them. Thus by the time MacGregor's period of governorship came to an end in 1898 there had been significant developments administratively and socially which had their effects on the linguistic situation. At first, MacGregor, like his predecessors in the Protectorate, had to rely on the mission and a number of locally appointed "servants" to maintain some sort of law and order. How­ ever, after 1890, Government influence and control rapidly outstripped mis­ sion influence and became the decisive factor in linguistic developments. By this time "Motu" had become the unofficial language of administra­ tion. This "Motu" was that that subsequently became known as Police Motu (and now Hiri Motu) because of its association with the police, notwithstand­ ing the fact that it was just as much part of the prison and village constable systems as of the police force. Thereafter this language spread (along with "English") with expanding contact with Europeans, so that today the language is spoken throughout most of what used to be called Papua. In 1971, in response to changing political circumstances in Papua New Guinea, a Study Conference on Police Motu was organized by the Government in Port Moresby to consider various aspects of the language's future. One of the results of that conference was the adoption of a new name for Police Motu. It was felt that because the term Police Motu had "become an anachronism" (Chatterton 1971:1) the name ought to be changed to something more in keeping with the origin of the language. As a result it was changed to Hiri Motu because, as has already been pointed out, it was thought that the language was a continuation of the hiri trading language. This new name was subsequently adopted as the official name of the language despite the misunderstood oral tradition that lies behind it.

26 5.

Tom Dutton Sources of the structure and vocabulary of Hiri Motu

Sourcing the structure and vocabulary of Hiri Motu is complicated by several factors — the generally poor quality and amount of relevant early data available, the complex nature of the external history of the language, and the number and nature of possible source languages available. Assuming, how­ ever, that Hiri Motu developed out of Simplified Motu as the social evidence outlined above strongly suggests, all discussion of the source of non-Motu features in Hiri Motu must begin with Simplified Motu. Given further that Simplified Motu was a register of Motu and not a separate pidgin language, then it is also reasonable to assume, especially when there is no evidence to the contrary, that this variety was composed solely of Motu elements even though some of these were apparently not used in strictly Motu ways and even though the variety probably varied from speaker to speaker and from context to context. Consequently all non-Motu features found in Simplified Motu, and later Hiri Motu, must have come from some external source or sources if they were not the result of some universal or internally generated tendencies. There are three languages most suspect of being probable source languages for such non-Motu features: Koita, the Papuan language spoken in and around Western Motu villages; Kiwai, the Papuan language spoken by MacGregor's first Papuan police recruits; and "broken" English, the initial lingua franca assumed to have been spoken by the founding members of MacGregor's police force. Each of the three aspects of language, phonology, morphosyntax, and vocabulary, are involved and will be considered separately. 5.1. Phonology In the beginning, as has just been suggested above, Hiri Motu, as Simplified Motu, is most likely to have had a phonology similar to that of Motu. As this variety expanded into a pidgin, however, it became increasingly diversified and the pronunciation in particular began reflecting the features of the mother tongues spoken by the individual users.15 Even so, and despite still observable variation within it, the language developed a number of features which are fairly consistent across the non-Central dialect (and which therefore help define it) irrespective of the nature of the background phonologies of the individual speakers. These features are:

Hiri Motu (i) (ii)

(iii) (iv) (v)

27

the reduction of the Motu kw/gw + au sequence to ko, as in koraia "inceptive aspect marker" (< Motu gwauraia "to talk about") ; the absence of a distinction between Motu /gl and /y/. Thus whereas Motu distinguishes between guria "to bury" and yuria "to pray", there is no contrast between these words in the non-Central dialect of Hiri Motu — they are homophonous; the contrast between Motu /r/ and /l/ is generally neutralized so that words like lau "I" and rau "leaf are both pronounced lau; /h/ is often lost, especially word-medially; the Motu contrast between /ao/ and /au/, between /ae/ and /ai/ and between /oe/ and /oi/ is lost so that each pair is realized as /au/, /ai/, and /oil respectively.16

Distinctive as these features are, it is not possible to trace them to any particular source. On social and historical grounds Koita and Kiwai are the most suspect, if these features were not progressively acquired from other Papuan languages as the dialect developed or if they did not develop indepen­ dently. Both Koita and Kiwai have a number of these features in common with each other and Motu — e.g., Koita has a contrast between /g/ and /y/, both Koita and Kiwai have /h/ and an open syllable structure and /a+u/ sequences, and Kiwai has a contrast between /r/ and /l/ (Wurm 1973:226). Consequently it is hard to see how they could have been sources. Likewise the languages spoken by the Solomon Islander and Fijian policemen are not likely to have been sources as they are related to Motu and are similar to it in general structure. 5.2. Morphosyntax This is the area of principal difference between Simplified Motu, Hiri Motu, and Motu. Compared with Motu, Simplified Motu and Hiri Motu are gener­ ally much simpler in structure. They also include a number of features not found in Motu. The features which distinguish Simplified Motu and Hiri Motu from Motu are those set out in the Appendix together with comparative notes for a selection of other languages that are relevant to the problem of sourcing.17 Comparing the Simplified Motu and Hiri Motu features, it is to be noted that, except for four features for which there is no evidence in Simplified Motu (viz. 11, 13, 17, 18), the only features that distinguish Hiri Motu from Simplified Motu are:

28

Tom Dutton 2 3 8 20

"Have" and "have not" Subject and focus markers ese and be Dekena(i) as a generalized locative postposition Word order flexibility

Such a correspondence between the two languages implies that many of the features found in Hiri Motu were acquired from Simplified Motu, and that Hiri Motu is indeed a continuation of Simplified Motu, as the social evidence presented above suggests.18 Where these features came from in Simplified Motu and Hiri Motu is, however, another question and one the linguistic evidence cannot answer unambiguously. On the one hand, most of the fea­ tures that distinguish Simplified Motu and Hiri Motu from Motu have paral­ lels in one or more of the languages the Motu were in contact with at the time of first European contact. On the other hand, all except perhaps feature 14 (gwauraia) can be accounted for as simplifications of Motu by the application of universal simplifying principles. But without further evidence it is not possible to make out a convincing case for the primacy of either of these two sources over the other. Indeed, it is highly likely that contact-induced changes and simplifying principles conspired with each other to give the observed results. Where Hiri Motu differs from Simplified Motu, however, it must be the case that Hiri Motu developed those features independently after it diverged from Simplified Motu, that is, after Simplified Motu began to be used in a wider context. As already noted there are four features involved (2, 3, 8, and 20). However, as two of these (3, 8) have similar structures in one or more of the languages that the Motu were in contact with at the time of the foundation of the first police force, or that the police themselves spoke, and as they are, at the same time, derivable from Motu by universal simplifying principles, it is clear that they are in very much the same category as the similarities dis­ cussed above, i.e. they most probably result from the interaction of simplify­ ing principles and contact-induced change. The remaining two features are of a different kind because they are not found in Motu and cannot be said to be simplifications of Motu. In fact feature 2 is as complex in Hiri Motu as in Motu and feature 20 is more complex in Hiri Motu than in Motu. Taking each of these in turn: (a)

Feature 2: "have" and "have not"

Hiri Motu

29

In Simplified Motu the method of indicating "have" and "have not" appears to have been similar to the Motu method in which the prepositions mai "with" and asi "without" are used without a verb, as in examples (1) and (2): (1)

Lau na mai egu ira. I be with my axe "I have an axe."

(2)

Lau na asi egu ira. I be without my axe "I have no axe."

In Hiri Motu "have" and "have not" are expressed either by using dekena(i) "at" with noho "stay" or noho lasi "not stay",19 as in example (3), or by using the possessive pronoun with noho, as in (4): (3)

Lau dekena(i) be ira ia noho (lasi). I at focus axe it stay (not) "I have (no) axe."

(4)

Lauegu ira ia noho (lasi). my axe it stay (not) "I have (no) axe."

Interesting as these differences are, however, it is not possible to source them unambiguously. The reason is that there are many competing possibili­ ties. First, they could represent independent developments. Second, they could represent changes induced by contact with the HTL(E), which has a (partly) similar construction. A third possibility is that they could represent developments induced by contact with the form of 'broken" English assumed to have been spoken by unofficial "visitors" and the founding members of MacGregor's police force recruited in Fiji. This is so because one way of expressing "have" was probably (judging by the fact that it is currently used in Bislama and other modern forms of the "broken" English spoken in the South Pacific in earlier times) to use a possessive construction with stap "to be", as in the following example from Bislama:20 (5)

Akis bilong yu i stap? axe of you be "Do you have an axe?"

30

Tom Dutton

A final possibility is that the constructions could have resulted from changes induced by HTL(E) and "broken" English constructions. Yet despite the uncertainty surrounding the origin of these "have" constructions, one thing is clear and that is that the constructions could not have come from the native languages spoken by MacGregor's policemen and those joining later, as these were Papuan languages which do not use similar constructions. Koita is excluded for similar reasons. (b) Feature 20: word order Word order in Hiri Motu is quite varied, depending on whether the subject (S) and object (O) are nouns or pronouns. There are four possibilities: (i) When both S and  are nouns the order is S ese O V, as in example (6): (6)

Sisia ese boroma ia itaia. dog pig he see "The dog saw the pig."

(ii) When S is a noun and  is a pronoun the order is  be S ia V, or less commonly SVO, as in (7): (7)

Lau be sisia ia itaia. ~ Sisia ia itaia lau. I dog it see dog it see I "The dog saw me."

(iii) When S is a pronoun and  is a noun the order is OSV, as illustrated in (8): (8)

Sisia lau itaia. dog I see "I saw the dog."

(iv) When S and  are pronouns the order is OSV, or less commonly SVO, as in (9): (9)

Oi lau itaia. ~ Lau itaia oi. you I see ~ I see you "I saw you."

In Simplified Motu, however, word order was apparently invariably SOV (although there is unfortunately no evidence available for pronominal ob­ jects), as in (10):

Hiri Motu

31

(10) Ia mero itaia. he boy see "He saw the boy." This means that sometime between the documentation of Simplified Motu and that of Hiri Motu word order developed considerable flexibility, a flexibility not found in Motu, which has the basic order SOV for nominal Ss and Os but uses prefixes and suffixes on the verb for pronominal subjects and objects. Excluding ese and be, which are Motu features, either this develop­ ment must have been an independent one or it derives from the pidgin Englishes and/or other languages spoken by unofficial "visitors" or the first policemen. A further possibility is that it derives from Koita. Of these the most likely source is Koita, if any one particular source was responsible for this development in Hiri Motu, as it has some of the same flexibility as exhibited by Hiri Motu. For example both SOV and OSV word orders occur depending on the nature of the sentence (Dutton 1975). Whether or not Koita was the only or principal source, it would seem that the other possible sources named above could hardly have been sources as (i) the "broken" Englishes presumed to have been spoken by "visitors" and the early policemen did not have this kind of variation (judging by modern forms of them) and (ii) the majority of native languages spoken by the "visitors" and early policemen were Austronesian (including Indonesian, Solomon Islands, Vanuatuan, and Fijian languages) and Papuan ones which did not have this kind of variation. Clearly then it is not possible to identify the source or sources of word order flexibility in Hiri Motu with any confidence. 5.3. Vocabulary The vocabulary of modern Hiri Motu comes from several different sources. The largest percentage of words comes from Motu itself, as the name sug­ gests. This percentage includes the language's most basic vocabulary (that is, its pronouns, its names for common objects, body parts, kinship terms, adjectives, and simple action verbs) as well as a number of other words introduced by the South Sea Island pastor-teachers of the London Missionary Society who spearheaded the introduction of Christianity into Papua in the 1870s. These words include such items as pakosi "scissors", tamaka "shoes", pavapavana "king", and mamoe "sheep". The remaining Hiri Motu vocabu­ lary is made up of words that come from other languages of Papua as well as

32

Tom Dutton

from "broken" or other varieties of English that have been spoken in Papua throughout its history. Thus Hiri Motu contains such words as namba wan "first", bulamakau "beef, rais "rice", ti "tea", kesikesi or bisikesi "biscuit", traim "to try", and mikisim "to mix", which were part of the contact vocabu­ lary used by foreigners speaking "broken" or other forms of English, who came to the Port Moresby area in increasing numbers in the 1870s and 1880s. Subsequently other English words were added as new ideas and goods flowed into the country. The introduction of some of these can be dated quite precisely and their nature has changed over time as changes in technology and in English usage have occurred. The best examples of these are words for such introduced vehicles as aeroplanes, motor vehicles, and boats. Thus when aeroplanes were first introduced into Papua in the 1930s Papuans learned to refer to them as plaimasini or plaimasi "flying machines" in Hiri Motu. However, as aeroplanes became more popular and English speakers referred to them more generally as "aeroplanes" and eventually "planes", so the Hiri Motu word changed from plaimasini or plaimasi to elopleni and pleni. Simi­ larly trucks {traka in Hiri Motu today) were apparently first referred to as lori (from English "lorry"), and ships {boti in Hiri Motu today) as sisima or sitima (from English "steamer"). A small part of Hiri Motu vocabulary is also made up of words that come from languages both related and unrelated to Motu in Papua, and it is sus­ pected that many more such words were once part of the language.21 Some, such as kamkam "fowl, bush fowl", o kapore "oh sorry", dimdim "white man", and possibly susu "milk", can be traced to Suau, the Austronesian language related to Motu along the southern coast of the mainland just west of Milne Bay — see map. Others have come from Koriki (nakimi "brother-inlaw"), the substratum language of the HTL(K) trade language referred to above, from Binandere on the north coast (e.g. kiki "to yarn, tell a story"), 22 and from languages of the Torres Straits (mamoos or mamus "chief, police­ man"), all of which are unrelated to Motu. Thus linguistic features of Hiri Motu support the view suggested by social evidence that Hiri Motu is a descendant of SM, although it is not possible to identify the source or sources of most features of Hiri Motu that distinguish it from Motu and its predecessor Simplified Motu. Many of these were common to other languages that the Motu were in contact with at the time of first European contact but could equally well, and most probably do, represent the results of several processes — the application of universal

Hiri Motu

33

simplifying processes to Motu, changes induced by contact with one or more other languages which have similar structures, and independent develop­ ments.

6.

Conclusion

To recapitulate, in this paper I have argued, mainly on social grounds sup­ ported by comparative linguistic evidence, that Hiri Motu is not, as its name is meant to suggest, a continuation of one of the hiri trading languages used by the Motu in trading with linguistically unrelated peoples of the Gulf of Papua. Rather it is a descendant of a simplified form of Motu that was used by the Motu as a contact language with anyone coming to visit them in their own area. This language, unlike the hiri trading languages used by them, which were simplified forms of languages spoken by their Gulf trade partners, was a simplified form of their own language. This Simplified Motu was not related to those spoken on the hiri and was quite separate from them. It existed before the arrival of Europeans and other foreigners but was used by the Motu to communicate with the first European missionary in the area, Dr. Lawes, and taught to him as their language. The origin of this Simplified Motu is un­ known, but it most probably developed initially out of the contact between the Motu and their closest neighbors, the Koita, although others may have been involved. After the 1870s this Simplified Motu was extended in use by the arrival of a relatively large number of unofficial "visitors" who came to the area before a strong Government presence was established there. Some of these foreigners were later employed by the Government as interpreters, guides, boatmen, and unofficial policemen and were probably largely respon­ sible for a form of this language being adopted by members of the first official police force, from which the language took its early name, Police Motu. Thereafter the language became associated with the Government and the enforcement of law and order and spread with expanding Government control over the country. From that time on the history of the language became one of increasing competition with the other main lingua franca spoken in Papua before the Second World War, "broken" English. During this time also the language came into contact with other languages besides "broken" English as recruits were drawn into the police force from different areas and as the language expanded into different geographical areas. As a result the language

34

Tom Dutton

developed into two main dialects which varied in the sounds used and in some grammatical features (e.g. word order, object marking on verbs, possessive case) but especially in vocabulary (e.g. in non-Motu items obtained from other languages of Papua, "broken" English, and English). With the coming of the Second World War Police Motu was recorded and described for the first time and given official recognition. At the same time it was spread farther afield, was regularized, and was used increasingly for mass communication. After the war the language increased in status and "broken" English lost its identity as standard English was promoted officially with great vigor. In 1971 the name Police Motu was changed to Hiri Motu and the language began to acquire new ranges of vocabulary in keeping with increased contact with English and Tok Pisin, the other major lingua franca of Papua New Guinea since the Second World War, and in keeping with the changing social condi­ tions in an increasingly complex society. In 1975 it acquired equal status with Tok Pisin as one of two unofficial national languages.

Hiri Motu

35

Appendix Pidgin Features of Hiri Motu Compared with Those of Simplified Motu and Other Languages In the chart below, HM = Hiri Motu, SM = Simplified Motu, HTL(E) = Hiri Trading Language, Eleman variety, HTL(K) Hiri Trading Language, Koriki variety, and PPE

=

=

Papuan Pidgin English. Feature of HM

SM

HTL(E)

HTL(K)

PPE

l. No irregular verbs

Yes7 - most common ones like "corne, go" appear in simple form as in HM

Yes, no irregular verbs

Yes, no irregular verbs

Yes, no irregular verbs

2. "Have" and "have not" expressed by dekenai + noho orby POSS + N + noho

U ses both full form and short form ofMotu construetion with postposition mai "with"

Sometimes uses No verb used ita "with", sometimes amumuti "stay", but neither in negative sentenees

7- no evidence but presumably gat and stap as in Torres Straits Broken (of whieh PPE was an extension) and Bislama

3. Uses ese and be as subjeet mrkrs or foeus elements

No subjeet or foeus mrkrs (although evidenee very limited)

No subjeet mrkr No subjeet or mrkrs but foeus mrkr maro most eommonly used after subjeet

No subjeet or foeus marking elements

4. Restrieted dual pronoun form

Yes, same as in

HM

7 (free pronoun + oraore "two")

7 (free pronoun + 7 - no evidence rearea "two")

5. No bound

Yes, no bound forms; uses free pronouns taken from Eleman languages

Yes, no bound forms; uses free pronouns taken from Koriki

Yes, same as HM, exeept uses free pronouns derived from English

6. Transitive verbs Yes, same as in are marked by HM final-a

No, transiti ve verbs are not eonsistently marked

No, there is no transitive marker on verbs

No, although -im is often used in similar way

7. No noun classes Yes, same as in HM based on pos session (e.g. alienable vs. inalienable)

Yes, no noun classes but uses a relator enane derived from Eleman languages

Yes, no noun classes and no relators

Yes, same as in HM

Yes, same as in HM pronoun subjeets or objeets on verbs (exeept in Central dialeet); uses free pronouns derived from "true" Motu

Tom Dutton

36 8. Dekena(i) a generalized postposition

No, uses Motu postpositions

? - evidence patchy No, generalized No, uses along/long postposition used as generalized locabut when case is tive preposition marked there is for directions variation between suffix -ai « Motu) and preposed forms ta, ba and la « Eleman languages)

? - no evidence 9. Generalized forms for "like this, like that" and "how"

? - no evidence No, uses Eleman forms although these are similar in structure to HM ones, e.g. ma-feare (lit. this-like) "like this"

?- available 10. Restricted forms for "all" evidence suggests and "the whole" same as in HM

?- no evidence

? - no evidence

No, uses all and altugeta

? - no evidence

11. Restricted adjective agreement

? - no evidence

No, no adjective agreement at all

No, no adjective agreement at all

No, no adjective agreement at all

12. Simple imperative

Yes, same as in HM

Yes, but using Eleman material

Yes, but using Koriki material

Yes

13.Dohore and variants used before verb to indicate future

? - no evidence of ? - no tense No, but Koriki No, but baimbai such use in distinctions in material can be used in similar way available materials verbs, these being used in similar way indicated by although future tense is mostly indicated by adverbs of time such as aire "later, -varia on verbs afterwards", or vevere "tomorrow" for future; mamaro "now" for present. There is no evidence for past tense. If there is no adverb of time in sentence the tense can be read as past, present or future

14. Uses vadaeni, noho and gwauraia after verb to indicate completed, continuous, and intentive/inceptive aspect

No evidence of use of noho, but vada is used before verbs to indicate completed action (as in Motu) and gwauraia after verb (as in HM) to indicate intention

No, uses kaolaia ? - but -varia may after verbs for may be used for completed action intention and sinanaia (which is translation equivalent of gwauraia in HM) or mamaro after verbs for intention (which is translation equivalent of vadaeni in HM)

No? - but some evidence thatfinis is used in similar way for completive; evidence lacking for continuous .and inceptive aspects

Hiri Motu

1

15. Uses lasi as Yes, same as in generalized neg-HM: (a) lasi after ative: (a) after verbs; (b) lasi after verbs; (b) after adjectives adjectives 16.Reflexivity Yes, same as in expressed by HM sibona without special verb forms 17. Temporal ? - no evidence in clauses marked available materials. by negana(i) Juxtaposition the main relating device 18. Uses bema as conditional clause marker

? - no evidence in available materials. Juxta­ position the main relating device

19. Uses vadaeni Some use of "enough, okay' vadaeni similar to HM as sentence connective 20. Flexible word order depend­ ing on whether subjects and objects are

Yes, same as in HM: (a) lasi after verbs; (b) lasi after adjectives - no evidence

No,butpea is No, uses: (a) no be­ used in similar fore verbs; (b) no way: (a) pea after before adjectives verbs; (b) pea after adjectives ? - no evidence

■ no evidence

? - no evidence, Juxtaposition the main relating device

? - V + ane "and, No, uses juxtaposiwhen, if' or tion only juxtaposition used in available materials

- no evidence

No, uses V + ane No, although uses "and, when, if' sapos " i f in similar or juxtaposition way in available materials

No, although form tora "enough, okay" used in similar way

? invariable SOV SOV like Motu although no evidence for pronouns available

No, although form oaio "enough, okay" used in similar way

No, although finis "finished" and orait "okay" used in similar way sometimes

SOV

SVO

nouns or pronouns

Notes 1.

This paper is a condensed version of my book Police Motu: iena sivarai (1985).

2.

For descriptions of Motu see Lister-Turner & Clark (1930) and Taylor (1970). The standard variety is taken to be that described by Lister-Turner & Clark.

3.

Tok Pisin is the fastest growing language in Papua New Guinea. Until independence it was the principal lingua franca and unofficial language of administration of the northern half of the country, or that part of it that was known as the Territory of New Guinea. For detailed descriptions of it see Wurm & Mühlhäusler (1985) and references therein.

4.

For example, Chatterton (1950:5) says, "its origin dates back to before European penetra­ tion of Papua, when it was the trading language between the Motuans and their customers along the shores of the Gulf of Papua, to whom they bartered pots for sago."

38

Tom Dutton

5.

This was a movement led by Tommy Kabu. As a result of his wartime experiences Tommy Kabu felt that the only way to improve the living standard of his and other linguistically related peoples in the area was to adopt a single unifying language, Hiri Motu. At the time the population of the delta was estimated to be 6000, most, if not all, of whom belonged to the movement (Hitchcock & Oram 1967:5, 18).

6.

As will be indicated below, a form of Pidgin English now often referred to as Papuan Pidgin English following Mühlhäusler (1978) — but actually only an extension of the pidgin English formerly spoken in the Torres Straits and now creolized and referred to as broken (Shnukal 1988) — was spoken in some parts of Papua before Police Motu began to be disseminated.

7.

This is so mainly (a) because of its historical association with its parent Motu; (b) because this was the form of the language most similar to that propagated in early primers of the language; (c) because Central dialect speakers were the best educated (having been the first to be brought under mission and government control); and (d) because Central dialect speakers were readily available to act as translators and interpreters in Govern­ ment offices (on account of their distribution around Port Moresby, the admininstrative centre). Hence the impression usually given is that the Central dialect is the standard dialect, whereas there is no official standard dialect.

8.

The evidence for this claim is to be found in the various handbooks and language learning courses available on the language (e.g. Chatterton 1946 and others, Dutton & Voorhoeve 1974, Wurm & Harris 1963).

9.

Even so it was only with difficulty that Lawes was able to learn the true language, because many of the villagers were still opposed to imparting this knowledge to strangers, a position some of the older men maintained until the 1920s (Chatterton 1970:95).

10.

The use of the term "broken" English dates back to the very early days in Queensland. However, despite its name this language was really a pidgin language with features similar to those found in pidgin Englishes in the South-West Pacific and elsewhere today. In fact it is still referred to as Broken in the Torres Straits, where it has become creolized, as already noted.

11.

For example, Lawes notes in his journal on 4 April 1876 at Mailu (some 250 or more kilometers east of Port Moresby) that "as most of them know a little of the Port Moresby dialect I find I can communicate with them pretty freely", and again in June of the same year at Hula (about 100 kilometers east of Port Moresby) he notes, "most of the Hula natives understand more or less of the Port Moresby dialect".

12.

I say "presumably" here because without any evidence of the language learning abilities of these foreigners it is not possible to be more precise. Yet we know that individuals in general do display different language learning capabilities, and that level of education is not necessarily an indication of that ability. Thus, for example, J.H.P. Murray, the renowned Lieutenant-Governor of Papua for thirty-three years, admitted to attempting to learn "true" Motu but did not get very far with it. "It is said to be easy," he wrote in a letter to his brother George on 14th March 1904, "but for some reason or other I cannot get the hang of it" (West 1970:35). This lack of success annoyed him, the more so because of his knowledge of Latin, Greek, French and German — he was a judge and classical scholar. By way of contrast, most Papua New Guineans are excellent language learners and many are multilingual without ever having been to school.

Hiri Motu

39

13.

Bau and Kadavu are dialects of Eastern Fijian, which is generally distinguished from Western Fijian or Wayan.

14.

Hornbills are not native to Fiji.

15.

In general these languages are similar to Motu in having open syllables, five-vowel systems, and consonant systems that differ from it mainly in the number of distinctions made at the alveolar point of articulation. Probably the most extreme cases are to be found in some languages of the Gulf of Papua which do not have contrasts between stops, liquids, and/or nasals. As a result common Motu words like lau "I" or lao "go", dala "road", vanagi "canoe", and lasi "no, not" which contain some of these sounds are generally pronounced as , dana, maragi, and nasi respectively.

16.

These tendencies are apparently what has led some observers to claim that "it seems that there is some canon of "correctness" to which good speakers try to conform even when this involves distinctions which are not part of their native linguistic habits" (Wurm & Harris 1963:1). It has long been suggested by some that Hiri Motu, unlike Motu, has an /s/ phoneme. This latter claim must, however, be rejected as not substantiated by the data. Traditionally Motu had no /s/ phoneme — [s] was merely a distributional allophone of /t/ — although s was written by Lawes in reducing the language to writing and has become established. However, with the increasing familiarity with spoken and written English, from which many words are now borrowed into Motu, /s/ has become an added phoneme — e.g., sobea "to survey", so "saw, show", Sabati "Sabbath". The same has happened in Hiri Motu.

17.

The Simplified Motu features are based on those enumerated and discussed in Taylor (1978) and Dutton (1986) and those for Hiri Motu in Dutton (1985:8-16), which, for space limitation reasons, cannot be included here.

18.

Any other assumption is much more complex and therefore less preferable. For example, an obvious alternative is that Hiri Motu acquired its features independently of Simplified Motu and that both accidentally ended up with similar features because of similar social and linguistic forces acting on them.

19.

oth dekena(i) and noho come from Motu although dekena(i) in Hiri Motu is a generalized form of Motu deke + pronoun + ai, SL combination used to indicate nearness to persons, e.g. lau deke-gu-ai (lit. I towards-me-at) "near me" or "beside me". In Hiri Motu dekenaii) does not vary for persons and is used to express a wide range of locational meanings (such as "to, at, from") as well as the instrumental meaning "with".

20.

The most common way of expressing "have" in Bislama and Broken is by using the verb gat (< English got), but the following construction using stap is still very common, especially when the speaker is referring to an object that is on or about his/her person at the time of speaking.

21.

Thus, for example, Lock (n.d. : 3) noted that "nearly all the different tribes have their own formation of Police Motu" and (p.l) "the language takes its name from the Native Police Force whose members picked up more Motu words and added them to the Bastard language".

22.

See my 1980 article (fn. 5, pp. 194-95) for a discussion of the origin of nakimi and my 1987 article for an account of the origin of kiki.

40

Tom Dutton

References Brett, R., et al. 1962. A survey of Motu and Police Motu. Port Moresby: Department of Information and Extension Services. Chatterton, Percy. 1946. A primer of Police Motu. Port Moresby: The Department of Education. Chatterton, Percy. 1950. A Primer of Police Motu. Cairns: Cairns Post. Chatterton, Percy. n.d. (but  1968). A primer of Police Motu. Sydney: Pacific Publica­ tions. Chatterton, Percy. 1970. "The origin and development of Police Motu". Kivung 3.95-98. Chatterton, Percy. 1971. Hiri Motu {Police Motu). Port Moresby: UFM Press. Chatterton, Percy. 1975. Say it in Motu: An introduction to the common language of Papua. Sydney: Pacific Publications. Dutton, .. 1969. The peopling of Central Papua: Some preliminary observations. Canberra: Pacific Linguistics B-9. Dutton, .. 1975. "A Koita grammar sketch and vocabulary". Studies in languages of Central and South-East Papua, ed. by T. E. Dutton, 281-412. Canberra: Pacific Linguistics C-29. Dutton, Tom. 1980. "A Note on the Origin of the Motu term hirï". Kivung 12/2: 189-206. Dutton, Tom. 1985. Police Motu: iena sivarai [= its story]. Port Moresby: University of Papua New Guinea Press. Dutton, Tom. 1986. "More on Simplified Motu". Language in global perspective: Papers in honour of the 50th anniversary of the Summer Institute of Linguistics, 1935-1985, ed. by Benjamin F. Elson, 525-48. Dallas, Texas: The Summer Institute of Linguis­ tics. Dutton, Tom. 1987. "The Origin of kiki 'to tell a story, yarn' in Police (now Hiri) Motu". Language and Linguistics in Melanesia 15(1-2). 112-24. Dutton, .., and C.L.Voorhoeve. 1974. Beginning Hiri Motu. Canberra: Pacific Linguis­ tics D-24. Hitchcock, N.E., and N.D. Oram. 1967. Rabia Camp: A Port Moresby migrant settlement {= New Guinea Research Bulletin, No. 14). Canberra: The Australian National University. Lawes, W.G. 1896. Grammar and vocabulary of the language spoken by the Motu tribe, New Guinea. (Third and enlarged edition.) Sydney: Government Printer. Lister-Turner, R., and J.B. Clark. 1930. Revised Motu grammar and vocabulary. Port Moresby: Government Printer. Lock, Maynard. n.d.(but  1942). Police Motu. Unpublished MS held in the Australian War Memorial, Canberra. Mühlhäusler, Peter. 1978. "Papuan Pidgin English rediscovered". Second International Conference on Austronesian Linguistics: Proceedings, ed. by S.A.Wurm & Lois Carrington, 1377-1446. Canberra: Pacific Linguistics C-61. Shnukal, Anna. 1988. Broken: An introduction to the creole language of the Torres Strait. Canberra: Pacific Linguistics C-107. Taylor, A.J. 1970. Syntax and phonology of Motu {Papua): A transformational approach. Canberra: The Australian National University dissertation.

Hiri Motu

41

Taylor, A.J. 1978. "Evidence of a Pidgin Motu in the earliest written Motu materials". Second International Conference on Austronesian Linguistics: Proceedings, ed. by S.A.Wurm & Lois Carrington, 1325-60. Canberra: Pacific Linguistics C-40. West, Francis. 1970. Selected letters of Hubert Murray. Melbourne: Oxford University Press. Wurm, Stephen A. 1973. "The Kiwaian language family". The linguistic situation in the Gulf District and adjacent areas, Papua New Guinea, ed. by Karl Franklin, 219260.Canberra: Pacific Linguistics C-26. Wurm, S.A., and J.B. Harris. 1963. Police Motu: An introduction to the trade language of Papua (New Guinea) for anthropologists and other fieldworkers. Canberra: Pacific Linguistics B-l. Wurm, Stephen A., and P. Mühlhäusler, eds. 1985. Handbook of Tok Pisin (New Guinea Pidgin). Canberra: Pacific Linguistics C-70.

Pidgin Delaware Ives Goddard Smithsonian

1.

Institution

Introduction

Pidgin Delaware was used as a contact language between Indians and Euro­ peans on the Middle Atlantic Coast of North America during the Colonial Period. The Indians involved in its use were, in all known cases, speakers of varieties of Delaware, a grouping of two contiguous and closely related Eastern Algonquian languages, Munsee and Unami, each of which comprised a dialect continuum. Munsee was the language of the central and lower Hudson River Valley, westernmost Long Island, the upper Delaware River Valley, and the northern third of New Jersey; Unami was spoken to the south of Munsee in the Delaware River Valley and New Jersey.1 The European users were Dutch, Swedes, and English of the colonies of New Netherland, New Sweden, Pennsylvania, and New Jersey.2 Pidgin Delaware developed in the 1620s, and almost all documentation of it dates from the seventeenth century. It seems unlikely that, even at the height of its use, it would have been known by more than a few thousand people. Pidgin Delaware must have gradually gone out of use during the eighteenth century, although there is little documentation from this period. Presumably it was functionally displaced by American Indian Pidgin English, English, in some local situations other European languages, and, for a few non-Indians, Delaware. As late as 1785, however, the Delaware learned by Whites on the frontier contained Pidgin Delaware features. Pidgin Delaware was a pidginized form of Unami. Nearly all of its attested vocabulary has clear origins in Unami words and expressions. The dialects of Unami were most noticeably differentiated by phonological inno­ vations in Southern Unami that, under various conditions, dropped w, synco-

44

Ives Goddard

pated vowels, and deleted initial syllables. The Pidgin Delaware materials exhibit phonological features of both Northern and Southern Unami, but some words have only a Southern Unami shape or only a Northern Unami shape in all attestations. There are two major sources for Pidgin Delaware. The most extensive is from the Swedish Lutheran minister Johannes (or Johan) Campanius (b. 1601, d. 1683), who compiled a vocabulary of words and phrases in Pidgin Dela­ ware during his service in New Sweden, 1642-1648, and translated into it an adapted version of Martin Luther's Little Catechism, which he completed in Sweden in 1656. A small book containing the Catechism and vocabulary was published in 1696 for a Swedish Lutheran mission sent out the following year to the former Swedish colony, where it is reported to have been of some use in preaching to local Indians (Campanius 1696; Collijn 1937; Campanius 1938). The second major source is an anonymous vocabulary titled "The Indian Interpreter" compiled in West New Jersey in the late seventeenth century and surviving in a book of land records from Salem County, New Jersey (Anony­ mous 1684).3 Other sources for Pidgin Delaware include a few words and phrases in Lindeström's Geographia Americœ (Lindeström 1925:203; written in 1691 from notes made 1654-1656) and in a 1683 report by William Penn (Myers 1937:26-28), a few lines of dialogue and some numbers published by Gabriel Thomas (1698, 1:47, 2:7-13), and scattered words in various early records. It will be argued below (§8) that two other sources contain Pidgin Delaware and are in fact respectively the first and last documentations of it: an ostensible Northern Unami vocabulary from below the Falls of the Delaware at Trenton (de Laet 1633:75) and a Southern Unami vocabulary recorded among Dela­ ware emigrants in western Pennsylvania in 1785 (Denny 1860:478-481).4 This paper first surveys selected features of the phonology (§2) and morphosyntax (§3) of Unami and then describes the corresponding features of Pidgin Delaware — the phonology, as reflected in the different sources (§4), and the grammatical categories and word order (§5). Readers may wish to read the two phonology sections (§2 and §4) and the two sections on grammar (§3 and §5) in sequence. The sources of the Pidgin Delaware lexicon are examined in two sections, one treating more or less straightfor­ ward adaptations from Unami Delaware (§6) and one treating words that have idiosyncratic and miscellaneous derivations (§7). Agreement on these words in the various sources is considered as part of the evidence for the status of

Pidgin Delaware

45

Pidgin Delaware as a language, together with some aspects of Pidgin Dela­ ware vocabulary that shed light on the cultural context of its development. A final section summarizes the external history of this language (§8).

2.

Delaware phonology

The phonemes of Southern Unami (SU), as exemplified by twentieth-century Oklahoma Delaware, are displayed in Table l. 5 Some of the differences between Southern Unami and Northern Unami are relevant to understanding the sources of Pidgin Delaware. The Southern Unami vowels ɔ and ɔ reflect earlier wa and wa-, attested in Northern Unami (NU) and Munsee (Mu);6 in the corresponding sequences in Northern Unami the w is retained but the vowels are apparently rounded to [ɔ] and [ɔ] (generally written (o) and (oa) by the Moravians [e.g. Luckenbach 1838: 33]).7 (1)

SU -k 'and' (Table 2: 'three'). 23 One word consistently has (32; 35, 36), (tz) (33; 37, 38), and (in some prevocalic occurrences) ((s)si) (33; 39).24 Even š before low or back vowels, which might be expected to be readily transcribed with Swedish (si), Campanius usually fails to write distinctly (Table 2: 'two', 'seven'; 37). (32)  (142) (sijs) 'more'; cf. SU čí.č. (33)  (137) (dakótze) 'cold'; cf. SU ntákɔhči 'I am cold'. 25 (34)  (140) (nissiaanus) 'child'; cf. SU *nničá.əs 'my child'.26 (35)  (3) (suck)'but'; cf. SU šúkw. (36)  (136, 137) (chíssogh) 'sun'; cf. SU kíšox. (37)  (146) (tzackamææs) 'eel', L (syackameck);  (sehacameck);27 cf. SU šɔxamekw. (38)  (142) (étzkans) 'needle'; cf. SU éškans,

éskanš.

(39)  (5-6) (mosiuttamen) 'bring down' (cf. 120, 199, note 107); cf. SU nəməš.ó'tamən 'I hit it with a shot'. In a few words in which Campanius desired to write prevocalic š he eschewed Swedish (si) and used German (sch) (40, 41, 42, 262), a spelling that can also represent sk or sk (Table 2: 'nine').

Pidgin Delaware

53

(40)  (136) (scháckhan), (144) (sáckhang) 'wind'; cf. SU kSàx.ə 'the wind blows'. (41)  (150) (scháta), (siaata) 'tobacco', II (shauta); cf. SU kwšá.tay. (42)  (149) (schameera), (schaméra) 'run',  (shamahala); cf. SU ksa.méhəle. 'he runs'. 28 In two cases (ch) before (i) is used corresponding to č (43, 246); this spelling probably reflects the affricated pronunciation of Swedish (k) before front vowels, which must have been familiar to Campanius even if it was not his own (Haugen 1976:268-70). (43)  (143) (chirῶkha) 'leather strap'; cf. SU či . ló . he . s. 29 One example seems to have (ch) before (e) for š: (44)  (37, 144) (tackcheene) 'open, open the door'; cf. SU tunksé.ni 'open it'. The nearly complete overlap in the spellings of č and š in the Campanius and de Laet vocabularies suggests that even when they were distinguished from s they were not clearly distinguished from each other. (45)  (139) (siææt) 'tendons', L (cheet); cf. SU wčé .t 'sinew'. (46)  (138) (siættæn) 'lips', L (chettoen); cf. SU wsé.t.o.n. There is some variation in the forms recorded by Campanius that sug­ gests that he drew on information from different sources. He gives the variants of some words side by side in the same entry (41, 47, 48, 49), and some variants differ more than seems likely for renderings of the same pronunciation. (47)  (138) (alænskan), (laénskan) 'hand', 30 L (rinskan) 'fingers'; cf. NU (linschgan) [-lənčkan] 'finger' (Brinton & Anthony 1889: 64). (48)  (147) (homijmi), (mijmi) 'passenger pigeons', 31 L (mymy); cf. SU amí.mi, Mu mí.mi.w.32 (49)  (150) (jaésskung), (jésquem) 'corn'; cf. SU xáskwi-m.33 In at least one case dialect variants (cf. 7-10) have been borrowed with differentiated meanings.

Ives Goddard

54

(50)  (150) (assísko) 'black or soft stone', 34 (145) (sísko) 'clay'; cf. NU (assisku) [ahsí-sku] 'mud, clay' (Brinton & Anthony 1889: 23), SU sí-sku 'mud' (SU ɔpahsí.sku 'clay', lit. 'white mud'). Initials beginning with SU ɔ. and NU [wo-] (1) appear in Pidgin Delaware with or without the initial w; stems made with Iwa-p-l 'white' are attested in both forms. (51)  (142) (wopæk) 'white'; 35 cf. SU  . p . e . k '(inan.) which is white', Mu wá.pe.k. (52)  (142) (opijmi) 'chestnuts'; cf. SU 5.p.i.m, Mu wa.pi.m. Corresponding to modem Unami / Campanius usually writes (r) (Table 2: 'five', 'ten'), but he has (1) consistently in some words (47) and writes a few words both ways. (53)  (146) (tulpa), (turpa) 'tortoise'; cf. SU tó.lpe 'water turtle (sp.)', NU (tulpe) 'sea turtle' (Zeisberger 1887:206). (54)  (138) (hwijl), (wijr) 'head', II (wheel) "the head"; cf. SU wíl '(his) head'. Some words have an (h) added before an initial vowel (48, 55), though in most cases initial vowels are kept (52) and initial (h) represents Unami h (56) or x (Table 2: 'eight'). 36 (55)  (145) (hóckus) 'fox'; cf. SU .k.wds, some speakers (cf. 63).

h.k.wəs

(56)  (8) (hirússus) 'old man'; cf. SU hiló.s.ds. Certainly, different speakers seem to be represented in these variants, and very likely, in some cases, different dialects of Unami. The transcription in "The Indian Interpreter" has both English and nonEnglish features; in fact, some words appear as doublets, with both kinds of transcription (57, 60, 67). (57)  (dogwatcha), (taquatse) 'cold' (cf. 33). Some of the non-English features could have been acquired orally by an English speaker learning Pidgin from a Swede or a Dutchman. For example,

Pidgin Delaware

55

some words have (s) for Unami š or č, a highly unlikely substitution for an English speaker but a common one in the Dutch and Swedish sources. (58)  (sinques) 'wildcat',  (145) (síngwæs), (singues), L (synquoy); cf. NU (uchtschìnque) [wcínkwe.(w)], (tschínque) [čínkwe.(w)] (Zeisberger 1887:33,231),°SU čínkwe. 37 (59) II (sickenom) 'turkey',  (147) (síckenem), L (sickenum), T (chekenip);38 cf. SU čí.k.əə, NU (tschìkenum) (Zeisberger 1887:206). For some of these there are also spellings using English conventions that render these sounds correctly. (60) II (singa) 'when',  (152) (singa), II (chingo) 'when'; cf. SU cinke. Assuming that the numbers in "The Indian Interpreter" (Table 2) were learned orally from a Swede or Dutchman would explain why Unami š appears as II (sh) before vowels, where [š] would be possible in Swedish or Dutch, but as II (s) word-finally or before a consonant, where it does not occur. The assumption of learning from Swedish speakers would also explain why the Pidgin reflex of SU ní.s.a 'two' has a short (i) in the English sources; the shortening would reflect the Swedish syllable-balance rule (note 21), which typically results in the shortening of long vowels before long conso­ nants in Campanius's recordings of Pidgin. Other variations in the way Unami sounds are reflected in 'The Indian Interpreter" include, as in Campanius, the use of both (r) and (1) and the sporadic appearance of (h) before initial vowels. (61) II (renus) 'man',  (19) (rhénus), L(renoes), II(leno) 'man'; cf. SU l. 39 (62) II (hapitcha) "by and by",  (4) (apítzi) 'right away', T (a peechi); cf. NU (apitschi) 'after a little while, by and by' (Brinton & Anthony 1889:22). (63) II (hoccus) 'fox'; cf. (55). The phoneme x is written like h (Table 2: 'five', 'eight') or  (Table 2: 'three').

56

Ives Goddard

Some words in "The Indian Interpreter" exhibit distinctively Swedish spelling conventions, in some cases strikingly similar to spellings used by Campanius (cf. also 55 and 63, 60). (64) II (hwissamech) 'catfish',  (146) (hwíssamick); cf. SU wí.s.ame.kw, NU (wísameek) (Zeisberger 1887:34). (65)  (hwissase) 'afraid',  (23) (hwisásse); cf. SU wi.s.á.s.u afraid'.40

'he is

(66) II (twn) 'mouth',  (138) (tῶn); cf. SU wtó.n '(his) mouth'. (67)  (mwes) 'elk',  (6) (moos) 'animal', T (moos) 'cows'; cf. SU mos 'elk', NU (moos) 'elk; cow' (Zeisberger 1887:48, 65); 41 cf. II (muse) 'cow'. (68)  (iwse) 'meat, flesh',  (48, 138) (ῶs) 'meat, flesh, body', T (youse) 'venison'; cf. SU wiyo.s 'meat'. It seems improbable that it is only by chance that these spellings match so closely ones used by Campanius. In particular, the use of (hw-) as a lexically determined orthographic equivalent of (w-) is a spelling convention only in Swedish (and perhaps cf. 54). 42 Since "The Indian Interpreter" was entered into the Salem land records before Campanius's vocabulary appeared in print, it is likely that a written Swedish wordlist with some relation to Campanius's work was one of its ultimate sources. The extent to which other non-English features in "The Indian Interpreter" (58, 59, 60) may derive from written rather than oral Swedish sources is inherently difficult to estimate, however. In contrast to the earlier sources, Gabriel Thomas's recordings of Pidgin Delaware are internally more consistent and seem to be original. English spellings for the vowels, with their inherent ambiguity, appear to be universal. The digraph (ch) is used not only for č but also to render x in most words (Table 2: 'three', 'five', 'eight'). (69) T (chase) 'skin(s)',  (145) (hææs), II (hayes) 'an undressed skin'; cf. SU xé.s 'skin'. (70) T (marchkec) 'red',  (142) (mæckhhchæck) 'red' (cf. note 35). This use of (ch) must reflect a knowledge of Dutch, German, or Welsh. The spelling (ch) is also used for preconsonantal h.

Pidgin Delaware

57

(71) T (ochqueon) 'matchcoats',  (142) (äquijvan), (quijvan) 'cloth, article of cloth clothing', II (aquewan) 'coat, woolen clothing'; cf. SU ahkwí.'blanket', NU (achquíwannes) [ahkwí.wanəs] (Zeisberger 1887:24). In other words Thomas writes (h) for x and omits preconsonantal h. (72) T (hayly) 'much', 43  (ahalea); cf. SU xahé.li,44 NU (mcheli) (Brinton & Anthony 1889:49). (73) T (andogowa), (undogwa) 'yonder',  (149) (úndaque), II (undoque); cf. SU wntahkwi 'direction' (used with deictics). Thomas writes (1) in five words (e.g. 74) and two personal names and (r) in two words (e.g. 75) and the placename Arwaymouse. Initial (h) is added in two words (74, 75), one occurring as a doublet without it. (74) T (halopa) 'tomorrow', II (alloppau), (aloppau); cf. SU (some speakers haláp.a).

aláp.a

(75) T (orit), (horit) 'good',  (13) (urítt), (21) (hurít), (1) (hωrítt), II (olet), (Denny 1860:479, 480, 481) (wellett), (willet) (268, 269); cf. SU wəlt 'it is good'.

5.

Pidgin Delaware grammatical categories and word order

Pidgin Delaware makes essentially no use of the extensive inflectional mor­ phology of Unami. Instead, what few grammatical categories it has are indicated lexically and syntactically. To be sure, there are traces of many Unami inflectional morphemes frozen in invariant Pidgin Delaware words. But even in the few cases in which Pidgin Delaware attests more than one inflectional form made on the same Unami stem, it appears likely that the Pidgin forms are descriptively unanalyzable units. The pronominal categories are reduced to first, second, and third person, undifferentiated for gender or number, and are marked syntactically by independent words. There appears to be no indication of the contrast between singular and plural. The three most generally used pronouns reflect the first and second singular of the Unami emphatic pronoun set and an emphatic form of the inanimate singular proximal deictic.

58

Ives Goddard (76) First person:  (2) (nijr),  (2) (nijre); II alen "rain", (English or Dutch) rat > alata "rat", rice (or Dutch rijst) > alisi "rice". Likewise, some Ndyuka personal names appear to alternate between forms with and without initial a, e.g. Puyenke ~ Apuyenke.

19.

Smith (1987:150) suggests that Ndyuka and Sranan split around the first decade of the eighteenth century.

20.

Dipi is a term that is commonly used in Suriname, including in Sranan as spoken by various ethnic groups, to refer to language phenomena that are relatively difficult to understand, and hence all the more mysterious. Hindi of India, for example, is dipi compared to Sarnami Hindi spoken in Suriname; Sranan speakers consider Saramaccan more dipi than Ndyuka, with which it shares a much higher number of cognates.

21.

The corresponding form in modern N, baiba, means "chin", while Sranan barba means "beard".

22.

De Goeje (1908a:205) suggests that this form, mekolo, comes from negro, but "by what route cannot be determined". He gives mékolo as the Wayana form for (Bush)negro (1946:147), and mékolo as one of three T forms for a Ndyuka (1906:74).

23.

De Goeje (1908a:205) suggests Trio ekelama "buy" as a possible source of pidgin atemete. Crevaux (1882b:39) gives arépété for "bagages, objets d'échanges".

24.

The N-derived form is pikinini; the other form, kami, is presumably from T kami "term of address for younger sibling or grandchild" (Leavitt 1971:61-62). But cf. T sikami "boy" (de Goeje 1906:73) and W kahmi "child, term of address" (de Goeje 1946:145).

25.

Called "Ethnographie" in the original list presented by de Goeje (1908a:205).

26.

De Goeje (1980a:206) considers this form, kula, to have come into the pidgin from "Negerengelsch" (which can be either N or Sranan in his scheme), which in turn borrowed it from Kalinha. But since there is a likely kiKongo source for this word in N (see Huttar 1985:59), we assume that Kalinha was the borrowing language rather than the source language. With regard to the pidgin, we assume for simplicity's sake that the pidgin got the form directly from N.

122

George L. Ruttar and Frank J. Velantie

27.

De Goeje says (1908a:206) that this form, kwama, is from Kalinha; but it is also found in N.

28.

The form is kawalimbo. Cf.  kaipïrïmbo, from ka:pïrï "make" and -mbo "former" (Hoff 1968:137). Kawembo also occurs alongside gron as an alternative form for "garden".

29.

The form from N, , may in turn derive from Arawak directly or via Portuguese.

30.

"Morning/tomorrow" is toko-toko, which de Goeje (1908a:207) suggests derives from T or N tokro, a quail (Odontophorus guianensis) whose call is heard shortly before daybreak. The modern N word is tokoo; Ndyukas also consider this bird an announcer of daybreak.

31.

Kaikushi; cf.  kaiku:si, "jaguar", T kaikui "jaguar, dog". For W, de Goeje (1946:160) gives kaikui for both meanings, but de Goeje (1909:293) gives both kaikúsi and kaikui for "jaguar". De Goeje (1946:248) gives kaikushi for "jaguar" in Triometesem (Wayaricule), a Cariban language of southeastern Suriname and southwestern French Guiana.

32.

De Goeje (1908a:209) gives T joti "game" and W woto "fish" as possible sources of pidgin oto "game, meat".

33.

Sama is presumably from N sáma "who?"; cf. Ndyuka sama "person" (with low tone on both syllables, rather than the high-low tone sequence of sáma), Sranan suma "who?".

34.

Sungu means "sink" in N, but de Goeje gives "dronken" ("drunk"), not "verdronken" ("drowned"), as the meaning of the form in the pidgin.

35.

Komopo, from Sranan komopo "come from, come out of' (cf. N komoto).

36.

De Goeje (1906) gives ñami, de Goeje (1908a) miang or miáng, and our own fieldwork nyan; cf. N nyan (with N alternant nyami).

37.

Krutu,  Sranan krutu "hold council, criticize" (cf. Ndyuka kuutu).

References Aloema, Nardo, M. J. Pierre, and . N. van der Ziel. 1987. Kalinha — Nederlands Woordenboek. Paramaribo: Instituut voor Taalwetenschap. de Bruijning, C. F. A., and J. Voorhoeve. 1977. Encyclopedie van Suriname. Amsterdam: Elsevier. van Coll, C. 1903. "Gegevens over Land en Volk van Suriname". Bijdragen tot de Taal-, Land- en Volkenkunde van Nederlandsch-Indie, 55.451-650. Coudreau, Henri. 1895. Chez Nos Indiens: Quatre Années dans la Guayane Française (1887-1891). Paris: Librairie Hachette et Cie. Crevaux, J. 1882a. "Vocabulaire Français-Roucouyenne". Grammaires et vocabulaires roucouyenne, arrouague, piapoco et d'autres langues de la région des Guyanes, ed. by J. Crevaux, P. Sagot, and L. Adam, 1-20. Paris: Maisonneuve et Cie, Libraireséditeurs. Crevaux, J. 1882b. "Quelques Mots de la Langue des Indiens Trios". Grammaires et vocabulaires roucouyenne, arrouague, piapoco et d'autres langues de la région des

Ndyuka-Trio Pidgin

123

Guyanes, ed. by J. Crevaux, P. Sagot, and L. Adam, 39-40. Paris: Maisonneuve et Cic, Libraires-éditeurs. Crevaux, J. 1883. Voyages dans l'Amérique du Sud. Paris: Librairie Hachette et Cie. Franssen Herderschee, A. 1905a. Verslag van de Gonini-Expeditie. Leiden: E. J. Brill. (Originally in Tijdschrift van het Koninklijk Nederlandsch Aardrijkskundig Genoot­ schap, 1905.1-174.) Franssen Herderschee, A. 1905b. Verslag der Tapanahoni-Expeditie. Leiden: E. J. Brill. (Originally in Tijdschrift van het Koninklijk Nederlandsch Aardrijkskundig Genoot­ schap, 1905.847-1032.) de Goeje, C. H. 1906. Bijdrage tot de Ethnographie der Surinaamsche Indianen. Leiden: E. J. Brill. de Goeje, C. H. 1908a. Verslag der Toemoekhoemak-expeditie. Leiden: E. J. Brill. de Goeje, C. H. 1908b. Beiträge zur Völkerkunde von Surinam. Leiden: E. J. Brill. de Goeje, C. H. 1909. Etudes Linguistiques Caraïbes. Amsterdam: Johannes Müller. de Goeje, C. H. 1939. "Bij primitieve volken". De Gids, pp. 1-31. de Goeje, C. H. 1946. Etudes linguistiques Caribes, vol. 2. (= Verhandelingen der Koninklijke Nederlandsche Akademie van Wetenschappen, Afd. Letterkunde, New Series, Part 49, No. 2.) Amsterdam: N. V. Noord-Hollandsche Uitgevers Maat­ schappij . Green, Edward Crocker. 1974. The Matawai Maroons: An acculturating Afro-American society. Washington, DC: Catholic University of America dissertation. Grimes, Joseph E., ed. 1972. Languages of the Guianas. Norman, OK: Summer Institute of Linguistics. Hoff, B. J. 1968. The Carib language. The Hague: Martinus Nijhoff. Huttar, George L. 1982. A Creole-Amerindian pidgin of Suriname. Society for Caribbean Linguistics Occasional Paper No. 15. Huttar, George L. 1985. "Sources of Ndyuka African vocabulary". De Nieuwe WestIndische Gids 59.45-71. Huttar, George L. 1987. "The Afaka script: An indigenous creole syllabary". The Thirteenth LACUS forum 1986, ed. by Ilah Fleming, 167-77. Lake Bluff, IL: Linguis­ tic Association of Canada and the United States. Huttar, George L. 1988. Notes on Kwinti. Society for Caribbean Linguistics Occasional Paper, No. 19. Huttar, George L. 1989. "The Portuguese contribution to the Ndyuka lexicon". Estudios sobre Español de América y lingüística Afroamericana: ponencias presentadas en el 45 Congreso Internacional de Americanistas {Bogotá, julio de 1985), 263-79. Bogotá: Instituto Caro y Cuervo. Huttar, George L., and Mary L. Huttar. 1972. "Notes on Djuka phonology". Grimes 1972.1-11. Huttar, George L., and Mary L. Huttar. 1994. Ndyuka. London & New York: Routledge. Jackson, Walter S. 1972. "A Wayana grammar". Grimes 1972.47-77. Jones, Morgan W. 1972. "Trio phonology". Grimes 1972.42-46. Kappler, August. 1887. Surinam: Sein Land, seine Natur, Bevölkerung und seine KulturVerhältnisse mit Bezug auf Kolonisation. Stuttgart: J. G. Cotta'sche Verlagsbuch­ handlung.

124

George L. Ruttar and Frank J.

Velantie

Leavitt, Claude W. 1971. Trio language analysis. Alalapadoe, Surinam: Stencil. Mosis, André, and Ben Scholtens. 1988. "Bosnegers en Overheid: Ontwikkeling van de politieke verhoudingen, 1650-1988 (4): Vernieuwing van de vredesverdragen in de 19e eeuw". De Ware Tijd, 21 October 1988, 3. Neumann, Peter. 1967. Wirtschaft und materielle Kultur der Buschneger Surinames: Ein Beitrag zur Erforschung afroamerikanischer Probleme. (= Abhandlungen und Berich­ te des Staatlichen Museums für Völkerkunde Dresden, Band 26.) Berlin: AkademieVerlag. Sack, Albert von. 1821. Beschreibung einer Reise nach Surinam und des Aufenthaltes daselbst in den Jahren 1810, 1811, 1812, so wie von des Verfassers Rückkehr nach Europa über Nord-Amerika, Second Part. Berlin: Haude & Spener. Smith, Norval S. H. 1987. The genesis of the creole languages of Surinam. Amsterdam: University of Amsterdam dissertation. Thoden van Velzen, H. U. E., and W. van Wetering. 1988. The great father and the danger: Religious cults, material forces, and collective fantasies in the world of the Surinamese Maroons. Dordrecht: Foris Publications.

Arabic-based pidgins and creoles Jonathan Owens Universität

1.

Bayreuth

Introduction: Social and historical background

In this paper I will discuss all the varieties of pidgin and creole Arabic known to date.* In §1 I will give the historical background to the development of Arabic-based /'s in the two periods and areas where they are documented, the early Islamic period (ca. 700-1000 A.D.) and the 19th-century southern Sudan and Chad. In §2 I survey briefly the / varieties, and in §3 I explore in some detail the social and demographic factors which led to the emergence of the Arabic-based /'s in the southern Sudan, Chad, and East Africa. Section 4 provides a linguistic sketch of one creole Arabic variety, Kenyan Nubi, and a discussion of the linguistic interrelations among the different varieties of Sudanic / Arabic, Sudanic colloquial Arabic, and substratai languages. 1.1. Classical Arabic and Arabic dialects It is well known that Classical Arabic (CA, attested between ca. 550-800) differs considerably from the modern Arabic dialects. That varieties sepa­ rated by over 1000 years should differ is hardly cause for discussion. What is not so clear is how the differences arose. There is agreement on some basic points. First, modern dialects differ considerably among themselves and in very many respects cannot be historically derived from CA via traditional comparative methods. Second, the modern dialects are phonologically and morphologically less complex than CA. And third, CA was but one variety of Old Arabic existing alongside Old Arabic dialects (Rabin 1951).

126

Jonathan Owens

At this point opinion splits. One school holds that CA was a variety spoken by Bedouin tribes (Fiick 1950). A second sees CA as an intertribal koine of some sort (Spitaler 1953, Corriente 1976), which perhaps was originally a functionally-restricted variety used in poetry (Zwettler 1978); it was not the language of everyday use. If one adopts the second viewpoint (which I personally favor), the modern dialects can be seen to a large degree as a continuation of old Arabic dialects, whose character is inferrable in part by the comparative method. If one adopts the first viewpoint, that CA was the spoken variety of the Bedouin Arabs, the question to be answered is how the dialects came to diverge from CA. One answer invokes the role of a / variety. Versteegh (1984) proposes that pidginization/creolization played a key role in the development of early Arabic dialects. He notes that as the conquer­ ing Arabs spread over a large area of the Mediterranean and the Near East, they came into contact with non-Arabs — Persians, Copts in Egypt, Berbers in North Africa, etc. — and that these peoples had to learn Arabic. Since, however, Arabs were generally a minority in the newly-conquered territories, the subordinate populations did not have adequate opportunity to properly learn Arabic, and hence used language-learning strategies that produced intermediate pidginized varieties. The modern dialects derive ultimately from this pidginization process, and in the end the Arabs themselves, at least'the city dwellers, took over this variety as their own (Versteegh 1984:69). There are at least three problems with Versteegh's thesis taken in its strongest form (cf. below), that the modern dialects at some point have pidgin ancestors. First, although Arabic dialects are structurally simpler than CA, the differences between them and CA are far less than what one finds between modern-day Arabic creoles and modern dialects (see §4.4). This is also true, apparently, for earlier times (ca. 1050). A brief fragment of an early Pidgin Arabic variety discussed by Thomason & Elgibali (1986) deviates almost as much from a modern Arabic dialect as it does from CA (see §2.1). If pidginization/creolization played a significant role, then one might expect to find a greater divergence between CA and Arabic dialects.1 Secondly, /'s generally arise in areas with a high degree of linguistic heterogeneity; but in most conquered areas there were only one or two languages other than Arabic (Aramaic and Greek in the Levant, Coptic in Egypt). Moreover, as Goodman (1986) points out, the Arabic language often penetrated into lands, such as the Levant, well before the areas had been

Arabic-based Pidgins and Creoles

127

brought under Arabic political control. These factors would have favored the growth of bilingualism more than pidginization/creolization. Thirdly, Versteegh's thesis, entailing that the Arabs gave up their own "high" variety in favor of the low one which the conquered peoples used, and made it their own, is not what one usually finds in pidginization/creolization situations, though Versteegh suggests a similar process in the spread of Vulgar Latin in the Roman Empire. One could, perhaps, ask for well-docu­ mented parallels to substantiate his suggestions. It should be noted that by pidginization Versteegh understands not neces­ sarily an end point represented by a stable variety, but rather a process of untutored second language learning (1984:40). In these terms Versteegh is certainly correct in pointing out the relevance of pidginization to the develop­ ment of Arabic in the eighth and ninth centuries, for at this time Arabic acquired large groups of native speakers. I think his thesis that pidginization played a CRUCIAL role awaits a more precise examination of the linguistic and demographic situation in the early Islamic period. 1.2. Sudanic / Arabic For over 1000 years, starting from about 700 A.D., the southern boundary of Arabic speakers lay in the Sudanic region, first around Aswan (see §2.1 below) and then in a region stretching from central Sudan across central Chad to Lake Chad. A major trade and pilgrimage route followed an East-West direction over the area, from West Africa to Egypt and/or Mecca, and in the 16th century the first of a number of centralized kingdoms,2 the Funj empire, developed in east-central Sudan. It is virtually certain that Arabic served as a trade language for the entire region throughout the period, as the data dis­ cussed in §2.1 below confirms. Unfortunately, before the 19th century no other documentation for the use of nonstandard Arabic in the region has been reported. Since it is in the Sudanic region that the best-attested Arabic-based /'s are found, it is necessary to summarize the historical background of the region. Demographics are discussed in §3. The Sudan underwent drastic political and economic changes in the course of the nineteenth century, changes which left their linguistic mark as well, most particularly in the southern third of the country, approximately south of the Sobat and Bahr el-Ghazal rivers. It is in this region that the present-day Sudanic / Arabic varieties developed in the second half of the

128

Jonathan Owens

Arabic-based Pidgins and Creoles

129

19th century. A brief historical sketch of this region in the 19th and 20th centuries is thus in order.3 In 1820 Muhammad Ali, ruler of Egypt, turned his expansionist plans southwards and established suzerainty in the northern and central Sudan in an area centered roughly around Khartoum. Trade contacts with the northern­ most of the southern tribes, the Shilluk, were established, and within a few years, 1839-1841, Muhammad Ali dispatched a series of three expeditions under a Major Selim to explore the Nile south of Khartoum. These expedi­ tions reached Gondokoro near modern-day Juba, and instituted the first trade contacts between the North and the southern hinterland. Initially the main object of export from the South was ivory, and this was effected on a seasonal basis: every year during the Nile high water, traders sailed to the South and returned before the dry season rendered the Sud swamp impassable. The ivory trade was initially a government monopoly, but the increasinglypowerful trading community, a mixed group of Europeans and Arabs of different nationalities, was eventually granted the right to trade directly in the South, and after 1854 the first permanent trading settlements were established there.4 These proliferated quickly, and by 1869 the German explorer Schweinfurth observed traders' camps or 'zeribas' connected by a transport system (Junker 1888:33) at a distance of every 18-21 miles in the Bahr el-Ghazal region and across into what is today Zaire. Zeribas were also established east and north of Gondokoro, and south into northern Uganda. In addition, itinerant Sudanic colloquial Arabic-speaking traders or 'jallabi' (pl. jallaba) began coming in increasing numbers from Darfur into the South, some of them returning annually to the North, others settling permanently in the South. Frequent intercamp contact continued into the period when the South was ruled directly by the Egyptian government (Junker 1891,111:378). The trading settlements were initially manned largely by northern Sudanese Arabs and Nile Nubians (not to be confused with the Nubi of East Africa), many of whom married local women and settled permanently in the South. Relations between the camps and the local tribes were, allowing for local variations, bad. As early as 1863, for instance, Grant (1864:357) reports that the Bari used every opportunity they could to steal from the "Turks", and later Gessi (1892:211) observed that, although the Egyptian government had been established at Rumbek for some years, the Dinka-inhabited area around the outpost was unsafe for travel. The militarily superior northern traders freely commandeered cattle and provisions for themselves from among the

130

Jonathan Owens

local groups, and because the traders' agents were paid partly in slaves (Gray 1961:46 ff.), and because it provided a lucrative export commodity, slave taking became common. Many of the slaves themselves became integrated into the military structure of the camps, so that soon the southern population outnumbered the northerners (see (2) below). The trading camps thus became entities unto themselves, supporting populations of up to 10,000, cut off from the indigenous population except to the extent that the camps had to requisi­ tion their supplies and labor in one way or another from the local population. In response to the increased trafficking in slaves, which by 1870 (Schweinfurth 1918:479) had outstripped ivory as the premier export from the South, the Egyptian government sought to place the South directly under its control, a mission entrusted successively to Baker (1870-1873), Gordon (18731876), and Emin Pasha (1878-1888). Because of their lack of resources and manpower, the government often simply commissioned the standing army of one trader or another to serve as a government force, and in many respects, e.g. the practice of plundering local villages for provisions, relations between the camps and the native populations remained unchanged (see e.g. Junker 1891, 1:279, 322, 355, 340, 438, 512, Gessi 1892:430). The government established a measure of control along the Nile; in the western areas and in Bahr el-Ghazal province its control was far more tenuous (Junker 1888: 49). During the period of Egyptian rule, a significant number of Egyptian functionaries entered the southern Sudan (Kaye 1988:47), their number reaching perhaps 300-400. However, the greater number of government soldiers were locally recruited (see e.g. Stanley 1890:136). The anti-slavery efforts, and even more the encroachment of the Egyp­ tian government, which from 1878 made a concerted effort to keep northern traders out of the South, were enough to antagonize many of the permanent traders, particularly those in the western areas, and some rebelled. The most serious revolt was that of Sulaiman, son of Zubeir Rahman, in 1877-79, and although the revolt was put down, one of Zubeir's lieutenants, Rabeh, took a sizable group with him into the Chari-Logone river basin, eventually captur­ ing the kingdom of Bornu. It is this group which Zeltner (1988:265) quite plausibly suggests was responsible for bringing Turku (see §§2.2, 4.2) to the Chari river basin.5 The Egyptian government never established wholly effective control over the South, and during the Mahdist uprising, with Khartoum falling in 1885, the South was invaded as well (in 1884); most of it was lost either to the

Arabic-based Pidgins and Creoles

131

Mahdi's troops or to indigenous groups. Soon only a few government settle­ ments remained along the Nile south of Gondokoro. Stanley (1890:246) reports that at this time the governor, Emin Pasha, had about 1300 regular soldiers more or less under his command, along with 3,000 irregulars and camp followers. Jephson (1890:222) gives a higher figure for Emin's Equatoria followers, putting the figure at 8,000. In 1888 Emin was forced to abandon the last settlements, and he withdrew to northern Uganda, from where in 1889 most of his Egyptian administrators and officers returned first to the East African coast and then to Egypt or Europe.6 A number of Emin's soldiers and camp followers (700 according to Saghayroon 1981:28) stayed in Uganda, however, and these were quickly co-opted first by the British East Africa Company as early as 1893, then by the British government to help establish British control in Uganda and Kenya. Some of his former troops served the Germans in Tangyanika as well (Furley 1959:326; see §3.2). These soldiers eventually settled in Uganda and Kenya, where their descendants, the (East African) Nubi, still live. In the southern Sudan former soldiers and camp followers of the traders, as well as some former government troops (Junker 1891, III:405, 454, 479), stayed on. Some of these joined the Mahdists, while others formed themselves into private armies with local spheres of influence (Collins 1962:71), and through these groups the / Arabic was kept alive. The British established control of the Sudan in 1898, though the Mahdists had been driven out of parts of the South by Belgian forces as early as 1896. After numerous local rebellions were suppressed in the South, that area enjoyed a measure of peace until around independence in 1956. At that time a rebellion against the central government broke out, continuing intermittently until the present day.

2.

Varieties of modern Arabic-based p/c's

There are four varieties of modern Arabic-based p/c's. Two of these are still spoken today. One is known only through a work written around 1930, and the other, Maridi Arabic, has been termed the oldest pidgin on record (Thomason & Elgibali 1986). Beyond summarizing them in this section, I will have nothing to say about them — for Maridi because there is no new data to add, and for Turku because a detailed treatment can be found in Tosco & Owens (1993).

132

Jonathan Owens

2.1. Maridi Arabic The existence of early Arabic pidgins, or pidgin-like varieties, is confirmed by an eleventh-century text discussed by Thomason & Elgibali (1986). It appears in an 11th-century work on geography written by the Andalusian (Spanish) Al-Bakri, a geographer who himself never left his native land. It is a short text, a little over 50 words, which is said to represent a very bad form of Arabic spoken by people in the region of Maridi. Mainly on the basis of plausible historical and textual evidence, Thomason & Elgibali conclude that the pidgin was spoken in Mauritania. There are other indications, however, pointing to a textual source in Upper Egypt or northern Sudan. In particular, the text contains linguistic features which today are found either uniquely or largely in Egyptian Arabic, e.g. CA q = ', fw' < fwq "on", and CA j = d (dwmaa < jwmaa "Juma"); the last is a correspondence which as a perfectly regular feature is attested today only in Upper Egypt (Behnstedt & Woidich 1985:1 Off.; see Kaye 1985). The text, moreover, was recounted to Al-Bakri by a trader from Aswan who had heard it in the course of his travels. It is, then, quite plausible to assume that the text represents a variety spoken somewhere in Egypt or the Sudan but which Al-Bakri or his source transposed to Mauritania to conventionally represent a bad form of Arabic spoken there. The text itself is likely a caricature which is partly based on an actual spoken pidgin variety, but which is also partly taken from local Arabic dialects and partly perhaps deliberate distortion in order to produce a "bad" Arabic (e.g. incongruous use of the dual; see Owens 1989 for discussion). The preponderance of evidence thus points to a source for the Maridi text in the upper Nile valley. If this is the case, it would indicate the possible presence of Arabic-based pidgins in the region for nearly a thousand years. 2.2. Turku Turku, as mentioned in §1.2 above, was brought to Chad and NE Nigeria by Rabeh's soldiers ca. 1900. Writing of the colloquial Arabic spoken in the Lake Chad area, Lethem (1920:xiii) reports that words and phrases intro­ duced by the followers of Rabeh "are generally known as 'turuk', this being the term applied in Bornu to Rabeh's non-Arab Sudanese troops". Even today the Kotoko in northern Cameroon refer to Rabeh's troops as 'turks' (Zeltner & Tourneux 1986:11). Turku was thus a nonnative Arabic which had its

Arabic-based Pidgins and Creoles

133

origins in the same social milieu as did the / varieties of the Sudan (see §1.2). I have no information about the number of Turku speakers, though Maigret (1931, cited in Samarin 1955:259) reported that Turku was one of the three trade languages (along with Sango and Bangala) used in French Equato­ rial Africa. It was used primarily in Chad and northern Cameroon, though I believe it never played an important role in Nigeria (see Tosco & Owens 1993 for discussion as to why). Precise linguistic details of Turku are known through a lexicon (about 1500 entries), a very inadequate five-page grammar, and collection of short phrases/texts by Gaston Muraz (ca. 1929), a French colonial medical officer. Though not the work of a professional linguist, the practically-oriented gram­ mar, the oldest detailed grammar of a Sudanic Arabic / available, contains enough data to construct a basic grammar and provide comparative material. It can, for example, be ascertained that Turku was an expanded and probably stable pidgin (see Prokosch 1986, Tosco & Owens 1993). (I have no informa­ tion about its status as a creole, and suspect it was not one.) In addition, Derendiger (1912), describing the variety of Arabic spoken in Bagirmi, notes that a continuum existed from a colloquial Arabic dialect to a pidginized form. Derendiger offers only fragmentary evidence of this continuum, con­ centrating on the standard dialect in the area. What examples he does give of the pidginized form resemble Turku and other Sudanic p/c's. So far as is known, there is no variety in present-day Chad or the Central African Republic known as Turku. Turku was, however, a vehicular language which the French colonial government found useful (hence Muraz's book), and in present-day Chad Arabic continues to survive in various pidginized forms (alongside the colloquial Chadian Arabic spoken by ca. 300,000 people), none of which, unfortunately, have yet been described. Future research will thus have to determine whether modern Arabic pidgins continue Turku or arose independently.7 2.3. Juba Arabic The / of the southern Sudan is often conventionally termed "Juba Arabic" (Watson 1984, Miller 1983, 1985). This term unfortunately appears to restrict the variety to the area of Juba, or to suggest some special historical link between it and the area of Juba. The language in fact is the lingua franca of the

134

Jonathan Owens

entire southern Sudan, and is spoken by southerners living outside the south (e.g. in the Khartoum area). Other terms however, like "southern Sudanese Arabic" or "lingua franca Arabic of the Southern Sudan", are either too unwieldy (second alternative) or misleading in their own way (colloquial Arabic dialects are also spoken in the South). Although it is mainly a second language, Mahmud (1983) reports that in Juba (and no doubt elsewhere) the language has gained native speakers; Miller (1985:155) reports that as many as 40% of the population of Juba (total ca. 200,000) now speak Juba Arabic as a mother tongue. According to Mahmud (1983:109), interethnic group marriage is the most important factor favoring the use of Arabic in southern households, with 40% of Mahmud's respondents from households of mixed marriages reporting Arabic as their first language, against 15% of those from households where the parents spoke the same language. The southern Sudanese languages spoken alongside Juba Arabic include Bari, Lotuxo, Moru, Madi, Zande, Dinka, and Nuer, as well as a host of smaller languages. Juba Arabic is structurally quite distinct from and mutually unintelligible with any variety of colloquial Arabic. This structural distinctness will be briefly illustrated in §4.4. below. The structural differences are reflected in the social status of the languages. Miller (1985, 1988:304) says that southern Sudanese consider colloquial Arabic and Juba Arabic different languages and that those southerners with a knowledge of both mix and manipulate the two according to the exigencies of the situation. The difference between the two can take on strong symbolic significance. Bell (1974:4), for instance, notes that one objective propagated by the Azania Liberation Front, the southern Sudanese movement which fought the central government up to 1972, was furthering of a common southern Sudanese nationality through the develop­ ment of the "pidgin Arabic spoken in the South". No important structural differences between pidginized and creolized varieties of Juba Arabic have been reported, although detailed linguistic and sociolinguistic data is lacking. Under influence from colloquial Sudanese Arabic a continuum has developed (Mahmud 1979, Miller 1985, 1986) rang­ ing from pidginized varieties on to colloquial dialects, a situation which renders an adequate characterization of Juba Arabic all the more difficult. Watson (1984, also 1989:108), for example, makes valuable attempts to render a description of the basilectal Juba Arabic variety; he presents a number of constructions with variant realizations, one closer to Sudanic

Arabic-based Pidgins and Creoles

135

colloquial Arabic than the other(s) (e.g. on pp. 17 and 39 on the passive), without any comment on the sociolinguistic implications of using one form or the other. No dialect surveys of Juba Arabic exist, though there are clearly dialect differences, and grammatical sketches and sociolinguistic surveys are limited mainly to the Juba variety (Nhial 1975, Mahmud 1979, 1983, Miller 1983, 1985, 1986, 1988, Watson 1984, 1989, Vincent 1986). 2.4. East African Nubi East African Nubi, also known as KiNubi (Kaye 1991:5), is spoken as a first language by 10,000-15,000 Kenyans and an unknown number of Ugandans. The number of Ugandan speakers was for many years considerably larger than those in Kenya, though because of their close association with the Amin regime (he himself is a Kakwa) many fled Uganda to Kenya at his downfall. There were once also East African Nubi speakers in Tanzania and Somalia; whether some still remain is unknown. In East Africa Nubi is almost exclu­ sively a first language, except perhaps in northern Uganda, where, because of proximity to the southern Sudan, Nubi functions to some degree as a lingua franca. Juba Arabic and Nubi are mutually intelligible, and Nubi, like Juba Arabic, is not mutually intelligible with colloquial Arabic. Almost all Nubi are bilingual in Swahili and Nubi, and it is not uncommon for them to speak other languages as well (Heine 1982:16). The Nubi are mainly urban dwell­ ers, the largest concentration in Kenya being in the Kibera area of Nairobi; smaller settlements are found in Eidama Ravine, Kisumu, Nakuru, and Mom­ basa. In Uganda the largest settlements were at Bombo and Kampala/Mengo, as well as smaller settlements in most other cities.

3.

The genesis of Sudanic / Arabic

Discounting possible, though presently unverifiable, influence from pidgin/ trade languages prior to 1820, the formative period of Sudanic / Arabic has to be placed between 1854 and 1888, the period between the opening of the South to large-scale northern influence and the expulsion of the Egyptian government from the South. I take 1888 to be the endpoint because at about this time East African Nubi became isolated from the Arabic of the South, so present-day similarities between East African Nubi and Juba Arabic (see §4.2

136

Jonathan Owens

below) must be accounted for by their common pre-1888 origin. That is, the Arabic of the South had crystallized in a stable pidgin or creole variety by 1888. Two important characteristics of Sudanic / Arabic are thus of general interest for creole studies (Carrington 1987:78): it stabilized (a) within little more than one generation and (b) within a social milieu where native speakers of Arabic constituted no more than 10-25% of the total population. To gain a closer insight into the genesis of Sudanic / Arabic in the 19th century (§3.4), it is necessary to look more closely at the language situation of the 19th-century South (§§3.1, 3.5), its demography (§3.2), and the social structure of the population (§3.3). 3.1. The languages of the South The southern Sudan is (and in the 19th century was) a linguistically heterogeneous area. Its languages fall mainly into two large clusters, Eastern and Central Sudanic (Greenberg 1966:85ff.). Niger-Congo is represented by Zande (Adamawa-Eastern branch). (1) Language of the southern Sudan (partial list):

While there is some mutual intelligibility among these languages — I have placed brackets around those which are to a degree mutually intelligible8 — the diversity is considerable.

Arabic-based Pidgins and Creoles

Languages of the Southern Sudan and adjoining regions

137

138

Jonathan Owens

3.2. The 19th-century population Thanks to the observations of the German explorer Schweinfurth (1918:50508), who traveled around the Bahr el-Ghazal area during 1868-1871, we have some rough estimates of the population in the region. The picture he paints is of a great upheaval. He divides the population into the following categories; the list in (2) is supplemented by the observations of other writers. (2)

A. Indigenous populations, not connected with trading camps: 190,000. B. Camp populations: (i) The traders' soldiers, officials, and translators: 5,000-10,000. Particularly important here are the Nile Nubians, who consti­ tuted the largest single northern group.9 Translators were mainly southern Sudanese from various ethnic groups born and brought up in the North, though there were also southerners, particularly in the early years of southern penetration, who were sent by chiefs of various southern groups to the camps of traders in the South for the express purpose of learning Arabic (Petherick & Petherick 1869, I:141,274). (ii) Permanently settled traders (jallaba): 2,000. (iii) Seasonally migrant traders: 2,000. These moved back and forth between Darfur and the South. They were mainly from western Sudanese Arabic tribes, like the Baggara. Gessi (1892:386), who was largely responsible for excluding northerners from the South after 1878, estimated that up to 20,000 jallaba had been active in the South. Being itinerant, this group had the least intense contact with the southern Sudanese. (iv) Slave soldiers (called faruch, basingers, or nyarakik): 5,000, from various southern ethnic groups. These were mainly offspring of the soldiers. They worked as gun carriers and comprised the main fighting force of the zeribas (traders' camps). During the period of government rule some of these troops became regular government soldiers, some continued to serve local traders, and others constituted a floating element owing allegiance variously to the govern­ ment/traders/Mahdists/local groups, according to the prevail­ ing political climate (see below).

Arabic-based Pidgins and Creoles

139

(v) Field and household slaves from southern Sudanese ethnic groups: 40,000. Every soldier had at least one slave; Gessi (1892:222), who admittedly was prone to some exaggeration, gives a ratio of 10 or 20 to one (slaves to soldiers). In this category belong concu­ bines. Writing during the period of Egyptian rule, Gessi reports in a letter (1892:433), "I must not forget to mention that 100 regular troops have been sent from Khartoum to Lado, another 100 will follow shortly. These men will have no women with them. As soon as they arrive in the interior they will each lay claim to 3 wives, 3 concubines, and many servants." (vi) Other camp followers: 1,000.10 Total camp population: 55,000-60,000. Of the ca. 250,000 inhabitants of Bahr el-Ghazal in 1870, 60,000, or about one quarter, belonged to the trading camps. Of those in the camps, only 9,00014,000 would have been native or near-native Arabic speakers (see n. 9 for Nile Nubians) — that is, (2) B.iii. Schweinfurth's estimates, it should be noted, describe the situation in 1870, sixteen years after the establishment of the first trading camp in the South. Two factors would have favored a gradual decrease in the percentage of native Arabic speakers. First, given the population growth through birth within the camps (see §3.4), the percentage of native Arabic speakers would have tended to decline. Secondly, the government policy (after 1878) of prohibiting northern traders in the South and of favoring the recruitment of native southerners to the army would have led to a relative decrease in the number of native Arabic speakers. Reports by writers after Schweinfurth tend to confirm this development. Junker (1891, III:526, III:572, I:462), for ex­ ample, writes that in the government post at Bor the Egyptian officer in charge had a household of 32 persons, including 6 concubines; a clerk had 28 in his household, a sergeant 12, and so on. The bulk of these household followers were from the South. Casati, an Italian who served the Egyptian government for ten years in the South, reported that he had 3,000 troops under him, two-thirds of whom were volunteers from the area or freed slaves (1891, I:49). Jephson, who accompanied the Stanley/Emin Pasha relief expedition and spent a number of months alone with Emin and his followers, reports (1890:48) that most of Emin's soldiers in the South did not want to accom-

140

Jonathan Owens

 Emin to the coast because, though they were called 'Sudanese', they were really recruited from southern tribes — Dinka, Madi, Bari, Azande, Bongo, Makaraka, Mangbetu, Moru and others. The situation among Emin's followers in Uganda was similar. Furley (1959:316), for instance, reports that in their camps in the 1890s there were 10 camp followers — one can assume not mother-tongue colloquial Arabic speakers — for every soldier, and the soldiers themselves would not all have been native colloquial Arabic speakers. The ratios can be judged with some accuracy for the Sudanese serving the Germans in Tanganyika, from figures found in the German Reichstag Report of 1897-98 {Stenographische Berichte, pp. 248-49), which contains basic statistics on the colonial police and army forces; these are divided between Bantu and Sudanese troops, and the Sudanese are divided in turn into various subgroups.11 As of April 1898 there were 762 Sudanese troops, 466 soldiers, and 296 police. The soldiers were divided into 14 companies and the police were stationed around 13 different regions. In the preceding year 126 Sudanese had died, deserted, or (the large majority) been discharged, ostensibly because they were weary of service, though more likely because the German administration feared that a rebellion by the Sudanese in Uganda, begun in 1897, would spread into their own territory. A breakdown is given of the 762 Sudanese troops according to 50 different groups, only some of which clearly correspond to known ethnic or linguistic entities, and some of which would allow overlap with other groups in the count. I could identify about 35 of the groups, though the largest, termed "Sudanese" (= 134), is too general to ascertain what language they spoke. Subtracting this from the total, one is left with 628. Of these, 11 groups can be positively identified as possibly having a good command of colloquial Arabic. It is in fact unlikely, however, that all would have known Arabic well. Nubi clans are reckoned patrilineally according to tribal origin, so to be called a "Jaali" (an Arabic tribe), for instance, one would have needed only to have a "Jaali" father or grandfather. Nothing can be directly inferred about the language competence of the individual. Of the appellations used in the report, the following are all groups from northern Sudan or from Arabic countries: Nubians (= 107), Jaali (Ja'aliin, a northern Sudan Arab tribe, 19), Dagalani (?Dongola Nubians, 5), Barbarani (?from the city of Berber in n. Sudan, 7), Dongolani (Dongola Nubians, 3), Abbadi (Abbaabda, an Arabicized Beja group, 1), Chami (Shami, from Syria/Damascus, 1), Arabs (18), Turks (6), Kurds (2), and Syrians (1): The subtotal of these is 170, or 27% of the total (of

Arabic-based

Pidgins and

Creoles

141

628). This means that only 27% of the troops at most are likely to have had a good command of Arabic. When one recalls that each soldier had a household of 5-10 followers, most of whom would have been from the southern Sudan, with some perhaps from East Africa, the percentage of fluent Arabic speakers falls well below 10%. Allowing, conservatively, 5 followers for every soldier, 628 x 5 = 3140 + 628 = 3768, the percentage of potential Arabic speakers falls to 4.5%. If the Tanganyika situation is representative of that in Britishcontrolled East Africa, one can understand that post-creolizing processes would have been virtually nonexistent. 3.3. Social structure in the southern Sudan Junker presents evidence for an emerging tripartite social structure in the South by 1885. His best evidence pertains to the better-documented Egyptian government-controlled areas, though there is some indication that the situa­ tion was mirrored in other areas as well. At the bottom were local groups who had only transitory contact with the government camps and traders — at best brief trading exchanges or employment as porters, at worst victimization at the hands of a plundering expedition. At the top were the northerners, Sudanese, Egyptian, and Nubian officers and the northern traders who had managed to remain in the South. In the middle were the southern soldiers, whether born in the camps or recruited from the indigenous population. On the one hand they distanced themselves from the southern villagers, a process observed by Baker as early as 1864: All the best slave hunters, and the boldest and most energetic scoundrels were the negroes who had at one time themselves been kidnapped. These fellows aped a great and ridiculous importance. On the march they would seldom condescend to carry their own gun; a little slave boy invariably attended to his master, keeping close to his heels, and trotting along on foot during a long march, carrying a musket much longer than himself; a woman generally carried a basket with a cooking pot and a gourd of water and provisions, while a hired native carried the soldier's change of clothes and ox-hide upon which he slept. Thus the man who had been kidnapped became the kidnapper and the slave became the master... (Baker, quoted in Bradnum 1969:236)

On the other hand, according to Junker there was a tension between the northerners in the South and the southern black troops. He contends that the northerners tended to eschew the company of southerners, even if they were

142

Jonathan Owens

their own soldiers, and that the southerners in turn got along most easily with officers of black origin (1891, I:543), for instance those from the Nuba Hills (I:340, 356). This point is further substantiated by Gessi (1892:429), who observed that "during the war [against Sulaiman; see §1.2] we had clear proofs that it was the Arabs and not the natives who revolted... the negroes helped the Egyptian government." Although Gessi betrays an open anti-northern bias, his observation does bear witness to an Arab-southern Sudanese division. Tensions were greatest between the southern troops and lower ranking Nubians (Junker 1891, I:543, III:427, Emin/J. 180, 182, 190), particularly after the closing of the South to northerners and the repression of the slave trade (III:136; see Jephson 1890:110 for further evidence). In non-government-controlled areas Junker reports (1888:33) that, early on in the development of the network of trading camps, southerners living along the transport network joined, or were forced to join, the camps, forming an army (see (2) B.iv) which became the core of Sulaiman's troops (see n. 5 above), as well as participating in the Mahdist invasion of the South (Junker 1888:33, Emin/J. 1888:196, 199, 208). Junker observes (1888:33) that these were a "terror to the free indigenous groups" (cf. the Baker quote above). But Emin also reports (185, 208) that southerners in the army of the Mahdist leader in the South, Karamallah, found him excessively strict and fled his service. In 1884-85 there were at least two mutinies by southern troops in Mahdist employ in the South (Collins 1962:47). Here too evidence suggests a class of southerners distinct from both the local groups and the ruling elite. Junker's three-class analysis of southern society finds confirmation in the cleavages that became evident during the Mahdist invasion of the South. The indigenous population rose against the government forces, though they generally did not work in close alliance with the Mahdists (Junker 1888:145, 165 n. 2, 1891, III:458), and frequently fought them as well (Collins 1962:73, 75, 83, 114, 133). Aside from those who were officers of the Egyptian government, though even here there were exceptions, the Nubians and Arabs supported the Mahdists (Junker 1891, III:459, Jephson 1890:283). The south­ ern troops in government employ strongly opposed the Mahdists (Junker 1891, III:427, 478) and frequently had to fight the indigenous population as well. Irregular southern soldiers tended to change their loyalty according to the exigencies of the situation, sometimes supporting the government (Junker 1888:73, Emin/J. 214) or the Mahdists (Emin/J. 199, 205), sometimes joining

Arabic-based Pidgins and Creoles

143

the local groups against either (Emin/J. 185, 204, 220), and sometimes estab­ lishing their own local areas of control (Collins 1962:71). The three-cornered fight among indigenous groups, Mahdists/Northerners, and southern govern­ ment troops reflected three basic divisions in southern society. 3.4. Emergence of a / Four factors are frequently, if not always, relevant to the development of a /: speakers of diverse languages must be (i) mixed together (Whinnom 1971) and (ii) in daily contact, and (iii) either this mixture occurs in the presence of a dominant people using another language (the European-based creoles) or at least none of the languages of the various groups has a 'prestige' status and another language is available as the prestige model (Lubumbashi Swahili — Polomé 1968:14, Tok Pisin — Stoller 1979:75). Finally, (iv) the prestige target either cannot be learned by the subordinate groups, for what­ ever reason (when the prestige language is the language of the superordinate group), or, with the acquiescence of the superordinate group, a simplified version of a more complex variety becomes established.12 I now examine these points relative to the emergence of Sudanic / Arabic. These four factors obtained in the traders'/government camps that facili­ tated the emergence of a Sudanic Arabic-based /. First, the soldiers in the southern camps, as noted above, quickly gathered a slave and soldier popula­ tion taken from various southern ethnic groups. Many intermarriages in the camps would have been intertribal, a factor reported to favor creolization (Mahmud 1983; Polomé 1983:130 on Lubumbashi Swahili). Secondly, the contact among the different groups within the camps was regular. It had to be, for in daily life the camps were largely cut off from the local populations. The traders were universally hated by the southern tribes, even if alliances with local groups were regularly formed to further the activities of the traders and the local chiefs. Intensity of contact is further indicated by the presence of female slaves — male slaves more often being sent north for further exportation — who often became the wives of the northern soldiers. Nearly all writers remark on the large number of wives among the camp soldiers. Grant, for example, describing a seasonal trading camp at Faloro in northern Uganda (1864:332), reports that female slaves far outnumbered males, and Stanley, describing the situation over 25 years later (1890:136), writes that some of Emin's officers

144

Jonathan Owens

had between 18 and 100 women, children, and servants working in their homes. The women slaves had an important domestic function. Casati (1891, I:267) observes that they were the effective household heads, a situation Saghayroon reports as continuing among the East African Nubi (1981:86). When traveling, the government troops of the South never went without their wives and servants (Junker 1891, II:216, 254). The male offspring of these marriages were integrated into the socioeconomic milieu as soldiers, soon numbering as many as the northerners themselves (see §3.2 above). The slaves, concubines, and their offspring would undoubtedly have continued to speak a southern Sudanese language, though at the same time a language developed, the / Arabic, which would have been used between the different groups. Thirdly, the language of the dominant group was Arabic (§3.2). Finally, the question has to be posed why a / Arabic rather than a colloquial Arabic dialect developed in the South. This brings us to the fourth point, whether or not it was possible for colloquial Arabic to be learned. The answer is no, for demographic and social reasons. In the camps (traders', later government) a mixed population emerged which by 1880 is reported in many places to be roughly on the order of 10:1 in favor of native southerners vs. northerners, a large preponderance of non-colloquial-Arabic speakers. Fur­ ther hampering the acquisition of a colloquial Arabic dialect was the exist­ ence of a social stratification separating southerners from the northerners (see §3.3 above). Moreover, the important domestic role of southern women would have meant that the first language learned by children was either an indigenous language or a nonstandard form of Arabic spoken by the women and their offspring, or both. The reason for the emergence and stabilization of Sudanic / Arabic is to be sought in the crystallization of a class of southern­ ers who no longer belonged to indigenous southern groups, but who were also not part of the ruling northern elite. Parallels with the early development of the Atlantic creoles on the West African coast (Hancock 1986) are important, perhaps crucial. In both places the emergence of a stable / coincided with the formation of a social class that stood between an indigenous population and a dominant class. On the West African coast these were the creoles, who had a position between the whites and the African majority. In the 19thcentury southern Sudan the new class comprised the soldiers of the traders and the Egyptian government who had a station between the northern Arabs and Nile Nubians, on the one hand, and the surrounding African groups on the

Arabic-based Pidgins and Creoles

145

other. What I would hypothesize, then, is that for both the creoles and the southern Sudanese soldiers, the / was an expression of social class. 3.5. Language use in the 19th-century South Amidst all the figures and descriptions of life in 19th-century southern Sudan, it is difficult to determine exactly what languages/dialects were spoken by whom. Outside of the trading camps it appears that what little Arabic was used was restricted to some of the chiefs (Saghayroon 1981:4). In describing his travels in Equatoria, Casati, for example, reports on numerous occasions (1891, I:102, 144, 149, 196) conversations with local inhabitants in African languages, but none in Arabic, and he emphasizes the language and cultural differences between southerners and Arabs (1891, I:263). Both Schweinfurth and Junker emphasize the role of translators (Arabic-southern Sudanese languages) in their travels. Junker never traveled without translators for the local languages. As an extreme example, Jephson (1890:406), traveling south of Equatoria province, describes one chain of translation he needed to speak to a Munyoro chief that went from Swahili to Arabic to Alur to Lunyoro (and back). Within the camps both Arabic and local languages were used. Casati (1891, I:58) notes that among the southern soldiers divisions ran along ethnic lines, and Junker (1891, I:303) reports that no officer of the Egyptian govern­ ment would have marched against 'rebellious' southerners with a contingent composed solely of troops from one ethnic group, the danger of insubordina­ tion from a homogeneous group being too great. Junker further writes (1891, I:315) of large government caravans divided into discrete groups of Bari, Moru, Kakwa, Nubians, and others, and their wives and servants arranged "partly according to tribe, partly according to where their men were in the caravan". Ethnically-based divisions and solidarities such as the ones repre­ sented in these reports were no doubt symbolized inter alia by language differences (see also Jephson, cited in Stanley 1890:134). Crabtree (1913: 155) notes that among the East African Nubi there could always be found a speaker of any of the languages of the southern Sudan.13 Arabic was of course also important. Gordon, observing the traders' camps in Equatoria in 1875, wrote, "no nation could uproot Egypt from these lands even if they possessed them. Arabic [what variety is not stated] must be the language of these countries" (letter quoted in Gray 1961:113), and

146

Jonathan Owens

Mahmud reports (1983:13) that among Emin's southern soldiers who entered Uganda with him were those who spoke only Arabic. One reason for a mutiny by the Sudanese troops in Uganda in 1897 was that none of the British officers could communicate with their men in Swahili or Arabic (Furley 1959:321). It would appear, then, that in the camps Arabic was an important lan­ guage, though it had not displaced the local languages. What forms of Arabic were used is not entirely known. Colloquial Arabic was certainly used by Arabs and many Nile Nubians, and their use of dialect features is reflected today in the many traces of frozen morphological material found in East African Nubi (Owens 1985a:250 ff.). Among the southerners, knowledge of Arabic varied from place to place. Junker (1891, I:283), for example, writing about his 1876 travels, notes that around the newly-founded (1874) government station of Lado few of the Bari had learned Arabic, whereas just across the Nile near the longer-established Gondokoro they had "acquired the vehicularly necessary Arabic words". When describing the actual Arabic used by southerners, Junker is inconsis­ tent. In places he speaks of "vehicular" or "indigenized" ("negrisiertes" — 1891, I:172) Sudan Arabic, and elsewhere of two Azande princesses who spoke "some Arabic" (I:505). There were also southerners who had learned fluent colloquial Arabic, both in Khartoum (I:375) and perhaps in the South as well (III:150), though reading between the lines it would appear that refer­ ences to "Arabic" in the South by Junker and other European observers often refers to something other than a colloquial Sudanic Arabic dialect. One report explicitly confirms this. Stanley (1890:168) relates the complaints of his Zanzabari porters, who object to being called "slaves" by Emin Pasha's troops: "...we know enough Arabic to know what they mean, bad as their slang Arabic is...". However, direct comments on the Arabic used became available only in the 20th century, these mainly from colonial officers and missionaries. Crabtree (1913:154), writing about the languages of Uganda, speaks of a "modified form of Arabic" called Nubian, noting that it forms a lingua franca of sorts along the Nile valley. Other than Muraz (see §2.2), the only actual early textual data comes from Thorburn (1924).

Arabic-based Pidgins and Creoles 4.

147

Structural features

In this section I will first (§4.1) give a skeletal outline of Kenyan Nubi (KN), which will serve as a reference point for a comparative discussion of other Sudanic / Arabic varieties (§4.2) and the relationships between the / varieties, substratai languages (§4.3), and colloquial Arabic (§4.4). 4.1.  sketch of Kenyan Nubi In this section I will give a skeletal outline of Kenyan Nubi (my own data). The choice of Kenyan Nubi (KN) is one of convenience: it is the bestdescribed Arabic-based / in print (Heine 1982) and the one I am most familiar with (Owens 1977). 'UN' marks examples from Uganda Nubi. 4.1.1. Phonology The phonemic inventory is given in (3). (3)

p b f v m w

t d s z n l,r

c j sh

k g

i e h

u o , 

ny y

Stress is distinctive, both lexically and grammatically; compare ína "we, us" and iná "there", amsúku "catching" and amsukú "caught". The main syllable types are CV and CVC. Sandhi phenomena are important in Nubi. East African Nubi words undergo a fair degree of alternation in fast speech. I briefly mention (nonexhaustively of course) three important phenomena here. First, VOWEL HAR­ MONY is a lexically-defined phenomenon. Leaving out of account a word-final vowel, a word can have only front vowels i, e or back vowels o, u; the low vowel a occurs in both contexts: tegíli "heavy", shunú "what", and áshrubul ásurubu "drink", but kábitu "write" and boyí "far", where word-final vowel does not fall under the rule. Vowel harmony is also a morphological rule. Two sets of morphemes can be distinguished. The morphemes gi- "progressive", bi- "future", and fi "at" have the form  before back vowels (4a), while ta

148

Jonathan Owens

"of, ma "with", and na/la "to, for" have the form Ce before front vowels, Ca before a, and Co before back vowels (4b): (4)

a. fi bé "at home", fu sú "at the market" bi-jére "will run", bu-fútu "will pass" b. te bé "of home", ta rági "of a man", to sunú "of what"

This is more typical of UN than of KN. In close juncture a high vowel i, u of a morpheme-final CV syllable tends to be lost: (5)

gi ásadu→ gásadu "is asking", bi ágderi→ bágderi "will be able", bu-dugu íta → budugíta "you will be hit"

Before pronoun objects, final high vowels of verbs tend to be dropped, with the stress of the pronoun (always initial) remaining on the "new" syllable. In disyllabic verb stems the lexical stress of the verb is lost; in tri- or more syllabic stems the stem stress becomes secondary. (6) bu-dugú íta → bu-dugíta "will hit you" ásadu úmwon → ásadúmwon "asked them" gi-nadi ítakum → gi-nadítakum "are calling you (pl.)" Similar rules apply to further combinations with pronouns (e.g. preposition + pronoun, as in na íta → néyta "to you"). 4.1.2. Morphology In the verbal noun, the penultimate syllable is stressed, as in áshrubu "drink": ashrúbu "drinking", jére "run" : jére "running". The most significant deriva­ tional process is compounding. Compounds are syntactically similar to in­ alienable related nouns, though ta " o f can never occur between the nouns in a compound, as it can in a sequence of nouns, and in some compounds the first noun may lose its secondary stress; examples are mamá arús "bride", waja rás "headache", and gálam híbiri "pen". A predicative adjective (used only as a finite predicate) is formed from a transitive verb by the addition of ma-, e.g. séretu "tear" : maséretu "torn". Reduplication in Nubi has an intensive or distributive meaning; many, though not all, verbs can be reduplicated, and other word classes allow reduplication on a less regular basis. Examples are given in (7):

Arabic-based Pidgins and Creoles (7)

149

gátagáta "cut in pieces", kubu "spill" : kubkub "spill carelessly", jérejére "run all over", kurukum "yellow" : kúrukumkúrukum "dark yellow", guwóm guwóm "very fast" (UN).

The verb has the following structure: (8)

ANT

FUT

PROG (stem)

(kán) (bi) (gi) úwo áshrubu "he drank", úwo ma bi-gi-ashrubu "he won't be drinking", kán úwo bi-áshrubu "he would have drunk". The sequence is as stated, bi and gi are prefixes, while kan can occur before the subject. All combinations of prefixes + stem occur. Subject + bare stem = past, kán + bi = counterfactual. The stem has a single segmental form. The imperative is the bare stem (+ kum pl.): rúwa "go!", rúwa- "go (pl.)!". The negative imperative consists of máta(-kum) + stem, as in mata(kum) rúwa "don't (pl.) run!", or mata + V-k in pl., as in máta ágilib-kum "don't turn (pl.)!". 4.1.3. Clause structure The canonical clause structure is Adv S V O Adv. The basic typology is SVO; adverbial elements (time, place, manner) occur freely initially or, more com­ monly, at the end of the clause. A very few verbs allow SVOO (wedí "give, send", nadí "call, name"), and a larger number allow SVOAdj (e.g. so "make into something"). Generally an IO is marked by lalna. (9)

a. Kásulu gumásh dé nedíf. wash clothes the clean "Wash those clothes clean." b. Úwo wedí ána ája dé or Wedí he gave me thing the gave "He gave me that thing."

aja thing

dé na ána. the to me

The optative is formed with ke(de) "let, should" (UN also kelé): (10) a. Ké ítakum rúwa. let you/pl go "You (pl.) should go." b. Kelé gusú. let looked "Let it be looked for." (UN)

150

Jonathan Owens

An unspecified subject is indicated by means of a verb stressed on the final syllable with the subject deleted. This is generally translated as a passive, though intransitives occur as well: (11) a.

Úmwon gi-jere "they are running" : gi-jeré "someone is run­ ning". b. Adurb ína iná kamán ína sídu géliba ténna. shot us there again we closed hearts our "We were attacked, but we dug ourselves in." (UN)

Equational sentences have a zero verb in the present and kan in the past: (12) a. b.

úwo tájirilmálim. "he is rich/a teacher." úwo kán tájirilmálim. "He was rich/a teacher."

Existential sentences are marked by fi. This is used with all locative predi­ cates, and with a VN marked by the preposition fi it signals an action occurring at the time of speaking: (13) a.

Úwo fi fi bé/iná. he exist at home/there "He is at home/there." b. Bagará fí na ana. cows exist at me "I have cattle." c. Uwofi fi ashrúbu. he exist at drinking "He is drinking (right now)."

There also exists a "contingent be", which either represents a state of affairs as temporary or conveys an element of speaker subjectivity to the statement. (14) Úwo tájiri. "He is rich." vs. Úwo kún tájiri. "He is temporarily rich/ appears (to me) to be rich." Besides the use of existential sentences to indicate possession (second ex­ ample), there is a morphological verb "have", éndiléndis(i): (15) Úwo bi-éndi mótoka. he FUT-have car "He will have a car."

Arabic-based Pidgins and Creoles

151

The most important FOCUS construction consists of a left-topicalized constituent followed by the morpheme yd. This morpheme can in principle mark any constituent in any part of the sentence (one per major S), and used alone it serves as a presentative locative (in variation with the common yadál yadé). (16) a. Nyerekú fi fi skul yadá. child exist at school "The child is at the school right there." b. Úw wósul yá/yadá. he arrived "Here he has arrived." (UN)  Ana biyo yd tinén. I

bought FOCUS

two

"I sold/bought TWO." As a focus marker the ya-marked constituent is normally S-initial, and any major constituent, including the verb, can be topicalized with yd. S-initial Q markers are generally marked by yd (focused constituent is in boldface): (17) a. Kátifu wáraga yá mdriya dé so. write letter FOCUS wife the did "Write a letter is what my wife did." b. Mára táki dé bará yá ána áj. wife your the only FOCUS I want "It is only your wife whom I want."  Munú yd gi-já? who

FOCUS

PROG-come

"Who is coming?" Some DEPENDENT CLAUSE MARKERS are adverbial — báda "after", bakán "when", láman/náman/ládi "up to", and mishán/misén "because" — and one is conditional: kán "if. (18) a. Fi bakán dé ína dúrubu ládi ína ámurugu járman. at place the we shot until we removed Germans "At that place we attacked until we removed the Germans." (UN)

152

Jonathan Owens b. Bakán when úmwon they "When

úmwon ásuma kúra gi-kóre bumára dé they heard feet PROG-run much DEF jére. ran they heard our footsteps coming loud, they fled." (UN)

RELATIVE CLAUSES consist of allab + S. The markers allab are optional (though they usually occur). The DO extraction site is optionally marked by a resumptive pro, the locative site by fógo,14 and a definite relative clause is usually closed by dé.

(19) a.

Kéli (al) gi-gén jám séderi dé tayí. dog (that) PROG-stay near tree DEF mine "The dog which is sitting next to the tree is mine." (UN) b. bagara al íta áyinu (úwo) dé cow which you saw (it) this "the cow which you saw" c. ká ø ána kán kéytu gumás tayí fógo yá iná. had sewn clothes mine in-it FOCUS there place ø I "The place where I had sewn my clothes is over there."

The NOUN CLAUSE MARKERS are ø, zalja (gal), gal, kélkedé (after verbs of wishing and kútu "make"). (20) a. Ana árufu za gal kalám táki batál. I know that word your bad "I know that what you say is bad." b. Ana asuma min sabí tayí za (gal) azól nade I heard from friend mine that man that lísa má já. still not come "I heard from my friend that that man didn't come yet." c. Kalám al kútu anás náde ké jére bára áwaja. word which put people those that run outside wrong "The thing that made those people run outside is stupid." (UN) Verbs of perception and logó/ligó "find" take a verbal complement marked by

Arabic-based Pidgins and Creoles

153

(21) Ána shúf/ligó úmwon gi-álabu mpíra. I saw/found them PROG-play football. "I saw/found them playing football." 4.1.4. Negation The negative marker is ma in KN and ma or mafi in UN. It occurs immedi­ ately before the verb prefixes, starting with bi-, in KN, and before a preverb prefix or S-finally in UN. The existential negative is mafi (see §4.1.2 for the negative imperative). (22) Uwo má gi-já. he not PROG-came "He is not coming." = úwo gi-ja mafi (UN) 4.1.5. NP and Nominals Sequence: N Poss Adj Num RC Dem "all": (23) bagará máriya táki kubar-ín náde kúlu cattle woman your big-pl. those all "all those big cows of yours" In genitives, a possessor N is marked by ta. In inalienable relations the ta is optional. (24) silá ta rági dé "the weapon of this man", ída (ta) mulódo "hoe handle" Pronoun paradigms are given in (25): (25) Subject/Object SG

1st: ána 2nd: íta 3rd: úwo

Possessive

PLU

SG

PLU

ína/nína ítakum úmwon

tayí táki tó

téyna/ténna tákum tómwon

Nouns, adjectives, and demonstratives vary for number, though the only consistent agreement occurs between N and demonstrative. Moreover, only a few adjectives and not all nouns have separate sg/pl forms. The most common plural marker is -a; singular nouns ending in -a shift to final stress. Examples are bagara "cow" : bagará "cows", shéder "tree" : shederá "trees", ída "arm(s)", kabír "big", kubár or kubará "big (pl)".

154

Jonathan Owens

4.1.6. Prepositions, comparative, and question words The Nubi prepositions are gerí ma "near to", boyí ma "far from",fi "at", la/na "at/to" (human objects), bén "between", gidám "in front", wára "behind", jam "near", min "from", ládi "up to", té/téte "under", báda "after", kábla ta "before", fi rás "on top", fu ústu "in the middle", fijo ( 3rd person > 3' person (obviative)

This means two things. First, if the agent is lower in the hierarchy than the patient, then the verb receives an "inverse direction marker", which indicates that the action goes against the hierarchy. Otherwise there is a "direct" marker. Second, whenever two or more person affixes are present in one complex verb, the relative order is as stated in (8). Some examples are given in (9): (9)

a.

nı-wa:pam-a:-w 1-see-DIR-3 "I see him"

b. nı-wa:pam-ık[u] 1-see-INV.3 "He sees me"

c.

kı-wa:pam-ıtı-n 2-see-INV-NON3 "I see you"

d. kı-wa:pam-/-n 2-see-DIR-NON3 "You see me"

It should also be mentioned that verbs in Cree agree in gender with their objects (in transitive sentences) and their subjects. Gender in Cree is based on animacy/inanimacy. Verbs often have different stems, e.g. wa:paht- "see it", wa:pam- "see him/her". They also have different suffixes: wa:pam-e:w "he sees him", wa:paht-am "he sees it"; nıwa:pam-a:w "I see him", nıwa:pahtæ:n "I see it"; etc. This is also true for Michif. Because Michif nouns derive from French, they must be marked for regular French gender (but cf. §4.4.1 below) in order to select the correct French determiner; in addition, however,

316

Peter Bakker and Robert A. Papen

they must be marked for Cree gender in order to guarantee the correct verb forms. How this is done is theoretically problematic, particularly as most Michif speakers do not know Cree (see also §4.4.4). Here is an example to illustrate the animacy distinction: (10) a. b.

ki:-mıčımm-e:w aitiht laržã. PAST-hold-.→3' some money ki:-mıčımm-am aitiht la pej. the payment PAST-hold-TI.3→4 "He kept part of the money/payment."

The verb endings differ according to the animacy of the object (in the example the verb stem happens to be the same for animate and inanimate arguments). In (10), French laržã "money" is animate because Cree šu:nıja:w is animate; but la pej is inanimate, presumably following Cree (inanimate) nominalized verbals in -wın, such as tıpahamakowm "payment". Rhodes (1977) gives a full list of Michif verb paradigms. Some forms not elicited by Rhodes are generally known, however, even at Turtle Mountain. He says, for instance, that the exclusive/inclusive distinction for first person plural is lost in the conjunct order in Michif; specifically, he says that exclusive -ja:hk and inclusive -jahk have merged into -jahk. Our data show that this is not the case, at least not for most speakers. Here are two examples of inclusive and exclusive verb forms in both the conjunct order and the independent order: (11) gi:-ša:pu-st-a:na:n lɨ rũ-d port ši-pi:stıkwe:-ja:hk. lPAST-pass-go-lpl.EXCL the circle-of door -enter-lpl.EXCL "We [but not you] walked through the archway to come in." (12) la munisjũ či: kı-t-aja:~na:n šı-papa:-ma:či:-jahk? the ammunition Q 2-r-have-lpl.INCL -about-hunt-lpl.INCL "Do we have enough ammunition to go hunting?" However, it is true that some speakers do not use the inclusive/exclusive forms in all the cases where they would be required in Cree. Another slight difference is the Michif form ki~V-a:na:n (lpl incl.) versus Cree kı-V-a:naw. Rhodes (1977) gives only immediate imperatives. In our data we have some examples of delayed imperatives as well:

Michif: A Mixed Language based on Cree and French

317

(13) te:pwe:-hk-an pari e:-aja:-jan-ı. call-DIMP.AI-2 ready COMP-BE-AI.2-COND "Call when you're ready." (14) wi:pat pe:-ki:wke:-hk-e:k. soon come-visit-DIMP.-TA.2pl→1 "Come visit me soon." Although Michif Cree verbs are in principle identical to Plains Cree verbs, the Michif Cree verb seems to be subject to a tendency to use analytic constructions where synthetic constructions are possible and perhaps somewhat more common in non-Métis Cree. Benefactives, for example, tend to become analytic in Michif Cree. Example (15)a shows the normal Cree synthetic form, while (15)b has the analytic form, with pur nija "for me" (15) a. b.

en sæcyr IAFsg belt

d-ušt-am-a:-šu-n. l-make-BEN-TA-REFL-NON3

en sæcyr d-ust-am-e:n pur nıja. l-make.it-BEN-4sg for me "I am making myself a belt."

4.3.2. Non-Cree verb forms Michif uses two types of French verbs, though rather marginally. First, some French (or English) verbal lexical forms are integrated into Cree verb mor­ phology; and second, some French verbs are conjugated as they would be in French. Both types will be discussed. Some French and English verbal lexical items can take the full set of Cree verbal inflectional affixes. Many of these verbs seem to refer to nontraditional actions, e.g. bénir "to bless", témoigner "to witness in court", and bousiller "to plaster". Others, however, do not fall into this category: astiner "to argue", broder "to embroider", arranger "to fix", etc. The way in which these verbs are integrated into the Cree verb pattern is rather odd. It is always the infinitive form of the French verb that is used, together with a prefixed dummy element which is homophonous with the French masculine definite article le and the French object clitic le. We gloss it as a masculine article, as it is not used with intransitive verbs and as it varies in gender with similar nominal constructions (exx. 20 - 23). The French verb retains its original phonological structure. These verbs have Cree person and

318

Peter Bakker and Robert A. Papen

can also have Cree tense/mood marking and preverbs. Examples are given in (16) and (17): (16)

pε:r müd. DAMsg priest PAST-DAMsg-bless-INF-TA.3→3' DAMsg people "The priest blessed the people."

(17) kiskejım-e:w su mεtr know.TA-TA.3→3' John his master C0MP-DAMsg-argue-INF-3pl "John knows who's his master when they argue." (French obstiner) Some English verbs (such as to celebrate, to haul, to settle, to deal, to can, to box, and to gamble), many of them denoting nontraditional activities, can be used in a similar way: they retain English phonemes, but require the French masculine definite article as well as the French infinitive ending -er [i], and also Cree prefixes and suffixes. (18)

kat di žyjεt gi:- -sεlɨbre:t-t-na:n. DAMsg four of July l.PAST-DAMsg-celebrate-iNF-AI.lpl "We celebrated the Fourth of July."

(19) nu ki:-kıške:ht-am la bãk NEG PAST-know.TI-TI.3→4 DAFsg bank e:-ki:- -rab-i-hk. COMP-PAST-DAMsg-rob-INF-IMPRS

"He didn't know about the bank being robbed." The combination of Cree verbal affixes and French elements is also found in the formation of Cree verbs from French nouns or adjectives. Some of these are derived by means of the Cree suffix -ıwı "to be"; the appropriate Cree inflectional endings are then added. This is a productive process. The resulting verbs have the approximate meaning "to be X-y". The suffix -ıwı does not seem to be used with Cree stems, however. The French definite article, in the appropriate gender, must always precede the French noun or adjective. (20) ti suji la-bu-rw-an-wa. your.pl shoe DAFsg-mud-BE-4-pl "Your shoes are muddy." (French la boue "the mud")

Michif: A Mixed Language based on Cree and French

319

(21) ki:-te:pIšk-am e:-la-mitrεs-rwI-t. PAST-succeed-TI.3→4 COMP-DAFsg-teacher-- "She succeeded in becoming a teacher." (French la maîtresse "the teacher") (22)

ki:-li-fu-Iw-Iw e:-li-pči-Iw-it. PAST-DAMsg-crazy-BE-3sg.ANIM COMP-DAMsg-little-BE-3sg "He was crazy when he was small." (French le fou "the crazy", le petit "the small")

(23) ma:ma: gi:-It-Iku-na:n e:-ka mother l.PAST-speak-INV-TA.3→lpl NEG la-tεt-ǰyr-iwi-ja:hk. DAFsg-head-hard-BE- 1p1 "Mother told us not to be boneheads." (French la tête dure "the hard head") In another productive construction, the Cree verb suffix -Ihke: "to make" may be added to French adjectives and nouns. Again, the French definite article is required: (24) li-žali-Ihke:-w. DAMsg-pretty-MAKE-TA.3→3' "S/he is decorating him/her." (French joli "pretty") Apart from the integrated French verbs that take Cree affixes, there are French verbs that can be used in bare forms. Probably all Michif speakers use at least some French verbs, often frozen expressions like alũ "let's go", sa prā "one has to" (lit. "it takes"), and the French copula (see §4.3.3 and §4.6.1). Speakers who know French can, in principle, use any French verb, but French verbs are nevertheless relatively infrequent. Speakers sometimes use forms that are not the ones that whould be expected in French, e.g. participles instead of infinitives or vice versa. In (25)a-b below the verb is ambiguous between the French infinitive form in -er and the participial form in -é(e) but in (25)c-d participles are used, where French would have an infinitive: (25) a.

i va pas-i. "It will pass."

320

Peter Bakker and Robert A. Papen b.

i va bɔk-i. "He will buck."

c.

i va ufєr sa plas. "He will offer his seat." (cf. French il va offrir "he will offer (inf.)"/// a offert "he has offered (part.)")

d. i va dispary. "He will disappear." (cf. French il va disparaître (inf.)/// est disparu (part.)) Standard French imperatives can consist of the root (in the singular) or the root plus suffix Id (in the plural). The latter, written -ez in French, is phonologically homophonous with participial and infinitive suffixes (-é(e) and -er, respectively). Michif imperatives can be derived from French singu­ lar or plural imperatives. There seems to be no plural/singular distinction between these forms, and they are sometimes used in combination. (26) ta:š-i ãǰyr li pwæ. attempt endure the pain "Try to endure the pain." French imperatives can trigger a Cree conjunct clause: (27) dikuraž ka:ja šI-nakat-a:t. discourage.IMP.2sg NEG coMP-leave-TA.3-^3' "Discourage him from leaving her." Our data provide a few examples of verbs with cliticized pronouns, but most of them seem to be more or less fixed expressions: (28) i-m-a-ufer "He offered me." i-1-a-noz-amyzi "He amused us." sa-s-akord pa "They don't get along." ži-1-flεt / ži flεt "I flatter him." i-vœ-s-marji "He wants to get married." The only French verb forms found in Michif are infinitives/participles and analytic forms with past tense and future tense. No other forms (notably the synthetic future and several synthetic past tenses) are used in Michif. There is a special place for the verbs "to be" and "to have"; these are the only French verbs that have rather elaborate French paradigms. Only 2pl forms are lacking in our data.

Michif: A Mixed Language based on Cree and French (29)

321

"to be": INFINITIVE pur jet and: PRESENT

PAST

lsg: žł /žł swi 2sg: tł 3sg: ilł/ sł lpl: ũ-1-ł 3pl: i-sũ:

žł-t-ε t-łt-ε i-l-łt-e / ũ -1-łt-ε i-sũt-ε

FUTURE

žł tł s(ł)t-ε ũ

va jεt va jεt i / sa va jεt va jεt i vũ jεt

(30) "to have" (future not attested) lsg 2sg 3sg lpl 3pl

PRESENT

PAST

ž-a t-a i-l-a, sa-l-a ũ:-l-a i/sa-l-ü

ž-av-ε t-av-ε i-l-av-ε ũ-1-av-ε i/sa-1-ũv-ε

The use of both of these verbs is rather limited. The forms derived from French "to be" are often used as a copula (beside corresponding Cree verbs in equative clauses). The inflected forms of the French verb avoir "to have" are used mostly as a tense auxiliary with French verbs and also in a variety of (fossilized) verbal expressions such as avoir besoin "need" and il y a "there are". It is used as an auxiliary even where French would use a form of être "be", e.g. in i-l-a vny "he came" and i-l-a arivi "he arrived" (French il est venu, il est arrivé). 4.3.3. Reanalyzed French verb forms Modality is the only area in which Michif shows significant innovation compared to French and Cree. It is not clear why the modality system is the locus of so much innovation. In Michif a French adverb or phrasal expression is used to introduce a proposition in the conjunct order (also, rarely, a French infinitive or complement clause), as in (31) and (32). (31) æ sǣ äkur a

dog WISH

e:-ja:w-ak. coMP-have-TA.l-»3

"I wish I had a dog." (32) saprä

ti äjyr li træ. NECESSITY you endure the noise "You have to endure the noise."

322

Peter Bakker and Robert A. Papen

Below is a list, probably not exhaustive, of the modality markers found in our data: (33)

ãkur pa mwajæ magre fu(lε)-bæ sa-prã sa-s-pura-bæ sa-sεr-da-rjæ si-pa-si

(< F. encore "again") "I wish" (< F. pas moyen "no means") "it is impossible" (< F. malgré "in spite of') "oblige, coerce" (< F. fallait/faut bien) "it is/was necessary" (< F. ça prend "it takes") "it is necessary" (< F. ça se pourrait) bien) "it is possible" (< F. ça sert delà rien) "it is useless" (< F. je ne sais pas si) "it is doubtful"

Numerous French (and some English) adjectives and, more rarely, adverbs and nouns occur with the impersonal French expression si (< F. c'est "it is"). These can be in the present or in the past tense, and they can be positive or negative: (34) s-ɨ-d-valœr (< F. c'est de valeur) s-ɨ-tεrib (< F. c'est terrible) s-ɨ-bæ-rar (< F. c'est bien rare) sɨ-b -n-a-swεti(< F. c'est bien à souhaiter) s-ɨtε-pa-nɨsɨsεr (< F. c'était pas nécessaire) (< F./E. c'est pas fair) s-ɨ-pa-fεr s-ɨ-t-izi (< F./E. c'est easy)

"it's too bad" "it's awful" "it's rare" "it's to be hoped" "it was unnecessary" "it's not fair" "it's easy"

Cree has impersonal constructions — also used in Michif — which might have served as a model for the reanalysis of the French forms. These Cree constructions have a conjunct verb form, but this does not seem to be obliga­ tory; in contrast, the French-based forms always have a complement with a conjunct verb. Some Michif examples: (35) ki:ja:m či-pe:htwa:-hk uta. allright coMP-smoke-IMPRS here "It is all right to smoke here." (36) piku NECESSITY

či-šıpwe:hte:-ja:n. COMP-leave-AI.1

"I have to go." (37) ma:škut ka-ıšpajı-n. apt FUT-happen-II.4 "It will probably happen."

Michif: A Mixed Language based on Cree and French

323

4.4. The noun phrase The Michif noun phrase is fundamentally French, in all its syntactic functions — subject, direct or indirect object, object of an adposition, etc. Nouns are obligatorily accompanied by a French possessive element or article (definite or indefinite, masculine or feminine, singular or plural). Cree demonstratives (animate or inanimate) can be added to the French definite article. In all dialects a few Cree nouns are also used, but they are rare. Except in Camperville, Manitoba, the number of Cree nouns used in Michif now­ here exceeds fifteen. Cree nouns generally have a French determiner: lɨ we:pınıke:win "the garbage", æ wi:htıkuhka:n "a wolverine", εn wi:nma:kuš "a dowdy woman" (cf. §4.4.2). A great number of English nouns are used. These may be phonologically adapted to French or not (see §4.4.3), and they almost always have a French determiner. All adjectives are French (see §4.4.4); Cree has no distinct class of adjectives. Other categories within the NP can be either French or Cree. These include quantifiers, numerals and degree adverbials (see below). The internal order of the noun phrase is given in (38): (38) NUM - DEM - ART - ADJ - noun - ADJ - S 4.4.1. French nouns Most French nouns in Michif are identical to nouns found in archaic dialects of Canadian French, except for phonological differences (since a number of French nouns have been reanalyzed phonetically in Michif). Some vowelinitial nouns are reanalyzed with an initial consonant, either by agglutination of a consonant segment taken from a French determiner (39)a or by truncation of the initial syllable (39)b. See §4.2.1 for an argument that the initial consonants in (39)a are part of the noun rather than the result of a liaison rule. (39)

a.

AGGLUTINATION:

zafer zavœgl za:br lamu:r nivrɔñ nil

"business" "blind person" "tree" "love" "drunkard" "island"

( < les affaires ) (< les aveugles "the blind ones") (< les arbres "the trees") (< l' amour "(the) love") (< un ivrogne "a drunkard") (< une île "an island")

324

Peter Bakker and Robert A. Papen lɔm/nɔm/zɔm/tɔm "a man" (< l'homme "the man", etc.) lur/zur/nur "bear" (< l'ours "the bear" etc.) b.

TRUNCATION:

spjũ

"spy" "impolite" la lεksjũ "election" la sjεt; en sjεt "plate" la pɨtal, en pɨtal "hospital" früti

(espion) (effronté) (élection) (l'assiette) (l'hôpital)

The gender assigned to the truncated nouns in (39)b seems to be identical to the gender of the original French noun. The gender assigned to the noun in Michif, however, is not necessarily that of European French or even of Canadian French where the two varieties differ. Some gender differences are illustrated in (40) and (41): (40) Masculine in Michif, feminine in French: byt "butte, hill" muš "fly" ševr, šεv "goat" sum "whitewash" grif "claw" (41) Feminine in Michif, masculine in French: batũ "stick" mask "mask" bol "bowl" 4.4.2. Cree nouns There are only a few Cree nouns in Michif. Most of them denote berries, plants, animals, kinship terms, and common household objects. With few exceptions, they behave as if they were French nouns, although they retain a Cree phonological shape. For instance, they are pluralized by French articles instead of Cree suffixes. Only historical Cree borrowings into Canadian French (e.g. lepišu "bobcat" [Canadian French: "woolen slippers"], mocassin "moccasin", and mitasse "leggings") have French phonology. The Cree nouns musum "grandfather" and kuhkum "grandmother" (lit. "your grandmother") are used by Métis and Cree alike, even when speaking English or French. All Michif speakers make ample use of Cree nominalization possibilities. In Cree any verb can be nominalized by means of the nominalizing suffixes -kan ("instrument"), -win (neutral nominalizer), and -šk/-hk "repetitive V-er" or (-hk only) "action by unspecified actor(s)".

Michif: A Mixed Language based on Cree and French (42) lɨ pakamahı-kan

325

"the striker"

(. pakamahw-e:-w "s/he beats him") (43) lɨ "the toy" (. me:taw-e:-w "s/he is play­ ing") (44) lɨ we:pm-ıke:-wın "the garbage" (C.we:pm-ıke:-w "s/he throws things") (45) en pu:ju:-šk "a (fem.) quitter" (. pu:ju:-w "s/he quits") (46) æ dutama:ke:-hk "a panhandling" (. nıtutama:-ke:-w (AI) "s/he asks people for things") Cree nouns in -wın are always inanimate. All our examples have French masculine gender. Cree relative clauses can also be used as Michif noun phrases. Conjunct verb forms are in themselves relative clauses; these can take French determin­ ers. (47) (æ) ka:-ata:we:-t IAMsg C0MP-buy-3sg

"one who buys, a buyer"

(48) (æ) ka:-pe:hkısčı-ke:-t "one who cleans, a custodian" IAMsg C0MP-clean-DETRANS-3sg All in all, Cree nouns remain a marginal part of most of the Michif dialects. 4.4.3. English nouns As noted above, English influence on Michif is now pervasive. In fact, any English noun can be used in Michif, and it is hard to distinguish between code mixing and borrowing. English nouns usually have French (masculine or feminine) determiners. They agree in animacy/inanimacy with Cree demonstratives and verbs (compare Hogman 1981 for agreement in a Cree dialect with heavy French borrowing). Stress is often assigned according to Cree stress assignment rules (see §4.2.2.). French gender assignment is not well understood; it seems to be more or less arbitrary. It is often the same as in the equivalent French nouns (49), but sometimes it isn't (50). (49) la light lɨ base

"(electrical) light" (F. la lumière, une ampoule) (F. le but, le coussin)

(50) la fun æ binder

(F. le plaisir) (F. une moissonneuse)

326

Peter Bakker and Robert A. Papen

Plurality is only marked in the French article, not in the English noun: li stepsister "the stepsisters", li drug "drugs", li bεŋ "bengs" (traditional Métis food, from Canadian French beigne < French beignet). It should also be mentioned here that some French words that are plural in English also take a plural French article: li policik "politics" (French la politique). li polis "po­ lice" (also la polis; French la police). Lilla polis, even with singular article, always triggers plural agreement on the verb. 4.4.4. Noun modifiers Noun modifiers in Michif include articles, possessive elements, demonstra­ tives, quantifiers, relative clauses, adjectives, and adverbs. We will discuss all of these here. The French definite and indefinite article forms are used in Michif, in both genders. Only the plural indefinite article des is rare or nonexistent; the etymologically definite form li is used instead. The liaison rules of French do not exist in Michif (see §4.2.1). The forms of the article in the Turtle Mountain dialect are given in Table 3. Michif uses the complete French possessive paradigm (see Table 4), but see also §4.2.1 for discussion of the prevocalic forms and liaison. Cree nouns mostly have Cree possessive prefixes and suffixes (see Table 4). There are a few exceptions, such as musum "grandfather" and kuhkum "grandmother", which can have French possessive determiners, and probably also nouns formed by productive nominalizing (mentioned in §4.4.2). Both Cree and Michif (the Cree component as well as the French compo­ nent) distinguish between alienable and inalienable possession. Cree inalien­ ably-possessed nouns (mostly kinship and body parts) have an obligatory possession marker; nonpossession is marked by a prefix mi-, which cannot be used with other nouns. In French words, inalienably-possessed nouns always have a possessive element. (See also §4.6.7 below for possessive/genitive syntax.)

Table 3. French articles in Michif. INDEF.Sg MASC:

FEM:

INDEF.pl

DEF.Sg

DEF.pl

(F. un)

li (rarely di) (F. des)

li (F. le)

li (F. les)

en (F. une)

li (rarely di) (F. des)

la (F. la)

li (F. les)

Michif: A Mixed Language based on Cree and French

327

Table 4. French and Cree possessive elements in Michif. FRENCH POSSESSIVES: SINGULAR

masc. 1sg.: mu/mũ 2sg.: tu/tu 3sg.: su/sũ 1pl: nɔt/nut [2pl.: vɔt/vut 3pl: lœ

PLURAL

pre-V mũn tũn sũn nɔt/nut vɔt/vut lœr

fem. ma ta sa nɔt/nut vɔt/vut lœ

pre -C mi ti si nu vu lœ: /ly

pre-V miz tiz siz nuz vuz] 12 lœz/lyz

MICHIF AND CREE POSSESSIVES, with musum "grandfather" as stem:

lsg.:

nı-mušum

2sg.: 3sg.:

kı-mušum u-mušum-a

(51)

CREE:

lpl.excl.: nı-mušum-ına:n lpl.incl.: kı-mušum-ına:n (Cree: -maw) 2pl.: kı-mušum-ıwa:w 3pl.: u-mušum-ıwa:w-a

ALIENABLE POSSESSION:

nı-paskısıkan paskısıkan * mı-paskısıkan

"my gun" "a/the/someone's gun" "someone's gun"

INALIENABLE POSSESSION:

nı-ste:s * ste:s ? mı-ste:s

"my brother" "brother" "someone's brother".

FRENCH INALIENABLE POSSESSION:

mũ frεr * frεr

"my brother" "a brother"

Demonstratives in Michif are virtually always Cree. This produces a number of interesting phenomena, since Cree distinguishes animate and inanimate gender for nouns and demonstratives must agree in gender and number with the noun. Moreover, when a Cree demonstrative is used in Michif, it is accompanied by a form of the French definite article or a possessive. This article must agree in French (masculine/feminine) gender and number with the noun. The Cree verb also agrees in animacy/inanimacy with the (French or Cree) subject and/or the object. This means that nominal

328

Peter Bakker and Robert A. Papen

Table 5. Cree demonstratives in Michif.

ANIM.Sg. ANIM.pl. INAN.Sg. INAN.pl.

PROXIMAL

INTERMEDIATE

DISTANT

awa ũ:kık/u:kık u:ma ũ:hi:/ũ:hi

ana anĩkĩ/anıkık/enekık anıma anıhı/anĩhĩ/enehĩ

naha ne:kık ne:ma ne:hĩ, nεhi

forms must be doubly marked for two types of gender in Michif speakers' mental lexicons; Cree gender must be marked in the Cree verb and demon­ strative, and French gender must be marked in the French article. Some typical examples are given in (52): (52) awa lɨ garsũ awa la fij u:ma lɨ papji u:ma la bwεt anima la m zũ ne:ma lɨ šã ũ:hĩ li zafεr

"this boy" "this girl" "this paper" "this box" "that house" "that field" "those businesses"

Some speakers also use French demonstratives, but always marginally and in a limited number of contexts — specifically, only in temporal expressions. They do not seem to be very productive. The form of the French demonstra­ tive varies: [ɨsti, stɨ, st, ɨs, s, ɨst] (from popular French c'te). These may be fossilized forms in many cases. Some examples: (53) la sãdr ki:-mišı-pahkıstı-n is præta. the ash PAST-much-fall-II.3 DEM spring "We had an ash fallout this spring." (54) wi:-wi:wı-w st-atɔn. voL-marry-AI.3 DEM-fall "She'll get married this fall." Alternatively, the French demonstrative may be used to express the notion "such an X": (55) nu čı-ki:-tu:tam-an st-afεr. NEG COMP-PAST-do.TI-TI.2—4Sg DEM-thing "You shouldn't do such a thing."

Michif: A Mixed Language based on Cree and French

329

Quantifiers may be either French or Cree. Most quantifiers, whether French or Cree, require the presence of the French article between the quantifier and the noun, as in (56)-(60): (56)

li žεn d d ten plA young turkey "ten young turkeys"

(57) plys kɨ hen lɨ sɔkl more than one DAMsg plowshare "more than one plowshare" (58) æ pči bræ la sup IAMsg little bit DAFsg soup "a bit of soup" (59) misce:t lɨ mãži many DAMsg food "a lot of food" (60) j k pe:jak norãž dawe:jım-a:w. only one IAMsg orange 1.want-TA.1→3' "I want only one orange." Algonquian languages in general allow the preposing of certain NP elements, often quantifiers, thus creating discontinuous constituents (Dahlstrom 1987). This is also possible in Michif, as illustrated in (61): (61) kat gi:-aja:w-ak li mud la prɔmji ninety l.PAST-have-3pl DAMsg people DAFsg first žurni dɨ lã. day of the.year "I had 90 people over on New Year's day." Most relative clauses modifying nominais are Cree; these are typically conjunct clauses introduced by the complementizer prefix ka:-. (62) bakwat-a:w-ak li mũd ka:-kımutı-čik. l.hate-TA-3pl DAMsg people coMP-steal-3pl "I hate people who steal."

330

Peter Bakker and Robert A. Papen (63) ke:kwaj ka:-itwe:-ja:hk something -say-AI. 1EXCL "something that we said"

A few occurrences of French relative forms, possibly fossilized, have also been noted, with the shapes kɨ, kɨla, kɨlɨ; these are almost always followed by French predicates (adjectives, nouns). (64) æ ku en parson kili vjœ.... DAMsg blow IAFsg person REL.s/he.is old "Once a person is old..." (65) dilu kili bũ water REL.s/he.is good "fresh water" (66) awıjak kili drul someone REL.s/he.is funny "someone funny" There is no separate class of adjectives in Cree. All Michif adjectives are therefore of French origin. In Cree, noun-modifying elements are either noun prefixes or verbal constructions. These verbal constructions can have the form of relative clauses, which can be roughly equivalent in meaning to adjectives; as mentioned in §4.4.2, Cree relative clauses are common in Michif. (67) dilu ka:-wi:hkaši-hk water coMP-taste.good-II.3 "fresh water" (68) æ morsu:-d-vjäd IAMsg piece-of-meat "a big piece of meat"

ka:-mıša:-k C0MP-big-IL3

It also happens (fairly rarely) that Cree "adjective-like" elements are prefixed to French nouns: (69)

napakı-lɨ-pwεsũ flat-DAMsg-fish "flatfish" (Cree napakaw "it is flat"; compare napakı-tâpânask "a flat sled")

Michif: A Mixed Language based on Cree and French

331

French adjectives are typically either pre- or postposed according to the French pattern. When they precede the noun, they agree in gender and number with the noun they modify, as in French (70)-(71), but when they follow the noun they don't (unlike in French), as in (72)-(74): (70)

gru sarpã IAMsg big.MASC snake.MASC "a big snake"

(71) la grus DAFsg big.FEM "the big tent"

tat tent.FEM

(72) la m zũ blã DAFsg house.FEM white.MASC (French: la maison blanche [blãš]) "the white house" (73) en fam IAFsg wife.FEM "a new wife" (74)

nœ new.MASC

(French: une femme neuve?)

lɨ bwa: šεš DAMsg wood.MASC dry.FEM (French: le bois sec) "the dry wood"

French adjectives may be reinforced with vre (< F. vrai(e) "real") or by reduplication. Both these reinforcements are also known in Canadian French dialects, and in Michif: (75) en vre žali rob IAFsg real nice dress "a real(ly) pretty dress" (76) li vjœ-vjœ kı-musum-ma:n DAMsg old-old 2-grandfather-21 "our (incl) very old grandfather" Comparatives and superlatives can be French or Cree or mixtures of both systems. (77) li ste:k s-i la ply mijœr the steak it-is DAFsg most best "Steak is the best meat."

vjãd. meat

(French: la meilleure viande)

332

Peter Bakker and Robert A. Papen (78)

tũ nawat mıšıkıtı-w DAMsg horsefly more big-AI.3 "A horsefly is bigger than a fly."

ašpıcı εn muš. compared.to DAFsg fly

In Cree there are uninfected adverbs, e.g. ki:htwa:m "again" and mıtunı "much". These are used in Michif too, along with French adverbs, such as a pi pre "approximately", di kučym "usually", and surtu "especially". Things that are expressed by verb-modifying adverbs in other languages can be expressed in Cree by preverbs. In Michif this is also common: (79)

mıyu-atuške:-w. well-work-AI.3 "He works well."

(80)

li bɔn mark a likɔl mıšı-mıjustam-ıh-ıku-w. plA good mark at the.school big-be.glad-CAUS-INV-3 "Good marks in school make him very happy."

In French, manner adverbs can be formed from adjectives by addition of the suffix -ment, which is productive in French but does not exist in Michif. Michif often uses the Cree preverbs, but adverbs are also apparently derived from French adjectives by the addition of the preposition ã (< F. en) in front of the adjective: (81)

či:stastam si parɔl ã grus. pick.TI.3→4 POSS-3pl word in big "She speaks in an affected way."

(82) ä rof pi:kıškwe:-w. in rough speak-AI.3 "He talks in a rough way." 4.4.5. French derivational and compositional processes New concepts seem to be expressed in Michif by circumlocution rather than composition or derivation. French compositional rules are severely reduced in Michif, and only a few French derivational suffixes are used productively. These are added to French stems and sometimes to English stems; most of them are nominalizing suffixes that have been used to create words not found in French dialects:

Michif: A Mixed Language based on Cree and French (83) a. lɨ vɔl-až "theft" (F. vol- "steal" + F. -age; cf. F. le vol "theft") b.

sykr-až "a sweet" (F. sucre "sugar" + F. -age; cf. F. bonbon)

c.

trɔst-ab "trustworthy" (Engl, trust + F. able; cf. F. fiable)

d. bεg-œr, bεg-œz "beggar" (Engl, beg + F. eur/euse; cf. F. mendiant(e)) e.

avar-isœ "miser" (F. avare + F. isseux, isseuse ; cf. F. avare)

f.

ažãsri "office" (F. agence + F. -erie; cf. F. agence)

g. kũtãtri "happiness" (F. content "happy" + F. -erie; cf. F. bonheur) h. lãtũ, lãtɔn "slowpoke" (F. lent "slow" + F. -on, -onne; cf. F. lambin{e)) i.

kũfidasjũ "confidence" (F. confid(ence) + F. -ation; cf. F. confidence)

j.

pwεsunjĩ "fisherman" (F. poisson "fish" + F. ier; cf. F. pêcheur)

k. pruvas "proof (F. prouv(er) + F. asse; cf. F. preuve) We have very few examples of composition-like contructions: (84) a. b. c.

fe-lɨ-s "a hypocrite" (lit. F. faire le saint; cf. un(e) hypocrite) fε-lɨ-grɔs "a boaster" (lit. F. faire le gros; cf. un vantard) en fε-la-bɔn "a prude" (lit. F. faire la bonne; cf. une prude)

333

334

Peter Bakker and Robert A. Papen (85) 1-aržã pa-vrε "fake money"

(Lit. F. l'argent pas-vrai)

The pattern most often used for extending the lexicon is circumlocution, creating analytic constructions rather than compositions (usually N P N, often N à N or N de N). (86) a.

en butej ã vitr DAFsg bottle in glass "a carafe" (French: une carafe)

b.

en m zu d log DAFsg house of log "a log house" (French: une cabane)

c.

pči pwesu d-or DAMsg little fish of-gold "a goldfish" (French: poisson rouge)

4.5. Prepositons and postpositions French has prepositions; Cree has postpositions, prepositions and a locative nominal affix -ıhk/-ohk. Michif uses both the French and the Cree forms, but again there is considerable dialectal variation. Some speakers (e.g. many of those who live in Turtle Mountain and southern Manitoba) use French prepositions almost exclusively, others (e.g. those of Camperville, Manitoba) mostly use Cree adpositions, and still others use both systems. Michif preposi­ tions of French origin often derive from the related French adverbs rather than prepositions — as also happens sometimes in Canadian French (and in older stages of European French). For instance, instead of the French preposition derrière "behind" Michif uses anarjcr "behind", derived from the French adverb en arrière "in the back" (87b). Some French prepositions have been reanalyzed to express typically Cree functions. For example, the Michif preposition da/dã/sa (< F. dans "in") 13 no longer means uniquely "in(side)", as it does in French. Rather, it has taken on the Cree locative meaning: (87) a.

dã lɨ uhčı in the fridge from "out of the fridge" (Cf. Cree tahkascikan-ihk ohci, lit. "fridgeLoc from")

Michif: A Mixed Language based on Cree and French

335

b. ãnarjεr dã lɨ magaz behind in the store "behind the store" (not "in the back of the store") c.

utahk da lɨ ša:r "behind the car" (not "in the back of the car") (Cf. Cree otahk ota:pa:na:sk-ohk, lit. "behind car-LOC")

(88) a:tıht manıša li brãš dã lɨ za:br uhčı. some cut.off.IMP.TI2 plA branch LOC DAMsg tree from "Cut off some of the branches from the tree." In Michif the French-derived preposition da seems to have absorbed the function of the Cree locative suffix, as both are used in combination with certain prepositions and postpositions. Some prepositions merge with articles in French, but they merge differ­ ently in Michif: (89)

de + le → de + les → à + le → à + les → dans + un → d+un →

FRENCH

MICHIF

du dɨ des au aux dans un d'un

(dy in some fossilized expressions) li (di in some fossilized expressions) - ( in some fossilized expressions) dæ dæ

The preposition di (< F. de) is not always used in Michif as it is in French. The French semantic distinction between partitive and nonpartitive forms is lost in Michif; in addition, it seems to be possible to omit di in a number of cases: (90) la šeyer ili pl li sab. DAFsg pail it.is full the sand "The pail is full of sand." (French: pleine de sable) (91) mi suyi i-sü pl d sab. my.pl shoes are-3pl full of sand "My shoes are full of sand." (92)

žyp di rob pl li pli d akorjijü IAMsg skirt of dress full plA ruck of accordion "a skirt full of accordion rucks" (French: pleine de plis)

336

Peter Bakker and Robert A. Papen

There are some fossilized partitive forms, e.g. dilu "water" (from French de Veau "water"; cf. Veau "the water"), which has indefinite and definite mean­ ing — as does lu "water", which also occurs in Michif. Finally, preposition stranding occurs with some prepositions in Michif, just as it does in Canadian French: (93) la mɨčɨn e:-nıpe:-h-ıwe:-hk DAFsg drug coMP-sleep-CAUS-INDFZR-IMPRS "The drug people are put to sleep with."

avεk. with

4.6. Michif syntax Cree is a nonconfigurational language with relatively free word order. Stan­ dard French has a rather rigid word order and a clearly articulated syntactic structure; spoken French, however, often diverges from this pattern and has much freer word order. Michif syntax is basically that of Cree — that is, its word order is relatively free. Of course, the more French elements are used, the closer the syntax conforms to spoken or written French norms. Some examples of different sentential word orders are given in (94)-(99). (94) SVO: la žyma: ki:-aja:w-e:w pči pul . DAFsg mare PAST-have-TA.3→3' IAMsg little foal "The mare had a foal." (95) SOV: kahkıja:w awıjak la pwi dawe:stam-w-ak. all somebody DAFsg rain want.it-TI3pl-4sg "All the people want rain." (96) VSO:

ki:-wanıst-a:w lɨ žwal su liku. PAST-lose.it-3 DAMsg horse his halter "The horse lost its halter."

(97) OVS: la kuron ka-kiskaw-e:w Prince Charles. DAFsg crown.ANIM FUT-wear-TA.3→3' Prince Charles "Prince Charles will wear the crown." (98) VOS: ki:-ucıpıt-am sa tεt la tɔrčy. PAST-pull-TI.3.4sg his head.INAN DAFsg turtle "The turtle pulled back his head."

Michif: A Mixed Language based on Cree and French (99) OSV:

337

be:bi la pr ses ki:-aja:w-e:w. IAMsg baby DAFsg princess PAST-have-.→3' "The princess had a baby."

Sentences which consist only of French elements normally follow colloquial French word order: (100) žɨ rɨspεk mu nuk. "I respect my uncle." (101)

li žard d čwɨzɨn i furni tutli žardinaž. DAMsg garden of kitchen 3SUBJ furnish all plA vegetable "The kitchen-garden furnishes all the vegetables."

If a French verb is used, all nominal subjects require a resumptive pronoun (subject reprise), which is always based on the French masculine form, seemingly always singular: (102) Delia il-a fe sãbiã. Delia he-has made seeming "Delia used to pretend." (Fr.: Delia (elle) a fait semblant) (103) žyl il-a ãfarži lɨ žwal. Jules he-has hobble(d) DAMsg horse "Jules hobbled the horse." English NP syntax sometimes intrudes in an otherwise French NP: (104)

la tεt ruž pik-dɨ-bwa il-ɨ gru. DAFsg head red woodpecker he-is big "The red-headed woodpecker is big." (F.: le pic de bois à tête rouge)

(105)

lɨ pči šavæž bibi DAMsg little Indian baby sauvages)

(F.: le petit bébé sauvage/ des

(106) æ vre bũ balãsi rɨpa IAMsg real good balanced meal "a really well balanced meal" (F.: un très bon repas équilibré or un repas très bien équilibré)

338

Peter Bakker and Robert A. Papen

4.6.1. Copula constructions There are four kinds of copula constructions in Michif: a Cree-based predica­ tive construction using verbal adjectives or nouns, an innovative equative construction without any verb or copula, a Cree-based identifying construc­ tion using the Cree demonstrative determiner, and a French-based equative clause construction using the French copula. Cree has no copula as such. Copula constructions are expressed by an NP and a verbal adjective or noun, in any order. The Cree verb can be either a stative verb or a verbalized noun with a copula-like suffix -ıwı. This is the first type of Michif copula construction. The Cree verb agrees in animacy with the noun (object or subject), in both Cree and Michif. Michif examples: (107) papıkwa:-w lɨ 1 ž. be.rough-AI.3 DAMsg cloth "The cloth is rough." (108) la bwet misa:-w. DAFsg box be.big-II.3 'The box is big." (109)

lɨ za:br mıšıkıtı-w. DAMsg tree be.big-AI.3 "The tree is big."

The -IWI suffix does not seem to be used with Cree stems in Michif, strangely enough, although it is quite common in Cree. But in Michif it is used productively only with French nouns and adjectives, always preceded by the French definite article, as in (110): (110) la pusjer-Av-an lɨ pɔrtmãtũ. DAFsg dust-BE-. DAMsg suitcase "The suitcase is dusty." In the second type, which is relatively rare, the noun and its predicate (adjective or noun) may simply be juxtaposed in any order, producing an equative sentence: (111) en krem la žymã. IAFsg cream DAFsg mare "The mare is a buckskin."

Michif: A Mixed Language based on Cree and French

339

The third type is an identifying construction, in which a Cree demonstra­ tive is used along with a (usually French) noun as a subject. The predicate always precedes the demonstrative. (112)

za:br IAMsg

tree

ana. DEM.ANIM.sg

"That is a tree." (113) ma tãt my

aunt

ana.14 DEM.ANIM.Sg

"That is my aunt." (114)

n-arabjæ ana lɨ IAMsg Arabian DEM.ANIM.Sg DAMsg "That horse is an Arabian."

žwal. horse

Rhodes (1977:13) found an interesting variant of this construction. When the predicate noun is possessed (alienably), the possessor functions as the predi­ cate, and the logical predicate noun is displaced to the right. This is an emphatic construction. (115) nıja u:ma mu papji. 1sg DEM.INAN.sg my-MASC paper "This is my paper." The final type of copula construction involves a French copula, derived from the inflected forms of the French verb être, with cliticized pronouns (a list of these forms can be found in §4.3.2). Both predicate-subject and subjectpredicate orders occur: (116) e:waku la rob ili kwarεk. that DAFsg dress COP correct "That dress is appropriate." (117)

sɨ žali larãžmã d flœr. COP nice DAFsg.arrangement of flower "The flower arrangement is nice."

4.6.2. Interrogatives In this section we will discuss polar questions (yes-no questions), WH-questions, and embedded questions.

340

Peter Bakker and Robert A. Papen

Polar questions in Michif differ from declarative sentences in having a question element čĩ or či (Crée ci),15 either after the first major constituent or at the end of a sentence. (118)

la munisjü: čĩ kıt-a:ja:-na:n šı-papa:-ma:či:-jahk ? DAFsg ammunition Q 2-have-21 coMP-about-hunt-AI.21 "Do we (incl.) have ammunition for hunting?"

(119) kı:-kı:muci-natustaw-a:w-ak ti: zami: čĩ? 2.PAST-secretly-listen.TA-3-pl your.pl friend Q "Did you eavesdrop on your friends?" There are also cases where we don't find the question element: (120) ki:-kaškıst-e:n kekwaj ka:-wi:-ušt-a:-jan? 2.PAST-be.able-4sg what coMP-voL-do-DIR-2sg "Were you able to do what you wanted to do?" Interrogative pronouns and interrogative adjectives are fundamentally Cree. The latter agree in gender (animate/inanimate) and in number with the noun they modify. The most important question words are listed in (121)-(122) and some examples are given in (123)-(124). (121) awi:na, awe:na ki:kwe:, ke.'kwaj ta:nde ta:ne:hkı, ta:n hkı ta:nšı, ta:šı, ta:šıšı ta:nšpi:

"who" "what" "where" "why" "how" "when"

(122) "which": ta:na ta:nima ta:nıhı

(anim. sg.) (inan. sg.) (inan. pl. /anim. obv.)

(123) tama lɔm ka:-naskwe:ht-ahk? which man C0MP-answer-TI.3 "Which man answered?" (124)

ta:nıhı li livr ka:-utınam-a:n? which.pl plA book coMP-take-TI.l→4 "Which books shall I take?"

Michif: A Mixed Language based on Cree and French

341

A few French question words are also used, but they are rare: (125)

kɨ laž kɨ-t-a? which age that-you-have "How old are you?"

(126) kubjæ di fwe ka:-wa:pastam-an la niž? how.many of time coMP-see.it-2 DAFsg snow "How many times did you see the snow?" (i.e. "how old are you?") (127)

kεl sort di flœr anima? which kind of flower that "What kind of flower is that?"

Embedded polar questions sometimes use the Cree form ki:spm "if, although in Cree itself embedded questions are given in the form of direct questions. The use of these embedded questions in Michif seems to be due to French influence (although Drapeau, personal communication 1993, says these con­ structions also exist in the closely related Algonquian language Montagnais). (128) gwe:čım-ık ki:spm čı-wı:čıhıh-ıwe:-ja:n. 1 .ask-TA.INV.3 if COMP-help.TA-INDEFZR-1 "He asked me if I would help people." In most cases embedded WH-questions are identical to nonembedded sen­ tences, with the exception of the use of ıšpi: instead of taspi: and ıta or ıte: instead of ta:nde: as relative question markers (see §4.6.3), the forms that would be required in Cree also. (129) nı-nıšıtut-e:n ta:nšı ıtwe:-jan. l-understand-TI.4sg how speak-AI.2 "I understand the way you speak." 4.6.3. Subordinate clauses Apart from the relative clauses mentioned in §4.4.2. and §4.4.4., there are also local/temporal relatives and adverbial clauses. First, there is a locative rela­ tive form ıte:/ıta "where", and there is a temporal relative pronoun ıšpi: which is sometimes combined with a French relative clause:

342

Peter Bakker and Robert A. Papen (130) ka-ıtuhte:-h-ıtı-n ıte: e:-wi:-tuhte:-jan. 2.FUT-go-CAUS-TA.INV-NON3 where COMP-VOL-go-AI.2sg "I will take you where you want to go." (131) bakwat-e:n čı-tuhte:-ja:n d-æ nikol ıta 1.hate-TI.3→4 -go-AI.1 in-IAMsg school where strik. k-ɨl-ɨ C0MP-3sg-BE strict "I don't like to go to a school where they are strict."

Most adverbial conjunctions are from French; they control either a Cree conjunct clause or a French complement clause: (132) nu pe:ht-am akuz e:-pa:kı-payın-ıyı-k su NEG hear-TI.3→4 because COMP-thick-become-PS-IL4 his zaraj. ear "S/he doesn't hear it because of an ear infection." (133) pe:hta: zyskatã či-takušımı-ja:n. wait until coMP-arrive-AI.1 "Wait until I arrive." (134) sit brav la fij apre k-ɨl-a su:vi 3.BE.IAMsg brave DAFsg girl after -3-havesave.PART be:bi. lɨ DAMsg baby "She was a heroine after she saved the baby." A very peculiar construction, using French lexical material but nonexistent in French, is used as a manner adverbial: (135) pa mwaj lœz aržã šı-akım-ım-ıht kisũ no means their money coMP-count-PO-IMPRS COMP.3.BE.pl riš. rich "They can't count their money, they're so rich." (136) ki:-tu:hka:pı-w k-ɨl-ɨtε syrpri. PAST-have.eyes.open-AI.3 -3-was surprised "S/he was open-eyed with surprise."

Michif: A Mixed Language based on Cree and French

343

(137) nu mıhče:t meri či-lɨ-bud -hke:-t k-ɨl-ɨ NEG much Mary COMP-DAMsg-pout-make-AI.3 -3-is tužur budœz. always moody "It doesn't take a lot for Mary to pout, she's so moody." 4.6.4. Coördinating elements In Michif two NPs can be connected by pi "and" (from French puis "and then") or, rarely, by avik (F. avec "with"). French et is apparently used rarely or not at all. Sometimes the Cree forms e:kwa and mi:na are used to connect NPs. French ub or ubædũ "or" (from F. ou bien/ou bien donc "or"; these are emphatic forms — the normal French form is ou) is used for "or". Sentences can also be connected by e:kwa "and" and pi, and also by ub dũ. Cree and French verbs are sometimes connected by pi: (138) pi:kupıč-ıka:t-e:-wa pi dɨsk-i. fallow-PASS.INAN-DIR-4pl and disk-INF "They (inan.) are fallowed and disked." 4.6.5. Obviation Cree obviation marking is reduced in Michif. Obviation in Cree, which is marked in the verb as well as in the noun, distinguishes two (or more) third persons within a phrase, a sentence, or a stretch of discourse. The third person not mentioned previously is marked "obviative" (-a after consonants, -wa after vowels), while the known third person remains unmarked and is called "proximate". Obviation is not marked on inanimate nouns in Plains Cree, and number is not marked on obviative nouns. Animate third persons possessed by third persons obligatorily have the obviative suffix in Cree, but in Michif only Cree nouns have obviative markers, not French nouns: (139)

u-ma:ma:-wa his/her-mother-OBV "his/her mother"

(140) *su š -wa his/her dog-OBV "his/her dog"

344

Peter Bakker and Robert A. Papen

When two third persons do not have a relation of possession, but are part of the same discourse, the newly mentioned noun phrase receives obviation marking as "Irene" does in (141): (141) John ki:-wa:pam-e:w Irene-a. John PAST-see-→3' Irene-OBV "John saw Irene." This use of what we could call syntactic obviation is reduced by comparison to Plains Cree, where it is obligatorily used for animates (and never for inanimates). In Michif, obviation is always marked on personal names, often on other nouns referring to humans, sometimes on nouns referring to animals, and never on inanimate entities. (142)

nɔm dã lɨ bãd dɨ mud IAMsg man in DAMsg crowd of people ki:-nipa-h-e:w Bobby Kennedy-wa. PAST-die-CAUS-→3' B.K.-OBV "A man in the crowd killed Bobby Kennedy."

(143) Pitœrli bũ y-wa ki:-pimičišahw-e:w. Peter DAMsg good god-OBV PAST-follow-TA.3-^3' "Peter was a disciple (lit. "followed the Lord")." It must be stressed, though, that even when French forms are not marked for obviation, an obviative agreement marker is nevertheless usually present on the Cree verb, as in (144) (and see 159 for an examples without -ji "obviative"). (144) bæcis su frer nipa:-ji-w. John his brother sleep-OBV-3sg "John's brother is sleeping." If a third person possessed noun is the object of a verb, it has an agreement marker -ım- on the verb (see ex. 135 above). 4.6.6. Negation Both French and Cree negation occur in Michif. Cree has different negators for main clauses (namo:ja and variants), for embedded clauses (e:ka:wija, mostly shortened to e:ka:), and for imperative clauses (ka:ja). Namo:ja is also used to answer polar questions.

Michif: A Mixed Language based on Cree and French

345

Standard French has two main negators: non is used as a negative answer to polar questions and ne...pas is used elsewhere, the two elements separated by the verb and its clitics. Other French negative adverbs are ne...rien "nothing", ne...personne "no one", ne...plus "no longer", etc. In spoken French ne is usually omitted. The word pas is also used to negate nouns, which are given in the partitive form with de "of, e.g. pas de pommes "no (of) apples". By far the most common negative particle in Michif is nu (most likely from French non). This particle is typically used with nonimperative Cree verbs. It usually occurs immediately before the verb in independent or main clauses, and it immediately precedes the complementizer in negative embed­ ded clauses. In both cases Cree would use e:ka. If there is an adverb, nu immediately precedes it, in an innovative use of the French-origin negative element. (145) nu

ka-pakıtm-ıka:šu-w

NEG FUT-allow-PASS.ANIM-3

čı-ıtuhte:-t. COMP-go-3

"She won't be allowed to go." (146) nu kı-tıpe:jım-a:w lɨ minuš. NEG 2-own-TA.l/2→3 DAMsg cat "The cat doesn't belong to you." (147) li nasjŭ æ treti aja:-w-ak nu plA nation IAMsg treaty haveTI-3-pl NEG šı-nu:tm-itu-čık. COMP-fight-RECIP-3pl

"Nations have treaties not to fight each other." (148) nu tultã aja:-w a lɨkɔl. NEG always be-AI.3 in the.school "S/he is not always at school." In Michif, nouns are usually negated by the French particle pat (< F. pas de), and pa is used to negate adjectives. French verbs are sometimes negated with pa (never accompanied by the element ne, just as in colloquial French), but these may all be fixed expressions. Here, negation follows French syntax, preceding the noun and adjective but following the verb:

346

Peter Bakker and Robert A. Papen (149) lɔm pat barb ana. man NEG beard DEM.ANIM.sg "That's the man without a beard." (150) pat wicıhıwe:wın. NEG help "There is no help." (151) žɨ pãs

pa

šı-ıtuhte:-t.

I think NEG COMP-go.AI-3

"I don't think he will go." The most common Cree negator in Michif is namu(:), which precedes the verb, the noun, or the adverb it negates. It occurs in Michif in the same position as nu, with one exception: nu is not found in embedded sentences. (152) namu NEG

nı-wi:-ituhta:-n. l-VOL-go.AI-NON3

"I won't go/I don't want to go." (153) mısče:t laržã aja:w-e:w ma:ka namu: much the.money have.TA-3→3' but NEG fatas. braggart "He has a lot of money but he is no braggart." (154) namu: wa:hıjaw d-aja:-na:n a Bismarck NEG far 1-BE-AI.lpl to Bismarck "We are not far from Bismarck."

æ IAMsg

unčı. from

The Cree negator e:ka: is also used in Michif. As in Cree, it always modifies embedded clauses and precedes the conjunct verb. In Michif these can be Cree-based or French-based: (155) ahkıkı-n la mus ita lɨ salej e:ka: grow.II-4sg DAFsg moss where DAMsg sun NEG e:-wa:še:n-ike:-t. COMP-shine.AI-DETRANS-3

"Moss grows where the sun doesn't shine." As in Cree, imperative and subjunctive verbs (and rarely conjunct verbs) are negated by the Cree form ka:ja.

Michif: A Mixed Language based on Cree and French

341

(156) ka:ja napač-im-in. NEG bother-with words-IMP.TA.2→l "Don't hassle me." It should also be mentioned that some Cree negative + noun/ adverb construc­ tions are used in Michif too. In Cree (and Michif) these form a phonological unit, e.g. Cree namo:ja na:nita:w "nowhere, nothing", Michif nama na:nda:w or nu: na:nda:w; Cree namoja ki:kwe:, often nama: ki:kwe: or ma:ki:kwe: "nothing, not at all", Michif namake:kwaj or no ke:kwaj; and Cree namawijak "nobody", Michif nu awijak. These combinations can function as negators too. They seem to have scope only over nouns, not over verbs or sentences. (157) namake:kwaj 1 syk ašte:-w. nothing DAMsg sugar stand-3 "There is no sugar." 4.6.7. Possessive constructions We noted in §4.4 that the Michif noun phrase is fundamentally French; this includes syntactic ordering. The only exception to this is in possessive and genitive constructions, where syntactic ordering is more typical of Cree than of French (although Métis French also has this construction). This is true even if French possessive determiners are used. Here is a typical example: (158)

mữ pči gan Isg.poss little boy "my son's pony"

pči žwal P0ss.3sg little horse

The equivalent French phrase would be le petit cheval de mon petit garçon/ fils. Cree would have ni-kosis o-mistatim-wa (lit. my-son his-horse-OBv). Note that in Cree the possessed noun requires the obviative marker if it is possessed by a third person. In Michif the possessed noun will be marked for obviation only if it happens to be a Cree noun, as in (159): (159) Joe u-ma:ma:-wa ki:-api-w dā la šãbr Joe 3sg.POSS-mother-OBV PAST-sit-3 in DAFsg room ka:-pe:ta:-hk. COMP-listen?-3lNDEF.ACTOR

"Joe's mother was sitting in the parlour."

Peter Bakker and Robert A. Papen

348 5.

Origin of Michif

We turn now to the origin of Michif, discussing first the lexical, phonological, semantic, morphological, and syntactic components and then the historical context in which Michif came into being. 5.1. Lexicon As is clear from the discussion above, the Michif lexicon comes from two main sources: French and Cree. English is now also a major influence. Speakers clearly know which elements originate from which language. It seems that French verbs can (in principle) always be used by speakers who know (some) French and who suppose the listener understands it. The ad­ dressee plays a major role in the lexical choice of the speaker. People who do not know French also use French verbs, some of them integrated into Cree verb morphology (§4.3.2). The same is true for Cree nouns: the less Cree a speaker knows, the smaller the chance of Cree borrowings in the nominal component. One could argue that these nouns behave as if they were borrow­ ings into French. The approximate distribution of French and Cree elements in Michif is given in Table 6. The different ranges within some categories refer to ex­ tremes of 17 individual speakers from five different communities; see Bakker 1996 for details. Table 6. Distribution of French and Cree elements in Michif. The figures are percent­ ages, drawn from Bakker's corpus of questionnaire data based on 17 Michif speakers in five communities

Nouns: Verbs: Question words: Personal pronouns: Adverbials: Postpositions: Coörd, conj.: Prepositions: Numerals: Demonstratives: Negators:

FRENCH

CREE

OTHER

83-94