Improving Literacy by Teaching Morphemes (Improving Learning)

KirupaVFile Attachment2000f0a9coverv05b.jpg

Improving Literacy byTeaching Morphemes

Words consist of units of meaning, called morphemes. These morphemeshave a striking effect on spelling that has been largely neglected untilnow. For example, nouns that end in -ian are words that refer topeople, and so when this ending is attached to magic we can tell thatthe resulting word means someone who produces magic. Knowledge ofthis rule, therefore, helps us with spelling: it tells us that this word isspelled as magician and not magicion.

This book by Terezinha Nunes, Peter Bryant and their colleaguesshows how important and necessary it is for children to find out aboutmorphemes when they are learning to read and to spell. The bookconcentrates on how to teach children about the morphemic structureof words and on the beneficial effects of this teaching for childrensspelling and for the breadth of their vocabulary. It reports the results ofseveral studies in the laboratory and in school classrooms of the effectsof teaching children about a wide variety of morphemes. These projectsshowed that schoolchildren enjoy learning about morphemes and thatthis learning improves their spelling and their vocabulary as well. Thebook, therefore, suggests new directions in the teaching of literacy. Itshould be read by everyone concerned with helping children to learn toread and to write.

Terezinha Nunes is Professor of Educational Studies at the Universityof Oxford and Fellow of Harris-Manchester College, Oxford.

Peter Bryant is Visiting Professor of Psychology at Oxford BrookesUniversity and Emeritus Fellow of Wolfson College, Oxford.

Improving Learning TLRP

Series Editor: Andrew Pollard, Director of the ESRC Teaching andLearning Programme

Improving Learning How to Learn: Classrooms, schoolsand networksMary 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 MorphemesTerezinha Nunes and Peter Bryant

Improving Schools, Developing InclusionMel Ainscow, Alan Dyson and Tony Booth

Improving Subject Teaching: Lessons from research inscience educationJohn Leach, Robin Millar, Jonathan Osborne and Mary Radcliffe

Improving Workplace LearningKaren Evans, Phil Hodkinson, Helen Rainbird and Lorna Unwin

Improving Literacy byTeaching Morphemes

Edited by Terezinha Nunes and Peter Bryantwith Ursula Pretzlik and Jane Hurry

First published 2006by Routledge2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN

Simultaneously published in the USA and Canadaby Routledge270 Madison Ave, New York, NY 10016

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 DataA catalogue record for this book is available from the British Library

Library of Congress Cataloging in Publication DataA catalog record for this book has been requested

ISBN10: 0415383129 (hbk)ISBN10: 0415383137 (pbk)

ISBN13: 9780415383127 (hbk)ISBN13: 9780415383134 (pbk)

This edition published in the Taylor & Francis e-Library, 2006.

To purchase your own copy of this or any of Taylor & Francis or Routledgescollection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.

We dedicate our book to Nick Pretzlik whose kindnessand cheerful support we remember with greatpleasure

Contents

List of illustrations ixSeries editors preface xiiiAcknowledgements xv

Part IWhat is the issue? 1

1 Morphemes and literacy: A starting point 3PETER BRYANT AND TEREZINHA NUNES

2 What knowledge of morphemes do children and adults show in the way that they spell words? 35TEREZINHA NUNES, PETER BRYANT, URSULA PRETZLIK,

DEBORAH EVANS, DANIEL BELL, AND JENNY OLSSON

PART IIWhat does the research tell us? 63

3 From the laboratory to the classroom 65PETER BRYANT, TEREZINHA NUNES, URSULA PRETZLIK,

DANIEL BELL, DEBORAH EVANS, AND JENNY OLSSON

4 An intervention program for teaching children about morphemes in the classroom: Effects on spelling 104FREYJA BIRGISDOTTIR, TEREZINHA NUNES, URSULA PRETZLIK,

DIANA BURMAN, SELLY GARDNER, AND DANIEL BELL

5 An intervention program for classroom teaching about morphemes: Effects on the childrens vocabulary 121TEREZINHA NUNES, PETER BRYANT, URSULA PRETZLIK,

DIANA BURMAN, DANIEL BELL, AND SELINA GARDNER

6 Can we increase teachers awareness of morphology and have an impact on their pupils spelling? 134JANE HURRY, TAMSIN CURNO, MARY PARKER, AND

URSULA PRETZLIK

PART IIIWhat are the overall implications? 155

7 Morphemes and literacy: Context and conclusions 157TEREZINHA NUNES AND PETER BRYANT

Appendix 183The four research strategies in this research programPETER BRYANT AND TEREZINHA NUNES

References 191Index 195

viii Contents

Illustrations

Figures

1.1 The first two pages of a 712-year-old girls story 261.2 Overgeneralizations of the -ed ending by a 712-year-old

boy 282.1 Percentage of children who spelled each suffix (-ion,

-ness, and -ed) correctly, by age level 392.2 Percentage of children who spelled each suffix (-ion

and -ian) in words and pseudowords correctly, by age level 41

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 45

2.4 Proportion of past regular verb endings spelled correctly and produced correctly for pseudowords in an oral task 49

2.5 Pictures of dinosaurs with their names, which the children were asked to spell 52

2.6 Proportion of word and pseudoword pairs whose stems were spelled in the same way at each age level 53

2.7 Proportion of real verb endings spelled correctly with -ed and proportion of stems spelled consistently across two words 55

2.8 Percentage of correct pseudowords with -ion and -ian spelled correctly and percentage of correct explanations, by age level 57

2.9 Percentage of correct spellings of one-morpheme and two-morpheme words, by age level 60

3.1 Design of the first teaching study 68

3.2 The mean number (out of 16) of correctly spelled -ion and -ian endings in real words in Study 1 79

3.3 The mean number (out of 8) of correctly spelled -ion and -ian endings in pseudowords in Study 1 80

3.4 The mean number (out of 16) of correctly spelled -ion and -ian endings in real words in Study 2 84

3.5 The mean number (out of 8) of correctly spelled -ion and -ian endings in pseudowords in Study 2 85

3.6 Items from a task used to make children aware of how places in a sentence frame define grammatical categories 89

3.7 Examples of items used to teach the category of prefixes that refer to number 91

3.8 Focusing on verbs 923.9 Examples of items used to practice identification of

stems and creation of person words. Playing with pseudowords was fun 93

3.10 Adjusted means at pretest and for both posttests by group for the correctness of spelling suffixes in Study 3 100

3.11 Adjusted means at pretest and for both posttests by group for the spelling of suffixes in pseudowords in Study 3 101

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 111

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 112

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 116

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 118

5.1 A description and two sample items from the vocabulary test 125

5.2 Mean scores (adjusted for pretest differences) in the vocabulary test for each testing occasion and group (maximum score = 40) 129

x Illustrations

5.3 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 130

5.4 Percentage of correct pseudoword definitions (adjusted for pretest differences) by group and testing occasion 131

5.5 Percentage correct in the pseudoword-definition test (adjusted for pretest differences) by group and testing occasion 132

6.1 One-year teacher follow-up 1486.2 Childrens scores on spelling test: A comparison of

morphology, National Literacy Strategy, and standard conditions 149

7.1 Writing of a 6-year-old boy who seems to attribute to the digraph ck the function of the split digraph V+C+e 171

Tables

2.1 Number of children in each year group and their mean age 39

2.2 Proportion of use of -ion and -ian spellings for each of the types of word and pseudoword 42

3.1 Mean age and standard deviation for the intervention and control groups in Study 1 68

4.1 Mean age in years (and standard deviation) by type of group 106

5.1 Number of children, mean age in years (and standard deviation) by year group in school and type of group in the project 124

6.1 Number of children in each teaching condition, by year group 146

6.2 Childrens average scores before the course, by teaching condition and year group 147

6.3 Average percentage increase in the childrens scores by the end of the course, by teaching condition and year group 147

Boxes

1.1 A crash course in roots and stems (and bases) 51.2 A crash course in affixes 5

Illustrations xi

1.3 How psychologists measure morphological awareness 111.4 A collision course with schwa vowels 172.1 Childrens spellings of -ness and -ion by year group in

school 383.1 The word- and pseudoword-spelling tasks used in

Studies 1 and 2 693.2 The analogy game 713.3 The correction game 743.4 The items used for the word- and pseudoword-spelling

tests in Study 3 953.5 Sample of items from the spelling test showing one

childs answers 974.1 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 108

4.2 Examples of the segmentation used in scoring the word- and pseudoword-spelling tests 113

4.3 A sample of the same boys spelling in the pretest and posttest 114

5.1 The instructions and the items in the pseudoword-definition task 127

6.1 Teachers talking about -ed endings 1366.2 Lack of awareness of -ed rule 1376.3 Teachers thinking about morphemes with connection to

meaning 1386.4 Teachers thinking about morphemes without connection

to meaning 1396.5 Teachers talking about -ion 1416.6 Theories about morphology and spelling 143

xii Illustrations

Series editors preface

The Improving Learning series showcases findings from projects withinthe Economic and Social Research Councils Teaching and LearningResearch Programme (TLRP), the UKs largest ever coordinated edu-cational research initiative.

Books in the Improving Learning series are explicitly designed tosupport evidence-informed decisions in educational practice andpolicymaking. In particular, they combine rigorous social and edu-cational science with high awareness of the significance of the issuesbeing researched.

Working closely with practitioners, organizations, and agenciescovering all educational sectors, the program has supported many of theUKs best researchers to work on the direct improvement of policy andpractice to support learning. Over sixty projects have been supported,covering many issues across the life course. We are proud to present theresults 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 infor-mation is required, the books may be used as a gateway to academicjournals, monographs, websites, etc. On the other hand, shortersummaries and research briefings on key findings are also available via the programs website at www.tlrp.org.

We hope that you will find the analysis and findings presented inthis book are helpful to you in your work on improving outcomes forlearners.

Andrew PollardDirector, TLRP

Institute of Education, University of London

Acknowledgements

As we wrote this book, we became steadily more aware of the hugeeffort by very many colleaguesresearchers, teachers, and illustra-torsand many institutions that made this publication possible. The Economic and Social Research Council (ESRC) was the majorsupporter of the intervention studies through the Teaching and LearningResearch 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 essentialsupport for the investigations that made it possible for us to develop theinterventions. We are very grateful for the support of these researchcouncils, without which the research reported here would not have beenpossible.

Many teachers and children in different schools participated in thelongitudinal phases of this work. In Oxford: Wolvercote First School,Botley Primary School, Cassington Primary School, Kennington PrimarySchool. In London: William Tyndale Primary School, Honeywell Infantsand Junior School, Ravenstone Primary School, and Trinity St. MarysChurch of England School. Miriam Bindman and Gill Surman workedin this initial project and were excellent collaborators.

The early stages of the development of interventions received theinestimable cooperation of teachers and children in eight schools inLondon and thirteen schools in the Oxford area. In London: BessemerGrange Primary School, Dulwich Hamlet Primary School, Hargrave ParkPrimary 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 Schoolin Wheatley; and in Oxford: St. Nicholas, Marston, Bayswater MiddleSchool, Larkrise Primary School, Marston Middle School, SS Philip and

James Primary School, East Oxford Primary School, Frideswide MiddleSchool, St. Andrews Primary School, Cutteslowe Primary School, NewHinksey Primary School, and Woodfarm Primary School.

The Directors of the Hillingdon Cluster of Excellence, Rodney Staffordand Peter Shawley, as well as the teachers and children in the schoolsthat participated in the collaboration with Oxford Brookes Universitysupported the largest part of the intervention studies carried out in theclassroom. These were Brookside Primary School, Charville PrimarySchool, Cherry Lane Primary School, Colham Manor Primary School,Grange Park Infant School, Grange Park Junior School, LongmeadPrimary School, Minet Infant School, John Penrose Primary School,Pinkwell Primary School, Wood End Park Primary School, and YeadingJunior School.

We are very grateful to all these teachers and children whoseparticipation made our research possible.

Very special thanks are directed to our colleagues and long-timecollaborators in Lauriston Primary School, including the Principal,Heather Rockhold, whose rock-solid collaboration for more than tenyears has taught us so much. Her team over these years included HillaryCook and Sue Dobbing, who worked alongside us in each project,Gwenan Thomas, Aidan OKelly, Natasha Nevison, Alison Rosica, andAaron Bertran. We feel privileged to have been able to work with themfor so long. They were occasionally, but not always, supported byDepartment for Education and Skills (DfES) Best Practice grants, whichhelped them to develop their research skills and to analyze their practicein greater depth.

The teachers who attended the morphology course were also truly ourpartners in this research, testing their children, marking and enteringdata, teaching the interventions, nagging us about the rigor of ourresearch and the management of the intervention sessions. Participantteachers and schools were: Maggie Bacon, Nick Bonell, Kay Croft, KarenHenry (Kingswood Primary); Louisa Lochner (Gateway); KathyThornton (Kingsgate Primary); Stephen Buzzard (New End Primary);Lucinda Midgely (Linton Mead Primary); Rachel Webber (WatersideSocial, Emotional and Behavioural Difficulties (SEBD) Primary);Sameena Bashir (Fullwood Primary School); Karen Bloomfield (CoppicePrimary School); Caroline Havers (Mayespark Primary School); CarinaMcleod (Cleveland Junior School); Caroline Rogers (Beckford JuniorSchool); Sophia Shaikh, Barbara Turner (Woodlands Junior School);Deborah Walters (Christchurch Primary School); Nathalie Allexant(Gallions Primary School); Bryony Roberts (Edith Neville School).

xvi Acknowledgements

All the teachers who participated in our research were tremendouslygenerous 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 askedfor their feedback and suggestions.

Our assessments and interventions included illustrations that led togreater enjoyment by the children. Eldad Druks drew the dinosaurs forour pseudoword testing, and Adelina Gardner did all the illustrationsfor all the remaining materials. We, and the children, were fortunate tobenefit from Addys talents and imagination.

So many children generously agreed to give their time freely so thatother children in the future could benefit from what they helped us tofind out. Their participation was essential and made our work in schoolsgreat fun.

So, THANK YOU EVERYONE!

Acknowledgements xvii

Part I

What is the issue?

Chapter 1

Morphemes and literacyA starting point

What morphemes are

Take a fairly simple word like unforgettable. Its meaning is clear andwidely understood, but the word has three different parts to it, and itis the combination of these three parts that gives the word its final andoverall meaning.

The three parts to unforgettable are un- and forget and -able.Forget is actually a verb, because it refers to an action. Putting -ableon the end of this verb makes it into an adjective (forgettable), whichtells us that one can easily forget the person or event that the adjectiveis describing. The addition of un- at the beginning of the adjectivegives 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 differentmeaning or has no meaning at all. Each of the three parts in

We all know that words have meanings, but not everyone under-stands that the meaning of any word depends on its underlyingstructure. Words consist of morphemes, which are units of meaning.These morphemes, in our view, are of immense importance inchildrens learning of the meaning of new words and also in theirlearning how to read and write familiar and novel words. The aim of our book is to show how important morphemes can be inchildrens education and how easy it is to enhance their knowledgeabout morphemes and thus to increase the richness of theirvocabulary and the fluency of their reading and writing.

Authored by Peter Bryant and Terezinha Nunes

unforgettable therefore is a unit of meaning. The technical term for aunit of meaning is a morpheme. Some words contain one morphemeonly, but many other words in English and in other languages containmore than one. Forget is a one-morpheme word, forgettable a two-morpheme word and unforgettable, as we have seen, contains threemorphemes. So, when more than half a century ago thousands of peoplecrooned the popular Nat King Cole song Unforgettable, they wererepeating a three-morpheme word whose meaning they understoodperfectly, though they may not have been completely aware that theword had three separate units to it or that these units were calledmorphemes.

In general, people do have some awareness of morphemes, although,as we shall be showing later on in the book, this awareness tends to behazy and incomplete. Nevertheless, we can easily work out the meaningof entirely new words if these words are combinations of morphemeswhose meaning we already understand. All of us immediately knewwhat Toni Braxton meant when we heard her desperate, but charming,plea Unbreak my heart, uncry my tears. None of us had met the worduncry before, but because we knew that adding un- to the beginningof a word reverses the meaning of this word (untie, untidy, unfor-gettable) 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 greatimportance is between roots or stems (see Box 1.1) and affixes (Box1.2). Every word with more than one morpheme in it contains a root,and this is combined with one or more than one affix morpheme (seeBoxes 1.1 and 1.2 for a more detailed description of these morphemes).The words meaning starts with its root in the sense that the word wouldbe meaningless without this particular morpheme. Forget is the rootmorpheme in unforgettable and un- and -able are both affixes.Affixes that precede the root are called prefixes and those that followthe root are called suffixes. These are the only kinds of affix that we have in English, but other languages, such as Swahili, also haveinfixes, which are added-on morphemes that appear in the middle ofthe root.

Another essential distinction is between derivational and inflec-tional affixes. Inflectional-affix morphemes, or inflections for short,tell us what kind of a word we are dealing withwhether 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

4 What is the issue?

end of kissed and the -er and -est at the end of kinder andkindest are inflections, and they combine with the root to producetwo-morpheme words with a root and an affix.

Morphemes and literacy: A starting point 5

BOX 1.1

A crash course in roots and stems (and bases)

There is a distinction to be made between roots and stems, althoughfrom the point of view of this book it is not a particularly importantone. The root is the basic part of the word that remains when all derivational and inflectional affixes have been removed. Forexample, teach is the root for the word teacher and also for theword unteachable. The stem, on the other hand, is the part of theword that remains when all inflectional affixes have been removed.Teacher therefore is the stem for teachers. Thus, sometimes theroot and the stem are the same, but sometimes they are different.Cat is both the root and the stem for the plural word cats, butteach is the root and teacher the stem for the plural wordteachers. In all the examples and the tasks that we shall describein 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 isconcerned.

The base or base word is another related term and it is relevant toour book. This refers to the word from which a complex word isderived (for example, touchable is the base for untouchable).Thus in the word unbearable, bear is the root, bearable is thebase, 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 orthe root of a word (see Box 1.1 for the distinction between stems androots). These affixes either come before the root or follow it. Those


that come before the root are called prefixes and those that follow itare suffixes.

There are two types of affix: Inflectional and derivational affixes.Inflectional affixes, or inflections, give you essential information aboutthe word. For instance, all nouns are either singular or plural, and inEnglish the presence of an /s/ or a /z/ sound at the end of a nounusually means that the word is in the plural, whereas its absenceusually signals that it is a singular noun. This end sound is the pluralinflection. When you hear the word cats or the word dogs theinflection at the end of each word tells you that it refers to morethan one animal. Similarly, the absence of the s at the end of anEnglish noun means, in most cases, that the noun is a singular one.

There are inflections in English for nouns (the plural -s and thepossessive -s), adjectives (the comparative -er and thesuperlative -est), and for verbs (the past tense -ed, the third-person 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 moreinflected than English. In these other two languages, for example,there are plural inflections for adjectives as well as for nouns. Somelanguages also mark gender in adjectives as well as nouns withinflections.

Derivational affixes are different. Adding a derivational affix to a word creates a different word, which is based on the original wordbut not the same. Sometimes the difference between the base wordand the derived word is that they belong to different grammaticalclasses: For example, the derivational suffix -ness changes adjectivesinto abstract nouns ( for example happyhappiness) and thesuffix -ion changes verbs, again, into abstract nouns (for example,educateeducation). The suffix -ful changes nouns intoadjectives (for example, helphelpful, hopehopeful). Otherderivations such as un- and re- bring about a radical change inthe meaning of the base words to which they are attached (forexample, un-helpful, re-born) but do not affect their grammat-ical class. Some derivational affixes are prefixes and others suffixes.

Derived words include the base word from which they are derivedbut in many cases the pronunciation of the base word changes in the derivation, as in fifth, which is derived from five, andelectricity which is derived from electric.

Derivational morphemes create new words based on old ones. Un-, which, as we have seen, reverses the meaning normally given tothe 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 Englishwords, such as unbearable. Consider the relatively new coinage of theword doable (do-able), which we ourselves have heard our studentsand our builder use: Its doable, they say, and we instantly under-stand what they mean, even though they often turn out to be wrong.This suffix is a derivational morpheme because it changes the word fromthe verb, represented by the base word, to an adjective, which says thatthe action referred to by the verb is entirely possible.

By now you should know, if you did not know before, how manymorphemes there are in education or in uneducated (there are twoin the first word and three in the second). You should be able to workout whether the affix at the beginning of incompetent and the affix atthe end of kisses are derivational or inflectional (derivational inincompetent and inflectional in kisses). You should also have notedthat there is a strong connection between morphemes and grammar:You can use the -ed at the end of verbs, in order to convey the meaningof past tense, but you cannot use the -ed ending with nouns; nounsdont have a past tense. Once you are completely clear about roots andaffixes, prefixes and suffixes, and derivations and inflections, we knowthat you will want us to justify our claim that these morphemes play a crucial but neglected role in childrens development and in theireducation.

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.Wodehouses joke about the word disgruntled is so very funny. Hewrote of a man who was consumed with anger: If not actuallydisgruntled, he certainly wasnt gruntled. This understatement isamusing because although he followed strict morphemic principles,Wodehouse managed to create a word that we never use. The mor-pheme dis-, like the morpheme un-, reverses the meaning of thebase that it is attached to, and so gruntled should be the opposite ofdisgruntled, but this is an unused word. We know both these things,and it is the tension between them that makes us laugh.

Wodehouses joke helps us make another point about morphemes,which is that morphemes and grammar, these inseparable friends, forma basis on which we build the learning of new words. Philosophers,linguists, and psychologists have pondered at the marvel that it is to


learn a word with all that this learning implies. If a mother points to adog sniffing the lamp post, and says to her baby Look at the dog, howis the child to know that the mother means by dog the animal and notthe 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 anidea about what a new word means (Brown 1957). In the sentenceLook 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 theaction, as the meaning for the word dog. Browns studies actuallyrequired much more from the children than the distinction betweennouns and verbs. He created a technique, which we will use often in ourresearch, of observing how children learn a made-up word. The reasonfor studying how children learn made-up words, which are calledpseudowords or nonsense words by researchers, is that because theword is made up by the researcher, we can be certain that the child hasnot come across it beforejust like gruntled in Wodehouses joke.

To clarify how the technique works, consider one of the examplesused by Brown in his research. He showed children in the age range 35a picture of a pair of hands kneading a strange substance in a strangecontainer. To some children he said In this picture you can see somesibbing; to other children he said In this picture you can see somesib; 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 thatsibbing refers to the action; those who were told some sib shouldconclude that sib refers to the substance; those who were told a sibshould conclude that sib refers to the container. Each of the childrenwas then shown three pictures, one that depicted the same action on adifferent substance and with a different container, one depicting thesame substance but a different action and container, and one depictingthe same container but a different substance and action. The childrenwere able to choose the correct picture more often than one wouldexpect if they were just guessing. With three pictures to choose from, ifthey were just guessing they could be right one-third of the time, butthey were right more than two-thirds of the time for any of thesedifferent presentations.

Later work by many other researchers interested in childrens learningof vocabulary (see further readings by Gleitman and colleagues:Gleitman 1990, Gleitman and Gleitman 1992) confirmed that childrendo use their implicit knowledge of grammar in learning vocabulary.They referred to this idea as the Syntactic Bootstrapping Hypothesis,


to indicate that children use grammar to help narrow down the meaningof 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 grammarto bootstrap word learning. There is some, albeit limited, evidence forthis. It is known that older children, in the age range 1113, learn the definitions of pairs of pseudowords better if the pseudowords sharea stem (for example flur and flurment) than if they do not. Thisindicates that they can use what they learned about one stem whenlearning the derived pseudowordsthat is, they can use morphemicbootstrapping, not only syntactic bootstrapping.

In summary,

Words are formed with units of meaning, termed morphemes. Morphemes and grammar are strongly connected because inflec-

tional morphemes can only be applied to particular grammaticalcategories and derivational morphemes are used to form words ofparticular grammatical categories.

Research shows that morphemes are not just ways in whichlinguists analyze words: People use knowledge of morphemes andgrammar to learn the meanings of new words.

Why are morphemes important in education?

For our answer to this question we turn to childrens explicit knowledgeor awareness of the language that they speak and to which they listen.We shall argue that schoolchildren need to become explicitly aware ofprinciples of language, which at earlier ages they learned and obeyedat an implicit level only. Once at school they need to develop explicitknowledge of language, in general, and of morphemes, in particular,which they can think about and can even talk about much more openlyand explicitly than they had before.

We shall be arguing that schoolchildren need this new explicitknowledge about morphemes for two main reasons. One is that it isessential in learning to read and to spell. The other is that morphemicknowledge plays a central role in the growth of schoolchildrensvocabulary, because large numbers of the words that they have to learnat school are derived (with the help of derivational morphemes) fromother words.


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

why we concluded that childrens knowledge of morphemes is at first implicit and that there might be ways of increasing the levelof childrens explicit knowledge about these units;

why explicit knowledge of morphemes may be an essential ingre-dient of learning to read and to write;

why children also need explicit knowledge about morphemes tokeep to a respectable level of vocabulary growth while they are atschool.

Implicit and explicit knowledge of morphemes

Young children begin to understand and to use morphemes from anearly age. English-speaking children usually begin to produce two-morpheme 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 forpossessive words (Adams ball), for the plural (dogs), for presentprogressive verbs (I walking), for third-person singular present tense verbs (he walks), and for past tense verbs, although not alwayswith complete correctness (I brunged it here) (Brown 1973). Noticethat these new morphemes are all of them inflections. Children tend tolearn derivational morphemes a little later and to continue to learnabout them right through childhood, as we shall show in later chapters.Nevertheless, from their third year on, with little or often no explicit helpfrom other people, they master the system of roots, prefixes, and suffixeswith ease. By the time that they go to school they are morphemicexperts. They are, to derive a new word, morphemists.

They are experts, however, only at an implicit level. They are soon ata loss when given quite simple tasks that need some explicit judgmentabout morphemes. These tasks do not require children to know anythingabout the terms that we set out in the previous section (morphemes,roots, prefixes, suffixes, etc.), but they do require the children to reflectabout some fairly basic morphemic similarities between words, andyoung children find them very hard indeed.

One such task is a simple analogy task that we devised ourselves andgave to a large group of children in the 69 age range (Nunes et al.


1997b). Our aim in this task was to find out how well individual childrencan 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: Thedog scratched the chair. Immediately after that, we said another verysimilar sentence, also with a present-tense verbThe dog is chasing thecatand we asked the children to say it back to us, but like the puppetwould say it. We wanted to see if the child, as the puppet had done withthe previous sentence, could change the verb by removing the present-continuous-tense inflection -ing and adding instead the past-tenseinflection -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 thepresent-continuous inflection and add the past-tense inflection instead.However, this apparently simple task was quite difficult even for manyof the oldest children in the group. We found that 6-year-old children,all of whom could spontaneously produce present- continuous- andpast-tense verbs in the right places in their own speech, only managedto get 31 percent of the items right in this morphological test. For the7-year-old group this figure rose to 41 percent and for the 8-year-oldsto 56 percent. So, children get better at this task as they grow older, buteven the oldest make many mistakes.

Many other morphological awareness tasks that were invented byour team and still others that were devised by other research teamshave produced the same results. Most young schoolchildren fluentlyspeak and effortlessly understand words that are quite complicated froma morphemic point of view. Yet, they are usually completely, albeit quitecheerfully, at sea when asked to make simple comparisons of themorphemes in different words. Box 1.3 presents a sample of differenttasks used to assess childrens awareness of morphology.


BOX 1.3

How psychologists measure morphological awareness

The aim of all morphological awareness tasks is to measure childrensor adults conscious knowledge of the morphemic structure ofspoken words. There is a wide variety of such tasks.


Productive morphology (Nunes et al. 1997a,adapted from Berko 1958)

The tester says two sentences, which contain an entirely unfamiliarpseudoword and then invites the child to complete a sentence usingthat pseudoword with the target inflection. Each item is presentedalong with a picture. The picture used for the first item is includedhere to illustrate the method.

1. This is a man who knows how to snig; he is snigging onto hischair. He did the same thing yesterday. What did he doyesterday? 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 thesame thing. What does he do every day? Every day he?

3. This person is always tigging his head. Today, as he falls to theground, he tigs his head. Yesterday he did the same thing. Whatdid he do yesterday? Yesterday he?

4. Be careful, said the farmer. Youre always clomming on yourshoelace. Youre 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. Todayhe will do the same thing. What will he do today? Today he will?


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 handsand shouted with?

9. It was night-time and the moon was shining. He danced luggilyand smiled with lugginess. He felt very?

10. When the sun shines he feels very chowy. He dances chowilyand laughs with?

The sentence analogy task (Nunes et al. 1997a)

The tester uses puppets to present the sentences. The first puppetsays the first sentence in the pair; the second puppet says thesecond sentence. Then the first puppet says the first sentence in thesecond pair and the child is encouraged to help the second puppetand 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 saysthe first word in the pair; the second puppet says the second word.

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 psy-chologist, did a classic experiment in which she used pseudowords (as did Roger Brown) like wug to avoid testing childrens specificknowledge and thus to arrive at some conclusion about their knowledgeof morphemic principles (Berko 1958). In her best-known question,


Then the first puppet says the first word in the second pair and thechild is encouraged to help the second puppet and say its word.Each item presents the corresponding pairs.

1. anger : angry :: strength : ________2. teacher : taught :: writer : ________3. walk : walked :: shake ________4. see : saw :: dance ________5. cried : cry :: drew ________6. work : worker :: write : ________7. sing : song :: live ________8. happy: happiness :: high : ________

Test of morphological production (Fowler andLiberman 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 bigracehorse came in ________) or they are given the derived formand have to produce the base form (for example, Fourth. When he counted the puppies, there were ________). The word pairseither fit into the phonologically neutral or phonologically complexcondition. The same suffix was used for each pair to make theconditions more comparable.

Phonologically neutral Phonologically complexdanger : dangerous courage : courageousshine : shiny anger : angryfour : fourth five : fifthagree : agreeable respond : responsibleexamine : examination combine : combinationsuggest : suggestion decide : decision

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 thesame creatures and asked them to describe it. Most children, down tothe age of 4, produced the correct answer of wugs, and from this Berkoconcluded that they knew about the nature of the plural -s inflection.

The children were as successful in their answers to some otherquestions that involved past verbs. For example, referring to a manexercising, Jean Berko told the children This is a man who knows howto gling. He is glinging. He did the same thing yesterday. What did hedo yesterday? Yesterday, he, . . . and then she asked the children tocomplete 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-schoolchildrenare explicitly aware of the past-tense inflectional morphemes as well asof the plural ending.

Yet, some other results from the same Berko study gave the lie to thisoptimistic claim. When the children were asked to make singularnonsense words with /s/ or /z/ sound endings into plural words (forexample niznizzes), and present-tense nonsense verbs ending in /t/or /d/ sounds into past verbs (for example, motmotted), theynearly always failed to do so. It was as if the children had some vagueidea that plural words end in /s/ or /z/ sounds and that past verbs endin /t/ or /d/ sounds, and yet do not understand that plural words consistof two parts: the root or stem (see Box 1.1), which is the same as thesingular 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 deriva-tional morphemes on nonsense words, but she found that the childrenwere strikingly unsuccessful. For example, she asked the children, andsome adults too, what would they call a man whose job is to zib. Allthe adults formed a new noun by adding a derivational suffix -er toform a new word, zibber, but only 11 percent of the children were ableto come up with this word. They simply found it too difficult toconsciously derive an agent from a verb.

The answer to our first question about young schoolchildrens explicit (as opposed to their implicit) morphemic knowledge is thereforemostly negative. When children arrive at school and during their firstfew years there, they have some awareness of the morphemic system,which they themselves use in their own conversations with extra-ordinary proficiency. But this awareness is only a weak one. Morphemesare an essential part of the young childrens everyday life, but these


youngsters are barely conscious of them or of their importance. Whatimplications does this have for the learning that they have to do atschool?

Explicit knowledge of morphemes may be anessential ingredient of learning to read and to write

It is an important, though shockingly neglected, fact that one of the bestways to help children to become experts in reading and spelling is tomake sure that they are thoroughly familiar with the morphemic systemin their own language. This kind of knowledge may not be an absoluterequirement for learning how to read and write English, Portuguese,Greek, French, German, Arabic, and Hebrew, but it certainly will makethis learning an easier and a more successful task.

The main reason why morphemic structure is so important for readingand writing in these and in many other languages is that morphemesaffect the ways in which words are spelled. If you want to know whatmany written words are, particularly new words, and if you want toknow how to write words, and, again, new words in particular, youreally have to be able to work out their morphemic structure.

Morphemes have such a powerful effect on spelling for three goodreasons, which we shall look at in turn.

1. The same sounds are spelled in different ways in differentmorphemes.

2. It is often the case that a particular morpheme is spelled in the sameway, even though it is represented by different sounds in differentwords.

3. Some morphemes are represented in writing but not in speech.

The same sounds are spelled in different ways in differentmorphemes

The first of these points needs particular attention from those who aretempted to think that the be-all and end-all of teaching children to read is to encourage them to learn about the relationship betweensounds and letters or sequences of letters. This doctrine is no help at allto a child who wants to know why the ending of locks and fox soundexactly 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 its final /s/ sound is the plural inflection, which isrepresented by -s in regular plural words in English. The second word,fox, on the other hand, is singular and therefore there is no reason tobreak up the /ks/ sound at the end. In such words, this ending isrepresented 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 atwo-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. Thisactually is a clear and inflexible principle in English spelling: Take everyword 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 andBryant 2003).

So far, we have contrasted two- with one-morpheme words, but it iseasy to show rather similar effects of morphemes on spelling in contrastsbetween different two-morpheme words. The point here is that thereare affix morphemes that have quite different functions and yet soundexactly the same. Sometimes these different affixes are spelled in thesame way, like the -er ending. When -er represents an affix, it is acomparative (bigger, braver, cleverer) in some words and anagentive in others (baker, sweeper, cleaner).

No problem there, but what about the -ion and -ian endings ineducation and magician? These endings sound exactly the same (ifyou dont believe this, say both words out aloud and listen carefully),but they are spelled quite differently. Both ending syllables contain aschwa vowel followed by the /n/ sound. (See Box 1.4 for an explanationof schwa vowels and for an object lesson in why children need to knowabout morphemes when they are learning to spell.)


BOX 1.4

A collision course with schwa vowels

Educationalists and psychologists are fond of the term spellingdemons, a term that they use to describe words whose spellingflouts conventional spelling rules. In our view, however, the worstdemon in English spelling is not a word, but a particular sound. This


is the schwa vowel sound, which, to take one pair as an example, isthe last sound both in Bognor and in picture. The schwa vowelis easily the most frequently used vowel sound in the Englishlanguage, and yet there is no set way of spelling it on the basis oflettersound rules. It is also known as a weak vowel soundonethat is rather poorly articulated.

Schwa vowels crop up in profusion in words of more than onesyllable, and they are always in the unstressed part of the word. Hereare 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-fourvowels that we have pinpointed in this list are all the same vowelsound, since the sound is spelled in so many different ways in thedifferent words. But with a moments reflection, and perhaps withthe help of pronouncing the words out loud, you will see that theyare all one and the same sound. The variety of ways in which thissound is spelled in English is truly astonishing. In this small list we counted six different spellings for the schwa vowel (a, e,

There must be some reason for this difference in the spelling of thesetwo affixes, and the chances are that it is a morphemic one, since thesewritten endings usually do represent morphemes. Yet none of the tomeson English spelling, no educational textbook, nor any one of the manyaccounts of the psychology of reading and spelling provide any kind ofa clue to the reason for the two different spellings for this ending, eventhough 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 thisending with nouns. If the noun refers to a person or an animal, its endingis spelled as -ian (magician, mathematician). If it does not referto a person, it is spelled as -ion (education, institution). There arehardly any exceptions to this principle, and these few exceptions are allwords that are quite uncommon ones (radian, centurion).

This is a distinction that should cause no particular difficulty to 7- and8-year-old children. Teachers, therefore, should be able to put it across


o, ia, ia and ou), and there are many other spellings for thispromiscuous vowel.

This range of spellings for the same sound is as good an illustrationas one can possibly find for the inadequacy of treating or teachingEnglish spelling as a system of rules about lettersound relationshipswith a few exceptions. With the schwa vowel there is no lettersoundrule. At the level of phonology, every spelling is an exception.

On another level, however, which is the level of morphemes, thereis a set of principles that can guide the spelling of this sound in verymany words. These are morphemic spelling principles. Look againat the list and you will see that the first thirteen words (happinessto rehearsal) are all two- or three-morpheme words with a stemor base followed by a suffix, and with a schwa vowel in each suffix.The spelling of each of these thirteen endings is highly consistentacross different words with the same suffix (glorious and furious,magician and logician, happiness and sadness) and whentwo different suffixes sound exactly the same they are sometimesspelled differently (attraction and mathematician). Thus, thespelling of the schwa vowel is often determined by the meaning,rather than the sound, of the word. Meaning, and the morphemeswhich convey that meaning, can often tame this particular spellingdemon.

to their pupils quite easily. Yet, as far as we know, no one teaches ourprinciple about -ion and -ian endings in schools in England, and, asthe next chapter will show, the pupils continue to make frequent andrather serious mistakes when writing words that ought to have one orthe other of these two endings.

The -ion/-ian issue is something of a test case for us. We areinterested in morphemic spelling principles and, particularly, principlesthat could be, but are not, taught at school. We are also interested inspelling patterns that cause children great, and possibly quite unnec-essary, difficulties. The -ion/-ian endings fit both these requirementsand raise two clear and pressing questions:

1. Can schoolchildren be taught this morphemic spelling principle?2. Will this teaching help them to spell these difficult words?

We shall present our answers to these questions in the chapters thatfollow.

It is worth mentioning at this point that the same questions can beasked about other languages. In Portuguese there are several instancesof the same sound being spelled in different ways in different mor-phemes. For example, the endings of the words princesa (princess)and pobreza (poverty) sound exactly the same (they both rhymewith the English word blazer), but they are spelled differently becausethe -esa ending is the right one for the derivational morpheme thatrepresents a female, while the -eza ending is the conventional spellingin abstract nouns that end with that suffix.

In modern Greek, which is known as a highly regular script, childrenstill have to learn to pay particular attention to morphemes (Aidinisand Nunes 2001, Bryant et al. 2000). They need to do so because theGreek language has few vowel sounds and many ways to spell them. For example, there are many different ways to spell the ee vowelsound, as in feet, in Greek words, which becomes a problem for Greekchildren because they have to learn which spelling to choose for thissound in different words. The best help that Greek children get inmaking this choice is from morphemes (Bryant et al. 1999, Chliounakiand Bryant 2003). Greek root morphemes are always spelled in thesame way, of course, and so whole families of words always use thesame spelling for the vowel or vowels in the root that they have incommon. Also, different Greek inflections are spelled differently evenwhen they sound the same (as with -ian and -ion in English). Fourdifferent inflections are signalled by the ee sound at the end of Greek


words and there is a different way of spelling each of them. The spellingfor feminine singular endings on nouns and adjectives is , for masculineplural endings is , for neuter singular endings is and for third-person-singular present-tense verb endings is .

It is often the case that a particular morpheme is spelled inthe same way, even though it is represented by differentsounds in different words

The central point of this chapter is that in English and in many otherlanguages there is a system of relationships between morphemes andspelling and that it will help children immensely to know what thissystem is. One of the most compelling reasons why schoolchildren needto know about this system is that in many cases there is a constantspelling for a particular morpheme, even though the sound of thatmorpheme differs from word to word.

For instance, we take medicines to heal ourselves, and we worry aboutour health. These two words share the same root morpheme, and thespelling for this morpheme is the same in both words even though thevowel sound is long in the one-morpheme word, heal, and short in thetwo-morpheme word, health. Muscle and muscular form a similarpair: the sc sequence represents one sound, /s/, in the first word buttwo, /sk/, in the second. The reason for this apparent inconsistency inlettersound correspondences is that the two words share the same root.The relationship between letters and sounds is inconsistent, but therelationship between letters and morphemes is entirely consistent. Onceagain, 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 connectionsbetween spelling sequences and morphemes, despite inconsistentconnections between these same spelling sequences and sounds. The past-tense ending in verbs is the most powerful example, and aninteresting one from our point of view, because it is one of the fewconnections between morphemes and spelling that teachers tell theirpupils about at school. In regular past verbs there are three differentpronunciations for the past-tense ending /t/ as in kissed, /d/ as inkilled, 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 getto grips with this particular spelling principle, despite being taughtabout it in the classroom. One problem for them is that they have to


distinguish not just between past verbs and similar-sounding wordsthat 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 inspeech

Morphemes are important in reading and writing for a third reason,which is that some morphemic distinctions are explicit and clearlysignaled in writing but not in speech. In some ways, this point is at leastas important for childrens acquisition of spoken language as it is fortheir learning about written language, because it is entirely possiblethat they may eventually become aware about these particularmorphemic distinctions in speech through seeing them in print.

The apostrophe, which is notorious for the difficulties that it causesadults and children alike, is a case in point. In the English script itrepresents either an elision (cant for cannot; its for it is) or thepossessive (the boys cousin; the girls teacher). The possessiveending is an affix morpheme, and so we will concentrate on that for themoment.

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 theboys drink have entirely different meanings in their written form. Inone phrase, boys is a plural noun and drink refers to what they aredoing. The other phrase is about a boy in the singular and drink is anoun. Both these fundamental differences are signaled simply by theabsence of an apostrophe in one passage and its presence in the other.

In spoken language it would be quite a different matter: Although thetwo passages have quite different meanings, they sound exactly thesame. Of course, listeners who hear one of the passages would soon beable to infer what the person speaking to them had meant by it, butthey would have to use the context to do that. The spoken words on theirown are ambiguous; the written words are not.

Precisely because writing represents a distinction here which spokenlanguage does not, we should expect it to be quite hard for people tolearn about the possessive apostrophe. In fact, many people, and not justgreengrocers, have real difficulties with this morphemic spelling:Adults, as well as childrens, knowledge of when and when not to usethe apostrophe is often distinctly sketchy.


Lest you should think that we are dealing with a peculiarity of theEnglish language here, we shall show you now that the French writtenlanguage also signals morphemic distinctions that are completelyhidden in spoken French. Plural endings in French nouns, adjectives,and verbs are for the most part silent. The word for house soundsexactly the same in the plural as it does in the singular (la maison, lesmaisons), even though the two forms have different spellings, and thisis true of most other nouns as well. It is the same with verbs: Third-person singular, and plural verbs have exactly the same sound /e/ in thepresent tense (il aime, ils aiment) but are spelled differently. Thus,the plural affix appears in writing as -s at the end of plural nouns andadjectives and as -nt at the end of plural verbs, but not in speech.

These silent plurals cause French schoolchildren a lot of difficultywhen they first learn to write. In an intriguing series of studies, MichelFayol, a French psychologist, and his colleagues have clearly shown asequence 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 theplural ending out: They write the words as they sound, and, sincethe plural endings have no sound, they do not represent them intheir spelling at all. So, they usually make the mistake of writingles arbres, for example, as les arbre.

2. Later on, they do learn about the plural -s ending, but they tendto use it altogether too frequently, since they often put it at the endof verbs as well as at the end of plural nouns and adjectives. Theywrite ils aimes instead of the correct ils aiment.

3. Next, they learn about the -nt ending as well, but again their use of this ending is often indiscriminate. Sometimes, they putthe -nt ending on some nouns and adjectives too. For instance,they sometimes write les maisonent instead of the correct lesmaisons.

4. Finally, and usually with the help of a great deal of instruction inthe classroom, they manage to make the distinction between pluralnoun and verb endings.

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


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

Cause and effect in the connectionsbetween childrens knowledge ofmorphemes and their learning to spellmorphemes

Our last point about the effect on French-speaking children of findingout about morphemes through learning how to spell raises a generalquestion about the direction of cause and effect. In most of this chapterwe have been emphasizing possible causes and effects in one directiononly. We have argued that childrens knowledge about morphemes mustbe a powerful and necessary resource in learning to read and write.Morphological knowledge, according to this view, should have a strongeffect on childrens reading and spelling. Now we should also considerthe possibility that cause and effect might take the opposite directionas well. Learning to read and write might alert children to morphemicdistinctions that had escaped them before: Their experiences withwritten language also cause a change in their explicit awareness ofmorphemes.

The idea of a two-way streetfrom reading and writing to mor-phemic knowledge as well as from morphemic knowledge to readingand writingis at its most plausible when morphemic distinctions areexplicit in writing but hidden in speech, as happens with the possessiveapostrophe in English and with plural endings in French. But it mightalso be true of morphemic distinctions that are explicit both in writtenand in spoken language. Here, too, childrens experiences with writtenlanguage might alert them to the structure of morphemes in spokenlanguage.

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 childrens morphological knowledge and theprogress that they make in learning to read? Fortunately, such evidencedoes 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 amorphological production task (see Box 1.3 for a description of thistask) and related it to a measure of their reading comprehension(Carlisle 1995). She found that the first-grade childrens scores in this


morphological task were quite strongly related to the level of theirreading comprehension a year later when they were in the second grade.Anne Fowler and Isabelle Liberman, two U.S. psychologists, also foundan impressive relationship between childrens success in a morpho-logical production task and their reading levels (Fowler and Liberman1995). From Denmark, Carsten Elbro reported a high correlationbetween the number of mistakes that children made with inflections ina Danish version of Jean Berkos task and the mistakes that they madein reading inflected words in a written text (Elbro 1989). In France,Sverine Casalis and Marie-France Louis-Alexandre also found thatkindergarten childrens scores in a variety of morphological productiontasks predicted their progress in reading two years later at school(Casalis and Louis-Alexandre 2000). One interesting aspect of this laststudy was that the morphology tasks dealt both with inflectional andderivational morphemes and the scores with the inflectional items dida much better job of predicting reading than the scores with thederivational problems. This pattern of relations would almost certainlybe different in the case of older children. The inflectional system is farless varied than the derivational system, and young children are morelikely to understand and use their knowledge of inflections whenreading than their knowledge of derivations.

We could go on, but we think that we have said enough to make thepoint that a relationship between morphological knowledge and literacydoes exist. Now we can consider the question of the direction of causeand effect in this relationship.

Some of the evidence on this question comes from a study that weourselves carried out several years ago (Nunes et al. 1997a, 1997b).This was a longitudinal study in which we looked at the same childrensspellings over a 3-year period. We were interested in childrens spellingof inflections, in particular of the past tense -ed inflection, and westudied how their spelling of this morpheme changed as they grew olderand how these changes were related to their knowledge of morphemes.

Let us begin with the first question: How does childrens spelling ofthe past-tense morpheme change over time? The childrens ages at thebeginning of the study ranged from 6 to 9 years. So, by the end of theproject, the youngest children were 9 years old and the oldest were 12years old. We found that during this period their spelling of the past-tense inflections changed radically. The very youngest childrens spellingof past verbs, as of other words, was often quite unsystematic. However,we found that as soon as their spelling of past-tense endings becameconsistent, it invariably followed the same pattern. These children began


by spelling the ending phonetically and, therefore, incorrectly. Figure1.1 illustrates this phonetic spelling, using a childs free writing of astory. Pikt for picked, opund for opened and suckt for suckedare mistakes familiar to the point of banality to anyone who works withyoung schoolchildren, but they are no less important for that. Thesemistakes clearly show children obeying one kind of spelling principleand ignoring another. Their use of phonologically based spellingprinciples is ingenious but too pervasive. Their complete disregard formorphemically based spelling principles is obvious. The -ed spellingtransgresses phonological correspondences, and so they ignore it.

Later on this changes. It is hard to assign a particular age to thischange, for it varies so much between children, but usually, according


Figure 1.1 The first two pages of a 712-year-old girls story.

to our results, children begin to put the -ed ending on past verbs sometime between the ages of 7 and 8. At first they do so with some regularpast verbs and not with others. However, the really interesting thingabout their initial sporadic use of the -ed ending is that they often alsoput it at the end of quite inappropriate words, as well as on regular pastverbs, where it belongs. Figure 1.2 gives an example of how a verytypical 712-year-old boy spelled a set of words for us, many of whichended in /t/ or in /d/. Some of these words were past verbs such assold, slept and told, but others, like next, were not. Notice thatthis 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 projectwe worked with over 350 children, and the majority of them made suchmistakes at some time during the study.


What is going on here? Our interpretation is that at this stage childrenare still treating the -ed sequence as some kind of a lettersound rule.They pick up the idea that -ed is another way of representing thesound /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.


Figure 1.2 Overgeneralizations of the -ed ending by a 712-year-old boy.

It is as though they have to learn about the actual spelling and have topractice 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 awhile, but later on they do learn to confine the -ed spelling to pastverbs. For some time they continue to use the ending with irregular pastverbs as well as regular ones, writing slept as sleped for instance,but at least this is a grammatically appropriate kind of mistake in a waythat necsed is not. They even manage quite well to use the -edending with entirely unfamiliar past-tense pseudoverbs (Yesterday heprelled his car) and not with other pseudowords (There is a preld atthe end of the road). Thus they eventually learn a genuine morphemicspelling principle: That -ed is the correct spelling for the regular past-tense ending.

This brings us to the second question. How is this developmentrelated to childrens morphemic knowledge? In our study we used threemain measures of this knowledge. The first was the sentence analogytask that we described earlier in the chapter. In this the child heard asentence followed by a transformed version of the same sentence(present to past verb, or vice versa); then the child heard anothersentence, rather similar to the first one, and had to transform thissentence 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 thesentence 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 Berkos morphological study. We calledit the productive morphology task. We gave the children pictures andwe used pseudowords as part of our description of what was going onin the pictures. So, for example, one picture showed a man performingan unusual action and a little story, which the child had to complete byusing 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 issnigging onto his chair. He did the same thing yesterday. What did hedo yesterday? He, . . . and the child was encouraged to produce thepseudoverb in the past tense.

Our project was a longitudinal one, which means that we saw andtested 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


to other measures over time. Time is important in analyzing theserelationships. If, for example, a childs morphemic knowledge doesdetermine how well she or he learns morphemic spelling principles, agood measure of different childrens morphemic knowledge taken earlyon in the project should predict how well children will learn thesespelling principles later on. If A determines B, A should precede B, and,therefore, the strength of A at one time should predict the strength ofB later on.

In fact, if you are examining how much one variable determinesanother over time, you have to take one extra step to be sure yourhypothesis is right. Suppose, for example, that you want to see if thestrength of childrens morphemic knowledge in one session (Session 1)has an effect on how much they have learned about morphemic spellingrules in a later session a year or so on (Session 2). Your hypothesis isreally about the changes in the childrens learning of morphemic rulesbetween the two sessions and not about how much they had learnedabout these rules at the beginning of the project. So you have to ruleout the effect of their earlier knowledge of these spelling rules. The wayto do this is to control for differences among the children in how wellthey knew the morphemic spelling rules in Session 1 before you examinethe relationship between their morphemic knowledge in Session 1 andtheir use of the morphemic spelling rules in Session 2. This statisticalmaneuver, which is called autoregression, may sound a complicatedone. In fact, it is quite easy to do.

We used this way of analyzing relations between our differentmeasures over time in our study. First, we looked at the relationshipbetween our measures of morphemic knowledge at the beginning ofthe project and the childrens success in spelling at the beginning of theproject and also 18 months later. We found that the childrens scoresfor morphemic knowledge in the first session predicted their success in spelling the past-tense inflection 18 months later, even after we hadcontrolled for their spelling prowess in the first session. We concludedthat this was strong evidence that morphemic knowledge plays a rolein how well children learn about morphemic spelling rules. This is not a surprising discovery, but it is an important one because of itsimplications for teaching spelling, which of course is the subject of thisbook. If morphemic knowledge partly determines how well childrenlearn morphemic spelling principles, one should take seriously thepossibility that steps should be taken to increase childrens explicitawareness of the morphemic structure of the words that they speak andhear and read and write.


The discovery that A affects B does not rule out the possibility that Balso affects A. Taking regular exercise may make people happier andmore relaxed than before, but being happy and relaxed may makepeople more inclined to take exercise rather than mope around at home.So, morphemic knowledge probably does affect how well children learnabout morphemic spelling rules, but it is also possible that intensiveexperience with morphemic spelling rules could increase childrensawareness of how words are constructed from morphemes. Here wemight find a two-way street.

Our data suggest that this too is true. We also looked at the rela-tionship between childrens 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 theretoo, even after we had controlled for differences among the children intheir morphemic knowledge in the first session. So, our results suggesta strong and thriving two-way relationship between these differentaspects of childrens linguistic knowledge.

Our confidence in the existence of this two-way connection wasstrengthened by a similar study of a very different language and script.Iris Levin, Dorit Ravid, and Sharon Rapaport worked with 5- to 6-year-old Israeli children, who were just starting to read and write Hebrew(Levin et al. 1999). In Hebrew the morphological system is rich andcomplex, and its effect on Hebrew spelling is at least as pervasive andimportant as the effect of English morphology on English spelling. So,it is useful to see if the relationships between childrens morphologicalknowledge and their literacy skills are much the same in this languageas 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 childrens knowledge of morphemes inspoken Hebrew by asking them to transform words morphemically. Anexample (translated from Hebrew to English) from one of their tasks is A baby who looks like an angel is an ________ baby. Here the childhas 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 addthe appropriate derivational suffix. The research team also measured the childrens progress in writing Hebrew at the same time. In Hebrew,as in English, children tend to concentrate on the phonological prin-ciples (graphemephoneme correspondences) before they adopt morecomplex correspondences such as the correspondences betweenmorphemic units and spelling. In this project the measures of childrens


progress in writing Hebrew charted the extent to which they hadprogressed from using basic phonological principles to the more difficultprinciples based on morphemes.

This project clearly established a two-way street, to use theresearchers own term, which we have already borrowed. The Israelichildrens knowledge of morphemes at the beginning of the projectpredicted their level of writing at the end of the project, even aftercontrols for initial differences between the children in their writingskills. There were also strong relationships in the opposite direction:The childrens level of writing at the beginning of the project predictedtheir knowledge about morphemes in spoken Hebrew at the end of theproject, even after controls for differences between the children at thestart of the project in their knowledge about Hebrew morphemes. Thisimpressive set of results establishes that in this language, too, childrenssensitivity to the way in which words are constructed from morphemesand the progress that the children make in literacy interact andstrengthen each other. It is a relationship of the greatest importance inEnglish and in Hebrew, and, almost certainly, in many other languagesas well, and it needs to be nurtured.

Teaching morphology: Improving spelling

The mention of nurturing brings us to our final question in thischapter, which is about how to nurture childrens understanding and use of the valuable morphemic spelling rules. The evidence thatwe have been reviewing suggests very strongly that one good way ofhelping children to learn about morphemic spelling principles would beto bolster their morphological awareness. Yet, tests of this simple ideaare remarkably thin on the ground.

This gap really is surprising because any study in which theresearchers manage to improve childrens morphological awareness andthen go on to examine the effect of doing so on the childrens learningof the correspondence between morphemes and spelling could yieldtwo most valuable insights. The first insight would be into the causalrelationship between these two. If the study establishes that teachingmorphological awareness leads to an improvement in spelling, it willhave provided the strongest evidence possible for the causal hypothesisthat we have been considering. But the second insight is even moreimportant than that, and it is the subject of this book. A successfulintervention study like this would have immediate educational sig-nificance. It would establish how possible and practicable it is to teach


children about morphemes, and it would show us whether this sort ofinstruction does have beneficial effects on childrens spelling and,perhaps, on their vocabulary as well.

Most of the rest of this book (Chapters 36) will be about a series of intervention studies that we carried out on the effects of raisingchildrens morphological awareness. The results of a previous project ofours on the effects of teaching children about morphemes (Nunes et al.2003) encouraged us to embark on this new program of research. In thisstudy, 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 anidentical posttest soon after the intervention was finished, in which wetested their ability to spell certain affixes, such as -ion and -ment,which normally cause children of this age a great deal of difficulty, andalso to follow particular phonologically based spelling principles, suchas how to represent short and long vowels.

In the intervention itself, we taught the children in small groups intwelve different sessions. We taught some groups about morphologicaldistinctions and others about phonological ones, and we also includeda 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, thechildren blended either morphemes or phonological segments, madeanalogies either about morphemes or about sounds, and classified wordsinto groups that shared the same morphemes or shared the samesounds.

Our morphological teaching did have a powerful effect, particularlyon the childrens success in spelling affixes. The study established, wethink for the first time, that it is possible to teach children aboutmorphemes and that this teaching has a direct effect on their knowledgeand use of morphemic spelling principles. The way was clear for us tobegin the program of studies that started in the laboratory and endedin the classroom. These are the studies that we shall tell you about inChapters 36.


Summary and conclusions

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

continued


1. Some of the most important links between spoken and writtenlanguage are at the level of the morpheme. The morphemicstructure of words in English and several other written languagesoften determines their spelling.

2. The system of morphemes, therefore, is a powerful resource forthose learning to read: The more schoolchildren know aboutmorphemes, the more likely it is that they will learn aboutspelling principles based on morphemes.

3. However, childrens knowledge of morphemes is largely implicit.It is quite likely that they need explicit knowledge aboutmorphemes in order to learn about the connection betweenmorphemes and spelling. Yet, many quite simple morphemicspelling principles are not taught at school. We need to knowhow easy it is to teach these principles explicitly and howeffective this teaching will be.

Chapter 2

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

Our book has two aims. Its first is to persuade our readers thatmorphemes are extremely important for children learning to readand write. Our second aim is to describe a set of studies that wecarried out on teaching children about morphemes and their relationto written words. The first two chapters in the book are all about thefirst of these two aims. In the remaining chapters we will try to fulfillour second aim by describing our work on teaching children aboutmorphemes 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 existenceof a strong relationship between childrens knowledge of themorphemic structure of spoken words and the progress that theymake in learning about written words. In this second chapter we shall look at a series of studies on childrens actual spellings andexamine what they tell us about their morphemic knowledge and about the way that they are using this knowledge in theirwriting. So this chapter focuses on what we can find out aboutpeoples knowledge of morphemes if we treat spelling as a win-dow on their knowledge of morpheme

Documents

Improving Literacy by Teaching Morphemes (Improving Learning)