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
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
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
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
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
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
1.1 A crash course in roots and stems (and bases) 51.2 A crash course in affixes 5
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
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
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).
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!
What is the issue?
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
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.
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
6 What is the issue?
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
Morphemes and literacy: A starting point 7
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,
8 What is the issue?
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.
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.
Morphemes and literacy: A starting point 9
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.
10 What is the issue?
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.
Morphemes and literacy: A starting point 11
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.
12 What is the issue?
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?
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 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,
14 What is the issue?
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
Morphemes and literacy: A starting point 15
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
16 What is the issue?
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.)
Morphemes and literacy: A starting point 17
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
18 What is the issue?
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
Morphemes and literacy: A starting point 19
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
20 What is the issue?
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
Morphemes and literacy: A starting point 21
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.
22 What is the issue?
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 and literacy: A starting point 23
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
24 What is the issue?
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
Morphemes and literacy: A starting point 25
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
26 What is the issue?
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.
Morphemes and literacy: A starting point 27
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.
28 What is the issue?
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
Morphemes and literacy: A starting point 29
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.
30 What is the issue?
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
Morphemes and literacy: A starting point 31
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
32 What is the issue?
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.
Morphemes and literacy: A starting point 33
Summary and conclusions
Our review of the ways in which morphemes and written languageare connected has led us to three simple conclusions.
34 What is the issue?
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.
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 morphemes. We shall show howchildrens spellings tell us a great deal about their knowledge ofmorphemes.
In the studies that we shall describe we used three differenttechniques to detect how people use morphemes in spelling. Inthese tasks we ask children:
Authored by Terezinha Nunes, Peter Bryant, Ursula Pretzlik, Deborah Evans, Daniel Bell, and Jenny Olsson
Spelling suffixes: Is it easy because theyhave a fixed form?
Some years ago, we started investigating childrens use of morphemesin spelling. In our first study we focused on only a few morphemes. Ouraim was to investigate how children spelled three different suffixes thathad a fixed spelling and a clear function: -ion, -ness, and -ed.
36 What is the issue?
1. to spell words that contain a morpheme whose spelling cannotbe entirely predicted from the way that it sounds but can bepredicted on the basis of how that morpheme is spelled: Forexample, the past-tense ending;
2. to spell pseudowords which contain particular morphemes: We create invented words using actual stems and affixes, placethem 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 explaintheir choice of spelling.
We carried out many different studies using these techniques. Theresults of these studies tell us a great deal about what children (andadults) know about morphemes without much explicit teaching,because there is currently little teaching about English morphemesin English schools.
This chapter presents a summary of what we found out. Thesections are organized by the target spellings and the aims of eachof the studies. The first section focuses on suffixes that have a fixedspelling which is not completely predictable from oral language.The second section focuses on the spelling of stems. The third sectionfocuses on how children and adults spell pseudowords made withreal stems and suffixes and the explanations that they give for theirchoice of spellings. The final section presents an overview of theresults and raises questions about the possibility of improving chil-drens knowledge of morphemes through teaching. Although thefocus of the chapter is on spelling, the aim of our investigations is tounderstand the connection between knowledge of morphemes andliteracy in a broader and a better way.
The derivational suffixes -ion and -ness are used to form abstractnouns. Neither can be spelled on the basis of their sounds: In neither isthe 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 waythat it sounds in different words. Sometimes this -ed ending representsthe sound /t/, as in kissed, sometimes /d/, as in killed, and some-times /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 wordsendings sound. So we thought that we needed to ask two questionsabout childrens use of these suffixes. First, we wanted to know whetherchildren understand that there are spellings for the ends of words thatoften dont accurately represent their sounds from the point of view ofthe traditional correspondences between letters and sounds. Second, wewanted to know whether children have a good grasp of when to usethese spellings.
Childrens awareness of suffix spellings that do notcorrespond to ending sounds
In one study we asked 710 children from a total of eight different schoolsin London and in Oxford to spell, among other words, four abstractnouns 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 inthe context of sentences, in order to make absolutely clear the meaningof each word that we asked the children to spell.
In Chapter 1, we gave some examples of childrens 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 werepresented in the context of sentences to help the children identify theirfunction. For example, the pseudoverb nelled was presented in thesentence We usually nell in the morning but yesterday we nelled in theafternoon. This sentence makes it clear that nelled ought to be treatedas a regular past verb. All the words in the sentences were written on apage, except for the target one, which we said in the context of the
Knowledge shown in way children and adults spell 37
sentence, and then repeated once so that the child could write it on aline that marked its place in the sentence. The children were in Years 3to 6 of primary school and their ages ranged from 7 to 10 years. Thenumber of children in each year group and their mean age is presentedin Table 2.1.
Figure 2.1 presents the proportion of correct uses of each of thesuffixes for each year group. We scored the spelling as correct wheneverthe children used the exact spelling for the suffix, irrespective of theway that they spelled the rest of the word.
The figure shows that in spite of the perfect predictability of thespelling of each of these morphemes, children do not seem to find it
38 What is the issue?
Childrens spellings of -ness and -ion byyear group in school
The correct suffix spelling -ness was used by 49 percent of thechildren in Year 3, 40 percent in Year 4, 68 percent in Year 5, and80 percent in Year 6. The second most used spelling was -nes,which was used by 15 percent of the children in Year 3, 25 percentin Year 4, 11 percent in Year 5 and 6 percent in Year 6. Spellings withother vowels comprised 21 percent of the spellings in Year 3, 13percent in Year 4, and 3 percent in Years 5 and 6. Other endingsobserved include -ners, -nace, -nece, -ns, -nerse. Thetotal number of different spellings for the -ness ending was twenty-two.
The correct suffix spelling -ion was used by 23 percent of thechildren 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 wastwenty-five.
easy to spell them. Only about one-third of the spellings of -ion werecorrect when the children are in their fourth year in school. Evenchildren in their sixth year at school, almost at the age of 11, do notmanage to spell any of these suffixes perfectly.
It is worth pointing out that the children are unlikely to be spellingthese suffixes simply on the basis of a specific learning of each of thewords. The fact that they spell the suffix -ed at the end of regular pastpseudoverbs as well as at the end of real past verbs suggests that theyhave learned more than specific spellings of well-practiced words. Ourhypothesis is that this generalization from real to pseudoverbs suggeststhat the children might have some knowledge of the representation ofmorphemes in spelling.
This first study led us to conclude that spelling even those morphemesthat have a fixed form is not easy for children. The fact that the spellingis not a good representation of the way that the words sound does
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 Mean age Number of children(years:months) (N=710)
Year 3 7:9 122Year 4 8:10 124Year 5 9:9 229Year 6 10:9 235
7:9 years 8:10 years 9:9 years 10:9 years
Age in years:months
"-ed" in realverbs
Figure 2.1 Percentage of children who spelled each suffix (-ion,-ness, and -ed) correctly, by age level.
interfere with the childrens spelling. This was demonstrated by ananalysis of the errors the children made: In the word emotion, forexample, whose final vowel is not pronounced clearly, about 98 percentof the errors were due to the use of an incorrect vowel or to missing outthe vowel altogether, and only 2 percent of the spellings failed to includethe letter n (although sometimes n was not the last letter in theword and it was followed by an e or a g).
Do children know when to use and when not to usethe different suffixes?
The case of -ion and -ian
In the study just described, we considered only the use of suffixes in theirappropriate place. We also carried out two other investigations thatanalyzed whether the children restricted the use of these suffixes to theright type of word. Consider, for example, the two words emotion andmagician. Their end sounds are the same. However, emotion is anabstract noun, and thus it is spelled with -ion, whereas magician isa person who does something (magic) and thus it is spelled with -ian.We had found in the previous study that about 70 percent of thespellings of children aged 11 correctly represented the suffix -ion atthe end of the abstract nouns we asked them to write. Would they beable to discriminate between the use of the two suffixes, -ion and -ian, although the words that contain these suffixes sound the sameat the end? If they are able to discriminate between the two types ofwords, we can conclude that they have some insight into the way thatthese suffixes are used to represent different meanings.
In another study, we asked 176 children from three schools in theOxford area to spell eight real words ending in -ion, another eightreal words ending in -ian, four pseudowords ending in -ion and fourending in -ian. As in the previous study, the words and pseudowordswere presented in sentences, which were written on the page. A gap inthe sentence marked the place where the word (or pseudoword) wasto be written. The children in this study were in Years 4 and 5 in primaryschool. Their mean ages were 8 years 8 months and 9 years 9 months,respectively. We also asked the children to spell eight other words andfour pseudowords with completely different endings, so that the wordsin the list would not all have the same sound at the end. A final noteabout this study: This study took place after the British Governmenthad introduced the National Literacy Strategy (NLS), whereas the first
40 What is the issue?
study that we described in this chapter was carried out before this policyhad been implemented. The NLS includes in the teaching of spellinglists of words ending in ion to be taught to children in the year groupsthat participated in this study, so we can assume that the children weretaught about ion endings.
Figure 2.2 shows the percentage of correctly spelled suffixes for wordsending in -ion and -ian.
Three results from this study will be stressed here. First, the obser-vations from the previous study are replicated in this study: A largeproportion of the spellings of the suffix -ion is correct at about age 10,but the level of success is not close to 100 percenta level of successthat could be achieved if the children were using knowledge of mor-phemes that makes this spelling predictable. Second, the children aremuch less successful in spelling correctly the suffix -ian: Their level ofsuccess with these words is about half their level of success with theion suffix. Third, the level of correct spellings of -ion in pseudowordsis lower than that observed for words, but it is not much lower (approx-imately 605 percent). This suggests that the children are learningsomething more general about spelling than the specific memories ofhow to spell particular words. However, it is not clear from this analysiswhat the children are learning. Because they are explicitly taught aboutthe existence of -ion at the end of words, they could not only be using
Knowledge shown in way children and adults spell 41
8:8 years 9:9 years
Age in years:months
"-ion" in words
"-ian" in words
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 betweenthe two age levels, a significant difference between the words and pseudowords(p
this as an ending where it is appropriate but also where it is not. Theycould, for example, be using the -ion ending both at the end of abstractnouns and for the agentive ending, -ian, because these two endingssound exactly the same.
Thus it was necessary to investigate the types of errors that thechildren made in writing the endings of these words in order to under-stand better what it was that children were learning about the -ionending. Is it possible that children simply learn that -ion is a possiblespelling for certain end sounds but have no clue where it is the appro-priate spelling? Are they just as likely to use -ion for words that endin -ian as for those that genuinely should be spelled with -ion?
We then calculated the proportion of use of -ion and -ian spellingsfor each of the types of word and pseudoword. These results arepresented in Table 2.2. The sections of the table that have a graybackground are those where the wrong suffix was used; that is, the childused -ion where it should have been -ian or vice versa.
The comparisons of greater interest in this table are between the rightand the wrong use of the two spellings. The children are roughly threetimes more likely to use the -ion spelling with the right words (abstractnouns) than with the wrong ones (words about people). This is trueboth for words and for pseudowords. So we can conclude that thechildren are learning not only that -ion is a possible ending but alsosomething about its morphemic value. However, they do use the -ionending in about one-quarter of the words that should be spelled with
42 What is the issue?
Table 2.2 Proportion of use of -ion and -ian spellings for each of thetypes of word and pseudoword
Year group ending in ending in ending in ending in in school -ion -ian -ion -ian
Spelled with ion (%)Year 4 77 25 65 16Year 5 77 23 60 14
Spelled with ian (%)Year 4 0 34 3 26Year 5 7 44 7 30
Note: Two ANOVAs showed that the proportion of -ion and -ian spellings inthe correct place was significantly greater than the use of these spellings in thewrong words (p
-ian, and this shows that there is much room for learning more aboutthe use of -ion and -ian.
Children seem much less aware of the -ian ending. Overall, they useit far less than they do the -ion ending. However, when they do useit, they also use it more often with the right words (words about people)than with the wrong words (abstract nouns). In fact, they hardly everuse -ian in the wrong place. Children in Year 5 were about three timesas likely to use -ion for -ian words than to use -ian for ion words.
These analyses suggest that children have to learn at least two thingsabout suffixes. They need to learn that a particular ending is a possiblespelling, and they also need to learn when to use it. Year 4 and 5 childrenknow that -ion is a plausible end spelling better than they know that-ian is also one. However, they seem to be no better at knowing whento use -ion than when to use -ian.
Finally, we want to make the case that the phenomenon that we arediscussing is a general one. The study that we have just describedinvolved a relatively good sample size (N=176), but all the childrenwere in the Oxford area. We also have data from a very large sample of children (N=7,377) who participated in the longitudinal study ofthe children born in the county of Avon in 19901. The children weretested when they were in the age range 910. Because of the largesample size, it was only possible to include two words ending in -ion(emotion and election) and two ending in ian (magician andelectrician). The level of correct spellings for the words with -ion was74 percent and for the words with -ian was 19 percent. More thanhalf of the pupils (66 percent) spelled both -ion words correctly, butonly 11 percent spelled both -ian words correctly. Thus, the results forthis large sample confirm that children are more aware of and readierto use the -ion ending than the -ian one.
The error analysis in this study showed that the children were morelikely to use -ion in the right than in the wrong place. The endings of the words election and emotion were spelled correctly with the-ion ending by 73 percent and 75 percent of the children, respectively.The words electrician and magician were spelled incorrectly with-ion at the end by 53 percent and 30 percent of the children, respec-tively. Thus, the children were aware of the -ion ending and used itmore often in the right than in the wrong place. However, the rate ofoveruse of this ending is not small and neither is the exact number of children who used -ion in the wrong place. It is no small deal thatthat 3,938 children in the sample misspelled the ending of electrician,and 2,216 misspelled the ending of magician, by using -ion instead
Knowledge shown in way children and adults spell 43
of -ian. This should give teachers and schools cause for concern. In fact, more children used the -ion than the -ian ending for both ofthese words.
The error analysis in the large-scale study showed, like that in our previous studies, that the children used the -ian ending more oftenin the right than in the wrong words. The word magician was spelledwith the -ian ending by 25 percent of the children; for the wordelectrician, 12 percent of the children used the correct -ian spelling.This is not an impressive rate of success for 10-year-olds. Nevertheless,the children did use the -ian ending more often in the right thanin the wrong words. The words emotion and election were spelledwith the -ian ending by less than 1 percent of the children (0.33percent and 0.14 percent, respectively). Electrician proved to be verydifficult for the children. The spelling most often used was incorrect:1,812 children spelled electricion and electrition, and only 785spelled it correctly. The children were slightly more successful with theword magician: The correct spelling was the one used most oftenit was used by 1,659 children. However, magition was used by 780children, about 11 percent. These incorrect spellings were, perhaps,used by making a wrong analogy to words like emotion and election,ignoring the meaning category and focusing on the similarity of the endsounds.
The results of this large survey support our initial idea: Children needto become aware of the possibility of using a letter string at the end ofthe words, but this does not ensure that they will use it in the right butnot in the wrong type of words. The second type of progress, learningwhen to use and when not to use a letter string, seems to be attainedlater, when they understand the role of suffixes in spelling. This learningcan be described in more general terms by saying that they have to learnthat some forms are possiblethat is, they have to learn that -ion and-ian endings are possible endingsbut they also have to learn theirfunction. Children who make the wrong analogy between the spellingof electrician and emotion and use the -ion letter string, spellingelectrition for electrician, have learned a form but not its function.They seem to be using this ending as a possible phonological represen-tation, missing the point of morphemes, which are units of meaning.
The case of -ed
Let us now go back to the longitudinal study that we mentioned inChapter 1 and consider how well the children learned about the -ed
44 What is the issue?
past-tense ending. Children get better at using the -ed at the end ofregular past verbs, but do they know when to use this spelling, ratherthan a phonological spelling, for words that end in /t/ or /d/ sounds?
In the study we asked 365 children in four schools in Oxford and fourin London to spell a number of words, which included ten regular verbsin the past (for example, kissed, killed, opened), ten irregularverbs in the past (for example, lost, sent, kept, slept), and tenwords that were not verbs and ended in /t/ or /d/ sounds (for example,soft, except, field, ground). The study took place before theintroduction of the NLS and so there was no explicit guidance at the timefor teachers to practice the use of the suffix -ed with their pupils. Themethod for the spelling test was the same described earlier on: Thewords were presented in a sentence to ensure that the children knewwhich word we wanted them to spell; they were said in the context ofthe sentence, and then repeated so that the children could write them.The children were in five age groups, shown in Figure 2.3, which alsosummarizes their performance.
On the left, the figure shows the mean correct spelling of the endingswith t, d or -edobtained by the children. At all age levels, thechildren were considerably better at using the endings t and d, whichare the right way of spelling the irregular past verbs and the nonverbs,
Knowledge shown in way children and adults spell 45
Figure 2.3 On the left: Number of correct spellings of regular and irregularverbs in the past and nonverbs ending in /t/ or /d/. On theright: Generalization of -ed to the wrong words.
to irregular verbs
nonverbs irregular verbs
(a) Correct endings (out of 10) (b) Proportion of errors using -ed in the wrong place
than at using the -ed ending, which is the correct spelling for theregular past verbs. This indicates that they were better able to use theirknowledge of sounds to guide their spelling of the ends of words thanto use their knowledge of morphemes. Although this is a differentsample of children from those in the previous studies, the similarity inthe rates of success for the age groups that are comparable across studiesis remarkable: At the age of about 10, the children show a high level ofsuccess, but there is considerable room for further learning.
The graph on the right of Figure 2.3 shows how many errors madewith irregular past verbs and nonverbs were due to the children using -ed at the end of these words. These are the mistakes that we referred to as overgeneralization errors in Chapter 1. Instead ofspelling the endings with the letters t or d, which would have beencorrect, the children spelled these endings with ed: For example, they wrote sofed for soft and helded for held. Similarly to whatwas observed with -ion, the children used the spelling -ed correctlybut also used it in words where it was not appropriate: They generalizedit to irregular past verbs and even to other words which were not verbs. They had learned that the form -ed is a possible ending butseemed to attribute to it a phonological function: Words that end in /t/or /d/ sounds could be spelled with -ed. They had not grasped itsmorphological function.
By now, we can attempt to answer the questions posed at the start ofthis section.
Is it easy to spell suffixes because they have a fixed form? The answerseems to be that it is harder than one would expect. If a suffix cannotbe spelled on the basis of its soundsfor example, because it containsa schwa vowel, which is not pronounced clearlythen it takes childrensome time to spell the suffix correctly. On average, children in theirfourth and fifth year in school, who are aged 9 and 10, still do not showclose to 100 percent correct responses when spelling suffixes such as -ness, -ion, -ian and -ed.
In the case of -ion, -ian and -ed, used at the end of words thatsound like other words which are spelled differently, the children needto become aware of the suffixes as possible endings and also to learnwhen to use these spellings at the end of words. Many children use thesesuffixes in the right as well as the wrong words. This generalization tothe wrong type of words suggests that perhaps they are learning
46 What is the issue?
something about the existence of spelling patternsthat is, sequencesof letters that appear together at the end of wordswithout necessarilygrasping that these sequences of letters have meaning and are used toconvey these meanings in the written language.
It is difficult to end this section without wondering why it takeschildren so many years in school to learn to spell suffixes, consideringthat the spelling of suffixes is so predictable. Is it possible that childrenare not such good morphemists after all and that the studies about howgood they are in oral language overestimate their ability? Or could therebe two different levels of knowledge of morphemes, one used in oral andthe other in written language?
Different types of linguistic knowledge
In the first chapter we discussed the possibility that knowing how to usean aspect of a languagesay, the different types of past tenseisdifferent from explaining how the same aspect of language is used. Thetype of knowledge of language that is used for speaking is implicit.When we are doing something, like speaking, we dont really thinkabout how language works. Tasks that we perform without having toexplain why we do things the way that we do are referred to as onlinetasks.1 To explain what we do would be considered as an offline task:we need to represent what we do in order to explain it.
In this sense, spelling is different from speaking because lettersrepresent the spoken language. In its relation to oral language, spellingcan be seen as an offline task. This is one reason why children may bebetter morphemists when speaking than when spelling.
But there is another reason: Some layers of language construction are not manifested in oral language. Children may be good mor-phemists, and may be able to inflect words in oral language, but stillmight not know enough about the morphemic elements that are notpronouncedand the -ed, as we have seen, is not pronounced; it isonly written.
In the previous chapter, we described a technique developed by Jean Berko that is considered to be a good assessment of childrensknowledge of morphemes in online oral tasks. The children are asked
Knowledge shown in way children and adults spell 47
1 Annette Karmiloff-Smith (1992) discusses the distinction between online and offlinetasks and the relation between these types of tasks and the level of awareness required for performance. It is beyond the scope of this book to present the issue at length.
to complete sentences containing pseudowords. In order for thesentences to be completed correctly, the children must apply amorphemic transformation to the pseudoword; sometimes they arerequired to inflect verbs and nouns (for example, say the verbs in thepast or the nouns in the plural) and sometimes they are required to usea derivational suffix (for example, make an agentive out of a verb).Because the transformations have to be applied to pseudowords, thechildren cannot use specific memories of the items; they have togenerate the new item themselves. Thus, this is an online task, but theknowledge it draws on is likely to be a more explicit form of knowledgeof morphemes than the use of inflections in familiar words. In somesense, this is more a language game than speaking. But it is not anunrealistic game: Children will often encounter unfamiliar words intheir base form and will want to use them in their inflected form later,without having heard the inflected form.
Our idea that there are different forms of knowledge of morphemes,which can involve more or less awareness of the morphemes, led us tocompare performance on these two tasks. Spelling is an offline task inits relation to morphemes: It requires that we represent the morphemeswith letters. Inflecting spoken pseudowords is an online task: We canuse the morphemes without having to think about them. If this is correct,then children should show better performance in the pseudowordinflection tasks than in spelling.
In another of our studies we gave 6-, 7-, and 8-year-old children oralpseudoword tasks along with the spelling tasks described earlier. Insome of the items, they were asked to complete sentences that containeda pseudoverb; the pseudoverb needed to be in the past for the sentenceto be completed correctly. In these oral tasks we showed the children apicture, gave them clues from spoken sentences, and asked them to usethese clues in order to complete the sentence. They had to come up withthe past for the pseudoverb. This task was described in Box 1.3, with anexample of an item that required the children to inflect the pseudoverbto snig.
Our task contained a variety of pseudowords, three of which wereverbs. We calculated the proportion of correct answers to thepseudoverb items in the oral task and the proportion of correct spellingsof the endings of real past regular verbs in the spelling task. Figure 2.4shows the comparison between their ability to produce the inflection inthe oral pseudoword task and their ability to use the -ed when spellingthe end of regular past verbs. The results are quite clear: Children arebetter morphemists in oral than in written language. The same children
48 What is the issue?
who found it so difficult to spell the ends of verbs using -ed had asignificantly higher level of success in producing the pseudoverbs in thepast correctly.
So why could they not spell with the -ed suffix correctly? We thinkthat the answer is in the discrepancy between oral and written language:Some of the information about morphemes is either not present or notclear in oral language.
In Chapter 1, we argued that there are many cases where a correctspelling cannot be produced on the basis of sounds: Knowledge ofmorphemes is what allows us to master these spellings. This is exactlythe case with the morphemes considered here: Information from orallanguage is insufficient for correct spelling. We think that in order tomaster these spellings the children have to develop some awareness of the morphemes that compose these words and try to represent themorphemes and not just the sounds. In the case of the past tense, orallanguage does not represent the -ed: Past regular verbs sound like /t/or /d/ or /id/ at the end. The children need information about writtenlanguage in order to realize what the morpheme is. It is not sufficientto know that some words in English sound like /t/ or /d/ at the end butare written with -ed. This increases the likelihood that the children will
Knowledge shown in way children and adults spell 49
6yrs 7yrs 8yrs
pseudoverbs correctly inflected in oral language
regular verbs correctly spelled
Figure 2.4 Proportion of past regular verb endings spelled correctly andproduced correctly for pseudowords in an oral task.
Note: An ANOVA with repeated measures produced significant main effects for agegroup and task and a significant age by task type interaction (all differencessignificant at the 0.001 level).
use the -ed ending in spelling, but it does not guarantee that they willdo so in the right place. If they grasp the morphemic significance of the-ed ending, they should be more likely to succeed.
Our ideas about morphemes and spelling are a hypothesis about theeffect of teaching. We think that becoming aware of morphemes shouldmake it easier for children to learn to use these suffixes in the right, andnot in the wrong, words. We also think that teaching can be an effectiveway of making children more aware of this aspect of language. This iswhy in later studies we looked at different ways in which children canbe taught about morphemes. These studies are presented in Chapter 3.Before we move to the analysis of teaching in Chapter 3, though, we stillwant to look through the spelling window to find out more about whatchildren and adults know about morphemes.
Spelling different words with the samestem: Do children conserve the stem?
In the first section of this chapter our attention was focused on howchildren spell suffixes to gain insight into what they know aboutmorphemes. We now turn our attention to word stems, which form thecore of the word. Words cannot be made only with a suffix: A suffix isalways attached to a stem. So, although suffixes are important, theycannot tell the whole story about our awareness of morphemes. It ispossible that children find it easier to realize that words with the samestem (for example, art and artist) bear some relation to each otherthan to realize that words ending in -ion are abstract nouns and thoseending in -ian are agentives.
The best approach to spelling a stem is arguably to analyze its soundsand to try to represent them. However, this is not the only thing weshould think about. Sometimes a stem is clearly pronounced when it isa word on its own, but it is not as clearly pronounced once a suffix isadded to it. Think of how we say the words magic and magician. Theword magic has two clearly pronounced vowels that are easy torepresent in spelling, and the last letter, c, represents the sound thatit always represents at the end of words. When we add the suffix -ianto it, the stress pattern changes and the letter a now represents aschwa vowel. (See Box 1.4 for an explanation about schwa vowels.)The letter c now stands for a sound most often spelled with sh. Thismeans that children who might spell the first vowel in the word magiccorrectly could well make mistakes in the same place when spelling theword magicianand they often do.
50 What is the issue?
The survey of childrens spelling in the county of Avon for Year 6children showed that the majority of the children (75 percent) wereable to spell the beginning of magician correctly, with ma-. But one-quarter of the children did not spell this beginning correctly. Thenumber of children is such that it should not be neglected: 565 childrenspelled the beginning of magician with mu-, 736 spelled it with mi-, and a small number (45) actually used no vowel between theinitial letter m and a subsequent consonant. So even 10-year-oldsmight not realize that they should conserve the stem when a suffix isadded to it. However, for those who realize that magician is composedwith the stem magic- and the suffix -ian, this spelling becomescompletely predictable, although the vowel is not clearly pronounced.
This sort of analysis led us to investigate how consistently childrenspell words that are different but have the same stem. Do they, on thewhole, conserve the stem across words? If they do, does this consistencyin spelling the same stem across words reveal anything about theirknowledge of morphemes?
The children in one of our studies spelled ten pairs of words with the same stem and ten wordpseudoword pairs that also had the samestem. All the words included in these pairs had stems with somethingthat was unpredictable either in both words or in one of them. Examplesof the pairs of words we used are magicmagician, knowknowledge, naughtynaughtiness and treasuretreasures. Thedifficulties in the word magician have already been pointed out. In thewords know and knowledge, there is a silent k and the stem soundsdifferently across the words. The words naughty and treasure alsocontain difficulties: Naughty has an unusual spelling for the first vowelsound and the letter s in treasure does not represent its mostcommon pronunciation. However, these spellings should be maintainedin the derived forms.
Previous research (Fowler and Liberman 1995) has shown thatchildren find it difficult to detect the common stem when it soundsdifferently across the base and the derived form, so we thought it wouldbe important to include in our task some pairs where the stem soundsthe same and other pairs where the stem sounds differently.
We thought it was also necessary to include pseudowords in this taskbecause childrens consistency in spelling stems across words couldresult from rote learning of each word. However, if the children spelleda word and a pseudoword that contained the same stem consistently,we would know that they could not be spelling the pseudoword on thebasis of a specific memory. The pseudowords that we created were
Knowledge shown in way children and adults spell 51
dinosaur names. The children were shown a picture of a dinosaur thatmade it very clear why there should be a connection between its nameand the stem we wanted them to think about. Figure 2.5 illustrates theitems used in this task.
In order to make sure that the children did not copy the stem for thesecond word after they had already written the first word, we dictatedone element from the pair on one day and the second element on adifferent day. So, if they spelled, say, the stem of magician in the sameway as they spelled magic, we would consider it likely that they hadthought of the word magic on their own rather than were promptedby seeing it on the page. Half of the children wrote half of the basicwords and half of the derived words on the first day and the other halfon the second day. The remaining children had the dictation in theopposite order.
For each of the word pairs, the children were given one point if thestems were spelled exactly in the same way, independently of whetherthey had been spelled correctly or not. We realize that this couldoverestimate the childrens awareness of the connection between thestems. For example, if a child spelled naughty as noty and also usednot for the stem of naughtiness, the consistency might result notfrom the childs realization that the stems are the same but from the useof a phonological strategy to spell both words. However, we felt that we
52 What is the issue?
(a) Knotosaurus (b) Combosaurus
Figure 2.5 Pictures of dinosaurs with their names, which the children wereasked to spell.
could not exclude from the consistency score those words that had beenspelled phonetically.
Figure 2.6 shows two graphs with results from this task. Graph (a)displays the results for a sample of about 320 children from Londonand Oxford schools. Graph (b) displays the results for the same Oxfordchildren approximately one year later; 169 children completed the taskson this second occasion.
Three results stand out in these graphs. First, as in the previousstudies, the childrens performance improves with age. This can be seen
Knowledge shown in way children and adults spell 53
6:6 years 7:7 years 8:10 years
Age in years:months
consistency in words consistency in pseudowords
7:4 years 8:4 years 9:4 years
Age in years:months
consistency in words consistency in pseudowords
Figure 2.6 Proportion of word and pseudoword pairs whose stems werespelled in the same way at each age level.
on Graph (a) where the results are for different children of differentages, and also from a comparison between the two graphs, when someof the same children were tested at a different age. Second, the perfor-mance of the 9-year-olds, like in the previous studies, can be consideredgood but not excellent. The majority of the spellings conserve the stemacross words but there is still much room for improvement. The thirdaspect of these results was unexpected: The children at all age levelsperformed significantly better on the consistency of stems when onestimulus was a pseudoword than when they were both words. Thedifference in performance between the tasks could be due to the fact thatthe dinosaur pictures gave such a strong clue to the stem that thechildren could not avoid making a connection between the derivedpseudoword and the basic word. In psychological experiments, we refer to using stimuli to provoke connections, even if implicitly, aspriming. In spite of the difference in level of performance, the twotasks do not seem to assess different aspects of the childrens knowledgeof morphemes: The correlation between them was 0.83, which is a high correlation (a perfect correlation has the value of 1). This highcorrelation would be unexpected if the processes assessed by thedinosaur task were different from those assessed by the task with realwords.
At first glance, the results do not seem to indicate that children findit easier to make a connection between stems than between suffixes:They do not seem to spell them consistently more often than they spell correctly the suffixes examined in the previous section. A directcomparison is only possible if we look at the results for a sample ofchildren from the different studies who were of comparable age levelswhen they did the spellings of suffixes and the consistency in spellingstems tasks. This sample comprises 172 children with a mean age of 8 years 2 months and 164 children with a mean age of 9 years 3 months.Figure 2.7 shows the mean results for the two age groups in the twotasks. We think that these results leave little room for doubt thatchildrens performance in tasks that involve spelling a fixed form,whether it is a stem or a suffix, is very similar. It does not seem to beeasier to conserve the spelling of a stem than to use a suffix consistently.However, it should be kept in mind that this is only one example of asuffix and that other suffixes may be easier or more difficult: It may beeasier to understand the connection between past verbs, for example,than the connection between abstract nouns.
So far, we have looked at childrens knowledge of suffixes and stemsas revealed by their spelling and their ability to inflect pseudowords. We
54 What is the issue?
have not asked them about why they spell words and pseudowords theway that they do. Asking people to explain why they do something the way that they do, we argued, assesses a different process: It requiresa greater level of awareness. But we have not provided any evidence for this idea. In the section that follows, we will consider differencesbetween spelling and explaining why pseudowords should be spelledin particular ways.
What childrens and adults spellings ofpseudowords tell us about their knowledgeof morphemes
We have, in previous work, asked children and adults to spell somewords and to explain why they spelled them the way that they did. It isoften difficult to obtain an explanation that goes further than Becausethats the way that this word is spelled or Thats how Ive seen theword spelled in the past. We have found that it is much easier to obtainexplanations when we ask people to explain why they spelled a pseudo-word in a particular way. When people spell pseudowords, they realizethat they cannot say This is the right way or Ive seen it spelled likethis in the past. They created the spelling, so they try to think why theyused that particular spelling. So, if we ask children and adults to spell
Knowledge shown in way children and adults spell 55
8:2 years 9:3 years
Age in years:months
use of -ed in regular past verbs
consistency in spelling stems
Figure 2.7 Proportion of real verb endings spelled correctly with -ed andproportion of stems spelled consistently across two words.
pseudowords and to justify the way that they spelled them, we can havetwo ways of assessing their knowledge of morphemes: The first by thecorrectness of their spelling and the second by analyzing what sort ofexplanations they give for their spellings. We expect that explaining ismore difficult than making correct choices, but also that there is aconnection between the ability to explain why the pseudowords shouldbe spelled in certain ways and the correctness of the spelling. So weused pseudoword spellings and explanations as another window forlooking at childrens awareness of morphemes.
Freyja Birgisdottir and her students Victoria Brierley and Peter Sherrelasked twenty undergraduate students (mean age 27 years 7 months),fifty-one children in their fourth year in primary school (mean age 8years 3 months) and fifty-six children in their sixth year in primaryschool (mean age 10 years 6 months) to spell thirty-five pseudowords.Unlike the pseudowords used in the dinosaur task, which had stemsfrom real words, the pseudowords used in this task did not draw onstems from real words: They were phonological sequences formed by changing the initial consonant of a word (for example, crill was formed using drill by substitution of the initial consonant), bychanging the initial syllable (for example, fompect was created bychanging the initial syllable in inspect) or by omitting a consonant in a word (for example, Marid was created by omitting the sound /d/in the middle of Madrid). Each of the pseudowords was presented ina sentence that made the grammatical status of the pseudoword clear,and, consequently, if the pseudoword had a suffix in it, what the suffixwas. The whole sentence was presented in writing; a gap designated theplace where the pseudoword should be written. The pseudoword wasfirst said in the linguistic context and then repeated, in order to help theparticipants remember what the pseudoword was.
All the items were chosen to reveal something about the knowledgeof morphemes used by the participants. They belonged to sevendifferent categories of words; five examples of each type were included.In order to illustrate the difference between spelling and explainingones choices for spelling pseudowords, we will consider here two cases,where the endings are pronounced in the same way but the spelling isdifferent. We will discuss each case separately.
Pseudowords with -ion or -ian
The first case has been discussed earlier in this chapter: It relates to the distinction between using -ion or -ian at the end of words.
56 What is the issue?
Two categories of pseudowords were used. As in the previous studies,the sentences were used to clarify what category the pseudowordbelonged to.
Figure 2.8 summarizes the findings for this study. These resultsreplicate our earlier findings. At about age 10, children perform well inspelling these pseudowords, but there is much room for improvement.As in the previous studies, the children were more successful in spellingthe pseudowords ending in -ion correctly than those ending in -ian.
They were also more likely to use -ion in -ian pseudowords thanthe other way around. In the group of children aged 8, 39 percent of thepseudowords that should have been spelled with -ian were writtenwith -ion, whereas the reverse mistake appeared in only 9 percent ofthe spellings. The 8-year-olds seemed aware of the existence of the -ion ending but did not seem to know when to use it and when not to use it: They spelled with -ion 42 percent of the pseudowords that should have ended in -ion but, as indicated earlier on, they alsoused this ending in 39 percent of the pseudowords that should end in-ian.
In the group of 10-year-olds, the percentage of spellings of -ianwords with -ion endings increased to 41 percent, and the use of
Knowledge shown in way children and adults spell 57
8-year-olds 10-year-olds adults
Figure 2.8 Percentage of correct pseudowords with -ion and -ianspelled correctly and percentage of correct explanations, byage level.
Note: An ANOVA showed that there were significant differences in performanceacross the age levels and word type, with -ion and -ian endings (p
-ian in -ion pseudowords also increased, reaching 19 percent. Thiswas due to the fact that the use of other spellings, different from -ionand -ian endings, decreased. This suggests that the children were moreaware of the existence of both suffixes and also more aware of when touse and when not to use these suffixes, because the difference betweenthe right and wrong use of each of the suffixes is larger than thatobserved for the 8-year-olds.
The performance of the adults shows much less difference betweenthe level of success across word type: Their spellings were correct for78 percent of the -ion pseudowords and for 76 percent of the -ianpseudowords. So, they seem equally aware of the existence of bothsuffixes. However, they were still more likely to use -ion spellings in-ian pseudowords: They spelled 31 percent of the -ian pseudowordswith -ion, and only 19 percent of the -ion words with -ian.
The explanations provided for the spellings were classified into threecategories. If there was a reference to at least part of the rule or anexplicit analogy with words of the appropriate category, the explana-tion was considered correct. This is, arguably, a lenient criterion butvery few responses would satisfy a strict criterion. Many explanationsreferred to the way that the word sounds: These were classified asphonological explanations and were, for all the cases considered here,insufficient. A third category was created with other types of expla-nations.
Figure 2.8 shows that none of the 8- and 10-year-old childrenprovided correct explanations for their choice of spelling in thesepseudowords. In contrast, about one-third of the explanations providedby the adults referred to the underlying rule in some way. For example,one participant wrote to justify an -ion spelling: a nounfrom rulein English: contagious contagion. To justify an -ian spelling, the same participant wrote: Noun with a in the ending. Library librarian; similar as example. Another participant spelled thepseudoword Maridian with -ian (which was used in the sentenceSomeone who comes from Marid is a ________) and wrote the followingexplanation: Peoples names, i.e. where they were from. This sameparticipant could not provide a correct explanation for -ion words andjustified the spelling on the basis of how it was pronounced.
Thus, a comparison between the level of correct spellings and correctexplanations confirms the idea that these tasks assess different types ofknowledge: The level of awareness required for explanations is certainlygreater than that required for spelling pseudowords.
58 What is the issue?
The use of s as a morpheme that marks plural andas a morpheme that marks the third person singularof verbs
The second morphological distinction included in this study was about pseudowords that end in the sound /ks/. Two types of wordswere included: One-morpheme and two-morpheme words. If a one-morpheme word ends in the sound /ks/, it is spelled with an x at the end, like box, fox, fix, and mix. An example of a singularpseudoword used is: There is only one spix in this recipe. Two sets oftwo-morpheme words were used in this study. One set was formed byplural nouns, which are spelled with -cks at the end: the s is amorpheme that marks the plural, added to a singular noun than endsin the /k/ sound. An example of this type of pseudoword used in thestudy is, My Grandma has macks in her kitchen. The second set of two-morpheme words used in the study was formed by verbs in thethird-person singular: The morpheme s, added to a stem that ends in the sound /k/, creates a pseudoword ending in the /ks/ sound. Anexample used in the study is, Mary smicks the trees every day. One-morpheme words that end with the sound /ks/ are spelled with an x;two-morpheme words ending in these sounds are spelled with an -sadded to the basic form (which could be spelled with -k, -ck or -keat the end). This is a very consistent spelling rule, but we have not seenit taught so far.
Previous studies have shown that the distinction between these casesis problematic for children. Peter Bryant and his students have observedin different studies that children use -x for plural words and also pluralendings for singular words. They do not seem to know when to use oneor the other type of spelling. However, the specific frequencies of errorsvaried across studies. It is quite possible that this variation depends onwhether the children consider -x as the appropriate phonologicalrepresentation for the /ks/ sound or whether they think that any wordending with an /s/ sound should have the letter -s at the end. If theythink that -x is the appropriate representation, they will spell one-morpheme words correctly and make more errors in two-morphemewords. If they think that an -s is needed when the words end with an/s/ sound, then they will make more mistakes in one-morpheme words.In either case, there should be a negative correlation between the scoresin the two types of words: The better you perform on one word type,the worse you perform on the other. Hopefully, if the children becomeaware of the morphological significance of these endings, the negative
Knowledge shown in way children and adults spell 59
correlation should disappear, and their spelling of all three word typeswould improve simultaneously.
Figure 2.9 displays the levels of performance observed in this study. As in previous studies, there is general improvement with age,but the pattern of improvement is not straightforward. This improve-ment is uneven because the level of performance across the types ofpseudowords is significantly different for the younger age group. The8-year-olds performed significantly worse in singular pseudowordsending in -x than in the other two word types. Although the 10-year-olds perform better than the 8-year-olds, their performance is only atchance level: There are only two phonologically appropriate spellingsfor these pseudoword endings, and they are right about 50 percent ofthe time for the singular and plural nouns. Their performance is slightlybetter on the verbs in the third-person singular. The adults spelling is above chance level but not close to 100 percent. They performedbetter with the distinction between -ion and -ian. These results weresurprising to us: We had thought that the distinction between singularand plural is such an easy one that its representation in spelling wouldbe mastered at an earlier age.
60 What is the issue?
8-year-olds 10-year-olds adults
Pseudowords ending in "-x"
Pseudowords in the plural
Pseudoverbs in thirdperson
Explanations, singular vs. plural
Figure 2.9 Percentage of correct spellings of one-morpheme and two-morpheme words, by age level.
Note: An ANOVA showed that there were significant differences in performanceacross the age levels and word type; the use of the s to mark the third-personsingular of pseudowords was easier than the spelling of the other pseudowords forthe children (p
When we analyzed the explanations provided for the spellings of the pseudowords, we found trends that were quite similar to what wedescribed before for the -ion and -ian pseudowords. The children didnot refer to grammar or to morphemes, nor did they use analogies toappropriate words justifying their choices of spelling. Approximatelyone-third of the explanations provided by the adults referred to the factthat the words were singular or plural. Some examples of explanationsconsidered to satisfy the criterion for correct responses are: Pix seemsto be a singular and This is more than one. Not all explanations givenby adults using grammar were correct. One participant distinguishedbetween her use of -x and -ks by saying that she used -x for a nounand -ks for a verb. This reference to grammar was not consideredcorrect. She then wrote pix, a singular noun, with -x, and spucks,a plural noun, with -ks. Next to this she wrote: Both nounsnotlogical! Her difficulty in identifying the different types of morphemerepresented by a final -s appeared to confirm the often-held view thatEnglish spelling is not logical.
If the participants were not thinking of morphemes to produce theirspelling, what were they thinking of? The majority of the explanationsreferred to how the words sound. The 8-year-olds justified their spellingsby referring to the way that the words sound 43 percent of the time. Thistype of explanation increased to 70 percent in the 10-year-old group butdecreased to 44 percent in the group of adults, who gave 36 percent ofexplanations using morphological rules or analogies to words thatrepresented these rules.
It is possible that the strong tendency to focus on phonology revealedin the explanations interferes with the awareness of rules that are basedon a different layer of language organization. The overall performancein pseudoword spelling shown by both children and adults is muchbetter than what one might expect from their explanations. However,it should not be forgotten that in none of the examples was the per-formance of any of the groups close to 100 percent. If they had had agreater awareness of the significance of morphology for spelling inEnglish, would they have been able to attain 100 percent?
Knowledge shown in way children and adults spell 61
Summary and conclusions
In this chapter, we have used spelling as a window to find out moreabout childrens and adults knowledge about morphemes. Our own
62 What is the issue?
studies involved more than 1,000 children at different age levels.The study of children in the county of Avon involved over 7,000children. These different studies on the spelling of morphemesconcur in showing that children make progress in the spelling ofsuffixes and the conservation of stems across words in primaryschool. However, performance does not reach a level of accuracythat we would expect good spellers to reach after five or six years ofschooling.
The conclusion that we need to teach children about morphemesseems inescapable. The question is what is the best way of doing this.Chapter 3 summarizes a set of studies that investigated this question.
What does the researchtell us?
From the laboratory to the classroom
Study 1: Teaching pairs of children aboutthe difference between -ion and -ianendings
Our first example is a project that we ourselves did on teaching childrenabout the distinction between -ion and -ian endings. This distinctionhad become a test case for us. We had identified a simple and usefulmorphemic principle that had, we thought, the potential to help childrenconquer an aspect of spelling that normally causes them much trouble.The principle is that when the word is a noun that ends with a schwavowel followed by an /n/ sound, the correct spelling for this ending is-ian if the noun refers to a person and -ion if it does not.
Normally, as you saw in Chapter 2, children get into a muddle withwords like election and magician. The spellings for the -ion and-ian endings bear little relationship to their sounds, and this is plainlythe cause of the young childrens problems with such words. Children
We turn now to teaching children about morphemes, and we shallfocus first on spelling. Our main aim in this chapter will be to showyou that it is quite possible to teach children about spelling principleswhich are based on morphemes and which no one would otherwisetell them about at school. We shall try to convince you that this sortof teaching is relatively easy to do, that it is great fun for children,and that it is highly effective: Children are able to learn morphemicspelling principles quickly, and by and large they remember whatthey learn long after they have been taught about it.
Authored by Peter Bryant, Terezinha Nunes, Ursula Pretzlik, Daniel Bell, Deborah Evans, and Jenny Olsson
who write words like elecshon or elexon (about 28 percent of the9-year-olds in the study in Avon, which we mentioned in Chapter 2,produced these or similar spellings) and magishon or magishen (11 percent of the same children spelled magician with sh followedby a single vowel followed by n and the most often used vowels wereo and e) are defeated by the clash between the lettersound corre-spondence rules, which they have learned so well, and by these twodifferent ways of spelling the same sound (-ion and -ian), both ofwhich flout the lettersound correspondence rules.
The solution, we argue, is not to rely on asking children to learn the spelling of words like election by rote, which is the normal practice, but to teach them that there is a principlea genuine spellingprinciplefor such words. In this case, the regularity is at the level ofmorphemes, not at the level of sounds or sequences of sounds.
We have no doubts, no hesitations at all, about this approach. If thereis a spelling principle that could remove a stubborn difficulty, then oneshould teach this principle to children to see if it helps them to spell. Theonly remaining question is how best to teach it.
When we considered this last question, we realized that there werethree main candidates. One was explicit teaching. This is the up-frontapproach. To teach children the principle explicitly, we would have to explain the spelling principle to them clearly and completely, withsuitable examples to provide them the relevant experience. The exam-ples would be -ion and -ian words that we would place into twoclear categories, and we would go through the difference between thesecategories. All the arguments that we have been through in Chapters 1and 2 pushed us in the direction of explicit methods, and so we decidedto teach some of the children in the study in this way.
We also wanted to look at the possible effects of what we calledimplicit teaching. This is to give children examples of words with -ionand -ian endings but not to explain the principle. So, the aim herewould be to give the children the chance to work out the principlesthemselves, by going through the examples. In some ways, this methodis quite close to present practice, since schoolchildren usually are askedto learn the spellings of words like election and magician and couldpossibly work out the person/nonperson principle for themselves. But,in other ways, our ideas for implicit teaching are quite different fromnormal school practice. Our plan was to put the two kinds of words (the-ion and the -ian words) into separate categories and then to makedirect comparisons between these categories, so that the child might seewhat the words in each category had in common, and what was the
66 What does the research tell us?
difference between the two. Schoolchildrens experiences with wordswith these two endings are usually a great deal more haphazard thanthat.
Our third idea for teaching the morphemic principle was to miximplicit and explicit teaching. We decided that a quite reasonable wayof teaching the principle would be to start our teaching of some of the children by giving them implicit experiences by categorizing thewords but not explaining the difference between the categories explicitlyat first. Halfway through, however, we would move from implicit toexplicit teaching: We would ask the children to explain what the prin-ciple was and, if they could not do so, we would explain to them howthe two word categories differed.
We need to make one more point before we describe the first of theteaching studies in which we looked at ways of promoting childrenslearning of the principle about --ion and -ian endings. This principlehas two parts to it, since it is a principle about two morphemes and twospellings. One morpheme (spelled -ian) signals that the word repre-sents people, the other (spelled -ion) that the word is not about people,and it usually represents an abstract noun (for example election,education, institution). Plenty of psychological evidence suggeststhat children would find it easier to learn a principle that involvedpeople, since this is an easy category for them, than a principle that ismostly about abstract ideas. If this is so, children who are taught theprinciple might find the -ian part of it easier to learn, since it involvesthe easier category of the two. Although it is possible to examine this,we will not carry out any analysis in this study to investigate thisconjecture.
Thus we set out to compare three different teaching methods, whichmeant that we had to compare three groups of children, each of whichwe taught in different ways. Since this was a tightly controlled experi-ment, we also included a fourth group of children, a control group, whoactually received exactly the same amount of instruction as the childrenin the other groups, but were taught about something entirely different,strategies that might help them to understand passages of prose.
The 200 children who took part in this study were all 9-year-olds andtheir average age was 912. They came from two schools in Oxford, andthe numbers of children from each of the two schools was the same inall four groups. Table 3.1 gives some detail about the four groups.
From the laboratory to the classroom 67
The study had a simple traditional design, which we have set out inFigure 3.1. First, we gave the children a pretest to measure how wellthey could spell the endings that interested us. Then we had twoteaching and learning sessions, in which we taught the different groupsin different ways. Very soon after the end of the second training sessionwe gave the children a posttest, which we called the immediateposttest and which was identical to the pretest and thus measured
68 What does the research tell us?
Table 3.1 Mean age and standard deviation for the intervention and control groups in Study 1
Group Mean Standard N(years) deviation
Morphology explicit 9.5 0.60 40Morphology implicit 9.7 0.67 43Morphology mixed 9.6 0.53 42Control 9.5 0.71 75
Two 20-minute teaching sessionswith pairs of children
Two month interval
Delayed posttest Figure 3.1 Design of the firstteaching study.
the different levels of progress in the four groups as a result of thedifferent kinds of teaching that they were given. Two months later wegave the children a delayed posttest. This was the same as the previoustests, and its aim was to test how well children had remembered the spelling principle a considerable time after they had been taughtabout it.
The pretest and posttest measures of spelling
Our pretest provided a measure of how well children spelled the endingsthat interested us. This consisted of two tasks that were designed to measure the childrens ability to spell words that end in -ion and -ian. Box 3.1 gives details of both pretest tasks. In one task, we askedthe children to write some real words ending in -ion and -ian. Ourtechnique was to present the children with a set of written sentenceswith a gap in each of them (see Box 3.1). This gap represented a missingword. Then we read out the whole of each sentence, including eachmissing word, and we asked the children to write in that word. So, forexample, the missing word in the sentence Kate was the only onewithout an ________ was invitation, and this was the word that weasked the children to write in after we had dictated it.
From the laboratory to the classroom 69
The word- and pseudoword-spelling tasks usedin Studies 1 and 2
Words1. The vegetarian grew her own vegetables.2. The submarine dived into action.3. The Egyptian rode past the pyramids.4. The woman was glad that she had a good education.5. Kate was the only one without an invitation.6. The cut had an infection.7. The guardian stood at the gate.8. Grandma had a celebration on her birthday.9. The discussion went on and on!
10. I hate injections.11. The magician performed a magic show.
In the other task, we asked the children to write pseudowords. Box 3.1gives the items used in this task. We introduced pseudowords becausethese make it possible for us to test whether or not the children dounderstand a spelling principle. With spellings of real words, there isalways a chance that someone might produce the correct spellingsbecause he or she has learned the correct spelling for each individualword by rotewhat psychologists call word-specific learningandnot by learning any principle. Pseudowords, in contrast, rule out thepossibility of word-specific learning completely: Such words, bydefinition, are completely unfamiliar. If children put -ion and -ianendings on appropriate pseudowords, they must have been using aprinciple to do so.
In this pseudoword test we used the same missing word technique aswith real words, and here the context provided by the surroundingsentence was essential. As you can see from the examples in Box 3.1,these surrounding sentences made it clear that in some sentences, thepseudowords referred to a person. For example, Someone who worksin a korpect is a ________. The pseudoword we dictated was korpect-ian. In other sentences, the pseudoword did not represent a person. For example, The place where you find foppats is a ________. Thepseudoword we dictated was foppation. We wanted to know whether
70 What does the research tell us?
12. Congratulations to the winner.13. The electrician fixed the lights. 14. The Brazilian scored again. 15. The librarian checked out my book.16. The comedian was funny.
Pseudowords1. A person who comes from Barrim is called a Barrimian.2. A saughty baby is full of saughtiness.3. If you often tekate that shows that you like tekation.4. A dinosaur with a knot in its tail is a knotosaurus.5. A fenagious person thinks that fenagion is important.6. When people denass they are making a denassion.7. Someone who works in a korpect is a korpectian.8. The place where you find foppats is called a foppation.9. Someone who does lagic is a lagician.
10. Anyone who organizes a golbary must be a golbarian.
children would be more likely to put an -ian ending when the pseudo-word meant a person (korpectian) than a nonperson (foppation).
Both tasks contained other words, whose spellings were not based onthese morphemes, as distractors. These are not analyzed here.
The two teaching and learning sessions
Soon after the pretest, all the children, including the children in thecontrol group, went through two teaching and learning sessions (wetaught, they learned). The second session came a few days after thefirst. Notice that there were two of these sessions only. Since eachsession lasted for around 20 minutes, the total teaching time for eachchild was really quite small.
We used the same general plan and exactly the same material to teachthe explicit, implicit and mixed groups about the -ion and -ian mor-phemes. We saw all these children in pairs, and we based our teachingon two different activities, which we called games. Boxes 3.2 and 3.3give details of these games and of how the procedures in them weredifferent for the three groups.
From the laboratory to the classroom 71
The analogy game
The actors in the intervention games were an adult experimenter,who was a psychologist, and a pair of children. During each game,the attention of all three was focused on a laptop computer on whichthe experimenter, using PowerPoint, presented the children withthe material needed for the game.
In the analogy game we showed the pairs of children several setsof four puppets. In every set, there was a word box immediatelyunderneath each of the four puppets. Here, as examples, are twosuch sets.
When we presented a set, all four puppets in the set appeared onthe screen at the same time, but their word boxes were empty. Thenthe top puppet on the left waved its arms and apparently spoke aone-morpheme word (for example, magic). The written form ofthe word immediately appeared in the word box below this puppet.Next the top-right puppet spoke a two-morpheme word, also
72 What does the research tell us?
accompanied by a few arm waves, which was derived from the first word (for example, magician), and, of course, the writtenform of the word appeared below in this puppets word box. Thethird step was for the bottom puppet on the left to produce a newwordagain, a one-morpheme word (for example, music). At
From the laboratory to the classroom 73
this point, a question mark appeared in the word box underneaththe fourth puppet, and the experimenter asked the two children towork out what the missing word would be and how it should bespelled, making it clear to the children that the relation betweenmusic and the missing word had to be the same as the relationbetween magic and magician. (The illustrations actually showthe display as it was at this point of the game.)
Then the two children discussed what the word was and how towrite it and, after they had come up with their own agreed solution,the fourth puppet spoke its word (the correct word) and theappropriate written word replaced the question mark in the puppetsword box.
At this point of the game, the procedure diverged for the differentgroups. In all the trials experienced by the explicit group and in theexplicit trials experienced by the implicit group (the second halfof the sets given to them in this game), the experimenter gave thechildren an explanation of why the correct word and the other two-morpheme word in the set ended in -ion and not in -ian, or viceversa. This explanation was not given to the children in the implicitgroup, nor to the mixed group during the implicit trials that wegave to this group (the first half of the sets given to them in thisgame).
Finally, we presented all the children in the explicit, mixed, andimplicit groups with a PowerPoint rsum of the whole game in theform of all the sets of four words used in the game. We arranged this display, which is shown below, to make as striking a contrast aspossible by putting all the -ian sets on one side of the display andall the -ion sets on the other. We discussed the reason for thedifference between the two endings in these two collections of wordswith the children in the explicit and mixed groups, but not with thechildren in the implicit group.
magic protectionmagician protect
India confessionIndian confess
library subtractionlibrarian subtract
history additionhistorian add
music infectionmusician infect
74 What does the research tell us?
technical collectiontechnician collect
Italy educationItalian educate
Hungary imitationHungarian imitate
political suggestionpolitician suggest
Egypt discussionEgyptian discuss
The correction game
The idea in the correction game was to encourage the pairs ofchildren to be the judges of someone elses spelling. We presentedthem with a set of written sentences, again in PowerPoint. All thewords in the sentence were typed except for one word which waswritten in rather childish handwriting. We told the children thatJoe had written these words (this was literally true: Joe was part of our research team and had written all these words), but that he wasnt always right in his spelling, and that it was their job todecide whether Joe had got the spelling of each of the words rightor not. The words were all words that should end in -ion or -ian.In some cases, the handwritten word was spelled correctly, and inothers the ending was incorrect. Here are some examples. Noticethat there is a box under each sentence which contains the correctspelling of the word in question.
The was wonderful.
From the laboratory to the classroom 75
With each word, our procedure was to show the children thesentence with the handwritten word on the laptop screen. At thisstage the box with the correct spelling was not visible. We asked thechildren to discuss with each other whether or not Joe had spelledthe handwritten word correctly and, if they thought it was incorrect,to write down the correct spelling. When they had done this, thecorrect spelling of the word appeared under the sentence, and thedisplay looked like the ones in our illustrations.
At this point of the game, the procedure diverged for the differentgroups. In all the trials experienced by the explicit group and in theexplicit sets experienced by the implicit group (the second half ofthe sets given to them in this game), the experimenter gave thechildren an explanation of why the word ended in -ion and not in-ian, or vice versa. This explanation was not given to the childrenin the implicit group, nor to the mixed group during the implicittrials that we gave to this group (the first half of the sets given tothem in this game).
Our final step was to give all the children in the explicit, mixed, and implicit groups a PowerPoint rsum of the whole game in the form of all the words whose spellings they had had to judge. We arranged this display, which is shown below, to makeas striking a contrast as possible by putting all the -ian sets on one side of the display and all the -ion sets on the other. Wediscussed the reason for the difference between the two endings inthese two collections of words with the pairs of children in theexplicit and mixed groups, but not with the children in the implicitgroup.
The gang made a to the police.
We called one of these games the analogy game (see Box 3.2). Westarted with a simple word, like electric and then we presented asecond word: A two-morpheme word, like electrician, which wasderived from the first word in the pair. After that we immediatelypresented the children with another simple word like magic and askedthem to produce a new word, which had the same connection to themagic as electrician had to electric. The pair of children had towork out these analogous words together and write down the new word.The children were then allowed to check on the computer screen wewere using whether they had produced the right word, and, if not, whatthat word should have been. The researcher who worked with thechildren also provided an explanation for the correct choice, but herewhat the researcher said varied between the three groups, as explainedearlier on.
The second task was the correction game (see Box 3.3). In this weasked the pairs of children to decide between themselves whether thespellings of another (fictional) child were right or not. If they thoughtthat the spelling was wrong, which it often was, we asked the childrento write down what it should have been. After that, we told the childrenwhether they had got it right or not, and, as Box 3.3 shows, in some caseswe explained why.
In fact, the only difference between the three groups was in what theinstructors told the children after they had made their decisions in bothgames. The instructors told the children in the explicit group not justwhether their answer was right or wrong, but also why. Thus, after each
76 What does the research tell us?
answer, the instructor stated clearly and explicitly the person versusnonperson principle about -ion and -ian endings.
We did not give the children in the implicit group this explanation.Their experiences were exactly the same as those of the children in theexplicit group (and, as we shall see, of the children in the mixed group)up to the point where the instructor had told them whether their answerwas correct or not. That was the end of each problem, as far as thechildren in the implicit group were concerned. At that point, theinstructor moved them on to the next problem.
The third group was the mixed group. With these children, we dividedeach of the two games into two halves. In the problems in the first halfof each game, we treated these children in the same way as we had thechildren in the implicit group: We told them whether they were rightor not, we gave them the correct answer if their answer had beenincorrect, but we explained nothing about the reasons for the differencebetween -ion and -ian endings. In the second half of each game, weswitched to the explicit method with the children in the mixed group.Now we not only told them whether they were right or not, we alsogave them the reasons why.
Finally, a reminder about the control group: The children in thisgroup also received two teaching and learning sessions in between the pretest and the immediate posttest, but the intervention was aboutsomething entirely different. These children were being taught strate-gies that we designed to improve their text comprehension and theyformed the experimental groups in a study of reading comprehension.
Vice versa, the three experimental groups in the study on morphemesthat we are describing here also formed the control group for the projecton reading comprehension. We managed to combine the two studies inthis fruitful way by giving all the children the pre- and posttests for theproject on spelling morphemes and also for the project on readingcomprehension. Thus, all the children who took part in these projectsstood to gain something from them, some in spelling morphemes andothers in reading comprehension. For once, the control-group childrenwere not the losers.
Since the posttests were the same as the pretest (see Table 3.2), weonly need to tell you about their timing at this point. We gave thechildren the immediate posttest on the first school day after they wentthrough the second and last teaching and learning session. We gave
From the laboratory to the classroom 77
them the delayed posttest two months later, and it is worth mentioningthat in that two-month period neither their teachers nor we ourselvestaught them anything more about the morphemic spelling principle forthe -ion and -ian endings.
The results of the first study
We should like to set the scene by talking about the pretest scores. Thesewere our baseline, of course, against which we could check the extentto which the children in the various groups improved in spelling afterthe teaching and learning sessions. The pretests also tell us somethingabout childrens knowledge, or rather their lack of knowledge, of theprinciple about -ion and -ian endings.
We need not spend much time on this point or on these data since wehave discussed them already in Chapter 2. We simply want to remindyou that the childrens performance in the pretest gave us no confidenceat all that any of the children in the study had grasped the morphemicspelling principle for -ion and -ian endings. The children did not doat all well with real words like education and magician, even thoughthey had had the chance to learn the spelling of these familiar words byrote learning. They made even more mistakes in spelling the pseudo-words, and, since the pseudowords task is the acid test of principlelearning, we felt that we could quite reasonably conclude that thesechildren started off the experiment with little, or more likely with no,understanding of the principle that we wanted to teach them. This wasnot a surprise. We know that there are many aspects of our nativelanguage of which we are totally unaware. So we had expected thatchildren might use words like magician to refer to persons andeducation as abstract nouns but have no awareness of the distinctionbetween these categories or of the fact that this distinction is marked inspelling by the use of different suffixes.
We can turn now to the posttests to see whether the children man-aged any better after our instruction. It is best to consider the real wordsand the pseudowords separately. Figure 3.2 gives pre- and posttestresults for spelling the real words. These scores are for the childrenssuccesses with both kinds of ending. The figure tells a clear story.
The two teaching and learning sessions on morphemes had animpressive effect on the childrens spelling of the real words. The scoresfor the three groups who were taught about morphemes were consis-tently better in the posttests than the pretest scores. In the immediateposttest this improvement was at its strongest, with the children in the
78 What does the research tell us?
explicit group and, next, in the mixed groupthe two groups of childrenwho had received explicit teaching about the -ian and -ion endings.In this posttest, all three groups who had been taught in one way oranother about the two endings fared better than the control group.
None of the three groups who had learned about the spelling of these morphemes fared quite as well in the delayed posttest as in theimmediate posttest, but this was not at all surprising given that thesechildren had had no systematic instruction on these suffixes during thetwo-month interval. In fact, the decline, if that is the right expression,was very small indeed among the children in the mixed and implicitgroups, but slightly larger in the case of the explicit group. However, thescores of the children in these three groups were still higher in thisdelayed posttest than they had been in the pretest at the beginning ofthe study. At the time of the delayed posttest, the explicit group was stillahead of the mixed and implicit groups, but not so far ahead.
These results with real words show us two things.
1. We did improve childrens spelling of the -ian morpheme, andthis improvement survived a two-month period during which thechildren were given no instruction about morphemic spellingprinciples at all.
From the laboratory to the classroom 79
PretestImmediate posttestDelayed posttest
Figure 3.2 The mean number (out of 16) of correctly spelled -ion and -ian endings in real words in Study 1.
Note: The scores in this figure are the adjusted scores from an analysis of covarianceof the scores in the two posttests, in which the pretest scores were the covariate. Inthe immediate posttest the scores of all three taught groups were significantly higherthan those of the control group.
2. Explicit teaching seemed to be the most effective way of producingthis improvement, particularly in the short term. Telling childrendirectly about the principle worked.
Now we must consider the results for the equivalent spellings in pseudo-words. There are two reasons why these are particularly important. One is the point, which we have made already, that pseudowordsprovide us with the critical test for principle learning, because they ruleout word-specific learning. The second is that we only worked with realwords in the teaching and learning sessions. We knew, therefore, thatthe childrens scores in the pseudoword posttests would give us a goodmeasure of how well they transfer their learning to a completelydifferent kind of material.
Figure 3.3 gives the childrens scores for spelling the pseudowords.The first thing to say about this figure is that it shows that the childrenwho had been taught about -ion and -ian endings easily extendedwhat they had learned to pseudowords, even though none of theinstruction that they had received included any experience with thiskind of word. All three groups of children given instruction on theseendings did better in the posttests than the children in the control group.
80 What does the research tell us?
Figure 3.3 The mean number (out of 8) of correctly spelled -ion and -ian endings in pseudowords in Study 1.
Note: The scores in this figure are the adjusted scores from an analysis of covarianceof the scores in the two posttests, in which the pretest scores were the covariate. Inthe immediate posttest the scores of all three taught groups were significantly higherthan those of the control group. In the delayed posttest the explicit and mixedgroups scored significantly better than the control group.
Another interesting result was that the two groups of children who weregiven some kind of explicit teaching (explicit and mixed methods) didbetter in the delayed posttest than the implicit group. Finally, we were pleased to see no decline over the two-month period betweenimmediate and delayed posttests among the children in the mixed andimplicit groups, even though they were not taught about the twoendings during those two months. There was a slight decline in theexplicit groups scores from the immediate to the delayed posttest, butthis groups scores were still far ahead of their pretest scores in thedelayed test.
So, we drew much the same conclusions about the pseudowordspelling as about real-word spelling.
1. Teaching children about the morphemic principle did help them touse the morphemic spelling, even with a type of word that had notbeen part of the teaching sessions.
2. Explicit teaching, on the whole, worked better than implicitteaching. For the second time, we can say that telling children aboutthe principles and giving them a chance to use these principlesworks.
Study 2: Teaching children about the -ianand -ion endings in the classroom
The results of the teaching project that we have just described arecertainly encouraging. The children clearly did learn something newand important about the link between different kinds of meaning anddifferent forms of spelling. Such results lead to an obvious but essentialnext step: This step is out of the laboratory and into the classroom.
We had conducted a laboratory-style study. In fact the study tookplace in the childrens school, but we took all the childrenone pair ata timeout of the classroom in order to teach them our morphemicspelling principle. So we had the fortune to be able to give each pair of children our undivided attention for the whole of each teaching and learning session in a quiet atmosphere with no distractions and noother demands on the children we were teaching. Having establishedthat our games and our material worked in these well-controlled andadvantageous circumstances, we now wanted to find out whether theywould be as effective when used by teachers working in real classrooms.Would it be possible, we asked, for teachers to put across the samemorphemic principles in these quite different circumstances but using
From the laboratory to the classroom 81
exactly the same material and the same methods that we had in ourquiet laboratory room?
A great deal depends on the answer to this question. The main aimof our research was to find out whether children can and should betaught about morphemic principles at school. So we had to find out ifour methods worked in school classrooms as well as in the laboratory.
The study that we shall now describe is, as far as we know, unique.Our friends and colleagues in a school in an eastern area of London,Lauriston Primary School, repeated our laboratory study in theirclassrooms. It was an almost exact repetition of our first study. TheLauriston teachers
worked with children in the same age group as the children in ourfirst teaching study;
administered exactly the same pretest, with the same words, asours;
conducted the same immediate posttest and the same delayedposttest two months later;
taught the children about the morphemic principle in two sessionsonly;
arranged for the children to work in pairs during the two teachingsessions;
gave the children the analogy and the correction games in these twoteaching sessions in the same way and with the same words as ours;
even formed the control group in the same way in their study as wehad in ours: Our Lauriston colleagues also taught some childrenabout morphemes and others about strategies for understandingtext, and again the children who were taught about comprehensionformed the control group in the morpheme-spelling study (and viceversa);
assigned the children to the intervention and control groupsrandomly, as we had also done in our study.
As far as we know, this is the first time that a laboratory study has beentransferred to a classroom and repeated so exactly there.
There were two main differences only between our first study and the Lauriston study. The Lauriston teachers used only one of our three teaching methods. They only taught children about morphemesexplicitly. We had established that this was the most successful methodin our first study, and our results led the teachers to adopt that methodonly. So in their study there was one morpheme group only: An explicit
82 What does the research tell us?
group. This allowed them to compare this explicit group with the controlgroup.
The second difference was about the presentation of the material in the two teaching sessions. In the classroom teaching sessions, thechildren worked in pairs, as in our earlier study, but the presentationof the material for the analogy and correction games was to the wholeclass at the same time. As in our study, the presentation was in Power-Point, but now it was projected onto a screen, whereas in the previousstudy the children saw it on a laptop screen.
The results of the second study
Again we wanted to know whether children can learn a new morphemicspelling principle and apply it to new words. As before, we looked at theeffects of teaching the morphemic principle on the childrens spellingof:
real words ending in -ion and -ian; pseudowords ending in -ion and -ian.
In Figure 3.4 we present scores for the two groups spelling of theendings in the real words. It shows a sharp improvement in the scoresof the children in the taught group from the pretest to the immediateposttest. There was a slight decline in these childrens scores during thetwo-month period between the two posttests, but they still spelled thewords endings a great deal better at the end of the experiment than atits beginning. In both the posttests the spelling scores of the childrenwho had been taught about the two endings were higher than those ofthe children in the control group. The scores of the control groupactually improved between the immediate and the delayed posttests,which meant that the difference between the two groups was appre-ciably smaller in the second posttest than in the first. This had an effecton the statistical analysis, which produced a significant differencebetween the taught and control groups in the immediate posttest butnot in the delayed posttest. So, we can say with certainty that theclassroom teaching led to an immediate improvement in the childrensmorphemic spellings. We cannot be so sure about the lasting effects ofthis teaching. On the one hand, the children in the taught group still hadfar higher scores in the final posttest than in the pretest. On the otherhand, the apparently spontaneous improvement over time in the scoresof the control group, though not as strong as in the taught groups scores,
From the laboratory to the classroom 83
was strong enough to make us wonder how much of the taught groupsgood performance was due to the teaching and how much of it wouldhave happened spontaneously.
One distinct possibility is that the spontaneous improvement in the control groups scores was not after all completely spontaneous.There could have been some contamination, as it is commonly butinelegantly called, between the two groups. In other words the childrenin the different groups could have talked to each other about what theyhad been taught, and this could have had a positive effect on the controlgroups spellings. The idea may seem far-fetched, but there are well-documented examples of this sort of effect in psychological research onyoung children.
So, we can conclude that
explicit teaching about this morphemic spelling principle has animmediate and impressive effect on childrens spelling;
we still need to know more about how to maintain this improve-ment over time.
84 What does the research tell us?
PretestImmediate posttestDelayed posttest
Figure 3.4 The mean number (out of 16) of correctly spelled -ion and -ian endings in real words in Study 2.
Note: The scores in this figure are the adjusted scores from an analysis of covarianceof the scores in the two posttests, in which the pretest scores were the covariate.There was an overall difference between the taught and the control groups in thetwo tests. The difference between the taught and control groups was significant inthe immediate posttest (with an effect size of 0.16 of a standard deviation, which issmall but quite good, considering that only two teaching sessions were used) butnot in the delayed posttest.
We shall have more to say about this second conclusion in followingchapters. At this point, we can look at the effects of the same classroomteaching on the spelling of pseudowords. The scores for these pseu-doword spellings are shown in Figure 3.5. Here again we can see a striking immediate effect of teaching: The children who were taughtabout the two morphemic endings spelled them much better after thisteaching than before and much better than the control group did in theimmediate posttest. The children in the taught group did not do so wellin the delayed posttest as in the immediate posttest, but they stillmanaged to spell the two endings much better at the end of the studythan at the beginning.
The control group children did better in the two posttests than in the pretest. They did not ever catch up with the children in the taughtgroup, but by the time of the delayed posttest there was no significantdifference between the two groups. Again, we must wonder whether theimprovement in the control groups scores was a spontaneous one or theresult of some talk between the children in the two groups.
Thus, our main conclusion from the childrens spellings of pseudo-words is that the children applied what they were taught with real words
From the laboratory to the classroom 85
Figure 3.5 The mean number (out of 8) of correctly spelled -ion and -ian endings in pseudowords in Study 2.
Note: The scores in this figure are the adjusted scores from an analysis of covarianceof the scores in the two posttests, in which the pretest scores were the covariate.The overall difference between groups fell just short of significance. In a separateanalysis of covariance of the immediate posttest the scores of the taught group weresignificantly higher than those of the control group, with an effect size of 0.42 of astandard deviation.
to pseudowords as well and dealt with these pseudowords in much thesame way as with real words.
Our general conclusions from the first twostudies
When we put the results of these two studies together, we thought thatthey told an interesting and useful story. This was the first time, wethink, that a study originally done in a laboratory setting had beenrepeated so exactly in the classroom and with such similar results. Thetwo studies show how a method, first developed in a laboratory study,for teaching children an entirely new spelling principle worked just aswell in the classroom as in the laboratory.
These two studies also established that explicit teaching of thisparticular principle works extremely well. The consistent success ofexplicit teaching also suggested that that this kind of teaching might bethe most effective way of explaining morphemic principles to school-children and persuading them to adopt these principles themselves. Sothe studies were a spur to further research on teaching children aboutmorphemes in the real world of the classroom.
We decided to widen our brief. Our basic claim is a general one: Weare arguing that it is possible, and beneficial, to teach children about thelinks between a large number of morphemes and their spelling. Thestudies that we have described so far dealt with only one of these links.The aim of our next study was to develop an instruction package thatwould deal with a variety of morphemic principles and would assess theeffects of this broader instruction on childrens spelling of differentsuffixes.
Study 3: Constructing and assessing apackage for teaching morphemic principles
Teaching children about spelling involves two different decisions. Oneis what to teach, and we have made our point here quite clear: We thinkthat children should be taught about spelling principles based onmorphemes and their meaning in order to improve their spelling beyondphonological principles. The second decision is how to teach.
In order to design a wide-ranging teaching package, we had to makemany choices about how to teach. Some were based on the researchabout teaching that we had carried out before. For example, we knewfrom our previous work that it is better to teach a contrast between two
86 What does the research tell us?
spellings (as in the -ian and -ion case) than to teach each of theseseparately. The first study described in this chapter further showed thatit is better to be explicit about the principles than to leave it all for thechildren to discover. However, there was still much to be decided aboutwhat to include in the package and how to involve the childrens logicand imagination so that they would enjoy the teaching and learn aboutmorphemes. We feel that it is important to make many of our decisionsexplicit: Teachers and researchers will be in a better position to assessthem if they realize that there were decisions being made. When ateaching package is built, it is not possible to assess each of thesedecisions separately, so the best that we can do is to be clear about ourown principles in making choices.
Principles used in our teaching materials
The groundwork we used to develop this teaching program was doneinitially in another study, which we ourselves carried out (Nunes et al.2003), about how improving childrens awareness of morphemes affectstheir literacy skills. In our previous study we used word reading andspelling as outcome measures. We found that our teaching interventionabout morphemes was as powerful at improving childrens word reading(measured by a standardized test) as a teaching intervention wherethey were taught about phonological spelling principles. We also foundthat only the morphological intervention improved the childrensspelling of words whose spelling is determined by their morphologicalstructure.
Our previous intervention had been carried out in a laboratory-likesetting: The children worked in small groups with researchers outsidethe classroom. In this study, we wanted to develop this package so thatit would be possible to carry out the intervention in the classroom. Thus,we examined carefully what we had done in the previous study in orderto maintain what we thought were the principles of the teachingintervention, while adapting it for classroom presentation. We discussthese principles below.
It is important to develop childrens awareness ofgrammar
Morphemic principles and grammar are connected in many ways. In the case of the -ian and -ion suffixes, the words that we used wereall nouns. This is not a coincidence: Derivational suffixes, such as -ian
From the laboratory to the classroom 87
and -ion, are used to form nouns, such as mathematician andinstitution (-ian is also used to form adjectives like mammalian).Inflectional morphemes, which do not change the grammatical categoryof the words that they are attached to, are also connected to particularword classes: For example, we use the suffix -s (or its allomorph -es)to mark the plural of nouns but not of adjectives, which do not haveplural marks in English.
Grammar is also connected to derivational morphemes in a differentway. If we want to form an agent, for example, there are differentderivational morphemes that could be used: -er, -ian, or -ist. Whichsuffix is the correct one depends on the word class of the base formused. Agentsor, more generally, nouns that refer to personscan be formed from verbs, nouns, or adjectives. Usually, the suffix -er (but sometimes also -or) is used to form agents from verbs: Forexample, readreader, cleancleaner, swimswimmer. Thesuffixes -ian and -ist are used to form person words from nouns and adjectives: musicmusician, magicmagician, electricelectrician, BrazilBrazilian, ItalyItalian are examples of -ian words, and sciencescientist, artartist, femininefeminist, special specialist, communecommunist areexamples of -ist words.
This analysis of permissible combinations of base forms and suffixessuggested that we should increase the childrens awareness of gram-matical categories in order to create a basis for the development of theirawareness of morphemes.
It is very difficult to explain grammatical categories because they arenot defined by content but by relationships in sentences. For example,the commonly used definition of verbs as action words provesconfusing for children and is inaccurate: We have witnessed childrenarguing that in the sentence I had a fight in school, the word fightis an action word and therefore a verb. However, children are able tounderstand that sentences have frames and that some words fit, whileothers dont fit, into particular gaps in these frames. So, we createdexercises where the children had to decide whether certain words could fit into particular places in sentences and had to discuss why. To illustrate how this type of exercise works, some examples taken from one task used at the beginning of the program are presented inFigure 3.6.
88 What does the research tell us?
From the laboratory to the classroom 89
We saw a in the town centre.
Do the words fit into the sentences? Tick those that fit in thesentence. Discuss with your friend why some do and some dont.
We before we go to school
Figure 3.6 Items from a task used to make children aware of how placesin a sentence frame define grammatical categories.
It is important to engage childrens reasoning indifferent ways
Our aim was to teach the children to think about morphemes in general,not about a list of particular words formed with specific morphemes.There is a large body of psychological research about the differencebetween learning principles versus learning specific facts. This researchis referred to by different terms: Studies of learning set, general-ization or transfer are all about how learning can become general,rather than remain tied to specific instances. Although there are manydifferences in opinion, there is a certain amount of convergence acrossstudies.
There are two crucial elements in promoting the learning of principlesrather than specific facts. The first is to promote categorization: Whenlearners form a category that involves different specific instances, theycan respond to the category and ignore the differences between thespecific examples. We used this principle in a variety of ways. Oneexample was to help the children form a category of suffixes that formperson words. The children would, for example, see on the screen onwhich the task was projected the three suffixes, -ian, -er and -ist,and would be asked to indicate Who is the person that . . . ? Thiswould help them recognize that different suffixes can be used to formperson words.
Another example was to help the children realize that some prefixesrefer to number. Here we asked the children to consider the mostimportant difference between a bicycle and a tricycle and to discusswhether there was an element in the word that gave a clue to thisdifference. Once they were able to think about the possibility that someprefixes give a clue to number, they were presented with other words(for example, triangle, pentagon, uniform, binoculars) and askedto identify prefixes that give clues about numbers. Figure 3.7 showssome examples of this task.
The second principle in creating tasks with the aim of helping childrento generalize is to promote the engagement of different operations ofthought in solving problems. Thus we created examples where thechildren had to make analogies (as in our analogy game described onp. 71), to choose the best word for a sentence, to focus on grammaticalcategories (see Figure 3.8), to correct spelling errors (as described in thecorrection game), to count morphemes in multimorphemic words, toanalyze novel words (to name just a few examples).
90 What does the research tell us?
From the laboratory to the classroom 91
What is the most important difference between a bicycle and atricycle? What is it about the word that gives you a clue to thisdifference?
Are there similar clues in these words? Identify these clues.What do these clues suggest?
Today in town I saw a biheaded monster. Can you draw a biheadedmonster?
Figure 3.7 Examples of items used to teach the category of prefixes thatrefer to number.
It is important to engage childrens imagination
Some of our tasks involved pseudowords, where the children were asked to think about nonexisting words that combined real suffixes withmade-up base forms. These tasks involved the children in imaginingwhat a word could mean, if it existed. Some examples are presented inFigure 3.9.
It is important to maintain childrens motivation
In order to maintain the childrens motivation, we designed tasks thatwere presented with the support of a computer, so that stimuli could becolorful and interesting. The childrens tasks were varied: They receivedsupport to formulate their own ideas about the morphemes in the task.
92 What does the research tell us?
You can add -en to some words and form a new word.Sometimes you add it at the beginning, sometimes youadd it at the end. What type of word do you form?
Figure 3.8 Focusing on verbs.
Feedback was provided in positive ways using the PowerPoint presen-tations: The answers moved onto the screen in different ways and oftenwith different sounds. They also received feedback from each otherwhen they discussed the explanations and presented them to the class.
From the laboratory to the classroom 93
What jobs do these people do on Mars?
They are spamters. Theythe spaceship for flying too low.
They are montists. They look after , which they keep in theirlittle black boxes.
At work his job is to loment the towers.He is a .
Figure 3.9 Examples of items used to practice identification of stems andcreation of person words. Playing with pseudowords was fun.
They had something to talk about and learned ways of talking aboutlanguage.
The design of the study
In order to assess whether our package had an impact on childrensspelling and vocabulary, we needed to have an experimental and acontrol group. We did not want the children in our control group to bethe losers in the study: As before, we wanted to make sure that they alsobenefited from participation. However, it was not possible to work witha parallel intervention for the control group in this study because oftime constraints. We decided instead to use a waiting-list model, underwhich the control group receives the same learning opportunities as the taught group but at a later time. The taught and the control groupscompleted a pretest, an immediate posttest and a delayed posttest,which took place about 8 weeks after the teaching for the taught grouphad been completed. After this delayed posttest, the teacher of thecontrol group received the same materials for use with her class.
The teaching program was implemented by two teachers in a schoolin Oxford with their Year 4 and 5 classes, which included a total oftwenty-eight children. The activities were included in the literacy hour(the time specifically aimed at teaching literacy in English primaryschools) so that the taught groups did not receive extra literacy instruc-tion in comparison with the control group; they received the sameamount of literacy instruction, but worked on a special curriculumduring the two-week period in which the teaching was implemented.The teachers who worked with the control groups also taught Year 4 and5 classes, which included a total of fifty children.
Before the teaching program started, the researchers showed theteachers some of the activities and discussed their aim briefly. Thisorientation took place in a single session. The teachers received a CD-ROM that contained the activities and response sheets for the childrento complete during the sessions.
In order to support the teachers in case they had questions and todocument how the tasks worked in the classroom, one researcher wasalways present during the teaching sessions. It turned out that theteachers did not require further clarification beyond the initial orien-tation. The activities were easily understood by the teachers andchildren alike.
94 What does the research tell us?
The pretests and posttests
We used a design similar to that employed in the previous studies. Thechildren went through a pretest just before the teaching programstarted, an immediate posttest immediately after this program wascompleted, and a delayed posttest approximately 8 weeks later. Thesetests were identical and included the spelling of suffixes in words andpseudowords. The items used are presented in Box 3.4.
From the laboratory to the classroom 95
The items used for the word- and pseudoword-spellingtests in Study 3
The words and pseudowords in italic were not on the childrensanswer sheet and were dictated by the teacher. The childrensanswer sheet contained the rest of the sentence. The children spelledthe missing word above the line that marked its place in thesentence.
Word spelling test1. On Sunday we are going to see the magician.2. The policeman asked me to make a statement.3. To tease a gorilla is complete madness.4. Walk on the pavement, John!5. My sister wants to be a musician.6. You must not be careless when driving. 7. The richness of the colors made the picture attractive.8. He was allowed to park near the school because he had a
disability sticker.9. He was overcome by emotion and began to cry.
10. The soft chair was very comfortable.11. Tim was so cheerful during the tour! 12. Politeness is important when you ask for something.13. We will hear a combination of sounds.14. She checked the measurement before writing it down.15. The favourite singer won the popularity contest.16. The politician was often on television.17. I had an enjoyable visit with my aunt.
The method used for eliciting the spellings was the same as the onedescribed previously: The childrens answer sheets contained sentenceswith a gap where the dictated word was to be written. Box 3.5 (oppo-site) shows a portion of the test with a childs spelling. Because it wasimpossible to carry out the delayed posttest with one of the controlclasses, the number of control children who participated in the delayedposttest was reduced to twenty-five.
A brief overview of the teaching program
The intervention program consisted of a set of activities designed tooccupy approximately seven sessions of about 50 minutes each(although this could vary because the time taken up for discussionvaried). Because the program was to be implemented in the classroom,
96 What does the research tell us?
18. There was a great similarity between the twins.19. He is a graceful dancer.20. The librarian found the book I had lost.21. When he won the lottery, he cried with happiness.22. Everyone voted in the election.23. The lion tamer was famous for being fearless.24. The footballer was very skilful.25. Our destination is Athens.26. George brought an inflatable mattress to our sleepover.
The pseudoword spelling test1. She never grats when she dances. She is a gratless dancer.2. A dinosaur skin is blaged. The blageness protects it from the
cold.3. I cleaned my car with great mape. I was mapeful.4. A saughty baby is full of saughtiness.5. The teacher told Harry to senk to his work. That work needed a
lot of senktion.6. A person who does lagic is a lagician.7. Mary krings her cat. He is such a kringable cat!8. He is always granging things. He is full of grangination.9. He soamed my car yesterday. He is a good soamer.
10. My friend usually prells very well. He is a good prellian.
the tasks were prepared in PowerPoint and answer sheets were createdfor the children. The teachers task was to project the slides, to ask thechildren to fill in their answers, to discuss the answers in pairs, and,later, to present the pairs answer to the class. Feedback on specificresponses was provided using the slides. Feedback on more generalissues was to be provided by the teacher throughout the tasks. Forexample, one of the tasks used involved the distinction between -ianand -ion. The item was projected on the screen (for example, in theanalogy game, the word pair protectprotection appeared on thescreen (see Box 3.2). The children were asked to provide the missingword in the pair infect________).
When the children had written the word on their answer sheet, the teacher activated the feedback on the computer. The children then discussed why -ian or -ion was correct in that word and theteacher provided further feedback. The children were also expected todiscuss how the base forms ended and how the last consonant in thebase form related to the middle consonant in the derived form (forexample, protect ends in -t and this is the same consonant inprotection; confess ends in -ss and this is also the consonant in confession).
Session 1 contained four activities, the first two focused on wordclasses. Activity 1 is exemplified in Figure 3.6. Activity 2 presented the
From the laboratory to the classroom 97
Sample of items from the spelling test showing onechilds answers
children with four words, all of which could belong to the samegrammatical category, and the childrens task was to find more wordsof the same category. The choice of words was important as we neededto avoid having more than one item that could belong to more than onecategory. Pictures were used to help make the meaning of words clearerand to define the category: For example, book can be a noun or a verb,but the ambiguity disappears if a picture of a book is used to define the meaning of the word. Occasionally, there were words that couldbelong to more than one category even if presented with a picture: Forexample, the word dance with a picture of someone dancing can stillbe treated as a noun or a verb. The children would have to use otherwords from the list to decide which of the possibilities would be rightfor that item (for example, learn and sing can be verbs but not nouns,so dance in this case would have to be treated as a verb). Theseexamples led to interesting discussions when the children noted that one item could fit into more than one category but not if the set of words was considered. The last two activities in this session required the children to choose the correct suffix to form person words (-er, -ian, or -ist, Activity 3) and for abstract nouns (-ness or -ion,Activity 4).
Session 2 contained further exercises on word class, choosing theappropriate suffix for abstract nouns (including -ment, -ness, and -ion), and using prefixes that make negatives (in-, un-, and dis-). The discussion of suffixes that change the meaning of words inthis way also led to interesting discoveries by the children: For example,the combinations unarmed and disarmed as well as uncover anddiscover are both possible but have slightly different meanings.
Session 3 included exercises on identifying the stems of words (seeone example in Figure 3.9, which was preceded by an exercise with realwords and a variety of affixes), counting morphemes and discussingwhen the different suffixes for person words are used (-er and -orwith verbs, -ian and -ist with nouns and adjectives).
Session 4 covered a comparison of words ending in -ian and -ionby starting from the childrens knowledge that -ian is for persons andasking them to attempt to classify -ion words. These were classifiedas feelings (for example, satisfaction), eventsor, as some childrensaid, things that happen(for example, infection, confession),mental processes (for example, imagination, addition), etc. Thediscussion can come to a notion that these are not concrete things. Thissession also included trying out different suffixes and prefixes with somestems to see how many words could be formed; adding -en to see
98 What does the research tell us?
what word class is formed (see Figure 3.8 for examples) and prefixesthat give a clue about numbers (Figure 3.7).1
Session 5 included activities that required the children to choose the right word form for a sentence when all the words had the same stem (for example, happy, happiness, happily), reinforcing theconcept of word class. It also included transformations from base formsto abstract nouns using a variety of endings (for example, intelligence,distance, ability, misery) and a discussion about the fact that theending -y has different functions (for example, in thirsty and funnyit is used to form adjectives, and in misery and poverty to formabstract nouns) but -ly is only used for adverbs.
Session 6 contained exercises of fitting the appropriate words from alist of words with the same stems into sentences, building words withstems and affixes, and comparing the effect of adding -less and -fulto words.
Session 7 contained exercises of forming adjectives (with -y, -al,and -able), adding prefixes to change the meaning (re- and negativeprefixes) and some revision exercises such as adding -en to form verbsand different endings to form abstract nouns.
The results of the third study
There were some differences between the taught and the control groupsat pretest: The control group performed significantly better than thetaught group both in spelling suffixes in words and in pseudowords.For this reason, the statistical analysis used to assess the effectivenessof our teaching program was analysis of covariance, which controlsstatistically for the differences that existed at the beginning between thegroups.
Figure 3.10 shows the adjusted means at pretest and for both posttestsby group for the correctness of spelling suffixes in real words. Becausethese are adjusted means, the means at pretest are set as the same for both groups. In this way, it is possible to make direct visual compar-isons between the estimated means for the posttests, controlling for the pretest performance. The figure shows that both groups madeprogress between the testing occasions. However, the taught group
From the laboratory to the classroom 99
1 The prefixes en- and in- are often spelled as em- and im- when the subsequentletter is p or m. We prefer not to distract the children from the classification targetand do not spend time discussing this at length, but the children may note this and askabout it.
made significantly more progress than the control group from pre- toposttest. Although there was no decrement in the taught groups perfor-mance between the immediate and the delayed posttest, the controlclasses that participated in the delayed posttest showed an unexpectedimprovement between the two posttests. This result was not expectedas there is no systematic teaching of suffixes in the literacy hour cover-ing the range of words that we included in our spelling assessment.However, it is quite possible that the teachers coincidentally gave theirchildren practice in some of the words in our assessment and that thishad a positive effect on their performance.
If this is the case, we should observe a different pattern of results inthe spelling of suffixes in pseudowords. We think that the spelling ofsuffixes in pseudowords is the acid test that shows that the childrenunderstand the principle of analyzing words into morphemes and thusare able to spell the suffixes in pseudowords, although these pseudo-words were not practiced in the classroom. The results of the analysisof spellings of suffixes in pseudowords are presented in Figure 3.11.These results show that the children in the taught group made animprovement from the pretest to the immediate posttest and maintainedthis improvement in the spelling of suffixes in pseudowords, just as theyhad maintained the improvement in the spelling of words. The statisticalanalysis showed that at both posttests the taught group performedsignificantly better than the control group, and that the control groupmade negligible improvement across testing occasions. Therefore, we
100 What does the research tell us?
Taught group Control groupNum
Figure 3.10 Adjusted means at pretest and for both posttests by group forthe correctness of spelling suffixes in Study 3.
Note: Effect size for immediate posttest = 0.42 SD. The difference between groupsat the delayed posttest was not significant.
can conclude that the unexpected improvement in performance in thecontrol group at the delayed posttest is likely to result from specificpractice. The evidence suggests that the taught group learned how touse principles of morphemic analysis in order to spell rather than simplylearned how to spell particular words that contain particular suffixes.
From the laboratory to the classroom 101
Taught group Control group
Figure 3.11 Adjusted means at pretest and for both posttests by group forthe spelling of suffixes in pseudowords in Study 3.
Note: Effect size for immediate posttest = 0.65 SD; for delayed posttest = 0.42 SD.
Summary and conclusions
The research presented in this chapter makes unique contributionsto the understanding of the development of childrens spelling. The first study demonstrated the importance of making childrenexplicitly aware of suffixes and their meaning in order to help themdistinguish spellings that cannot be distinguished from the way thewords sound. It is not sufficient to know that words than end in thesound /un/ are sometimes spelled with -ian and sometimes spelledwith -ion, it is also important to be aware of the basis upon whichit is possible to predict which of these two spellings is correct. Thiscan only be accomplished if the children become aware of themeanings of these derivational suffixes. As discussed in Chapter 2,children have to learn the forms used in spelling and also theirfunction. If their function is to represent units of meaning, they have
102 What does the research tell us?
to make a connection between the letter strings -ion and -ianand their meanings.
The second study is a very important next step in the constructionof evidence-based practice: The study was carried out in theclassroom and replicates in detail the results that we observed in thelaboratory-like situation, when the children were taught in pairs byresearchers, outside the classroom. In the second study, the childrenwere taught by their teachers in their classrooms. Although astringent random assignment was made to the taught and controlgroups, and although the children in the control group had anequivalent amount of additional literacy experiences and exposureto written text, the taught group made significantly more progressthan the control group.
The third study also makes a unique contribution to the knowl-edge available on the teaching of spelling. We developed a teachingpackage that utilized a variety of suffixes and prefixes to show howthese morphemes can be used to analyze words and how they relateto the meaning of words. The trial of this teaching program showedthat it can easily fit with the regular literacy instruction that is offeredin schools and that it is readily understood by teachers, with aminimal amount of training. The use of a waiting-list model to assignthe classes to taught and control conditions replaced the randomassignment of the children used previously and ensured that allteachers were highly motivated to collaborate with the research.The children in the control group did make progress in word spelling,but there is no evidence that they had learned much about the useof morphemic principles for analyzing and spelling words. Thechildren in the taught group made significantly more progress inspelling suffixes in pseudowords, our acid test for the learning ofgeneral principles of morphemic analysis, and this improvement wassustained in a delayed posttest, even though the teachers were nolonger using activities from our program.
These results encouraged us to move to the third phase in thiswork, which was to make the teaching program available to teachersin a wider way. In the classroom studies that we have just described,the teachers, who had collaborated with us on other occasions,adopted the methods that we had designed very closely, and wewere always at hand to help them to do this. We are very grateful
From the laboratory to the classroom 103
that the teachers decided to follow our methods closely for thereasons that we gave earlier, but we realize that other teachers inother schools who adopt these methods in the future would notapply them in one prescribed way.
We hope that many teachers will adopt these new games andexercises and new ways of measuring their pupils progress, but weknow that they will do so in different ways. The amount of time thatteachers will spend on each game, the distribution of the lessons, theway they organize the classroom (with children working bythemselves or in groups), how much explicit teaching they willprovide and how much discussion among the children they willstimulate will vary from teacher to teacher, and quite rightly so, forthe teachers will be the best judges of these details. So, we decidedthat in our next classroom studies we would look at how well ourprogram worked in schools when teachers had complete freedomabout how and when to administer the new program. We wantedto establish that the program still worked with this degree ofvariation and flexibility, and so in the classroom studies that we shalldescribe in the subsequent chapters, we encouraged the teacherswho used our teaching program to do it their way.
The results of the studies that we have described in this chapterencouraged us in another way too. Their success led us to think ofother possible benefits of teaching children about morphemes.Vocabulary was the obvious next step, for reasons that we have givenin the first two chapters, and so we set ourselves another question:Could our methods of teaching children about the morphemicstructure of words also help them to learn new words? Most of the new words that schoolchildren learn are derived words, as wementioned in Chapter 1. If children know more about how manywords are derived from other words, perhaps they will be able to addnew words to their vocabulary more rapidly and effectively as aresult.
The results of the classroom studies where teachers used theirown approach in implementing the teaching program we designedare presented in the two chapters that follow. Chapter 4 focuses onspelling measures as outcomes and analyzes a variation across twotypes of program using the same principles. Chapter 5 considers theimpact of the teaching program on vocabulary measures.
An intervention program for teaching children about morphemes in the classroomEffects on spelling
Overview of the classroom spelling project
The classroom project included 201 children in Year 5 and consisted ofa spelling pretest, followed by an intervention program, and then animmediate posttest, which was the same as the pretest.
The teaching intervention was always delivered in the childrensclassroom by their teacher. We simply gave the teachers a CD-ROMcontaining all the intervention tasks and some brief instructions on howto work with a sample of the tasks. The tasks were organized as a set of seven sessions, each lasting about 50 minutes. The teachers were
In the previous chapter, we saw that teaching children aboutmorphemes has a strong and positive effect on their ability to analyzewords into morphemes and to spell their suffixes. This seems to applynot only when children are taught on a one-to-one basis or in small groups but also when the teaching is delivered to the wholeclass.
However, in the studies we have discussed so far, the teaching wasalways delivered by researchers, unaffected by the daily pressures of school life, or by teachers working in close connection withresearchers. In this and in the following chapter we will describe aproject that assessed how our ideas work when used by teachers inreal-life classroom settings and as part of their normal teaching. Thischapter concentrates on the effects that the teaching program hadon the childrens ability to spell.
Authored by Freyja Birgisdottir, Terezinha Nunes, Ursula Pretzlik, Diana Burman,Selly Gardner, and Daniel Bell
free to deliver the intervention following their own pace, depending ontheir teaching schedules. Thus, while some of them went quite rapidlythrough the teaching program, others preferred to have longer intervalsbetween sessions or split some of the sessions into two or three parts.Our only request was that the teachers would complete all the sessionsand deliver them in the order we had planned.
The teachers and children who participated in the study came fromfive different schools; three were in London and two in Cheltenham. Wedivided the children in these schools into three groups: Two groupsreceived an intervention program and the third was an unseen controlgroup. It was not possible to develop activities for this control group aswe had done in our previous studies.
The children in the intervention groups, all of whom were pupils at the three London schools, received one of two versions of the inter-vention. One version was the same as that used in the intervention study described in Chapter 3 and concentrated both on enhancingchildrens awareness of morphology and on promoting the link betweenmorphology and spelling. We refer to this intervention group as themorphemes-plus-spelling group.
The other version contained the same tasks but concentrated onenhancing the childrens awareness of morphemes in oral languagewithout placing emphasis on the connection between morphemes andspelling. We call this intervention group the morphemes-only group.We developed this second form of the program because we wanted tosee how teaching children explicitly about morphemes affects theirability to spell morphologically complex words when they have to workout the spelling principles more or less for themselves. We will describethe difference between the two versions of the intervention in moredetail later on in the chapter.
The control group comprised children from two schools in Cheltenhamand one of the London schools; in the latter, the children were randomlyassigned either to the control group or to the morphemes-plus-spellingintervention. Table 4.1 gives the details of the childrens age at the timeof the posttest. Although the differences between the groups were small,they were statistically significant. So our statistical analyses will controlfor these age differences.
All the schools that participated in this study were inner-city schools.However, at the beginning of the project, the vocabulary and literacyskills of the children turned out to be different across the groups.Consequently, we will also need to control for these differencesstatistically.
Intervention program: Effects on spelling 105
The intervention sessions
We need not describe the general pattern of the intervention programin detail here because we have already set it out in Chapter 3. In short,each session in the program in the CD-ROM given to the teachers wasdivided into three or four different tasks, each of which was presentedin a context of a word game. Every task had a special educationalcontent, such as learning about word classes, how to break words intomorphemes, or how different morphemes influence the meaning andthe grammatical status of words. The tasks also involved a number ofdifferent cognitive operations, such as counting the number of mor-phemes in a word, adding prefixes and suffixes to a stem to make a newword, or working with analogies. We prepared the tasks in PowerPoint,so that they could be projected on a screen or on a whiteboard in frontof the whole class. The children doing these tasks typically work in pairsor small groups and produce their answers either orally or in writing on their work sheets. We took special care to make the tasks enjoyablefor the children and prepared them in a way that would encouragediscussion between them.
We used two different versions of the intervention in this study. Thetwo versions contain exactly the same tasks as each other, but the stressgiven to spelling in each task differs between the two versions. Themorphemes-only version was designed primarily to develop childrensawareness of morphemes in spoken language and does not make anexplicit link between morphology and spelling. Thus, although thechildren have to write down the words when they produce theiranswers, the emphasis in this version is on discussing how morphemesare used to form words and how they influence meaning and gram-matical categories.
The morphemes-plus-spelling version, in contrast, makes the connec-tion between morphemes and spelling an explicit one. The children gothrough the same tasks as in the morphemes-only intervention, but are
106 What does the research tell us?
Table 4.1 Mean age in years (and standard deviation) by type of group
Group Mean Standard deviation N
Control 10.2 .31 75Morphemes only 10.1 .33 26Morphemes plus spelling 9.9 .34 100For the total sample 10.00 .35 201
always encouraged to write down the words they are working with andto discuss the meanings as well as the spellings with their peers andwith their teacher. The tasks are also often followed by a discussion ofa specific spelling rule (for example, that words that end in the suffix -ful are always spelled with a single l and words that end in thesuffix -less are always spelled with a double s). The central goal isto make the children aware that there is an underlying system in thespelling of morphemesthat the same morpheme tends to be spelledthe same, even if the sound of that morpheme changes from word toword. This is often useful in the case of stem spellings because stemsoften change their pronunciation when a derivational morpheme isadded and the stress pattern changes. For example, in the word magic,the stress is on the first vowel and it is well articulated; in the wordmagician the stress is on the second vowel and the first vowel is nowa schwa vowel, not clearly articulated. Further, the last consonant in theword magic is clearly articulated; when it becomes a middle consonantin the word magician, it represents the sound which we wouldnormally spell with sh.
Box 4.1 shows two examples of the contrast between the morphemes-only and the morphemes-plus-spelling activities. In one of them, forinstance, the children saw a word pair, such as magic and magicianon the screen and then a third word, such as electric, appearedunderneath (see Box 3.1 for a description of this task). In both versionsof the intervention, their task was to use the first-word pair as a clue tofigure out what the missing word should be (in this case electrician).The children in both groups had the opportunity to find out that thereis a morpheme, -ian, that is used to form person nouns, and a differentone, -ion, that is used for nouns that do not refer to the person.However, the children who received the morphemes-plus-spellingintervention discussed, at the end of the task, how they could work out the different spellings of the words, thus adding an emphasis to the idea that there is a connection between morphemes and spelling.They were presented with a list of all the words that had appeared inthe session and were encouraged to find the stems and endings, todiscuss the fixed spellings of these endings, and to describe how theymight teach a friend to spell these words. The main challenge is to make the children understand that the best strategy in spelling these morphologically complex words correctly is to think about howthey relate to the other words on the screen. For example, they couldfocus on the fact that the word electrician is derived from the wordelectric and therefore has to be spelled with a c in the middle, or
Intervention program: Effects on spelling 107
108 What does the research tell us?
Examples of suggestions for discussion used to focus onspelling used with the morphemes-plus-spelling group,which were added to the basic activities in themorphemes-only group
Example A (see Box 3.1 for a description of this task)
Person words end with: Abstract nouns end with:
musician confessionlibrarian protectionHungarian discussiontechnician subtractionvegetarian imitation
Note: The suffixes were highlighted in red. The children were invitedto create further examples of words spelled with the same endingsand discuss how they might teach a friend a good method forspelling these words correctly.
In the task common to both groups, the children had to makeabstract nouns from different words. The abstract nouns in the task were: sickness, misery, intelligence, thirst, ability,fairness, greed, foolishness, pity, mercy, patience,untidiness, similarity, beauty.
At the end of the task, the children in the training group thatfocused on spelling was assigned one further task: To find the suffixesthey had used and find other abstract nouns spelled in the sameway. They were also asked whether they could find another suffix forabstract nouns that did not appear in this task.
on the fact that the word electrician is a person word, just like the word magician, and should therefore be spelled with an -ian at the end.
The underlying principle of this and of other tasks used in theintervention is that if children understand the morphological structureof words and are able to use that knowledge when they spell, they oftenhave a much greater chance of succeeding than if they spell wordspurely on the basis of sound or from memory.
The pretest and posttest assessmentsWe used the same two spelling tests to assess the childrens progress as a result of the intervention as we had in the previous classroomintervention study (Study 3) described in Chapter 3, Table 3.3. The firsttest included twenty-six two- or three-morpheme words, each con-taining either a prefix or a suffix. As we showed in Chapter 3, the wordscannot be spelled correctly purely on the basis of sound. For example,in the word election, both the suffix and the medial consonant arespelled differently from the way they sound. The same applies to thewords politician and destination. Words such as happiness andfearless seem very simple, but why should their endings have ssrather than a single s? Words like inflatable and comfortablecontain a schwa vowel in the suffix, which leaves children in doubtabout the spelling. Of course, the reason is that these are morphemeswith fixed spellings.
We used the same method in the spelling assessments as the one thatwe described in previous chapters. The children heard each word that they had to spell first in a sentence and then again on its own. Thesame sentence was also provided on an answer sheet in front of them,with the target word replaced by a blank line. The childrens task wasto write the missing word on the line. So, for example, when they hadto spell the word musician, we read out to them, My sister wants tobe a musician. Musician.
The second spelling task was similar to the first one, except that now we asked the children to spell pseudowords. There were ten ofthem in total and each was composed of a made-up stem and a realsuffix (for example, mape-ful, lagic-ian). The pseudowords werealways presented in two different forms and were always presented in a sentence context. For example, when the children had to spell thepseudoword lagician they heard, A person who does lagic is alagician. Lagician. The children then had to write lagician on theiranswer sheet.
Intervention program: Effects on spelling 109
Our reason for including this test is that pseudowords cannot bespelled from memory and provide, therefore, an excellent way ofchecking whether children really use their morphological knowledgewhen they spell novel words. The test also enables us to see whether thechildren were able to preserve the spelling of unfamiliar stems acrossdifferent word forms, even when the sound of that stem changes (forexample, from magic to magician).
Our main hypothesis was, again, that teaching children about mor-phology improves their ability to spell, especially when the connectionbetween spelling and morphemes is explicitly pointed out to them. Wepredicted, therefore, that the two intervention groups would make moreprogress in spelling between pretest and posttest than would the controlgroup. We also predicted that the performance of the children whoreceived the morphemes-plus-spelling intervention would improvemore between the two test sessions than would the performance of themorphemes-only group.
As mentioned before, there were differences between the groups at pretest. In the pretest, the control children were already able to spell 65 percent of the suffixes correctly on average, whereas the children in the two intervention groups were only able to spell approximately 56 percent of the same suffixes correctly. So, as we did in Chapter 3, wedealt with this discrepancy between the groups by using analyses ofcovariance, which control for the differences between the groups atpretest statistically.
Figure 4.1 summarizes the childrens performance in spelling thesuffixes in real words before and after the intervention. Because themeans displayed are the adjusted means, controlling for differences atpretest, it is possible to make direct visual comparisons between thegroups. As you can see in Figure 4.2, where we compare the childrensaverage spelling scores before and after the intervention, all threegroups seem to have made some progress. This was to be expected,since all the children had taken the same test twice, with a relativelyshort time interval between them. However, the two intervention groupsimproved significantly more than did the control group. Thus, our mainhypothesis that teaching children about morphemes enhances theirspelling skills was supported.
In contrast, our prediction that the morphemes-plus-spellingintervention group would show more improvement than would the
110 What does the research tell us?
morphemes-only group was not confirmed by the results. The twogroups made similar progress between the pretest and posttest in bothspelling tests. This came as a surprise to us, especially since our earlierfindings indicate that teaching children directly about the link betweenmorphemes and spelling helps them to spell morphologically complexwords (see Chapter 3).
In this assessment, we considered only whether the children hadspelled the suffixes correctly. We thought that perhaps the morpheme-plus-spelling group would have an advantage over the morphemes-onlygroup if we considered the spelling of the whole words. We reasonedthat they would have had more opportunity to think about why words such as confession and education have a different spelling for the consonant that precedes the -ion ending, because they had the opportunity to discuss this more explicitly than the children in
Intervention program: Effects on spelling 111
Figure 4.1 The adjusted mean scores on the test of spelling suffixes inwords (out of a maximum of 26) on each testing occasion foreach group.
Note: The overall effect of the intervention was significant (p
the morpheme-only group. In order to obtain a score that reflected thechildrens use of morphological information in detail, we divided the words into stem and suffix and gave the children one point for each of these segments. We separated words such as magician anddestination, where the middle consonant represented an addedhurdle, into three segments. One was the stem up to the final consonant,the second segment was the final consonant in the stem, and the thirdthe suffix (for example, magi/c/ian and destina/t/ion). Eachsegment was awarded one point if correctly spelled. Box 4.2 shows a fewexamples of how we divided and scored the words. If you look at theword politician for instance, you can see that the children receivedthree separate points for this word: one point for spelling the stemcorrectly, another for spelling the final /sh/ sound of the stem correctly(that is with a c), and then a third point for spelling the suffix correctly.By contrast, the children only received two separate points for words
112 What does the research tell us?
Figure 4.2 The adjusted mean scores on the test of spelling polymor-phemic words (out of a maximum of 61) on each testingoccasion for each group.
Note: The overall effect of the intervention was significant. Both intervention groupsdiffered significantly from the control group, after we controlled for pretest differ-ences and the age differences at posttest. The effect size for both groups was 0.2SD. The difference between the two interventions was not statistically significant.
like comfortable and politeness: One for the stem and one for thesuffix.
We then ran a similar analysis as the one described previously, butthis time we used the childrens scores for correctly spelling the wholewords in the test, rather than the scores for the suffixes only. Figure 4.2shows the results of this analysis.
This second analysis largely confirms the results of the previous one: There was a significant effect of the intervention because bothintervention groups made more progress than the control group frompretest to posttest. The two interventions did not differ significantlyfrom each other.
The effect of the intervention on the childrens ability to spell suffixesbecomes even more apparent when we look at how they spelled indi-vidual words before and after the intervention. Box 4.3 shows how oneboy in the morphology-plus-spelling group spelled the first twelve of the twenty-six words in the real-word spelling test. If you look at hisresponses in the pretest you can see that his primary strategy seems tohave been to spell the words according to the way they sound. Thus, hespelled magician as magishon, musician as musison, richnessas richnes, and disability as disabilatey. His phonological strategydoes not even seem to have been entirely consistent. For example, hespelled the suffix of magician and musician differently, despite thefact that these sound exactly the same. The same applies to the suffixof richness and politeness.
However, if you look at how he spelled these same words in the posttest you can see that, although his spelling was not perfect, his
Intervention program: Effects on spelling 113
Examples of the segmentation used in scoring on theword- and pseudoword-spelling tests
Words PseudowordsMagi/c/ian Grat/lessElec/t/ion Blage/nessPoliti/c/ian Saught/i/nessHapp/i/ness Senk/t/ionComfort/able Prell/ianPolite/ness Kring/able
approach had completely changed. He stopped relying so much on howthe words sound and he seems to have understood that some words aremade up of different parts and that these parts tend to be spelled thesame from word to word. This means that he is now able to spell difficultsuffixes such as -ian, -ness, -ity and -able correctly in differentwords. In fact, the percentage of suffixes he spelled correctly rose from27 percent in the pretest to 81 percent in the posttest.
114 What does the research tell us?
A sample of the same boys spelling in the pretest and posttest
Although this particular boy may be an extreme example of thesuccess of the intervention, the same basic trend always appeared whenwe examined the spellings produced by the other children in the twointervention groups, both in the spelling of words and of pseudowords.Thus, they started off spelling the suffixes more or less phonologically,but, after receiving the intervention, they became less reliant on thisstrategy for spelling suffixes and started to use their newly acquiredmorphological knowledge.
As discussed in the previous chapter, we think that the critical test ofchildrens learning of a new strategy, rather than the learning of specificword, is their performance in the spelling of suffixes in pseudowords.
Intervention program: Effects on spelling 115
Because these were not included in the teaching intervention, progressin spelling suffixes in pseudowords indicates the acquisition of a spellingprinciple, rather than the learning of specific words. We ran an analysisthat controlled for differences in pretest scores in order to compare theintervention and control groups. The results of this analysis are sum-marized in Figure 4.3. They confirm the results of the previous analysesby showing significant effects of both interventions and extend theseresults to support the idea that the children were learning a spellingprinciple not only the spellings of specific words.
To summarize, the children who received the two versions of theintervention made significantly more progress between the pretest andposttest than did the children who received no teaching in morphology,both in the spelling of suffixes and in the spelling of morphologicallycomplex words whose spelling is not predictable from the way they
116 What does the research tell us?
Figure 4.3 The adjusted mean scores on the spelling of suffixes in pseudo-words (out of a maximum of 12) on each testing occasion foreach group.
Note: The overall effect of the intervention was significant (p
sound. Most importantly, without an exception, this difference betweenthe intervention and the control groups appeared irrespectively ofwhether the children had to spell words or pseudowords. This meansthat the greater improvement of the children in the two interventiongroups cannot be explained in terms of them simply getting better atspelling specific words from memory.
Does the intervention help both low and highachievers?
One question that we are asked often is whether the interventions wedesigned are effective both for high and low achievers. Some teachersmake the interesting suggestion that our intervention should work bestwith high achievers, because these are difficult concepts, which mightgo above the head of low achievers. Other teachers hold the view thathigh achievers cannot learn much from a teaching intervention thatfocuses on morphemes, because they are likely to get there on theirown, without explicit teaching.
Picking up on the teachers questions and hypotheses, we decided toanalyze our data in search for an answer. Because we had a relativelylarge number of children and they were all in the same year group atschool, this study offered a good opportunity for analyzing the effectsof the intervention with low and high achievers. We did so by splittingthe children into two groups according to their pretest scores: Thosewhose scores were up to the median were placed in the lower group and those who scored above the median were placed in the highergroup. The groups differed considerably in two ways: They had, ofcourse, different means, but they also had very different degrees ofvariability. The lower scoring group showed a significantly largerstandard deviation than the high scoring group because there was alarge tail of children with very low scores in the group classified as belowthe median at pretest. It should also be remembered that the numberof children in each group is now, by definition, half of what it was inthe previous analyses. So, weak effects would no longer be significant.
Figure 4.4 presents the results of the analysis of the childrens scores in spelling suffixes in words. Both interventions had a significantimpact on the scores of children who scored below the median at pretest.For the children who scored above the median, the morphemes-plus-spelling intervention was more effective, although there was a significantoverall effect of the interventions. Our intervention, therefore, worksacross the ability range with high achievers and with low achievers.
Intervention program: Effects on spelling 117
These results also suggest that teaching about the explicit connectionsbetween morphemes and spelling is relatively more effective with themore advanced children, who probably already have some awarenessof morphemes.
We ran the same analysis with the spelling of suffixes in pseudowords.In the analysis for children who scored below the median at pretest, thedifference between the performance of the morphemes-plus-spellinggroup and the control group just fell short of significance (p=.07), but
118 What does the research tell us?
Figure 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 thepretest.
Notes: The results on the top graph are for children who scored below the medianat pretest (N=93) and those on the bottom graph are for children who scored abovethe median at pretest (N=86). The overall effect of the intervention was significantfor both groups. Both intervention groups differed significantly from the controlgroup for the children who had low scores at pretest; only the morphemes-plus-spelling group differed significantly from the control group for the children who hadscored above the median in the pretest.
this difference was not significant for the below-median children in themorphemes-only group. For the children who scored above the medianat pretest, both types of intervention had a significant effect.
Thus, both types of intervention helped the low and high achieverswith real words. With the pseudoword spellings, both types of inter-vention improved the high achievers spellings, but only the morphemes-plus-spelling intervention helped the low achievers.
Intervention program: Effects on spelling 119
Summary and conclusions
The central aim of this classroom project was to explore whetherinstruction about morphology in the classroom increases childrensability to use morphemically based spelling principles. Much to our delight, the results of the project strongly indicate that it does.Almost without an exception, the spelling scores of the children whoreceived the morphology intervention improved much more thandid the spelling scores of the children who received no extrainstruction in morphology. This indicates that the intervention wassuccessful in its specific aim to increase childrens ability to use theirknowledge of morphology when they spell.
Thus, we can now safely conclude that the effect of morphologicalinstruction on childrens ability to spell is not limited to tightlycontrolled experimental settings but also applies when the instruc-tion is delivered by teachers in busy classrooms and as part of theirday-to-day teaching schedule.
However, there is one finding that we have yet to explain, and thatis why the children who received the morphemes-plus-spellingintervention did not improve more than the children who receivedthe morphemes-only version.
Because of the nature of this study, it is difficult for us at this pointto come up with a definite explanation. Since we wanted to see howthe morphological instruction worked in real-life classrooms, itwas important to us that the teachers should deliver the interventionsin their own way, without any assistance from researchers. However,the consequence is that we have no way of knowing what exactlywent on in the intervention sessions or the extent to which theteachers followed our advice of how to use the tasks. This is ofparticular importance when we compare the success of the twoversions of the intervention, because the central difference between
120 What does the research tell us?
them lies not in the construction of the tasks themselves but in howthey are implemented. For example, it was really up to the teacherswho used the morphemes-plus-spelling intervention whether theyconcentrated mainly on the childrens awareness of morphemes inspoken language or whether they took the extra step of emphasizingthe link between morphemes and spelling. It is possible, therefore,that the difference in the actual implementation of the two inter-ventions was not so great after all.
The only way to determine whether the two teaching regimesreally influence childrens spelling skills in different ways is to take onestep back and to try them out in a more tightly controlled environ-ment where we can observe how they are actually implemented.This will have to be left for the future, because at this point in ourresearch we decided to focus on a different outcome of childrenslearning about morphemes, their vocabulary growth.
An intervention programfor classroom teachingabout morphemesEffects on the childrens vocabulary
We have now described, in Chapters 3 and 4, how our ideas aboutteaching morphemes fared when we took the path to the classroomand asked schoolteachers to use our methods with whole classes ofchildren. In this chapter we shall describe what happened when we took the same path with our idea that teaching children aboutmorphemes should help them to learn new words as well.
We had three reasons for investigating whether teaching childrenabout morphemes has a positive effect on their vocabulary growth.The first is that various theories (see Chapter 1) suggest that growthin vocabulary is a possible outcome of enhancing childrens aware-ness of morphemes. Morphemes are units of meaning, and most ofthe words that children learn from the middle of primary school onare polymorphemic words. So, if children have a way of analyzingthese words, they might find them easier to learn. The second reasonis that vocabulary is an important part of literacy learning: Childrenstext comprehension is highly related to the size of their vocabulary.The third reason is also important: Research on how to teach childrenabout vocabulary has dealt predominantly with the question of howto teach specific new words. This research focuses mainly on howmany repetitions are required and whether it is best to get thechildren to encounter new words in a text or in isolation, to providethem with definitions or with different sentences in which themeaning and use of the word is illustrated. Our own idea is differentto this. We argue that teaching children word-attack strategies basedon morphemes should help them to analyze new words in order to understand their meaning. We carried out the next study to testthis idea.
Authored by Terezinha Nunes, Peter Bryant, Ursula Pretzlik, Diana Burman, Daniel Bell, and Selina Gardner
The context of the study
The teaching in this study was done by schoolteachers who wereparticipating in a broader project that aimed at improving childrensspeaking and listening skills in the Hillingdon Cluster of Excellence, aprogram run in a cluster of schools in Hillingdon (West London) andsupported by the Department for Education and Skills (DfES) in theU.K. We will refer to this project as the Speaking and Listening Skillsproject.
All schools participating in the project did so voluntarily. The projectcontinued for two years. In the first year of the project, we approachedthe teachers and discussed two possibilities with them. One was thatthey would design their own interventions for improving childrensspeaking and listening. The other possibility was for them to sign up fora project, designed by us, which aimed at improving childrens vocab-ulary by making the children more aware of how morphemes are usedto construct words.
Of the nine schools that participated during the first year, fourdesigned their own interventions, none of which involved teachingchildren about morphemes. The remaining five schools opted for themorpheme-teaching program that we had designed. All participatingteachers agreed that the children in their projects would be assessed onmeasures that would be relevant for evaluating the different teachinginterventions that had been designed. Thus, the children who wentthrough our own morpheme-teaching program were the interventiongroup in this study, whereas those who received some form of interven-tion that had been chosen by the teachers themselves became the controlgroup in the study.
The first year was a development year. The teachers who had createdtheir own interventions sharpened their ideas by becoming clearer aboutwhat they thought they could achieve with their teaching interventionand developed expertise in evaluation. They became more aware of theimportance of finding ways of assessing the outcomes of their teachingand developing an evidence-based approach to the teaching. We had the opportunity to create measures that the teachers felt were fairassessments of their aims, to test these measures, and to analyze theirreliability.
The results reported here are from the second project year, when allteachers were implementing interventions that had been developed andassessed in the first year.
In the second year, some new schools joined the program; twelveprimary schools now participated in the Speaking and Listening Skills
122 What does the research tell us?
project. The teachers who had already been involved in the project had the option of continuing to work with the intervention that they hadimplemented in the previous year or of moving on to something new.The new teachers joining the project found out about all the ongoingteaching programs and had the chance of choosing a program for theyear. Seven schools chose to work with the morpheme intervention andfive worked with other intervention programs. Once again, the partici-pating teachers agreed to collect assessments that would allow for anevaluation of all the teaching programs being implemented, and eachgroup acted as a control for other interventions. So, all the childrenwere receiving some intervention, their teachers had opted for workingwith the teaching program that they were using, and the schools werein the same region of the country. Here, as in most of our studies, thecontrol children were not losers in a research program.
The teachers who used the morphemes intervention that we designedreceived a CD-ROM with the activities and a brief induction to the useof the intervention materials. This took the form of a session duringwhich the researchers showed some of the activities to the teachers,explained their aims, and gave the teachers the opportunity to askquestions. It was made clear to the teachers that they could implementthe program in the way that they thought would suit their class best. Thedistribution of sessions over time and whether the children worked inpairs or small groups were all decisions made by the teacher. They wereasked only to ensure that all activities were implemented in the orderthat they appeared in the CD-ROM and to encourage the children todiscuss their responses, because the program aimed to increase thechildrens awareness of morphemes and speaking about their ideaswould help enhance this awareness.
The design of the study
The children in the project were either in a morpheme-interventiongroup, which means that they were given especial instruction aboutmorphemes, or they were in a control group and did not get any extraexperiences with morphemes but had experience with one of the otherteaching programs in the project. Four teaching programs were used.Telling and acting out well-known stories, developing emotional lan-guage, analyzing poems, and describing paintings reproduced in postersobtained from the National Art Gallery in London. Table 5.1 presents asummary of the numbers of participants at pretest for the differentgroups by their year group in school.
Intervention program: Effects on vocabulary 123
The children participated in a pretest, followed by the interventionthat was described in Chapter 4 as morphemes only, which wasfollowed by an immediate posttest and a delayed posttest approximately8 to 10 weeks after the completion of the program. All tests wereadministered and scored by the researchers. Scoring was blind to thetype of program that the children had participated in.
The pretests and posttests
The pretest and immediate and delayed posttests were identical. Theycontained a vocabulary test and a pseudoword-definition test.
The vocabulary test was designed to be easy to administer in theclassroom. The children are presented with a picture booklet thatcontains the same pictures projected onto a screen by the researcher.The researcher reads a sentence, which has a missing word, and thechildren have three choices for what the missing word will be. Theresearcher says the sentence three times, each time completing it withone of the words. Each of the choices appears on the screen when theresearcher reads it and each appears in a different color so that thechildren can use color as a cue if they find the words difficult to read.The children tick on their booklet the word that they think best fits intothe sentence. Some examples of items and information about theassessment are presented in Figure 5.1.
The test contained a total of forty items. In the previous project year,we established that the internal consistency of the test, which was given
124 What does the research tell us?
Table 5.1 Number of children, mean age in years (and standard deviation) by year group in school and type of group in the project
Control Intervention Total
School Mean Mean Mean year age SD N age SD N age SD N
3 7.73 .39 71 7.83 .37 78 7.80 .38 1494 9.10 .26 38 8.90 .29 75 8.95 .30 1135 9.90 .30 34 9.85 .33 119 9.86 .32 1536 11.00 .35 35 10.73 .28 38 10.85 .34 737 11.76 .43 4 11.50 .30 9 11.62 .38 13N 182 319 501
Intervention program: Effects on vocabulary 125
She always arrived late. She was
The doctor told Georgia not to worry because the injectionwould be
Figure 5.1 A description and two sample items from the vocabulary test.
The pictures are projected on the screen. The researcher reads out the sentencewith each of the words fitting into it, one at a time. The words appear on the screenas the researcher reads them. The children tick on their books the box next to theword that they think is the correct choice.
to 207 children in this age range, according to the Cronbachs alpha is0.9. This is very high: Acceptable consistency levels must be 0.7 or more.We also gave the vocabulary test and the British Picture VocabularyScale (BPVS) to a smaller group of children (N=24) and found that ithad a high (r=0.66) and significant correlation with the BPVS. Thus,we established that this assessment is reliable and valid in the first yearof the project.
The pseudoword-definition test was designed for reasons that aresimilar to those that motivated us to design the pseudoword-spellingtest. Our program aimed at promoting childrens vocabulary growth not just at promoting the learning of specific new words. We wanted to provide them with a word-attack strategy that would allow them tointerpret polymorphemic words that they had never encountered before.A vocabulary test tells us about how many words children have learned,but it does not tell us how they have learned them: They could havelearned them by direct instruction or by getting at their meaning throughanalyzing them into morphemes. So, we designed a test that we thoughtwould get more directly at their ability to analyze polymorphemic wordsin search for their meaning. We created pseudowords by combining astem and an affix in a nonexisting combination. We presented thechildren with these pseudowords, told them that these made-up wordsdo not really exist, and asked them to try to think what these made-upwords would mean if they did exist. The instructions for the assessmentsand the list of pseudowords used are presented in Box 5.1.
The children received one point if their answer took into account themeaning of the stem and of the affixes; no points were awardedotherwise. We analyzed the reliability of this assessment in the first yearof the project by looking at the reliability of the scoring procedure andat the internal consistency of the items. The method of scoring provedreliable: Two researchers who scored the same tests independentlyshowed 98 percent of agreement in their scores for the items. The itemsshowed a high internal consistency; the Cronbachs alpha variedbetween 0.75 and 0.85 with different samples. A value of 0.7 isconsidered acceptable. Finally, this test showed a significant correlationwith our vocabulary test (r=.57), and, thus, there is evidence for itsvalidity.
For practical reasons, not all the children participated in all of theassessments, as this would have taken a very large amount of time fromteaching. So, each class in one intervention was assigned as a matchedcontrol class for another intervention, and only a small number ofchildren participated in the assessment of both outcome measures for
126 What does the research tell us?
the evaluation of the morpheme intervention. The vocabulary test wasgiven to 206 children in the morpheme group and to 121 children in thecontrol group on all three testing occasions (N=327 for this analysis).The pseudoword-definition test was given to 170 children in the
Intervention program: Effects on vocabulary 127
The instructions and the items in the pseudoword-definition task
Sometimes it is possible to make up words that seem to make sense.Although they do not exist, if we say these made-up words tosomeone else, they might think that they know just what we mean.For example, if I said to you that I saw a childreny chair, what typeof chair do you think it would be?
If the children come up with answers such as a chair for childrenor a child-size chair or a small chair, they are encouraged toexplain how they knew that. If no answer is elicited from the class,the researcher can ask whether this would be a large chair, forgrown-ups, and then ask how they know the answer.
I am now going to show you some made-up words, and your jobis to try to think what these words could mean. Write the answer inyour booklets.
For children in the third year of school, aged about 8, the test wasadministered orally and individually by a researcher; the responseswere also oral.
The pseudowords are projected onto the screen and the researcherreads them out loud. The children write their answers and theresearcher verifies that they have finished before moving on to thenext item.
List of words used in the pseudoword-definitiontask
bricker bookist rewet uncombunlie unwork biheaded shoutistunclimb triwinged resleep chickener
morpheme group and 104 in the control group on all three testingoccasions (N=274 for this analysis).
The morpheme teaching intervention
The program of instruction that we presented to the teachers wascontained in a CD-ROM, which was the same as the one used for themorpheme-only groups described in the previous chapter. To remindyou, the CD-ROM consisted of a number of tasks, all of which werepresented as games with words. We organized these games in sessionsfor the previous study, but in this study the sessions were entirelynotional ones because the teachers were free to pace the games indifferent ways. The flexibility that we allowed in the teachers use of theteaching CD-ROM meant that there was a great deal of variation in thedetails of what the teachers did and how long they took to do it. Oneteacher reported using the games as rewards when the children hadbehaved well in the literacy hour. Others used the games as whole-classactivities during the literacy hour. Only a few used the sessions as awhole once or twice a week.
The effects of the morpheme interventionon vocabulary
Our hypothesis was simply that teaching children about morphemesshould have the effect of improving their vocabulary. So, we predictedthat the children in the morpheme group would do better than theothers when their vocabulary was measured in posttests after con-trolling for the differences in the pretest.
This was what we found. As in the previous chapters, we usedstatistical analyses that control for differences between groups at pretestwhen making the comparisons between the different groups at the twoposttests. Figure 5.2 shows that all three groups made progress overtime. It can be seen clearly from the graph that the morpheme groupmade significantly more progress than the control group. This differencein progress was sustained in the delayed posttest and was statisticallysignificant.
In Chapter 4, it was seen that for children in one year group, themorphology intervention was effective both for high and low achievers.In this study, we worked with different school years and fewer classesin each year group. As a result, it was not possible to conduct an exactly
128 What does the research tell us?
parallel analysis because the children in the group scoring below themedian might be either younger than the others or low achievers.However, it is still possible to ask: Is the intervention as effective forchildren who start out with lower scores as it is for those who start outwith higher scores, independently of whether they have low scoresbecause they are low achievers or younger?
We ran the same kind of analysis as we had done in Chapter 4: Wedivided the children into a group that scored below the median and asecond group that had scores equal to or above the median in thepretest. The analysis was then run separately for these two groups. Theresults of these analyses are presented in Figure 5.3. Although therewas a difference between the morpheme and control groups in bothcases, this difference was statistically significant only for the childrenwho started out with low scores.
Thus, our program may not have stretched the children who alreadyhad a good level of knowledge of morphemes or who already had bettervocabularies at the outset. We think that this finding is actually quiteinteresting: Schools and teachers always find it difficult to designprograms that are effective for those who have accomplished less, butthe morpheme intervention is shown to be more effective in improvingthe vocabulary of exactly the children who start out with lower levelsof performance.
Intervention program: Effects on vocabulary 129
17.5Time 2 Time 3
Figure 5.2 Mean scores (adjusted for pretest differences) in the vocabularytest for each testing occasion and group (maximum score =40).
Note: The difference between the groups was statistically significant (F=5.08; p
It is particularly exciting to see in this figure that there are no signsof regression at delayed posttest; in fact, for the children who startedout with low scores, there is further progress at delayed posttest.
Similar analyses were carried out in order to evaluate the effects of the intervention using the pseudoword-definition test as the measureof progress. The aim of this measure, as we pointed out earlier, is to testwhether the children can analyze novel words in morphemes in orderto have a way of discovering their meanings. Thus it is the critical testfor us: Children who can use such word-attack strategies should be ableto learn more new polymorphemic words than those who depend onsome form of direct instruction about the meaning of each new wordthey encounter. Figure 5.4 shows the results of the analysis for all thechildren.
Both the morpheme and the control group made some progress frompretest to posttest, suggesting an effect of practice in the test. However,
130 What does the research tell us?
Time 1 Time 2 Time 3
Time 1 Time 2 Time 3
Figure 5.3 Mean scores by testing occasion and group (adjusted for pretestdifferences) in the vocabulary test for children who scored upto the median (left) or above (right) in the pretest.
Note: The difference between the morpheme and control group was significant forthe children with scores up to the median in the pretest but not for the children withscores above the median. For children who started out with low scores, the effectsize was 0.25 SD.
the control group did not progress any further from the immediate post-test to the delayed posttest, whereas the morpheme group continued to make progress. The rate of progress differed between the groups: The children in the morpheme group made significantly more progressthan those in the control group. This suggests that the children in themorpheme group learned a word-attack strategy and possibly used it ontheir own after the intervention had been concluded, because theyseemed to become better at using this strategy at a time when they werereceiving no further teaching about morphemes.
We also ran an analysis similar to that presented for the vocabularytest, where we separated out the children on the basis of their pretestscores into two groups, one which scored up to the median and thesecond which scored above the median. The results of these analyses arepresented in Figure 5.5.
The analyses showed that the children who started out with a lowscore and also those who started out with a high score benefited fromthe intervention. The differences between the morpheme and thecontrol groups were significant in both cases. There was a particularlylarge effect size for the children who started out with high scores, andthis indicates that they benefited from instruction more than those whostarted out with a low score.
Intervention program: Effects on vocabulary 131
Time 2 Time 3
Figure 5.4 Percentage of correct pseudoword definitions (adjusted forpretest differences) by group and testing occasion.
Note: The difference between the groups was statistically significant (F=17.5;p
132 What does the research tell us?
Time 1 Time 2 Time 3
Time 1 Time 2 Time 3
Figure 5.5 Percentage correct in the pseudoword-definition test (adjustedfor pretest differences) by group and testing occasion.
Notes: The graph on the left shows the results for children who scored up to themedian in the pretest; on the right are the results for children who scored above themedian in the pretest. The differences between the morpheme and control groupwere significant in both analyses (p
Intervention program: Effects on vocabulary 133
are measured through the childrens performance on our acid test,the pseudoword-definition test. The importance of sustained gainsshown in the pseudoword-definition task is that they show that thechildren learned a word-attack strategy that they can use to learnmore polymorphemic words in the future.
This takes us to the end of the work with children. This workshows, without a shade of doubt, that primary-school childrenbenefit enormously from learning about grammar and morphemes.When they participated in our interventions, they learned not onlyhow to spell and analyze specific words but also developed a moregeneral form of knowledge, based on their awareness of morphemesas elements that form words, which helps them to understand themeaning of novel words and helps them predict the new wordsspelling. The implication is that developing childrens knowledgeabout morphemes and helping them to use this knowledge todevelop their vocabulary and spelling should be an important partof the curriculum in primary school.
In the two remaining chapters, we ask the question: Is it alreadythe case that children are taught about morphemes in school? Arewe proposing something that is already part of practice but for whichthere was, so far, no solid evidence? Or is it possible that teachersand teaching programs have not yet incorporated morphemes asunits of meaning into their way of thinking about how to developchildrens literacy?
Chapter 6 considers evidence directly from practice in Londonschools. Teachers were interviewed about how to help children learnthe spelling of words whose spelling cannot be predicted from theway they sound, though it can be predicted from their constituentmorphemes. Their teaching strategies in the classroom were alsoobserved. Finally, they were made more aware of the importance ofmorphemes for literacy and had the opportunity to use our materialsin the classroom. Chapter 7 analyzes the consequences of theseinteractions for primary-school teachers and their children.
Chapter 7 contains an overview of how the teaching of spellinghas been treated in the U.K. and in the U.S.A. and the evidence thathas been provided for the effectiveness of the different ways ofteaching spelling. Future developments are discussed in light of theresearch presented throughout this volume.
Can we increase teachersawareness of morphology and have an impact on theirpupils spelling?
So far, we have discussed the fact that morphology is one of thebuilding blocks of English language and orthography. We haveshown that when children learn how to spot morphemes and to use the rules of morphology in their spelling, their spelling improves.This happens when children work as whole-class groups withteachers as well as when they learn one-to-one with researchers.Now that we know that our techniques are appropriate for classroomteaching, we need to find out teachers views and practices regard-ing morphology. In this chapter, we explore what these views andpractices are, to what extent they can be changed, and whethersuch a change has any effect on the teachers pupils.
From what we have already seen in Chapter 2, although adults are skilled morphemists, they do not seem to have a high level ofawareness of the morphemic regularities of English. Taking this a little further, in the English context, it is also possible to explore thedegree to which the teaching establishment explicitly identifiesmorphemic strategies as important aids to spelling. Since the Englishsystem of teaching literacy is centrally defined through the NationalCurriculum and the more detailed and practically oriented NLS, the place of morphology is reasonably transparent. Both of thesesets of policy documentation mention morphemes. The NationalCurriculum promotes teaching morphemes in the context of spell-ing (Qualifications and Curriculum Authority, 1999). In the NLSdocumentation, morphemes are identified as one of the principlesunderpinning word construction and are seen as having a place in teaching spelling (DfEE 1998, DfES 2001, 2003). However, as
Authored by Jane Hurry, Tamsin Curno, Mary Parker, and Ursula Pretzlik
Teachers strategies for teaching spelling
We interviewed fifty teachers from the London area to explore theirviews on spelling and, in particular, to see if they explicitly referred tomorphemes in their teaching. To try to get at their working knowledgeand practice, we asked them quite concretely about the difficulties that their pupils had with spelling and the ways in which they wouldaddress these difficulties. Teachers were told that we were interestedin the sort of difficulties that 7- to 11-year-old children have when tryingto spell and the ways teachers help them with their spelling. They were then presented with the following list of words, which illustrate arange of challenges for spellers: white, opened, pavement, base-ball, richness, motion, combination, slept, prepare, smoke,dark, uncovered. For each word, they were asked what sort of errorstheir pupils would make and what they would teach the children to helpthem correct their mistakes.
The teachers mentioned a range of mistakes that children could make in spelling. The majority were attributed to problems relating to phonology (for example, problems with spelling silent letters such as the h or the e in white and representing vowel sounds whichcould be difficult to hear unambiguously). Difficulty with letter blendsand familiarity with the meaning of a word were also mentioned.However, not surprisingly, considering our selection of spelling words,teachers also mentioned error types directly related to morphemes and,in particular, problems with past-tense -ed endings, irregular past-tense endings (slept) and prefixes and other suffixes. Overall, 1,930statements were coded. Forty-five percent (n=878) of these statementsreferred directly to problems with phonology. This is not surprising, aswe know that the dominant spelling strategy is a phonological one. Thenext most significant category was morphology, which accounted for afurther 20 percent (n=386) of teachers statements, and this was no
Can we increase teachers awareness of morphology? 135
teacher guidance becomes more detailed and practical, the place of morphemes in the teachers repertoire becomes less clear.For example, the word morpheme is not used. Arguably, it is this practical level of guidance that will have the most impact on teachers practice, but it also reveals the lack of research andtheorizing in this area available to the authors of the model lessonplans.
doubt influenced by the fact that eight of the twelve prompt words inthe interview were chosen for reasons relating to morphemes. Moreinteresting is the way in which the teachers discussed the role ofmorphemes in spelling and teaching spelling.
Despite our rigged interview, the word morpheme was conspicuousby its absence; it was not once mentioned by any of the teachers.However, teachers did refer to morphemes in other ways. The vastmajority (82 percent, n=41) talked about prefixes and suffixes. Theyalso talked about past-tense verbs in response to the prompt wordsopened, slept, and uncovered (62 percent, n=31). When teacherstalked about the -ed morpheme, they almost always linked it with a change in meaning. They explicitly taught their pupils that adding -ed changes the verb from the present to the past tense (see Box 6.1).
136 What does the research tell us?
Teachers talking about -ed endings
We wrote every word and its past participle on the same piece ofcard and then we sorted them according to how theyve beenturned into their past participle. Theyre living in (four) houses;theres the house when you just add -ed and its nice andsimple; theres the house where you dont need to add an -ebecause its already there; theres also the house where you haveto double the consonant, as in stopped or shopped, so itsdouble the consonant and then add -ed. And then on the otherside of the track is the nasty house where things like write andwrote live and bite and bit. All the irregular ones live onthe wrong side of the tracks.
Opened, they wouldnt think to put the -ed on the end,although we do do lots of work with past tense, but its drip, drip,drip. We have to constantly remind them. . . . So, for example,weve just been to the National Gallery and we wrote it as whatwe were going to do, so Tomorrow we are going to go to . . .And then we went, so the next day, we wrote On Monday
Occasionally, teachers were apparently unaware of the significanceof the -ed suffix and thus were unable to explain the spelling rule totheir pupils (see Box 6.2).
Can we increase teachers awareness of morphology? 137
we went to . . . , so we changed the whole, so its a concreteexperience.
This term we have been focusing specifically on suffixes to dowith verbs particularly, so they know that the -ed becomes thepast tense. Its got better since weve done that actually.
Lack of awareness of -ed rule
I would probably talk to them about how it is spelled and how itis pronounced and how we dont always pronounce all the lettersthat were in a word. There is a unit [in the NLS] that talks aboutunstressed vowels in words and makes a list of other words thathave similar missing sounds and keeps a bank of those sorts ofwords on display, so that we say it but its got a letter in that ismissing and then every time we need to use a word like that,dont forget the missing letter, the letter we cant hear, so justreminding them that its there.
The ones that got it wrong would put a d on the end . . .because of the way it is pronounced . . . pt arent very commonblends and can easily be pronounced with a d on the endinstead of a t so, concentrating on the way a word looks andpronouncing the last letter.
When teachers referred to prefixes and suffixes other then the -ed ending, they were much less likely to make a link with meaning.Only 36 percent of teachers (n=18) did this (see Box 6.3). If they didtalk about meaning, it was more likely to be in the context of a prefixlike un- or pre- rather than in the context of derivational morphemessuch as -ness or -ion, which are used to form words of a specificgrammatical category. It would seem that English speakers tend to bevery uneasy with the subtleties of grammar.
138 What does the research tell us?
Teachers thinking about morphemes with connectionto meaning
Again similar to the one with -ed at the end Id perhaps havepay and payment, pave and pavement, and all wordswith and without the suffix, where they make sense. Talk aboutthe rule for changing it; talk about how we use one set of words,i.e. theyre verbs, how we use the other set and theyre nouns.And then actually model changing develop into develop-ment and then get them to do it themselves.
I always like to have the prefixes change the meaning of the wordand then to identify the common ones so you know for un- youcould say it makes the word opposite and looking at how itchanges the word.
It would be a whole class thing that wed do. Theyd all havedifferent prefixes to work with and come back at the end anddiscuss how its changed the word and then just say thatsometimes, like prepare, youd have to find out . . . look theroot of prepare anyway. They love things like, is it called
The majority talked about morphemes in terms of letter strings orletter patterns (Box 6.4). Teachers observed that their children speltrichness with a single s, or pavement as pavemint or pavemnt,and, as we have seen in previous chapters, -ion words are particularlydifficult. Here, teachers were addressing the fact that prefixes andsuffixes are frequently occurring letter chunks, which are difficult to spell by relying entirely on their sound. Memory of visual patternsoffers a viable spelling strategy. One teacher was clear that prefixes are usefully linked to meaning, but she thought that only her bright-est pupils found this helpful (Teacher 27). Another teacher commentedon the difficulty of identifying the root word in a word like prepare,a bit like the gruntled character in P. G. Wodehouse mentioned inChapter 1.
Can we increase teachers awareness of morphology? 139
etymology? They love finding out where these words come from,so it is worth it.
Well, theyve done lots of work, the entire class at different levels,on suffixes and prefixes, what goes before a word, what goesafter a word, what does it mean and so on . . . We talked aboutif something comes before a word or after how it can change themeaning of the word, but we did quite a chunk on that.
Teachers thinking about morphemes withoutconnection to meaning
I simply pulled out a whole heap of words ending in like, we had -ness and -less and what sorts of patterns could they
140 What does the research tell us?
see? And they all went Hey, they all end in -ness or they all end in -less.
With that one, you could focus on the -ness ending. Thatsquite a common string, suffix.
I tend to do things like, recognizing it as being a suffix. I mean,we did some looking at prefixes last week and we look at trans- and tele- and some of them, the more able ones, weretrying to think up what it meant, but for the majority of them itsjust recognizing it and thinking I know that Ive come across thatpattern.
Within the literacy hour there is focus on suffixes in Year 5 and inYear 3 and 4 so they would look at groups of words with the samesuffix and I have some suffixes written up for display in my roomas well so they can look for patterns.
And pre-, that would come under prefixes, prevent, pre-pare . . . But then pare is not a word on its own, or maybe itis. Im getting a bit foggy now, theres actually root words, do wetake the root word away? If you took the prefix off, does this haveto be a root word? Id have to look it up.
Basically, thats a suffix thing . . . Looking at the root word andadding the suffix (-ment). So it would be a lot of word cards,where theyre literally moving it into it and then writing it down. . . as soon as they recognize that thats how the word looks,then thats when they spell it correctly.
Finally, when it came to -ion, teachers had no difficulty inrecognizing that this was a real problem for childrens spelling, but onlythree teachers referred to suffixes in this context, and none of theteachers mentioned the meaning function of -ion, changing a verb toan abstract noun. There was a general unease about teaching -ion, andthose who had a go at describing a strategy suggested either a visual one,involving rote learning, or analogy (Box 6.5). Some teachers mentionedthe -ion and -ian confusion. As we have seen, there is a rule for dif-ferentiating the spelling of -ion (abstract nouns) and -ian (personwords, or, as one colleague suggested, things that a Scottish lad calledIan could do), but this was not articulated.
Can we increase teachers awareness of morphology? 141
Teachers talking about -ion
Thats just about learning that . . . the [middle] t can make the/sh/ sound. Again, its practicing it, learning words . . . groupingwords that would have that kind of spelling, -tion words. Wedo quite a lot of work with that in literacy hour . . . We link it with-cian, you know, magician and stuff, and we look at thedifferences in groups of words. [Are there ways of identifyingwhich type of spelling it would be?] There are, probably, but Icant think what they are at the moment though. There probablyis a rule, but unless Im teaching it that day, I dont learn it!
I would probably do, either learn a group of words that end in -tion, or specifically learning -tion as a phrase and having itdisplayed in class or in the word bank. And it could be anythingridiculous, it could be Tigers itch or not, you know, it could beanything to help them remember it.
Teach this as a chunk as they will know some through morefamiliar words like station.
It seems that teachers have explicit knowledge of some aspects of morphology but not others, and this explicit knowledge reflects boththe context in which they teach (in this case under the mantle of theNLS) and aspects of morphology that are most transparent.
We are arguing that this antimorphemism is a problem and thatchildren would be spelling better if adults indulged in more explicitmorphology. In fact, we have already presented some evidence tosupport our case. But can we change the teachers? If we could do that,we would be more confident that interventions would have a life outsidea research project.
The teacher intervention
We designed a ten-session literacy course for teachers, which empha-sized two neglected but, we think, important dimensions of explicitinstruction for 7- to 11-year-olds: Comprehension and morphology.Here we will tell you just about the morphology work. The course was
142 What does the research tell us?
They just . . . separated them out and wrote down as many -tions and -sions and I think there was some rule, but I cantremember, as a rule in my head, Im not sure, I cant rememberit, I cant think of anything that comes straight to my mind so itwould just be, it is just the case of remembering it, probably betterto rely on the visual than to . . . theres no things from the sound.
I tell the children, Think of a word that sounds the same and doyou know how to spell that? So how do you think that that willhelp to spell this? It doesnt always work but I think that it willgive them an idea of where they should be going with it.
Brainstorming work, similar endings, -tion, -cian because itsexactly the same sound, its so confusing isnt it? Some of thebetter spellers might put -cian but they would probably know-tion. That would be a smaller mistake, that would be the onesthat almost got it and theyre just not familiar enough with -tion so there are quite different levels of mistakes.
delivered at university over one school term. There were three mainaspects to the morphology part of the course:
1. an introduction to theories about morphology and literacy; 2. involvement of teachers in the intervention and research process; 3. the provision of a practical set of materials to enable teachers to
do explicit morphology in their classrooms.
The core of our introduction to theories about morphology and spellingis given in Box 6.6, although this was elaborated in various ways andthe teachers were also given readings. At the beginning of the course,teachers assessed their children on our word and pseudoword spellingtests (similar to those described in previous chapters) and agreed toassess them again at the end of the school term. We discussed theresearch design and teachers identified suitable control classes withintheir schools, so that we could tell whether any gains that their childrenmade were above average. We gave each teacher the plans and materialfor a seven-session spelling intervention, similar to the one describedin Chapter 3, and we asked them to try the intervention out in theirclassroom over the term. As the teachers taught the sessions, they wereasked to reflect on their experience and to observe the way that theirpupils responded. When teachers attended the course at university, theydiscussed how their children did in the tests, how the sessions weregoing, and how they would approach the next sessions. We examinedsamples of childrens writing and we discussed how this tied in with thetheoretical side and the readings. The teachers heard from otherteachers who had tried the same system, and they discussed approachesto teaching spelling between themselves. We wrestled with the practi-calities of handing out information, teaching materials, etc. Throughoutthis process we attempted to integrate theory and practice.
Can we increase teachers awareness of morphology? 143
Theories about morphology and spelling
The definition of a morpheme is the smallest part of a word thatcarries meaning, so that can be a whole word, like cat or just partof a word like the s in cats or the suffix less in careless. Wethink that children at Key Stage 2 will be helped in both reading and
How teachers changed
Twenty-two teachers and three literacy advisors (literacy experts withresponsibility for managing and supporting the teaching of literacy inlocal authorities) attended the course. We wanted to know whetherengagement in this process made a difference to how the teachersthought and behaved. To begin the first session, we asked them to writedown and to discuss the strategies that they use in order to teachspelling. This served the dual function of allowing teachers to discuss
144 What does the research tell us?
spelling if they are explicitly taught about morphemes. Heres why.Three broad stages of development have been observed in bothreading and spelling: (1) the whole-word stage (logographic); (2)decoding letter by letter (the alphabetic stage); (3) recognizingchunks in words (the orthographic stage) (Frith 1985). Particularly,in reference to spelling, these translate into the following (Ehri 1986):
1. Semi-phonetic: Spellers represent sounds or syllables with lettersthat match their letter names
R (are) U (you) LEFT (elephant)
2. Phonetic: The child can symbolize the entire sound structure ofwords in their spellings but the letters are assigned strictly on thebasis of sound. Some sounds (for example, /r/, /n/, and /l/) aremore difficult than others to represent in this way (Treiman andCassar 1997).
3. Morphemic: The child becomes more aware of conventionalspellings and employs visual and morphological information inspelling. For example, children learn to represent the /t/ soundat the end of past-tense regular verbs with -ed (Nunes et al.1997b).
Seven- and 8-year-old children should be entering the morphemicstage in spelling. We think that teaching them explicitly aboutmorphemes at this stage is important and will help them to tacklelong words in their reading, to understand word meanings ofunfamiliar words and to help them to spell. Some tricky issues inspelling, for example, apostrophes, are only really successfullytackled through a discussion of morphemes.
the range of spelling strategies that they already used and allowing usto document their practice before intervention. Lots of methods werementioned by the teachers, but the most common were:
look, cover, write, check (65 percent of teachers mentioned this); letter strings/letter patterns (60 percent mentioned this); spelling rules (magic e, plurals of words like baby), particularly
using mnemonic strategies (for example, ought: O u great hairyteacher (60 percent mentioned this);
phonic strategies (55 percent mentioned this); learning whole words (high frequency words, technical words,
words identified as difficulties for individual children) (45 percentmentioned this);
proofreading of various kinds (45 percent mentioned this); spelling investigations (35 percent mentioned this); spelling banks and dictionaries (30 percent mentioned this); kinetic learning of various kinds (25 percent mentioned this).
The use of prefixes, suffixes, or roots was only spontaneously mentionedfour times, and there was only one reference to morphemes.
We also asked the teachers to write down their definition of amorpheme. Of the twenty teachers who completed this pre-course ques-tionnaire, 25 percent knew that it was a small chunk of a word whichhad meaning. Other responses varied: For example, God knows,Something to do with spelling. The most common definition was a unit of sound. Clearly, morphemes are a minority sport.
At the end of the course, we asked the teachers to give us theirdefinition of a morpheme again. Of the seventeen teachers for whomwe had data at the beginning and the end of the course, three haddefined a morpheme fairly accurately at the beginning but all but oneproduced perfect definitions by the end. All but one of the teachersreported that the course had changed their approaches to teachingspelling. As might be expected, most mentioned that they would teachmore explicit morphology, making connections between spelling,grammar, and meaning. However, they changed in a number of otherways too. Several teachers mentioned that they would take spellingmore seriously, for example, focusing one (1-hour) literacy session perweek on spelling and taking a more structured approach to teach-ing spelling. They also increasingly saw spelling as having creativepossibilities, such as class discussions and investigations. A number ofteachers thought that using computer-generated materials had been
Can we increase teachers awareness of morphology? 145
good fun and effective and this had encouraged them to try this againin the future. Seven out of the seventeen teachers spontaneouslymentioned that they would like to introduce the use of spelling journals.This was not something that we had suggested but something that theteachers had discussed themselves during the course. We had hopedthat the course would be an interaction between us and the teachers butworried that in the end, the tremendous pressure of dealing with allthe practicalities would squeeze this luxury out. It seems that somedialogue did go on between teachers despite everything.
How children changed
The sample of children
Seventeen of the teachers attending the course assessed their childrenat the beginning and end of the course (the morphology children). Wealso collected similar data from fifteen classes where the teacher had notattended the course (the control children). The control classes wererecruited from the schools of the teachers attending the course (n=8)and from classes taught by teachers who had attended a similar coursefocusing on numeracy (n=7) in the previous term. Table 6.1 shows thenumbers of children in these two groups by year group.
The children were assessed using two specially devised spelling tests.The first test was made up of thirty-two words, almost all of whichcontained one of the morphemes targeted by the intervention (spellingtest). The second test comprised ten invented words, which alsoincluded the target morphemes (pseudoword-spelling test). Childrencould only spell the words in this second test correctly by applying therules, as they had never seen the words before. The reasons for using
146 What does the research tell us?
Table 6.1 Number of children in each teaching condition, by year group
Year group Control Morphology Total
Year 3 112 110 222Year 4 57 132 189Year 5 128 88 216Year 6 56 30 86Total 353 360 713
pseudoword-spelling tests in order to evaluate this teaching programhave already been discussed in Chapters 3 and 4. The scores for eachof these tests have been converted to percentages to make them easierto interpret.
Changes in childrens spelling
Before the course, the children in the control classes were fairly evenlymatched with the morphology children, except for Year 3, where thecontrol group were substantially better (Table 6.2).
Around 7 weeks later, all the children had improved, but the mor-phology group had made (statistically) significantly larger gains than the control group (overall, four times as much, see Table 6.3) inboth spelling and pseudoword spelling. In the pseudoword test, the
Can we increase teachers awareness of morphology? 147
Table 6.2 Childrens average scores before the course, by teaching condition and year group, percentage correct
Year group Control children spelling Morphology children spelling test test
Word Pseudoword Word Pseudoword
Year 3 62 50 51 38Year 4 65 54 63 40Year 5 74 58 71 52Year 6 72 62 74 62Mean 68 56 62 44
Table 6.3 Average percentage increase in the childrens scores by theend of the course, by teaching condition and year group
Year group Control children spelling Morphology children spelling test test
Word Pseudoword Word Pseudoword
Year 3 0 0 4 6Year 4 3 7 5 16Year 5 2 3 4 7Year 6 2 5 9 10Mean 1.3 2.5 5.1 10.1
morphology group had to apply an understanding of rules to the task;they couldnt merely have learned the words through the course of theintervention. The only time they had seen these words before was in thepretest.
In the following school year, one of the teachers who had attendedthe course organized a range of different spelling conditions in herschool and assessed their impact on a new cohort of children. Sheinvolved three parallel Year 4 classes. Two of these classes were testedon their spelling in the fall term. Thirteen weeks later, in the springterm, they were retested and the third parallel class was also tested.Over the next 13 weeks, she used the morphology-intervention mate-rials with her class in a special time slot (n=23, the morphologycondition: The children had not used these materials previously). One parallel class spent the same dedicated time on spelling usingmaterials from the NLS (n=26, the NLS condition). In a third parallelclass, children followed the standard program (n=19, the standardcondition). All three classes were retested in the summer term. Figure6.1 gives a schema of the design used in this study.
Between fall and spring, children made slight (nonsignificant) gainsin their spelling. Between spring and summer, the morphology and NLS classes made noticeably greater (statistically significant) improve-ments and the standard classroom group made slight (nonsignificant)gains.
148 What does the research tell us?
Figure 6.1 One-year teacher follow-up.
One-year teacher follow-up
Three Year 4 classes
One class, morphology intervention
One class, NLS intervention
One class, standard program
Teachers comments on the intervention process
The morphology course seemed to work well overall, having an impactboth on the teachers and their pupils. We asked teachers to keep diariesthroughout and to comment on the intervention process at the end.Generally, they felt it had been a very worthwhile experience both forthem and for their children, but there were some problems.
Teachers felt under pressure to test their children, to learn about the morphology sessions and to teach a new session each week, to testtheir children again and to try to come to grips with some of the techni-calities of the research process. Ideally, it would be preferable to spreadthe course over a longer period, though this may have resulted in lesscommitment.
We had aspirations to include the teachers at all stages of the researchprocess as well as the intervention process. The teachers were verysophisticated in their reasoning about the research. They were con-cerned to set up suitable comparison/control groups; they wereconscious that each teacher used the morphology sessions in slightlydifferent ways, some doing more, some less, some reinforcing learning
Can we increase teachers awareness of morphology? 149
Pretest (fall) First posttest(spring)
Figure 6.2 Childrens scores on spelling test: A comparison of morphology,National Literacy Strategy, and standard conditions.
Note: Statistically significant improvements from first to second posttest:
Morphology condition: Paired t-test, t = 7.0, df = 22, p
in other parts of their literacy teaching, others not, etc. They wereconcerned that it was difficult to be sure that the improvementsobserved in their children were as a result of linking morphemes withmeaning or merely due to increased exposure to letter patterns such as-ion. However, they did not have the technical skills to analyze thedata by themselves, and this turned out to be too much to take onalongside running the interventions in their classrooms.
Working with the children, there were two particular concerns. First,the use of the technology was an issue. On the whole, both teachersand children approved of the lively, interactive PowerPoint presentationof the morphology materials. In particular, it was seen as motivating forchildren who were more difficult to engage. However, the downsideswere that sometimes children were too passive and that some teachersfound the program hard to organize. The second major issue wasdifferentiation. The teachers on the course had pupils spanning the age range 811. Some of the younger children struggled with graspingwhat was a noun (especially abstract nouns), a verb, and an adjective.They probably needed more practice and, perhaps, focusing onmorphemes such as un-, -less, and -ful that dont involve any fancygrammar. Some of the more able children found the tasks too easy andneeded an extension of the program. This is always the tension withproviding materials. On the one hand, they offer an efficient vehicle tofamiliarize teachers with ideas and a practical manifestation of moreabstract concepts relating to their teaching. On the other hand, theywill always need some adaptation or selective use or extension to takeaccount of individual differences.
Two case studies
Teachers attending the course were given the option to gain creditstoward a Masters course if they completed an assignment. Two teacherscompleted assignments on the morphology intervention. Here, onereflected that much of her knowledge of grammar was picked up ratherthan taught.
I learned the majority of my English grammar indirectly. I dontconsciously remember being taught present, past and future tensesin the same way as they were taught in my French classes. [French]was taught through a structure specific approach and informed myunderstanding of the English language.
150 What does the research tell us?
They considered phonological awareness as an essential foundation inthe learning of reading and spelling, but saw that teaching childrenabout morphology had important benefits for older children.
Differentiation was one issue that was raised. One teacher com-mented that the morphology sessions worked particularly well forchildren for whom English is an additional language.
It was clear they benefited from the direct teaching which enabledthem to clarify misconceptions and underpin their existing (andsometimes shaky) knowledge of word classifications and changingmeaning.
She felt it was less suitable for her children with special educationalneeds who did not have a broad enough basic vocabulary to benefit atthe level the intervention had been set. The other teacher found thata mixed-ability group worked well.
The teacher could use the more able children in the groups asexperts and encourage them to explain their thinking and the howpart of the sessions. By this I mean that when the penny drops forchildren, they can have a unique way of explaining how they came to understand, which can often help the less able children tosee the light. In the intervention group there was a particular childwho is on a behavior strategy plan and who finds it difficult tointeract with other children. This child, who was one of the moreable children, found the sessions a good way to express his ownknowledge and was excellent at giving the other children cues andclues as to what they needed to do.
This teacher commented particularly on the advantage of children beingable to think about underlying concepts or rules.
The teacher did encourage the children to take ownership of their learning and at the end of the session there was a reflectiveperiod where the children could evaluate their own learning. It was interesting to note that with some of the rules, the childrenunderstood why they had previously made errors and some eventried to come up with their own memory cues. For example, onechild suggested that we could remember -ian as an ending
Can we increase teachers awareness of morphology? 151
pertaining to a person because it is similar to I am. So, therefore,if you could say, I am in front of a word, it meant it most likelyended in -ian.
These teachers final reflections on the intervention are informative.
There is no doubt in my mind that an adaptation of this interven-tion would be beneficial for works with further groups of pupils inthe future. This adaptation would involve the use of shorter sessionsregularly repeated over a longer period of time and reinforcementactivities which would incorporate word classification games andapplication within a range of written contexts.
Although as a teacher I try to think of innovative ways to helpchildren acquire spelling success, there are many administra-tive pitfalls, such as time and resources. In the literacy hour we doaddress spelling in the word and sentence level section of the hour,but is 15 minutes a day really enough time to allow children tolearn experientially and firmly concrete their knowledge? I guesssome may say, Yes, if youre organized! But the reality of time is a big issue. Children need to have time to discover the patterns and trends for themselves and then formulate their own rules,(which should naturally correlate with the teachers) as well as thetime to then test their new knowledge. [Year 5] children were stillexcited to play word games. [. . .] I think that in Key Stage 2 we tendto lose sight at times of the value of games and the importance ofexperiential learning. The groups results, as well as their refreshedattitudes toward spelling, speak for themselves.
152 What does the research tell us?
Summary and conclusions
Although it can be seen from previous chapters that doing focusedwork with children improves their spelling, transforming researchinto teacher practice is a complicated business. In the research thatis the focus of this study, the process of transformation had already
Can we increase teachers awareness of morphology? 153
begun. English policy documents do identify the role of morphologyin teaching spelling.
However, when we looked at teachers practice, reference tomorphemes was limited and patchy. No teacher spontaneouslymentioned the word morpheme, and, when asked, most teacherswere unaware of its meaning. This suggests the absence of an explicitknowledge of the concept of the way morphology governs thespelling construction of English. Although teachers talked aboutaspects of morphology, most commonly in the context of verbendings and prefixes and suffixes, they normally focused on thevisual patterns, failing to make a link between this and the meaningfunction. Observation in the classroom confirmed that children wererarely taught about the morphological dimension of our language.
When teachers attended a ten-session course on morphology andcomprehension at Key Stage 2, their practice changed, and this had a positive impact on their pupils spelling. The pupils of teachersattending the course made significant gains in spelling compared tochildren in similar classrooms receiving standard instruction,particularly impressive for a whole-class intervention delivered byteachers just learning a technique for the first time. The interventionis quite a focused and practical one, despite its conceptual base, andthis probably contributed to its impact.
Exactly what aspect of the intervention caused the change is lessclear. We would like to say that it was due to teachers conceptualawareness of morphology. However, there are a number of othercontenders. The classic alternative explanation is that the childrendid better just because they were being exposed to something new.This seems unlikely. The teachers of half of the control children wereattending a mathematics course that was also exposing the pupilsto novel practices. Another alternative explanation for childrensspelling gains is that teachers spent more classroom time teachingspelling. This was certainly an effect of the intervention. We had seenfrom the survey phase of our research that usually not a great dealof teaching time is dedicated to spelling in the standard classroomand that the teachers on our course commented that they werespending more time teaching spelling than they normally would.
In the year following the course, one example clarified what aspectof our intervention was improving childrens spelling. Just improving
A more technical version of this chapter can be found in J. Hurry, P. Bryant, T. Curno, T. Nunes, M. Parker, and U. Pretzlik (2005)Teaching and learning literacy, Research Papers in Education, 20:187206. The website for Research Papers in Education is http://www.tandf.co.uk/journals
154 What does the research tell us?
teachers explicit knowledge of morphology was not enough. In thefall term, a teacher who had been on the course did not have anopportunity to use the morphology materials. During this term, her new group of pupils made no greater gains in spelling than theother children in the school, despite the fact that this teacher didhave explicit understanding of the role of morphology. Increasing the amount of time on spelling did not explain the gains entirelyeither. When her class was compared with a parallel class (the NLScondition) receiving the same amount of additional spelling instruc-tion, her morphology group made significantly more progress thanthe children having additional NLS spelling activities. However,additional curriculum time was helpful. Both these classes madesignificant spelling gains compared to a control class and to theirown progress in the previous term. The ingredients for change inpupils performance appear to be teacher knowledge and dedicatedteacher time.
We conclude that it is possible to transform the teaching ofmorphemes developed in a research context into teacher practice,and this in turn has an impact on childrens learning.
What are the overallimplications?
Morphemes and literacyContext and conclusions
In this final chapter, we shall write about our findings in the contextof current theory and practice in the teaching of spelling. When we began our research on teaching children about morphemes and spelling (the research that we have described in Chapters 36of this book), we already knew that there was a strong and impor-tant connection between childrens knowledge of morphemes andtheir understanding and use of the English spelling system. Ourprevious work and the work of several other researchers (describedin Chapters 1 and 2) had led us to three main conclusions about this connection. We thought that these conclusions were clear andconvincing enough for us to use them as starting points for aneducational research program about how best to promote childrensmorphemic knowledge in the classroom in order to help them tomaster the English spelling system and vocabulary. Our three startingpoints were:
1. Some of the most important correspondences between spokenand written language are at the level of the morpheme. Themorphemic structure of words in English, and in several otherorthographies, often determines their spelling. The system ofmorphemes, therefore, is a powerful resource for those learningliteracy: The more schoolchildren know about morphemes, themore likely it is that they will learn about spelling principles basedon morphemes.
2. Childrens knowledge of morphemes is largely implicit. Theconnection between this implicit knowledge and childrens
Authored by Terezinha Nunes and Peter Bryant
158 What are the overall implications?
literacy learning is not a strong one. However, there is a strongconnection between explicit knowledge and literacy learning.Those children who do have a relatively strong and explicit levelof awareness of morphemes and grammar at younger ages learnto use morphemes in spelling more systematically at a later age.This explicit knowledge helps them to use the spelling for suffixescorrectly, such as the -ed for past verbs, and also to preservethe spelling of stems across words. The connection betweenmorphemes and literacy is a two-way street: Awareness ofmorphemes strengthens childrens spelling and their knowledgeof morphemic-spelling correspondences also promotes theirawareness of morphemes at a later age.
3. It is possible to enhance childrens explicit knowledge ofmorphemes through systematic instruction and to improve theirperformance in word reading and spelling. These teachinginterventions are effective for normal readers and poor readerswho are seriously underperforming for their age and intellectualability. As things stand at the moment in our educational system,one of the main reasons for childrens generally low level ofexplicit awareness of morphemes may be that they are taughtvery little about morphemes at school.
All three of these statements have been amply supported by theresults that we have reported here. The evidence that we havepresented in this book has shown for the first time that it is possibleto teach children in interesting ways about morphemes in theclassroom and also that this learning has a good effect on childrensliteracy development. All the studies that we have reported point inthe same direction. They show that it is possible and also highlydesirable to teach children about the connection between spellingand morphemes, and that the essential ingredient of this teachingmust be the promotion of childrens explicit understanding of themorphemic structure of the words in their spoken vocabulary. Wenow wish to set these results in the context of contemporaryeducational theory and practice.
The path from the laboratory to the classroom
When we began the research (Chapters 36) on improving school-childrens morphemic knowledge, we had reasons for optimism but alsofor caution. We were optimistic because of the example of previouswork on improving another form of explicit linguistic awareness inschoolchildren, which was their phonological awareness. In the past 25years of the twentieth century, many research studies established thatthe awareness that children have of sounds before they reach school agepredicts their progress in reading and spelling and that increasing theirawareness of sounds has a positive effect on how well they learn to readand spell. Teachers and policymakers then made use of this knowledgeabout phonological awareness and, as a result, literacy teaching hasnow a much more solid foundation than before. With this example inmind, we felt that it was time for a similar effort to be made withchildrens awareness of morphemes and grammar.
We were cautious at the same time because we knew that the trans-formation of basic knowledge from research into classroom practiceand educational policy is never simple. Edmund Henderson, one of theprominent figures in the study of the development of childrens spelling,warned about the seductive belief in a simple path from researchfindings to educational practice. He wrote,
One of the best ways to learn about English spelling is to studythe developmental stages that achieving children follow as theygradually master the system. Information of this kind tells whatchildren learn about words and the order in which they learn it.Developmental studies do not, of course, tell how children learnthese things or how best to teach them to do so.
(Henderson 1990: 40, emphasis in the original)
Our previous research had suggested that it is worth ones while toincrease childrens explicit knowledge of morphemes and grammar.However, past practice had shown that teaching children aboutgrammar and morphology can be boring and might therefore have littleeffect on literacy. In order to make the connection between research andpractice, we needed to find good ways to teach morphemic principlesto children and see whether it is possible to do so in a way that maintainsthe childrens interest in the classroom.
We made the transition from the laboratory to the classroom in sixsmall steps, and this strategy was a definite success.
Morphemes and literacy: Context and conclusions 159
Carrying out a laboratory study to compare the effectiveness of different kinds of teaching
Our first step was to do an orthodox laboratory-style training study(Chapter 3, Study 1). This study had all the usual features of alaboratory-style intervention. It included pretests and posttests, threedifferent intervention groups and a control group, and two comparableintervention sessions for all the groups. The instructorpupil ratio inthe intervention sessions was small (1:2); some would say that it was ridiculously small. Nonetheless, the study showed that it is possibleto increase childrens awareness of two suffixes with identical soundsbut quite different meanings and to improve their success in finding the right spelling for these suffixes. It also showed that explicit methodsare generally the most effective way of teaching children about thisdistinction.
Repeating in the classroom a form of training that had worked in the laboratory
Could this success in a laboratory-style experiment be repeated in theclassroom where the instructorpupil ratio is very much larger (around1:30) and there are many distractions? The second step on our progressfrom laboratory to classroom was taken when our friends and colleaguesin Lauriston Primary School set out to answer this question. The answerthat came from their study (Chapter 3, Study 2) was that the two-sessionintervention definitely does have an effect in the classroom as well,though one that is weaker than in our laboratory study. We concludedthat more than two intervention sessions might be needed in classroomstudies.
Extending and broadening the classroomintervention
Our third step, therefore, was to go back to the classroom and to set upanother study (Chapter 3, Study 3) with more intervention sessions. At the same time, we decided to broaden our intervention radically by teaching children about a wide range of affixes and their spellings.We had already found that if children work with a variety of stems andaffixes and apply different cognitive operations on morphemic units,they can use their knowledge of morphemes to bolster their wordreading and the spelling of morphemic units (Nunes et al. 2003). So we
160 What are the overall implications?
developed an intervention program to be tested in the classroom. Theintervention was run by classroom teachers, but we prescribed quitetightly what the teacher should do, and a member of our research teamwas always at hand during the intervention sessions to advise theteacher if advice was needed. Again, we found that the teaching worked.The children who had been taught about morphemes and spelling faredbetter in a morpheme-spelling task than children in a control group.Thus, we had successfully made the step from laboratory to theclassroom, but the teaching in the classroom projects had been rathertightly constrained. The methods and the limited scope of the firstclassroom study were very similar indeed to those of our laboratorystudy. In the second classroom study, the teacher followed a set programvery closely and a member of our research team was always present inthe intervention. We were acutely aware we would never have suchcontrol if our methods and ideas were widely adopted.
Leaving the details and the timing of the teaching to the teachers themselves
So, we wanted to know what would happen if several teachers adoptedand applied our program with no more than a small amount ofpreliminary advice and the use of a CD-ROM which contained ourdifferent tasks or games in PowerPoint. If our ideas are to be used in theclassroom, we would never have any more control over the teachingthan this. Therefore our fourth step was to set up a study (Chapter 4)in which we simply gave the teachers a CD-ROM with our program,with a minimum of advice about how to use the CD-ROM and about theaims of the project. Our only constraint was that the teacher shouldadminister the different tasks in the order in which they appeared on the CD-ROM. This project also produced strong effects of teachingabout morphemes on the childrens spelling, as measured by pretestsand posttests. Thus, our methods also worked when the tight constraintsof our earlier interventions were no longer there. Our methods work inreal classrooms.
Teaching vocabulary as well as spelling
Schools are at least as likely as university departments to produce goodresearch questions about teaching. Now that we were working inclassrooms, we thought a great deal about questions raised by theteachers themselves. The teachers in a group of schools that we worked
Morphemes and literacy: Context and conclusions 161
with in London were concerned about their pupils vocabulary andwanted to find ways of improving it. Our view was that helping childrento understand how words are constructed from morphemes should be an effective way of promoting their vocabulary for reasons that weset out in Chapter 1, and so our fifth step was to use our interventiontasks in a training program designed to improve childrens vocabulary.The answer to this question, generated by the schools themselves, wasagain positive. The program, again administered with minimal con-straints imposed by ourselves, did improve the childrens knowledgeand understanding of polymorphemic words.
Studying the effects of a teachers course about morphemes and spelling on the teachers and on their pupils
If children in our schools are to be taught about morphemes in the ways that we have been advocating, these teaching methods would haveto be part of teacher training too. People training to be teachers andpeople who are teachers already would have to learn about the kind oftasks that we have developed. It would be essential, also, to increasetheir awareness and understanding of morphemes. Thus our sixth step(Chapter 6) was to institute a course for teachers about morphemes andto study the effect of this course on the teachers and on their pupils too.The study showed that the course did radically increase the teachersawareness of morphology and its links with spelling and that their pupilsspelling also benefited, provided that the teachers used some of ourtasks to teach the children about morphemic spelling principles.
Our journey from the laboratory to the classroom and to courses forteachers was intricate but productive. It is right to be cautious aboutmoving from one environment to the other, but we believe that we haveshown that it is possible and also profitable to make the transitionproperly.
We now turn to the context of ideas about teaching spelling in orderto examine in what ways these could change under the impact of ourresults.
Word spelling: caught or taught?
We have just dealt with a question about research methods that can be used to make a connection between the laboratory and the class-room. Our next question is about educational theory and practice. Do
162 What are the overall implications?
educators think that children ought to be, and need to be, taughtspelling? If they need to be taught, what should they be taught and howshould they be taught it? As we shall see, views on what and howchildren should be taught have changed over time. Our answer, to alarge extent, is a new one.
The question whether spelling is caught or taught was posed byMargaret Peters (1985) some time ago. Most authors then seemed toagree that there are basically only two alternative possibilities abouthow children learn spelling: Either they catch spelling, which meansthat they learn it implicitly as they read, or they have to be systematicallytaught how to spell.
Those who thought that spelling was simply caught did not feel thatanything had to be done to teach children to spell. But, according toMargaret Peters, it was only possible for some children to catchspelling and only under special circumstances. She argued that parentsand teachers have to direct childrens attention to spelling patterns forthis catching to happen. This guidance should emphasize the proba-bilities of some letters occurring in sequence, thereby facilitating visuallearning of words.
In fact, there is little evidence that children actually can catch spellingwithout any teaching. It is likely that everyone receives some teaching,as even the most radical believers in the possibility of children catchingspelling will teach them some spelling. It is also true that many uni-versity students continue to find English spelling a challenge, perhapseven throughout life. Many confess to changing their choice of wordswhen writing, due to their uncertainty about how to spell words thatthey had intended to use in their texts.
It is also quite likely that there is much implicit learning of spellingbut, again, there is little evidence for how much and what is learnedimplicitly. Over time, the hope that children would simply catch spellingseems to have waned. Teachers and policymakers have turned to whatto teach and how to teach.
Some time ago, those who thought that spelling has to be taught were impressed by the intricacies of English spelling. They focused their efforts on the memorization of the spellings of the individual words because they believed that easy words (that is, regular words)would be learned without difficulty and that children just had tomemorize the problem wordsor, as they were often called, thespelling demons. Now, each child would fall under the spell of differentdemons, and there is no point in protecting one child against someoneelses enemy, so each child had to try to memorize the words that he
Morphemes and literacy: Context and conclusions 163
or she found difficult. Therefore, lists of words to be learned would haveto be compiled. This was the first step required to transform these ideasabout spelling into practice. What should children be taught? Theanswer was the spelling of words.
There was some disagreement among those who advocated theteaching of spelling about how the lists of words should be compiled.Peters, for example, favored the idea that the lists should be based onthe childs own writing: Children should learn those words that theymisspelled when writing their own texts. This form of instruction wouldbe individualized, and it was believed that it would be of greater use andmore motivating to individual learners: They would be learning thespelling of the words they had wanted to use in their own writing. Incontrast, Fred Schonell in 1932, another pioneer of the study of spelling,thought that standard word lists could be used (Schonell 1957). Hecompiled the frequencies of words from material written by childrenaged 712 and recommended that children should work their waythrough these lists, learning the spelling of words from more to lessfrequent ones. This, he argued, should make it easier for children tospell words that they would be using all the time, would increase theirchance of success in spelling in early grades, would provide them witha basis to enjoy writing, and would result in an increase of their writtenvocabulary at their own pace.
The first step in the transformation into educational practice of either of these ideas about spelling was simple: Teachers had to compilelists of words whose spellings children had to learn. In order to createindividualized lists, following the views put forward by Peters, teacherswould simply have to ask the children to write texts with some fre-quency, correct them, single out the words to be learned, and ask thechildren to learn them. The construction of individualized lists wasattainable by teachers for their pupils. Similarly, it was possible forteachers to draw on Schonells lists and to use these in the teaching ofspelling, because he divided the words into six groups, each representinga chronological age, which would give teachers a starting point forchildren at particular age levels. Schonell also indicated that teacherscould replace some words with others that might be more appropriatefor their pupils. Both Peters and Schonell recommended frequent testingand targeted learning, so the spelling lists, whether individualized orstandardized, were used to assess the childrens level of spelling and todetermine their learning program.
There is, however, a further question to answer for anyone takingthis approach to the teaching of spelling. What should children do to
164 What are the overall implications?
learn the spelling of the words they were presented with? Once teacherschose words for the children to learn, how could they best help thechildren to memorize the spellings?
Peters supported the idea of visual learning. Children, she argued,should look carefully at the word spelled correctly by the teacher andnote what in its spelling might be difficult, cover the word with a pieceof paper, write the word in their own books, and then check that thespelling was correct by matching it against the teachers spelling. This became known as the lookcoverwritecheck method. If theword was misspelled, the child should rub it out, write it correctly, and then repeat the whole process. The aim of this method is to formvisual memories and motor habits for each word; it does not attributeimportance to correspondences between letters and sounds in thismemorization process because the spelling demons were exactly thosewords whose spelling could not be predicted from the way the wordsounded. Saying the word, it was thought, might actually interfere withlearning spellings which were not predictable from phonology.
Others, in contrast, advocated multisensory methods. MariaMontessori (1915) was one of the first educators to suggest that it isimportant for children to establish multiple connections between thesounds, visual shapes, and kinetic representations of letters and words.Although Montessoris method was developed in the phonic tradition,multisensory ideas were also used in the whole word tradition, perhapsmost notably by Grace Fernald (1943). She suggested that the childrenshould think of the word they wanted to learn to spell, the teacher wouldthen write the word on a card (usually in the context of a sentence), andthen the child would trace the word with a finger, saying each syllableas it was traced. The process would be repeated until the child was ableto write the word from memory.
This very brief description allows us to identify the underlyingassumptions to these approaches to the teaching of spelling. Theiranswer to the question What should children learn? is whole words.Their answer to the question How should children learn? is by mem-orization. Their shared and implicit theory is, in some sense, a form of behaviorism: Written words are spelling responses to be learned as individual and unconnected items. There seems to be nothing to belearned from childrens mistakes. The form of memory to be engaged inthe teaching of spelling is brute memory, as Henderson (1990) calledit. There is no place for understanding or for the creative role of childrenin generating spellings. We shall discuss the idea that children cangenerate spellings rather than memorize each one in the next section.
Morphemes and literacy: Context and conclusions 165
Childrens understanding of English spelling
Ideas have moved back and forth in educational theory, but onetheoretical change, which we think is unlikely to revert to past views,is the way in which childrens mistakes are conceptualized. We thinkthat this is one of the many positive impacts of Piagetian theory oneducation. Childrens mistakes, which were treated in the past asresponses to be eliminated, are now considered as valuable pieces ofinformation about how children think.
The linguist and former teacher Charles Read provided the break-through for this view in the domain of spelling in 1971, when he showedthat young children were able to invent spellings for words that no onehad taught them to write. These spellings, though often mistaken,revealed an understanding of the idea that letters represent sounds.Although not all sounds were represented, and although childrenswritings revealed that their analysis of sounds was not exactly the sameas the analysis a literate adult would produce, there was no question thatthe childrens spellings were based on what is now termed an alphabeticor phonetic conception of writing. Children created spellings from theirunderstanding of the spelling system, they did not just reproduce(correctly or incorrectly) memorized spellings. Reads findings havesince been confirmed by many researchers working in English (forexample, Ehri 1997, Treiman 1993) as well as other languages (forexample, Ferreiro and Teberosky 1983, working in Spanish).
Further research has attempted to elaborate on this discovery, either by trying to identify earlier stages that precede the childrensalphabetic conception or by seeking ways of describing progress inchildrens conception of spelling beyond the alphabetic stage. There arecontroversies about what is the best way to describe the developmentof childrens conceptions of English orthography, but these do not needto be discussed here. Our focus is actually on the convergence that canbe seen across many researchers toward the idea that spelling devel-opment is best described not as the acquisition of a spelling vocabulary,word by word, but as the acquisition of spelling principles.
Among the many pieces of evidence against the idea that spellingdevelopment is characterized by the acquisition of correct spellings forparticular words, one stands out: This is the finding that children mayactually appear to learn to spell words worse than before as they growolder, if their progress is judged on a word-by-word criterion, becauseas they grow older they make spelling mistakes with words that theyused to spell correctly before. We have found clear evidence for this
166 What are the overall implications?
apparent regression in our studies of childrens acquisition of the -edending. When children first start to use -ed at the end of words, theyput it at the end of the right as well as of the wrong words. Because ourstudy was longitudinal, we were able to show that the same childrenwould make the mistake of using -ed on words that they had spelledphonetically and correctly earlier on. They had discovered a new factthat some words ending in /t/ and /d/ sounds are spelled with -edand had started to use this knowledge, but they had not quiteunderstood the principle that the -ed ending is reserved for regularverbs in the past.
Results like these suggest that we should no longer think that childrenlearn spelling by adding one word at a time to their memorized list. Itis to some extent irrelevant whether spelling principles are acquiredsuddenly or in small steps and whether they can be learned as rulesthat are taught or must be learned from experiences with words. It nowseems certain that childrens understanding of the relations between thespellings of different words plays an important role in the developmentof spelling.
Henderson (1990) suggested that this new view of the develop-ment of spelling does not refute the notion that memory plays a role in learning spelling: The question is how memory should be used inteaching and learning spelling. According to him, there is good reasonto doubt that poor spellers can mend their ways by even the most heroicefforts to memorize the spellings of all the words they need. Never-theless, he agrees that they can start from learning groups of words thatillustrate principles and increase their ability to use these principlesthrough further experience. There is room for word-specific learning,but there is evidence that children spell words that they have not learned by using their knowledge of relations across different writtenwords.
So, if we want to engage childrens understanding of principles along with their memory efforts when we teach spelling, we mustchoose well what they study.
When teachers and researchers came to view childrens under-standing of how written language works as an important part of learningto spell, they came up with a new answer to the question What shouldchildren learn?: They should learn relations between the spelling ofdifferent words. They also had a new answer to the question Howshould children learn?, which was that they should learn throughunderstanding, not just through memory traces.
Morphemes and literacy: Context and conclusions 167
Spelling units or spelling principles
One-to-one correspondence between letters and sounds cannot tell the whole story of how to spell English words because there are morephonemes in English than letters of the alphabet to represent them with.So inconsistency in lettersound correspondences is inevitable. Butthere is also more inconsistency than strictly necessary: Some wordssound exactly the same and yet they are spelled differently. To confusematters further, the same word might be pronounced differently bydifferent speakers, although they all have English as their first language.Where does one begin to put some order into English spelling so thatchildren can use their understanding when studying written Englishwords?
One possibility that has been suggested by many is that phonologicaland graphic units larger than letters and phonemes should be used inteaching. It has been suggested that teaching children the regularitiesthat can be observed by dividing words into onset and rime (that is,dividing them into two parts, the first which contains its beginning andgoes up to the first vowel, and the second which starts with the voweland includes what follows it) is a good way to deal with inconsistencyat the alphabetic level. Rebecca Treiman and her colleagues (Treimanet al. 1995) argue that the level of sound to spelling consistency at thelevel of rimes is much higher than at other phonological levels, such asphonemes.
Others (for example, Cardoso-Martins 1994) have suggested the useof syllables rather than of rimes or phonemes as units in spelling. Theuse of syllables as spelling units is discussed less often in English thanin other languages, such as Portuguese and Spanish, possibly becausechildren have difficulty in identifying syllable boundaries in English.The syllable is undoubtedly a different unit of analysis that could beused in the search of consistencies in spelling. However, it should bepointed out that there are so many monosyllabic words in English thatthis solution might not be very practical: Using the syllable as a unit ofanalysis for so many words could turn out to be carrying out no analysisat all.
These two ideas differ insofar as they focus on the use of differentunits of analysis, but they have one thing in common, the idea thatspelling can be taught through the use of different phonological units.
A different approach is suggested by those who think, not in terms ofphonological units but in terms of the acquisition of different principlesto be used in conjunction with the alphabetic principle. This is a positionthat was adopted by teachers such as Edmund Henderson (1990) and
168 What are the overall implications?
linguists such as Richard Venezky (1970, 1995); it is also our ownposition.
A practical example might help clarify the distinction between these two approaches. If a child is learning to use rimes in spelling, for example, the child would have to learn different things in spellingword-pairs such as hathate, madmade, taptape,planplane, hophope, rotrote, tontone, pinpine,sitsite, tubtube, ususe, etc. In contrast, if the child islearning to use principles, rules, or generalizations in spelling, all ofthese pairs can be summarized under one principle. This would reducethe number of specific facts to be learned and would provide thechildren with many connected experiences that would help themunderstand the relevance of one spelling principle. Rimes might be away of assembling a list of words that exemplify a principle, but wesuggest that what children should be learning is not single rimes as ifthey were spelling responses: They should be learning principles. Ourresearch has shown that children learn better if they are exposed to thecontrast between the pairs than to each set of instances of rimes (Nunesand Bryant 2006). The exposure to the contrasts helps children learnthe principles.
Richard Venezky argued that the regularities of English spelling couldbe understood better if, instead of thinking of the relations betweensingle letters and single phonemes, we thought of graphemephonemecorrespondences, where the graphemes might be composed by morethan one letter and these letters might not actually be in sequence. Thisapproach is very useful for the description of regularities in the Englishorthography but less useful for an account of the development ofchildrens spelling. It seems to us that children take a major step whenthey start to think that there are different types of regularities in Englishand shift from using almost exclusively the alphabetic principle tocombining it with other types of regularities. We will refer to the regu-larities that go beyond single lettersingle phoneme correspondences as higher-order spelling principles, to differentiate them from thealphabetic principle.
There has been a great deal of discussion about whether such spelling regularities should be called rules, generalizations, orprinciples. When teachers use these regularities, they are trying to helpchildren understand that, under certain conditions, regular lettersoundrelations are replaced by other regular and predictable relationsbetween groups of letters and sounds. There is agreement regardingwhat to teach. Disagreements are actually about how to teach: Should
Morphemes and literacy: Context and conclusions 169
children be taught these regularities as rules or should they learn themfrom experience?
We argue that there are two aspects to higher-order spellingprinciples: These are their form and function. We will argue that thelearning of higher-order spelling principles cannot be simply thelearning of a visual pattern. In the case of the higher-order phonologicalprinciples that we have briefly exemplified in this section, each formhas a phonological function.
Most researchers seem to agree that the development of childrensspelling shows the acquisition of higher-order principles to be a laterachievement than the discovery of the alphabetic principles. There aretwo important sources of evidence for the learning of higher-orderspelling principles: The appropriate use of these principles in pseudo-words and their creative, though incorrect, use in the spelling of words.
Research by George Marsh and his colleagues (Marsh et al. 1980)has shown that childrens reading and spelling of pseudowords changesfrom relying on the alphabetic principle to drawing also on higher-orderspelling principles. Higher-order principles are, in fact, unavoidable inEnglish because, as pointed out earlier on, there are not as many lettersas there are phonemes in English. So, some phonemes are presented by combinations of letters and introduce a higher-order principle. If a child is asked to spell a pseudoword that contains one such sound, forexample the vowel sound in the word goat, we can test whether thechild is using a higher-order principle in spelling. Because pseudowordsdo not exist in the language, some argue that there is no correct spellingfor them. This is so, but only to a certain extent. If you were asked tospell a pseudoword that contains the vowel sound like the one in goat,you could actually spell it correctly in different ways (for example, you could write floab or flobe), but you would not represent thissound correctly if you tried to use a one-to-one correspondence betweenletters and sounds. If you spelled it as flob, your spelling would have to be considered incorrect, because the vowel sound would havebeen incorrectly represented. Because pseudowords cannot have beenencountered before, as they are not part of the language, a child thatadequately represents this vowel sound in a pseudoword is usingknowledge of regularities in spelling that go beyond lettersoundrelations. The child cannot be using the specific memory of a wordlearned previously, whether this memory was visual or kinetic.George Marsh pioneered the use of pseudowords in the investigation of childrens use of higher-order principles in spelling and showed that younger children inadequately used simple lettersound corre-
170 What are the overall implications?
spondences to spell all pseudowords, whereas older children wouldadequately use these higher-order spelling principles and differentiatebetween pseudowords such as flob and flobe. His observations havebeen confirmed many times since (see, for example, Templeton andBear 1992).
Childrens errors in the use of higher-order principles also tell us abouttheir understanding of the insufficiency of the alphabetic principle inEnglish orthography. Our view is that in order to use the digraphs oaor o+C+e (o plus consonant plus e) and spell floab or flobe,it is not sufficient to learn that these are possible spelling patterns: It isalso necessary to connect them to a specific vowel sound, that is, tolearn their function. This is no small accomplishment because there areso many digraphs in English. Thus, it is quite possible that children learna new pattern without being completely clear about what its functionin spelling is. Figure 7.1 shows the writing of a 6-year-old boy whoseems to attribute to the digraph ck the function of the split digraphV+C+e (vowel plus consonant plus e). He is certainly using itcreatively in the sense that he has generated novel spellings for thewords like, take and make. He is also using this pattern consis-tently, with a specific function, and is generalizing this across the words.He has learned the form but has misunderstood its function. Becauseform and function are tightly connected in spelling, learning the formis not sufficient.
To summarize, we have argued that analyses of the development ofchildrens spelling reveal that they go beyond the alphabetic principleby learning higher-order principles, which are generalizations, ratherthan sets of specific rimes with their spellings. Their mistakes in the use of the new forms that they learn show that there are two aspects
Morphemes and literacy: Context and conclusions 171
Figure 7.1 Writing of a 6-year-old boy who seems to attribute to thedigraph ck the function of the split digraph V+C+e.
that children have to master in the acquisition of higher-order spellingprinciples: The new forms and their functions. Learning the forms only (visual learning only) does not suffice for mastery. Form andfunction are essential for the mastery of higher-order spelling principles.Children are also able to use these generalizations in spelling pseudo-words, thereby showing that their learning goes beyond word-specificlearning.
Higher-order spelling principles in teaching spelling
We have argued that the ideas about what children need to learn haveshifted, from words to relations between the spellings of different words.But deciding that children should learn higher-order spelling principlesinstead of lists of words does not solve the question, How can we besthelp children to learn these relations between the spellings of differentwords?
One approach that seems, even if implicitly, to be favored by manyis the idea of visual learning. Many references to teaching letter stringsand attention to visual patterns are common in books about theteaching of spelling (for example, Mudd 1994: 148). Some authors goon to recommend that the children say the words whereas others areadamant that the children should simply try to learn the visual patternsbecause of the inconsistencies of English. The use of letter stringswithout attention to the function of higher-order forms can lead toproblems: Words that have the same letter strings for different reasonscan be put together, and this will make it more difficult for children to understand the functions of particular forms. One example of this use of letter strings is to treat -ough as a visual unit, a letter string,irrespective of whether it is followed by a -t or not. Using thisprocedure, a teacher compiled the word list rough, cough, thoughta list that does not help the children to understand the possible functionof -gh at the end of words as opposed to -ght.
However, it has become increasingly common since the work ofHenderson, Venezky and their colleagues to use form and functiontogether in the study of words. We carried out a literature search to findout whether there is support for the idea that children should learnvisual patterns without learning their phonological function. We couldfind no empirical studies where a visual without phonology conditionwas included in teaching. So, the claim that children should learn formsvisually without learning their function seems not to be widely accepted,at least by educational researchers.
172 What are the overall implications?
What differences are there between the methods proposed to getchildren to learn more about higher-order spelling principles? Manyseem to have a complete horror of the idea of teaching children rules.Henderson argued that it is not rules that children need but experi-ences. Their capacity as human learners will bring them to a feel for, or tacit knowledge of, words long before they are able to understand the rules (1990: 59). So, according to this view, teachers shouldassemble for study those words that exemplify patterns in order toprovide children with experiences about what these patterns mean.Although Henderson spoke of spelling patterns, he insisted that thesewould have meaning only if their function was understood. Hendersonwrote extensively about what children should study, but he gives noclear indication for how they would be able to learn, beyond the notionthat they should be given experience.
Following on Hendersons work, Jerry Zutell (1993) proposed theDirected Spelling Thinking Activity method, where children are pre-sented with contrasts rather than patterns and encouraged to analyzethe phonological differences between the contrasting spelling patterns.Do we know whether it is better to learn patterns separately or to learnthe contrast between them?
Our own research (Nunes and Bryant 2006) has shown that childrenlearn higher-order principles better if they are exposed to the contraststhan if they analyze each spelling pattern separately. This was true evenwhen they have the same amount of experience with each patternduring the lessons. Children who first studied short vowels (for example,a list of words such as hat, lap, lad, clam) and then long vowels(for example, a list of words such as hate, ape, blade, lame)separately learned less and retained less over time than those who hadthe opportunity to study the contrast between the two types of spellingpatterns. The possibility of establishing relations between differentspelling patterns and their functions helped retention over a 10-weekperiod.
The importance of establishing connections between forms andfunctions at a more abstract level is supported even further by anotherstudy that we conducted on the use of different sequences for theteaching of spelling principles. Our collaborators from Lauriston PrimarySchool taught some children the hathate contrast on one day andthe hophope contrast on a subsequent day. They taught anothergroup of children the hathate contrast on one day and on thesubsequent day the children learned the distinction between using -ckand -k at the end of words (for example, pickpink, lockhonk,
Morphemes and literacy: Context and conclusions 173
packpark). The first group of children learned two contrasts thatcould be brought under the same higher-order spelling principle, theV+C versus V+C+e distinction; the second group learned unrelatedprinciples. The children who learned two related principles profitedfrom establishing this relation between the two patterns: They spelledwords with a+C and a+C+e better than the children in the secondgroup, although they had the same amount of experience with thehathate contrast. This is, in fact, our reason for referring to thesegeneralizations as higher-order spelling principles: It is possible tocategorize them in more abstract terms and to use these abstractcategories in our understanding.
So, we seem to know something about how to help children to learnspelling principles.
It is better to help them to understand the functions of thesedifferent forms than to let them learn the forms without a focus ontheir functions.
Children learn more from establishing relations between differentforms with similar functions than from learning unrelated sets offorms and functions.
Higher-order spelling principles based on meaning
The higher-order spelling principles that we have discussed up to thispoint in our analysis of what children need to learn are connected tophonology. Phonology is important for spelling and it is not surprisingthat most of the research on spelling and most of the guides to teachingspelling focus on phonological spelling principles. But, as the researchthat we present in the previous chapters shows, these principles do nottell the whole story of English orthography. There is a connectionbetween morphemes and English orthography, and this connection canbe used by children in learning to spell.
The idea of using morphemes and meaning in teaching spelling hasreceived much less attention than higher-order phonological rulesand, when it did, we think that there was something missing.
Among the pioneers of the idea that systematic instruction onmorphemes could benefit spelling are three U.S. educatorsDixon,Henderson, and Henrywhose ideas have been used in the teaching ofspelling and have been subjected to some empirical test.
Dixon (1979) developed an extensive program for teaching spelling
174 What are the overall implications?
using morphemes, Corrective Spelling through Morphographs, whichincluded 140 lessons to be taught over a period of 8 months. The termmorphograph was created and used to refer to the smallest unit of meaning identifiable in written English, perhaps shifting the focusfrom the commonalities between oral and written language to anexclusive focus on spelling. Morphographs include stems and affixes.The focus of the lessons in Dixons program is on teaching rules foranalyzing words into morphographs, learning the meaning of thedifferent affixes, and learning rules for combining morphographs inorder to spell words correctly. The program was evaluated by Robinsonand Hesse (1981) using measures that were developed to assesswhether it attained its aim to teach students how to analyze words intomorphographs and whether this improved the students spelling ofpolymorphemic words. A general spelling test was also included. Thecomparison between the taught and control groups, which included atotal of 172 students in their seventh year in school (approximately1314 years of age), showed that the intervention students scored muchbetter in the tests of rule application and morphographic analysis. Theyalso scored better in the spelling test that contained words selected bythe authors because they were formed with the morphographs that thestudents had learned. However, the benefit of learning these rules andmorphographic analysis did not translate into significant differencesbetween the intervention and control students in a general spellingachievement test (the Stanford Achievement Test) although more thanhalf of the words in this standardized test were either part of theteaching program that had been followed by the students or containedmorphographs that had been covered in the program. After 140 lessonsand 8 months of work, this can hardly be considered an encouragingresult.
We think that this program is an example of teaching that includes agood idea about what children need to learn but it falls into difficultywhen it comes to the crucial question of how children should learn.Research on expertise in different domains of psychology has shownthat it is possible to learn rules that describe a skilled persons behaviorwithout benefiting from this knowledge when implementing the skill(see, for example, Anderson 1981 for a compilation of papers on thedevelopment of expertise).
Hendersons approach on what to teach was quite similar to Dixons,but the two of them had radically different ideas about how to teach:Henderson was entirely opposed to the idea of teaching rules.
Henderson proposed a meaning principle for spelling stems: If two
Morphemes and literacy: Context and conclusions 175
stems have the same pronunciation and different meanings, they areusually spelled differently, but stems with the same meaning anddifferent pronunciations are typically spelled in the same way. The firstpart of his rule refers to homophones (different words that sound thesame), such as weak and week, maid and made. The second partrefers to stems and derived forms, such as heal and health, magicand magician.
Henderson focused many of his suggestions for the teaching ofspelling on getting children to understand the constancy of the root.This is quite a complex process because sometimes the pronunciationof the stem changes and nothing is done to its spelling in order tomaintain the pronunciation, whereas sometimes letters are doubled or deleted in order to preserve the pronunciation. For example, whenwe add -ed to stop, we also add a p to maintain the pronunciationof the vowel: We would read stoped differently from the way that we read stopped. However, if we add -ed to verbs ending in e, wedelete the final e from the stem in order to preserve the pronunciation(as in deleted), because the double e would move the stress to thesuffix (think how you would read deleteed), and this suffix is neverstressed.
Henderson paid much less attention to prefixes and suffixes than tostems, probably because he considered affixes to be relatively unprob-lematic. The exception was the -ed for past regular verbs. He diddescribe a progression in childrens use of the -ed ending, which hasbeen confirmed by many since, and which we have ourselves observed:children seem not to use the -ed ending much at first, then they useit as a form without knowing its correct function, and later they restrictit to the appropriate places (see Chapters 1 and 2). But Henderson didnot consider that it might be necessary to increase childrens awarenessof morphology and grammar in order for them to master the use of thepattern with meaning. He seemed to take for granted that childrensimplicit knowledge of morphology and grammar would suffice for themto know when and where to use the -ed ending and focused on whathappened to the spelling of the root. He did suggest an interesting usefor suffixes in raising childrens awareness of their own grammaticalknowledge. The activity that he proposed was to ask children to place-ed or -s at the end of words and then discuss why these endingsmight not fit with certain words: For example, you can add s to maidbut not to made (his example, which we think illustrates that his focuswas on the stem; you cant add either -s or -ed to made because itis already a verb in the past).
176 What are the overall implications?
Henry (1989) also made a direct contribution to the idea thatmorphology is important for literacy teaching. She argued for teachingchildren, particularly those with reading problems, lists of prefixes andsuffixes so that they could be treated as units in word recognition andin spelling. She also focused on the etymology of morphemes, arguingthat there are common spelling difficulties in morphemes that have the same origin: for example, Greek morphemes may have a silent pat the start, as in pneumonia and psychology. When she refers tosuffixes, she treats them more as units of pronunciation than as units ofmeaning. For example, she refers to -tion and -cian as suffixes, andthus she focuses on the syllables formed by the suffix and the final letterin the stem. She does not discuss the function of these suffixes.
This past work offers a starting point for the idea of using morphologyin literacy teaching. However, there are many issues that require furtheranalysis, as our research has shown.
The first weakness of the proposals that we have just reviewed is thatthey take childrens awareness of morphology for granted. In order toimplement teaching that relies on childrens awareness of morphemes,we must be able to also promote it. This lesson was learned aboutchildrens awareness of phonology and should be also learned nowabout childrens awareness of morphemes and grammar. Implicitknowledge of morphemes and grammar is not a solid enough basis forthe teaching of spelling. Our own work on the use of apostrophes, asilent morpheme whose use is entirely determined by grammar, showshow difficult it is for children to use an apostrophe in the absence ofexplicit awareness of its function, and how much progress children canmake by becoming aware of grammar and the function of apostrophesto indicate possession (as in the mans house) (Bryant et al. 2000,Bryant et al. 1997).
A second weakness of these proposals is that the suggestions areabout what to teachlists of stems, prefixes, and suffixes are offeredbut there is a scarcity of research on how to teach these things. Thiscould be the result of taking childrens awareness of morphemes forgranted. But we have witnessed how the treatment of prefixes andsuffixes as letter strings without reference to meaning can hinder, andnot help, childrens efforts at understanding language. Teaching theprefix un- has become part of the curriculum in the U.K., but this isnot necessarily done in the context of raising childrens awareness ofmorphemes. We have seen some children, supporters of the footballteam Manchester United, struggle when asked to read the word unitedby itself on a page. They often wrongly treat the un- as a prefix and
Morphemes and literacy: Context and conclusions 177
then cannot figure out what to do with the rest of the word. We thinkthat more explicit knowledge of morphemes would have been helpfulto these children; the isolated knowledge of un- as a prefix was moreof an obstacle.
A third, and very significant, weakness is that the evidence for theeffectiveness of these practices was weak, and sometimes there wasnone at all. In education, as in many other applied settings, we mighthave knowledge of what works, but not why; or we might haveknowledge that leads us to expect that something will work and why,but dont know whether it will actually work in the classroom. In thecase of the connection between morphemes and literacy, there wereprograms that could be used in the classroom, but there was no evidencethat they worked. There were also suggestions from linguistic andpsychological studies that teaching children about morphemes couldbe a good thing, but the path from the laboratory to the classroom hadnot been traced.
Finally, the outcome measures used in past research were limited. If children perform better when spelling words that they were taught,we dont know whether they learned the specific words or whether they learned principles. A program for developing childrens use ofmorphology should provide evidence for the use of morphologicalprinciples, not simply for learning specific words.
The contribution of our own project
Our own research program has changed much of this picture insignificant ways. First, we did not take childrens explicit awareness ofmorphemes for granted. Our previous research, described in Chapters1 and 2, had shown that childrens explicit knowledge of morphemesand grammar differs from their implicit knowledge. When morphemesare part of the regular functions of spoken language, children arereasonably good at using them creatively, as they did when theyinflected pseudowords in our version of the Berko task. In contrast, inorder to spell past regular verbs with -ed, without ever using the -edending in other words, children would have to be able to represent thecategory of past verbs in a highly explicit form. We showed in Chapters1 and 2 that there is a considerable difference in the difficulty of thespoken and the written tasks.
Second, our contribution to promoting childrens awareness ofmorphology and grammar differs radically from that used in the pre-vious studies and, as our results show, this conceptual advance was
178 What are the overall implications?
translated into practical benefits for the children. Dixon and Hendersondiffer in their ideas about how to teach, but they seem to have a similarconception of what it means to make knowledge explicit: For them, itmeans to put knowledge into words. As our studies have shown, inparticular Chapter 2, adults might spell pseudowords correctly usingmorphological principles and be quite unable to explain why theyspelled the pseudoword in the way they did.
Our approach to how to teach in order to develop childrensawareness of morphemes was to engage the children in a variety ofproblem-solving activities that required different operations of thought.We did not design tasks that required the same response over and over,produced by the application of one and the same rule. We already knewfrom our own research that the same amount of experience has differentresults depending on whether the children applied the same rule overand over or whether they had to make decisions about the differentspelling possibilities.
In order to implement these ideas, we developed several tasks that,our evidence shows, encourage children, in a way that is interesting tothem, to perform various operations on morphemes. These tasks demandthat children make some explicit representation of the morphemes,because they require children to think about the morphemic structureof words rather than simply to use them.
Analogical reasoning is one of the operations of thought that we used in order to promote (see Chapter 3) as well as to measure (seeChapter 1) childrens awareness of morphemes. In analogy tasks, thechildren are given one pair of words (for example, readreader)and asked to produce the missing word to complete the pairmagic?. Phonological strategies, such as rhyming, cannot lead tothe correct answer. In order to solve this analogy problem, childrenneed to realize that there is a morphological transformation from thefirst to the second word in the first pair, which results in forming a person word, and then apply a similar transformation to the first inthe second pair.
Counting the number of morphemes in particular words (see Chapter3) is another operation that requires children to make morphemes intoan object of thinking: How many morphemes are in unforgettable? It is difficult to count something if you do not become explicitly awareof the items that you are counting.
Putting morphemes into categories (see Chapter 3) also requiresawareness of morphemes: If you have to sort words into those thatcontain suffixes that form person words and those with suffixes that
Morphemes and literacy: Context and conclusions 179
form other words, you need to think about the words differently fromthe way that you think about them when you are just using them inspeaking.
Subtracting morphemes from pseudowords (see Chapter 3), we think,also requires greater awareness than inflecting pseudowords. When we asked children to decide What jobs do these people do on Mars?,and they needed to take away the suffixes in spamters or montists,they would have to be pretty clear about the boundaries of the suffixesin order to arrive at the base word.
We found that all of these tasks challenged the pupils whom we were working with, but did not daunt them. We also established that it was useful to discuss childrens solutions in the classroom with themand their classmates and to provide them with a metalanguage thatrepresented the elements of language under discussion and thuspromoted the development of explicit awareness of morphemes andgrammar. Discussion does help to put thoughts into words, but it is notthe memorization of a rule.
We think that this is a major contribution to a new way of thinkingabout teaching the use of morphemes and grammar in spelling. Throughthese games, teachers now have the support that they need for engagingchildren in a variety of intellectually challenging and motivatingactivities, which draw on much more than memory and definitions.These activities draw on childrens experiences and on their reasoning.There is, we believe, no need to be concerned about boring the children.The classroom activities that we have described in Chapters 36 alwaysattracted and maintained childrens involvement and interest.
The use of a variety of intellectual operations is probably also one ofthe reasons the program worked. Psychological research, referred to inChapter 3, suggests that the key to generalization is variation. We thinkthat the variety of learning tasks played an important part in improvingthe childrens understanding of the relatively abstract morphemicspelling principles that they were learning about.
We have strong evidence to show that the program that we designedworks in promoting childrens use of a new set of spelling principles. Ourlaboratory experiments were designed with the greatest care, using thehighest scientific standards: Random assignment to conditions, teachingprocedures described in detail, control groups exposed to appropriateexperiences, and relatively large samples so that the danger of rejectinggood hypotheses was minimized and the possibility of generalizing theresults was maximized.
180 What are the overall implications?
Our outcome measures give further weight to the nature of theevidence for the validity of our interventions. Childrens improvementin word spelling is undoubtedly a main measure: This is one of the outcomes that we expect of education. However, by itself, thismeasure does not tell us whether the children learned isolated wordsor a principle. The fact that we have significant results in childrensspelling of pseudowords, and that, when it was possible to obtaindelayed posttest measures, these tended to remain significant, gives us confidence in the childrens learning of a principle rather than the acquisition of isolated spellings of words. Our use of an outcomemeasure that assessed childrens vocabulary growth, with a comple-mentary pseudoword-definition measure, shows that learning aboutmorphemes has the potential to result in a wider impact on childrensliteracy. Childrens vocabulary is consistently related to their literacyperformance, and so it could be expected that, in the long run, thebenefits of learning about morphemes would show up in other measuresof literacy. However, we think that this would be more likely if aconsistent program of study were part of the curriculum rather than a set of lessons, as we used in this study.
The potential benefits of including the teaching about morphemes inthe curriculum could go well beyond literacy learning in the firstlanguage. Our past research (Castro et al. 2004) has shown that youngPortuguese students awareness of morphemes and grammar in theirfirst language is a predictor of how well they perform in learning asecond languagein the case of our research, English. These studentsreceived the same teaching of English by the same teacher, but thosewho at the start of the year had better awareness of morphology andgrammar in their first language learned the second language better thanthose who had lower levels of awareness. Thus, children can transferawareness of grammar and morphology from their first language to thesecond language they learn.
In an interview (presented in Chapter 6), a teacher from Londondescribed how she had had no awareness of English morphology untilshe started studying French! She was able to transfer what she learnedabout French grammar and morphology back into her own language,English, and this is what informed her knowledge of morphemes inEnglish.
So we know that teachers, as well as children, can use their knowl-edge of morphology across languages. The effects of this transfer arebound to be positive.
Morphemes and literacy: Context and conclusions 181
182 What are the overall implications?
We think that there is much more that can still be done in designingactivities for teaching children about morphemes, and we wouldlike to end with an idea that still needs testing. Think of the wordpairs tidetied and guestguessed. They are homophonepairs, so they sound the same. However, their spelling can bedistinguished on the basis of morphology. In the sentences I trippedbecause my shoes were not tied and I did not know the answerbut I guessed, only one of the spellings could be correct. The wordstied and guessed are two-morpheme words and the suffixes arethere for grammatical reasons. The use of morphology should makeit easier for us to distinguish between these homophone pairs thanto distinguish between weakweek or greatgrate becauseboth words in these latter pairs are one-morpheme words. Somerecent research of ours (Nunes and Bryant 2006) has shown thatchildren aged 7 or 8 cannot use their knowledge of morphemes toimprove their ability to distinguish between homophones when oneof them is a two-morpheme word. Even with older children, aged 9 or 10, the evidence is not unambiguous, and though they mightdo better on some measures, other measures do not show thedifference. Can children learn to use their knowledge of morphemesto distinguish between the homophone pairs where this knowledgeis relevant?
We conclude by saying that this book has provided solid evidenceto suggest that English children benefit from becoming more awareof morphemes and grammar. This awareness enhances both theirliteracy and their spoken vocabulary. The path from the laboratoryto the classroom has been opened so that classroom practice andpolicy can be put in place. The difficulties that have been associatedwith the idea of teaching rules to be memorized and with boredomno longer need to be an obstacle. We know that there are benefitsto be harvested, but we dont yet know how far they extend. In the preceding section, we touched on further possible benefits thatcan be obtained in second-language learning from awareness ofmorphology and grammar. There is a lot to be gained by improvingthe teaching of English.
The four research strategies in this research program
All of our proposals about teaching reading and spelling and morphol-ogy stem directly from the evidence that we have collected over theyears on childrens morphological knowledge. The strength of our case for change, therefore, depends on the quality of this evidence, andthis in turn depends on the power and suitability of our researchmethods. In this appendix, we will summarize our methods and ourreasons for adopting them. Our aim is to show you that these were theright methods and that they provided good evidence for the worth ofour educational proposals.
The main part of our book deals with a set of research studies thatwere part of the ESRCs Teaching and Learning Research Programme(TLRP). However, quite naturally, these TLRP studies of ours werepreceded by other studies from our research group that led to thequestions which prompted our TLRP research. Part I of our book dealswith some of the earlier work, and Part II describes the TLRP inter-vention studies.
Together, the studies that we describe in Parts I and Part II add up toa research agenda that combines several different research strategies inorder to study morphological knowledge and its relation to reading andspelling. In the end we adopted four different strategies in differentparts of our research.
Strategy 1: Using the longitudinal method toestablish the connection between childrensknowledge about morphemes and thedevelopment of literacy
We started our research on children and morphemes with a causalquestion. Does childrens morphological knowledge affect their progress
Authored by Peter Bryant and Terezinha Nunes
in learning to spell? Our hypothesis was that morphological knowledgeis important even when, as was the case then, children are taught verylittle about morphemes. We decided that the best way to investigatethis hypothesis was to do a large-scale longitudinal study, that is, to seethe same children several times over a long period and to look at therelation between their morphological knowledge and their spelling.
Longitudinal studies are an essential part of testing causal hypotheses.If, for example, childrens morphological knowledge affects how theylearn to spell, then their scores on a test of morphological knowledgeshould predict how well they spell later on. The results of our openinglongitudinal study established this predictive link. The childrens scoresin the initial tests of morphological knowledge did predict their spellingperformance over a year later, even when our statistical analysescontrolled for differences in their age, IQ and their original spellinglevels. Thus this longitudinal study established the relevance ofmorphological knowledge to childrens literacy.
Strategy 2: Using laboratory interventions toimprove childrens morphological knowledge
The longitudinal study established an important connection betweenchildrens morphological knowledge and their spelling. Our idea is thatthis connection is a causal one. In other words, we think that childrensknowledge of morphology in spoken language determines how wellthey learn and use morphological spelling rules in written language. Inour laboratory intervention studies we looked at this possible causalconnection in two ways.
The first way was to use the intervention method to see if this causalconnection really does exist. It may seem obvious that cause and effectmust go in this way, but it might not be true. Some other factor, suchas childrens general linguistic development, might determine both howwell they understand morphology in spoken language and also howquickly they learn morphological spelling rules. To establish a causallink such as the one that we had in mind, one must resort to a differentmethod, which is to carry out one or more tightly controlled interventionstudies.
The rationale for doing so is a simple one. Suppose that childrensmorphological knowledge really does determine or partly determinehow well and quickly they learn about morphologically determinedspellings. In that case one effect of an increase in a childs morphological
184 Four research strategies
knowledge will be a consequent improvement in this childs use ofmorphological spelling rules. If you were to teach children effectivelyabout morphemes, you should also be helping them with spellingatany rate if the hypothesis is right.
Therefore, the questions that we asked in our first interventionexperiments were whether it is possible to enhance morphologicalknowledge and what effect this would have on their reading andspelling. Notice that this is not a directly educational question. We setup several traditional intervention experiments (described in Part I ofthis book), which is the term that psychologists use to describe studiesin which they directly intervene by teaching at least one group ofparticipants and then comparing them to at least one other group whohave not received this instruction.
It is not our purpose here to describe either the instruction that wegave or the results of these studies. In this appendix our concern is withthe design of the studies that led to our conclusions. Intervention studieshave to include at least three phases.
1. The first of these is the pretest, in which all the participants aregiven the same set of tasks, including tasks that test the skill, forexample, of spelling the sort of polymorphemic words that theintervention is to be aimed at.
2. The next phase is the intervention itself, which might last for a brieftime, like 1 hour, or might go over several weeks or even months.Only some of the participants are given this intervention and theyform the intervention or experimental group(s). The otherparticipants do not receive the intervention or instruction whichthe experimenters predict will enhance the skill in question, and are usually called the control group(s). There are two types ofcontrol group: Those that are given as much attention and as muchinstruction as the intervention groups and those (sometimes calledunseen control groups) that are not given any extra attention orinstruction. On the whole, the first type of control group makes fora more satisfactory comparison.
3. The final phase is the posttest in which the participants are giventhe same or the same kind of test of the skill in question as they were given in the pretest. In fact there can be more than oneposttest, and it is quite common in intervention studies to arrangefor two posttests, one an immediate posttest and the other adelayed posttest which the participants are given two months orso later.
Four research strategies 185
If the experimenters hypothesis is right, there should be moreimprovement from pretest to posttest among the participants in theintervention group(s) than in the control group(s). In the interventionstudies that we describe in Part I, we did find consistent signs of thispattern, and so we concluded that it is, in principle, possible to instructchildren about morphemes in a way that affects their reading andspelling.
This was not a claim about teaching in the classroom. During theintervention the actual instruction period was given for the most parton a one-to-one basis in these first intervention studies. Our aim was tofind out whether this instruction worked, before we looked at whetherit worked in the classroom.
Our second way of using laboratory intervention research to pursue ourhypothesis was to compare the effectiveness of different teachingmethods. Intervention experiments are excellent instruments, not justfor studying the effects of enhancing a particular ability, but also forcomparing different forms of intervention. This was the main purposeof the first intervention study that we describe in Chapter 3. Here, weused the method of laboratory intervention to settle a theoreticalquestion about explicit and implicit learning. (We give the reasons forasking this question in Chapter 3.)
We decided to give our intervention groups instruction about onehighly specific morphological spelling rulethe rule about -ion and-ian endings. Our reason for doing this was that we knew that thisspelling rule is never taught in schools in the U.K., and thus ourintervention would take the form of teaching children about somethingthat they had never been taught before.
We adopted the familiar pretestinterventionposttest (immediateand delayed) design, but we now had three intervention groups as well as a large control group. The three different interventions involved(1) entirely explicit instruction, or (2) entirely implicit instruction, or(3) a mixture of the two.
Thus we designed this experiment to settle not just whether instruc-tion is effective but also whether one form of instruction is better thananother. The answer to this second question depended on the relativesuccess of the three intervention groups in the two posttests. You canfind this answer in Chapter 3.
One other aspect of this study is worth mentioning. The study was stilla laboratory one, and a long way from the classroom, but we no longerstuck to one-to-one instruction. Instead, we taught the children in pairs
186 Four research strategies
and we encouraged them to discuss their solutions to the variousintervention tasks that we gave them among themselves. This was ourfirst move away from one-to-one intervention and toward instructionin groups.
Strategy 3: Using classroom interventions tofind out whether it is possible to repeat thesuccess of the laboratory interventions in aclassroom setting
Our laboratory interventions were about teaching, but children on thewhole are taught in classrooms. Although it was essential for us to dothe laboratory studies, it was just as important to check that our positiveresults could be repeated in the realistic context of the classroom. Wedid this in three ways.
One way was to repeat, fairly exactly, part of one of our laboratorystudies in a classroom setting. We did this in the second study that we describe in Chapter 3. We, and our teacher colleagues at Lauristonschools, wanted to find out whether the instruction, which hadenhanced childrens spellings of the -ion and -ian endings in ourlaboratory study, would work as well in the classroom.
So the Lauriston team suggested that they could take the mostsuccessful method, the explicit intervention, and give it to their studentsin their classrooms. The teams idea, which we enthusiastically accepted,was that they would use the same material, the same pretest, inter-vention period and posttest (immediate and delayed) design, and thesame intervention tasks as we had, and that they would also comparean intervention group and a control group of children who wereinstructed about something else during the intervention period.
A great deal depended on the results of this classroom study, since ifthe effects of the intervention had been different in the Lauriston studyfrom those in our laboratory study we would have had the difficult taskof working out the reasons for the discrepancy. So, we were gratifiedby the similarity of the results of the Lauriston classroom study to ourown intervention. This, it seemed to us, gave us a warrant to start full-scale intervention studies in school settings.
Our second way of adapting our laboratory intervention work to theclassroom was to broaden our teaching about morphemes. If schools doradically increase their teaching about morphology and morphological
Four research strategies 187
spelling rules, as we are now advocating they should do, this instructionwould have to be about a wide range of inflectional and derivationalmorphemes. Therefore, our next study (the third study in Chapter 3)was on teaching schoolchildren about morphemes in general and aboutseveral different morphological spelling rules.
This was also a classroom study, in which teachers gave students the actual instruction about morphemes and morphological spellingrules. Again we adopted a pretestinterventionposttest design andcompared an intervention group with a control group. However we, theresearchers, prescribed the methods and the timetable and we werealways at hand during the intervention period. Again our interventionwas successful.
Our third way of bringing intervention about morphemes and spellinginto the classroom was to study what would happen if we relinquisheddirect control of the intervention. In our other classroom research we exercised a great deal of control over what the teachers did, theorder in which they administered the interventions, and the timing ofthe different parts of the interventions. Even just by being availableduring the intervention period we had had an effect on what went onin the classroom. This level of control had been absolutely necessary atthis stage of the research, but it was not real life. If teachers are toadopt our methods generally, we will not be able to, or even want to,be in charge of the way that they apply these methods, and there arebound to be great differences between different teachers in how theydo so.
In the studies that we describe in Chapters 4 and 5, we stuck to thepretestinterventionposttest design, with comparisons between inter-vention and control groups, and, of course, we organized and carriedout the pretests and posttests ourselves. We delivered the interventionsthat we had devised to the schools with some explanatory material, butwe had no other control than this. Thus, this really was a realistic study.What happened in the intervention was what would normally happenwhen a school adopts a new method of teaching.
Strategy 4: Using interviews andintervention to study the teachers as well
In the three strategies that we have described so far, the focus of thework was always on the childrens morphological knowledge and onhow to enhance it, but we were aware all the while that there was
188 Four research strategies
another variable of great importance to take into account. In the end itwill be the task of teachers to tell children about the link betweenmorphemes and spelling and to interest them in this connection.
We needed, therefore, to pay attention to teachers as well, andparticularly to their interest in morphology and their knowledge ofmorphological spelling rules. Since at present morphemes and morpho-logical spelling rules are either not taught at all at school or taughtextremely patchily, we could not take this knowledge for granted inteachers. We had to find out how aware teachers are of morphologicalstructure and what effect their awareness has on the students whomthey teach.
The research strategy in our work with teachers, which we describein Chapter 6, was a two-pronged one. First, we investigated teachersknowledge of morphemes and of morphological spelling rules. We didthis by interviewing a group of experienced teachers, and the interviewsrevealed a marked lack of explicit knowledge about morphemes amongthese teachers.
Our second move was to devise a rather interesting form of theintervention study. The focus of our intervention was a group of teacherswho were enrolled on a course, which ran for several weeks, onmorphology and its impact on spelling. Before the course and at its end,we tested children in these teachers classes and other children whowere not taught by the teachers who took our course on morphemes.Thus, the intervention was given to teachers, but the pretests andposttests which told us about the effectiveness of the intervention weregiven to schoolchildren who either were or were not the students ofthese particular teachers.
In this study we showed that teachers themselves usually need someinstruction about morphemes, and we also demonstrated that it ispossible to give them the instruction that they need in a way that helpsthem and eventually their students as well.
Thus, to summarize, we decided on a combination of four differentresearch strategies to discover (a) whether there is a need to teachchildren systematically about morphemes and spelling and (b) how todo this teaching.
1. Our first research strategy was longitudinal research that we did inorder to establish that there is a connection between childrens
Four research strategies 189
morphological knowledge and their learning about morphologicalspelling rules.
2. Our second strategy took the form of a set of laboratory interventionstudies, and here we had two aims. One aim was to discover if thelink between childrens morphological knowledge and their spellingis a causal one. Our other aim was to compare the effects of explicitand implicit teaching about morphemes and spelling.
3. Our third strategy was a series of classroom intervention studies,the purpose of which was to discover whether the teaching methodsthat we had developed in the laboratory could be adapted to theclassroom.
4. In our fourth research strategy we turned to teachers and in partic-ular to their knowledge of the link between morphemes andspelling. We used interviews to assess this knowledge and then welooked at the effect of improving the teachers knowledge ofmorphological spelling rules on their pupils spelling.
It seems to us that none of these research strategies on its own wouldbe a good enough basis for applying the insights of psychologists toclassroom practice. It is the combination of the four different approachesthat allows us to claim that it is possible and highly worthwhile to teachschoolchildren about morphemes, and that there are no insurmountableobstacles to doing so.
190 Four research strategies
Aidinis, A. and Nunes, T. (2001) The role of different levels of phonologicalawareness in the development of reading and spelling in Greek, Readingand Writing, 14: 14577.
Anderson, J. R. (1981) Cognitive Skills and their Acquisition. Hillsdale, N.J.: L. Erlbaum Associates.
Berko, J. (1958) The childs learning of English morphology, Word, 14:15077.
Brown, R. (1957) Linguistic determinism and parts of speech, Journal ofAbnormal and Social Psychology, 55: 15.
Brown, R. (1973) A First Language. Cambridge, Mass.: Harvard UniversityPress.
Bryant, P., Devine, M., Ledward, A., and Nunes, T. (1997) Spelling withapostrophes and understanding possession, British Journal of EducationalPsychology, 67: 93112.
Bryant, P., Nunes, T., and Aidinis, A. (1999) Different morphemes, samespelling problems: Cross-linguistic developmental studies, in M. Harrisand G. Hatano (eds) Learning to Read and Write: A Cross-LinguisticPerspective. Cambridge: Cambridge University Press, pp. 11233.
Bryant, P., Nunes, T., and Bindman, M. (2000) The relations between readingability and morphological skills: The case of the apostrophe, Reading andWriting, 12: 25376.
Cardoso-Martins, C. (1994) Rhyme perception: global or analytical? Journalof Experimental Child Psychology, 57: 2641.
Carlisle, J. (1988) Knowledge of derivational morphology and spelling abilityin fourth, sixth and eighth graders, Applied Psycholinguistics, 9: 24766.
Carlisle, J. F. (1995) Morphological awareness and early readingachievement, in L. B. Feldman (ed.) Morphological Aspects of LanguageProcessing. Hillsdale N.J.: L. Erlbaum Associates, pp. 189209.
Casalis, S. and Louis-Alexandre, M.-F. (2000) Morphological analysis,phonological analysis and learning to read, Reading and Writing, 12:30335.
Castro, A., Nunes, T., and Strecht-Ribeiro, O. (2004) Relao entreconscincia gramatical na linguagem materna e progresso naaprendizagem de uma lngua estrangeira, Da Investigao s prticas.Estudos de Natureza Educacional, 5: 5166.
Chliounaki, K., and Bryant, P. (2003) Choosing the right spelling in Greek:Morphology helps, Revue Franaise de Linguistique Applique, 8: 14252.
Department for Education and Employment (1998) The National LiteracyStrategy: Framework for Teaching YR to Y6. Suffolk: DfEE Publications.
Department for Employment and Skills (2001) The National Literacy Strategy:Spelling Bank. London: DfES Publications.
Department for Employment and Skills (2003) Year 2 and Year 3 PlanningExemplification and Spelling Programme. London: DfES Publications.
Dixon, R. (1979) Corrective Spelling through Morphographs. Chicago, Ill.:Science Research Associates.
Ehri, L. (1997) Learning to read and learning to spell are one and the same.Almost, in C. A. Perfetti, L. Rieben and M. Fayol (eds) Learning to Spell.Mahwah, N.J.: L. Erlbaum Associates, pp. 23770.
Ehri, L. (1986) Sources and difficulty in learning to spell and read, in M.L.Wolraich and D. Routh (eds) Advances in Developmental and BehaviouralPaediatrics. Greenwich, Conn.: JAI Press, Vol. VII, 12195.
Elbro, C. (1989) Morphological awarness in dyslexia, in C. von Euler, I. Lundberg, and G. Lennerstrand (eds) Brain and Reading: Structural andFunctional Anomalies in Developmental Dyslexia. Basingstoke: Macmillan,pp. 27991.
Fayol, M., Hupet, M., and Largy, P. (1999) The acquisition of subjectverbagreement in written French: From novices to experts errors, Reading andWriting, 11: 15374.
Fayol, M., Thenevin, M.-G., Jarousse, J.-P., and Totereau, C. (1999) Fromlearning to teaching to learn French written morphology, in T. Nunes(ed.), Learning to Read: An Integrated View from Research and Practice.Dordrecht: Kluwer, pp. 4364.
Fernald, G. (1943) Remedial Techniques in Basic School Subjects. New York:McGraw-Hill.
Ferreiro, E. and Teberosky, A. (1983) Literacy before Schooling. Exeter, N.H.and London: Heinemann Educational Books.
Fowler, A. E. and Liberman, I. Y. (1995) The role of phonology andorthography in morphological awareness, in L. B. Feldman (ed.),Morphological Aspects of Language Processing. Hillsdale N.J.: L. ErlbaumAssociates, pp. 15788.
Frith, U. (1985) Beneath the surface of developmental dyslexia, in K. Patterson, M. Coltheart, and J. Marshall (eds) Surface Dyslexia. London:L. Erlbaum Associates, pp. 30130.
Gleitman, L. R. (1990) The structural sources of verb meaning, LanguageAcquisition, 1: 355.
Gleitman, L. R. and Gleitman, H. (1992) A picture is worth a thousand words,but thats the problem, Current Directions in Psychological Science, 1: 315.
Henderson, E. (1990) Teaching Spelling, 2nd edn. Dallas, Tex.: HoughtonMifflin.
Henry, M. K. (1989) Decoding instruction based on word structure andorigin, in P. G. Aaron and R. M. Joshi (eds) Reading and Writing Disordersin Different Orthographic Systems. Dordecht: Kluwer, pp. 2549.
Karmiloff-Smith, A. (1992) Beyond Modularity: A Developmental Perspective onCognitive Science. Cambridge, Mass.: MIT Press.
Kemp, N. and Bryant, P. (2003) Do beez buzz? Rule-based and frequency-based knoweldge in learning to spell plural -s, Child Development, 74:6374.
Levin, I., Ravid, D., and Rapaport, S. (1999) Developing morphologicalawareness and learning to write: A two-way street, in T. Nunes (ed.)Learning to Read: An Integrated View from Research and Practice. Dordrecht:Kluwer, pp. 77104.
Marsh, G., Friedman, M. P., Welch, V., and Desberg, P. (1980) Thedevelopment of strategies in spelling, in U. Frith (ed.), Cognitive Processesin Spelling. London: Academic Press.
Montessori, M. (1915) The Montessori Method. London: Heinemann.Mudd, N. (1994) Effective Spelling: A Practical Guide for Teachers. London:
Hodder & Stoughton.Nunes, T. and Bryant, P. (2006) Childrens Reading and Spelling: Beyond the
First Steps. Oxford: Blackwell.Nunes, T., Bryant, P., and Bindman, M. (1997a) Spelling acquisition in
English, in C. Perfetti, L. Rieben, and M. Fayol (eds.) Learning to Spell.London: L. Erlbaum Associates, pp. 15170.
Nunes, T., Bryant, P. E., and Bindman, M. (1997b) Morphological spellingstrategies: Developmental stages and processes, Developmental Psychology,33: 63749.
Nunes, T., Bryant, P., and Olsson, J. (2003) Learning morphological andphonological spelling rules: An intervention study, Reading and Writing,7: 289307.
Peters, M. (1985) Spelling: Caught or Taught? (A New Look). London:Routledge & Kegan Paul.
Qualifications and Curriculum Authority (1999) National Curriculum online,EN3 Writing. Available at . Accessed22 January 2006.
Read, C. (1971) Pre-schoolchildrens knowledge of English phonology,Harvard Educational Review, 41: 134.
Robinson, J. W. and Hesse, K. D. (1981) A morphemically based spelling
programs effect on spelling skills and spelling performance of seventhgrade students, Journal of Educational Research, 75: 5662.
Schonell, F. J. (1957) Essentials in Teaching and Testing Spelling, firstpublished in 1932. Basingstoke: Macmillan.
Templeton, S. and Bear, D. (1992) Development of Orthographic Knowledgeand the Foundations of Literacy. Hillsdale, N.J.: L. Erlbaum Associates.
Treiman, R. (1993) Beginning to Spell: A Study of First-Grade Children. NewYork: Oxford University Press.
Treiman, R. and Cassar, M. (1997) Spelling acquisition in English, in C. Perfetti, L. Rieben, and M. Fayol (eds) Learning to Spell. London: L. Erlbaum Associates, pp. 6180.
Treiman, R., Mullenix, J., Bijeljac-Babic, R., and Richmond-Welty, E. D.(1995) The special role of rimes in the description, use and acquisition of English orthography, Journal of Experimental Psychology-General, 124:10736.
Venezky, R. L. (1970) The Structure of English Orthography. The Hague:Mouton.
Venezky, R. L. (1995) How English is read: Grapheme-phoneme regularityand orthographic structure in word recognition, in I. Taylor and D. R.Olson (eds) Scripts and Literacy. Dordrecht: Kluwer, pp. 11130.
Zutell, J. (1993) Directed spelling thinking activity: A developmental,conceptual approach to advance spelling word knowledge, paperpresented at the Nineteenth National Conference of the Australian ReadingAssociation, Melbourne, Australia.
ability 11719, 12830, 1312-able 7abstract nouns 67, 98, 107, 108achievement groups 11719,
12830, 1312adjectives 6, 99affixes 47; see also prefixes; suffixesage: conserving stems 534, 55;
different types of linguisticknowledge 489; effect ofclassroom intervention onspelling 105, 106; spelling ofpseudowords with -ian and -ion578; spelling of suffixes and389, 41, 456, 55; and use ofmorphemes 10; use of s as amorpheme marking plural andthird-person singular of verbs 60
agents 88Aidinis, A. 20alphabetic conception of writing 166alphabetic principles 168, 169, 170;
childrens understanding ofinsufficiency of 171
alphabetic stage (decoding letter byletter) 144
analogy tasks 90, 179; laboratorystudy 714, 76; sentence analogytask 13, 29; transformingpresent-tense verbs into past-tense verbs 1011; word analogytask 1314, 29
analyzing novel words 90Anderson, J.R. 175
apostrophes 22, 144, 177autoregression 30
bases (base words) 5, 6, 88Bear, D. 171Berko, J. 12, 1415, 47Birgisdottir, F. 56Braxton, T. 4Brierley, V. 56British Picture Vocabulary Scale
(BPVS) 126Brown, R. 8, 10Bryant, P. 17, 20, 59, 169, 173, 177,
Cardoso-Martins, C. 168Carlisle, J. 14, 245Casalis, S. 25Cassar, M. 144Castro, A. 181categories, grammatical 88, 90, 92,
978categorization 90, 17980cause and effect 1846, 190; in
connections betweenmorphological knowledge andlearning to spell morphemes2432
CD-ROM 104, 106, 123, 128, 161Chliounaki, K. 20classroom interventions 1878, 190;
Lauriston Primary School study816, 102, 160, 187; teachingpackage see teaching package
contamination effects 84contrasts 867, 169, 1734correction game 746, 76, 90Corrective Spelling through
Morphographs 1745counting the number of morphemes
delayed posttests 185; classroomintervention and vocabulary1248, 12832; laboratory study689, 78, 789, 801; LauristonPrimary School study 834, 85;teaching package 94, 956,1001
Department for Education andEmployment (DfEE) 134
Department for Education and Skills(DfES) 122, 134
derivations 6, 7, 10, 15, 25, 878,103, 176
development: of spelling 1667;stages of for reading and spelling144
dictionaries 145differentiation 150, 151digraphs 171dinosaur names 514Directed Spelling Thinking Activity
method 173Dixon, R. 1745, 179
-ed ending 7, 55, 176; changes inspelling over time 259;development of spelling 1667;spelling 3640, 447; teacherslack of awareness of -ed rule137; teachers strategies forteaching spelling 1367
educational theory and practice15782; childrens understandingof English spelling 1667;contribution of classroom spellingproject 17881; higher-orderprinciples based on meaning1748; higher-order spellingprinciples in teaching spelling1724; need to teach spelling1625; path from laboratory to
classroom 15962; spelling unitsor spelling principles 16872
Ehri, L. 144, 166Elbro, C. 25-en ending 92, 989English as an additional language
151-er ending 17, 88errors in spelling 135; analysis of
errors and spelling suffixes 424,45, 46; and development ofspelling 1667; use of higher-order principles 171
ESRC Teaching and LearningResearch Programme (TLRP) 183
experimental group(s) 1856expertise 175explanations for spellings of
pseudowords 5561explicit knowledge of morphemes
924, 158, 1789; as essentialingredient of learning to read andwrite 1624; measuringmorphological awareness 1015;teachers 13542, 145, 1534,189
explicit teaching 86, 87; laboratorystudy 66, 67, 68, 73, 75, 767,7881, 1012; Lauriston PrimarySchool study 816
facts, learning specific 90Fayol, M. 23feedback 924, 97Fernald, G. 165Ferreiro, E. 166form and function 170, 1712,
1724Fowler, A.E. 14, 25, 51French 6, 234, 150, 181Frith, U. 144function and form 170, 1712,
generalization 902; see alsoprinciples
Gleitman, L.R. 8grammar 79; importance of
developing awareness of 879;teachers knowledge of 150
grammatical categories 88, 90, 92,978
graphic units 1689Greek 6, 201
Hebrew 312Henderson, E. 159, 165, 167, 168,
173, 174, 1756, 179Henry, M.K. 174, 177Hesse, K.D. 175high achievers 11719, 12830,
1312higher-order spelling principles
16978; based on meaning1748; in teaching spelling 1724
Hillingdon Cluster of Excellence 122homophones 176, 182Hupet, M. 23
-ian endings 17, 98, 141, 186;effect of classroom interventionon spelling 1079; pseudowordswith 568, 701, 78, 801, 85;spelling -ion and 404, 467;teaching pairs of children about difference between -ionendings and 6581, 86, 1012;teaching about in the classroom816, 102; teaching package 97,98
imagination, importance of engaging92, 93
immediate posttests 185; classroomintervention and vocabulary1248, 12832; laboratory study689, 77, 789, 801; LauristonPrimary School study 834, 85;teaching package 94, 956,1001
implicit knowledge of morphemes 9,1016, 1578, 177
implicit learning of spelling 163implicit teaching 667, 67, 68, 73,
75, 77, 7881
individualized word lists 164infixes 4inflections 45, 6, 10, 15, 25, 489,
88internal consistency of tests 1246,
126intervention group(s) 1856intervention studies 1856-ion endings 17, 98, 186; effect of
classroom intervention onspelling 107, 108; knowing whento use 404; pseudowords with568, 701, 78, 801, 85;spelling suffixes 3644, 467;teachers strategies for teachingspelling 1412; teaching about inthe classroom 816, 102;teaching pairs of children aboutdifference between -ian endingsand 6581, 86, 1012; teachingpackage 97, 98
irregular verbs 456
Jarousse, J.-P. 23journals, spelling 146
Karmiloff-Smith, A. 47Kemp, N. 17kinetic learning 145knowledge of morphemes 3562,
177; different types of 4750;explicit see explicit knowledge ofmorphemes; implicit 9, 1016,1578, 177; and literacy learning1578; longitudinal studies1834, 18990; spelling suffixes3650; spelling words with thesame stem 505; spellings ofpseudowords and 5561;teachers 13542, 145, 1534,189
laboratory study 6581, 86, 1012,160, 180; children 678;posttests 68, 6971, 778, 789,801; pretest and posttestmeasures of spelling 6971;research strategy 1847, 190;results 7881; study design 689;
two teaching and learningsessions 717
Largy, P. 23Lauriston Primary School study
816, 102, 160, 187letter patterns 467, 13940, 145,
1724letter strings 44, 13940, 145, 1723Levin, I. 312Liberman, I.Y. 14, 25, 51literacy: connection between
morphemes and 2432, 158;development of 1834; see alsoreading; writing
literacy hour 152logographic (whole-word) stage 144longitudinal method 1834, 18990;
morphological knowledge anddevelopment of literacy 2932
look-cover-write-check method 145,165
Louis-Alexandre, M.-F. 25low achievers 11719, 12830,
1312-ly ending 99
Marsh, G. 1701meaning: higher-order spelling
principles based on 1748;morphemes as units of 34;teachers strategies for teachingspelling 13840
memorization 1634, 165, 167memory cues 1512missing word tests see sentence
completion tasksmixed-ability teaching 151mixed implicit and explicit teaching
67, 68, 73, 75, 77, 7881Montessori, M. 165morpheme-teaching programme see
teaching packagemorphemes: cause and effect in
connections between knowledgeof morphemes and learning tospell morphemes 2432; impactof teaching about morphemes onliteracy 158; importance ineducation 924; knowledge of see
knowledge of morphemes;morphemic structure and spelling157; nature of 39; representedin writing but not in speech 224; same sounds spelled indifferent ways in differentmorphemes 1621; spelled insame way but represented bydifferent sounds 212; teachersdefinitions 145; teachingmorphology and improvingspelling 323
morphemes-only intervention:effects on spelling 105, 106,11019, 11920; effects onvocabulary 124, 12832
morphemes-plus-spellingintervention 105, 1069, 11019,11920
morphemic stage in spelling 144morphographs 175morphological awareness see explicit
knowledge of morphemesmorphological knowledge see
knowledge of morphemesmorphological production tasks 14,
245morphology, theories about 1434morphology children (teacher
intervention) 1469motion 135motivation, importance of
maintaining 924Mudd, N. 172multisensory methods 165
National Curriculum 134National Literacy Strategy (NLS)
401, 134, 1489, 154negatives, prefixes that make 98-ness suffix 3640, 467nonsense words see pseudowordsnouns 6, 878; abstract nouns 67,
98, 107, 108; person nouns 93,98, 1079
novel words, analyzing 90number 90, 91, 99Nunes, T. 1011, 1214, 20, 25, 33,
87, 144, 160, 169, 173, 182
offline tasks 4750online tasks 4750onset and rime 168oral language see speechorthographic stage (recognizing
chunks in words) 144outcome measures 178, 181overgeneralization errors 279, 45,
pairs of children laboratory study seelaboratory study
past-tense endings 15, 212;changes in spelling over time259; see also -ed ending
patterns 467, 13940, 145, 1724people, -ian morpheme and 67person nouns 93, 98, 1079Peters, M. 163, 164, 165phonemegrapheme
correspondences 169phonetic conception of writing 166phonetic stage in spelling 144phonic strategies 145phonological awareness 151phonological explanations 58phonological principles 312, 33, 87phonological spelling 256, 11315,
135phonological units 1689phonology 174; explanations for
spelling of pseudowords 61;problems with relating to 135
Piagetian theory 166plurals 6, 15; silent endings in
French 234; use of s as amorpheme marking 5961
Portuguese 20, 181possessive apostrophe 22posttests 1856; effects of classroom
intervention on spelling 104,10910, 11016, 11719; effectsof classroom intervention onvocabulary 1248, 12832;laboratory study 68, 6971, 778,789, 801; Lauriston PrimarySchool study 834, 85; teacherintervention 1489; teachingpackage 94, 956, 99101
prefixes 4, 56, 139; effect ofclassroom intervention onspelling 109; teaching package86101, 1023; teaching lists of177
pretests 1856; effects of classroomintervention on spelling 104,10910, 11016, 11719; effectsof classroom intervention onvocabulary 1248, 12832;laboratory study 68, 6971, 78,79, 80; Lauriston Primary Schoolstudy 83, 84, 85; teacherintervention 1489; teachingpackage 94, 956, 99101
priming 54principles 1667, 16878;
higher-order spelling principlesbased on meaning 1748; higher-order spelling principles inteaching spelling 1724; -ionand -ian endings 1920, 656,67; learning of vs learningspecific facts 90; phonological312, 33, 87
problem words 1720, 1634productive morphology task 1213,
29pronunciation 176proofreading 145pseudoword-definition test 126, 127,
1278, 1302, 1323pseudoword-spelling tests: impact of
classroom intervention onspelling 10910, 113, 11516,119; teacher intervention 1468;teaching package 95, 96, 1001,102
pseudowords 89; engagingchildrens imagination 92, 93;higher-order spelling principles1701; knowledge of morphemicprinciples 1415; laboratorystudy 701, 78, 801; LauristonPrimary School study 85;productive morphology task1213, 29; sentence completiontask with 1415, 479; spellingdifferent words with the same
stem 514; spelling suffixes 379,402; subtracting morphemesfrom 92, 93, 180; use of s tomark plural or third-personsingular of verbs 5961; whatspellings of pseudowords tellabout knowledge of morphemes5561; with -ion or -ian 568,701, 78, 801, 85
Qualifications and CurriculumAuthority 134
Rapaport, S. 312Ravid, D. 312Read, C. 166reading: explicit knowledge of
morphemes as essentialingredient of learning to read1624; impact of learning to readon morphological knowledge2432; stages of development144
reading comprehension 77, 121real words, spelling see word spellingreasoning: importance of engaging
childrens reasoning in differentways 902
regular verbs 456, 489, 55research: contribution to educational
theory and practice 17881;transformation into educationalpractice 15962
research strategies 18390;classroom intervention studies1878, 190; laboratoryintervention studies 1847, 190;longitudinal method 1834,18990; studying teachers 1889,190
rimes 168, 169Robinson, J.W. 175roots 4, 5; see also stemsrules 1512; see also principles
s, as morpheme marking plural andthird-person singular of verbs5961
Schonell, F. 164
schwa vowels 1720, 50segmentation of words 11213semi-phonetic stage in spelling 144sentence analogy task 1011, 13, 29sentence completion tasks 1415,
478, 90, 96, 97; laboratorystudy 6971; vocabulary test1246, 127
sentence frames 88, 89Sherrel, P. 56silent plurals 234sounds: awareness of suffix spellings
that do not correspond to endingsounds 3740; morphemesspelled the same way butrepresented by different soundsin different words 212; samesounds spelled in different waysin different morphemes 1621
Speaking and Listening Skills project1223
special educational needs 151specific facts, learning of 90speech: knowledge of morphemes
and 4750; morphemesrepresented in writing but not inspeech 224
spelling 3562, 10420, 161, 181;cause and effect in connectionsbetween morphologicalknowledge and learning to spellmorphemes 2432; childrensunderstanding of English spelling1667; different words with thesame stem 505; higher-orderprinciples in learning spelling1724; impact of classroomintervention on high and lowachievers 11719; interventionsessions 1069; morphemesspelled in the same way butrepresented by different soundsin different words 212; need tobe taught 1625; overview ofteaching package 1046; pretestand posttest assessments 10910;of pseudowords and knowledgeof morphemes 5561;pseudowords with -ion or -ian
568; results of classroomintervention 11019; samesounds spelled in different waysin different morphemes 1621;stages of development 144; s asmorpheme marking plural andthird-person singular of verbs5961; suffixes 3650; teachersstrategies for teaching 13542,145; teaching morphology andimproving 323; theories about1434
spelling banks 145spelling demons 1720, 1634spelling investigations 145spelling journals 146spelling lists 1645spelling patterns 467, 13940, 145,
1724spelling principles see principlesspelling rules 145spelling units 1689standard word lists 164Stanford Achievement Test 175stems 4, 5; conserving 505;
identifying 93, 98; learningpseudowords 9; meaningprinciple for spelling 1756;pronunciation change when aderivational morpheme is added107; segmentation of words11213; spelling different wordswith the same stem 505
suffixes 4, 56, 10, 878, 139;awareness of suffix spellings thatdo not correspond to endingsounds 3740; effect of classroomintervention on spelling 10910,11019; higher-order spellingprinciples based on meaning 176,177; knowledge of when andwhen not to use 406; spelling3650; spelling compared withstems spelling 54, 55; teachingpackage 86101, 1023; thatchange meaning 98
syllables, as spelling units 168Syntactic Bootstrapping Hypothesis
teacher intervention 14252, 1534,162, 189; case studies 1502;changes in childrens spelling1479; impact on children1469; impact on teachers1446; sample of children 146;spelling assessment 1467;teachers comments on theintervention process 14950
teachers 13454, 181; knowledge ofmorphemes 13542, 145, 1534,189; research strategy 1889,190; strategies for teachingspelling 13542, 145; see alsoteacher intervention
teaching activities 17980; CD-ROM104, 106, 123, 128, 161;laboratory study 717; LauristonPrimary School study 82;teaching package 969, 1069
Teaching and Learning ResearchProgramme (TLRP) 183
teaching methods, comparingeffectiveness of 6681, 1867,190
teaching package 86101, 1023,1069, 1601, 1878; effects onspelling 10420, 161; effects onvocabulary 12133, 1612;overview of teaching programme969; pretests and posttests 94,956, 99101; principles used inteaching materials 8794; resultsof study 99101; study design949
Teberosky, A. 166technology, teachers concerns and
150Templeton, S. 171Thenevin, M.-G. 23third-person singular of verbs 5961thought operations 90, 92time 2931, 152, 153, 154Totereau, C. 23Treiman, R. 144, 166, 168
un- 6, 7, 1778
validity of tests 126
variety of learning tasks 180Venezky, R. 169verbs 88, 92; irregular 456; regular
456, 489, 55; spelling suffixes447, 4850; transformingpresent-tense into past-tense1011; use of s as a morphememarking third-person singular5961; see also past-tenseendings
visual learning 165, 1723vocabulary 9, 10, 103, 12133,
1612, 181; context of study1223; effects of morphemeintervention on 12832; learning89; morpheme teachingintervention 128; pretests andposttests 1248, 12832; studydesign 1238
vocabulary test 1246, 127, 12830,1323
vowels, schwa 1720, 50, 65
waiting-list model 94, 102whole-word stage 144whole words, learning 145, 164, 165
Wodehouse, P.G. 7, 139word analogy task 1314, 29word classes 978, 99word lists 1645word-specific learning 70word spelling: impact of classroom
intervention on spelling 109,11015, 11718; laboratorystudy 6970, 7880; LauristonPrimary School study 834;teacher intervention 1469;teaching package 956, 97,99100, 102
writing: impact of learning to writeon morphological knowledge2432; morphemes representedin writing but not in speech 224; morphological knowledge and learning to write1624
-x, pseudowords ending in 5961
-y ending 99
Zutell, J. 173
Annual subscription packages
We now offer special low-cost bulk subscriptions topackages of eBooks in certain subject areas. These areavailable to libraries or to individuals.
For more information please email@example.com
Were continually developing the eBook concept, sokeep up to date by visiting the website.
eBooks at www.eBookstore.tandf.co.uk
A library at your fingertips!
eBooks are electronic versions of printed books. You canstore them on your PC/laptop or browse them online.
They have advantages for anyone needing rapid accessto a wide variety of published, copyright information.
eBooks can help your research by enabling you tobookmark chapters, annotate text and use instant searchesto find specific words or phrases. Several eBook files wouldfit on even a small laptop or PDA.
NEW: Save money by eSubscribing: cheap, online accessto any eBook for as long as you need it.
Book CoverHalf-TitleSeries TitleTitleCopyrightDedicationContentsIllustrationsSeries editor's prefaceAcknowledgementsPart I What is the issue?Chapter 1 Morphemes and literacyChapter 2 What knowledge of morphemes do children and adults show in the way that they spell words?
Part II What does the research tell us?Chapter 3 From the laboratory to the classroomChapter 4 An intervention program for teaching children about morphemes in the classroomChapter 5 An intervention program for classroom teaching about morphemesChapter 6 Can we increase teachers' awareness of morphology and have an impact on their pupils' spelling?
Part III What are the overall implications?Chapter 7 Morphemes and literacy
Appendix The four research strategies in this research programReferencesIndex