Improving Literacy by Teaching Morphemes (Improving Learning)

  • 44 200 1
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up

Improving Literacy by Teaching Morphemes (Improving Learning)

Improving Literacy by Teaching Morphemes Words consist of units of meaning, called morphemes. These morphemes have a st

1,708 611 2MB

Pages 223 Page size 432 x 648 pts Year 2006

Report DMCA / Copyright


Recommend Papers

File loading please wait...
Citation preview

Improving Literacy by Teaching Morphemes

Words consist of units of meaning, called morphemes. These morphemes have a striking effect on spelling that has been largely neglected until now. For example, nouns that end in “-ian” are words that refer to people, and so when this ending is attached to “magic” we can tell that the resulting word means someone who produces magic. Knowledge of this rule, therefore, helps us with spelling: it tells us that this word is spelled as “magician” and not “magicion”. This book by Terezinha Nunes, Peter Bryant and their colleagues shows how important and necessary it is for children to find out about morphemes when they are learning to read and to spell. The book concentrates on how to teach children about the morphemic structure of words and on the beneficial effects of this teaching for children’s spelling and for the breadth of their vocabulary. It reports the results of several studies in the laboratory and in school classrooms of the effects of teaching children about a wide variety of morphemes. These projects showed that schoolchildren enjoy learning about morphemes and that this learning improves their spelling and their vocabulary as well. The book, therefore, suggests new directions in the teaching of literacy. It should be read by everyone concerned with helping children to learn to read and to write. Terezinha Nunes is Professor of Educational Studies at the University of Oxford and Fellow of Harris-Manchester College, Oxford. Peter Bryant is Visiting Professor of Psychology at Oxford Brookes University and Emeritus Fellow of Wolfson College, Oxford.

Improving Learning TLRP Series Editor: Andrew Pollard, Director of the ESRC Teaching and Learning Programme

Improving Learning How to Learn: Classrooms, schools and networks Mary James, Paul Black, Patrick Carmichael, Mary-Jane Drummond, Alison Fox, Leslie Honour, John MacBeath, Robert McCormick, Bethan Marshall, David Pedder, Richard Procter, Sue Swaffield, Joanna Swann and Dylan Wiliam Improving Literacy by Teaching Morphemes Terezinha Nunes and Peter Bryant Improving Schools, Developing Inclusion Mel Ainscow, Alan Dyson and Tony Booth Improving Subject Teaching: Lessons from research in science education John Leach, Robin Millar, Jonathan Osborne and Mary Radcliffe Improving Workplace Learning Karen Evans, Phil Hodkinson, Helen Rainbird and Lorna Unwin

Improving Literacy by Teaching Morphemes

Edited by Terezinha Nunes and Peter Bryant with Ursula Pretzlik and Jane Hurry

First published 2006 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016

This edition published in the Taylor & Francis e-Library, 2006. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to” Routledge is an imprint of the Taylor and Francis Group, an informa business © 2006 editorial matter and selection, Terezinha Nunes and Peter Bryant; individual chapters, the contributors All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book has been requested ISBN10: 0–415–38312–9 (hbk) ISBN10: 0–415–38313–7 (pbk) ISBN13: 978–0–415–38312–7 (hbk) ISBN13: 978–0–415–38313–4 (pbk)

We dedicate our book to Nick Pretzlik whose kindness and cheerful support we remember with great pleasure


List of illustrations Series editor’s preface Acknowledgements

ix xiii xv

Part I

What is the issue?




Morphemes and literacy: A starting point PETER BRYANT AND TEREZINHA NUNES


What knowledge of morphemes do children and adults show in the way that they spell words?




What does the research tell us?






An intervention program for teaching children about morphemes in the classroom: Effects on spelling FREYJA BIRGISDOTTIR, TEREZINHA NUNES, URSULA PRETZLIK, DIANA BURMAN, SELLY GARDNER, AND DANIEL BELL


viii Contents


An intervention program for classroom teaching about morphemes: Effects on the children’s vocabulary




Can we increase teachers’ awareness of morphology and have an impact on their pupils’ spelling?




What are the overall implications? 7

Morphemes and literacy: Context and conclusions




Appendix The four research strategies in this research program



References Index

191 195


Figures 1.1 The first two pages of a 71⁄2-year-old girl’s story 1.2 Overgeneralizations of the “-ed” ending by a 71⁄2-year-old boy 2.1 Percentage of children who spelled each suffix (“-ion”, “-ness”, and “-ed”) correctly, by age level 2.2 Percentage of children who spelled each suffix (“-ion” and “-ian”) in words and pseudowords correctly, by age level 2.3 On the left: Number of correct spellings of regular and irregular verbs in the past and nonverbs ending in /t/ or /d/. On the right: Generalization of “-ed” to the wrong words 2.4 Proportion of past regular verb endings spelled correctly and produced correctly for pseudowords in an oral task 2.5 Pictures of dinosaurs with their names, which the children were asked to spell 2.6 Proportion of word and pseudoword pairs whose stems were spelled in the same way at each age level 2.7 Proportion of real verb endings spelled correctly with “-ed” and proportion of stems spelled consistently across two words 2.8 Percentage of correct pseudowords with “-ion” and “-ian” spelled correctly and percentage of correct explanations, by age level 2.9 Percentage of correct spellings of one-morpheme and two-morpheme words, by age level 3.1 Design of the first teaching study

26 28 39


45 49 52 53


57 60 68

x Illustrations


The mean number (out of 16) of correctly spelled “-ion” and “-ian” endings in real words in Study 1 3.3 The mean number (out of 8) of correctly spelled “-ion” and “-ian” endings in pseudowords in Study 1 3.4 The mean number (out of 16) of correctly spelled “-ion” and “-ian” endings in real words in Study 2 3.5 The mean number (out of 8) of correctly spelled “-ion” and “-ian” endings in pseudowords in Study 2 3.6 Items from a task used to make children aware of how places in a sentence frame define grammatical categories 3.7 Examples of items used to teach the category of prefixes that refer to number 3.8 Focusing on verbs 3.9 Examples of items used to practice identification of stems and creation of person words. Playing with pseudowords was fun 3.10 Adjusted means at pretest and for both posttests by group for the correctness of spelling suffixes in Study 3 3.11 Adjusted means at pretest and for both posttests by group for the spelling of suffixes in pseudowords in Study 3 4.1 The adjusted mean scores on the test of spelling suffixes in words (out of a maximum of 26) on each testing occasion for each group 4.2 The adjusted mean scores on the test of spelling polymorphemic words (out of a maximum of 61) on each testing occasion for each group 4.3 The adjusted mean scores on the spelling of suffixes in pseudowords (out of a maximum of 12) on each testing occasion for each group 4.4 The adjusted mean scores on the test of spelling suffixes in words (out of a maximum of 26) on each testing occasion for each intervention group by achievement group in the pretest 5.1 A description and two sample items from the vocabulary test 5.2 Mean scores (adjusted for pretest differences) in the vocabulary test for each testing occasion and group (maximum score = 40)

79 80 84 85

89 91 92







118 125


Illustrations xi


5.4 5.5

6.1 6.2


Mean scores by testing occasion and group (adjusted for pretest differences) in the vocabulary test for children who scored up to the median (left) or above (right) in the pretest Percentage of correct pseudoword definitions (adjusted for pretest differences) by group and testing occasion Percentage correct in the pseudoword-definition test (adjusted for pretest differences) by group and testing occasion One-year teacher follow-up Children’s scores on spelling test: A comparison of morphology, National Literacy Strategy, and standard conditions Writing of a 6-year-old boy who seems to attribute to the digraph “ck” the function of the split digraph “V+C+e”

130 131

132 148

149 171

Tables 2.1 2.2 3.1 4.1 5.1

6.1 6.2 6.3

Number of children in each year group and their mean age Proportion of use of “-ion” and “-ian” spellings for each of the types of word and pseudoword Mean age and standard deviation for the intervention and control groups in Study 1 Mean age in years (and standard deviation) by type of group Number of children, mean age in years (and standard deviation) by year group in school and type of group in the project Number of children in each teaching condition, by year group Children’s average scores before the course, by teaching condition and year group Average percentage increase in the children’s scores by the end of the course, by teaching condition and year group

39 42 68 106

124 146 147


Boxes 1.1 1.2

A crash course in roots and stems (and bases) A crash course in affixes

5 5

xii Illustrations

1.3 1.4 2.1 3.1 3.2 3.3 3.4 3.5 4.1

4.2 4.3 5.1 6.1 6.2 6.3 6.4 6.5 6.6

How psychologists measure morphological awareness A collision course with schwa vowels Children’s spellings of “-ness” and “-ion” by year group in school The word- and pseudoword-spelling tasks used in Studies 1 and 2 The analogy game The correction game The items used for the word- and pseudoword-spelling tests in Study 3 Sample of items from the spelling test showing one child’s answers Examples of suggestions for discussion used to focus on spelling used with the morphemes-plus-spelling group, which were added to the basic activities in the morphemes-only group Examples of the segmentation used in scoring the word- and pseudoword-spelling tests A sample of the same boy’s spelling in the pretest and posttest The instructions and the items in the pseudoworddefinition task Teachers talking about “-ed” endings Lack of awareness of “-ed” rule Teachers thinking about morphemes with connection to meaning Teachers thinking about morphemes without connection to meaning Teachers talking about “-ion” Theories about morphology and spelling

11 17 38 69 71 74 95 97

108 113 114 127 136 137 138 139 141 143

Series editor’s preface

The Improving Learning series showcases findings from projects within the Economic and Social Research Council’s Teaching and Learning Research Programme (TLRP), the UK’s largest ever coordinated educational research initiative. Books in the Improving Learning series are explicitly designed to support “evidence-informed” decisions in educational practice and policymaking. In particular, they combine rigorous social and educational science with high awareness of the significance of the issues being researched. Working closely with practitioners, organizations, and agencies covering all educational sectors, the program has supported many of the UK’s best researchers to work on the direct improvement of policy and practice to support learning. Over sixty projects have been supported, covering many issues across the life course. We are proud to present the results of this work through books in the Improving Learning series. Each book provides a concise, accessible, and definitive overview of innovative findings from a TLRP investment. If more advanced information is required, the books may be used as a gateway to academic journals, monographs, websites, etc. On the other hand, shorter summaries and research briefings on key findings are also available via the program’s website at We hope that you will find the analysis and findings presented in this book are helpful to you in your work on improving outcomes for learners. Andrew Pollard Director, TLRP Institute of Education, University of London


As we wrote this book, we became steadily more aware of the huge effort by very many colleagues—researchers, teachers, and illustrators—and many institutions that made this publication possible. The Economic and Social Research Council (ESRC) was the major supporter of the intervention studies through the Teaching and Learning Research Programme (Grant #L139251015). A previous ESRC grant (#R000237752) and two others by the Medical Research Council (MRC) (G9214719 and G9900004/ID 47376) also gave us essential support for the investigations that made it possible for us to develop the interventions. We are very grateful for the support of these research councils, without which the research reported here would not have been possible. Many teachers and children in different schools participated in the longitudinal phases of this work. In Oxford: Wolvercote First School, Botley Primary School, Cassington Primary School, Kennington Primary School. In London: William Tyndale Primary School, Honeywell Infants’ and Junior School, Ravenstone Primary School, and Trinity St. Mary’s Church of England School. Miriam Bindman and Gill Surman worked in this initial project and were excellent collaborators. The early stages of the development of interventions received the inestimable cooperation of teachers and children in eight schools in London and thirteen schools in the Oxford area. In London: Bessemer Grange Primary School, Dulwich Hamlet Primary School, Hargrave Park Primary School, Brecknock Primary School, Honeywell Primary School, Lauriston Primary School, St. Joseph Roman Catholic Primary School, and St. Michael Church of England Primary School. In Oxfordshire: St. Nicholas Primary School in Abingdon and Wheatley Primary School in Wheatley; and in Oxford: St. Nicholas, Marston, Bayswater Middle School, Larkrise Primary School, Marston Middle School, SS Philip and

xvi Acknowledgements

James Primary School, East Oxford Primary School, Frideswide Middle School, St. Andrews Primary School, Cutteslowe Primary School, New Hinksey Primary School, and Woodfarm Primary School. The Directors of the Hillingdon Cluster of Excellence, Rodney Stafford and Peter Shawley, as well as the teachers and children in the schools that participated in the collaboration with Oxford Brookes University supported the largest part of the intervention studies carried out in the classroom. These were Brookside Primary School, Charville Primary School, Cherry Lane Primary School, Colham Manor Primary School, Grange Park Infant School, Grange Park Junior School, Longmead Primary School, Minet Infant School, John Penrose Primary School, Pinkwell Primary School, Wood End Park Primary School, and Yeading Junior School. We are very grateful to all these teachers and children whose participation made our research possible. Very special thanks are directed to our colleagues and long-time collaborators in Lauriston Primary School, including the Principal, Heather Rockhold, whose rock-solid collaboration for more than ten years has taught us so much. Her team over these years included Hillary Cook and Sue Dobbing, who worked alongside us in each project, Gwenan Thomas, Aidan O’Kelly, Natasha Nevison, Alison Rosica, and Aaron Bertran. We feel privileged to have been able to work with them for so long. They were occasionally, but not always, supported by Department for Education and Skills (DfES) Best Practice grants, which helped them to develop their research skills and to analyze their practice in greater depth. The teachers who attended the morphology course were also truly our partners in this research, testing their children, marking and entering data, teaching the interventions, nagging us about the rigor of our research and the management of the intervention sessions. Participant teachers and schools were: Maggie Bacon, Nick Bonell, Kay Croft, Karen Henry (Kingswood Primary); Louisa Lochner (Gateway); Kathy Thornton (Kingsgate Primary); Stephen Buzzard (New End Primary); Lucinda Midgely (Linton Mead Primary); Rachel Webber (Waterside Social, Emotional and Behavioural Difficulties (SEBD) Primary); Sameena Bashir (Fullwood Primary School); Karen Bloomfield (Coppice Primary School); Caroline Havers (Mayespark Primary School); Carina Mcleod (Cleveland Junior School); Caroline Rogers (Beckford Junior School); Sophia Shaikh, Barbara Turner (Woodlands Junior School); Deborah Walters (Christchurch Primary School); Nathalie Allexant (Gallions Primary School); Bryony Roberts (Edith Neville School).

Acknowledgements xvii

All the teachers who participated in our research were tremendously generous with their time and endlessly patient as we interviewed them, videoed their lessons, asked them questions during their lunchtime, tested their children, organized twilight meetings with them and asked for their feedback and suggestions. Our assessments and interventions included illustrations that led to greater enjoyment by the children. Eldad Druks drew the dinosaurs for our pseudoword testing, and Adelina Gardner did all the illustrations for all the remaining materials. We, and the children, were fortunate to benefit from Addy’s talents and imagination. So many children generously agreed to give their time freely so that other children in the future could benefit from what they helped us to find out. Their participation was essential and made our work in schools great fun. So, THANK YOU EVERYONE!

Part I

What is the issue?

Chapter 1

Morphemes and literacy A starting point

We all know that words have meanings, but not everyone understands that the meaning of any word depends on its underlying structure. Words consist of morphemes, which are units of meaning. These morphemes, in our view, are of immense importance in children’s learning of the meaning of new words and also in their learning how to read and write familiar and novel words. The aim of our book is to show how important morphemes can be in children’s education and how easy it is to enhance their knowledge about morphemes and thus to increase the richness of their vocabulary and the fluency of their reading and writing.

What morphemes are Take a fairly simple word like “unforgettable”. Its meaning is clear and widely understood, but the word has three different parts to it, and it is the combination of these three parts that gives the word its final and overall meaning. The three parts to “unforgettable” are “un-” and “forget” and “-able”. “Forget” is actually a verb, because it refers to an action. Putting “-able” on the end of this verb makes it into an adjective (“forgettable”), which tells us that one can easily forget the person or event that the adjective is describing. The addition of “un-” at the beginning of the adjective gives it the opposite meaning: The new adjective (“unforgettable”) means that it is impossible to forget someone or something. Remove one of these parts, and the word either takes on a different meaning or has no meaning at all. Each of the three parts in Authored by Peter Bryant and Terezinha Nunes

4 What is the issue?

“unforgettable” therefore is a unit of meaning. The technical term for a unit of meaning is a “morpheme”. Some words contain one morpheme only, but many other words in English and in other languages contain more than one. “Forget” is a one-morpheme word, “forgettable” a twomorpheme word and “unforgettable”, as we have seen, contains three morphemes. So, when more than half a century ago thousands of people crooned the popular Nat King Cole song “Unforgettable”, they were repeating a three-morpheme word whose meaning they understood perfectly, though they may not have been completely aware that the word had three separate units to it or that these units were called morphemes. In general, people do have some awareness of morphemes, although, as we shall be showing later on in the book, this awareness tends to be hazy and incomplete. Nevertheless, we can easily work out the meaning of entirely new words if these words are combinations of morphemes whose meaning we already understand. All of us immediately knew what Toni Braxton meant when we heard her desperate, but charming, plea “Unbreak my heart, uncry my tears”. None of us had met the word “uncry” before, but because we knew that adding “un-” to the beginning of a word reverses the meaning of this word (“untie”, “untidy”, “unforgettable”) we could grasp what the singer meant, and, at the same time, we could see that she was asking for a physical impossibility. There are different kinds of morpheme. One distinction of great importance is between roots or stems (see Box 1.1) and affixes (Box 1.2). Every word with more than one morpheme in it contains a root, and this is combined with one or more than one affix morpheme (see Boxes 1.1 and 1.2 for a more detailed description of these morphemes). The word’s meaning starts with its root in the sense that the word would be meaningless without this particular morpheme. “Forget” is the root morpheme in “unforgettable” and “un-” and “-able” are both affixes. Affixes that precede the root are called “prefixes” and those that follow the root are called “suffixes”. These are the only kinds of affix that we have in English, but other languages, such as Swahili, also have “infixes”, which are added-on morphemes that appear in the middle of the root. Another essential distinction is between “derivational” and “inflectional” affixes. Inflectional-affix morphemes, or “inflections” for short, tell us what kind of a word we are dealing with—whether it is a singular (“cat”) or a plural (“cats”) noun, a present (“kiss”) or a past (“kissed”) verb, an adjective (“kind”) or a comparative (“kinder”) or a superlative (“kindest”) adjective. So, the “-s” at the end of “cats”, the “-ed” at the

Morphemes and literacy: A starting point 5

end of “kissed” and the “-er” and “-est” at the end of “kinder” and “kindest” are inflections, and they combine with the root to produce two-morpheme words with a root and an affix.

BOX 1.1 A crash course in roots and stems (and bases) There is a distinction to be made between roots and stems, although from the point of view of this book it is not a particularly important one. The root is the basic part of the word that remains when all derivational and inflectional affixes have been removed. For example, “teach” is the root for the word “teacher” and also for the word “unteachable”. The stem, on the other hand, is the part of the word that remains when all inflectional affixes have been removed. “Teacher” therefore is the stem for “teachers”. Thus, sometimes the root and the stem are the same, but sometimes they are different. “Cat” is both the root and the stem for the plural word “cats”, but “teach” is the root and “teacher” the stem for the plural word “teachers”. In all the examples and the tasks that we shall describe in this book, the roots and the stems are always identical, which is why the distinction is not an important one as far as this book is concerned. The base or base word is another related term and it is relevant to our book. This refers to the word from which a complex word is derived (for example, “touchable” is the base for “untouchable”). Thus in the word “unbearable”, “bear” is the root, “bearable” is the base, and “un-” is the derivational prefix.

BOX 1.2 A crash course in affixes In English, affixes are morphemes that are attached to the stem or the root of a word (see Box 1.1 for the distinction between stems and roots). These affixes either come before the root or follow it. Those

6 What is the issue?

that come before the root are called prefixes and those that follow it are suffixes. There are two types of affix: Inflectional and derivational affixes. Inflectional affixes, or inflections, give you essential information about the word. For instance, all nouns are either singular or plural, and in English the presence of an /s/ or a /z/ sound at the end of a noun usually means that the word is in the plural, whereas its absence usually signals that it is a singular noun. This end sound is the plural inflection. When you hear the word “cats” or the word “dogs” the inflection at the end of each word tells you that it refers to more than one animal. Similarly, the absence of the “s” at the end of an English noun means, in most cases, that the noun is a singular one. There are inflections in English for nouns (the plural “-s” and the possessive “-’s”), adjectives (the comparative “-er” and the superlative “-est”), and for verbs (the past tense “-ed”, the thirdperson singular in the present tense (“-s”) and the continuous tense (“-ing”). All inflections in English are suffixes. Many other languages, such as French and Greek, are much more inflected than English. In these other two languages, for example, there are plural inflections for adjectives as well as for nouns. Some languages also mark gender in adjectives as well as nouns with inflections. Derivational affixes are different. Adding a derivational affix to a word creates a different word, which is based on the original word but not the same. Sometimes the difference between the base word and the derived word is that they belong to different grammatical classes: For example, the derivational suffix “-ness” changes adjectives into abstract nouns ( for example “happy”–“happiness”) and the suffix “-ion” changes verbs, again, into abstract nouns (for example, “educate”–“education”). The suffix “-ful” changes nouns into adjectives (for example, “help”–”helpful”, “hope”–”hopeful”). Other derivations such as “un-” and “re-” bring about a radical change in the meaning of the base words to which they are attached (for example, “un-helpful”, “re-born”) but do not affect their grammatical class. Some derivational affixes are prefixes and others suffixes. Derived words include the base word from which they are derived but in many cases the pronunciation of the base word changes in the derivation, as in “fifth”, which is derived from “five”, and “electricity” which is derived from “electric”.

Morphemes and literacy: A starting point 7

Derivational morphemes create new words based on old ones. “Un-”, which, as we have seen, reverses the meaning normally given to the root that it precedes, is a good example of a derivational morpheme. So is the suffix “-able”, which we have met once already at the end of “unforgettable” and which appears at the end of many other English words, such as “unbearable”. Consider the relatively new coinage of the word “doable” (do-able), which we ourselves have heard our students and our builder use: “It’s doable”, they say, and we instantly understand what they mean, even though they often turn out to be wrong. This suffix is a derivational morpheme because it changes the word from the verb, represented by the base word, to an adjective, which says that the action referred to by the verb is entirely possible. By now you should know, if you did not know before, how many morphemes there are in “education” or in “uneducated” (there are two in the first word and three in the second). You should be able to work out whether the affix at the beginning of “incompetent” and the affix at the end of “kisses” are derivational or inflectional (derivational in “incompetent” and inflectional in “kisses”). You should also have noted that there is a strong connection between morphemes and grammar: You can use the “-ed” at the end of verbs, in order to convey the meaning of past tense, but you cannot use the “-ed” ending with nouns; nouns don’t have a past tense. Once you are completely clear about roots and affixes, prefixes and suffixes, and derivations and inflections, we know that you will want us to justify our claim that these morphemes play a crucial but neglected role in children’s development and in their education. Before we move on to the next section where we will begin to make this claim in earnest, we should like you to ponder why P. G. Wodehouse’s joke about the word “disgruntled” is so very funny. He wrote of a man who was consumed with anger: “If not actually disgruntled, he certainly wasn’t gruntled”. This understatement is amusing because although he followed strict morphemic principles, Wodehouse managed to create a word that we never use. The morpheme “dis-”, like the morpheme “un-”, reverses the meaning of the base that it is attached to, and so “gruntled” should be the opposite of “disgruntled”, but this is an unused word. We know both these things, and it is the tension between them that makes us laugh. Wodehouse’s joke helps us make another point about morphemes, which is that morphemes and grammar, these inseparable friends, form a basis on which we build the learning of new words. Philosophers, linguists, and psychologists have pondered at the marvel that it is to

8 What is the issue?

learn a word with all that this learning implies. If a mother points to a dog sniffing the lamp post, and says to her baby “Look at the dog”, how is the child to know that the mother means by “dog” the animal and not the action that the dog is performing? The U.S. child psychologist Roger Brown suggested that children use grammatical information contained in the sentences in forming an idea about what a new word means (Brown 1957). In the sentence “Look at the dog”, the article “the” gives a clue that the word is a noun, not a verb, and this helps them come up with the dog, rather than the action, as the meaning for the word “dog”. Brown’s studies actually required much more from the children than the distinction between nouns and verbs. He created a technique, which we will use often in our research, of observing how children learn a made-up word. The reason for studying how children learn made-up words, which are called “pseudowords” or “nonsense words” by researchers, is that because the word is made up by the researcher, we can be certain that the child has not come across it before—just like “gruntled” in Wodehouse’s joke. To clarify how the technique works, consider one of the examples used by Brown in his research. He showed children in the age range 3–5 a picture of a pair of hands kneading a strange substance in a strange container. To some children he said “In this picture you can see some sibbing”; to other children he said “In this picture you can see some sib”; to a third group of children he said: “In this picture you can see a sib”. The children who were told “some sibbing” should conclude that “sibbing” refers to the action; those who were told “some sib” should conclude that “sib” refers to the substance; those who were told “a sib” should conclude that “sib” refers to the container. Each of the children was then shown three pictures, one that depicted the same action on a different substance and with a different container, one depicting the same substance but a different action and container, and one depicting the same container but a different substance and action. The children were able to choose the correct picture more often than one would expect if they were just guessing. With three pictures to choose from, if they were just guessing they could be right one-third of the time, but they were right more than two-thirds of the time for any of these different presentations. Later work by many other researchers interested in children’s learning of vocabulary (see further readings by Gleitman and colleagues: Gleitman 1990, Gleitman and Gleitman 1992) confirmed that children do use their implicit knowledge of grammar in learning vocabulary. They referred to this idea as the “Syntactic Bootstrapping Hypothesis”,

Morphemes and literacy: A starting point 9

to indicate that children use grammar to help narrow down the meaning of words. Because of the strong connection between grammar and morphemes, and because morphemes are units of meaning, it is reasonable to expect that morphemes help us to learn new words even when these are not in the context of sentences, and thus we cannot use grammar to bootstrap word learning. There is some, albeit limited, evidence for this. It is known that older children, in the age range 11–13, learn the definitions of pairs of pseudowords better if the pseudowords share a stem (for example “flur” and “flurment”) than if they do not. This indicates that they can use what they learned about one stem when learning the derived pseudowords—that is, they can use morphemic bootstrapping, not only syntactic bootstrapping. In summary, • •

Words are formed with units of meaning, termed morphemes. Morphemes and grammar are strongly connected because inflectional morphemes can only be applied to particular grammatical categories and derivational morphemes are used to form words of particular grammatical categories. Research shows that morphemes are not just ways in which linguists analyze words: People use knowledge of morphemes and grammar to learn the meanings of new words.

Why are morphemes important in education? For our answer to this question we turn to children’s explicit knowledge or awareness of the language that they speak and to which they listen. We shall argue that schoolchildren need to become explicitly aware of principles of language, which at earlier ages they learned and obeyed at an implicit level only. Once at school they need to develop explicit knowledge of language, in general, and of morphemes, in particular, which they can think about and can even talk about much more openly and explicitly than they had before. We shall be arguing that schoolchildren need this new explicit knowledge about morphemes for two main reasons. One is that it is essential in learning to read and to spell. The other is that morphemic knowledge plays a central role in the growth of schoolchildren’s vocabulary, because large numbers of the words that they have to learn at school are derived (with the help of derivational morphemes) from other words.

10 What is the issue?

The main purpose of the rest of this book will be to provide evidence, mostly from our own research, that these propositions are right, but before we do that we shall say more about what made us think that they might be right in the first place. We need to tell you first: •

• •

why we concluded that children’s knowledge of morphemes is at first implicit and that there might be ways of increasing the level of children’s explicit knowledge about these units; why explicit knowledge of morphemes may be an essential ingredient of learning to read and to write; why children also need explicit knowledge about morphemes to keep to a respectable level of vocabulary growth while they are at school.

Implicit and explicit knowledge of morphemes Young children begin to understand and to use morphemes from an early age. English-speaking children usually begin to produce twomorpheme words in their third year and during that year the growth in their use of affixes is rapid and extremely impressive. This is the time, as Roger Brown showed, when children begin to use suffixes for possessive words (“Adam’s ball”), for the plural (“dogs”), for present progressive verbs (“I walking”), for third-person singular present tense verbs (“he walks”), and for past tense verbs, although not always with complete correctness (“I brunged it here”) (Brown 1973). Notice that these new morphemes are all of them inflections. Children tend to learn derivational morphemes a little later and to continue to learn about them right through childhood, as we shall show in later chapters. Nevertheless, from their third year on, with little or often no explicit help from other people, they master the system of roots, prefixes, and suffixes with ease. By the time that they go to school they are morphemic experts. They are, to derive a new word, morphemists. They are experts, however, only at an implicit level. They are soon at a loss when given quite simple tasks that need some explicit judgment about morphemes. These tasks do not require children to know anything about the terms that we set out in the previous section (morphemes, roots, prefixes, suffixes, etc.), but they do require the children to reflect about some fairly basic morphemic similarities between words, and young children find them very hard indeed. One such task is a simple analogy task that we devised ourselves and gave to a large group of children in the 6–9 age range (Nunes et al.

Morphemes and literacy: A starting point 11

1997b). Our aim in this task was to find out how well individual children can transform a present-tense verb into a past-tense one, or vice versa. So, we said a sentence like “The dog is scratching the chair” to a puppet, who “repeated” it, transforming the verb into a past-tense one: “The dog scratched the chair”. Immediately after that, we said another very similar sentence, also with a present-tense verb—“The dog is chasing the cat”—and we asked the children to say it back to us, but like the puppet would say it. We wanted to see if the child, as the puppet had done with the previous sentence, could change the verb by removing the presentcontinuous-tense inflection “-ing” and adding instead the past-tense inflection “-ed” to make the new sentence, “The dog chased the cat”. The task was straightforward and contained no technicalities. We did not ask the children to write anything and so they had no reason to worry about spelling. All that they had to do was to remove the present-continuous inflection and add the past-tense inflection instead. However, this apparently simple task was quite difficult even for many of the oldest children in the group. We found that 6-year-old children, all of whom could spontaneously produce present- continuous- and past-tense verbs in the right places in their own speech, only managed to get 31 percent of the items right in this morphological test. For the 7-year-old group this figure rose to 41 percent and for the 8-year-olds to 56 percent. So, children get better at this task as they grow older, but even the oldest make many mistakes. Many other “morphological awareness” tasks that were invented by our team and still others that were devised by other research teams have produced the same results. Most young schoolchildren fluently speak and effortlessly understand words that are quite complicated from a morphemic point of view. Yet, they are usually completely, albeit quite cheerfully, at sea when asked to make simple comparisons of the morphemes in different words. Box 1.3 presents a sample of different tasks used to assess children’s awareness of morphology.

BOX 1.3 How psychologists measure morphological awareness The aim of all morphological awareness tasks is to measure children’s or adults’ conscious knowledge of the morphemic structure of spoken words. There is a wide variety of such tasks.

12 What is the issue?

Productive morphology (Nunes et al. 1997a, adapted from Berko 1958) The tester says two sentences, which contain an entirely unfamiliar pseudoword and then invites the child to complete a sentence using that pseudoword with the target inflection. Each item is presented along with a picture. The picture used for the first item is included here to illustrate the method.

1. This is a man who knows how to snig; he is snigging onto his chair. He did the same thing yesterday. What did he do yesterday? Yesterday he? 2. This is a person who know how to mab along the street. Yesterday he mabbed along the street. Every day he does the same thing. What does he do every day? Every day he? 3. This person is always tigging his head. Today, as he falls to the ground, he tigs his head. Yesterday he did the same thing. What did he do yesterday? Yesterday he? 4. “Be careful,” said the farmer. “You’re always clomming on your shoelace. You’re about to clom on it now.” Yesterday you? 5. Ever since he learned how to do it this man has been seeping his iron bar into a knot. Yesterday he sept it into a knot. Today he will do the same thing. What will he do today? Today he will?

Morphemes and literacy: A starting point 13

6. This is a zug. Now there is another one. There are two of them. There are two? 7. This is a nuz. Now there is another one. There are two of them. There are two? 8. It was a bazing day. He felt very bazed. He stuck out his hands and shouted with? 9. It was night-time and the moon was shining. He danced luggily and smiled with lugginess. He felt very? 10. When the sun shines he feels very chowy. He dances chowily and laughs with?

The sentence analogy task (Nunes et al. 1997a) The tester uses puppets to present the sentences. The first puppet “says” the first sentence in the pair; the second puppet “says” the second sentence. Then the first puppet says the first sentence in the second pair and the child is encouraged to help the second puppet and say its sentence. Each item presents the corresponding pairs. 1. Tom helps Mary : Tom helped Mary :: Tom sees Mary : ________ 2. Jane threw the ball : Jane throws the ball :: Jane kicked the ball : ________ 3. The cow woke up : The cow wakes up :: The cow ran away : ________ 4. The dog is scratching the chair : The dog scratched the chair :: The dog is chasing the cat : ________ 5. I felt happy : I feel happy :: I was ill : ________ 6. Bob is turning the TV on : Bob turned the TV on :: Bob is plugging the kettle in: ________ 7. She kept her toys in a box : She keeps her toys in a box :: She hung her washing on a line : ________ 8. Bob gives the ball to Ann : Bob gave the ball to Ann :: Bob sings a song to Ann: ________

The word analogy task (Nunes et al. 1997a) The tester uses puppets to present the words. The first puppet “says” the first word in the pair; the second puppet “says” the second word.

14 What is the issue?

Then the first puppet says the first word in the second pair and the child is encouraged to help the second puppet and say its word. Each item presents the corresponding pairs. 1. 2. 3. 4. 5. 6. 7. 8.

anger : angry :: strength : ________ teacher : taught :: writer : ________ walk : walked :: shake ________ see : saw :: dance ________ cried : cry :: drew ________ work : worker :: write : ________ sing : song :: live ________ happy: happiness :: high : ________

Test of morphological production (Fowler and Liberman 1995, adapted from Carlisle 1988) The children are either presented with the base form and have to use the derived form in a sentence (for example, “Four. The big racehorse came in ________”) or they are given the derived form and have to produce the base form (for example, “Fourth. When he counted the puppies, there were ________”). The word pairs either fit into the phonologically neutral or phonologically complex condition. The same suffix was used for each pair to make the conditions more comparable. Phonologically neutral danger : dangerous shine : shiny four : fourth agree : agreeable examine : examination suggest : suggestion

Phonologically complex courage : courageous anger : angry five : fifth respond : responsible combine : combination decide : decision

There is one apparent exception to this run of rather negative results, and it is an instructive one. In 1958 Jean Berko, a U.S. child psychologist, did a classic experiment in which she used pseudowords (as did Roger Brown) like “wug” to avoid testing children’s specific knowledge and thus to arrive at some conclusion about their knowledge of morphemic principles (Berko 1958). In her best-known question,

Morphemes and literacy: A starting point 15

she showed 4- to 7-year-old children a picture of an unfamiliar creature, told them it was a “wug” and then showed them a picture of two of the same creatures and asked them to describe it. Most children, down to the age of 4, produced the correct answer of “wugs”, and from this Berko concluded that they knew about the nature of the plural “-s” inflection. The children were as successful in their answers to some other questions that involved past verbs. For example, referring to a man exercising, Jean Berko told the children “This is a man who knows how to gling. He is glinging. He did the same thing yesterday. What did he do yesterday? Yesterday, he,” . . . and then she asked the children to complete the sentence. Even children who were still too young to go to school came up, for the most part, with the appropriate “glinged” in answer to her question. So, Berko argued, even pre-schoolchildren are explicitly aware of the past-tense inflectional morphemes as well as of the plural ending. Yet, some other results from the same Berko study gave the lie to this optimistic claim. When the children were asked to make singular nonsense words with /s/ or /z/ sound endings into plural words (for example “niz”–“nizzes”), and present-tense nonsense verbs ending in /t/ or /d/ sounds into past verbs (for example, “mot”–“motted”), they nearly always failed to do so. It was as if the children had some vague idea that plural words end in /s/ or /z/ sounds and that past verbs end in /t/ or /d/ sounds, and yet do not understand that plural words consist of two parts: the root or stem (see Box 1.1), which is the same as the singular word, and the added /s/ (“cats”) or /z/ (“dogs”) or /iz/ (“kisses”) sound, which signals that the word is in the plural. In the same study, Berko also tried to get the children to use derivational morphemes on nonsense words, but she found that the children were strikingly unsuccessful. For example, she asked the children, and some adults too, what would they call a man whose job is to “zib”. All the adults formed a new noun by adding a derivational suffix “-er” to form a new word, “zibber”, but only 11 percent of the children were able to come up with this word. They simply found it too difficult to consciously derive an agent from a verb. The answer to our first question about young schoolchildren’s explicit (as opposed to their implicit) morphemic knowledge is therefore mostly negative. When children arrive at school and during their first few years there, they have some awareness of the morphemic system, which they themselves use in their own conversations with extraordinary proficiency. But this awareness is only a weak one. Morphemes are an essential part of the young children’s everyday life, but these

16 What is the issue?

youngsters are barely conscious of them or of their importance. What implications does this have for the learning that they have to do at school?

Explicit knowledge of morphemes may be an essential ingredient of learning to read and to write It is an important, though shockingly neglected, fact that one of the best ways to help children to become experts in reading and spelling is to make sure that they are thoroughly familiar with the morphemic system in their own language. This kind of knowledge may not be an absolute requirement for learning how to read and write English, Portuguese, Greek, French, German, Arabic, and Hebrew, but it certainly will make this learning an easier and a more successful task. The main reason why morphemic structure is so important for reading and writing in these and in many other languages is that morphemes affect the ways in which words are spelled. If you want to know what many written words are, particularly new words, and if you want to know how to write words, and, again, new words in particular, you really have to be able to work out their morphemic structure. Morphemes have such a powerful effect on spelling for three good reasons, which we shall look at in turn. 1. 2.


The same sounds are spelled in different ways in different morphemes. It is often the case that a particular morpheme is spelled in the same way, even though it is represented by different sounds in different words. Some morphemes are represented in writing but not in speech.

The same sounds are spelled in different ways in different morphemes The first of these points needs particular attention from those who are tempted to think that the be-all and end-all of teaching children to read is to encourage them to learn about the relationship between sounds and letters or sequences of letters. This doctrine is no help at all to a child who wants to know why the ending of “locks” and “fox” sound exactly the same and yet are spelled quite differently from each other. The reason for the difference is a morphemic one. The first word has two

Morphemes and literacy: A starting point 17

morphemes and its final /s/ sound is the plural inflection, which is represented by “-s” in regular plural words in English. The second word, “fox”, on the other hand, is singular and therefore there is no reason to break up the /ks/ sound at the end. In such words, this ending is represented by an “-x” as in “fox” or “-xe” as in “axe”. For precisely the same reason, the final /z/ sound in “trees” and “freeze” is spelled quite differently in the two words. “Trees” is a two-morpheme plural word and so its /z/ ending is represented by “-s”, the conventional spelling for the plural inflection. “Freeze” is a one-morpheme word and the /z/ sound ending is spelled as “-ze”. This actually is a clear and inflexible principle in English spelling: Take every word that ends with a /z/ sound and you will find that this ending is always spelled as “-s” in plural words and always as “-zz” (“jazz”) or “-ze” (“froze”) or “-se” (“rose”) in one-morpheme words (Kemp and Bryant 2003). So far, we have contrasted two- with one-morpheme words, but it is easy to show rather similar effects of morphemes on spelling in contrasts between different two-morpheme words. The point here is that there are affix morphemes that have quite different functions and yet sound exactly the same. Sometimes these different affixes are spelled in the same way, like the “-er” ending. When “-er” represents an affix, it is a comparative (“bigger”, “braver”, “cleverer”) in some words and an agentive in others (“baker”, “sweeper”, “cleaner”). No problem there, but what about the “-ion” and “-ian” endings in “education” and “magician”? These endings sound exactly the same (if you don’t believe this, say both words out aloud and listen carefully), but they are spelled quite differently. Both ending syllables contain a schwa vowel followed by the /n/ sound. (See Box 1.4 for an explanation of schwa vowels and for an object lesson in why children need to know about morphemes when they are learning to spell.)

BOX 1.4 A collision course with schwa vowels Educationalists and psychologists are fond of the term “spelling demons,” a term that they use to describe words whose spelling flouts conventional spelling rules. In our view, however, the worst demon in English spelling is not a word, but a particular sound. This

18 What is the issue?

is the schwa vowel sound, which, to take one pair as an example, is the last sound both in “Bognor” and in “picture”. The schwa vowel is easily the most frequently used vowel sound in the English language, and yet there is no set way of spelling it on the basis of letter–sound rules. It is also known as a “weak” vowel sound—one that is rather poorly articulated. Schwa vowels crop up in profusion in words of more than one syllable, and they are always in the unstressed part of the word. Here are some examples of words with one or more schwa vowels: “happiness” (schwa vowel in the last syllable) “election” (schwa vowel in last syllable) “magician” (schwa vowel in first and last syllables) “hasten” (schwa vowel in last syllable) “glorious” (schwa vowel in last syllable) “attraction” (schwa vowel in first and last syllable) “psychology” (schwa vowel in third syllable) “bigger” (schwa vowel in last syllable) “painter” (schwa vowel in last syllable) “embarrassment” (schwa vowel in third and fourth syllables) “incredible” (schwa vowel in last syllable) “unforgettable” (schwa vowel in the last two syllables) “rehearsal” (schwa vowel in last syllable) “banana” (schwa vowel in first and last syllables) “onion” (schwa vowel in last syllable) “tomato” (schwa vowel in first syllable) “Stilton” (schwa vowel in last syllable) “exaggerate” (schwa vowel in third syllable) “photography” (schwa vowel in the first and third syllables). It may come as something of a surprise that the twenty-four vowels that we have pinpointed in this list are all the same vowel sound, since the sound is spelled in so many different ways in the different words. But with a moment’s reflection, and perhaps with the help of pronouncing the words out loud, you will see that they are all one and the same sound. The variety of ways in which this sound is spelled in English is truly astonishing. In this small list we counted six different spellings for the schwa vowel (“a”, “e”,

Morphemes and literacy: A starting point 19

“o”, “ia”, “ia” and “ou”), and there are many other spellings for this promiscuous vowel. This range of spellings for the same sound is as good an illustration as one can possibly find for the inadequacy of treating or teaching English spelling as a system of rules about letter–sound relationships with a few exceptions. With the schwa vowel there is no letter–sound rule. At the level of phonology, every spelling is an exception. On another level, however, which is the level of morphemes, there is a set of principles that can guide the spelling of this sound in very many words. These are morphemic spelling principles. Look again at the list and you will see that the first thirteen words (“happiness” to “rehearsal”) are all two- or three-morpheme words with a stem or base followed by a suffix, and with a schwa vowel in each suffix. The spelling of each of these thirteen endings is highly consistent across different words with the same suffix (“glorious” and “furious”, “magician” and “logician”, “happiness” and “sadness”) and when two different suffixes sound exactly the same they are sometimes spelled differently (“attraction” and “mathematician”). Thus, the spelling of the schwa vowel is often determined by the meaning, rather than the sound, of the word. Meaning, and the morphemes which convey that meaning, can often tame this particular spelling demon.

There must be some reason for this difference in the spelling of these two affixes, and the chances are that it is a morphemic one, since these written endings usually do represent morphemes. Yet none of the tomes on English spelling, no educational textbook, nor any one of the many accounts of the psychology of reading and spelling provide any kind of a clue to the reason for the two different spellings for this ending, even though the schwa vowel followed by an /n/ is a very common ending, which is notoriously hard for children to spell. In fact, there is a clear and rather simple principle for spelling this ending with nouns. If the noun refers to a person or an animal, its ending is spelled as “-ian” (“magician”, “mathematician”). If it does not refer to a person, it is spelled as “-ion” (“education”, “institution”). There are hardly any exceptions to this principle, and these few exceptions are all words that are quite uncommon ones (“radian”, “centurion”). This is a distinction that should cause no particular difficulty to 7- and 8-year-old children. Teachers, therefore, should be able to put it across

20 What is the issue?

to their pupils quite easily. Yet, as far as we know, no one teaches our principle about “-ion” and “-ian” endings in schools in England, and, as the next chapter will show, the pupils continue to make frequent and rather serious mistakes when writing words that ought to have one or the other of these two endings. The “-ion”/“-ian” issue is something of a test case for us. We are interested in morphemic spelling principles and, particularly, principles that could be, but are not, taught at school. We are also interested in spelling patterns that cause children great, and possibly quite unnecessary, difficulties. The “-ion”/“-ian” endings fit both these requirements and raise two clear and pressing questions: 1. 2.

Can schoolchildren be taught this morphemic spelling principle? Will this teaching help them to spell these difficult words?

We shall present our answers to these questions in the chapters that follow. It is worth mentioning at this point that the same questions can be asked about other languages. In Portuguese there are several instances of the same sound being spelled in different ways in different morphemes. For example, the endings of the words “princesa” (“princess”) and “pobreza” (“poverty”) sound exactly the same (they both rhyme with the English word “blazer”), but they are spelled differently because the “-esa” ending is the right one for the derivational morpheme that represents a female, while the “-eza” ending is the conventional spelling in abstract nouns that end with that suffix. In modern Greek, which is known as a highly regular script, children still have to learn to pay particular attention to morphemes (Aidinis and Nunes 2001, Bryant et al. 2000). They need to do so because the Greek language has few vowel sounds and many ways to spell them. For example, there are many different ways to spell the “ee” vowel sound, as in “feet”, in Greek words, which becomes a problem for Greek children because they have to learn which spelling to choose for this sound in different words. The best help that Greek children get in making this choice is from morphemes (Bryant et al. 1999, Chliounaki and Bryant 2003). Greek root morphemes are always spelled in the same way, of course, and so whole families of words always use the same spelling for the vowel or vowels in the root that they have in common. Also, different Greek inflections are spelled differently even when they sound the same (as with “-ian” and “-ion” in English). Four different inflections are signalled by the “ee” sound at the end of Greek

Morphemes and literacy: A starting point 21

words and there is a different way of spelling each of them. The spelling for feminine singular endings on nouns and adjectives is η, for masculine plural endings is οι, for neuter singular endings is ι and for third-personsingular present-tense verb endings is ει.

It is often the case that a particular morpheme is spelled in the same way, even though it is represented by different sounds in different words The central point of this chapter is that in English and in many other languages there is a system of relationships between morphemes and spelling and that it will help children immensely to know what this system is. One of the most compelling reasons why schoolchildren need to know about this system is that in many cases there is a constant spelling for a particular morpheme, even though the sound of that morpheme differs from word to word. For instance, we take medicines to heal ourselves, and we worry about our health. These two words share the same root morpheme, and the spelling for this morpheme is the same in both words even though the vowel sound is long in the one-morpheme word, “heal”, and short in the two-morpheme word, “health”. “Muscle” and “muscular” form a similar pair: the “sc” sequence represents one sound, /s/, in the first word but two, /sk/, in the second. The reason for this apparent inconsistency in letter–sound correspondences is that the two words share the same root. The relationship between letters and sounds is inconsistent, but the relationship between letters and morphemes is entirely consistent. Once again, if we are going to teach children the principles of English spelling, we shall have to tell them about morphemes too. In affix morphemes as well we can find consistent connections between spelling sequences and morphemes, despite inconsistent connections between these same spelling sequences and sounds. The past-tense ending in verbs is the most powerful example, and an interesting one from our point of view, because it is one of the few connections between morphemes and spelling that teachers tell their pupils about at school. In regular past verbs there are three different pronunciations for the past-tense ending /t/ as in “kissed”, /d/ as in “killed”, and /id/ as in “waited”. Yet we spell all three endings as “-ed”, despite the notable differences in the ways that we pronounce them. In the next chapter, we shall see that children take a long time to get to grips with this particular spelling principle, despite being taught about it in the classroom. One problem for them is that they have to

22 What is the issue?

distinguish not just between past verbs and similar-sounding words that are not verbs (“peeled” versus “field”; “kissed” versus “list”) but also between regular and irregular past verbs (“tipped” versus “slept”; “frowned” versus “found”).

Some morphemes are represented in writing but not in speech Morphemes are important in reading and writing for a third reason, which is that some morphemic distinctions are explicit and clearly signaled in writing but not in speech. In some ways, this point is at least as important for children’s acquisition of spoken language as it is for their learning about written language, because it is entirely possible that they may eventually become aware about these particular morphemic distinctions in speech through seeing them in print. The apostrophe, which is notorious for the difficulties that it causes adults and children alike, is a case in point. In the English script it represents either an elision (“can’t” for “cannot”; “it’s” for “it is”) or the possessive (“the boy’s cousin”; “the girls’ teacher”). The possessive ending is an affix morpheme, and so we will concentrate on that for the moment. We mention the possessive apostrophe at this point because it makes an explicit morphemic distinction that spoken language fails to do (Bryant et al. 2000). The two phrases “the boys drink” and “the boy’s drink” have entirely different meanings in their written form. In one phrase, “boys” is a plural noun and “drink” refers to what they are doing. The other phrase is about a boy in the singular and “drink” is a noun. Both these fundamental differences are signaled simply by the absence of an apostrophe in one passage and its presence in the other. In spoken language it would be quite a different matter: Although the two passages have quite different meanings, they sound exactly the same. Of course, listeners who hear one of the passages would soon be able to infer what the person speaking to them had meant by it, but they would have to use the context to do that. The spoken words on their own are ambiguous; the written words are not. Precisely because writing represents a distinction here which spoken language does not, we should expect it to be quite hard for people to learn about the possessive apostrophe. In fact, many people, and not just greengrocers, have real difficulties with this morphemic spelling: Adults’, as well as children’s, knowledge of when and when not to use the apostrophe is often distinctly sketchy.

Morphemes and literacy: A starting point 23

Lest you should think that we are dealing with a peculiarity of the English language here, we shall show you now that the French written language also signals morphemic distinctions that are completely hidden in spoken French. Plural endings in French nouns, adjectives, and verbs are for the most part silent. The word for “house” sounds exactly the same in the plural as it does in the singular (“la maison”, “les maisons”), even though the two forms have different spellings, and this is true of most other nouns as well. It is the same with verbs: Thirdperson singular, and plural verbs have exactly the same sound /e/ in the present tense (“il aime”, “ils aiment”) but are spelled differently. Thus, the plural affix appears in writing as “-s” at the end of plural nouns and adjectives and as “-nt” at the end of plural verbs, but not in speech. These “silent” plurals cause French schoolchildren a lot of difficulty when they first learn to write. In an intriguing series of studies, Michel Fayol, a French psychologist, and his colleagues have clearly shown a sequence in the way that children learn to spell plural nouns, adjectives, and verbs (Fayol, Hupet, and Largy 1999, Fayol, Thenevin, Jarousse, and Totereau 1999). There are four steps in this sequence. 1.




At first, young French-speaking schoolchildren simply leave the plural ending out: They write the words as they sound, and, since the plural endings have no sound, they do not represent them in their spelling at all. So, they usually make the mistake of writing “les arbres”, for example, as “les arbre”. Later on, they do learn about the plural “-s” ending, but they tend to use it altogether too frequently, since they often put it at the end of verbs as well as at the end of plural nouns and adjectives. They write “ils aimes” instead of the correct “ils aiment”. Next, they learn about the “-nt” ending as well, but again their use of this ending is often indiscriminate. Sometimes, they put the “-nt” ending on some nouns and adjectives too. For instance, they sometimes write “les maisonent” instead of the correct “les maisons”. Finally, and usually with the help of a great deal of instruction in the classroom, they manage to make the distinction between plural noun and verb endings.

We cannot be surprised by the problems that French-speaking children have with learning how to represent in writing a morpheme that they do not hear in speech. But their pain may be worthwhile, for it is quite likely that French-speaking children learn a lot about these silent

24 What is the issue?

morphemes from seeing them so explicitly there in print. Through seeing and writing these plural endings, they should become much more aware than they were before of singular and plural distinctions in spoken language as well.

Cause and effect in the connections between children’s knowledge of morphemes and their learning to spell morphemes Our last point about the effect on French-speaking children of finding out about morphemes through learning how to spell raises a general question about the direction of cause and effect. In most of this chapter we have been emphasizing possible causes and effects in one direction only. We have argued that children’s knowledge about morphemes must be a powerful and necessary resource in learning to read and write. Morphological knowledge, according to this view, should have a strong effect on children’s reading and spelling. Now we should also consider the possibility that cause and effect might take the opposite direction as well. Learning to read and write might alert children to morphemic distinctions that had escaped them before: Their experiences with written language also cause a change in their explicit awareness of morphemes. The idea of a two-way street—from reading and writing to morphemic knowledge as well as from morphemic knowledge to reading and writing—is at its most plausible when morphemic distinctions are explicit in writing but hidden in speech, as happens with the possessive apostrophe in English and with plural endings in French. But it might also be true of morphemic distinctions that are explicit both in written and in spoken language. Here, too, children’s experiences with written language might alert them to the structure of morphemes in spoken language. Before we find out whether the street is a two-way one, we first have to establish whether the street exists at all. Is there evidence for a connection between children’s morphological knowledge and the progress that they make in learning to read? Fortunately, such evidence does exist, and in such abundance that we can only review some of it here. In the U.S.A., for example, Joanna Carlisle gave children a morphological production task (see Box 1.3 for a description of this task) and related it to a measure of their reading comprehension (Carlisle 1995). She found that the first-grade children’s scores in this

Morphemes and literacy: A starting point 25

morphological task were quite strongly related to the level of their reading comprehension a year later when they were in the second grade. Anne Fowler and Isabelle Liberman, two U.S. psychologists, also found an impressive relationship between children’s success in a morphological production task and their reading levels (Fowler and Liberman 1995). From Denmark, Carsten Elbro reported a high correlation between the number of mistakes that children made with inflections in a Danish version of Jean Berko’s task and the mistakes that they made in reading inflected words in a written text (Elbro 1989). In France, Séverine Casalis and Marie-France Louis-Alexandre also found that kindergarten children’s scores in a variety of morphological production tasks predicted their progress in reading two years later at school (Casalis and Louis-Alexandre 2000). One interesting aspect of this last study was that the morphology tasks dealt both with inflectional and derivational morphemes and the scores with the inflectional items did a much better job of predicting reading than the scores with the derivational problems. This pattern of relations would almost certainly be different in the case of older children. The inflectional system is far less varied than the derivational system, and young children are more likely to understand and use their knowledge of inflections when reading than their knowledge of derivations. We could go on, but we think that we have said enough to make the point that a relationship between morphological knowledge and literacy does exist. Now we can consider the question of the direction of cause and effect in this relationship. Some of the evidence on this question comes from a study that we ourselves carried out several years ago (Nunes et al. 1997a, 1997b). This was a longitudinal study in which we looked at the same children’s spellings over a 3-year period. We were interested in children’s spelling of inflections, in particular of the past tense “-ed” inflection, and we studied how their spelling of this morpheme changed as they grew older and how these changes were related to their knowledge of morphemes. Let us begin with the first question: How does children’s spelling of the past-tense morpheme change over time? The children’s ages at the beginning of the study ranged from 6 to 9 years. So, by the end of the project, the youngest children were 9 years old and the oldest were 12 years old. We found that during this period their spelling of the pasttense inflections changed radically. The very youngest children’s spelling of past verbs, as of other words, was often quite unsystematic. However, we found that as soon as their spelling of past-tense endings became consistent, it invariably followed the same pattern. These children began

26 What is the issue?

Figure 1.1 The first two pages of a 71⁄2-year-old girl’s story.

by spelling the ending phonetically and, therefore, incorrectly. Figure 1.1 illustrates this phonetic spelling, using a child’s free writing of a story. “Pikt” for “picked”, “opund” for “opened” and “suckt” for “sucked” are mistakes familiar to the point of banality to anyone who works with young schoolchildren, but they are no less important for that. These mistakes clearly show children obeying one kind of spelling principle and ignoring another. Their use of phonologically based spelling principles is ingenious but too pervasive. Their complete disregard for morphemically based spelling principles is obvious. The “-ed” spelling transgresses phonological correspondences, and so they ignore it. Later on this changes. It is hard to assign a particular age to this change, for it varies so much between children, but usually, according

Morphemes and literacy: A starting point 27

to our results, children begin to put the “-ed” ending on past verbs some time between the ages of 7 and 8. At first they do so with some regular past verbs and not with others. However, the really interesting thing about their initial sporadic use of the “-ed” ending is that they often also put it at the end of quite inappropriate words, as well as on regular past verbs, where it belongs. Figure 1.2 gives an example of how a very typical 71⁄2-year-old boy spelled a set of words for us, many of which ended in /t/ or in /d/. Some of these words were past verbs such as “sold”, “slept” and “told”, but others, like “next”, were not. Notice that this boy used the “-ed” ending but often put it at the end of non-verbs. So he writes “next” as “necsed” and “direct” as “direced”. In our project we worked with over 350 children, and the majority of them made such mistakes at some time during the study.

28 What is the issue?

Figure 1.2 Overgeneralizations of the “-ed” ending by a 71⁄2-year-old boy.

What is going on here? Our interpretation is that at this stage children are still treating the “-ed” sequence as some kind of a letter–sound rule. They pick up the idea that “-ed” is another way of representing the sound /t/ or /d/ at the end of a word, but they have no idea about the morphemic significance of this spelling pattern, and so they put it on the end of nouns and adjectives as well as of regular past verbs.

Morphemes and literacy: A starting point 29

It is as though they have to learn about the actual spelling and have to practice using it for some time before they tumble to its connection to the past-tense morpheme. Children make these “overgeneralizations”, as we call them, for a while, but later on they do learn to confine the “-ed” spelling to past verbs. For some time they continue to use the ending with irregular past verbs as well as regular ones, writing “slept” as “sleped” for instance, but at least this is a grammatically appropriate kind of mistake in a way that “necsed” is not. They even manage quite well to use the “-ed” ending with entirely unfamiliar past-tense pseudoverbs (“Yesterday he prelled his car”) and not with other pseudowords (“There is a preld at the end of the road”). Thus they eventually learn a genuine morphemic spelling principle: That “-ed” is the correct spelling for the regular pasttense ending. This brings us to the second question. How is this development related to children’s morphemic knowledge? In our study we used three main measures of this knowledge. The first was the “sentence analogy task” that we described earlier in the chapter. In this the child heard a sentence followed by a transformed version of the same sentence (present to past verb, or vice versa); then the child heard another sentence, rather similar to the first one, and had to transform this sentence in the same way as the first sentence had been transformed (see Box 1.3). We called the second task “word analogy”. This was much like the sentence task. The child heard a word and, after it, a transformation of this word (for example, “teacher”; “taught”); then the child was given another word and asked to make the same transformation to it (“writer”; ?) (see Box 1.3). The third task was based on Berko’s morphological study. We called it the “productive morphology task.” We gave the children pictures and we used pseudowords as part of our description of what was going on in the pictures. So, for example, one picture showed a man performing an unusual action and a little story, which the child had to complete by using the pseudoword we had used to describe the action (see Box 1.3). In this example, we said, “This is a man who knows how to snig. He is snigging onto his chair. He did the same thing yesterday. What did he do yesterday? He”, . . . and the child was encouraged to produce the pseudoverb in the past tense. Our project was a longitudinal one, which means that we saw and tested the same children many times over the 3-year period. So we were able to see how well each of our various measures was related

30 What is the issue?

to other measures over time. Time is important in analyzing these relationships. If, for example, a child’s morphemic knowledge does determine how well she or he learns morphemic spelling principles, a good measure of different children’s morphemic knowledge taken early on in the project should predict how well children will learn these spelling principles later on. If A determines B, A should precede B, and, therefore, the strength of A at one time should predict the strength of B later on. In fact, if you are examining how much one variable determines another over time, you have to take one extra step to be sure your hypothesis is right. Suppose, for example, that you want to see if the strength of children’s morphemic knowledge in one session (Session 1) has an effect on how much they have learned about morphemic spelling rules in a later session a year or so on (Session 2). Your hypothesis is really about the changes in the children’s learning of morphemic rules between the two sessions and not about how much they had learned about these rules at the beginning of the project. So you have to rule out the effect of their earlier knowledge of these spelling rules. The way to do this is to control for differences among the children in how well they knew the morphemic spelling rules in Session 1 before you examine the relationship between their morphemic knowledge in Session 1 and their use of the morphemic spelling rules in Session 2. This statistical maneuver, which is called autoregression, may sound a complicated one. In fact, it is quite easy to do. We used this way of analyzing relations between our different measures over time in our study. First, we looked at the relationship between our measures of morphemic knowledge at the beginning of the project and the children’s success in spelling at the beginning of the project and also 18 months later. We found that the children’s scores for morphemic knowledge in the first session predicted their success in spelling the past-tense inflection 18 months later, even after we had controlled for their spelling prowess in the first session. We concluded that this was strong evidence that morphemic knowledge plays a role in how well children learn about morphemic spelling rules. This is not a surprising discovery, but it is an important one because of its implications for teaching spelling, which of course is the subject of this book. If morphemic knowledge partly determines how well children learn morphemic spelling principles, one should take seriously the possibility that steps should be taken to increase children’s explicit awareness of the morphemic structure of the words that they speak and hear and read and write.

Morphemes and literacy: A starting point 31

The discovery that A affects B does not rule out the possibility that B also affects A. Taking regular exercise may make people happier and more relaxed than before, but being happy and relaxed may make people more inclined to take exercise rather than mope around at home. So, morphemic knowledge probably does affect how well children learn about morphemic spelling rules, but it is also possible that intensive experience with morphemic spelling rules could increase children’s awareness of how words are constructed from morphemes. Here we might find a two-way street. Our data suggest that this too is true. We also looked at the relationship between children’s success in spelling the past-tense ending at the beginning of the project and their morphemic knowledge in later sessions. There was, it transpired, a strong predictive relation there too, even after we had controlled for differences among the children in their morphemic knowledge in the first session. So, our results suggest a strong and thriving two-way relationship between these different aspects of children’s linguistic knowledge. Our confidence in the existence of this two-way connection was strengthened by a similar study of a very different language and script. Iris Levin, Dorit Ravid, and Sharon Rapaport worked with 5- to 6-yearold Israeli children, who were just starting to read and write Hebrew (Levin et al. 1999). In Hebrew the morphological system is rich and complex, and its effect on Hebrew spelling is at least as pervasive and important as the effect of English morphology on English spelling. So, it is useful to see if the relationships between children’s morphological knowledge and their literacy skills are much the same in this language as in English. The purpose of the project was to track these relationships over a 7-month period. At the beginning, and also at the end, of the project the researchers measured the children’s knowledge of morphemes in spoken Hebrew by asking them to transform words morphemically. An example (translated from Hebrew to English) from one of their tasks is “A baby who looks like an angel is an ________ baby.” Here the child has to derive an adjective (“angelic” in English) from the noun “angel”, and in Hebrew as in English this means that they have to find and add the appropriate derivational suffix. The research team also measured the children’s progress in writing Hebrew at the same time. In Hebrew, as in English, children tend to concentrate on the phonological principles (grapheme–phoneme correspondences) before they adopt more complex correspondences such as the correspondences between morphemic units and spelling. In this project the measures of children’s

32 What is the issue?

progress in writing Hebrew charted the extent to which they had progressed from using basic phonological principles to the more difficult principles based on morphemes. This project clearly established a “two-way street,” to use the researchers’ own term, which we have already borrowed. The Israeli children’s knowledge of morphemes at the beginning of the project predicted their level of writing at the end of the project, even after controls for initial differences between the children in their writing skills. There were also strong relationships in the opposite direction: The children’s level of writing at the beginning of the project predicted their knowledge about morphemes in spoken Hebrew at the end of the project, even after controls for differences between the children at the start of the project in their knowledge about Hebrew morphemes. This impressive set of results establishes that in this language, too, children’s sensitivity to the way in which words are constructed from morphemes and the progress that the children make in literacy interact and strengthen each other. It is a relationship of the greatest importance in English and in Hebrew, and, almost certainly, in many other languages as well, and it needs to be nurtured.

Teaching morphology: Improving spelling The mention of “nurturing” brings us to our final question in this chapter, which is about how to “nurture” children’s understanding and use of the valuable morphemic spelling rules. The evidence that we have been reviewing suggests very strongly that one good way of helping children to learn about morphemic spelling principles would be to bolster their morphological awareness. Yet, tests of this simple idea are remarkably thin on the ground. This gap really is surprising because any study in which the researchers manage to improve children’s morphological awareness and then go on to examine the effect of doing so on the children’s learning of the correspondence between morphemes and spelling could yield two most valuable insights. The first insight would be into the causal relationship between these two. If the study establishes that teaching morphological awareness leads to an improvement in spelling, it will have provided the strongest evidence possible for the causal hypothesis that we have been considering. But the second insight is even more important than that, and it is the subject of this book. A successful intervention study like this would have immediate educational significance. It would establish how possible and practicable it is to teach

Morphemes and literacy: A starting point 33

children about morphemes, and it would show us whether this sort of instruction does have beneficial effects on children’s spelling and, perhaps, on their vocabulary as well. Most of the rest of this book (Chapters 3–6) will be about a series of intervention studies that we carried out on the effects of raising children’s morphological awareness. The results of a previous project of ours on the effects of teaching children about morphemes (Nunes et al. 2003) encouraged us to embark on this new program of research. In this study, we made a direct comparison of the effects of teaching 7- and 8-year-old children about morphemes or about phonology. We gave the children a pretest before the intervention and an identical posttest soon after the intervention was finished, in which we tested their ability to spell certain affixes, such as “-ion” and “-ment”, which normally cause children of this age a great deal of difficulty, and also to follow particular phonologically based spelling principles, such as how to represent short and long vowels. In the intervention itself, we taught the children in small groups in twelve different sessions. We taught some groups about morphological distinctions and others about phonological ones, and we also included a control group of children in the study to whom we gave no teaching. We made sure that the activities given to the “morphological” and to the “phonological” groups had the same structure. So, for example, the children blended either morphemes or phonological segments, made analogies either about morphemes or about sounds, and classified words into groups that shared the same morphemes or shared the same sounds. Our morphological teaching did have a powerful effect, particularly on the children’s success in spelling affixes. The study established, we think for the first time, that it is possible to teach children about morphemes and that this teaching has a direct effect on their knowledge and use of morphemic spelling principles. The way was clear for us to begin the program of studies that started in the laboratory and ended in the classroom. These are the studies that we shall tell you about in Chapters 3–6.

Summary and conclusions Our review of the ways in which morphemes and written language are connected has led us to three simple conclusions.


34 What is the issue?

1. Some of the most important links between spoken and written language are at the level of the morpheme. The morphemic structure of words in English and several other written languages often determines their spelling. 2. The system of morphemes, therefore, is a powerful resource for those learning to read: The more schoolchildren know about morphemes, the more likely it is that they will learn about spelling principles based on morphemes. 3. However, children’s knowledge of morphemes is largely implicit. It is quite likely that they need explicit knowledge about morphemes in order to learn about the connection between morphemes and spelling. Yet, many quite simple morphemic spelling principles are not taught at school. We need to know how easy it is to teach these principles explicitly and how effective this teaching will be.

Chapter 2

What knowledge of morphemes do children and adults show in the way that they spell words?

Our book has two aims. Its first is to persuade our readers that morphemes are extremely important for children learning to read and write. Our second aim is to describe a set of studies that we carried out on teaching children about morphemes and their relation to written words. The first two chapters in the book are all about the first of these two aims. In the remaining chapters we will try to fulfill our second aim by describing our work on teaching children about morphemes and spelling. In the first chapter we showed how many spelling principles in English and in several other languages are based on morphemes, and we also reported some research that established the existence of a strong relationship between children’s knowledge of the morphemic structure of spoken words and the progress that they make in learning about written words. In this second chapter we shall look at a series of studies on children’s actual spellings and examine what they tell us about their morphemic knowledge and about the way that they are using this knowledge in their writing. So this chapter focuses on what we can find out about people’s knowledge of morphemes if we treat spelling as a window on their knowledge of morphemes. We shall show how children’s spellings tell us a great deal about their knowledge of morphemes. In the studies that we shall describe we used three different techniques to detect how people use morphemes in spelling. In these tasks we ask children:

Authored by Terezinha Nunes, Peter Bryant, Ursula Pretzlik, Deborah Evans, Daniel Bell, and Jenny Olsson

36 What is the issue?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

1. to spell words that contain a morpheme whose spelling cannot be entirely predicted from the way that it sounds but can be predicted on the basis of how that morpheme is spelled: For example, the past-tense ending; 2. to spell pseudowords which contain particular morphemes: We create invented words using actual stems and affixes, place them in sentences that clearly identify the word type, and ask the children to spell the pseudoword that we dictate to them; 3. to spell some real words and pseudowords and then to explain their choice of spelling. We carried out many different studies using these techniques. The results of these studies tell us a great deal about what children (and adults) know about morphemes without much explicit teaching, because there is currently little teaching about English morphemes in English schools. This chapter presents a summary of what we found out. The sections are organized by the target spellings and the aims of each of the studies. The first section focuses on suffixes that have a fixed spelling which is not completely predictable from oral language. The second section focuses on the spelling of stems. The third section focuses on how children and adults spell pseudowords made with real stems and suffixes and the explanations that they give for their choice of spellings. The final section presents an overview of the results and raises questions about the possibility of improving children’s knowledge of morphemes through teaching. Although the focus of the chapter is on spelling, the aim of our investigations is to understand the connection between knowledge of morphemes and literacy in a broader and a better way.

Spelling suffixes: Is it easy because they have a fixed form? Some years ago, we started investigating children’s use of morphemes in spelling. In our first study we focused on only a few morphemes. Our aim was to investigate how children spelled three different suffixes that had a fixed spelling and a clear function: “-ion”, “-ness”, and “-ed”.

Knowledge shown in way children and adults spell 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

The derivational suffixes “-ion” and “-ness” are used to form abstract nouns. Neither can be spelled on the basis of their sounds: In neither is the vowel clearly pronounced. The “-ness” suffix ends with a double “s”, which sounds no different from a single “s” at the end of words. The inflectional suffix “-ed” cannot be spelled on the basis of the way that it sounds in different words. Sometimes this “-ed” ending represents the sound /t/, as in “kissed”, sometimes /d/, as in “killed”, and sometimes /id/, as in “wanted”. For someone who understands the representation of morphemes in spelling, it should be easy to spell these suffixes because they have a clear morphemic function and are always written in the same way. However, they are unexpected spellings from the way that the words’ endings sound. So we thought that we needed to ask two questions about children’s use of these suffixes. First, we wanted to know whether children understand that there are spellings for the ends of words that often don’t accurately represent their sounds from the point of view of the traditional correspondences between letters and sounds. Second, we wanted to know whether children have a good grasp of when to use these spellings.

Children’s awareness of suffix spellings that do not correspond to ending sounds In one study we asked 710 children from a total of eight different schools in London and in Oxford to spell, among other words, four abstract nouns that ended in “-ion” (“emotion”, “destination”, “combination”, and “election”), four that ended in “-ness” (“madness”, “politeness”, “richness”, and “happiness”), and five regular verbs (“kissed”, “opened”, “laughed”, “stopped”, and “covered”). We presented all these words in the context of sentences, in order to make absolutely clear the meaning of each word that we asked the children to spell. In Chapter 1, we gave some examples of children’s spelling for the “-ed” suffix. Box 2.1 presents a summary of how the children spelled “-ness” in the word “madness” and “-ion” in the word “emotion”. We also asked the children to spell three “pseudowords”, which were presented in the context of sentences to help the children identify their function. For example, the pseudoverb “nelled” was presented in the sentence “We usually nell in the morning but yesterday we nelled in the afternoon.” This sentence makes it clear that “nelled” ought to be treated as a regular past verb. All the words in the sentences were written on a page, except for the target one, which we said in the context of the

38 What is the issue?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

BOX 2.1 Children’s spellings of “-ness” and “-ion” by year group in school “Madness” The correct suffix spelling “-ness” was used by 49 percent of the children in Year 3, 40 percent in Year 4, 68 percent in Year 5, and 80 percent in Year 6. The second most used spelling was “-nes”, which was used by 15 percent of the children in Year 3, 25 percent in Year 4, 11 percent in Year 5 and 6 percent in Year 6. Spellings with other vowels comprised 21 percent of the spellings in Year 3, 13 percent in Year 4, and 3 percent in Years 5 and 6. Other endings observed include “-ners”, “-nace”, “-nece”, “-ns”, “-nerse”. The total number of different spellings for the “-ness” ending was twentytwo.

“Emotion” The correct suffix spelling “-ion” was used by 23 percent of the children in Year 3, 33 percent in Year 4, 59 percent in Year 5, and 64 percent in Year 6. Other spellings observed include “-an”, “-en”, “-eon”, “-in”, “-on”, “-un”, “-ian”, “-ine”, “-ihon”, “-one”, “-oan”, “-une”, “-oone”, “-erne”, and also “-n” without a vowel. The total number of different spellings for the “-ion” ending was twenty-five.

sentence, and then repeated once so that the child could write it on a line that marked its place in the sentence. The children were in Years 3 to 6 of primary school and their ages ranged from 7 to 10 years. The number of children in each year group and their mean age is presented in Table 2.1. Figure 2.1 presents the proportion of correct uses of each of the suffixes for each year group. We scored the spelling as correct whenever the children used the exact spelling for the suffix, irrespective of the way that they spelled the rest of the word. The figure shows that in spite of the perfect predictability of the spelling of each of these morphemes, children do not seem to find it

Knowledge shown in way children and adults spell 39 Table 2.1 Number of children in each year group and their mean age Year group in school

Year 3 Year 4 Year 5 Year 6

Mean age (years:months) 7:9 8:10 9:9 10:9

Number of children (N=710) 122 124 229 235

100 80 % correct

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41


60 "-ness"

40 20

"-ed" in real verbs

0 7:9 years 8:10 years 9:9 years 10:9 years

"-ed" in pseudoverbs

Age in years:months

Figure 2.1 Percentage of children who spelled each suffix (“-ion”, “-ness”, and “-ed”) correctly, by age level.

easy to spell them. Only about one-third of the spellings of “-ion” were correct when the children are in their fourth year in school. Even children in their sixth year at school, almost at the age of 11, do not manage to spell any of these suffixes perfectly. It is worth pointing out that the children are unlikely to be spelling these suffixes simply on the basis of a specific learning of each of the words. The fact that they spell the suffix “-ed” at the end of regular past pseudoverbs as well as at the end of real past verbs suggests that they have learned more than specific spellings of well-practiced words. Our hypothesis is that this generalization from real to pseudoverbs suggests that the children might have some knowledge of the representation of morphemes in spelling. This first study led us to conclude that spelling even those morphemes that have a fixed form is not easy for children. The fact that the spelling is not a good representation of the way that the words sound does

40 What is the issue?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

interfere with the children’s spelling. This was demonstrated by an analysis of the errors the children made: In the word “emotion”, for example, whose final vowel is not pronounced clearly, about 98 percent of the errors were due to the use of an incorrect vowel or to missing out the vowel altogether, and only 2 percent of the spellings failed to include the letter “n” (although sometimes “n” was not the last letter in the word and it was followed by an “e” or a “g”).

Do children know when to use and when not to use the different suffixes? The case of “-ion” and “-ian” In the study just described, we considered only the use of suffixes in their appropriate place. We also carried out two other investigations that analyzed whether the children restricted the use of these suffixes to the right type of word. Consider, for example, the two words “emotion” and “magician”. Their end sounds are the same. However, “emotion” is an abstract noun, and thus it is spelled with “-ion”, whereas “magician” is a person who does something (magic) and thus it is spelled with “-ian”. We had found in the previous study that about 70 percent of the spellings of children aged 11 correctly represented the suffix “-ion” at the end of the abstract nouns we asked them to write. Would they be able to discriminate between the use of the two suffixes, “-ion” and “-ian”, although the words that contain these suffixes sound the same at the end? If they are able to discriminate between the two types of words, we can conclude that they have some insight into the way that these suffixes are used to represent different meanings. In another study, we asked 176 children from three schools in the Oxford area to spell eight real words ending in “-ion”, another eight real words ending in “-ian”, four pseudowords ending in “-ion” and four ending in “-ian”. As in the previous study, the words and pseudowords were presented in sentences, which were written on the page. A gap in the sentence marked the place where the word (or pseudoword) was to be written. The children in this study were in Years 4 and 5 in primary school. Their mean ages were 8 years 8 months and 9 years 9 months, respectively. We also asked the children to spell eight other words and four pseudowords with completely different endings, so that the words in the list would not all have the same sound at the end. A final note about this study: This study took place after the British Government had introduced the National Literacy Strategy (NLS), whereas the first

Knowledge shown in way children and adults spell 41

study that we described in this chapter was carried out before this policy had been implemented. The NLS includes in the teaching of spelling lists of words ending in “ion” to be taught to children in the year groups that participated in this study, so we can assume that the children were taught about “ion” endings. Figure 2.2 shows the percentage of correctly spelled suffixes for words ending in “-ion” and “-ian”. Three results from this study will be stressed here. First, the observations from the previous study are replicated in this study: A large proportion of the spellings of the suffix “-ion” is correct at about age 10, but the level of success is not close to 100 percent—a level of success that could be achieved if the children were using knowledge of morphemes that makes this spelling predictable. Second, the children are much less successful in spelling correctly the suffix “-ian”: Their level of success with these words is about half their level of success with the “ion” suffix. Third, the level of correct spellings of “-ion” in pseudowords is lower than that observed for words, but it is not much lower (approximately 60–5 percent). This suggests that the children are learning something more general about spelling than the specific memories of how to spell particular words. However, it is not clear from this analysis what the children are learning. Because they are explicitly taught about the existence of “-ion” at the end of words, they could not only be using

100 "-ion" in words 80 % correct

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41


"-ion" in pseudowords


"-ian" in words


"-ian" in pseudowords

0 8:8 years

9:9 years

Age in years:months

Figure 2.2 Percentage of children who spelled each suffix (“-ion” and “-ian”) in words and pseudowords correctly, by age level. Note: The ANOVA (Analysis of Variance) showed no significant difference between the two age levels, a significant difference between the words and pseudowords (p