Improving Literacy by Teaching Morphemes (Improving Learning)

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

Author: T. Nunes

286 downloads 847 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Improving Literacy by Teaching Morphemes

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

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

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

Improving Literacy by Teaching Morphemes

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

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

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

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

Contents

List of illustrations Series editor’s preface Acknowledgements

ix xiii xv

Part I

What is the issue?

1

1

3

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

2

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

35

TEREZINHA NUNES, PETER BRYANT, URSULA PRETZLIK, DEBORAH EVANS, DANIEL BELL, AND JENNY OLSSON

PART II

What does the research tell us?

63

3

65

From the laboratory to the classroom PETER 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 FREYJA BIRGISDOTTIR, TEREZINHA NUNES, URSULA PRETZLIK, DIANA BURMAN, SELLY GARDNER, AND DANIEL BELL

104

viii Contents

5

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

121

TEREZINHA 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?

134

JANE HURRY, TAMSIN CURNO, MARY PARKER, AND URSULA PRETZLIK

PART III

What are the overall implications? 7

Morphemes and literacy: Context and conclusions

155

157

TEREZINHA NUNES AND PETER BRYANT

Appendix The four research strategies in this research program

183

PETER BRYANT AND TEREZINHA NUNES

References Index

191 195

Illustrations

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

26 28 39

41

45 49 52 53

55

57 60 68

x Illustrations

3.2

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

79 80 84 85

89 91 92

93

100

101

111

112

116

118 125

129

Illustrations xi

5.3

5.4 5.5

6.1 6.2

7.1

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

130 131

132 148

149 171

Tables 2.1 2.2 3.1 4.1 5.1

6.1 6.2 6.3

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

39 42 68 106

124 146 147

147

Boxes 1.1 1.2

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

5 5

xii Illustrations

1.3 1.4 2.1 3.1 3.2 3.3 3.4 3.5 4.1

4.2 4.3 5.1 6.1 6.2 6.3 6.4 6.5 6.6

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

11 17 38 69 71 74 95 97

108 113 114 127 136 137 138 139 141 143

Series editor’s preface

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

Acknowledgements

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

xvi Acknowledgements

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

Acknowledgements xvii

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

Part I

What is the issue?

Chapter 1

Morphemes and literacy A starting point

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

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

4 What is the issue?

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

Morphemes and literacy: A starting point 5

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

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

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

6 What is the issue?

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

Morphemes and literacy: A starting point 7

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

8 What is the issue?

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

Morphemes and literacy: A starting point 9

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

•

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

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

10 What is the issue?

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

• •

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

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

Morphemes and literacy: A starting point 11

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

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

12 What is the issue?

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

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

Morphemes and literacy: A starting point 13

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

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

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

14 What is the issue?

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

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

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

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

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

Morphemes and literacy: A starting point 15

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

16 What is the issue?

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

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

3.

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

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

Morphemes and literacy: A starting point 17

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

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

18 What is the issue?

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

Morphemes and literacy: A starting point 19

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

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

20 What is the issue?

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

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

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

Morphemes and literacy: A starting point 21

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

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

22 What is the issue?

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

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

Morphemes and literacy: A starting point 23

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

2.

3.

4.

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

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

24 What is the issue?

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

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

Morphemes and literacy: A starting point 25

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

26 What is the issue?

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

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

Morphemes and literacy: A starting point 27

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

28 What is the issue?

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

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

Morphemes and literacy: A starting point 29

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

30 What is the issue?

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

Morphemes and literacy: A starting point 31

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

32 What is the issue?

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

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

Morphemes and literacy: A starting point 33

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

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

continued

34 What is the issue?

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

Chapter 2

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

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

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

36 What is the issue?

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

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

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

Knowledge shown in way children and adults spell 37

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

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

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

38 What is the issue?

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

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

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

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

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

Year 3 Year 4 Year 5 Year 6

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

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

100 80 % correct

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

"-ion"

60 "-ness"

40 20

"-ed" in real verbs

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

"-ed" in pseudoverbs

Age in years:months

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

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

40 What is the issue?

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

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

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

Knowledge shown in way children and adults spell 41

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

100 "-ion" in words 80 % correct

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

60

"-ion" in pseudowords

40

"-ian" in words

20

"-ian" in pseudowords

0 8:8 years

9:9 years

Age in years:months

Figure 2.2 Percentage of children who spelled each suffix (“-ion” and “-ian”) in words and pseudowords correctly, by age level. Note: The ANOVA (Analysis of Variance) showed no significant difference between the two age levels, a significant difference between the words and pseudowords (p<0.001) and a significant difference between spelling the “-ion” and the “-ian” suffixes (p<0.001).

42 What is the issue?

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

this as an ending where it is appropriate but also where it is not. They could, for example, be using the “-ion” ending both at the end of abstract nouns and for the agentive ending, “-ian”, because these two endings sound exactly the same. Thus it was necessary to investigate the types of errors that the children made in writing the endings of these words in order to understand better what it was that children were learning about the “-ion” ending. Is it possible that children simply learn that “-ion” is a possible spelling for certain end sounds but have no clue where it is the appropriate spelling? Are they just as likely to use “-ion” for words that end in “-ian” as for those that genuinely should be spelled with “-ion”? We then calculated the proportion of use of “-ion” and “-ian” spellings for each of the types of word and pseudoword. These results are presented in Table 2.2. The sections of the table that have a gray background are those where the wrong sufﬁx was used; that is, the child used “-ion” where it should have been “-ian” or vice versa. The comparisons of greater interest in this table are between the right and the wrong use of the two spellings. The children are roughly three times more likely to use the “-ion” spelling with the right words (abstract nouns) than with the wrong ones (words about people). This is true both for words and for pseudowords. So we can conclude that the children are learning not only that “-ion” is a possible ending but also something about its morphemic value. However, they do use the “-ion” ending in about one-quarter of the words that should be spelled with Table 2.2 Proportion of use of “-ion” and “-ian” spellings for each of the types of word and pseudoword Words

Pseudowords

Year group in school

ending in “-ion”

ending in “-ian”

ending in “-ion”

ending in “-ian”

Year 4 Year 5

77 77

Spelled with “ion” (%) 25 65 23 60

16 14

Year 4 Year 5

0 7

Spelled with “ian” (%) 34 3 44 7

26 30

Note: Two ANOVAs showed that the proportion of “-ion” and “-ian” spellings in the correct place was significantly greater than the use of these spellings in the wrong words (p<0.001 for the use of “-ion”; p=0.03 for the use of “-ian”).

Knowledge shown in way children and adults spell 43

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

“-ian”, and this shows that there is much room for learning more about the use of “-ion” and “-ian”. Children seem much less aware of the “-ian” ending. Overall, they use it far less than they do the “-ion” ending. However, when they do use it, they also use it more often with the right words (words about people) than with the wrong words (abstract nouns). In fact, they hardly ever use “-ian” in the wrong place. Children in Year 5 were about three times as 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 things about sufﬁxes. They need to learn that a particular ending is a possible spelling, and they also need to learn when to use it. Year 4 and 5 children know 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 when to use “-ion” than when to use “-ian”. Finally, we want to make the case that the phenomenon that we are discussing is a general one. The study that we have just described involved a relatively good sample size (N=176), but all the children were in the Oxford area. We also have data from a very large sample of children (N=7,377) who participated in the longitudinal study of the children born in the county of Avon in 1990–1. The children were tested when they were in the age range 9–10. Because of the large sample size, it was only possible to include two words ending in “-ion” (“emotion” and “election”) and two ending in “ian” (“magician” and “electrician”). The level of correct spellings for the words with “-ion” was 74 percent and for the words with “-ian” was 19 percent. More than half of the pupils (66 percent) spelled both “-ion” words correctly, but only 11 percent spelled both “-ian” words correctly. Thus, the results for this large sample conﬁrm that children are more aware of and readier to use the “-ion” ending than the “-ian” one. The error analysis in this study showed that the children were more likely 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, respectively. Thus, the children were aware of the “-ion” ending and used it more often in the right than in the wrong place. However, the rate of overuse 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 that that 3,938 children in the sample misspelled the ending of “electrician”, and 2,216 misspelled the ending of “magician”, by using “-ion” instead

44 What is the issue?

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

of “-ian”. This should give teachers and schools cause for concern. In fact, more children used the “-ion” than the “-ian” ending for both of these words. The error analysis in the large-scale study showed, like that in our previous studies, that the children used the “-ian” ending more often in the right than in the wrong words. The word “magician” was spelled with the “-ian” ending by 25 percent of the children; for the word “electrician”, 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 than in the wrong words. The words “emotion” and “election” were spelled with the “-ian” ending by less than 1 percent of the children (0.33 percent and 0.14 percent, respectively). “Electrician” proved to be very difﬁcult for the children. The spelling most often used was incorrect: 1,812 children spelled “electricion” and “electrition”, and only 785 spelled it correctly. The children were slightly more successful with the word “magician”: The correct spelling was the one used most often— it was used by 1,659 children. However, “magition” was used by 780 children, 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 end sounds. The results of this large survey support our initial idea: Children need to become aware of the possibility of using a letter string at the end of the words, but this does not ensure that they will use it in the right but not in the wrong type of words. The second type of progress, learning when to use and when not to use a letter string, seems to be attained later, when they understand the role of sufﬁxes in spelling. This learning can be described in more general terms by saying that they have to learn that some forms are possible—that is, they have to learn that “-ion” and “-ian” endings are possible endings—but they also have to learn their function. Children who make the wrong analogy between the spelling of “electrician” and “emotion” and use the “-ion” letter string, spelling “electrition” for “electrician”, have learned a form but not its function. They seem to be using this ending as a possible phonological representation, 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 in Chapter 1 and consider how well the children learned about the “-ed”

Knowledge shown in way children and adults spell 45

past-tense ending. Children get better at using the “-ed” at the end of regular past verbs, but do they know when to use this spelling, rather than 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 four in London to spell a number of words, which included ten regular verbs in the past (for example, “kissed”, “killed”, “opened”), ten irregular verbs in the past (for example, “lost”, “sent”, “kept”, “slept”), and ten words that were not verbs and ended in /t/ or /d/ sounds (for example, “soft”, “except”, “field”, “ground”). The study took place before the introduction of the NLS and so there was no explicit guidance at the time for teachers to practice the use of the sufﬁx “-ed” with their pupils. The method for the spelling test was the same described earlier on: The words were presented in a sentence to ensure that the children knew which word we wanted them to spell; they were said in the context of the sentence, and then repeated so that the children could write them. The children were in ﬁve age groups, shown in Figure 2.3, which also summarizes their performance. On the left, the ﬁgure shows the mean correct spelling of the endings —with “t”, “d” or “-ed”—obtained by the children. At all age levels, the children were considerably better at using the endings “t” and “d”, which are the right way of spelling the irregular past verbs and the nonverbs, (a) Correct endings (out of 10)

(b) Proportion of errors using “-ed” in the wrong place 1

10

irregular verbs

rs 10 yr s

9y

rs 8y

9y

8y

7y

nonverbs

rs

0 7y

0

6y

0.2 10 yr s

2 rs

0.4

rs

4

rs

0.6

rs

6

rs

0.8

8

6y

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

to nonverbs to irregular verbs

regular verbs

Figure 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.

46 What is the issue?

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

than at using the “-ed” ending, which is the correct spelling for the regular past verbs. This indicates that they were better able to use their knowledge of sounds to guide their spelling of the ends of words than to use their knowledge of morphemes. Although this is a different sample of children from those in the previous studies, the similarity in the rates of success for the age groups that are comparable across studies is remarkable: At the age of about 10, the children show a high level of success, but there is considerable room for further learning. The graph on the right of Figure 2.3 shows how many errors made with 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 of spelling the endings with the letters “t” or “d”, which would have been correct, the children spelled these endings with “ed”: For example, they wrote “sofed” for “soft” and “helded” for “held”. Similarly to what was observed with “-ion”, the children used the spelling “-ed” correctly but also used it in words where it was not appropriate: They generalized it to irregular past verbs and even to other words which were not verbs. They had learned that the form “-ed” is a possible ending but seemed to attribute to it a phonological function: Words that end in /t/ or /d/ sounds could be spelled with “-ed”. They had not grasped its morphological function.

Conclusion By now, we can attempt to answer the questions posed at the start of this section. Is it easy to spell sufﬁxes because they have a ﬁxed form? The answer seems to be that it is harder than one would expect. If a sufﬁx cannot be spelled on the basis of its sounds—for example, because it contains a schwa vowel, which is not pronounced clearly—then it takes children some time to spell the sufﬁx correctly. On average, children in their fourth and ﬁfth year in school, who are aged 9 and 10, still do not show close 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 that sound like other words which are spelled differently, the children need to become aware of the sufﬁxes as possible endings and also to learn when to use these spellings at the end of words. Many children use these sufﬁxes in the right as well as the wrong words. This generalization to the wrong type of words suggests that perhaps they are learning

Knowledge shown in way children and adults spell 47

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

something about the existence of spelling patterns—that is, sequences of letters that appear together at the end of words—without necessarily grasping that these sequences of letters have meaning and are used to convey these meanings in the written language. It is difficult to end this section without wondering why it takes children so many years in school to learn to spell sufﬁxes, considering that the spelling of sufﬁxes is so predictable. Is it possible that children are not such good morphemists after all and that the studies about how good they are in oral language overestimate their ability? Or could there be two different levels of knowledge of morphemes, one used in oral and the other in written language?

Different types of linguistic knowledge In the ﬁrst chapter we discussed the possibility that knowing how to use an aspect of a language—say, the different types of past tense—is different from explaining how the same aspect of language is used. The type of knowledge of language that is used for speaking is implicit. When we are doing something, like speaking, we don’t really think about how language works. Tasks that we perform without having to explain why we do things the way that we do are referred to as “online tasks”.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 letters represent the spoken language. In its relation to oral language, spelling can be seen as an “ofﬂine” task. This is one reason why children may be better 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 morphemists, and may be able to inﬂect words in oral language, but still might not know enough about the morphemic elements that are not pronounced—and the “-ed”, as we have seen, is not pronounced; it is only written. In the previous chapter, we described a technique developed by Jean Berko that is considered to be a good assessment of children’s knowledge of morphemes in online oral tasks. The children are asked

1

Annette Karmiloff-Smith (1992) discusses the distinction between online and ofﬂine tasks 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.

48 What is the issue?

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

to complete sentences containing pseudowords. In order for the sentences to be completed correctly, the children must apply a morphemic transformation to the pseudoword; sometimes they are required to inﬂect verbs and nouns (for example, say the verbs in the past or the nouns in the plural) and sometimes they are required to use a derivational suffix (for example, make an agentive out of a verb). Because the transformations have to be applied to pseudowords, the children cannot use specific memories of the items; they have to generate the new item themselves. Thus, this is an online task, but the knowledge it draws on is likely to be a more explicit form of knowledge of morphemes than the use of inﬂections in familiar words. In some sense, this is more a language game than speaking. But it is not an unrealistic game: Children will often encounter unfamiliar words in their 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 to compare performance on these two tasks. Spelling is an ofﬂine task in its relation to morphemes: It requires that we represent the morphemes with letters. Inﬂecting spoken pseudowords is an online task: We can use the morphemes without having to think about them. If this is correct, then children should show better performance in the pseudoword inﬂection tasks than in spelling. In another of our studies we gave 6-, 7-, and 8-year-old children oral pseudoword tasks along with the spelling tasks described earlier. In some of the items, they were asked to complete sentences that contained a pseudoverb; the pseudoverb needed to be in the past for the sentence to be completed correctly. In these oral tasks we showed the children a picture, gave them clues from spoken sentences, and asked them to use these clues in order to complete the sentence. They had to come up with the past for the pseudoverb. This task was described in Box 1.3, with an example of an item that required the children to inﬂect the pseudoverb “to snig”. Our task contained a variety of pseudowords, three of which were verbs. We calculated the proportion of correct answers to the pseudoverb items in the oral task and the proportion of correct spellings of the endings of real past regular verbs in the spelling task. Figure 2.4 shows the comparison between their ability to produce the inﬂection in the oral pseudoword task and their ability to use the “-ed” when spelling the end of regular past verbs. The results are quite clear: Children are better morphemists in oral than in written language. The same children

Knowledge shown in way children and adults spell 49

1 Proportion correct

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

0.8 0.6 0.4 0.2 0 6yrs

7yrs

8yrs

Age level

pseudoverbs correctly inflected in oral language regular verbs correctly spelled

Figure 2.4 Proportion of past regular verb endings spelled correctly and produced correctly for pseudowords in an oral task. Note: An ANOVA with repeated measures produced significant main effects for age group and task and a significant age by task type interaction (all differences significant at the 0.001 level).

who found it so difﬁcult to spell the ends of verbs using “-ed” had a signiﬁcantly higher level of success in producing the pseudoverbs in the past correctly. So why could they not spell with the “-ed” sufﬁx correctly? We think that the answer is in the discrepancy between oral and written language: Some of the information about morphemes is either not present or not clear in oral language. In Chapter 1, we argued that there are many cases where a correct spelling cannot be produced on the basis of sounds: Knowledge of morphemes is what allows us to master these spellings. This is exactly the case with the morphemes considered here: Information from oral language is insufﬁcient for correct spelling. We think that in order to master these spellings the children have to develop some awareness of the morphemes that compose these words and try to represent the morphemes and not just the sounds. In the case of the past tense, oral language does not represent the “-ed”: Past regular verbs sound like /t/ or /d/ or /id/ at the end. The children need information about written language in order to realize what the morpheme is. It is not sufﬁcient to know that some words in English sound like /t/ or /d/ at the end but are written with “-ed”. This increases the likelihood that the children will

50 What is the issue?

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

use the “-ed” ending in spelling, but it does not guarantee that they will do 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 the effect of teaching. We think that becoming aware of morphemes should make it easier for children to learn to use these sufﬁxes in the right, and not in the wrong, words. We also think that teaching can be an effective way of making children more aware of this aspect of language. This is why in later studies we looked at different ways in which children can be taught about morphemes. These studies are presented in Chapter 3. Before we move to the analysis of teaching in Chapter 3, though, we still want to look through the spelling window to ﬁnd out more about what children and adults know about morphemes.

Spelling different words with the same stem: Do children conserve the stem? In the ﬁrst section of this chapter our attention was focused on how children spell suffixes to gain insight into what they know about morphemes. We now turn our attention to word stems, which form the core of the word. Words cannot be made only with a sufﬁx: A sufﬁx is always attached to a stem. So, although sufﬁxes are important, they cannot tell the whole story about our awareness of morphemes. It is possible that children ﬁnd it easier to realize that words with the same stem (for example, “art” and “artist”) bear some relation to each other than to realize that words ending in “-ion” are abstract nouns and those ending in “-ian” are agentives. The best approach to spelling a stem is arguably to analyze its sounds and to try to represent them. However, this is not the only thing we should think about. Sometimes a stem is clearly pronounced when it is a word on its own, but it is not as clearly pronounced once a sufﬁx is added to it. Think of how we say the words “magic” and “magician”. The word “magic” has two clearly pronounced vowels that are easy to represent in spelling, and the last letter, “c”, represents the sound that it always represents at the end of words. When we add the sufﬁx “-ian” to it, the stress pattern changes and the letter “a” now represents a schwa vowel. (See Box 1.4 for an explanation about schwa vowels.) The letter “c” now stands for a sound most often spelled with “sh”. This means that children who might spell the ﬁrst vowel in the word “magic” correctly could well make mistakes in the same place when spelling the word “magician”—and they often do.

Knowledge shown in way children and adults spell 51

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

The survey of children’s spelling in the county of Avon for Year 6 children showed that the majority of the children (75 percent) were able to spell the beginning of “magician” correctly, with “ma-”. But onequarter of the children did not spell this beginning correctly. The number of children is such that it should not be neglected: 565 children spelled the beginning of “magician” with “mu-”, 736 spelled it with “mi-”, and a small number (45) actually used no vowel between the initial letter “m” and a subsequent consonant. So even 10-year-olds might not realize that they should conserve the stem when a sufﬁx is added to it. However, for those who realize that “magician” is composed with the stem “magic-” and the suffix “-ian”, this spelling becomes completely predictable, although the vowel is not clearly pronounced. This sort of analysis led us to investigate how consistently children spell words that are different but have the same stem. Do they, on the whole, conserve the stem across words? If they do, does this consistency in spelling the same stem across words reveal anything about their knowledge of morphemes? The children in one of our studies spelled ten pairs of words with the same stem and ten word–pseudoword pairs that also had the same stem. All the words included in these pairs had stems with something that was unpredictable either in both words or in one of them. Examples of the pairs of words we used are “magic”–“magician”, “know”– “knowledge”, “naughty”–“naughtiness” and “treasure”–“treasures”. The difﬁculties in the word “magician” have already been pointed out. In the words “know” and “knowledge”, there is a silent “k” and the stem sounds differently across the words. The words “naughty” and “treasure” also contain difﬁculties: “Naughty” has an unusual spelling for the ﬁrst vowel sound and the letter “s” in “treasure” does not represent its most common pronunciation. However, these spellings should be maintained in the derived forms. Previous research (Fowler and Liberman 1995) has shown that children find it difficult to detect the common stem when it sounds differently across the base and the derived form, so we thought it would be important to include in our task some pairs where the stem sounds the same and other pairs where the stem sounds differently. We thought it was also necessary to include pseudowords in this task because children’s consistency in spelling stems across words could result from rote learning of each word. However, if the children spelled a word and a pseudoword that contained the same stem consistently, we would know that they could not be spelling the pseudoword on the basis of a specific memory. The pseudowords that we created were

52 What is the issue?

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

(a) Knotosaurus

(b) Combosaurus

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

dinosaur names. The children were shown a picture of a dinosaur that made it very clear why there should be a connection between its name and the stem we wanted them to think about. Figure 2.5 illustrates the items used in this task. In order to make sure that the children did not copy the stem for the second word after they had already written the ﬁrst word, we dictated one element from the pair on one day and the second element on a different day. So, if they spelled, say, the stem of “magician” in the same way as they spelled “magic”, we would consider it likely that they had thought of the word “magic” on their own rather than were prompted by seeing it on the page. Half of the children wrote half of the basic words and half of the derived words on the ﬁrst day and the other half on the second day. The remaining children had the dictation in the opposite order. For each of the word pairs, the children were given one point if the stems were spelled exactly in the same way, independently of whether they had been spelled correctly or not. We realize that this could overestimate the children’s awareness of the connection between the stems. For example, if a child spelled “naughty” as “noty” and also used “not” for the stem of “naughtiness”, the consistency might result not from the child’s realization that the stems are the same but from the use of a phonological strategy to spell both words. However, we felt that we

Knowledge shown in way children and adults spell 53

could not exclude from the consistency score those words that had been spelled 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 London and Oxford schools. Graph (b) displays the results for the same Oxford children approximately one year later; 169 children completed the tasks on this second occasion. Three results stand out in these graphs. First, as in the previous studies, the children’s performance improves with age. This can be seen

% with same spelling

(a) 100 90 80 70 60 50 40 30 20 10 0 6:6 years

7:7 years

8:10 years

Age in years:months consistency in words

consistency in pseudowords

(b) % with same spelling

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

100 90 80 70 60 50 40 30 20 10 0 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 were spelled in the same way at each age level.

54 What is the issue?

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

on Graph (a) where the results are for different children of different ages, and also from a comparison between the two graphs, when some of the same children were tested at a different age. Second, the performance of the 9-year-olds, like in the previous studies, can be considered good but not excellent. The majority of the spellings conserve the stem across words but there is still much room for improvement. The third aspect of these results was unexpected: The children at all age levels performed signiﬁcantly better on the consistency of stems when one stimulus was a pseudoword than when they were both words. The difference in performance between the tasks could be due to the fact that the dinosaur pictures gave such a strong clue to the stem that the children could not avoid making a connection between the derived pseudoword and the basic word. In psychological experiments, we refer to using stimuli to provoke connections, even if implicitly, as “priming”. In spite of the difference in level of performance, the two tasks do not seem to assess different aspects of the children’s knowledge of morphemes: The correlation between them was 0.83, which is a high correlation (a perfect correlation has the value of 1). This high correlation would be unexpected if the processes assessed by the dinosaur task were different from those assessed by the task with real words. At ﬁrst glance, the results do not seem to indicate that children ﬁnd it 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 direct comparison is only possible if we look at the results for a sample of children from the different studies who were of comparable age levels when they did the spellings of sufﬁxes and the consistency in spelling stems 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 two tasks. We think that these results leave little room for doubt that children’s 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 be easier 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 a sufﬁx and that other sufﬁxes may be easier or more difﬁcult: It may be easier to understand the connection between past verbs, for example, than the connection between abstract nouns. So far, we have looked at children’s knowledge of sufﬁxes and stems as revealed by their spelling and their ability to inﬂect pseudowords. We

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

% correct

Knowledge shown in way children and adults spell 55 100 90 80 70 60 50 40 30 20 10 0 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” and proportion of stems spelled consistently across two words.

have not asked them about why they spell words and pseudowords the way that they do. Asking people to explain why they do something the way that they do, we argued, assesses a different process: It requires a greater level of awareness. But we have not provided any evidence for this idea. In the section that follows, we will consider differences between spelling and explaining why pseudowords should be spelled in particular ways.

What children’s and adults’ spellings of pseudowords tell us about their knowledge of morphemes We have, in previous work, asked children and adults to spell some words and to explain why they spelled them the way that they did. It is often difﬁcult to obtain an explanation that goes further than “Because that’s the way that this word is spelled” or “That’s how I’ve seen the word spelled in the past.” We have found that it is much easier to obtain explanations when we ask people to explain why they spelled a pseudoword in a particular way. When people spell pseudowords, they realize that they cannot say “This is the right way” or “I’ve seen it spelled like this in the past.” They created the spelling, so they try to think why they used that particular spelling. So, if we ask children and adults to spell

56 What is the issue?

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

pseudowords and to justify the way that they spelled them, we can have two ways of assessing their knowledge of morphemes: The ﬁrst by the correctness of their spelling and the second by analyzing what sort of explanations they give for their spellings. We expect that explaining is more difficult than making correct choices, but also that there is a connection between the ability to explain why the pseudowords should be spelled in certain ways and the correctness of the spelling. So we used pseudoword spellings and explanations as another window for looking at children’s awareness of morphemes. Freyja Birgisdottir and her students Victoria Brierley and Peter Sherrel asked twenty undergraduate students (mean age 27 years 7 months), ﬁfty-one children in their fourth year in primary school (mean age 8 years 3 months) and fifty-six children in their sixth year in primary school (mean age 10 years 6 months) to spell thirty-ﬁve pseudowords. Unlike the pseudowords used in the dinosaur task, which had stems from real words, the pseudowords used in this task did not draw on stems 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), by changing the initial syllable (for example, “fompect” was created by changing 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 in a sentence that made the grammatical status of the pseudoword clear, and, consequently, if the pseudoword had a sufﬁx in it, what the sufﬁx was. The whole sentence was presented in writing; a gap designated the place where the pseudoword should be written. The pseudoword was ﬁrst said in the linguistic context and then repeated, in order to help the participants remember what the pseudoword was. All the items were chosen to reveal something about the knowledge of morphemes used by the participants. They belonged to seven different categories of words; ﬁve examples of each type were included. In order to illustrate the difference between spelling and explaining one’s choices for spelling pseudowords, we will consider here two cases, where the endings are pronounced in the same way but the spelling is different. 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.

Knowledge shown in way children and adults spell 57

Two categories of pseudowords were used. As in the previous studies, the sentences were used to clarify what category the pseudoword belonged to. Figure 2.8 summarizes the findings for this study. These results replicate our earlier ﬁndings. At about age 10, children perform well in spelling these pseudowords, but there is much room for improvement. As in the previous studies, the children were more successful in spelling the pseudowords ending in “-ion” correctly than those ending in “-ian”. They were also more likely to use “-ion” in “-ian” pseudowords than the other way around. In the group of children aged 8, 39 percent of the pseudowords that should have been spelled with “-ian” were written with “-ion”, whereas the reverse mistake appeared in only 9 percent of the 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 also used 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 “-ian” words with “-ion” endings increased to 41 percent, and the use of 100 90 80 % correct spellings

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

70 60

Pseudowords with "-ion"

50 40

Pseudowords with "-ian"

30 20

Explanations

10 0 8-year-olds

10-year-olds

adults

Age group

Figure 2.8 Percentage of correct pseudowords with “-ion” and “-ian” spelled correctly and percentage of correct explanations, by age level. Note: An ANOVA showed that there were significant differences in performance across the age levels and word type, with “-ion” and “-ian” endings (p<0.001 in both cases).

58 What is the issue?

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

“-ian” in “-ion” pseudowords also increased, reaching 19 percent. This was due to the fact that the use of other spellings, different from “-ion” and “-ian” endings, decreased. This suggests that the children were more aware of the existence of both sufﬁxes and also more aware of when to use and when not to use these sufﬁxes, because the difference between the right and wrong use of each of the suffixes is larger than that observed for the 8-year-olds. The performance of the adults shows much less difference between the level of success across word type: Their spellings were correct for 78 percent of the “-ion” pseudowords and for 76 percent of the “-ian” pseudowords. So, they seem equally aware of the existence of both sufﬁxes. However, they were still more likely to use “-ion” spellings in “-ian” pseudowords: They spelled 31 percent of the “-ian” pseudowords with “-ion”, and only 19 percent of the “-ion” words with “-ian”. The explanations provided for the spellings were classiﬁed into three categories. If there was a reference to at least part of the rule or an explicit analogy with words of the appropriate category, the explanation was considered correct. This is, arguably, a lenient criterion but very few responses would satisfy a strict criterion. Many explanations referred to the way that the word sounds: These were classified as phonological explanations and were, for all the cases considered here, insufficient. A third category was created with other types of explanations. Figure 2.8 shows that none of the 8- and 10-year-old children provided correct explanations for their choice of spelling in these pseudowords. In contrast, about one-third of the explanations provided by the adults referred to the underlying rule in some way. For example, one participant wrote to justify an “-ion” spelling: “a noun—from rule in 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 the pseudoword “Maridian” with “-ian” (which was used in the sentence “Someone who comes from Marid is a ________) and wrote the following explanation: “People’s names, i.e. where they were from.” This same participant could not provide a correct explanation for “-ion” words and justiﬁed the spelling on the basis of “how it was pronounced.” Thus, a comparison between the level of correct spellings and correct explanations conﬁrms the idea that these tasks assess different types of knowledge: The level of awareness required for explanations is certainly greater than that required for spelling pseudowords.

Knowledge shown in way children and adults spell 59

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

The use of “s” as a morpheme that marks plural and as a morpheme that marks the third person singular of verbs The second morphological distinction included in this study was about pseudowords that end in the sound /ks/. Two types of words were included: One-morpheme and two-morpheme words. If a onemorpheme 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 singular pseudoword used is: “There is only one spix in this recipe.” Two sets of two-morpheme words were used in this study. One set was formed by plural nouns, which are spelled with “-cks” at the end: the “s” is a morpheme that marks the plural, added to a singular noun than ends in the /k/ sound. An example of this type of pseudoword used in the study is, “My Grandma has macks in her kitchen.” The second set of two-morpheme words used in the study was formed by verbs in the third-person singular: The morpheme “s”, added to a stem that ends in the sound /k/, creates a pseudoword ending in the /ks/ sound. An example used in the study is, “Mary smicks the trees every day.” Onemorpheme words that end with the sound /ks/ are spelled with an “x”; two-morpheme words ending in these sounds are spelled with an “-s” added to the basic form (which could be spelled with “-k”, “-ck” or “-ke” at the end). This is a very consistent spelling rule, but we have not seen it taught so far. Previous studies have shown that the distinction between these cases is problematic for children. Peter Bryant and his students have observed in different studies that children use “-x” for plural words and also plural endings for singular words. They do not seem to know when to use one or the other type of spelling. However, the speciﬁc frequencies of errors varied across studies. It is quite possible that this variation depends on whether the children consider “-x” as the appropriate phonological representation for the /ks/ sound or whether they think that any word ending with an /s/ sound should have the letter “-s” at the end. If they think that “-x” is the appropriate representation, they will spell onemorpheme words correctly and make more errors in two-morpheme words. 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 scores in the two types of words: The better you perform on one word type, the worse you perform on the other. Hopefully, if the children become aware of the morphological signiﬁcance of these endings, the negative

60 What is the issue?

correlation should disappear, and their spelling of all three word types would 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 improvement is uneven because the level of performance across the types of pseudowords is signiﬁcantly different for the younger age group. The 8-year-olds performed significantly worse in singular pseudowords ending in “-x” than in the other two word types. Although the 10-yearolds perform better than the 8-year-olds, their performance is only at chance level: There are only two phonologically appropriate spellings for these pseudoword endings, and they are right about 50 percent of the time for the singular and plural nouns. Their performance is slightly better on the verbs in the third-person singular. The adults’ spelling is above chance level but not close to 100 percent. They performed better with the distinction between “-ion” and “-ian”. These results were surprising to us: We had thought that the distinction between singular and plural is such an easy one that its representation in spelling would be mastered at an earlier age.

% correct spellings

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

Pseudowords ending in "-x"

100 90 80 70 60 50 40 30 20 10 0

Pseudowords in the plural Pseudoverbs in third person Explanations, singular vs. plural 8-year-olds

10-year-olds Age group

adults

Explanations, third-person singular

Figure 2.9 Percentage of correct spellings of one-morpheme and twomorpheme words, by age level. Note: An ANOVA showed that there were significant differences in performance across the age levels and word type; the use of the “s” to mark the third-person singular of pseudowords was easier than the spelling of the other pseudowords for the children (p<0.001).

Knowledge shown in way children and adults spell 61

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

When we analyzed the explanations provided for the spellings of the pseudowords, we found trends that were quite similar to what we described before for the “-ion” and “-ian” pseudowords. The children did not refer to grammar or to morphemes, nor did they use analogies to appropriate words justifying their choices of spelling. Approximately one-third of the explanations provided by the adults referred to the fact that the words were singular or plural. Some examples of explanations considered to satisfy the criterion for correct responses are: “Pix seems to be a singular” and “This is more than one.” Not all explanations given by adults using grammar were correct. One participant distinguished between her use of “-x” and “-ks” by saying that she used “-x” for a noun and “-ks” for a verb. This reference to grammar was not considered correct. She then wrote “pix”, a singular noun, with “-x”, and “spucks”, a plural noun, with “-ks”. Next to this she wrote: “Both nouns—not logical!” Her difﬁculty in identifying the different types of morpheme represented by a ﬁnal “-s” appeared to conﬁrm the often-held view that English spelling is not logical. If the participants were not thinking of morphemes to produce their spelling, what were they thinking of? The majority of the explanations referred to how the words sound. The 8-year-olds justiﬁed their spellings by referring to the way that the words sound 43 percent of the time. This type of explanation increased to 70 percent in the 10-year-old group but decreased to 44 percent in the group of adults, who gave 36 percent of explanations using morphological rules or analogies to words that represented these rules. It is possible that the strong tendency to focus on phonology revealed in the explanations interferes with the awareness of rules that are based on a different layer of language organization. The overall performance in pseudoword spelling shown by both children and adults is much better than what one might expect from their explanations. However, it should not be forgotten that in none of the examples was the performance of any of the groups close to 100 percent. If they had had a greater awareness of the significance of morphology for spelling in English, would they have been able to attain 100 percent?

Summary and conclusions In this chapter, we have used spelling as a window to find out more about children’s and adults’ knowledge about morphemes. Our own

continued

62 What is the issue?

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

studies involved more than 1,000 children at different age levels. The study of children in the county of Avon involved over 7,000 children. These different studies on the spelling of morphemes concur in showing that children make progress in the spelling of suffixes and the conservation of stems across words in primary school. However, performance does not reach a level of accuracy that we would expect good spellers to reach after five or six years of schooling. The conclusion that we need to teach children about morphemes seems inescapable. The question is what is the best way of doing this. Chapter 3 summarizes a set of studies that investigated this question.

Part II

What does the research tell us?

Chapter 3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

From the laboratory to the classroom

We turn now to teaching children about morphemes, and we shall focus first on spelling. Our main aim in this chapter will be to show you that it is quite possible to teach children about spelling principles which are based on morphemes and which no one would otherwise tell them about at school. We shall try to convince you that this sort of teaching is relatively easy to do, that it is great fun for children, and that it is highly effective: Children are able to learn morphemic spelling principles quickly, and by and large they remember what they learn long after they have been taught about it.

Study 1: Teaching pairs of children about the difference between “-ion” and “-ian” endings Our ﬁrst example is a project that we ourselves did on teaching children about the distinction between “-ion” and “-ian” endings. This distinction had become a test case for us. We had identiﬁed a simple and useful morphemic principle that had, we thought, the potential to help children conquer an aspect of spelling that normally causes them much trouble. The principle is that when the word is a noun that ends with a schwa vowel 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 with words like “election” and “magician”. The spellings for the “-ion” and “-ian” endings bear little relationship to their sounds, and this is plainly the cause of the young children’s problems with such words. Children Authored by Peter Bryant, Terezinha Nunes, Ursula Pretzlik, Daniel Bell, Deborah Evans, and Jenny Olsson

66 What does the research tell us?

who write words like “elecshon” or “elexon” (about 28 percent of the 9-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” followed by a single vowel followed by “n” and the most often used vowels were “o” and “e”) are defeated by the clash between the letter–sound correspondence rules, which they have learned so well, and by these two different ways of spelling the same sound (“-ion” and “-ian”), both of which ﬂout the letter–sound 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 principle—a genuine spelling principle—for such words. In this case, the regularity is at the level of morphemes, not at the level of sounds or sequences of sounds. We have no doubts, no hesitations at all, about this approach. If there is a spelling principle that could remove a stubborn difﬁculty, then one should teach this principle to children to see if it helps them to spell. The only remaining question is how best to teach it. When we considered this last question, we realized that there were three main candidates. One was explicit teaching. This is the up-front approach. To teach children the principle explicitly, we would have to explain the spelling principle to them clearly and completely, with suitable examples to provide them the relevant experience. The examples would be “-ion” and “-ian” words that we would place into two clear categories, and we would go through the difference between these categories. All the arguments that we have been through in Chapters 1 and 2 pushed us in the direction of explicit methods, and so we decided to teach some of the children in the study in this way. We also wanted to look at the possible effects of what we called implicit teaching. This is to give children examples of words with “-ion” and “-ian” endings but not to explain the principle. So, the aim here would be to give the children the chance to work out the principles themselves, by going through the examples. In some ways, this method is quite close to present practice, since schoolchildren usually are asked to learn the spellings of words like “election” and “magician” and could possibly work out the person/nonperson principle for themselves. But, in other ways, our ideas for implicit teaching are quite different from normal school practice. Our plan was to put the two kinds of words (the “-ion” and the “-ian” words) into separate categories and then to make direct comparisons between these categories, so that the child might see what the words in each category had in common, and what was the

From the laboratory to the classroom 67

difference between the two. Schoolchildren’s experiences with words with these two endings are usually a great deal more haphazard than that. Our third idea for teaching the morphemic principle was to mix implicit and explicit teaching. We decided that a quite reasonable way of teaching the principle would be to start our teaching of some of the children by giving them implicit experiences by categorizing the words but not explaining the difference between the categories explicitly at ﬁrst. Halfway through, however, we would move from implicit to explicit teaching: We would ask the children to explain what the principle was and, if they could not do so, we would explain to them how the two word categories differed. We need to make one more point before we describe the ﬁrst of the teaching studies in which we looked at ways of promoting children’s learning of the principle about “-”-ion” and “-ian” endings. This principle has two parts to it, since it is a principle about two morphemes and two spellings. One morpheme (spelled “-ian”) signals that the word represents 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 suggests that children would find it easier to learn a principle that involved people, since this is an easy category for them, than a principle that is mostly about abstract ideas. If this is so, children who are taught the principle might ﬁnd the “-ian” part of it easier to learn, since it involves the easier category of the two. Although it is possible to examine this, we will not carry out any analysis in this study to investigate this conjecture. Thus we set out to compare three different teaching methods, which meant that we had to compare three groups of children, each of which we taught in different ways. Since this was a tightly controlled experiment, we also included a fourth group of children, a control group, who actually received exactly the same amount of instruction as the children in the other groups, but were taught about something entirely different, strategies that might help them to understand passages of prose.

The children The 200 children who took part in this study were all 9-year-olds and their average age was 91⁄2. They came from two schools in Oxford, and the numbers of children from each of the two schools was the same in all four groups. Table 3.1 gives some detail about the four groups.

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 (years)

Standard deviation

N

Morphology explicit Morphology implicit Morphology mixed Control

9.5 9.7 9.6 9.5

0.60 0.67 0.53 0.71

40 43 42 75

The design The study had a simple traditional design, which we have set out in Figure 3.1. First, we gave the children a pretest to measure how well they could spell the endings that interested us. Then we had two teaching and learning sessions, in which we taught the different groups in different ways. Very soon after the end of the second training session we gave the children a posttest, which we called the “immediate posttest” and which was identical to the pretest and thus measured

Pre-test

Two 20-minute teaching sessions with pairs of children

Immediate posttest

Two month interval

Delayed posttest

Figure 3.1 Design of the first teaching study.

From the laboratory to the classroom 69

the different levels of progress in the four groups as a result of the different kinds of teaching that they were given. Two months later we gave the children a “delayed posttest.” This was the same as the previous tests, and its aim was to test how well children had remembered the spelling principle a considerable time after they had been taught about it.

The pretest and posttest measures of spelling Our pretest provided a measure of how well children spelled the endings that interested us. This consisted of two tasks that were designed to measure the children’s ability to spell words that end in “-ion” and “-ian”. Box 3.1 gives details of both pretest tasks. In one task, we asked the children to write some real words ending in “-ion” and “-ian”. Our technique was to present the children with a set of written sentences with a gap in each of them (see Box 3.1). This gap represented a missing word. Then we read out the whole of each sentence, including each missing word, and we asked the children to write in that word. So, for example, the missing word in the sentence “Kate was the only one without an ________” was “invitation”, and this was the word that we asked the children to write in after we had dictated it.

BOX 3.1 The word- and pseudoword-spelling tasks used in Studies 1 and 2 Words 1. 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.

70 What does the research tell us?

12. 13. 14. 15. 16.

Congratulations to the winner. The electrician fixed the lights. The Brazilian scored again. The librarian checked out my book. The comedian was funny.

Pseudowords 1. 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.

In the other task, we asked the children to write pseudowords. Box 3.1 gives the items used in this task. We introduced pseudowords because these make it possible for us to test whether or not the children do understand a spelling principle. With spellings of real words, there is always a chance that someone might produce the correct spellings because he or she has learned the correct spelling for each individual word by rote—what psychologists call “word-speciﬁc learning”—and not by learning any principle. Pseudowords, in contrast, rule out the possibility of word-specific learning completely: Such words, by deﬁnition, are completely unfamiliar. If children put “-ion” and “-ian” endings on appropriate pseudowords, they must have been using a principle to do so. In this pseudoword test we used the same missing word technique as with real words, and here the context provided by the surrounding sentence was essential. As you can see from the examples in Box 3.1, these surrounding sentences made it clear that in some sentences, the pseudowords referred to a person. For example, “Someone who works in a korpect is a ________”. The pseudoword we dictated was “korpectian”. In other sentences, the pseudoword did not represent a person. For example, “The place where you find foppats is a ________”. The pseudoword we dictated was “foppation”. We wanted to know whether

From the laboratory to the classroom 71

children would be more likely to put an “-ian” ending when the pseudoword meant a person (“korpectian”) than a nonperson (“foppation”). Both tasks contained other words, whose spellings were not based on these 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 the control group, went through two teaching and learning sessions (we taught, they learned). The second session came a few days after the first. Notice that there were two of these sessions only. Since each session lasted for around 20 minutes, the total teaching time for each child was really quite small. We used the same general plan and exactly the same material to teach the explicit, implicit and mixed groups about the “-ion” and “-ian” morphemes. We saw all these children in pairs, and we based our teaching on two different activities, which we called “games”. Boxes 3.2 and 3.3 give details of these games and of how the procedures in them were different for the three groups.

BOX 3.2 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 which the experimenter, using PowerPoint, presented the children with the material needed for the game. In the analogy game we showed the pairs of children several sets of four puppets. In every set, there was a word box immediately underneath each of the four puppets. Here, as examples, are two such sets. When we presented a set, all four puppets in the set appeared on the screen at the same time, but their word boxes were empty. Then the top puppet on the left waved its arms and apparently spoke a one-morpheme word (for example, “magic”). The written form of the 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? (a)

magic

magician

music

?

protect

protection

infect

?

(b)

accompanied by a few arm waves, which was derived from the first word (for example, “magician”), and, of course, the written form of the word appeared below in this puppet’s word box. The third step was for the bottom puppet on the left to produce a new word—again, 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 underneath the fourth puppet, and the experimenter asked the two children to work out what the missing word would be and how it should be spelled, making it clear to the children that the relation between “music” and the missing word had to be the same as the relation between “magic” and “magician”. (The illustrations actually show the display as it was at this point of the game.) Then the two children discussed what the word was and how to write it and, after they had come up with their own agreed solution, the fourth puppet spoke its word (the correct word) and the appropriate written word replaced the question mark in the puppet’s word box. At this point of the game, the procedure diverged for the different groups. In all the trials experienced by the explicit group and in the “explicit” trials experienced by the implicit group (the second half of the sets given to them in this game), the experimenter gave the children an explanation of why the correct word and the other twomorpheme word in the set ended in “-ion” and not in “-ian”, or vice versa. This explanation was not given to the children in the implicit group, nor to the mixed group during the “implicit” trials that we gave to this group (the first half of the sets given to them in this game). Finally, we presented all the children in the explicit, mixed, and implicit groups with a PowerPoint résumé of the whole game in the form 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 as possible by putting all the “-ian” sets on one side of the display and all the “-ion” sets on the other. We discussed the reason for the difference between the two endings in these two collections of words with the children in the explicit and mixed groups, but not with the children in the implicit group. magic

magician

protect

protection

music

musician

infect

infection

history

historian

add

addition

library

librarian

subtract

subtraction

India

Indian

confess

confession

74 What does the research tell us?

Egypt

Egyptian

discuss

discussion

political

politician

suggest

suggestion

technical

technician

collect

collection

Italy

Italian

educate

education

Hungary

Hungarian

imitate

imitation

BOX 3.3 The correction game The idea in the correction game was to encourage the pairs of children to be the judges of someone else’s spelling. We presented them with a set of written sentences, again in PowerPoint. All the words in the sentence were typed except for one word which was written in rather childish handwriting. We told the children that “Joe” had written these words (this was literally true: Joe was part of our research team and had written all these words), but that he wasn’t always right in his spelling, and that it was their job to decide whether Joe had got the spelling of each of the words right or not. The words were all words that should end in “-ion” or “-ian”. In some cases, the handwritten word was spelled correctly, and in others the ending was incorrect. Here are some examples. Notice that there is a box under each sentence which contains the correct spelling of the word in question.

The

was wonderful. musician

From the laboratory to the classroom 75

The gang made a

to the police. confession

With each word, our procedure was to show the children the sentence with the handwritten word on the laptop screen. At this stage the box with the correct spelling was not visible. We asked the children to discuss with each other whether or not Joe had spelled the handwritten word correctly and, if they thought it was incorrect, to write down the correct spelling. When they had done this, the correct spelling of the word appeared under the sentence, and the display looked like the ones in our illustrations. At this point of the game, the procedure diverged for the different groups. In all the trials experienced by the explicit group and in the “explicit” sets experienced by the implicit group (the second half of the sets given to them in this game), the experimenter gave the children an explanation of why the word ended in “-ion” and not in “-ian”, or vice versa. This explanation was not given to the children in the implicit group, nor to the mixed group during the “implicit” trials that we gave to this group (the first half of the sets given to them in this game). Our final step was to give all the children in the explicit, mixed, and implicit groups a PowerPoint résumé 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 make as 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. We discussed the reason for the difference between the two endings in these two collections of words with the pairs of children in the explicit and mixed groups, but not with the children in the implicit group.

76 What does the research tell us?

barbarian

confession

Christian

conversation

comedian

combination

magician

discussion

mathematician

imagination

musician

invitation

We called one of these games the “analogy game” (see Box 3.2). We started with a simple word, like “electric” and then we presented a second word: A two-morpheme word, like “electrician”, which was derived from the first word in the pair. After that we immediately presented the children with another simple word like “magic” and asked them to produce a new word, which had the same connection to the “magic” as “electrician” had to “electric”. The pair of children had to work out these analogous words together and write down the new word. The children were then allowed to check on the computer screen we were using whether they had produced the right word, and, if not, what that word should have been. The researcher who worked with the children also provided an explanation for the correct choice, but here what the researcher said varied between the three groups, as explained earlier on. The second task was the “correction game” (see Box 3.3). In this we asked the pairs of children to decide between themselves whether the spellings of another (ﬁctional) child were right or not. If they thought that the spelling was wrong, which it often was, we asked the children to write down what it should have been. After that, we told the children whether they had got it right or not, and, as Box 3.3 shows, in some cases we explained why. In fact, the only difference between the three groups was in what the instructors told the children after they had made their decisions in both games. The instructors told the children in the explicit group not just whether their answer was right or wrong, but also why. Thus, after each

From the laboratory to the classroom 77

answer, the instructor stated clearly and explicitly the person versus nonperson 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 the explicit 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 answer was correct or not. That was the end of each problem, as far as the children in the implicit group were concerned. At that point, the instructor moved them on to the next problem. The third group was the mixed group. With these children, we divided each of the two games into two halves. In the problems in the ﬁrst half of each game, we treated these children in the same way as we had the children in the implicit group: We told them whether they were right or not, we gave them the correct answer if their answer had been incorrect, but we explained nothing about the reasons for the difference between “-ion” and “-ian” endings. In the second half of each game, we switched to the explicit method with the children in the mixed group. Now we not only told them whether they were right or not, we also gave them the reasons why. Finally, a reminder about the control group: The children in this group also received two teaching and learning sessions in between the pretest and the immediate posttest, but the intervention was about something entirely different. These children were being taught strategies that we designed to improve their text comprehension and they formed the experimental groups in a study of reading comprehension. Vice versa, the three experimental groups in the study on morphemes that we are describing here also formed the control group for the project on reading comprehension. We managed to combine the two studies in this fruitful way by giving all the children the pre- and posttests for the project on spelling morphemes and also for the project on reading comprehension. Thus, all the children who took part in these projects stood to gain something from them, some in spelling morphemes and others in reading comprehension. For once, the control-group children were not the losers.

The posttests Since the posttests were the same as the pretest (see Table 3.2), we only need to tell you about their timing at this point. We gave the children the immediate posttest on the ﬁrst school day after they went through the second and last teaching and learning session. We gave

78 What does the research tell us?

them the delayed posttest two months later, and it is worth mentioning that in that two-month period neither their teachers nor we ourselves taught them anything more about the morphemic spelling principle for the “-ion” and “-ian” endings.

The results of the ﬁrst study We should like to set the scene by talking about the pretest scores. These were our baseline, of course, against which we could check the extent to which the children in the various groups improved in spelling after the teaching and learning sessions. The pretests also tell us something about children’s knowledge, or rather their lack of knowledge, of the principle about “-ion” and “-ian” endings. We need not spend much time on this point or on these data since we have discussed them already in Chapter 2. We simply want to remind you that the children’s performance in the pretest gave us no conﬁdence at all that any of the children in the study had grasped the morphemic spelling principle for “-ion” and “-ian” endings. The children did not do at all well with real words like “education” and “magician”, even though they had had the chance to learn the spelling of these familiar words by rote learning. They made even more mistakes in spelling the pseudowords, and, since the pseudowords task is the acid test of principle learning, we felt that we could quite reasonably conclude that these children started off the experiment with little, or more likely with no, understanding of the principle that we wanted to teach them. This was not a surprise. We know that there are many aspects of our native language of which we are totally unaware. So we had expected that children might use words like “magician” to refer to persons and “education” as abstract nouns but have no awareness of the distinction between these categories or of the fact that this distinction is marked in spelling by the use of different sufﬁxes. We can turn now to the posttests to see whether the children managed any better after our instruction. It is best to consider the real words and the pseudowords separately. Figure 3.2 gives pre- and posttest results for spelling the real words. These scores are for the children’s successes with both kinds of ending. The ﬁgure tells a clear story. The two teaching and learning sessions on morphemes had an impressive effect on the children’s spelling of the real words. The scores for the three groups who were taught about morphemes were consistently better in the posttests than the pretest scores. In the immediate posttest this improvement was at its strongest, with the children in the

From the laboratory to the classroom 79

14 12 10 8

Pretest Immediate posttest Delayed posttest

6 4 2 0 Explicit method

Mixed method

Implicit method

Control

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 covariance of the scores in the two posttests, in which the pretest scores were the covariate. In the immediate posttest the scores of all three taught groups were significantly higher than those of the control group.

explicit group and, next, in the mixed group—the two groups of children who had received explicit teaching about the “-ian” and “-ion” endings. In this posttest, all three groups who had been taught in one way or another 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 the immediate posttest, but this was not at all surprising given that these children had had no systematic instruction on these sufﬁxes during the two-month interval. In fact, the decline, if that is the right expression, was very small indeed among the children in the mixed and implicit groups, but slightly larger in the case of the explicit group. However, the scores of the children in these three groups were still higher in this delayed posttest than they had been in the pretest at the beginning of the study. At the time of the delayed posttest, the explicit group was still ahead of the mixed and implicit groups, but not so far ahead. These results with real words show us two things. 1.

We did improve children’s spelling of the “-ian” morpheme, and this improvement survived a two-month period during which the children were given no instruction about morphemic spelling principles at all.

80 What does the research tell us?

2.

Explicit teaching seemed to be the most effective way of producing this improvement, particularly in the short term. Telling children directly about the principle worked.

Now we must consider the results for the equivalent spellings in pseudowords. There are two reasons why these are particularly important. One is the point, which we have made already, that pseudowords provide us with the critical test for principle learning, because they rule out word-speciﬁc learning. The second is that we only worked with real words in the teaching and learning sessions. We knew, therefore, that the children’s scores in the pseudoword posttests would give us a good measure of how well they transfer their learning to a completely different kind of material. Figure 3.3 gives the children’s scores for spelling the pseudowords. The ﬁrst thing to say about this ﬁgure is that it shows that the children who had been taught about “-ion” and “-ian” endings easily extended what they had learned to pseudowords, even though none of the instruction that they had received included any experience with this kind of word. All three groups of children given instruction on these endings did better in the posttests than the children in the control group. 6 5 4 Pretest

3

Immediate posttest 2

Delayed posttest

1 0 Explicit method

Mixed method

Implicit method

Control

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 covariance of the scores in the two posttests, in which the pretest scores were the covariate. In the immediate posttest the scores of all three taught groups were significantly higher than those of the control group. In the delayed posttest the explicit and mixed groups scored significantly better than the control group.

From the laboratory to the classroom 81

Another interesting result was that the two groups of children who were given some kind of explicit teaching (explicit and mixed methods) did better in the delayed posttest than the implicit group. Finally, we were pleased to see no decline over the two-month period between immediate and delayed posttests among the children in the mixed and implicit groups, even though they were not taught about the two endings during those two months. There was a slight decline in the explicit group’s scores from the immediate to the delayed posttest, but this group’s scores were still far ahead of their pretest scores in the delayed test. So, we drew much the same conclusions about the pseudoword spelling as about real-word spelling. 1.

2.

Teaching children about the morphemic principle did help them to use the morphemic spelling, even with a type of word that had not been part of the teaching sessions. Explicit teaching, on the whole, worked better than implicit teaching. For the second time, we can say that telling children about the principles and giving them a chance to use these principles works.

Study 2: Teaching children about the “-ian” and “-ion” endings in the classroom The results of the teaching project that we have just described are certainly encouraging. The children clearly did learn something new and important about the link between different kinds of meaning and different forms of spelling. Such results lead to an obvious but essential next step: This step is out of the laboratory and into the classroom. We had conducted a laboratory-style study. In fact the study took place in the children’s school, but we took all the children—one pair at a time—out of the classroom in order to teach them our morphemic spelling 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 no other demands on the children we were teaching. Having established that our games and our material worked in these well-controlled and advantageous circumstances, we now wanted to ﬁnd out whether they would be as effective when used by teachers working in real classrooms. Would it be possible, we asked, for teachers to put across the same morphemic principles in these quite different circumstances but using

82 What does the research tell us?

exactly the same material and the same methods that we had in our quiet ‘laboratory’ room? A great deal depends on the answer to this question. The main aim of our research was to find out whether children can and should be taught about morphemic principles at school. So we had to ﬁnd out if our 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 their classrooms. It was an almost exact repetition of our first study. The Lauriston teachers • • • • • • •

•

worked with children in the same age group as the children in our ﬁrst teaching study; administered exactly the same pretest, with the same words, as ours; conducted the same immediate posttest and the same delayed posttest two months later; taught the children about the morphemic principle in two sessions only; arranged for the children to work in pairs during the two teaching sessions; gave the children the analogy and the correction games in these two teaching 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 we had in ours: Our Lauriston colleagues also taught some children about morphemes and others about strategies for understanding text, and again the children who were taught about comprehension formed the control group in the morpheme-spelling study (and vice versa); assigned the children to the intervention and control groups randomly, as we had also done in our study.

As far as we know, this is the ﬁrst time that a laboratory study has been transferred 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 morphemes explicitly. We had established that this was the most successful method in our ﬁrst study, and our results led the teachers to adopt that method only. So in their study there was one morpheme group only: An explicit

From the laboratory to the classroom 83

group. This allowed them to compare this explicit group with the control group. The second difference was about the presentation of the material in the two teaching sessions. In the classroom teaching sessions, the children worked in pairs, as in our earlier study, but the presentation of the material for the analogy and correction games was to the whole class at the same time. As in our study, the presentation was in PowerPoint, but now it was projected onto a screen, whereas in the previous study 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 morphemic spelling principle and apply it to new words. As before, we looked at the effects of teaching the morphemic principle on the children’s spelling of: • •

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 the endings in the real words. It shows a sharp improvement in the scores of the children in the taught group from the pretest to the immediate posttest. There was a slight decline in these children’s scores during the two-month period between the two posttests, but they still spelled the words’ endings a great deal better at the end of the experiment than at its beginning. In both the posttests the spelling scores of the children who had been taught about the two endings were higher than those of the children in the control group. The scores of the control group actually improved between the immediate and the delayed posttests, which meant that the difference between the two groups was appreciably smaller in the second posttest than in the ﬁrst. This had an effect on the statistical analysis, which produced a significant difference between the taught and control groups in the immediate posttest but not in the delayed posttest. So, we can say with certainty that the classroom teaching led to an immediate improvement in the children’s morphemic spellings. We cannot be so sure about the lasting effects of this teaching. On the one hand, the children in the taught group still had far higher scores in the ﬁnal posttest than in the pretest. On the other hand, the apparently spontaneous improvement over time in the scores of the control group, though not as strong as in the taught group’s scores,

84 What does the research tell us?

14 12 10 8

Pretest Immediate posttest Delayed posttest

6 4 2 0 Taught group

Control group

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 covariance of 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 the two tests. The difference between the taught and control groups was significant in the immediate posttest (with an effect size of 0.16 of a standard deviation, which is small but quite good, considering that only two teaching sessions were used) but not in the delayed posttest.

was strong enough to make us wonder how much of the taught group’s good performance was due to the teaching and how much of it would have happened spontaneously. One distinct possibility is that the “spontaneous” improvement in the control group’s scores was not after all completely spontaneous. There could have been some “contamination”, as it is commonly but inelegantly called, between the two groups. In other words the children in the different groups could have talked to each other about what they had been taught, and this could have had a positive effect on the control group’s spellings. The idea may seem far-fetched, but there are welldocumented examples of this sort of effect in psychological research on young children. So, we can conclude that • •

explicit teaching about this morphemic spelling principle has an immediate and impressive effect on children’s spelling; we still need to know more about how to maintain this improvement over time.

From the laboratory to the classroom 85

6 5 4 Pretest 3

Immediate posttest Delayed posttest

2 1 0 Taught group

Control group

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 covariance of 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 separate analysis of covariance of the immediate posttest the scores of the taught group were significantly higher than those of the control group, with an effect size of 0.42 of a standard deviation.

We shall have more to say about this second conclusion in following chapters. At this point, we can look at the effects of the same classroom teaching on the spelling of pseudowords. The scores for these pseudoword spellings are shown in Figure 3.5. Here again we can see a striking immediate effect of teaching: The children who were taught about the two morphemic endings spelled them much better after this teaching than before and much better than the control group did in the immediate posttest. The children in the taught group did not do so well in the delayed posttest as in the immediate posttest, but they still managed to spell the two endings much better at the end of the study than 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 taught group, but by the time of the delayed posttest there was no signiﬁcant difference between the two groups. Again, we must wonder whether the improvement in the control group’s scores was a spontaneous one or the result of some talk between the children in the two groups. Thus, our main conclusion from the children’s spellings of pseudowords is that the children applied what they were taught with real words

86 What does the research tell us?

to pseudowords as well and dealt with these pseudowords in much the same way as with real words.

Our general conclusions from the ﬁrst two studies When we put the results of these two studies together, we thought that they told an interesting and useful story. This was the ﬁrst time, we think, that a study originally done in a laboratory setting had been repeated so exactly in the classroom and with such similar results. The two studies show how a method, ﬁrst developed in a laboratory study, for teaching children an entirely new spelling principle worked just as well in the classroom as in the laboratory. These two studies also established that explicit teaching of this particular principle works extremely well. The consistent success of explicit teaching also suggested that that this kind of teaching might be the most effective way of explaining morphemic principles to schoolchildren and persuading them to adopt these principles themselves. So the studies were a spur to further research on teaching children about morphemes in the real world of the classroom. We decided to widen our brief. Our basic claim is a general one: We are arguing that it is possible, and beneﬁcial, to teach children about the links between a large number of morphemes and their spelling. The studies 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 that would deal with a variety of morphemic principles and would assess the effects of this broader instruction on children’s spelling of different sufﬁxes.

Study 3: Constructing and assessing a package for teaching morphemic principles Teaching children about spelling involves two different decisions. One is what to teach, and we have made our point here quite clear: We think that children should be taught about spelling principles based on morphemes and their meaning in order to improve their spelling beyond phonological principles. The second decision is how to teach. In order to design a wide-ranging teaching package, we had to make many choices about how to teach. Some were based on the research about teaching that we had carried out before. For example, we knew from our previous work that it is better to teach a contrast between two

From the laboratory to the classroom 87

spellings (as in the “-ian” and “-ion” case) than to teach each of these separately. The ﬁrst study described in this chapter further showed that it is better to be explicit about the principles than to leave it all for the children to discover. However, there was still much to be decided about what to include in the package and how to involve the children’s logic and imagination so that they would enjoy the teaching and learn about morphemes. We feel that it is important to make many of our decisions explicit: Teachers and researchers will be in a better position to assess them if they realize that there were decisions being made. When a teaching package is built, it is not possible to assess each of these decisions separately, so the best that we can do is to be clear about our own principles in making choices.

Principles used in our teaching materials The groundwork we used to develop this teaching program was done initially in another study, which we ourselves carried out (Nunes et al. 2003), about how improving children’s awareness of morphemes affects their literacy skills. In our previous study we used word reading and spelling as outcome measures. We found that our teaching intervention about morphemes was as powerful at improving children’s word reading (measured by a standardized test) as a teaching intervention where they were taught about phonological spelling principles. We also found that only the morphological intervention improved the children’s spelling of words whose spelling is determined by their morphological structure. Our previous intervention had been carried out in a laboratory-like setting: The children worked in small groups with researchers outside the classroom. In this study, we wanted to develop this package so that it would be possible to carry out the intervention in the classroom. Thus, we examined carefully what we had done in the previous study in order to maintain what we thought were the principles of the teaching intervention, while adapting it for classroom presentation. We discuss these principles below.

It is important to develop children’s awareness of grammar Morphemic principles and grammar are connected in many ways. In the case of the “-ian” and “-ion” sufﬁxes, the words that we used were all nouns. This is not a coincidence: Derivational sufﬁxes, such as “-ian”

88 What does the research tell us?

and “-ion”, are used to form nouns, such as “mathematician” and “institution” (“-ian” is also used to form adjectives like “mammalian”). Inﬂectional morphemes, which do not change the grammatical category of the words that they are attached to, are also connected to particular word 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 have plural marks in English. Grammar is also connected to derivational morphemes in a different way. If we want to form an agent, for example, there are different derivational morphemes that could be used: “-er”, “-ian”, or “-ist”. Which sufﬁx is the correct one depends on the word class of the base form used. Agents—or, more generally, nouns that refer to persons—can be formed from verbs, nouns, or adjectives. Usually, the suffix “-er” (but sometimes also “-or”) is used to form agents from verbs: For example, “read”–“reader”, “clean”–“cleaner”, “swim”–“swimmer”. The suffixes “-ian” and “-ist” are used to form person words from nouns and adjectives: “music”–“musician”, “magic”–“magician”, “electric”– “electrician”, “Brazil”–“Brazilian”, “Italy”–“Italian” are examples of “-ian” words, and “science”–“scientist”, “art”–“artist”, “feminine”– “feminist”, “special”– “specialist”, “commune”–“communist” are examples of “-ist” words. This analysis of permissible combinations of base forms and sufﬁxes suggested that we should increase the children’s awareness of grammatical categories in order to create a basis for the development of their awareness of morphemes. It is very difﬁcult to explain grammatical categories because they are not deﬁned by content but by relationships in sentences. For example, the commonly used definition of verbs as “action words” proves confusing for children and is inaccurate: We have witnessed children arguing that in the sentence “I had a ﬁght in school,” the word “ﬁght” is an “action word” and therefore a verb. However, children are able to understand that sentences have frames and that some words ﬁt, while others don’t fit, into particular gaps in these frames. So, we created exercises 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 in Figure 3.6.

From the laboratory to the classroom 89 (a)

Do the words fit into the sentences? Tick those that fit in the sentence. Discuss with your friend why some do and some don’t.

We saw a

in the town centre.

car

computer

sing

buses

(b) We

before we go to school

quickly

eat

ball

read

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

90 What does the research tell us?

It is important to engage children’s reasoning in different 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 difference between learning principles versus learning speciﬁc facts. This research is referred to by different terms: Studies of “learning set,” “generalization” or “transfer” are all about how learning can become general, rather than remain tied to speciﬁc instances. Although there are many differences in opinion, there is a certain amount of convergence across studies. There are two crucial elements in promoting the learning of principles rather than speciﬁc facts. The ﬁrst is to promote categorization: When learners form a category that involves different speciﬁc instances, they can respond to the category and ignore the differences between the specific examples. We used this principle in a variety of ways. One example was to help the children form a category of sufﬁxes that form person words. The children would, for example, see on the screen on which the task was projected the three sufﬁxes, “-ian”, “-er” and “-ist”, and would be asked to indicate “Who is the person that . . . ?” This would help them recognize that different sufﬁxes can be used to form person words. Another example was to help the children realize that some preﬁxes refer to number. Here we asked the children to consider the most important difference between a bicycle and a tricycle and to discuss whether there was an element in the word that gave a clue to this difference. Once they were able to think about the possibility that some preﬁxes give a clue to number, they were presented with other words (for example, “triangle”, “pentagon”, “uniform”, “binoculars”) and asked to identify preﬁxes that give clues about numbers. Figure 3.7 shows some examples of this task. The second principle in creating tasks with the aim of helping children to generalize is to promote the engagement of different operations of thought in solving problems. Thus we created examples where the children had to make analogies (as in our analogy game described on p. 71), to choose the best word for a sentence, to focus on grammatical categories (see Figure 3.8), to correct spelling errors (as described in the correction game), to count morphemes in multimorphemic words, to analyze novel words (to name just a few examples).

From the laboratory to the classroom 91 (a) What is the most important difference between a bicycle and a tricycle? What is it about the word that gives you a clue to this difference?

(b)

Are there similar clues in these words? Identify these clues. What do these clues suggest?

binoculars

triangle

uniform

octopus

(c) Today in town I saw a biheaded monster. Can you draw a biheaded monster?

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

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 you add it at the end. What type of word do you form? sharp

circle

strength

bright

rage

power

Figure 3.8 Focusing on verbs.

It is important to engage children’s imagination Some of our tasks involved pseudowords, where the children were asked to think about nonexisting words that combined real sufﬁxes with made-up base forms. These tasks involved the children in imagining what a word could mean, if it existed. Some examples are presented in Figure 3.9.

It is important to maintain children’s motivation In order to maintain the children’s motivation, we designed tasks that were presented with the support of a computer, so that stimuli could be colorful and interesting. The children’s tasks were varied: They received support to formulate their own ideas about the morphemes in the task.

From the laboratory to the classroom 93 What jobs do these people do on Mars? (a)

They are spamters. They the spaceship for flying too low.

(b) They are montists. They look after , which they keep in their little black boxes.

(c) At work his job is to loment the towers. He is a .

Figure 3.9 Examples of items used to practice identification of stems and creation of person words. Playing with pseudowords was fun.

Feedback was provided in positive ways using the PowerPoint presentations: The answers moved onto the screen in different ways and often with different sounds. They also received feedback from each other when they discussed the explanations and presented them to the class.

94 What does the research tell us?

They had something to talk about and learned ways of talking about language.

The design of the study In order to assess whether our package had an impact on children’s spelling and vocabulary, we needed to have an experimental and a control group. We did not want the children in our control group to be the losers in the study: As before, we wanted to make sure that they also beneﬁted from participation. However, it was not possible to work with a parallel intervention for the control group in this study because of time constraints. We decided instead to use a waiting-list model, under which the control group receives the same learning opportunities as the taught group but at a later time. The taught and the control groups completed a pretest, an immediate posttest and a delayed posttest, which took place about 8 weeks after the teaching for the taught group had been completed. After this delayed posttest, the teacher of the control group received the same materials for use with her class. The teaching program was implemented by two teachers in a school in Oxford with their Year 4 and 5 classes, which included a total of twenty-eight children. The activities were included in the literacy hour (the time specifically aimed at teaching literacy in English primary schools) so that the taught groups did not receive extra literacy instruction in comparison with the control group; they received the same amount of literacy instruction, but worked on a special curriculum during the two-week period in which the teaching was implemented. The teachers who worked with the control groups also taught Year 4 and 5 classes, which included a total of ﬁfty children. Before the teaching program started, the researchers showed the teachers some of the activities and discussed their aim briefly. This orientation took place in a single session. The teachers received a CDROM that contained the activities and response sheets for the children to complete during the sessions. In order to support the teachers in case they had questions and to document how the tasks worked in the classroom, one researcher was always present during the teaching sessions. It turned out that the teachers did not require further clariﬁcation beyond the initial orientation. The activities were easily understood by the teachers and children alike.

From the laboratory to the classroom 95

The pretests and posttests We used a design similar to that employed in the previous studies. The children went through a pretest just before the teaching program started, an immediate posttest immediately after this program was completed, and a delayed posttest approximately 8 weeks later. These tests were identical and included the spelling of sufﬁxes in words and pseudowords. The items used are presented in Box 3.4.

BOX 3.4 The items used for the word- and pseudoword-spelling tests in Study 3 The words and pseudowords in italic were not on the children’s answer sheet and were dictated by the teacher. The children’s answer sheet contained the rest of the sentence. The children spelled the missing word above the line that marked its place in the sentence. Word spelling test 1. 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.

96 What does the research tell us?

18. 19. 20. 21. 22. 23. 24. 25. 26.

There was a great similarity between the twins. He is a graceful dancer. The librarian found the book I had lost. When he won the lottery, he cried with happiness. Everyone voted in the election. The lion tamer was famous for being fearless. The footballer was very skilful. Our destination is Athens. George brought an inﬂatable mattress to our sleepover.

The pseudoword spelling test 1. 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 method used for eliciting the spellings was the same as the one described previously: The children’s answer sheets contained sentences with a gap where the dictated word was to be written. Box 3.5 (opposite) shows a portion of the test with a child’s spelling. Because it was impossible to carry out the delayed posttest with one of the control classes, the number of control children who participated in the delayed posttest was reduced to twenty-ﬁve.

A brief overview of the teaching program The intervention program consisted of a set of activities designed to occupy approximately seven sessions of about 50 minutes each (although this could vary because the time taken up for discussion varied). Because the program was to be implemented in the classroom,

From the laboratory to the classroom 97

BOX 3.5 Sample of items from the spelling test showing one child’s answers

the tasks were prepared in PowerPoint and answer sheets were created for the children. The teacher’s task was to project the slides, to ask the children to ﬁll in their answers, to discuss the answers in pairs, and, later, to present the pair’s answer to the class. Feedback on specific responses was provided using the slides. Feedback on more general issues was to be provided by the teacher throughout the tasks. For example, one of the tasks used involved the distinction between “-ian” and “-ion”. The item was projected on the screen (for example, in the analogy game, the word pair “protect”–“protection” appeared on the screen (see Box 3.2). The children were asked to provide the missing word 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 the teacher provided further feedback. The children were also expected to discuss how the base forms ended and how the last consonant in the base form related to the middle consonant in the derived form (for example, “protect” ends in “-t” and this is the same consonant in “protection”; “confess” ends in “-ss” and this is also the consonant in “confession”). Session 1 contained four activities, the first two focused on word classes. Activity 1 is exempliﬁed in Figure 3.6. Activity 2 presented the

98 What does the research tell us?

children with four words, all of which could belong to the same grammatical category, and the children’s task was to ﬁnd more words of the same category. The choice of words was important as we needed to avoid having more than one item that could belong to more than one category. Pictures were used to help make the meaning of words clearer and 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 could belong to more than one category even if presented with a picture: For example, the word “dance” with a picture of someone dancing can still be treated as a noun or a verb. The children would have to use other words from the list to decide which of the possibilities would be right for 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). These examples 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 the appropriate 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 in this way also led to interesting discoveries by the children: For example, the combinations “unarmed” and “disarmed” as well as “uncover” and “discover” are both possible but have slightly different meanings. Session 3 included exercises on identifying the stems of words (see one example in Figure 3.9, which was preceded by an exercise with real words and a variety of afﬁxes), counting morphemes and discussing when the different sufﬁxes for person words are used (“-er” and “-or” with verbs, “-ian” and “-ist” with nouns and adjectives). Session 4 covered a comparison of words ending in “-ian” and “-ion” by starting from the children’s knowledge that “-ian” is for persons and asking them to attempt to classify “-ion” words. These were classiﬁed as feelings (for example, “satisfaction”), events—or, as some children said, things that happen—(for example, “infection”, “confession”), mental processes (for example, “imagination”, “addition”), etc. The discussion can come to a notion that these are not “concrete things”. This session also included trying out different sufﬁxes and preﬁxes with some stems to see how many words could be formed; adding “-en” to see

From the laboratory to the classroom 99

what word class is formed (see Figure 3.8 for examples) and preﬁxes that 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 the concept of word class. It also included transformations from base forms to abstract nouns using a variety of endings (for example, “intelligence”, “distance”, “ability”, “misery”) and a discussion about the fact that the ending “-y” has different functions (for example, in “thirsty” and “funny” it is used to form adjectives, and in “misery” and “poverty” to form abstract nouns) but “-ly” is only used for adverbs. Session 6 contained exercises of ﬁtting the appropriate words from a list of words with the same stems into sentences, building words with stems and afﬁxes, and comparing the effect of adding “-less” and “-ful” to words. Session 7 contained exercises of forming adjectives (with “-y”, “-al”, and “-able”), adding preﬁxes to change the meaning (“re-” and negative preﬁxes) and some revision exercises such as adding “-en” to form verbs and different endings to form abstract nouns.

The results of the third study There were some differences between the taught and the control groups at pretest: The control group performed signiﬁcantly better than the taught group both in spelling sufﬁxes in words and in pseudowords. For this reason, the statistical analysis used to assess the effectiveness of our teaching program was analysis of covariance, which controls statistically for the differences that existed at the beginning between the groups. Figure 3.10 shows the adjusted means at pretest and for both posttests by group for the correctness of spelling sufﬁxes in real words. Because these 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 comparisons between the estimated means for the posttests, controlling for the pretest performance. The figure shows that both groups made progress between the testing occasions. However, the taught group

1

The preﬁxes “en-” and “in-” are often spelled as “em-” and “im-” when the subsequent letter is “p” or “m”. We prefer not to distract the children from the classiﬁcation target and do not spend time discussing this at length, but the children may note this and ask about it.

Number of suffixes spelled correctly

100 What does the research tell us?

18 16 14 12 10 8 6 4 2 0

Pretest Immediate posttest Delayed posttest

Taught group

Control group

Figure 3.10 Adjusted means at pretest and for both posttests by group for the correctness of spelling suffixes in Study 3. Note: Effect size for immediate posttest = 0.42 SD. The difference between groups at the delayed posttest was not significant.

made signiﬁcantly more progress than the control group from pre- to posttest. Although there was no decrement in the taught group’s performance between the immediate and the delayed posttest, the control classes that participated in the delayed posttest showed an unexpected improvement between the two posttests. This result was not expected as there is no systematic teaching of sufﬁxes in the literacy hour covering the range of words that we included in our spelling assessment. However, it is quite possible that the teachers coincidentally gave their children practice in some of the words in our assessment and that this had a positive effect on their performance. If this is the case, we should observe a different pattern of results in the spelling of sufﬁxes in pseudowords. We think that the spelling of sufﬁxes in pseudowords is the acid test that shows that the children understand the principle of analyzing words into morphemes and thus are able to spell the sufﬁxes in pseudowords, although these pseudowords were not practiced in the classroom. The results of the analysis of spellings of sufﬁxes in pseudowords are presented in Figure 3.11. These results show that the children in the taught group made an improvement from the pretest to the immediate posttest and maintained this improvement in the spelling of sufﬁxes in pseudowords, just as they had maintained the improvement in the spelling of words. The statistical analysis showed that at both posttests the taught group performed signiﬁcantly better than the control group, and that the control group made negligible improvement across testing occasions. Therefore, we

Number of suffixes spelled correctly

From the laboratory to the classroom 101

8 7 6 5

Pretest

4

Immediate posttest

3

Delayed posttest

2 1 0 Taught group

Control group

Figure 3.11 Adjusted means at pretest and for both posttests by group for the spelling of suffixes in pseudowords in Study 3. Note: Effect size for immediate posttest = 0.65 SD; for delayed posttest = 0.42 SD.

can conclude that the unexpected improvement in performance in the control group at the delayed posttest is likely to result from specific practice. The evidence suggests that the taught group learned how to use principles of morphemic analysis in order to spell rather than simply learned how to spell particular words that contain particular sufﬁxes.

Summary and conclusions The research presented in this chapter makes unique contributions to the understanding of the development of children’s spelling. The first study demonstrated the importance of making children explicitly aware of suffixes and their meaning in order to help them distinguish spellings that cannot be distinguished from the way the words sound. It is not sufficient to know that words than end in the sound /un/ are sometimes spelled with “-ian” and sometimes spelled with “-ion”, it is also important to be aware of the basis upon which it is possible to predict which of these two spellings is correct. This can only be accomplished if the children become aware of the meanings of these derivational suffixes. As discussed in Chapter 2, children have to learn the forms used in spelling and also their function. If their function is to represent units of meaning, they have

continued

102 What does the research tell us?

to make a connection between the letter strings “-ion” and “-ian” and their meanings. The second study is a very important next step in the construction of evidence-based practice: The study was carried out in the classroom and replicates in detail the results that we observed in the laboratory-like situation, when the children were taught in pairs by researchers, outside the classroom. In the second study, the children were taught by their teachers in their classrooms. Although a stringent random assignment was made to the taught and control groups, and although the children in the control group had an equivalent amount of additional literacy experiences and exposure to written text, the taught group made significantly more progress than the control group. The third study also makes a unique contribution to the knowledge available on the teaching of spelling. We developed a teaching package that utilized a variety of suffixes and prefixes to show how these morphemes can be used to analyze words and how they relate to the meaning of words. The trial of this teaching program showed that it can easily fit with the regular literacy instruction that is offered in schools and that it is readily understood by teachers, with a minimal amount of training. The use of a waiting-list model to assign the classes to taught and control conditions replaced the random assignment of the children used previously and ensured that all teachers 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 use of morphemic principles for analyzing and spelling words. The children in the taught group made significantly more progress in spelling suffixes in pseudowords, our acid test for the learning of general principles of morphemic analysis, and this improvement was sustained in a delayed posttest, even though the teachers were no longer using activities from our program. These results encouraged us to move to the third phase in this work, which was to make the teaching program available to teachers in 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 we were 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 the reasons that we gave earlier, but we realize that other teachers in other schools who adopt these methods in the future would not apply them in one prescribed way. We hope that many teachers will adopt these new games and exercises and new ways of measuring their pupils’ progress, but we know that they will do so in different ways. The amount of time that teachers will spend on each game, the distribution of the lessons, the way they organize the classroom (with children working by themselves or in groups), how much explicit teaching they will provide and how much discussion among the children they will stimulate will vary from teacher to teacher, and quite rightly so, for the teachers will be the best judges of these details. So, we decided that in our next classroom studies we would look at how well our program worked in schools when teachers had complete freedom about how and when to administer the new program. We wanted to establish that the program still worked with this degree of variation and flexibility, and so in the classroom studies that we shall describe in the subsequent chapters, we encouraged the teachers who used our teaching program to do it their way. The results of the studies that we have described in this chapter encouraged us in another way too. Their success led us to think of other possible benefits of teaching children about morphemes. Vocabulary was the obvious next step, for reasons that we have given in the first two chapters, and so we set ourselves another question: Could our methods of teaching children about the morphemic structure of words also help them to learn new words? Most of the new words that schoolchildren learn are derived words, as we mentioned in Chapter 1. If children know more about how many words are derived from other words, perhaps they will be able to add new words to their vocabulary more rapidly and effectively as a result. The results of the classroom studies where teachers used their own approach in implementing the teaching program we designed are presented in the two chapters that follow. Chapter 4 focuses on spelling measures as outcomes and analyzes a variation across two types of program using the same principles. Chapter 5 considers the impact of the teaching program on vocabulary measures.

Chapter 4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

An intervention program for teaching children about morphemes in the classroom Effects on spelling

In the previous chapter, we saw that teaching children about morphemes has a strong and positive effect on their ability to analyze words into morphemes and to spell their suffixes. This seems to apply not only when children are taught on a one-to-one basis or in small groups but also when the teaching is delivered to the whole class. However, in the studies we have discussed so far, the teaching was always delivered by researchers, unaffected by the daily pressures of school life, or by teachers working in close connection with researchers. In this and in the following chapter we will describe a project that assessed how our ideas work when used by teachers in real-life classroom settings and as part of their normal teaching. This chapter concentrates on the effects that the teaching program had on the children’s ability to spell.

Overview of the classroom spelling project The classroom project included 201 children in Year 5 and consisted of a spelling pretest, followed by an intervention program, and then an immediate posttest, which was the same as the pretest. The teaching intervention was always delivered in the children’s classroom by their teacher. We simply gave the teachers a CD-ROM containing all the intervention tasks and some brief instructions on how to 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 Authored by Freyja Birgisdottir, Terezinha Nunes, Ursula Pretzlik, Diana Burman, Selly Gardner, and Daniel Bell

Intervention program: Effects on spelling 105

free to deliver the intervention following their own pace, depending on their teaching schedules. Thus, while some of them went quite rapidly through the teaching program, others preferred to have longer intervals between sessions or split some of the sessions into two or three parts. Our only request was that the teachers would complete all the sessions and deliver them in the order we had planned. The teachers and children who participated in the study came from ﬁve different schools; three were in London and two in Cheltenham. We divided the children in these schools into three groups: Two groups received an intervention program and the third was an unseen control group. It was not possible to develop activities for this control group as we 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 intervention. One version was the same as that used in the intervention study described in Chapter 3 and concentrated both on enhancing children’s awareness of morphology and on promoting the link between morphology and spelling. We refer to this intervention group as the “morphemes-plus-spelling” group. The other version contained the same tasks but concentrated on enhancing the children’s awareness of morphemes in oral language without placing emphasis on the connection between morphemes and spelling. We call this intervention group the “morphemes-only” group. We developed this second form of the program because we wanted to see how teaching children explicitly about morphemes affects their ability to spell morphologically complex words when they have to work out the spelling principles more or less for themselves. We will describe the difference between the two versions of the intervention in more detail later on in the chapter. The control group comprised children from two schools in Cheltenham and one of the London schools; in the latter, the children were randomly assigned either to the control group or to the morphemes-plus-spelling intervention. Table 4.1 gives the details of the children’s age at the time of the posttest. Although the differences between the groups were small, they were statistically signiﬁcant. So our statistical analyses will control for 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 literacy skills of the children turned out to be different across the groups. Consequently, we will also need to control for these differences statistically.

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 Morphemes only Morphemes plus spelling For the total sample

10.2 10.1 9.9 10.00

.31 .33 .34 .35

75 26 100 201

The intervention sessions We need not describe the general pattern of the intervention program in 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 was divided into three or four different tasks, each of which was presented in a context of a word game. Every task had a special educational content, such as learning about word classes, how to break words into morphemes, or how different morphemes inﬂuence the meaning and the grammatical status of words. The tasks also involved a number of different cognitive operations, such as counting the number of morphemes in a word, adding preﬁxes and sufﬁxes to a stem to make a new word, or working with analogies. We prepared the tasks in PowerPoint, so that they could be projected on a screen or on a whiteboard in front of the whole class. The children doing these tasks typically work in pairs or small groups and produce their answers either orally or in writing on their work sheets. We took special care to make the tasks enjoyable for the children and prepared them in a way that would encourage discussion between them. We used two different versions of the intervention in this study. The two versions contain exactly the same tasks as each other, but the stress given to spelling in each task differs between the two versions. The morphemes-only version was designed primarily to develop children’s awareness of morphemes in spoken language and does not make an explicit link between morphology and spelling. Thus, although the children have to write down the words when they produce their answers, the emphasis in this version is on discussing how morphemes are used to form words and how they inﬂuence meaning and grammatical categories. The morphemes-plus-spelling version, in contrast, makes the connection between morphemes and spelling an explicit one. The children go through the same tasks as in the morphemes-only intervention, but are

Intervention program: Effects on spelling 107

always encouraged to write down the words they are working with and to discuss the meanings as well as the spellings with their peers and with their teacher. The tasks are also often followed by a discussion of a 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 the sufﬁx “-less” are always spelled with a double “s”). The central goal is to make the children aware that there is an underlying system in the spelling of morphemes—that the same morpheme tends to be spelled the same, even if the sound of that morpheme changes from word to word. This is often useful in the case of stem spellings because stems often change their pronunciation when a derivational morpheme is added 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 word “magician” the stress is on the second vowel and the ﬁrst vowel is now a schwa vowel, not clearly articulated. Further, the last consonant in the word “magic” is clearly articulated; when it becomes a middle consonant in the word “magician”, it represents the sound which we would normally spell with “sh”. Box 4.1 shows two examples of the contrast between the morphemesonly and the morphemes-plus-spelling activities. In one of them, for instance, the children saw a word pair, such as “magic” and “magician” on the screen and then a third word, such as “electric”, appeared underneath (see Box 3.1 for a description of this task). In both versions of the intervention, their task was to use the ﬁrst-word pair as a clue to ﬁgure out what the missing word should be (in this case “electrician”). The children in both groups had the opportunity to ﬁnd out that there is a morpheme, “-ian”, that is used to form person nouns, and a different one, “-ion”, that is used for nouns that do not refer to the person. However, the children who received the morphemes-plus-spelling intervention 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 in the session and were encouraged to find the stems and endings, to discuss the ﬁxed spellings of these endings, and to describe how they might 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 how they relate to the other words on the screen. For example, they could focus on the fact that the word “electrician” is derived from the word “electric” and therefore has to be spelled with a “c” in the middle, or

108 What does the research tell us?

BOX 4.1 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 Example A (see Box 3.1 for a description of this task) Person words end with:

Abstract nouns end with:

musician librarian Hungarian technician vegetarian

confession protection discussion subtraction imitation

Note: The suffixes were highlighted in red. The children were invited to create further examples of words spelled with the same endings and discuss how they might teach a friend a good method for spelling these words correctly.

Example B In the task common to both groups, the children had to make abstract 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 that focused on spelling was assigned one further task: To find the suffixes they had used and find other abstract nouns spelled in the same way. They were also asked whether they could find another suffix for abstract nouns that did not appear in this task.

Intervention program: Effects on spelling 109

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 the intervention is that if children understand the morphological structure of words and are able to use that knowledge when they spell, they often have a much greater chance of succeeding than if they spell words purely on the basis of sound or from memory.

The pretest and posttest assessments We used the same two spelling tests to assess the children’s progress as a result of the intervention as we had in the previous classroom intervention study (Study 3) described in Chapter 3, Table 3.3. The ﬁrst test included twenty-six two- or three-morpheme words, each containing either a preﬁx or a sufﬁx. As we showed in Chapter 3, the words cannot be spelled correctly purely on the basis of sound. For example, in the word “election”, both the sufﬁx and the medial consonant are spelled differently from the way they sound. The same applies to the words “politician” and “destination”. Words such as “happiness” and “fearless” seem very simple, but why should their endings have “ss” rather than a single “s”? Words like “inflatable” and “comfortable” contain a schwa vowel in the suffix, which leaves children in doubt about the spelling. Of course, the reason is that these are morphemes with ﬁxed spellings. We used the same method in the spelling assessments as the one that we described in previous chapters. The children heard each word that they had to spell ﬁrst in a sentence and then again on its own. The same sentence was also provided on an answer sheet in front of them, with the target word replaced by a blank line. The children’s task was to write the missing word on the line. So, for example, when they had to spell the word “musician”, we read out to them, “My sister wants to be 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 of them in total and each was composed of a made-up stem and a real sufﬁx (for example, “mape-ful”, “lagic-ian”). The pseudowords were always presented in two different forms and were always presented in a sentence context. For example, when the children had to spell the pseudoword “lagician” they heard, “A person who does lagic is a lagician. Lagician.” The children then had to write “lagician” on their answer sheet.

110 What does the research tell us?

Our reason for including this test is that pseudowords cannot be spelled from memory and provide, therefore, an excellent way of checking whether children really use their morphological knowledge when they spell novel words. The test also enables us to see whether the children were able to preserve the spelling of unfamiliar stems across different word forms, even when the sound of that stem changes (for example, from “magic” to “magician”).

The results Our main hypothesis was, again, that teaching children about morphology improves their ability to spell, especially when the connection between spelling and morphemes is explicitly pointed out to them. We predicted, therefore, that the two intervention groups would make more progress in spelling between pretest and posttest than would the control group. We also predicted that the performance of the children who received the morphemes-plus-spelling intervention would improve more between the two test sessions than would the performance of the morphemes-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, we dealt with this discrepancy between the groups by using analyses of covariance, which control for the differences between the groups at pretest statistically. Figure 4.1 summarizes the children’s performance in spelling the sufﬁxes in real words before and after the intervention. Because the means displayed are the adjusted means, controlling for differences at pretest, it is possible to make direct visual comparisons between the groups. As you can see in Figure 4.2, where we compare the children’s average spelling scores before and after the intervention, all three groups seem to have made some progress. This was to be expected, since all the children had taken the same test twice, with a relatively short time interval between them. However, the two intervention groups improved signiﬁcantly more than did the control group. Thus, our main hypothesis that teaching children about morphemes enhances their spelling skills was supported. In contrast, our prediction that the morphemes-plus-spelling intervention group would show more improvement than would the

Intervention program: Effects on spelling 111 22 21

Mean score

20 19 Pretest

18

Posttest

17 16 15 14

Control

Morpheme only

Morpheme with spelling

Groups

Figure 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. Note: The overall effect of the intervention was significant (p<0.001). Both intervention groups differed significantly from the control group after we controlled for pretest differences and the age differences at posttest. The effect size was equal to 0.3 of a standard deviation for both intervention groups. The difference between the two interventions was not statistically significant.

morphemes-only group was not confirmed by the results. The two groups made similar progress between the pretest and posttest in both spelling tests. This came as a surprise to us, especially since our earlier ﬁndings indicate that teaching children directly about the link between morphemes and spelling helps them to spell morphologically complex words (see Chapter 3). In this assessment, we considered only whether the children had spelled the sufﬁxes correctly. We thought that perhaps the morphemeplus-spelling group would have an advantage over the morphemes-only group if we considered the spelling of the whole words. We reasoned that 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

112 What does the research tell us? 49

44

Mean score

39

34 Pretest Posttest 29

24

19

14

Control

Morpheme only

Morpheme with spelling

Groups

Figure 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. Note: The overall effect of the intervention was significant. Both intervention groups differed significantly from the control group, after we controlled for pretest differences and the age differences at posttest. The effect size for both groups was 0.2 SD. The difference between the two interventions was not statistically significant.

the morpheme-only group. In order to obtain a score that reﬂected the children’s 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” and “destination”, where the middle consonant represented an added hurdle, into three segments. One was the stem up to the ﬁnal consonant, the second segment was the ﬁnal consonant in the stem, and the third the sufﬁx (for example, “magi”/“c”/“ian” and “destina”/“t”/“ion”). Each segment was awarded one point if correctly spelled. Box 4.2 shows a few examples of how we divided and scored the words. If you look at the word “politician” for instance, you can see that the children received three separate points for this word: one point for spelling the stem correctly, 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

Intervention program: Effects on spelling 113

BOX 4.2 Examples of the segmentation used in scoring on the word- and pseudoword-spelling tests Words Magi/c/ian Elec/t/ion Politi/c/ian Happ/i/ness Comfort/able Polite/ness

Pseudowords Grat/less Blage/ness Saught/i/ness Senk/t/ion Prell/ian Kring/able

like “comfortable” and “politeness”: One for the stem and one for the sufﬁx. We then ran a similar analysis as the one described previously, but this time we used the children’s scores for correctly spelling the whole words in the test, rather than the scores for the sufﬁxes only. Figure 4.2 shows 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 both intervention groups made more progress than the control group from pretest to posttest. The two interventions did not differ signiﬁcantly from each other. The effect of the intervention on the children’s ability to spell sufﬁxes becomes even more apparent when we look at how they spelled individual words before and after the intervention. Box 4.3 shows how one boy 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 his responses in the pretest you can see that his primary strategy seems to have been to spell the words according to the way they sound. Thus, he spelled “magician” as “magishon”, “musician” as “musison”, “richness” as “richnes”, and “disability” as “disabilatey”. His phonological strategy does not even seem to have been entirely consistent. For example, he spelled the sufﬁx of “magician” and “musician” differently, despite the fact that these sound exactly the same. The same applies to the sufﬁx of “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

114 What does the research tell us?

approach had completely changed. He stopped relying so much on how the words sound and he seems to have understood that some words are made up of different parts and that these parts tend to be spelled the same from word to word. This means that he is now able to spell difﬁcult sufﬁxes such as “-ian”, “-ness”, “-ity” and “-able” correctly in different words. In fact, the percentage of sufﬁxes he spelled correctly rose from 27 percent in the pretest to 81 percent in the posttest.

BOX 4.3 A sample of the same boy’s spelling in the pretest and posttest

Intervention program: Effects on spelling 115

Although this particular boy may be an extreme example of the success of the intervention, the same basic trend always appeared when we examined the spellings produced by the other children in the two intervention 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 this strategy for spelling sufﬁxes and started to use their newly acquired morphological knowledge. As discussed in the previous chapter, we think that the critical test of children’s learning of a new strategy, rather than the learning of speciﬁc word, is their performance in the spelling of sufﬁxes in pseudowords.

116 What does the research tell us?

Because these were not included in the teaching intervention, progress in spelling sufﬁxes in pseudowords indicates the acquisition of a spelling principle, rather than the learning of speciﬁc words. We ran an analysis that controlled for differences in pretest scores in order to compare the intervention and control groups. The results of this analysis are summarized in Figure 4.3. They conﬁrm the results of the previous analyses by showing signiﬁcant effects of both interventions and extend these results to support the idea that the children were learning a spelling principle not only the spellings of speciﬁc words. To summarize, the children who received the two versions of the intervention made signiﬁcantly more progress between the pretest and posttest than did the children who received no teaching in morphology, both in the spelling of sufﬁxes and in the spelling of morphologically complex words whose spelling is not predictable from the way they 10 9 8

Mean score

7 6 Pretest

5

Posttest

4 3 2 1 0

Control

Morpheme only

Morpheme with spelling

Groups

Figure 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. Note: The overall effect of the intervention was significant (p<0.05). Both intervention groups differed significantly from the control after we controlled for pretest differences and age differences at posttest. The difference between the two interventions was not statistically significant. The effect size for both intervention groups was 0.3 SD.

Intervention program: Effects on spelling 117

sound. Most importantly, without an exception, this difference between the intervention and the control groups appeared irrespectively of whether the children had to spell words or pseudowords. This means that the greater improvement of the children in the two intervention groups cannot be explained in terms of them simply getting better at spelling speciﬁc words from memory.

Does the intervention help both low and high achievers? One question that we are asked often is whether the interventions we designed are effective both for high and low achievers. Some teachers make the interesting suggestion that our intervention should work best with high achievers, because these are difﬁcult concepts, which might go above the head of low achievers. Other teachers hold the view that high achievers cannot learn much from a teaching intervention that focuses on morphemes, because they are likely to get there on their own, without explicit teaching. Picking up on the teachers’ questions and hypotheses, we decided to analyze our data in search for an answer. Because we had a relatively large number of children and they were all in the same year group at school, this study offered a good opportunity for analyzing the effects of the intervention with low and high achievers. We did so by splitting the children into two groups according to their pretest scores: Those whose scores were up to the median were placed in the lower group and those who scored above the median were placed in the higher group. The groups differed considerably in two ways: They had, of course, different means, but they also had very different degrees of variability. The lower scoring group showed a significantly larger standard deviation than the high scoring group because there was a large tail of children with very low scores in the group classiﬁed as below the median at pretest. It should also be remembered that the number of children in each group is now, by deﬁnition, half of what it was in the previous analyses. So, weak effects would no longer be signiﬁcant. Figure 4.4 presents the results of the analysis of the children’s scores in spelling sufﬁxes in words. Both interventions had a signiﬁcant impact on the scores of children who scored below the median at pretest. For the children who scored above the median, the morphemes-plusspelling intervention was more effective, although there was a signiﬁcant overall effect of the interventions. Our intervention, therefore, works across the ability range with high achievers and with low achievers.

118 What does the research tell us? 25

20 19

24 18 23

16

Mean score

Mean score

17

15 14 13

22

21

20

12 19 11 10

Control

Morpheme Morpheme only with spelling

18

Control

Groups

Morpheme Morpheme only with spelling

Groups

Pretest

Pretest

Posttest

Posttest

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 the pretest. Notes: The results on the top graph are for children who scored below the median at pretest (N=93) and those on the bottom graph are for children who scored above the median at pretest (N=86). The overall effect of the intervention was significant for both groups. Both intervention groups differed significantly from the control group for the children who had low scores at pretest; only the morphemes-plusspelling group differed significantly from the control group for the children who had scored above the median in the pretest.

These results also suggest that teaching about the explicit connections between morphemes and spelling is relatively more effective with the more advanced children, who probably already have some awareness of 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, the difference between the performance of the morphemes-plus-spelling group and the control group just fell short of signiﬁcance (p=.07), but

Intervention program: Effects on spelling 119

this difference was not signiﬁcant for the below-median children in the morphemes-only group. For the children who scored above the median at pretest, both types of intervention had a signiﬁcant effect. Thus, both types of intervention helped the low and high achievers with real words. With the pseudoword spellings, both types of intervention improved the high achievers’ spellings, but only the morphemesplus-spelling intervention helped the low achievers.

Summary and conclusions The central aim of this classroom project was to explore whether instruction about morphology in the classroom increases children’s ability 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 who received the morphology intervention improved much more than did the spelling scores of the children who received no extra instruction in morphology. This indicates that the intervention was successful in its specific aim to increase children’s ability to use their knowledge of morphology when they spell. Thus, we can now safely conclude that the effect of morphological instruction on children’s ability to spell is not limited to tightly controlled experimental settings but also applies when the instruction is delivered by teachers in busy classrooms and as part of their day-to-day teaching schedule. However, there is one finding that we have yet to explain, and that is why the children who received the morphemes-plus-spelling intervention did not improve more than the children who received the morphemes-only version. Because of the nature of this study, it is difficult for us at this point to come up with a definite explanation. Since we wanted to see how the morphological instruction worked in “real-life” classrooms, it was important to us that the teachers should deliver the interventions in their own way, without any assistance from researchers. However, the consequence is that we have no way of knowing what exactly went on in the intervention sessions or the extent to which the teachers followed our advice of how to use the tasks. This is of particular importance when we compare the success of the two versions of the intervention, because the central difference between

continued

120 What does the research tell us?

them lies not in the construction of the tasks themselves but in how they are implemented. For example, it was really up to the teachers who used the morphemes-plus-spelling intervention whether they concentrated mainly on the children’s awareness of morphemes in spoken language or whether they took the extra step of emphasizing the link between morphemes and spelling. It is possible, therefore, that the difference in the actual implementation of the two interventions was not so great after all. The only way to determine whether the two teaching regimes really influence children’s spelling skills in different ways is to take one step back and to try them out in a more tightly controlled environment where we can observe how they are actually implemented. This will have to be left for the future, because at this point in our research we decided to focus on a different outcome of children’s learning about morphemes, their vocabulary growth.

Chapter 5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

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

We have now described, in Chapters 3 and 4, how our ideas about teaching morphemes fared when we took the path to the classroom and asked schoolteachers to use our methods with whole classes of children. In this chapter we shall describe what happened when we took the same path with our idea that teaching children about morphemes should help them to learn new words as well. We had three reasons for investigating whether teaching children about morphemes has a positive effect on their vocabulary growth. The first is that various theories (see Chapter 1) suggest that growth in vocabulary is a possible outcome of enhancing children’s awareness of morphemes. Morphemes are units of meaning, and most of the words that children learn from the middle of primary school on are polymorphemic words. So, if children have a way of analyzing these words, they might find them easier to learn. The second reason is that vocabulary is an important part of literacy learning: Children’s text comprehension is highly related to the size of their vocabulary. The third reason is also important: Research on how to teach children about vocabulary has dealt predominantly with the question of how to teach specific new words. This research focuses mainly on how many repetitions are required and whether it is best to get the children to encounter new words in a text or in isolation, to provide them with definitions or with different sentences in which the meaning and use of the word is illustrated. Our own idea is different to this. We argue that teaching children word-attack strategies based on morphemes should help them to analyze new words in order to understand their meaning. We carried out the next study to test this idea.

Authored by Terezinha Nunes, Peter Bryant, Ursula Pretzlik, Diana Burman, Daniel Bell, and Selina Gardner

122 What does the research tell us?

The context of the study The teaching in this study was done by schoolteachers who were participating in a broader project that aimed at improving children’s speaking and listening skills in the Hillingdon Cluster of Excellence, a program run in a cluster of schools in Hillingdon (West London) and supported by the Department for Education and Skills (DfES) in the U.K. We will refer to this project as the Speaking and Listening Skills project. All schools participating in the project did so voluntarily. The project continued for two years. In the ﬁrst year of the project, we approached the teachers and discussed two possibilities with them. One was that they would design their own interventions for improving children’s speaking and listening. The other possibility was for them to sign up for a project, designed by us, which aimed at improving children’s vocabulary by making the children more aware of how morphemes are used to construct words. Of the nine schools that participated during the first year, four designed their own interventions, none of which involved teaching children about morphemes. The remaining ﬁve schools opted for the morpheme-teaching program that we had designed. All participating teachers agreed that the children in their projects would be assessed on measures that would be relevant for evaluating the different teaching interventions that had been designed. Thus, the children who went through our own morpheme-teaching program were the intervention group in this study, whereas those who received some form of intervention that had been chosen by the teachers themselves became the control group in the study. The ﬁrst year was a development year. The teachers who had created their own interventions sharpened their ideas by becoming clearer about what they thought they could achieve with their teaching intervention and developed expertise in evaluation. They became more aware of the importance of ﬁnding ways of assessing the outcomes of their teaching and developing an evidence-based approach to the teaching. We had the opportunity to create measures that the teachers felt were fair assessments of their aims, to test these measures, and to analyze their reliability. The results reported here are from the second project year, when all teachers were implementing interventions that had been developed and assessed in the ﬁrst year. In the second year, some new schools joined the program; twelve primary schools now participated in the Speaking and Listening Skills

Intervention program: Effects on vocabulary 123

project. The teachers who had already been involved in the project had the option of continuing to work with the intervention that they had implemented in the previous year or of moving on to something new. The new teachers joining the project found out about all the ongoing teaching programs and had the chance of choosing a program for the year. Seven schools chose to work with the morpheme intervention and ﬁve worked with other intervention programs. Once again, the participating teachers agreed to collect assessments that would allow for an evaluation of all the teaching programs being implemented, and each group acted as a control for other interventions. So, all the children were receiving some intervention, their teachers had opted for working with the teaching program that they were using, and the schools were in the same region of the country. Here, as in most of our studies, the control children were not losers in a research program. The teachers who used the morphemes intervention that we designed received a CD-ROM with the activities and a brief induction to the use of the intervention materials. This took the form of a session during which the researchers showed some of the activities to the teachers, explained their aims, and gave the teachers the opportunity to ask questions. It was made clear to the teachers that they could implement the program in the way that they thought would suit their class best. The distribution of sessions over time and whether the children worked in pairs or small groups were all decisions made by the teacher. They were asked only to ensure that all activities were implemented in the order that they appeared in the CD-ROM and to encourage the children to discuss their responses, because the program aimed to increase the children’s awareness of morphemes and speaking about their ideas would help enhance this awareness.

The design of the study The children in the project were either in a morpheme-intervention group, which means that they were given especial instruction about morphemes, or they were in a control group and did not get any extra experiences with morphemes but had experience with one of the other teaching programs in the project. Four teaching programs were used. Telling and acting out well-known stories, developing emotional language, analyzing poems, and describing paintings reproduced in posters obtained from the National Art Gallery in London. Table 5.1 presents a summary of the numbers of participants at pretest for the different groups by their year group in school.

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

School year

Mean age

SD

N

3 4 5 6 7 N

7.73 9.10 9.90 11.00 11.76

.39 .26 .30 .35 .43

71 38 34 35 4 182

Total

Mean age

SD

N

7.83 8.90 9.85 10.73 11.50

.37 .29 .33 .28 .30

78 75 119 38 9 319

Mean age

SD

N

7.80 8.95 9.86 10.85 11.62

.38 .30 .32 .34 .38

149 113 153 73 13 501

The children participated in a pretest, followed by the intervention that was described in Chapter 4 as “morphemes only”, which was followed by an immediate posttest and a delayed posttest approximately 8 to 10 weeks after the completion of the program. All tests were administered and scored by the researchers. Scoring was blind to the type of program that the children had participated in.

The pretests and posttests The pretest and immediate and delayed posttests were identical. They contained a vocabulary test and a pseudoword-deﬁnition test. The vocabulary test was designed to be easy to administer in the classroom. The children are presented with a picture booklet that contains the same pictures projected onto a screen by the researcher. The researcher reads a sentence, which has a missing word, and the children have three choices for what the missing word will be. The researcher says the sentence three times, each time completing it with one of the words. Each of the choices appears on the screen when the researcher reads it and each appears in a different color so that the children 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 into the sentence. Some examples of items and information about the assessment 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

Intervention program: Effects on vocabulary 125 (a) She always arrived late. She was

unrelated

reliable

unreliable

(b) The doctor told Georgia not to worry because the injection would be

painless

painful

pointless

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 sentence with each of the words fitting into it, one at a time. The words appear on the screen as the researcher reads them. The children tick on their books the box next to the word that they think is the correct choice.

126 What does the research tell us?

to 207 children in this age range, according to the Cronbach’s alpha is 0.9. This is very high: Acceptable consistency levels must be 0.7 or more. We also gave the vocabulary test and the British Picture Vocabulary Scale (BPVS) to a smaller group of children (N=24) and found that it had a high (r=0.66) and signiﬁcant correlation with the BPVS. Thus, we established that this assessment is reliable and valid in the ﬁrst year of the project. The pseudoword-deﬁnition test was designed for reasons that are similar to those that motivated us to design the pseudoword-spelling test. Our program aimed at promoting children’s 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 to interpret 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 have learned them by direct instruction or by getting at their meaning through analyzing them into morphemes. So, we designed a test that we thought would get more directly at their ability to analyze polymorphemic words in search for their meaning. We created pseudowords by combining a stem and an affix in a nonexisting combination. We presented the children with these pseudowords, told them that these made-up words do not really exist, and asked them to try to think what these made-up words would mean if they did exist. The instructions for the assessments and the list of pseudowords used are presented in Box 5.1. The children received one point if their answer took into account the meaning of the stem and of the affixes; no points were awarded otherwise. We analyzed the reliability of this assessment in the ﬁrst year of the project by looking at the reliability of the scoring procedure and at the internal consistency of the items. The method of scoring proved reliable: Two researchers who scored the same tests independently showed 98 percent of agreement in their scores for the items. The items showed a high internal consistency; the Cronbach’s alpha varied between 0.75 and 0.85 with different samples. A value of 0.7 is considered acceptable. Finally, this test showed a signiﬁcant correlation with our vocabulary test (r=.57), and, thus, there is evidence for its validity. For practical reasons, not all the children participated in all of the assessments, as this would have taken a very large amount of time from teaching. So, each class in one intervention was assigned as a matched control class for another intervention, and only a small number of children participated in the assessment of both outcome measures for

Intervention program: Effects on vocabulary 127

BOX 5.1 The instructions and the items in the pseudoworddeﬁnition 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 to someone 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 type of chair do you think it would be? If the children come up with answers such as “a chair for children” or “a child-size chair” or “a small chair”, they are encouraged to explain how they knew that. If no answer is elicited from the class, the researcher can ask whether this would be a large chair, for grown-ups, and then ask how they know the answer. I am now going to show you some made-up words, and your job is to try to think what these words could mean. Write the answer in your booklets. For children in the third year of school, aged about 8, the test was administered orally and individually by a researcher; the responses were also oral. The pseudowords are projected onto the screen and the researcher reads them out loud. The children write their answers and the researcher verifies that they have finished before moving on to the next item.

List of words used in the pseudoword-deﬁnition task bricker unlie unclimb

bookist unwork triwinged

rewet biheaded resleep

uncomb shoutist chickener

the evaluation of the morpheme intervention. The vocabulary test was given to 206 children in the morpheme group and to 121 children in the control group on all three testing occasions (N=327 for this analysis). The pseudoword-definition test was given to 170 children in the

128 What does the research tell us?

morpheme group and 104 in the control group on all three testing occasions (N=274 for this analysis).

The morpheme teaching intervention The program of instruction that we presented to the teachers was contained in a CD-ROM, which was the same as the one used for the morpheme-only groups described in the previous chapter. To remind you, the CD-ROM consisted of a number of tasks, all of which were presented as games with words. We organized these games in sessions for the previous study, but in this study the “sessions” were entirely notional ones because the teachers were free to pace the games in different ways. The ﬂexibility that we allowed in the teachers’ use of the teaching CD-ROM meant that there was a great deal of variation in the details of what the teachers did and how long they took to do it. One teacher reported using the games as rewards when the children had behaved well in the literacy hour. Others used the games as whole-class activities during the literacy hour. Only a few used the sessions as a whole once or twice a week.

The effects of the morpheme intervention on vocabulary Our hypothesis was simply that teaching children about morphemes should have the effect of improving their vocabulary. So, we predicted that the children in the morpheme group would do better than the others when their vocabulary was measured in posttests after controlling for the differences in the pretest. This was what we found. As in the previous chapters, we used statistical analyses that control for differences between groups at pretest when making the comparisons between the different groups at the two posttests. Figure 5.2 shows that all three groups made progress over time. It can be seen clearly from the graph that the morpheme group made signiﬁcantly more progress than the control group. This difference in progress was sustained in the delayed posttest and was statistically signiﬁcant. In Chapter 4, it was seen that for children in one year group, the morphology intervention was effective both for high and low achievers. In this study, we worked with different school years and fewer classes in each year group. As a result, it was not possible to conduct an exactly

Scores

Intervention program: Effects on vocabulary 129 22.5 22 21.5 21 20.5 20 19.5 19 18.5 18 17.5

Control group Morpheme group

Time 1

Time 2

Time 3

Figure 5.2 Mean scores (adjusted for pretest differences) in the vocabulary test for each testing occasion and group (maximum score = 40). Note: The difference between the groups was statistically significant (F=5.08; p<.02).

parallel analysis because the children in the group scoring below the median might be either younger than the others or low achievers. However, it is still possible to ask: Is the intervention as effective for children who start out with lower scores as it is for those who start out with higher scores, independently of whether they have low scores because they are low achievers or younger? We ran the same kind of analysis as we had done in Chapter 4: We divided the children into a group that scored below the median and a second group that had scores equal to or above the median in the pretest. The analysis was then run separately for these two groups. The results of these analyses are presented in Figure 5.3. Although there was a difference between the morpheme and control groups in both cases, this difference was statistically signiﬁcant only for the children who started out with low scores. Thus, our program may not have stretched the children who already had a good level of knowledge of morphemes or who already had better vocabularies at the outset. We think that this ﬁnding is actually quite interesting: Schools and teachers always find it difficult to design programs that are effective for those who have accomplished less, but the morpheme intervention is shown to be more effective in improving the vocabulary of exactly the children who start out with lower levels of performance.

130 What does the research tell us? (a)

(b)

18.5

27.5

18

27

17.5

26.5

17

26

16.5

25.5

16 15.5

25

15

24.5

14.5

24

14

23.5

13.5 Time 1

Time 2

Time 3

23 22.5

Control group

Time 1

Time 2

Time 3

Morpheme group

Figure 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. Note: The difference between the morpheme and control group was significant for the children with scores up to the median in the pretest but not for the children with scores above the median. For children who started out with low scores, the effect size was 0.25 SD.

It is particularly exciting to see in this ﬁgure that there are no signs of regression at delayed posttest; in fact, for the children who started out 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 measure of progress. The aim of this measure, as we pointed out earlier, is to test whether the children can analyze novel words in morphemes in order to have a way of discovering their meanings. Thus it is the critical test for us: Children who can use such word-attack strategies should be able to learn more new polymorphemic words than those who depend on some form of direct instruction about the meaning of each new word they encounter. Figure 5.4 shows the results of the analysis for all the children. Both the morpheme and the control group made some progress from pretest to posttest, suggesting an effect of practice in the test. However,

Intervention program: Effects on vocabulary 131 90 80 70 % correct

60 50 Control group

40

Morpheme group

30 20 10 0 Time 1

Time 2

Time 3

Figure 5.4 Percentage of correct pseudoword definitions (adjusted for pretest differences) by group and testing occasion. Note: The difference between the groups was statistically significant (F=17.5; p<0.0001); the effect size was between small and average (0.33 SD).

the control group did not progress any further from the immediate posttest 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 progress than those in the control group. This suggests that the children in the morpheme group learned a word-attack strategy and possibly used it on their own after the intervention had been concluded, because they seemed to become better at using this strategy at a time when they were receiving no further teaching about morphemes. We also ran an analysis similar to that presented for the vocabulary test, where we separated out the children on the basis of their pretest scores into two groups, one which scored up to the median and the second which scored above the median. The results of these analyses are presented in Figure 5.5. The analyses showed that the children who started out with a low score and also those who started out with a high score beneﬁted from the intervention. The differences between the morpheme and the control groups were signiﬁcant in both cases. There was a particularly large effect size for the children who started out with high scores, and this indicates that they beneﬁted from instruction more than those who started out with a low score.

132 What does the research tell us? (a)

(b)

70

92

65

90

60

88

55

86

50

84

45

82

40

80

35

78

30

76

25

74 72

20 Time 1

Time 2

Time 3

Time 1

Time 2

Time 3

Control group Morpheme group

Figure 5.5 Percentage correct in the pseudoword-definition test (adjusted for pretest differences) by group and testing occasion. Notes: The graph on the left shows the results for children who scored up to the median in the pretest; on the right are the results for children who scored above the median in the pretest. The differences between the morpheme and control group were significant in both analyses (p<.02). The effect size was 0.33 SD for the children who started out with low scores and 0.64 for the children who started out with high scores.

Summary and conclusions The overall conclusions regarding the effectiveness of the morphological intervention in terms of promoting pupils’ vocabulary growth are very clear. The morpheme intervention is effective and its effect can be shown to last over a period of 8 to 10 weeks. These effects are apparent not only when they are measured through the children’s performance on a vocabulary test, but also when they

Intervention program: Effects on vocabulary 133

are measured through the children’s performance on our acid test, the pseudoword-definition test. The importance of sustained gains shown in the pseudoword-definition task is that they show that the children learned a word-attack strategy that they can use to learn more polymorphemic words in the future. This takes us to the end of the work with children. This work shows, without a shade of doubt, that primary-school children benefit enormously from learning about grammar and morphemes. When they participated in our interventions, they learned not only how to spell and analyze specific words but also developed a more general form of knowledge, based on their awareness of morphemes as elements that form words, which helps them to understand the meaning of novel words and helps them predict the new words’ spelling. The implication is that developing children’s knowledge about morphemes and helping them to use this knowledge to develop their vocabulary and spelling should be an important part of the curriculum in primary school. In the two remaining chapters, we ask the question: Is it already the case that children are taught about morphemes in school? Are we proposing something that is already part of practice but for which there was, so far, no solid evidence? Or is it possible that teachers and teaching programs have not yet incorporated morphemes as units of meaning into their way of thinking about how to develop children’s literacy? Chapter 6 considers evidence directly from practice in London schools. Teachers were interviewed about how to help children learn the spelling of words whose spelling cannot be predicted from the way they sound, though it can be predicted from their constituent morphemes. Their teaching strategies in the classroom were also observed. Finally, they were made more aware of the importance of morphemes for literacy and had the opportunity to use our materials in the classroom. Chapter 7 analyzes the consequences of these interactions for primary-school teachers and their children. Chapter 7 contains an overview of how the teaching of spelling has been treated in the U.K. and in the U.S.A. and the evidence that has been provided for the effectiveness of the different ways of teaching spelling. Future developments are discussed in light of the research presented throughout this volume.

Chapter 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

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

So far, we have discussed the fact that morphology is one of the building blocks of English language and orthography. We have shown 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 with teachers as well as when they learn one-to-one with researchers. Now that we know that our techniques are appropriate for classroom teaching, we need to find out teachers’ views and practices regarding morphology. In this chapter, we explore what these views and practices are, to what extent they can be changed, and whether such 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 of awareness of the morphemic regularities of English. Taking this a little further, in the English context, it is also possible to explore the degree to which the teaching establishment explicitly identifies morphemic strategies as important aids to spelling. Since the English system of teaching literacy is centrally defined through the National Curriculum and the more detailed and practically oriented NLS, the place of morphology is reasonably transparent. Both of these sets of policy documentation mention morphemes. The National Curriculum promotes teaching morphemes in the context of spelling (Qualifications and Curriculum Authority, 1999). In the NLS documentation, morphemes are identified as one of the principles underpinning 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

Can we increase teachers’ awareness of morphology? 135

teacher guidance becomes more detailed and practical, the place of morphemes in the teacher’s 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 and theorizing in this area available to the authors of the model lesson plans.

Teachers’ strategies for teaching spelling We interviewed ﬁfty teachers from the London area to explore their views on spelling and, in particular, to see if they explicitly referred to morphemes in their teaching. To try to get at their working knowledge and practice, we asked them quite concretely about the difficulties that their pupils had with spelling and the ways in which they would address these difﬁculties. Teachers were told that we were interested in the sort of difﬁculties that 7- to 11-year-old children have when trying to spell and the ways teachers help them with their spelling. They were then presented with the following list of words, which illustrate a range of challenges for spellers: “white”, “opened”, “pavement”, “baseball”, “richness”, “motion”, “combination”, “slept”, “prepare”, “smoke”, “dark”, “uncovered”. For each word, they were asked what sort of errors their pupils would make and what they would teach the children to help them 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 which could be difﬁcult to “hear” unambiguously). Difﬁculty with letter blends and 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 pasttense endings (“slept”) and preﬁxes and other sufﬁxes. Overall, 1,930 statements were coded. Forty-ﬁve percent (n=878) of these statements referred directly to problems with phonology. This is not surprising, as we know that the dominant spelling strategy is a phonological one. The next most signiﬁcant category was morphology, which accounted for a further 20 percent (n=386) of teachers’ statements, and this was no

136 What does the research tell us?

doubt inﬂuenced by the fact that eight of the twelve prompt words in the interview were chosen for reasons relating to morphemes. More interesting is the way in which the teachers discussed the role of morphemes in spelling and teaching spelling. Despite our rigged interview, the word “morpheme” was conspicuous by its absence; it was not once mentioned by any of the teachers. However, teachers did refer to morphemes in other ways. The vast majority (82 percent, n=41) talked about preﬁxes and sufﬁxes. They also talked about past-tense verbs in response to the prompt words “opened”, “slept”, and “uncovered” (62 percent, n=31). When teachers talked 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).

BOX 6.1 Teachers talking about “-ed” endings “Opened” We wrote every word and its past participle on the same piece of card and then we sorted them according to how they’ve been turned into their past participle. They’re living in (four) houses; there’s the house when you just add “-ed” and it’s nice and simple; there’s the house where you don’t need to add an “-e” because it’s already there; there’s also the house where you have to double the consonant, as in “stopped” or “shopped”, so it’s double the consonant and then add “-ed”. And then on the other side of the track is the nasty house where things like “write” and “wrote” live and “bite” and “bit”. All the irregular ones live on the wrong side of the tracks. (Teacher 03) “Opened”, they wouldn’t think to put the “-ed” on the end, although we do do lots of work with past tense, but it’s drip, drip, drip. We have to constantly remind them. . . . So, for example, we’ve just been to the National Gallery and we wrote it as what we were going to do, so “Tomorrow we are going to go to . . . And then we went, so the next day, we wrote “On Monday

Can we increase teachers’ awareness of morphology? 137

we went to . . . ”, so we changed the whole, so it’s a concrete experience. (Teacher 11) This term we have been focusing specifically on suffixes to do with verbs particularly, so they know that the “-ed” becomes the past tense. It’s got better since we’ve done that actually. (Teacher 37)

Occasionally, teachers were apparently unaware of the signiﬁcance of the “-ed” sufﬁx and thus were unable to explain the spelling rule to their pupils (see Box 6.2).

BOX 6.2 Lack of awareness of “-ed” rule “Opened” I would probably talk to them about how it is spelled and how it is pronounced and how we don’t always pronounce all the letters that were in a word. There is a unit [in the NLS] that talks about unstressed vowels in words and makes a list of other words that have similar missing sounds and keeps a bank of those sorts of words on display, so that we say it but it’s got a letter in that is missing and then every time we need to use a word like that, don’t forget the missing letter, the letter we can’t hear, so just reminding them that it’s there. (Teacher 39)

“Slept” The ones that got it wrong would put a “d” on the end . . . because of the way it is pronounced . . . “pt” aren’t very common blends and can easily be pronounced with a “d” on the end instead of a “t” so, concentrating on the way a word looks and pronouncing the last letter. (Teacher 39)

138 What does the research tell us?

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 did talk about meaning, it was more likely to be in the context of a preﬁx like “un-” or “pre-” rather than in the context of derivational morphemes such as “-ness” or “-ion”, which are used to form words of a speciﬁc grammatical category. It would seem that English speakers tend to be very uneasy with the subtleties of grammar.

BOX 6.3 Teachers thinking about morphemes with connection to meaning “Pavement” Again similar to the one with “-ed” at the end I’d perhaps have “pay” and “payment”, “pave” and “pavement”, and all words with and without the suffix, where they make sense. Talk about the rule for changing it; talk about how we use one set of words, i.e. they’re verbs, how we use the other set and they’re nouns. And then actually model changing “develop” into “development” and then get them to do it themselves. (Teacher 03)

“Prepare” I always like to have the prefixes change the meaning of the word and then to identify the common ones so you know for “un-” you could say it makes the word opposite and looking at how it changes the word. (Teacher 36) It would be a whole class thing that we’d do. They’d all have different prefixes to work with and come back at the end and discuss how it’s changed the word and then just say that sometimes, like “prepare”, you’d have to find out . . . look the root of “prepare” anyway. They love things like, is it called

Can we increase teachers’ awareness of morphology? 139

etymology? They love finding out where these words come from, so it is worth it. (Teacher 20)

“Uncovered” Well, they’ve done lots of work, the entire class at different levels, on suffixes and prefixes, what goes before a word, what goes after a word, what does it mean and so on . . . We talked about if something comes before a word or after how it can change the meaning of the word, but we did quite a chunk on that. (Teacher 07)

The majority talked about morphemes in terms of letter strings or letter patterns (Box 6.4). Teachers observed that their children spelt “richness” with a single “s”, or “pavement” as “pavemint” or “pavemnt”, and, as we have seen in previous chapters, “-ion” words are particularly difficult. Here, teachers were addressing the fact that prefixes and suffixes are frequently occurring letter chunks, which are difficult to spell by relying entirely on their sound. Memory of visual patterns offers a viable spelling strategy. One teacher was clear that prefixes are usefully linked to meaning, but she thought that only her brightest pupils found this helpful (Teacher 27). Another teacher commented on the difﬁculty of identifying the root word in a word like “prepare”, a bit like the “gruntled” character in P. G. Wodehouse mentioned in Chapter 1.

BOX 6.4 Teachers thinking about morphemes without connection to meaning “Richness” 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’”. (Teacher 13) With that one, you could focus on the “-ness” ending. That’s quite a common string, suffix. (Teacher 19) 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, were trying to think up what it meant, but for the majority of them it’s just recognizing it and thinking I know that I’ve come across that pattern. (Teacher 27) Within the literacy hour there is focus on suffixes in Year 5 and in Year 3 and 4 so they would look at groups of words with the same suffix and I have some suffixes written up for display in my room as well so they can look for patterns. (Teacher 56)

“Prepare” And “pre-”, that would come under prefixes, “prevent”, “prepare” . . . But then “pare” is not a word on its own, or maybe it is. I’m getting a bit foggy now, there’s actually root words, do we take the root word away? If you took the prefix off, does this have to be a root word? I’d have to look it up. (Teacher 02)

“Pavement” Basically, that’s a suffix thing . . . Looking at the root word and adding the suffix (“-ment”). So it would be a lot of word cards, where they’re literally moving it into it and then writing it down . . . as soon as they recognize that that’s how the word looks, then that’s when they spell it correctly. (Teacher 04)

Can we increase teachers’ awareness of morphology? 141

Finally, when it came to “-ion”, teachers had no difficulty in recognizing that this was a real problem for children’s spelling, but only three teachers referred to suffixes in this context, and none of the teachers mentioned the meaning function of “-ion”, changing a verb to an abstract noun. There was a general unease about teaching “-ion”, and those who had a go at describing a strategy suggested either a visual one, involving rote learning, or analogy (Box 6.5). Some teachers mentioned the “-ion” and “-ian” confusion. As we have seen, there is a rule for differentiating the spelling of “-ion” (abstract nouns) and “-ian” (“person” words, or, as one colleague suggested, things that a Scottish lad called “Ian” could do), but this was not articulated.

BOX 6.5 Teachers talking about “-ion” “Combination” That’s just about learning that . . . the [middle] “t” can make the /sh/ sound. Again, it’s practicing it, learning words . . . grouping words that would have that kind of spelling, “-tion” words. We do quite a lot of work with that in literacy hour . . . We link it with “-cian”, you know, “magician” and stuff, and we look at the differences in groups of words. [Are there ways of identifying which type of spelling it would be?] There are, probably, but I can’t think what they are at the moment though. There probably is a rule, but unless I’m teaching it that day, I don’t learn it! (Teacher 05) I would probably do, either learn a group of words that end in “-tion”, or specifically learning “-tion” as a phrase and having it displayed in class or in the word bank. And it could be anything ridiculous, it could be “Tigers itch or not,” you know, it could be anything to help them remember it. (Teacher 09) Teach this as a chunk as they will know some through more familiar words like “station”. (Teacher 25)

142 What does the research tell us?

They just . . . separated them out and wrote down as many “-tion”s and “-sion”s and I think there was some rule, but I can’t remember, as a rule in my head, I’m not sure, I can’t remember it, I can’t think of anything that comes straight to my mind so it would just be, it is just the case of remembering it, probably better to rely on the visual than to . . . there’s no things from the sound. (Teacher 32) I tell the children, “Think of a word that sounds the same and do you know how to spell that? So how do you think that that will help to spell this?” It doesn’t always work but I think that it will give them an idea of where they should be going with it. (Teacher 48) Brainstorming work, similar endings, “-tion”, “-cian” because its exactly the same sound, it’s so confusing isn’t it? Some of the better spellers might put “-cian” but they would probably know “-tion”. That would be a smaller mistake, that would be the ones that almost got it and they’re just not familiar enough with “-tion” so there are quite different levels of mistakes. (Teacher 50)

It seems that teachers have explicit knowledge of some aspects of morphology but not others, and this explicit knowledge reﬂects both the context in which they teach (in this case under the mantle of the NLS) and aspects of morphology that are most transparent. We are arguing that this antimorphemism is a problem and that children would be spelling better if adults indulged in more explicit morphology. In fact, we have already presented some evidence to support our case. But can we change the teachers? If we could do that, we would be more conﬁdent that interventions would have a life outside a research project.

The teacher intervention We designed a ten-session literacy course for teachers, which emphasized two neglected but, we think, important dimensions of explicit instruction for 7- to 11-year-olds: Comprehension and morphology. Here we will tell you just about the morphology work. The course was

Can we increase teachers’ awareness of morphology? 143

delivered at university over one school term. There were three main aspects to the morphology part of the course: 1. 2. 3.

an introduction to theories about morphology and literacy; involvement of teachers in the intervention and research process; 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 spelling is given in Box 6.6, although this was elaborated in various ways and the teachers were also given readings. At the beginning of the course, teachers assessed their children on our word and pseudoword spelling tests (similar to those described in previous chapters) and agreed to assess them again at the end of the school term. We discussed the research design and teachers identiﬁed suitable control classes within their schools, so that we could tell whether any gains that their children made were above average. We gave each teacher the plans and material for a seven-session spelling intervention, similar to the one described in Chapter 3, and we asked them to try the intervention out in their classroom over the term. As the teachers taught the sessions, they were asked to reﬂect on their experience and to observe the way that their pupils responded. When teachers attended the course at university, they discussed how their children did in the tests, how the sessions were going, and how they would approach the next sessions. We examined samples of children’s writing and we discussed how this tied in with the theoretical side and the readings. The teachers heard from other teachers who had tried the same system, and they discussed approaches to teaching spelling between themselves. We wrestled with the practicalities of handing out information, teaching materials, etc. Throughout this process we attempted to integrate theory and practice.

BOX 6.6 Theories about morphology and spelling The definition of a morpheme is the smallest part of a word that carries meaning, so that can be a whole word, like “cat” or just part of a word like the “s” in “cats” or the suffix “less” in “careless”. We think that children at Key Stage 2 will be helped in both reading and

144 What does the research tell us?

spelling if they are explicitly taught about morphemes. Here’s why. Three broad stages of development have been observed in both reading and spelling: (1) the whole-word stage (logographic); (2) decoding letter by letter (the alphabetic stage); (3) recognizing chunks 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 letters that match their letter names R (are)

U (you)

LEFT (elephant)

2. Phonetic: The child can symbolize the entire sound structure of words in their spellings but the letters are assigned strictly on the basis of sound. Some sounds (for example, /r/, /n/, and /l/) are more difficult than others to represent in this way (Treiman and Cassar 1997). 3. Morphemic: The child becomes more aware of conventional spellings and employs visual and morphological information in spelling. For example, children learn to represent the /t/ sound at the end of past-tense regular verbs with “-ed” (Nunes et al. 1997b). Seven- and 8-year-old children should be entering the morphemic stage in spelling. We think that teaching them explicitly about morphemes at this stage is important and will help them to tackle long words in their reading, to understand word meanings of unfamiliar words and to help them to spell. Some tricky issues in spelling, for example, apostrophes, are only really successfully tackled through a discussion of morphemes.

How teachers changed Twenty-two teachers and three literacy advisors (literacy experts with responsibility for managing and supporting the teaching of literacy in local authorities) attended the course. We wanted to know whether engagement in this process made a difference to how the teachers thought and behaved. To begin the ﬁrst session, we asked them to write down and to discuss the strategies that they use in order to teach spelling. This served the dual function of allowing teachers to discuss

Can we increase teachers’ awareness of morphology? 145

the range of spelling strategies that they already used and allowing us to document their practice before intervention. Lots of methods were mentioned 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 hairy teacher” (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 percent mentioned 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 mentioned four times, and there was only one reference to morphemes. We also asked the teachers to write down their definition of a morpheme. Of the twenty teachers who completed this pre-course questionnaire, 25 percent knew that it was a small chunk of a word which had 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 their deﬁnition of a morpheme again. Of the seventeen teachers for whom we had data at the beginning and the end of the course, three had deﬁned a morpheme fairly accurately at the beginning but all but one produced perfect definitions by the end. All but one of the teachers reported that the course had changed their approaches to teaching spelling. As might be expected, most mentioned that they would teach more explicit morphology, making connections between spelling, grammar, and meaning. However, they changed in a number of other ways too. Several teachers mentioned that they would take spelling more seriously, for example, focusing one (1-hour) literacy session per week on spelling and taking a more structured approach to teaching spelling. They also increasingly saw spelling as having creative possibilities, such as class discussions and investigations. A number of teachers thought that using computer-generated materials had been

146 What does the research tell us?

good fun and effective and this had encouraged them to try this again in the future. Seven out of the seventeen teachers spontaneously mentioned that they would like to introduce the use of spelling journals. This was not something that we had suggested but something that the teachers had discussed themselves during the course. We had hoped that the course would be an interaction between us and the teachers but worried that in the end, the tremendous pressure of dealing with all the practicalities would squeeze this luxury out. It seems that some dialogue did go on between teachers despite everything.

How children changed The sample of children Seventeen of the teachers attending the course assessed their children at the beginning and end of the course (the “morphology children”). We also collected similar data from ﬁfteen classes where the teacher had not attended the course (the “control children”). The control classes were recruited from the schools of the teachers attending the course (n=8) and from classes taught by teachers who had attended a similar course focusing on numeracy (n=7) in the previous term. Table 6.1 shows the numbers of children in these two groups by year group.

Spelling assessment The children were assessed using two specially devised spelling tests. The first test was made up of thirty-two words, almost all of which contained one of the morphemes targeted by the intervention (spelling test). The second test comprised ten invented words, which also included the target morphemes (pseudoword-spelling test). Children could only spell the words in this second test correctly by applying the rules, as they had never seen the words before. The reasons for using Table 6.1 Number of children in each teaching condition, by year group Year group

Control

Morphology

Total

Year 3 Year 4 Year 5 Year 6 Total

112 57 128 56 353

110 132 88 30 360

222 189 216 86 713

Can we increase teachers’ awareness of morphology? 147

pseudoword-spelling tests in order to evaluate this teaching program have already been discussed in Chapters 3 and 4. The scores for each of these tests have been converted to percentages to make them easier to interpret.

Changes in children’s spelling Before the course, the children in the control classes were fairly evenly matched with the morphology children, except for Year 3, where the control group were substantially better (Table 6.2). Around 7 weeks later, all the children had improved, but the morphology group had made (statistically) significantly larger gains than the control group (overall, four times as much, see Table 6.3) in both spelling and pseudoword spelling. In the pseudoword test, the Table 6.2 Children’s average scores before the course, by teaching condition and year group, percentage correct Year group

Year 3 Year 4 Year 5 Year 6 Mean

Control children spelling test

Morphology children spelling test

Word

Pseudoword

Word

Pseudoword

62 65 74 72 68

50 54 58 62 56

51 63 71 74 62

38 40 52 62 44

Table 6.3 Average percentage increase in the children’s scores by the end of the course, by teaching condition and year group Year group

Year 3 Year 4 Year 5 Year 6 Mean

Control children spelling test

Morphology children spelling test

Word

Pseudoword

Word

Pseudoword

0 3 2 2 1.3

0 7 3 5 2.5

4 5 4 9 5.1

6 16 7 10 10.1

148 What does the research tell us?

morphology group had to apply an understanding of rules to the task; they couldn’t merely have learned the words through the course of the intervention. The only time they had seen these words before was in the pretest. In the following school year, one of the teachers who had attended the course organized a range of different spelling conditions in her school and assessed their impact on a new cohort of children. She involved three parallel Year 4 classes. Two of these classes were tested on their spelling in the fall term. Thirteen weeks later, in the spring term, they were retested and the third parallel class was also tested. Over the next 13 weeks, she used the morphology-intervention materials with her class in a special time slot (n=23, the morphology condition: The children had not used these materials previously). One parallel class spent the same dedicated time on spelling using materials from the NLS (n=26, the NLS condition). In a third parallel class, children followed the standard program (n=19, the standard condition). All three classes were retested in the summer term. Figure 6.1 gives a schema of the design used in this study. Between fall and spring, children made slight (nonsigniﬁcant) gains in their spelling. Between spring and summer, the morphology and NLS classes made noticeably greater (statistically signiﬁcant) improvements and the standard classroom group made slight (nonsigniﬁcant) gains.

One-year teacher follow-up

Thirteen weeks

Thirteen weeks

Standard program

One class, morphology intervention

Three Year 4 classes

One class, NLS intervention One class, standard program

Pretest

First posttest

Second posttest

mid-fall term

mid-spring term

mid-summer term

Figure 6.1 One-year teacher follow-up.

Can we increase teachers’ awareness of morphology? 149

60 50 40

Morphology NLS Standard

30 20 10 0 Pretest (fall)

First posttest (spring)

Second posttest (summer)

Figure 6.2 Children’s 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<0.001. NLS condition: Paired t-test, t = 2.4, df = 25, p<0.02. Statistically significant effects of the morphology condition: Compared with standard provision: B = 12.3, p<0.0001, effect size = 1.88 Compared with NLS condition: B = 9.73, p<0.002.

Teachers’ comments on the intervention process The morphology course seemed to work well overall, having an impact both on the teachers and their pupils. We asked teachers to keep diaries throughout and to comment on the intervention process at the end. Generally, they felt it had been a very worthwhile experience both for them 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 test their children again and to try to come to grips with some of the technicalities of the research process. Ideally, it would be preferable to spread the course over a longer period, though this may have resulted in less commitment. We had aspirations to include the teachers at all stages of the research process as well as the intervention process. The teachers were very sophisticated in their reasoning about the research. They were concerned to set up suitable comparison/control groups; they were conscious that each teacher used the morphology sessions in slightly different ways, some doing more, some less, some reinforcing learning

150 What does the research tell us?

in other parts of their literacy teaching, others not, etc. They were concerned that it was difficult to be sure that the improvements observed in their children were as a result of linking morphemes with meaning or merely due to increased exposure to letter patterns such as “-ion”. However, they did not have the technical skills to analyze the data by themselves, and this turned out to be too much to take on alongside 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 teachers and children approved of the lively, interactive PowerPoint presentation of the morphology materials. In particular, it was seen as motivating for children who were more difﬁcult to engage. However, the downsides were that sometimes children were too passive and that some teachers found the program hard to organize. The second major issue was differentiation. The teachers on the course had pupils spanning the age range 8–11. Some of the younger children struggled with grasping what was a noun (especially abstract nouns), a verb, and an adjective. They probably needed more practice and, perhaps, focusing on morphemes such as “un-”, “-less”, and “-ful” that don’t involve any fancy grammar. Some of the more able children found the tasks too easy and needed an extension of the program. This is always the tension with providing materials. On the one hand, they offer an efﬁcient vehicle to familiarize teachers with ideas and a practical manifestation of more abstract concepts relating to their teaching. On the other hand, they will always need some adaptation or selective use or extension to take account of individual differences.

Two case studies Teachers attending the course were given the option to gain credits toward a Masters course if they completed an assignment. Two teachers completed assignments on the morphology intervention. Here, one reﬂected that much of her knowledge of grammar was picked up rather than taught. “I learned the majority of my English grammar indirectly. I don’t consciously remember being taught present, past and future tenses in the same way as they were taught in my French classes. [French] was taught through a structure speciﬁc approach and informed my understanding of the English language.” (Teacher 55)

Can we increase teachers’ awareness of morphology? 151

They considered phonological awareness as “an essential foundation in the learning of reading and spelling,” but saw that teaching children about morphology had important beneﬁts for older children. Differentiation was one issue that was raised. One teacher commented that the morphology sessions worked particularly well for children for whom English is an additional language. “It was clear they beneﬁted from the direct teaching which enabled them to clarify misconceptions and underpin their existing (and sometimes shaky) knowledge of word classiﬁcations and changing meaning.” (Teacher 55) She felt it was less suitable for her children with special educational needs who “did not have a broad enough basic vocabulary to beneﬁt at the level the intervention had been set.” The other teacher found that a mixed-ability group worked well. “The teacher could use the more able children in the groups as experts and encourage them to explain their thinking and the ‘how’ part of the sessions. By this I mean that when the penny drops for children, they can have a unique way of explaining how they came to understand, which can often help the less able children to see the light. In the intervention group there was a particular child who is on a behavior strategy plan and who finds it difficult to interact with other children. This child, who was one of the more able children, found the sessions a good way to express his own knowledge and was excellent at giving the other children cues and clues as to what they needed to do.” (Teacher 46) This teacher commented particularly on the advantage of children being able 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 reﬂective period where the children could evaluate their own learning. It was interesting to note that with some of the rules, the children understood why they had previously made errors and some even tried to come up with their own memory cues. For example, one child suggested that we could remember ‘-ian’ as an ending

152 What does the research tell us?

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 likely ended in ‘-ian’.” (Teacher 46) These teachers’ ﬁnal reﬂections on the intervention are informative. “There is no doubt in my mind that an adaptation of this intervention would be beneﬁcial for works with further groups of pupils in the future. This adaptation would involve the use of shorter sessions regularly repeated over a longer period of time and reinforcement activities which would incorporate word classiﬁcation games and application within a range of written contexts.” (Teacher 55) “Although as a teacher I try to think of innovative ways to help children acquire spelling success, there are many administrative pitfalls, such as time and resources. In the literacy hour we do address spelling in the ‘word and sentence level’ section of the hour, but is 15 minutes a day really enough time to allow children to learn experientially and ﬁrmly concrete their knowledge? I guess some may say, ‘Yes, if you’re 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 teacher’s) as well as the time to then test their new knowledge. [Year 5] children were still excited to play word games. [. . .] I think that in Key Stage 2 we tend to lose sight at times of the value of games and the importance of experiential learning. The group’s results, as well as their refreshed attitudes toward spelling, speak for themselves.” (Teacher 46)

Summary and conclusions Although it can be seen from previous chapters that doing focused work with children improves their spelling, transforming research into teacher practice is a complicated business. In the research that is 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 morphology in teaching spelling. However, when we looked at teachers’ practice, reference to morphemes was limited and patchy. No teacher spontaneously mentioned the word “morpheme”, and, when asked, most teachers were unaware of its meaning. This suggests the absence of an explicit knowledge of the concept of the way morphology governs the spelling construction of English. Although teachers talked about aspects of morphology, most commonly in the context of verb endings and prefixes and suffixes, they normally focused on the visual patterns, failing to make a link between this and the meaning function. Observation in the classroom confirmed that children were rarely taught about the morphological dimension of our language. When teachers attended a ten-session course on morphology and comprehension at Key Stage 2, their practice changed, and this had a positive impact on their pupils’ spelling. The pupils of teachers attending the course made significant gains in spelling compared to children in similar classrooms receiving standard instruction, particularly impressive for a whole-class intervention delivered by teachers just learning a technique for the first time. The intervention is quite a focused and practical one, despite its conceptual base, and this probably contributed to its impact. Exactly what aspect of the intervention caused the change is less clear. We would like to say that it was due to teachers’ conceptual awareness of morphology. However, there are a number of other contenders. The classic alternative explanation is that the children did better just because they were being exposed to something new. This seems unlikely. The teachers of half of the control children were attending a mathematics course that was also exposing the pupils to novel practices. Another alternative explanation for children’s spelling gains is that teachers spent more classroom time teaching spelling. This was certainly an effect of the intervention. We had seen from the survey phase of our research that usually not a great deal of teaching time is dedicated to spelling in the standard classroom and that the teachers on our course commented that they were spending more time teaching spelling than they normally would. In the year following the course, one example clarified what aspect of our intervention was improving children’s spelling. Just improving

continued

154 What does the research tell us?

teachers’ explicit knowledge of morphology was not enough. In the fall term, a teacher who had been on the course did not have an opportunity to use the morphology materials. During this term, her new group of pupils made no greater gains in spelling than the other children in the school, despite the fact that this teacher did have explicit understanding of the role of morphology. Increasing the amount of time on spelling did not explain the gains entirely either. When her class was compared with a parallel class (the NLS condition) receiving the same amount of additional spelling instruction, her morphology group made significantly more progress than the children having additional NLS spelling activities. However, additional curriculum time was helpful. Both these classes made significant spelling gains compared to a control class and to their own progress in the previous term. The ingredients for change in pupils’ performance appear to be teacher knowledge and dedicated teacher time. We conclude that it is possible to transform the teaching of morphemes developed in a research context into teacher practice, and this in turn has an impact on children’s learning.

Note 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: 187–206. The website for Research Papers in Education is http://www. tandf.co.uk/journals

Part III

What are the overall implications?

Chapter 7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

Morphemes and literacy Context and conclusions

In this final chapter, we shall write about our findings in the context of 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 3–6 of this book), we already knew that there was a strong and important connection between children’s knowledge of morphemes and their understanding and use of the English spelling system. Our previous work and the work of several other researchers (described in Chapters 1 and 2) had led us to three main conclusions about this connection. We thought that these conclusions were clear and convincing enough for us to use them as starting points for an educational research program about how best to promote children’s morphemic knowledge in the classroom in order to help them to master the English spelling system and vocabulary. Our three starting points were: 1. Some of the most important correspondences between spoken and written language are at the level of the morpheme. The morphemic structure of words in English, and in several other orthographies, often determines their spelling. The system of morphemes, therefore, is a powerful resource for those learning literacy: The more schoolchildren know about morphemes, the more likely it is that they will learn about spelling principles based on morphemes. 2. Children’s knowledge of morphemes is largely implicit. The connection between this implicit knowledge and children’s

Authored by Terezinha Nunes and Peter Bryant

158 What are the overall implications?

literacy learning is not a strong one. However, there is a strong connection between explicit knowledge and literacy learning. Those children who do have a relatively strong and explicit level of awareness of morphemes and grammar at younger ages learn to use morphemes in spelling more systematically at a later age. This explicit knowledge helps them to use the spelling for suffixes correctly, such as the “-ed” for past verbs, and also to preserve the spelling of stems across words. The connection between morphemes and literacy is a two-way street: Awareness of morphemes strengthens children’s spelling and their knowledge of morphemic-spelling correspondences also promotes their awareness of morphemes at a later age. 3. It is possible to enhance children’s explicit knowledge of morphemes through systematic instruction and to improve their performance in word reading and spelling. These teaching interventions are effective for normal readers and poor readers who are seriously underperforming for their age and intellectual ability. As things stand at the moment in our educational system, one of the main reasons for children’s generally low level of explicit awareness of morphemes may be that they are taught very little about morphemes at school. All three of these statements have been amply supported by the results that we have reported here. The evidence that we have presented in this book has shown for the first time that it is possible to teach children in interesting ways about morphemes in the classroom and also that this learning has a good effect on children’s literacy development. All the studies that we have reported point in the same direction. They show that it is possible and also highly desirable to teach children about the connection between spelling and morphemes, and that the essential ingredient of this teaching must be the promotion of children’s explicit understanding of the morphemic structure of the words in their spoken vocabulary. We now wish to set these results in the context of contemporary educational theory and practice.

Morphemes and literacy: Context and conclusions 159

The path from the laboratory to the classroom When we began the research (Chapters 3–6) on improving schoolchildren’s morphemic knowledge, we had reasons for optimism but also for caution. We were optimistic because of the example of previous work on improving another form of explicit linguistic awareness in schoolchildren, which was their phonological awareness. In the past 25 years of the twentieth century, many research studies established that the awareness that children have of sounds before they reach school age predicts their progress in reading and spelling and that increasing their awareness of sounds has a positive effect on how well they learn to read and spell. Teachers and policymakers then made use of this knowledge about phonological awareness and, as a result, literacy teaching has now a much more solid foundation than before. With this example in mind, we felt that it was time for a similar effort to be made with children’s awareness of morphemes and grammar. We were cautious at the same time because we knew that the transformation of basic knowledge from research into classroom practice and educational policy is never simple. Edmund Henderson, one of the prominent ﬁgures in the study of the development of children’s spelling, warned about the seductive belief in a simple path from research ﬁndings to educational practice. He wrote, One of the best ways to learn about English spelling is to study the developmental stages that achieving children follow as they gradually master the system. Information of this kind tells what children learn about words and the order in which they learn it. Developmental studies do not, of course, tell how children learn these 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 one’s while to increase children’s explicit knowledge of morphemes and grammar. However, past practice had shown that teaching children about grammar and morphology can be boring and might therefore have little effect on literacy. In order to make the connection between research and practice, we needed to ﬁnd good ways to teach morphemic principles to children and see whether it is possible to do so in a way that maintains the children’s interest in the classroom. We made the transition from the laboratory to the classroom in six small steps, and this strategy was a deﬁnite success.

160 What are the overall implications?

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 a laboratory-style intervention. It included pretests and posttests, three different intervention groups and a control group, and two comparable intervention sessions for all the groups. The instructor–pupil ratio in the intervention sessions was small (1:2); some would say that it was ridiculously small. Nonetheless, the study showed that it is possible to increase children’s awareness of two sufﬁxes with identical sounds but quite different meanings and to improve their success in ﬁnding the right spelling for these sufﬁxes. It also showed that explicit methods are generally the most effective way of teaching children about this distinction.

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 the classroom where the instructor–pupil ratio is very much larger (around 1:30) and there are many distractions? The second step on our progress from laboratory to classroom was taken when our friends and colleagues in Lauriston Primary School set out to answer this question. The answer that came from their study (Chapter 3, Study 2) was that the two-session intervention definitely does have an effect in the classroom as well, though one that is weaker than in our laboratory study. We concluded that more than two intervention sessions might be needed in classroom studies.

Extending and broadening the classroom intervention Our third step, therefore, was to go back to the classroom and to set up another 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 and afﬁxes and apply different cognitive operations on morphemic units, they can use their knowledge of morphemes to bolster their word reading and the spelling of morphemic units (Nunes et al. 2003). So we

Morphemes and literacy: Context and conclusions 161

developed an intervention program to be tested in the classroom. The intervention was run by classroom teachers, but we prescribed quite tightly what the teacher should do, and a member of our research team was always at hand during the intervention sessions to advise the teacher if advice was needed. Again, we found that the teaching worked. The children who had been taught about morphemes and spelling fared better in a morpheme-spelling task than children in a control group. Thus, we had successfully made the step from laboratory to the classroom, but the teaching in the classroom projects had been rather tightly constrained. The methods and the limited scope of the first classroom study were very similar indeed to those of our laboratory study. In the second classroom study, the teacher followed a set program very closely and a member of our research team was always present in the intervention. We were acutely aware we would never have such control 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 adopted and applied our program with no more than a small amount of preliminary advice and the use of a CD-ROM which contained our different tasks or games in PowerPoint. If our ideas are to be used in the classroom, we would never have any more control over the teaching than 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 the aims of the project. Our only constraint was that the teacher should administer the different tasks in the order in which they appeared on the CD-ROM. This project also produced strong effects of teaching about morphemes on the children’s spelling, as measured by pretests and posttests. Thus, our methods also worked when the tight constraints of our earlier interventions were no longer there. Our methods work in “real” classrooms.

Teaching vocabulary as well as spelling Schools are at least as likely as university departments to produce good research questions about teaching. Now that we were working in classrooms, we thought a great deal about questions raised by the teachers themselves. The teachers in a group of schools that we worked

162 What are the overall implications?

with in London were concerned about their pupils’ vocabulary and wanted to ﬁnd ways of improving it. Our view was that helping children to understand how words are constructed from morphemes should be an effective way of promoting their vocabulary for reasons that we set out in Chapter 1, and so our ﬁfth step was to use our intervention tasks in a training program designed to improve children’s vocabulary. The answer to this question, generated by the schools themselves, was again positive. The program, again administered with minimal constraints imposed by ourselves, did improve the children’s knowledge and 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 have to be part of teacher training too. People training to be teachers and people who are teachers already would have to learn about the kind of tasks that we have developed. It would be essential, also, to increase their awareness and understanding of morphemes. Thus our sixth step (Chapter 6) was to institute a course for teachers about morphemes and to study the effect of this course on the teachers and on their pupils too. The study showed that the course did radically increase the teachers’ awareness of morphology and its links with spelling and that their pupils’ spelling also beneﬁted, provided that the teachers used some of our tasks to teach the children about morphemic spelling principles. Our journey from the laboratory to the classroom and to courses for teachers was intricate but productive. It is right to be cautious about moving from one environment to the other, but we believe that we have shown that it is possible and also profitable to make the transition properly. We now turn to the context of ideas about teaching spelling in order to examine in what ways these could change under the impact of our results.

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 classroom. Our next question is about educational theory and practice. Do

Morphemes and literacy: Context and conclusions 163

educators think that children ought to be, and need to be, taught spelling? If they need to be taught, what should they be taught and how should they be taught it? As we shall see, views on what and how children should be taught have changed over time. Our answer, to a large extent, is a new one. The question whether spelling is caught or taught was posed by Margaret Peters (1985) some time ago. Most authors then seemed to agree that there are basically only two alternative possibilities about how children learn spelling: Either they catch spelling, which means that they learn it implicitly as they read, or they have to be systematically taught how to spell. Those who thought that spelling was simply caught did not feel that anything had to be done to teach children to spell. But, according to Margaret Peters, it was only possible for some children to “catch spelling” and only under special circumstances. She argued that parents and teachers have to direct children’s attention to spelling patterns for this “catching” to happen. This guidance should emphasize the probabilities of some letters occurring in sequence, thereby facilitating visual learning of words. In fact, there is little evidence that children actually can catch spelling without any teaching. It is likely that everyone receives some teaching, as even the most radical believers in the possibility of children catching spelling will teach them some spelling. It is also true that many university students continue to ﬁnd English spelling a challenge, perhaps even throughout life. Many confess to changing their choice of words when writing, due to their uncertainty about how to spell words that they had intended to use in their texts. It is also quite likely that there is much implicit learning of spelling but, again, there is little evidence for how much and what is learned implicitly. Over time, the hope that children would simply catch spelling seems to have waned. Teachers and policymakers have turned to what to 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 to memorize the problem words—or, as they were often called, “the spelling demons.” Now, each child would fall under the spell of different demons, and there is no point in protecting one child against someone else’s enemy, so each child had to try to memorize the words that he

164 What are the overall implications?

or she found difﬁcult. Therefore, lists of words to be learned would have to be compiled. This was the ﬁrst step required to transform these ideas about spelling into practice. What should children be taught? The answer was the spelling of words. There was some disagreement among those who advocated the teaching of spelling about how the lists of words should be compiled. Peters, for example, favored the idea that the lists should be based on the child’s own writing: Children should learn those words that they misspelled when writing their own texts. This form of instruction would be individualized, and it was believed that it would be of greater use and more motivating to individual learners: They would be learning the spelling of the words they had wanted to use in their own writing. In contrast, Fred Schonell in 1932, another pioneer of the study of spelling, thought that standard word lists could be used (Schonell 1957). He compiled the frequencies of words from material written by children aged 7–12 and recommended that children should work their way through these lists, learning the spelling of words from more to less frequent ones. This, he argued, should make it easier for children to spell words that they would be using all the time, would increase their chance of success in spelling in early grades, would provide them with a basis to enjoy writing, and would result in an increase of their written vocabulary 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 compile lists of words whose spellings children had to learn. In order to create individualized lists, following the views put forward by Peters, teachers would simply have to ask the children to write texts with some frequency, correct them, single out the words to be learned, and ask the children to learn them. The construction of individualized lists was attainable by teachers for their pupils. Similarly, it was possible for teachers to draw on Schonell’s lists and to use these in the teaching of spelling, because he divided the words into six groups, each representing a chronological age, which would give teachers a starting point for children at particular age levels. Schonell also indicated that teachers could replace some words with others that might be more appropriate for their pupils. Both Peters and Schonell recommended frequent testing and targeted learning, so the spelling lists, whether individualized or standardized, were used to assess the children’s level of spelling and to determine their learning program. There is, however, a further question to answer for anyone taking this approach to the teaching of spelling. What should children do to

Morphemes and literacy: Context and conclusions 165

learn the spelling of the words they were presented with? Once teachers chose words for the children to learn, how could they best help the children 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 and note what in its spelling might be difﬁcult, cover the word with a piece of paper, write the word in their own books, and then check that the spelling was correct by matching it against the teacher’s spelling. This became known as the “look–cover–write–check method.” If the word was misspelled, the child should rub it out, write it correctly, and then repeat the whole process. The aim of this method is to form visual memories and motor habits for each word; it does not attribute importance to correspondences between letters and sounds in this memorization process because the “spelling demons” were exactly those words whose spelling could not be predicted from the way the word sounded. Saying the word, it was thought, might actually interfere with learning spellings which were not predictable from phonology. Others, in contrast, advocated multisensory methods. Maria Montessori (1915) was one of the ﬁrst educators to suggest that it is important for children to establish multiple connections between the sounds, visual shapes, and kinetic representations of letters and words. Although Montessori’s method was developed in the phonic tradition, multisensory ideas were also used in the whole word tradition, perhaps most notably by Grace Fernald (1943). She suggested that the children should think of the word they wanted to learn to spell, the teacher would then write the word on a card (usually in the context of a sentence), and then the child would trace the word with a ﬁnger, saying each syllable as it was traced. The process would be repeated until the child was able to write the word from memory. This very brief description allows us to identify the underlying assumptions to these approaches to the teaching of spelling. Their answer to the question “What should children learn?” is “whole words”. Their answer to the question “How should children learn?” is “by memorization”. 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 be learned from children’s mistakes. The form of memory to be engaged in the teaching of spelling is “brute memory,” as Henderson (1990) called it. There is no place for understanding or for the creative role of children in generating spellings. We shall discuss the idea that children can generate spellings rather than memorize each one in the next section.

166 What are the overall implications?

Children’s understanding of English spelling Ideas have moved back and forth in educational theory, but one theoretical change, which we think is unlikely to revert to past views, is the way in which children’s mistakes are conceptualized. We think that this is one of the many positive impacts of Piagetian theory on education. Children’s mistakes, which were treated in the past as responses to be eliminated, are now considered as valuable pieces of information about how children think. The linguist and former teacher Charles Read provided the breakthrough for this view in the domain of spelling in 1971, when he showed that young children were able to invent spellings for words that no one had 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 children’s writings revealed that their analysis of sounds was not exactly the same as the analysis a literate adult would produce, there was no question that the children’s spellings were based on what is now termed an alphabetic or phonetic conception of writing. Children created spellings from their understanding of the spelling system, they did not just reproduce (correctly or incorrectly) memorized spellings. Read’s findings have since been confirmed by many researchers working in English (for example, Ehri 1997, Treiman 1993) as well as other languages (for example, 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 children’s alphabetic conception or by seeking ways of describing progress in children’s conception of spelling beyond the alphabetic stage. There are controversies about what is the best way to describe the development of children’s conceptions of English orthography, but these do not need to be discussed here. Our focus is actually on the convergence that can be seen across many researchers toward the idea that spelling development 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 spelling development is characterized by the acquisition of correct spellings for particular words, one stands out: This is the ﬁnding that children may actually appear to learn to spell words worse than before as they grow older, if their progress is judged on a word-by-word criterion, because as they grow older they make spelling mistakes with words that they used to spell correctly before. We have found clear evidence for this

Morphemes and literacy: Context and conclusions 167

apparent regression in our studies of children’s acquisition of the “-ed” ending. When children ﬁrst start to use “-ed” at the end of words, they put it at the end of the right as well as of the wrong words. Because our study was longitudinal, we were able to show that the same children would make the mistake of using “-ed” on words that they had spelled phonetically and correctly earlier on. They had discovered a new fact —that some words ending in /t/ and /d/ sounds are spelled with “-ed”—and had started to use this knowledge, but they had not quite understood the principle that the “-ed” ending is reserved for regular verbs in the past. Results like these suggest that we should no longer think that children learn spelling by adding one word at a time to their memorized list. It is to some extent irrelevant whether spelling principles are acquired suddenly or in small steps and whether they can be learned as rules that are taught or must be learned from experiences with words. It now seems certain that children’s understanding of the relations between the spellings of different words plays an important role in the development of spelling. Henderson (1990) suggested that this new view of the development of spelling does not refute the notion that memory plays a role in learning spelling: The question is how memory should be used in teaching and learning spelling. According to him, there is good reason to doubt that poor spellers can mend their ways by even the most heroic efforts to memorize the spellings of all the words they need. Nevertheless, he agrees that they can start from learning groups of words that illustrate principles and increase their ability to use these principles through 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 written words. So, if we want to engage children’s understanding of principles along with their memory efforts when we teach spelling, we must choose well what they study. When teachers and researchers came to view children’s understanding of how written language works as an important part of learning to spell, they came up with a new answer to the question “What should children learn?”: They should learn relations between the spelling of different words. They also had a new answer to the question “How should children learn?”, which was that they should learn through understanding, not just through memory traces.

168 What are the overall implications?

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 more phonemes in English than letters of the alphabet to represent them with. So inconsistency in letter–sound correspondences is inevitable. But there is also more inconsistency than strictly necessary: Some words sound exactly the same and yet they are spelled differently. To confuse matters further, the same word might be pronounced differently by different speakers, although they all have English as their ﬁrst language. Where does one begin to put some order into English spelling so that children can use their understanding when studying written English words? One possibility that has been suggested by many is that phonological and graphic units larger than letters and phonemes should be used in teaching. It has been suggested that teaching children the regularities that can be observed by dividing words into onset and rime (that is, dividing them into two parts, the ﬁrst which contains its beginning and goes up to the ﬁrst vowel, and the second which starts with the vowel and includes what follows it) is a good way to deal with inconsistency at the alphabetic level. Rebecca Treiman and her colleagues (Treiman et al. 1995) argue that the level of sound to spelling consistency at the level of rimes is much higher than at other phonological levels, such as phonemes. Others (for example, Cardoso-Martins 1994) have suggested the use of syllables rather than of rimes or phonemes as units in spelling. The use of syllables as spelling units is discussed less often in English than in other languages, such as Portuguese and Spanish, possibly because children have difﬁculty in identifying syllable boundaries in English. The syllable is undoubtedly a different unit of analysis that could be used in the search of consistencies in spelling. However, it should be pointed out that there are so many monosyllabic words in English that this solution might not be very practical: Using the syllable as a unit of analysis for so many words could turn out to be carrying out no analysis at all. These two ideas differ insofar as they focus on the use of different units of analysis, but they have one thing in common, the idea that spelling can be taught through the use of different phonological units. A different approach is suggested by those who think, not in terms of phonological units but in terms of the acquisition of different principles to be used in conjunction with the alphabetic principle. This is a position that was adopted by teachers such as Edmund Henderson (1990) and

Morphemes and literacy: Context and conclusions 169

linguists such as Richard Venezky (1970, 1995); it is also our own position. 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 spelling word-pairs such as “hat”–“hate”, “mad”–“made”, “tap”–“tape”, “plan”–“plane”, “hop”–“hope”, “rot”–“rote”, “ton”–“tone”, “pin”–“pine”, “sit”–“site”, “tub”–“tube”, “us”–“use”, etc. In contrast, if the child is learning to use principles, rules, or generalizations in spelling, all of these pairs can be summarized under one principle. This would reduce the number of specific facts to be learned and would provide the children with many connected experiences that would help them understand the relevance of one spelling principle. Rimes might be a way of assembling a list of words that exemplify a principle, but we suggest that what children should be learning is not single rimes as if they were spelling responses: They should be learning principles. Our research has shown that children learn better if they are exposed to the contrast between the pairs than to each set of instances of rimes (Nunes and Bryant 2006). The exposure to the contrasts helps children learn the principles. Richard Venezky argued that the regularities of English spelling could be understood better if, instead of thinking of the relations between single letters and single phonemes, we thought of grapheme–phoneme correspondences, where the graphemes might be composed by more than one letter and these letters might not actually be in sequence. This approach is very useful for the description of regularities in the English orthography but less useful for an account of the development of children’s spelling. It seems to us that children take a major step when they start to think that there are different types of regularities in English and shift from using almost exclusively the alphabetic principle to combining it with other types of regularities. We will refer to the regularities that go beyond single letter–single phoneme correspondences as higher-order spelling principles, to differentiate them from the alphabetic principle. There has been a great deal of discussion about whether such spelling regularities should be called rules, generalizations, or principles. When teachers use these regularities, they are trying to help children understand that, under certain conditions, regular letter–sound relations are replaced by other regular and predictable relations between groups of letters and sounds. There is agreement regarding what to teach. Disagreements are actually about how to teach: Should

170 What are the overall implications?

children be taught these regularities as rules or should they learn them from experience? We argue that there are two aspects to higher-order spelling principles: These are their form and function. We will argue that the learning of higher-order spelling principles cannot be simply the learning of a visual pattern. In the case of the higher-order phonological principles that we have brieﬂy exempliﬁed in this section, each form has a phonological function. Most researchers seem to agree that the development of children’s spelling shows the acquisition of higher-order principles to be a later achievement than the discovery of the alphabetic principles. There are two important sources of evidence for the learning of higher-order spelling principles: The appropriate use of these principles in pseudowords 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 children’s reading and spelling of pseudowords changes from relying on the alphabetic principle to drawing also on higher-order spelling principles. Higher-order principles are, in fact, unavoidable in English because, as pointed out earlier on, there are not as many letters as 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, for example the vowel sound in the word “goat”, we can test whether the child is using a higher-order principle in spelling. Because pseudowords do not exist in the language, some argue that there is no correct spelling for them. This is so, but only to a certain extent. If you were asked to spell 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 this sound correctly if you tried to use a one-to-one correspondence between letters and sounds. If you spelled it as “flob”, your spelling would have to be considered incorrect, because the vowel sound would have been incorrectly represented. Because pseudowords cannot have been encountered before, as they are not part of the language, a child that adequately represents this vowel sound in a pseudoword is using knowledge of regularities in spelling that go beyond letter–sound relations. The child cannot be using the specific memory of a word learned previously, whether this memory was “visual” or “kinetic”. George Marsh pioneered the use of pseudowords in the investigation of children’s use of higher-order principles in spelling and showed that younger children inadequately used simple letter–sound corre-

Morphemes and literacy: Context and conclusions 171

spondences to spell all pseudowords, whereas older children would adequately use these higher-order spelling principles and differentiate between pseudowords such as “ﬂob” and “ﬂobe”. His observations have been conﬁrmed many times since (see, for example, Templeton and Bear 1992). Children’s errors in the use of higher-order principles also tell us about their understanding of the insufﬁciency of the alphabetic principle in English orthography. Our view is that in order to use the digraphs “oa” or “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 is also necessary to connect them to a speciﬁc vowel sound, that is, to learn their function. This is no small accomplishment because there are so many digraphs in English. Thus, it is quite possible that children learn a new pattern without being completely clear about what its function in spelling is. Figure 7.1 shows the writing of a 6-year-old boy who seems to attribute to the digraph “ck” the function of the split digraph “V+C+e” (vowel plus consonant plus “e”). He is certainly using it creatively in the sense that he has generated novel spellings for the words “like”, “take” and “make”. He is also using this pattern consistently, with a speciﬁc function, and is generalizing this across the words. He has learned the form but has misunderstood its function. Because form and function are tightly connected in spelling, learning the form is not sufﬁcient. To summarize, we have argued that analyses of the development of children’s spelling reveal that they go beyond the alphabetic principle by learning higher-order principles, which are generalizations, rather than 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

Figure 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”.

172 What are the overall implications?

that children have to master in the acquisition of higher-order spelling principles: The new forms and their functions. Learning the forms only (visual learning only) does not suffice for mastery. Form and function are essential for the mastery of higher-order spelling principles. Children are also able to use these generalizations in spelling pseudowords, thereby showing that their learning goes beyond word-speciﬁc learning.

Higher-order spelling principles in teaching spelling We have argued that the ideas about what children need to learn have shifted, from words to relations between the spellings of different words. But deciding that children should learn higher-order spelling principles instead of lists of words does not solve the question, “How can we best help children to learn these relations between the spellings of different words”? One approach that seems, even if implicitly, to be favored by many is the idea of visual learning. Many references to teaching “letter strings” and attention to “visual patterns” are common in books about the teaching of spelling (for example, Mudd 1994: 148). Some authors go on to recommend that the children say the words whereas others are adamant that the children should simply try to learn the visual patterns because of the inconsistencies of English. The use of letter strings without attention to the function of higher-order forms can lead to problems: Words that have the same letter strings for different reasons can 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 this procedure, a teacher compiled the word list “rough, cough, thought”— a list that does not help the children to understand the possible function of “-gh” at the end of words as opposed to “-ght”. However, it has become increasingly common since the work of Henderson, Venezky and their colleagues to use form and function together in the study of words. We carried out a literature search to ﬁnd out whether there is support for the idea that children should learn “visual patterns” without learning their phonological function. We could ﬁnd no empirical studies where a “visual without phonology” condition was included in teaching. So, the claim that children should learn forms visually without learning their function seems not to be widely accepted, at least by educational researchers.

Morphemes and literacy: Context and conclusions 173

What differences are there between the methods proposed to get children to learn more about higher-order spelling principles? Many seem to have a complete horror of the idea of teaching children rules. Henderson argued that “it is not rules that children need but experiences. 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 should assemble for study those words that exemplify patterns in order to provide children with experiences about what these patterns mean. Although Henderson spoke of spelling patterns, he insisted that these would have meaning only if their function was understood. Henderson wrote extensively about what children should study, but he gives no clear indication for how they would be able to learn, beyond the notion that they should be given experience. Following on Henderson’s work, Jerry Zutell (1993) proposed the Directed Spelling Thinking Activity method, where children are presented with contrasts rather than patterns and encouraged to analyze the phonological differences between the contrasting spelling patterns. Do we know whether it is better to learn patterns separately or to learn the contrast between them? Our own research (Nunes and Bryant 2006) has shown that children learn higher-order principles better if they are exposed to the contrasts than if they analyze each spelling pattern separately. This was true even when they have the same amount of experience with each pattern during 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 had the opportunity to study the contrast between the two types of spelling patterns. The possibility of establishing relations between different spelling patterns and their functions helped retention over a 10-week period. The importance of establishing connections between forms and functions at a more abstract level is supported even further by another study that we conducted on the use of different sequences for the teaching of spelling principles. Our collaborators from Lauriston Primary School taught some children the “hat”–“hate” contrast on one day and the “hop”–“hope” contrast on a subsequent day. They taught another group of children the “hat”–“hate” contrast on one day and on the subsequent day the children learned the distinction between using “-ck” and “-k” at the end of words (for example, “pick”–“pink”, “lock”–“honk”,

174 What are the overall implications?

“pack”–“park”). The ﬁrst group of children learned two contrasts that could be brought under the same higher-order spelling principle, the “V+C” versus “V+C+e” distinction; the second group learned unrelated principles. The children who learned two related principles profited from establishing this relation between the two patterns: They spelled words with “a+C” and “a+C+e” better than the children in the second group, although they had the same amount of experience with the “hat”–“hate” contrast. This is, in fact, our reason for referring to these generalizations as higher-order spelling principles: It is possible to categorize them in more abstract terms and to use these abstract categories in our understanding. So, we seem to know something about how to help children to learn spelling principles. •

•

It is better to help them to understand the functions of these different forms than to let them learn the forms without a focus on their functions. Children learn more from establishing relations between different forms with similar functions than from learning unrelated sets of forms and functions.

Higher-order spelling principles based on meaning The higher-order spelling principles that we have discussed up to this point in our analysis of what children need to learn are connected to phonology. Phonology is important for spelling and it is not surprising that most of the research on spelling and most of the guides to teaching spelling focus on phonological spelling principles. But, as the research that we present in the previous chapters shows, these principles do not tell the whole story of English orthography. There is a connection between morphemes and English orthography, and this connection can be used by children in learning to spell. The idea of using morphemes and meaning in teaching spelling has received much less attention than higher-order phonological rules— and, when it did, we think that there was something missing. Among the pioneers of the idea that systematic instruction on morphemes could benefit spelling are three U.S. educators—Dixon, Henderson, and Henry—whose ideas have been used in the teaching of spelling and have been subjected to some empirical test. Dixon (1979) developed an extensive program for teaching spelling

Morphemes and literacy: Context and conclusions 175

using morphemes, Corrective Spelling through Morphographs, which included 140 lessons to be taught over a period of 8 months. The term “morphograph” was created and used to refer to the smallest unit of meaning identiﬁable in written English, perhaps shifting the focus from the commonalities between oral and written language to an exclusive focus on spelling. Morphographs include stems and afﬁxes. The focus of the lessons in Dixon’s program is on teaching rules for analyzing words into morphographs, learning the meaning of the different affixes, and learning rules for combining morphographs in order to spell words correctly. The program was evaluated by Robinson and Hesse (1981) using measures that were developed to assess whether it attained its aim to teach students how to analyze words into morphographs and whether this improved the students’ spelling of polymorphemic words. A general spelling test was also included. The comparison between the taught and control groups, which included a total of 172 students in their seventh year in school (approximately 13–14 years of age), showed that the intervention students scored much better in the tests of rule application and morphographic analysis. They also scored better in the spelling test that contained words selected by the authors because they were formed with the morphographs that the students had learned. However, the beneﬁt of learning these rules and morphographic analysis did not translate into signiﬁcant differences between the intervention and control students in a general spelling achievement test (the Stanford Achievement Test) although more than half of the words in this standardized test were either part of the teaching program that had been followed by the students or contained morphographs that had been covered in the program. After 140 lessons and 8 months of work, this can hardly be considered an encouraging result. We think that this program is an example of teaching that includes a good idea about what children need to learn but it falls into difﬁculty when it comes to the crucial question of how children should learn. Research on expertise in different domains of psychology has shown that it is possible to learn rules that describe a skilled person’s behavior without beneﬁting from this knowledge when implementing the skill (see, for example, Anderson 1981 for a compilation of papers on the development of expertise). Henderson’s approach on what to teach was quite similar to Dixon’s, 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

176 What are the overall implications?

stems have the same pronunciation and different meanings, they are usually spelled differently, but stems with the same meaning and different pronunciations are typically spelled in the same way. The ﬁrst part of his rule refers to homophones (different words that sound the same), such as “weak” and “week”, “maid” and “made”. The second part refers to stems and derived forms, such as “heal” and “health”, “magic” and “magician”. Henderson focused many of his suggestions for the teaching of spelling on getting children to understand the constancy of the root. This is quite a complex process because sometimes the pronunciation of the stem changes and nothing is done to its spelling in order to maintain the pronunciation, whereas sometimes letters are doubled or deleted in order to preserve the pronunciation. For example, when we add “-ed” to “stop”, we also add a “p” to maintain the pronunciation of the vowel: We would read “stoped” differently from the way that we read “stopped”. However, if we add “-ed” to verbs ending in “e”, we delete the ﬁnal “e” from the stem in order to preserve the pronunciation (as in “deleted”), because the double “e” would move the stress to the sufﬁx (think how you would read “deleteed”), and this sufﬁx is never stressed. Henderson paid much less attention to preﬁxes and sufﬁxes than to stems, probably because he considered afﬁxes to be relatively unproblematic. The exception was the “-ed” for past regular verbs. He did describe a progression in children’s use of the “-ed” ending, which has been conﬁrmed by many since, and which we have ourselves observed: children seem not to use the “-ed” ending much at ﬁrst, then they use it as a form without knowing its correct function, and later they restrict it to the appropriate places (see Chapters 1 and 2). But Henderson did not consider that it might be necessary to increase children’s awareness of morphology and grammar in order for them to master the use of the pattern with meaning. He seemed to take for granted that children’s implicit knowledge of morphology and grammar would sufﬁce for them to know when and where to use the “-ed” ending and focused on what happened to the spelling of the root. He did suggest an interesting use for sufﬁxes in raising children’s awareness of their own grammatical knowledge. The activity that he proposed was to ask children to place “-ed” or “-s” at the end of words and then discuss why these endings might not ﬁt with certain words: For example, you can add “s” to “maid” but not to “made” (his example, which we think illustrates that his focus was on the stem; you can’t add either “-s” or “-ed” to “made” because it is already a verb in the past).

Morphemes and literacy: Context and conclusions 177

Henry (1989) also made a direct contribution to the idea that morphology is important for literacy teaching. She argued for teaching children, particularly those with reading problems, lists of preﬁxes and sufﬁxes so that they could be treated as units in word recognition and in spelling. She also focused on the etymology of morphemes, arguing that there are common spelling difficulties in morphemes that have the same origin: for example, Greek morphemes may have a silent “p” at the start, as in “pneumonia” and “psychology”. When she refers to sufﬁxes, she treats them more as units of pronunciation than as units of meaning. For example, she refers to “-tion” and “-cian” as sufﬁxes, and thus she focuses on the syllables formed by the sufﬁx and the ﬁnal letter in the stem. She does not discuss the function of these sufﬁxes. This past work offers a starting point for the idea of using morphology in literacy teaching. However, there are many issues that require further analysis, as our research has shown. The ﬁrst weakness of the proposals that we have just reviewed is that they take children’s awareness of morphology for granted. In order to implement teaching that relies on children’s awareness of morphemes, we must be able to also promote it. This lesson was learned about children’s awareness of phonology and should be also learned now about children’s awareness of morphemes and grammar. Implicit knowledge of morphemes and grammar is not a solid enough basis for the teaching of spelling. Our own work on the use of apostrophes, a silent morpheme whose use is entirely determined by grammar, shows how difﬁcult it is for children to use an apostrophe in the absence of explicit awareness of its function, and how much progress children can make by becoming aware of grammar and the function of apostrophes to indicate possession (as in “the man’s house”) (Bryant et al. 2000, Bryant et al. 1997). A second weakness of these proposals is that the suggestions are about what to teach—lists of stems, preﬁxes, and sufﬁxes are offered— but there is a scarcity of research on how to teach these things. This could be the result of taking children’s awareness of morphemes for granted. But we have witnessed how the treatment of prefixes and sufﬁxes as “letter strings” without reference to meaning can hinder, and not help, children’s efforts at understanding language. Teaching the preﬁx “un-” has become part of the curriculum in the U.K., but this is not necessarily done in the context of raising children’s awareness of morphemes. We have seen some children, supporters of the football team Manchester United, struggle when asked to read the word “united” by itself on a page. They often wrongly treat the “un-” as a preﬁx and

178 What are the overall implications?

then cannot ﬁgure out what to do with the rest of the word. We think that more explicit knowledge of morphemes would have been helpful to these children; the isolated knowledge of “un-” as a preﬁx was more of an obstacle. A third, and very signiﬁcant, weakness is that the evidence for the effectiveness of these practices was weak, and sometimes there was none at all. In education, as in many other applied settings, we might have knowledge of what works, but not why; or we might have knowledge that leads us to expect that something will work and why, but don’t know whether it will actually work in the classroom. In the case of the connection between morphemes and literacy, there were programs that could be used in the classroom, but there was no evidence that they worked. There were also suggestions from linguistic and psychological studies that teaching children about morphemes could be a good thing, but the path from the laboratory to the classroom had not been traced. Finally, the outcome measures used in past research were limited. If children perform better when spelling words that they were taught, we don’t know whether they learned the specific words or whether they learned principles. A program for developing children’s use of morphology should provide evidence for the use of morphological principles, not simply for learning speciﬁc words.

The contribution of our own project Our own research program has changed much of this picture in signiﬁcant ways. First, we did not take children’s explicit awareness of morphemes for granted. Our previous research, described in Chapters 1 and 2, had shown that children’s explicit knowledge of morphemes and grammar differs from their implicit knowledge. When morphemes are part of the regular functions of spoken language, children are reasonably good at using them creatively, as they did when they inﬂected pseudowords in our version of the Berko task. In contrast, in order to spell past regular verbs with “-ed”, without ever using the “-ed” ending in other words, children would have to be able to represent the category of past verbs in a highly explicit form. We showed in Chapters 1 and 2 that there is a considerable difference in the difﬁculty of the spoken and the written tasks. Second, our contribution to promoting children’s awareness of morphology and grammar differs radically from that used in the previous studies and, as our results show, this conceptual advance was

Morphemes and literacy: Context and conclusions 179

translated into practical beneﬁts for the children. Dixon and Henderson differ in their ideas about how to teach, but they seem to have a similar conception of what it means to make knowledge explicit: For them, it means to put knowledge into words. As our studies have shown, in particular Chapter 2, adults might spell pseudowords correctly using morphological principles and be quite unable to explain why they spelled the pseudoword in the way they did. Our approach to how to teach in order to develop children’s awareness of morphemes was to engage the children in a variety of problem-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 knew from our own research that the same amount of experience has different results depending on whether the children applied the same rule over and over or whether they had to make decisions about the different spelling possibilities. In order to implement these ideas, we developed several tasks that, our evidence shows, encourage children, in a way that is interesting to them, to perform various operations on morphemes. These tasks demand that children make some explicit representation of the morphemes, because they require children to think about the morphemic structure of 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 (see Chapter 1) children’s awareness of morphemes. In analogy tasks, the children are given one pair of words (for example, “read”–“reader”) and asked to produce the missing word to complete the pair “magic”–“?”. Phonological strategies, such as rhyming, cannot lead to the correct answer. In order to solve this analogy problem, children need to realize that there is a morphological transformation from the first to the second word in the first pair, which results in forming a “person word,” and then apply a similar transformation to the ﬁrst in the second pair. Counting the number of morphemes in particular words (see Chapter 3) is another operation that requires children to make morphemes into an object of thinking: How many morphemes are in “unforgettable”? It is difﬁcult to count something if you do not become explicitly aware of the items that you are counting. Putting morphemes into categories (see Chapter 3) also requires awareness of morphemes: If you have to sort words into those that contain sufﬁxes that form “person words” and those with sufﬁxes that

180 What are the overall implications?

form “other words,” you need to think about the words differently from the way that you think about them when you are just using them in speaking. 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 sufﬁxes in 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 children’s solutions in the classroom with them and their classmates and to provide them with a metalanguage that represented the elements of language under discussion and thus promoted the development of explicit awareness of morphemes and grammar. Discussion does help to put thoughts into words, but it is not the memorization of a rule. We think that this is a major contribution to a new way of thinking about teaching the use of morphemes and grammar in spelling. Through these games, teachers now have the support that they need for engaging children in a variety of intellectually challenging and motivating activities, which draw on much more than memory and definitions. These activities draw on children’s 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 3–6 always attracted and maintained children’s involvement and interest. The use of a variety of intellectual operations is probably also one of the reasons the program worked. Psychological research, referred to in Chapter 3, suggests that the key to generalization is variation. We think that the variety of learning tasks played an important part in improving the children’s understanding of the relatively abstract morphemic spelling principles that they were learning about. We have strong evidence to show that the program that we designed works in promoting children’s use of a new set of spelling principles. Our laboratory experiments were designed with the greatest care, using the highest scientiﬁc standards: Random assignment to conditions, teaching procedures described in detail, control groups exposed to appropriate experiences, and relatively large samples so that the danger of rejecting good hypotheses was minimized and the possibility of generalizing the results was maximized.

Morphemes and literacy: Context and conclusions 181

Our outcome measures give further weight to the nature of the evidence for the validity of our interventions. Children’s improvement in word spelling is undoubtedly a main measure: This is one of the outcomes that we expect of education. However, by itself, this measure does not tell us whether the children learned isolated words or a principle. The fact that we have significant results in children’s spelling of pseudowords, and that, when it was possible to obtain delayed posttest measures, these tended to remain signiﬁcant, gives us confidence in the children’s learning of a principle rather than the acquisition of isolated spellings of words. Our use of an outcome measure that assessed children’s vocabulary growth, with a complementary pseudoword-definition measure, shows that learning about morphemes has the potential to result in a wider impact on children’s literacy. Children’s vocabulary is consistently related to their literacy performance, and so it could be expected that, in the long run, the beneﬁts of learning about morphemes would show up in other measures of literacy. However, we think that this would be more likely if a consistent 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 in the curriculum could go well beyond literacy learning in the first language. Our past research (Castro et al. 2004) has shown that young Portuguese students’ awareness of morphemes and grammar in their first language is a predictor of how well they perform in learning a second language—in the case of our research, English. These students received the same teaching of English by the same teacher, but those who at the start of the year had better awareness of morphology and grammar in their ﬁrst language learned the second language better than those who had lower levels of awareness. Thus, children can transfer awareness of grammar and morphology from their ﬁrst language to the second language they learn. In an interview (presented in Chapter 6), a teacher from London described how she had had no awareness of English morphology until she started studying French! She was able to transfer what she learned about French grammar and morphology back into her own language, English, and this is what informed her knowledge of morphemes in English. So we know that teachers, as well as children, can use their knowledge of morphology across languages. The effects of this transfer are bound to be positive.

182 What are the overall implications?

Future developments We think that there is much more that can still be done in designing activities for teaching children about morphemes, and we would like to end with an idea that still needs testing. Think of the word pairs “tide”–“tied” and “guest”–“guessed”. They are homophone pairs, so they sound the same. However, their spelling can be distinguished on the basis of morphology. In the sentences “I tripped because my shoes were not tied” and “I did not know the answer but I guessed,” only one of the spellings could be correct. The words “tied” and “guessed” are two-morpheme words and the suffixes are there for grammatical reasons. The use of morphology should make it easier for us to distinguish between these homophone pairs than to distinguish between “weak”–“week” or “great”–“grate” because both words in these latter pairs are one-morpheme words. Some recent research of ours (Nunes and Bryant 2006) has shown that children aged 7 or 8 cannot use their knowledge of morphemes to improve their ability to distinguish between homophones when one of them is a two-morpheme word. Even with older children, aged 9 or 10, the evidence is not unambiguous, and though they might do better on some measures, other measures do not show the difference. Can children learn to use their knowledge of morphemes to distinguish between the homophone pairs where this knowledge is relevant? We conclude by saying that this book has provided solid evidence to suggest that English children benefit from becoming more aware of morphemes and grammar. This awareness enhances both their literacy and their spoken vocabulary. The path from the laboratory to the classroom has been opened so that classroom practice and policy can be put in place. The difficulties that have been associated with the idea of teaching rules to be memorized and with boredom no longer need to be an obstacle. We know that there are benefits to be harvested, but we don’t yet know how far they extend. In the preceding section, we touched on further possible benefits that can be obtained in second-language learning from awareness of morphology and grammar. There is a lot to be gained by improving the teaching of English.

Appendix

The four research strategies in this research program

All of our proposals about teaching reading and spelling and morphology stem directly from the evidence that we have collected over the years on children’s morphological knowledge. The strength of our case for change, therefore, depends on the quality of this evidence, and this in turn depends on the power and suitability of our research methods. In this appendix, we will summarize our methods and our reasons for adopting them. Our aim is to show you that these were the right methods and that they provided good evidence for the worth of our educational proposals. The main part of our book deals with a set of research studies that were part of the ESRC’s Teaching and Learning Research Programme (TLRP). However, quite naturally, these TLRP studies of ours were preceded by other studies from our research group that led to the questions which prompted our TLRP research. Part I of our book deals with some of the earlier work, and Part II describes the TLRP intervention studies. Together, the studies that we describe in Parts I and Part II add up to a research agenda that combines several different research strategies in order to study morphological knowledge and its relation to reading and spelling. In the end we adopted four different strategies in different parts of our research.

Strategy 1: Using the longitudinal method to establish the connection between children’s knowledge about morphemes and the development of literacy We started our research on children and morphemes with a causal question. Does children’s morphological knowledge affect their progress Authored by Peter Bryant and Terezinha Nunes

184 Four research strategies

in learning to spell? Our hypothesis was that morphological knowledge is important even when, as was the case then, children are taught very little about morphemes. We decided that the best way to investigate this hypothesis was to do a large-scale longitudinal study, that is, to see the same children several times over a long period and to look at the relation between their morphological knowledge and their spelling. Longitudinal studies are an essential part of testing causal hypotheses. If, for example, children’s morphological knowledge affects how they learn to spell, then their scores on a test of morphological knowledge should predict how well they spell later on. The results of our opening longitudinal study established this predictive link. The children’s scores in the initial tests of morphological knowledge did predict their spelling performance over a year later, even when our statistical analyses controlled for differences in their age, IQ and their original spelling levels. Thus this longitudinal study established the relevance of morphological knowledge to children’s literacy.

Strategy 2: Using laboratory interventions to improve children’s morphological knowledge The longitudinal study established an important connection between children’s morphological knowledge and their spelling. Our idea is that this connection is a causal one. In other words, we think that children’s knowledge of morphology in spoken language determines how well they learn and use morphological spelling rules in written language. In our laboratory intervention studies we looked at this possible causal connection in two ways. The ﬁrst way was to use the intervention method to see if this causal connection really does exist. It may seem obvious that cause and effect must go in this way, but it might not be true. Some other factor, such as children’s general linguistic development, might determine both how well they understand morphology in spoken language and also how quickly they learn morphological spelling rules. To establish a causal link such as the one that we had in mind, one must resort to a different method, which is to carry out one or more tightly controlled intervention studies. The rationale for doing so is a simple one. Suppose that children’s morphological knowledge really does determine or partly determine how well and quickly they learn about morphologically determined spellings. In that case one effect of an increase in a child’s morphological

Four research strategies 185

knowledge will be a consequent improvement in this child’s use of morphological spelling rules. If you were to teach children effectively about morphemes, you should also be helping them with spelling—at any rate if the hypothesis is right. Therefore, the questions that we asked in our first intervention experiments were whether it is possible to enhance morphological knowledge and what effect this would have on their reading and spelling. Notice that this is not a directly educational question. We set up several traditional “intervention experiments” (described in Part I of this book), which is the term that psychologists use to describe studies in which they directly intervene by teaching at least one group of participants and then comparing them to at least one other group who have not received this instruction. It is not our purpose here to describe either the instruction that we gave or the results of these studies. In this appendix our concern is with the design of the studies that led to our conclusions. Intervention studies have to include at least three phases. 1.

2.

3.

The ﬁrst of these is the pretest, in which all the participants are given the same set of tasks, including tasks that test the skill, for example, of spelling the sort of polymorphemic words that the intervention is to be aimed at. The next phase is the intervention itself, which might last for a brief time, like 1 hour, or might go over several weeks or even months. Only some of the participants are given this intervention and they form the “intervention” or “experimental” group(s). The other participants do not receive the intervention or instruction which the experimenters predict will enhance the skill in question, and are usually called the “control” group(s). There are two types of control group: Those that are given as much attention and as much instruction as the intervention groups and those (sometimes called “unseen control groups”) that are not given any extra attention or instruction. On the whole, the ﬁrst type of control group makes for a more satisfactory comparison. The ﬁnal phase is the posttest in which the participants are given the 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 one posttest, and it is quite common in intervention studies to arrange for two posttests, one an “immediate posttest” and the other a “delayed posttest” which the participants are given two months or so later.

186 Four research strategies

If the experimenters’ hypothesis is right, there should be more improvement from pretest to posttest among the participants in the intervention group(s) than in the control group(s). In the intervention studies that we describe in Part I, we did ﬁnd consistent signs of this pattern, and so we concluded that it is, in principle, possible to instruct children about morphemes in a way that affects their reading and spelling. This was not a claim about teaching in the classroom. During the intervention the actual instruction period was given for the most part on a one-to-one basis in these ﬁrst intervention studies. Our aim was to ﬁnd out whether this instruction worked, before we looked at whether it worked in the classroom. Our second way of using laboratory intervention research to pursue our hypothesis was to compare the effectiveness of different teaching methods. Intervention experiments are excellent instruments, not just for studying the effects of enhancing a particular ability, but also for comparing different forms of intervention. This was the main purpose of the ﬁrst intervention study that we describe in Chapter 3. Here, we used the method of laboratory intervention to settle a theoretical question about explicit and implicit learning. (We give the reasons for asking this question in Chapter 3.) We decided to give our intervention groups instruction about one highly speciﬁc morphological spelling rule—the rule about “-ion” and “-ian” endings. Our reason for doing this was that we knew that this spelling rule is never taught in schools in the U.K., and thus our intervention would take the form of teaching children about something that they had never been taught before. We adopted the familiar pretest–intervention–posttest (immediate and 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 instruction is effective but also whether one form of instruction is better than another. The answer to this second question depended on the relative success of the three intervention groups in the two posttests. You can ﬁnd this answer in Chapter 3. One other aspect of this study is worth mentioning. The study was still a laboratory one, and a long way from the classroom, but we no longer stuck to one-to-one instruction. Instead, we taught the children in pairs

Four research strategies 187

and we encouraged them to discuss their solutions to the various intervention tasks that we gave them among themselves. This was our ﬁrst move away from one-to-one intervention and toward instruction in groups.

Strategy 3: Using classroom interventions to ﬁnd out whether it is possible to repeat the success of the laboratory interventions in a classroom setting Our laboratory interventions were about teaching, but children on the whole are taught in classrooms. Although it was essential for us to do the laboratory studies, it was just as important to check that our positive results could be repeated in the realistic context of the classroom. We did this in three ways. One way was to repeat, fairly exactly, part of one of our laboratory studies in a classroom setting. We did this in the second study that we describe in Chapter 3. We, and our teacher colleagues at Lauriston schools, wanted to find out whether the instruction, which had enhanced children’s spellings of the “-ion” and “-ian” endings in our laboratory study, would work as well in the classroom. So the Lauriston team suggested that they could take the most successful method, the explicit intervention, and give it to their students in their classrooms. The team’s idea, which we enthusiastically accepted, was that they would use the same material, the same pretest, intervention period and posttest (immediate and delayed) design, and the same intervention tasks as we had, and that they would also compare an intervention group and a control group of children who were instructed about something else during the intervention period. A great deal depended on the results of this classroom study, since if the effects of the intervention had been different in the Lauriston study from those in our laboratory study we would have had the difﬁcult task of working out the reasons for the discrepancy. So, we were gratiﬁed by the similarity of the results of the Lauriston classroom study to our own 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 the classroom was to broaden our teaching about morphemes. If schools do radically increase their teaching about morphology and morphological

188 Four research strategies

spelling rules, as we are now advocating they should do, this instruction would have to be about a wide range of inﬂectional and derivational morphemes. Therefore, our next study (the third study in Chapter 3) was on teaching schoolchildren about morphemes in general and about several different morphological spelling rules. This was also a classroom study, in which teachers gave students the actual instruction about morphemes and morphological spelling rules. Again we adopted a pretest–intervention–posttest design and compared an intervention group with a control group. However we, the researchers, prescribed the methods and the timetable and we were always at hand during the intervention period. Again our intervention was successful. Our third way of bringing intervention about morphemes and spelling into the classroom was to study what would happen if we relinquished direct control of the intervention. In our other classroom research we exercised a great deal of control over what the teachers did, the order in which they administered the interventions, and the timing of the different parts of the interventions. Even just by being available during the intervention period we had had an effect on what went on in the classroom. This level of control had been absolutely necessary at this stage of the research, but it was not “real life”. If teachers are to adopt 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 are bound to be great differences between different teachers in how they do so. In the studies that we describe in Chapters 4 and 5, we stuck to the pretest–intervention–posttest design, with comparisons between intervention and control groups, and, of course, we organized and carried out the pretests and posttests ourselves. We delivered the interventions that we had devised to the schools with some explanatory material, but we had no other control than this. Thus, this really was a realistic study. What happened in the intervention was what would normally happen when a school adopts a new method of teaching.

Strategy 4: Using interviews and intervention to study the teachers as well In the three strategies that we have described so far, the focus of the work was always on the children’s morphological knowledge and on how to enhance it, but we were aware all the while that there was

Four research strategies 189

another variable of great importance to take into account. In the end it will be the task of teachers to tell children about the link between morphemes and spelling and to interest them in this connection. We needed, therefore, to pay attention to teachers as well, and particularly to their interest in morphology and their knowledge of morphological spelling rules. Since at present morphemes and morphological spelling rules are either not taught at all at school or taught extremely patchily, we could not take this knowledge for granted in teachers. We had to ﬁnd out how aware teachers are of morphological structure and what effect their awareness has on the students whom they teach. The research strategy in our work with teachers, which we describe in Chapter 6, was a two-pronged one. First, we investigated teachers’ knowledge of morphemes and of morphological spelling rules. We did this by interviewing a group of experienced teachers, and the interviews revealed a marked lack of explicit knowledge about morphemes among these teachers. Our second move was to devise a rather interesting form of the intervention study. The focus of our intervention was a group of teachers who were enrolled on a course, which ran for several weeks, on morphology and its impact on spelling. Before the course and at its end, we tested children in these teachers’ classes and other children who were not taught by the teachers who took our course on morphemes. Thus, the intervention was given to teachers, but the pretests and posttests which told us about the effectiveness of the intervention were given to schoolchildren who either were or were not the students of these particular teachers. In this study we showed that teachers themselves usually need some instruction about morphemes, and we also demonstrated that it is possible to give them the instruction that they need in a way that helps them and eventually their students as well.

Summary Thus, to summarize, we decided on a combination of four different research strategies to discover (a) whether there is a need to teach children systematically about morphemes and spelling and (b) how to do this teaching. 1.

Our ﬁrst research strategy was longitudinal research that we did in order to establish that there is a connection between children’s

190 Four research strategies

2.

3.

4.

morphological knowledge and their learning about morphological spelling rules. Our second strategy took the form of a set of laboratory intervention studies, and here we had two aims. One aim was to discover if the link between children’s morphological knowledge and their spelling is a causal one. Our other aim was to compare the effects of explicit and implicit teaching about morphemes and spelling. Our third strategy was a series of classroom intervention studies, the purpose of which was to discover whether the teaching methods that we had developed in the laboratory could be adapted to the classroom. In our fourth research strategy we turned to teachers and in particular to their knowledge of the link between morphemes and spelling. We used interviews to assess this knowledge and then we looked at the effect of improving the teachers’ knowledge of morphological spelling rules on their pupils’ spelling.

It seems to us that none of these research strategies on its own would be a good enough basis for applying the insights of psychologists to classroom practice. It is the combination of the four different approaches that allows us to claim that it is possible and highly worthwhile to teach schoolchildren about morphemes, and that there are no insurmountable obstacles to doing so.

References

Aidinis, A. and Nunes, T. (2001) “The role of different levels of phonological awareness in the development of reading and spelling in Greek,” Reading and Writing, 14: 145–77. Anderson, J. R. (1981) Cognitive Skills and their Acquisition. Hillsdale, N.J.: L. Erlbaum Associates. Berko, J. (1958) “The child’s learning of English morphology,” Word, 14: 150–77. Brown, R. (1957) “Linguistic determinism and parts of speech,” Journal of Abnormal and Social Psychology, 55: 1–5. Brown, R. (1973) A First Language. Cambridge, Mass.: Harvard University Press. Bryant, P., Devine, M., Ledward, A., and Nunes, T. (1997) “Spelling with apostrophes and understanding possession,” British Journal of Educational Psychology, 67: 93–112. Bryant, P., Nunes, T., and Aidinis, A. (1999) “Different morphemes, same spelling problems: Cross-linguistic developmental studies,” in M. Harris and G. Hatano (eds) Learning to Read and Write: A Cross-Linguistic Perspective. Cambridge: Cambridge University Press, pp. 112–33. Bryant, P., Nunes, T., and Bindman, M. (2000) “The relations between reading ability and morphological skills: The case of the apostrophe,” Reading and Writing, 12: 253–76. Cardoso-Martins, C. (1994) “Rhyme perception: global or analytical?” Journal of Experimental Child Psychology, 57: 26–41. Carlisle, J. (1988) “Knowledge of derivational morphology and spelling ability in fourth, sixth and eighth graders,” Applied Psycholinguistics, 9: 247–66. Carlisle, J. F. (1995) “Morphological awareness and early reading achievement,” in L. B. Feldman (ed.) Morphological Aspects of Language Processing. Hillsdale N.J.: L. Erlbaum Associates, pp. 189–209. Casalis, S. and Louis-Alexandre, M.-F. (2000) “Morphological analysis, phonological analysis and learning to read,” Reading and Writing, 12: 303–35.

192 References Castro, A., Nunes, T., and Strecht-Ribeiro, O. (2004) “Relação entre consciência gramatical na linguagem materna e progresso na aprendizagem de uma língua estrangeira,” Da Investigação às práticas. Estudos de Natureza Educacional, 5: 51–66. Chliounaki, K., and Bryant, P. (2003) “Choosing the right spelling in Greek: Morphology helps,” Revue Française de Linguistique Appliquée, 8: 142–52. Department for Education and Employment (1998) The National Literacy Strategy: 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 Planning Exempliﬁcation 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. 237–70. Ehri, L. (1986) “Sources and difﬁculty in learning to spell and read,” in M.L. Wolraich and D. Routh (eds) Advances in Developmental and Behavioural Paediatrics. Greenwich, Conn.: JAI Press, Vol. VII, 121–95. Elbro, C. (1989) “Morphological awarness in dyslexia,” in C. von Euler, I. Lundberg, and G. Lennerstrand (eds) Brain and Reading: Structural and Functional Anomalies in Developmental Dyslexia. Basingstoke: Macmillan, pp. 279–91. Fayol, M., Hupet, M., and Largy, P. (1999) “The acquisition of subject–verb agreement in written French: From novices’ to experts’ errors,” Reading and Writing, 11: 153–74. Fayol, M., Thenevin, M.-G., Jarousse, J.-P., and Totereau, C. (1999) “From learning to teaching to learn French written morphology,” in T. Nunes (ed.), Learning to Read: An Integrated View from Research and Practice. Dordrecht: Kluwer, pp. 43–64. 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 and orthography in morphological awareness,” in L. B. Feldman (ed.), Morphological Aspects of Language Processing. Hillsdale N.J.: L. Erlbaum Associates, pp. 157–88. 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. 301–30. Gleitman, L. R. (1990) “The structural sources of verb meaning,” Language Acquisition, 1: 3–55.

References 193 Gleitman, L. R. and Gleitman, H. (1992) “A picture is worth a thousand words, but that’s the problem,” Current Directions in Psychological Science, 1: 31–5. Henderson, E. (1990) Teaching Spelling, 2nd edn. Dallas, Tex.: Houghton Mifﬂin. Henry, M. K. (1989) “Decoding instruction based on word structure and origin,” in P. G. Aaron and R. M. Joshi (eds) Reading and Writing Disorders in Different Orthographic Systems. Dordecht: Kluwer, pp. 25–49. Karmiloff-Smith, A. (1992) Beyond Modularity: A Developmental Perspective on Cognitive Science. Cambridge, Mass.: MIT Press. Kemp, N. and Bryant, P. (2003) “Do beez buzz? Rule-based and frequencybased knoweldge in learning to spell plural -s,” Child Development, 74: 63–74. Levin, I., Ravid, D., and Rapaport, S. (1999) “Developing morphological awareness 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. 77–104. Marsh, G., Friedman, M. P., Welch, V., and Desberg, P. (1980) “The development of strategies in spelling,” in U. Frith (ed.), Cognitive Processes in 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) Children’s 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. 151–70. Nunes, T., Bryant, P. E., and Bindman, M. (1997b) “Morphological spelling strategies: Developmental stages and processes,” Developmental Psychology, 33: 637–49. Nunes, T., Bryant, P., and Olsson, J. (2003) “Learning morphological and phonological spelling rules: An intervention study,” Reading and Writing, 7: 289–307. 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 . Accessed 22 January 2006. Read, C. (1971) “Pre-schoolchildren’s knowledge of English phonology,” Harvard Educational Review, 41: 1–34. Robinson, J. W. and Hesse, K. D. (1981) “A morphemically based spelling

194 References program’s effect on spelling skills and spelling performance of seventh grade students,” Journal of Educational Research, 75: 56–62. Schonell, F. J. (1957) Essentials in Teaching and Testing Spelling, ﬁrst published in 1932. Basingstoke: Macmillan. Templeton, S. and Bear, D. (1992) Development of Orthographic Knowledge and the Foundations of Literacy. Hillsdale, N.J.: L. Erlbaum Associates. Treiman, R. (1993) Beginning to Spell: A Study of First-Grade Children. New York: 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. 61–80. 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: 107–36. Venezky, R. L. (1970) The Structure of English Orthography. The Hague: Mouton. Venezky, R. L. (1995) “How English is read: Grapheme-phoneme regularity and orthographic structure in word recognition,” in I. Taylor and D. R. Olson (eds) Scripts and Literacy. Dordrecht: Kluwer, pp. 111–30. Zutell, J. (1993) “Directed spelling thinking activity: A developmental, conceptual approach to advance spelling word knowledge,” paper presented at the Nineteenth National Conference of the Australian Reading Association, Melbourne, Australia.

Index

ability 117–19, 128–30, 131–2 ‘-able’ 7 abstract nouns 67, 98, 107, 108 achievement groups 117–19, 128–30, 131–2 adjectives 6, 99 afﬁxes 4–7; see also preﬁxes; sufﬁxes age: conserving stems 53–4, 55; different types of linguistic knowledge 48–9; effect of classroom intervention on spelling 105, 106; spelling of pseudowords with ‘-ian’ and ‘-ion’ 57–8; spelling of sufﬁxes and 38–9, 41, 45–6, 55; and use of morphemes 10; use of ‘s’ as a morpheme marking plural and third-person singular of verbs 60 agents 88 Aidinis, A. 20 alphabetic conception of writing 166 alphabetic principles 168, 169, 170; children’s understanding of insufﬁciency of 171 alphabetic stage (decoding letter by letter) 144 analogy tasks 90, 179; laboratory study 71–4, 76; sentence analogy task 13, 29; transforming present-tense verbs into pasttense verbs 10–11; word analogy task 13–14, 29 analyzing novel words 90 Anderson, J.R. 175

apostrophes 22, 144, 177 autoregression 30 bases (base words) 5, 6, 88 Bear, D. 171 Berko, J. 12, 14–15, 47 Birgisdottir, F. 56 Braxton, T. 4 Brierley, V. 56 British Picture Vocabulary Scale (BPVS) 126 Brown, R. 8, 10 Bryant, P. 17, 20, 59, 169, 173, 177, 182 Cardoso-Martins, C. 168 Carlisle, J. 14, 24–5 Casalis, S. 25 Cassar, M. 144 Castro, A. 181 categories, grammatical 88, 90, 92, 97–8 categorization 90, 179–80 cause and effect 184–6, 190; in connections between morphological knowledge and learning to spell morphemes 24–32 CD-ROM 104, 106, 123, 128, 161 Chliounaki, K. 20 classroom interventions 187–8, 190; Lauriston Primary School study 81–6, 102, 160, 187; teaching package see teaching package

196 Index contamination effects 84 contrasts 86–7, 169, 173–4 correction game 74–6, 76, 90 Corrective Spelling through Morphographs 174–5 counting the number of morphemes 90, 179 delayed posttests 185; classroom intervention and vocabulary 124–8, 128–32; laboratory study 68–9, 78, 78–9, 80–1; Lauriston Primary School study 83–4, 85; teaching package 94, 95–6, 100–1 Department for Education and Employment (DfEE) 134 Department for Education and Skills (DfES) 122, 134 derivations 6, 7, 10, 15, 25, 87–8, 103, 176 development: of spelling 166–7; stages of for reading and spelling 144 dictionaries 145 differentiation 150, 151 digraphs 171 dinosaur names 51–4 Directed Spelling Thinking Activity method 173 Dixon, R. 174–5, 179 ‘-ed’ ending 7, 55, 176; changes in spelling over time 25–9; development of spelling 166–7; spelling 36–40, 44–7; teachers’ lack of awareness of ‘-ed’ rule 137; teachers’ strategies for teaching spelling 136–7 educational theory and practice 157–82; children’s understanding of English spelling 166–7; contribution of classroom spelling project 178–81; higher-order principles based on meaning 174–8; higher-order spelling principles in teaching spelling 172–4; need to teach spelling 162–5; path from laboratory to

classroom 159–62; spelling units or spelling principles 168–72 Ehri, L. 144, 166 Elbro, C. 25 ‘-en’ ending 92, 98–9 English as an additional language 151 ‘-er’ ending 17, 88 errors in spelling 135; analysis of errors and spelling sufﬁxes 42–4, 45, 46; and development of spelling 166–7; use of higher-order principles 171 ESRC Teaching and Learning Research Programme (TLRP) 183 experimental group(s) 185–6 expertise 175 explanations for spellings of pseudowords 55–61 explicit knowledge of morphemes 9–24, 158, 178–9; as essential ingredient of learning to read and write 16–24; measuring morphological awareness 10–15; teachers’ 135–42, 145, 153–4, 189 explicit teaching 86, 87; laboratory study 66, 67, 68, 73, 75, 76–7, 78–81, 101–2; Lauriston Primary School study 81–6 facts, learning speciﬁc 90 Fayol, M. 23 feedback 92–4, 97 Fernald, G. 165 Ferreiro, E. 166 form and function 170, 171–2, 172–4 Fowler, A.E. 14, 25, 51 French 6, 23–4, 150, 181 Frith, U. 144 function and form 170, 171–2, 172–4 generalization 90–2; see also principles Gleitman, L.R. 8 grammar 7–9; importance of

Index 197 developing awareness of 87–9; teachers’ knowledge of 150 grammatical categories 88, 90, 92, 97–8 grapheme–phoneme correspondences 169 graphic units 168–9 Greek 6, 20–1 Hebrew 31–2 Henderson, E. 159, 165, 167, 168, 173, 174, 175–6, 179 Henry, M.K. 174, 177 Hesse, K.D. 175 high achievers 117–19, 128–30, 131–2 higher-order spelling principles 169–78; based on meaning 174–8; in teaching spelling 172–4 Hillingdon Cluster of Excellence 122 homophones 176, 182 Hupet, M. 23 ‘-ian’ endings 17, 98, 141, 186; effect of classroom intervention on spelling 107–9; pseudowords with 56–8, 70–1, 78, 80–1, 85; spelling ‘-ion’ and 40–4, 46–7; teaching pairs of children about difference between ‘-ion’ endings and 65–81, 86, 101–2; teaching about in the classroom 81–6, 102; teaching package 97, 98 imagination, importance of engaging 92, 93 immediate posttests 185; classroom intervention and vocabulary 124–8, 128–32; laboratory study 68–9, 77, 78–9, 80–1; Lauriston Primary School study 83–4, 85; teaching package 94, 95–6, 100–1 implicit knowledge of morphemes 9, 10–16, 157–8, 177 implicit learning of spelling 163 implicit teaching 66–7, 67, 68, 73, 75, 77, 78–81

individualized word lists 164 inﬁxes 4 inﬂections 4–5, 6, 10, 15, 25, 48–9, 88 internal consistency of tests 124–6, 126 intervention group(s) 185–6 intervention studies 185–6 ‘-ion’ endings 17, 98, 186; effect of classroom intervention on spelling 107, 108; knowing when to use 40–4; pseudowords with 56–8, 70–1, 78, 80–1, 85; spelling sufﬁxes 36–44, 46–7; teachers’ strategies for teaching spelling 141–2; teaching about in the classroom 81–6, 102; teaching pairs of children about difference between ‘-ian’ endings and 65–81, 86, 101–2; teaching package 97, 98 irregular verbs 45–6 Jarousse, J.-P. 23 journals, spelling 146 Karmiloff-Smith, A. 47 Kemp, N. 17 kinetic learning 145 knowledge of morphemes 35–62, 177; different types of 47–50; explicit see explicit knowledge of morphemes; implicit 9, 10–16, 157–8, 177; and literacy learning 157–8; longitudinal studies 183–4, 189–90; spelling sufﬁxes 36–50; spelling words with the same stem 50–5; spellings of pseudowords and 55–61; teachers’ 135–42, 145, 153–4, 189 laboratory study 65–81, 86, 101–2, 160, 180; children 67–8; posttests 68, 69–71, 77–8, 78–9, 80–1; pretest and posttest measures of spelling 69–71; research strategy 184–7, 190; results 78–81; study design 68–9;

198 Index two teaching and learning sessions 71–7 Largy, P. 23 Lauriston Primary School study 81–6, 102, 160, 187 letter patterns 46–7, 139–40, 145, 172–4 letter strings 44, 139–40, 145, 172–3 Levin, I. 31–2 Liberman, I.Y. 14, 25, 51 literacy: connection between morphemes and 24–32, 158; development of 183–4; see also reading; writing literacy hour 152 logographic (whole-word) stage 144 longitudinal method 183–4, 189–90; morphological knowledge and development of literacy 29–32 look-cover-write-check method 145, 165 Louis-Alexandre, M.-F. 25 low achievers 117–19, 128–30, 131–2 ‘-ly’ ending 99 Marsh, G. 170–1 meaning: higher-order spelling principles based on 174–8; morphemes as units of 3–4; teachers’ strategies for teaching spelling 138–40 memorization 163–4, 165, 167 memory cues 151–2 missing word tests see sentence completion tasks mixed-ability teaching 151 mixed implicit and explicit teaching 67, 68, 73, 75, 77, 78–81 Montessori, M. 165 morpheme-teaching programme see teaching package morphemes: cause and effect in connections between knowledge of morphemes and learning to spell morphemes 24–32; impact of teaching about morphemes on literacy 158; importance in education 9–24; knowledge of see

knowledge of morphemes; morphemic structure and spelling 157; nature of 3–9; represented in writing but not in speech 22–4; same sounds spelled in different ways in different morphemes 16–21; spelled in same way but represented by different sounds 21–2; teachers’ deﬁnitions 145; teaching morphology and improving spelling 32–3 ‘morphemes-only’ intervention: effects on spelling 105, 106, 110–19, 119–20; effects on vocabulary 124, 128–32 ‘morphemes-plus-spelling’ intervention 105, 106–9, 110–19, 119–20 morphemic stage in spelling 144 morphographs 175 morphological awareness see explicit knowledge of morphemes morphological knowledge see knowledge of morphemes morphological production tasks 14, 24–5 morphology, theories about 143–4 morphology children (teacher intervention) 146–9 ‘motion’ 135 motivation, importance of maintaining 92–4 Mudd, N. 172 multisensory methods 165 National Curriculum 134 National Literacy Strategy (NLS) 40–1, 134, 148–9, 154 negatives, preﬁxes that make 98 ‘-ness’ sufﬁx 36–40, 46–7 nonsense words see pseudowords nouns 6, 87–8; abstract nouns 67, 98, 107, 108; person nouns 93, 98, 107–9 novel words, analyzing 90 number 90, 91, 99 Nunes, T. 10–11, 12–14, 20, 25, 33, 87, 144, 160, 169, 173, 182

Index 199 ofﬂine tasks 47–50 online tasks 47–50 onset and rime 168 oral language see speech orthographic stage (recognizing chunks in words) 144 outcome measures 178, 181 overgeneralization errors 27–9, 45, 46 pairs of children laboratory study see laboratory study past-tense endings 15, 21–2; changes in spelling over time 25–9; see also ‘-ed’ ending patterns 46–7, 139–40, 145, 172–4 people, ‘-ian’ morpheme and 67 person nouns 93, 98, 107–9 Peters, M. 163, 164, 165 phoneme–grapheme correspondences 169 phonetic conception of writing 166 phonetic stage in spelling 144 phonic strategies 145 phonological awareness 151 phonological explanations 58 phonological principles 31–2, 33, 87 phonological spelling 25–6, 113–15, 135 phonological units 168–9 phonology 174; explanations for spelling of pseudowords 61; problems with relating to 135 Piagetian theory 166 plurals 6, 15; ‘silent’ endings in French 23–4; use of ‘s’ as a morpheme marking 59–61 Portuguese 20, 181 possessive apostrophe 22 posttests 185–6; effects of classroom intervention on spelling 104, 109–10, 110–16, 117–19; effects of classroom intervention on vocabulary 124–8, 128–32; laboratory study 68, 69–71, 77–8, 78–9, 80–1; Lauriston Primary School study 83–4, 85; teacher intervention 148–9; teaching package 94, 95–6, 99–101

preﬁxes 4, 5–6, 139; effect of classroom intervention on spelling 109; teaching package 86–101, 102–3; teaching lists of 177 pretests 185–6; effects of classroom intervention on spelling 104, 109–10, 110–16, 117–19; effects of classroom intervention on vocabulary 124–8, 128–32; laboratory study 68, 69–71, 78, 79, 80; Lauriston Primary School study 83, 84, 85; teacher intervention 148–9; teaching package 94, 95–6, 99–101 priming 54 principles 166–7, 168–78; higher-order spelling principles based on meaning 174–8; higher-order spelling principles in teaching spelling 172–4; ‘-ion’ and ‘-ian’ endings 19–20, 65–6, 67; learning of vs learning speciﬁc facts 90; phonological 31–2, 33, 87 problem words 17–20, 163–4 productive morphology task 12–13, 29 pronunciation 176 proofreading 145 pseudoword-deﬁnition test 126, 127, 127–8, 130–2, 132–3 pseudoword-spelling tests: impact of classroom intervention on spelling 109–10, 113, 115–16, 119; teacher intervention 146–8; teaching package 95, 96, 100–1, 102 pseudowords 8–9; engaging children’s imagination 92, 93; higher-order spelling principles 170–1; knowledge of morphemic principles 14–15; laboratory study 70–1, 78, 80–1; Lauriston Primary School study 85; productive morphology task 12–13, 29; sentence completion task with 14–15, 47–9; spelling different words with the same

200 Index stem 51–4; spelling sufﬁxes 37–9, 40–2; subtracting morphemes from 92, 93, 180; use of ‘s’ to mark plural or third-person singular of verbs 59–61; what spellings of pseudowords tell about knowledge of morphemes 55–61; with ‘-ion’ or ‘-ian’ 56–8, 70–1, 78, 80–1, 85 Qualiﬁcations and Curriculum Authority 134 Rapaport, S. 31–2 Ravid, D. 31–2 Read, C. 166 reading: explicit knowledge of morphemes as essential ingredient of learning to read 16–24; impact of learning to read on morphological knowledge 24–32; stages of development 144 reading comprehension 77, 121 real words, spelling see word spelling reasoning: importance of engaging children’s reasoning in different ways 90–2 regular verbs 45–6, 48–9, 55 research: contribution to educational theory and practice 178–81; transformation into educational practice 159–62 research strategies 183–90; classroom intervention studies 187–8, 190; laboratory intervention studies 184–7, 190; longitudinal method 183–4, 189–90; studying teachers 188–9, 190 rimes 168, 169 Robinson, J.W. 175 roots 4, 5; see also stems rules 151–2; see also principles ‘s’, as morpheme marking plural and third-person singular of verbs 59–61 Schonell, F. 164

schwa vowels 17–20, 50 segmentation of words 112–13 semi-phonetic stage in spelling 144 sentence analogy task 10–11, 13, 29 sentence completion tasks 14–15, 47–8, 90, 96, 97; laboratory study 69–71; vocabulary test 124–6, 127 sentence frames 88, 89 Sherrel, P. 56 ‘silent’ plurals 23–4 sounds: awareness of sufﬁx spellings that do not correspond to ending sounds 37–40; morphemes spelled the same way but represented by different sounds in different words 21–2; same sounds spelled in different ways in different morphemes 16–21 Speaking and Listening Skills project 122–3 special educational needs 151 speciﬁc facts, learning of 90 speech: knowledge of morphemes and 47–50; morphemes represented in writing but not in speech 22–4 spelling 35–62, 104–20, 161, 181; cause and effect in connections between morphological knowledge and learning to spell morphemes 24–32; children’s understanding of English spelling 166–7; different words with the same stem 50–5; higher-order principles in learning spelling 172–4; impact of classroom intervention on high and low achievers 117–19; intervention sessions 106–9; morphemes spelled in the same way but represented by different sounds in different words 21–2; need to be taught 162–5; overview of teaching package 104–6; pretest and posttest assessments 109–10; of pseudowords and knowledge of morphemes 55–61; pseudowords with ‘-ion’ or ‘-ian’

Index 201 56–8; results of classroom intervention 110–19; same sounds spelled in different ways in different morphemes 16–21; stages of development 144; ‘s’ as morpheme marking plural and third-person singular of verbs 59–61; sufﬁxes 36–50; teachers’ strategies for teaching 135–42, 145; teaching morphology and improving 32–3; theories about 143–4 spelling banks 145 spelling demons 17–20, 163–4 spelling investigations 145 spelling journals 146 spelling lists 164–5 spelling patterns 46–7, 139–40, 145, 172–4 spelling principles see principles spelling rules 145 spelling units 168–9 standard word lists 164 Stanford Achievement Test 175 stems 4, 5; conserving 50–5; identifying 93, 98; learning pseudowords 9; meaning principle for spelling 175–6; pronunciation change when a derivational morpheme is added 107; segmentation of words 112–13; spelling different words with the same stem 50–5 sufﬁxes 4, 5–6, 10, 87–8, 139; awareness of sufﬁx spellings that do not correspond to ending sounds 37–40; effect of classroom intervention on spelling 109–10, 110–19; higher-order spelling principles based on meaning 176, 177; knowledge of when and when not to use 40–6; spelling 36–50; spelling compared with stems spelling 54, 55; teaching package 86–101, 102–3; that change meaning 98 syllables, as spelling units 168 Syntactic Bootstrapping Hypothesis 8–9

teacher intervention 142–52, 153–4, 162, 189; case studies 150–2; changes in children’s spelling 147–9; impact on children 146–9; impact on teachers 144–6; sample of children 146; spelling assessment 146–7; teachers’ comments on the intervention process 149–50 teachers 134–54, 181; knowledge of morphemes 135–42, 145, 153–4, 189; research strategy 188–9, 190; strategies for teaching spelling 135–42, 145; see also teacher intervention teaching activities 179–80; CD-ROM 104, 106, 123, 128, 161; laboratory study 71–7; Lauriston Primary School study 82; teaching package 96–9, 106–9 Teaching and Learning Research Programme (TLRP) 183 teaching methods, comparing effectiveness of 66–81, 186–7, 190 teaching package 86–101, 102–3, 106–9, 160–1, 187–8; effects on spelling 104–20, 161; effects on vocabulary 121–33, 161–2; overview of teaching programme 96–9; pretests and posttests 94, 95–6, 99–101; principles used in teaching materials 87–94; results of study 99–101; study design 94–9 Teberosky, A. 166 technology, teachers’ concerns and 150 Templeton, S. 171 Thenevin, M.-G. 23 third-person singular of verbs 59–61 thought operations 90, 92 time 29–31, 152, 153, 154 Totereau, C. 23 Treiman, R. 144, 166, 168 ‘un-’ 6, 7, 177–8 validity of tests 126

202 Index variety of learning tasks 180 Venezky, R. 169 verbs 88, 92; irregular 45–6; regular 45–6, 48–9, 55; spelling sufﬁxes 44–7, 48–50; transforming present-tense into past-tense 10–11; use of ‘s’ as a morpheme marking third-person singular 59–61; see also past-tense endings visual learning 165, 172–3 vocabulary 9, 10, 103, 121–33, 161–2, 181; context of study 122–3; effects of morpheme intervention on 128–32; learning 8–9; morpheme teaching intervention 128; pretests and posttests 124–8, 128–32; study design 123–8 vocabulary test 124–6, 127, 128–30, 132–3 vowels, schwa 17–20, 50, 65

Wodehouse, P.G. 7, 139 word analogy task 13–14, 29 word classes 97–8, 99 word lists 164–5 word-speciﬁc learning 70 word spelling: impact of classroom intervention on spelling 109, 110–15, 117–18; laboratory study 69–70, 78–80; Lauriston Primary School study 83–4; teacher intervention 146–9; teaching package 95–6, 97, 99–100, 102 writing: impact of learning to write on morphological knowledge 24–32; morphemes represented in writing but not in speech 22–4; morphological knowledge and learning to write 16–24

waiting-list model 94, 102 whole-word stage 144 whole words, learning 145, 164, 165

‘-y’ ending 99

‘-x’, pseudowords ending in 59–61

Zutell, J. 173

eBooks – at www.eBookstore.tandf.co.uk

A library at your fingertips!

eBooks are electronic versions of printed books. You can store them on your PC/laptop or browse them online. They have advantages for anyone needing rapid access to a wide variety of published, copyright information. eBooks can help your research by enabling you to bookmark chapters, annotate text and use instant searches to find specific words or phrases. Several eBook files would fit on even a small laptop or PDA. NEW: Save money by eSubscribing: cheap, online access to any eBook for as long as you need it.

Annual subscription packages We now offer special low-cost bulk subscriptions to packages of eBooks in certain subject areas. These are available to libraries or to individuals. For more information please contact [email protected] We’re continually developing the eBook concept, so keep up to date by visiting the website.

www.eBookstore.tandf.co.uk