Early Trilingualism: A Focus on Questions 9781853598555

Citation preview

Early Trilingualism

SECOND LANGUAGE ACQUISITION Series Editor: Professor David Singleton, Trinity College, Dublin, Ireland This new series will bring together titles dealing with a variety of aspects of language acquisition and processing in situations where a language or languages other than the native language is involved. Second language will thus be interpreted in its broadest possible sense. The volumes included in the series will all in their different ways offer, on the one hand, exposition and discussion of empirical findings and, on the other, some degree of theoretical reflection. In this latter connection, no particular theoretical stance will be privileged in the series; nor will any relevant perspective – sociolinguistic, psycholinguistic, neurolinguistic, etc. – be deemed out of place. The intended readership of the series will be final-year undergraduates working on second language acquisition projects, postgraduate students involved in second language acquisition research, and researchers and teachers in general whose interests include a second language acquisition component. Other Books in the Series Learning to Request in a Second Language: A Study of Child Interlanguage Pragmatics Machiko Achiba Effects of Second Language on the First Vivian Cook (ed.) Age and the Acquisition of English as a Foreign Language María del Pilar García Mayo and Maria Luisa García Lecumberri (eds) Fossilization in Adult Second Language Acquisition ZhaoHong Han Silence in Second Language Learning: A Psychoanalytic Reading Colette A. Granger Age, Accent and Experience in Second Language Acquisition Alene Moyer Studying Speaking to Inform Second Language Learning Diana Boxer and Andrew D. Cohen (eds) Language Acquisition: The Age Factor (2nd Edition) David Singleton and Lisa Ryan Focus on French as a Foreign Language: Multidisciplinary Approaches Jean-Marc Dewaele (ed.) Second Language Writing Systems Vivian Cook and Benedetta Bassetti (eds) Third Language Learners: Pragmatic Production and Awareness Maria Pilar Safont Jordà Artificial Intelligence in Second Language Learning: Raising Error Awareness Marina Dodigovic Studies of Fossilization in Second Language Acquisition ZhaoHong Han and Terence Odlin (eds) Language Learners in Study Abroad Contexts Margaret A. DuFon and Eton Churchill (eds)

For more details of these or any other of our publications, please contact: Multilingual Matters, Frankfurt Lodge, Clevedon Hall, Victoria Road, Clevedon, BS21 7HH, England http://www.multilingual-matters.com

SECOND LANGUAGE ACQUISITION 16 Series Editor: David Singleton, Trinity College, Dublin, Ireland

Early Trilingualism A Focus on Questions Julia D. Barnes

MULTILINGUAL MATTERS LTD Clevedon • Buffalo • Toronto

To Mikel, Jon Ander, Jenny and Aitor

Library of Congress Cataloging in Publication Data Barnes, Julia D. Early Trilingualism: A Focus on Questions / Julia D. Barnes. Second Language Acquisition: 16 Includes bibliographical references and index. 1. Multilingualism in children. 2. Second language acquisition. 3. English language–Interrogative. 4. English language–Acquisition. 5. Children–Spain–País Vasco–Language. 6. País Vasco (Spain)–Languages. I. Title. II. Second Language Acquisition (Clevedon, England): 16. P115.2.B367 2006 306.44'6'09466–dc22 2005014691 British Library Cataloguing in Publication Data A catalogue entry for this book is available from the British Library. ISBN 1-85359-854-2 / EAN 978-1-85359-854-8 (hbk) Multilingual Matters Ltd UK: Frankfurt Lodge, Clevedon Hall, Victoria Road, Clevedon BS21 7HH. USA: UTP, 2250 Military Road, Tonawanda, NY 14150, USA. Canada: UTP, 5201 Dufferin Street, North York, Ontario M3H 5T8, Canada. Copyright © 2006 Julia D. Barnes All rights reserved. No part of this work may be reproduced in any form or by any means without permission in writing from the publisher. Typeset by Wordworks Ltd. Printed and bound in Great Britain by MPG Books Ltd.

Color profile: Disabled Composite Default screen

Contents Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Part 1: Theoretical Perspectives 1 Bilingual and Trilingual Acquisition . . . . . . . . . . . . . . . . . . . 9 Definition of Early Bilingualism and Early Trilingualism . . . . . 9 Language Separation in Bilingual Acquisition. . . . . . . . . . . 11 Language Choice and Cross-linguistic Influence in Early Bilingualism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Metalinguistic Awareness in Bilinguals . . . . . . . . . . . . . . 25 Pragmatics in Bilinguals . . . . . . . . . . . . . . . . . . . . . . . 26 Trilingual Acquisition . . . . . . . . . . . . . . . . . . . . . . . . 28 Early Trilingualism . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2 The Development of Interrogative Behaviour . . . . . . . . . . . Question Form. . . . . . . . . . . . . . . . . . . . . . . . . . . Studies on the Acquisition of Question Form . . . . . . . . . Comprehension of Questions and Input . . . . . . . . . . . . Pragmatic Aspects of Children’s Acquisition of Question Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Acquisition of the Functional Use of Questions Between the Ages of Two and Four . . . . . . . . . . . . . . . . . . . Interaction and the Acquisition of Question Function . . . . Interrogative Behaviour in Bilinguals. . . . . . . . . . . . . . The Acquisition of Questions in Basque . . . . . . . . . . . . The Acquisition of Questions in Spanish . . . . . . . . . . . .

. . . .

. . . .

37 37 39 49

. . 52 . . . . .

. . . . .

56 65 73 79 82

Part 2: The Acquisition of English Question Form and Function in a Trilingual Child 3 Research Questions and Design . . . . . . . . . . . . . . . . . . . . . 89 Specific Research Issues . . . . . . . . . . . . . . . . . . . . . . . 89 The Subject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

v 1 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:06

Color profile: Disabled Composite Default screen

Early Trilingualism

vi

Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Analysis of Language Development . . . . . . . . . . . . . . . . 95 Question Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Mother’s Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Cross-linguistic Influence . . . . . . . . . . . . . . . . . . . . . . 106 4 The Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Language Development. . . . . . . . . . . . . . . . Formal Development of Interrogative Behaviour . . . . . . Yes/no Questions . . . . . . . . . . . . . . . . . . . . . . . . . Wh-questions . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Questions . . . . . . . . . . . . . . . . . . . . . . . . . Pragmatic Development . . . . . . . . . . . . . . . . . . . . Question Functions as Part of the Child’s Communicative Competence . . . . . . . . . . . . . . . . . . . . . . . . . . The Mother’s Question Functions . . . . . . . . . . . . . . . Mother and Child Feedback to Questions. . . . . . . . . . . Cross-linguistic Influence in the Child’s General English . . Cross-linguistic Influence in Questions . . . . . . . . . . . .

. . . . . . .

. . . . . . .

109 109 114 119 128 149 151

. . . . .

. . . . .

161 166 178 188 198

5 Interpretation of the Findings . . . . . . . . . . . . . General Language Development. . . . . . . . . . Formal Development of Interrogative Behaviour Pragmatic Development . . . . . . . . . . . . . . Cross-linguistic Influence . . . . . . . . . . . . . . Conclusions and Implications . . . . . . . . . . .

. . . . . .

. . . . . .

208 208 209 214 219 226

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 Author Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

2 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 3 changes b.vp 08 December 2005 19:04:36

Color profile: Disabled Composite Default screen

Acknowledgements I have been fortunate to count on the support of numerous people in the preparation of this book, and I apologise for not mentioning everybody. First and foremost, I would like to express my gratitude to Jasone Cenoz for so generously sharing her knowledge and her time. Her interest has motivated me to complete my task, and her wisdom, patience and guidance have enabled me to do so. I would like to thank the Departamento de Filologia Inglesa y Alemana at the University of the Basque Country for starting me off, and to thank Humanitate eta Hezkuntza Zientzien Fakultatea at Mondragón University for their unconditional support in seeing the project through to its end. My thanks to all the colleagues and friends at both institutions who have encouraged me. I have been privileged to discuss aspects of my work with Colin Baker, Andoni Barreña, Maria Jose Ezeizabarrena, Maria Luisa Garcia-Lecumberri, Pilar Garcia Mayo, Fred Genesee, Jurgen Meisel, Arantza Mongelos, Pili Sagasta and David Singleton, and gratefully acknowledge their comments. Needless to say, mistakes are all my own. I would like to thank all the librarians who have helped me, especially Joxe Mari Egaña at HUHEZI, and Imanol Irizar and Ana Oyaga at the HABE library. On the technical side I would like to show my appreciation to the IKTB team at HUHEZI, and to Aitor Alberdi, Maider Bastida and Eider Alustiza for their valuable assistance. To my family and friends in Euskadi, eskerrik asko, for their help and understanding. To my family in England, thanks for their tireless interest and comfort whenever I needed it. I am also indebted to the nuns of Santa Rita and to the other people who provided me with a quiet place to work. Yet, my most heartfelt thanks of all must go to my trilingual family. My little girl Jenny who provided the data for this study, and her elder brothers Jon Ander and Mikel whose earlier talk gave me the inspiration. Without the selfless and unfailing support of my husband Aitor Alberdi this book would never have been possible. I cannot thank you enough for putting up with me and without me.

vii 3 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:02

Color profile: Disabled Composite Default screen

W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:02

Color profile: Disabled Composite Default screen

Introduction Interest in child language acquisition has a long history. Herodotus (circa 485–425BC) reports on an experiment by King Psammetchos, who left two infants in the care of shepherd with orders not to speak to them until they uttered their first word, which was allegedly bekos, the word for bread in the Phrygian language. Further experiments of this type were reportedly carried out by Frederick of Hohenstaufen (1194–1250), Charles IV of France (1294–1328), James IV of Scotland (1473–1513) and the Mogul Emperor Akbar Khan (1542–1605). More scientific studies have been carried out by linguists, psychologists and biologists who have kept diary records of their children’s language development. However, investigators were confined by lack of technology until the early 1960s, when the field was revolutionised by the coming of the tape recorder. The tape recorder enabled researchers, such as Brown (1968, 1973) and his team, to collect language samples from individuals and groups of children that are still used to this day both longitudinally and cross-sectionally. The samples were transcribed by hand and mimeographed for distribution to other colleagues. Technical improvements coincided with new theories about universals and creative hypothesis testing (Chomsky, 1957, 1965), which further motivated efforts in child language research. Technological developments in computers during the 1980s led to the design of systems for the storage and analysis of child language corpora – such as SALT (Systematic Analysis of Language Transcripts) (Miller & Chapman, 1985) and CHILDES (Child Language Data Exchange System) (MacWhinney, 1991) – and research into child language acquisition has gone from strength to strength. Research into the bilingual and trilingual acquisition of language has a somewhat shorter history since few studies were carried out until the late 1970s. The most notable of the early studies are those of Ronjat (1913) on his French/German bilingual son and of Leopold (1939–1949/1970) on his English/German bilingual daughter. In 1978 Volterra and Taeschner published a study on two Italian/ German bilingual girls in which it was claimed that initially the two languages formed one system before eventually becoming differentiated. This issue dominated research in the field of bilingualism throughout the

1 4 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:02

Color profile: Disabled Composite Default screen

Early Trilingualism

2

1980s and into the 1990s as researchers sought evidence of either single or differentiated systems in bilingual subjects (De Houwer, 1990, 1995; Genesee, 1989; Meisel, 1989). It coincided with a realisation that cases of bilingual acquisition were not uncommon and that the study of these cases could provide insights into language not found in monolingual research. Nowadays research into bilingual acquisition is a flourishing field, largely focused on how cross-linguistic evidence can contribute to understanding of the language acquisition process (Slobin, 1985, 1997). It may be that much of what has been learned about bilingualism also applies to trilingualism but, in the absence of studies on trilingualism to back up such claims (Hoffmann, 2001; Hoffmann & Widdicombe, 1999), this is speculation. Since Murrell (1966), there have been only a dozen or so descriptive reports and studies specifically related to the simultaneous acquisition of three or more languages that can shed light on this process. Very recently there has been some growth in studies on the acquisition of a third language (usually English), in relation to educational contexts, as English and multilingualism have become increasingly important as a result of globalisation (Byram & Leman, 1989; Cenoz & Genesee, 1998; Cenoz & Jessner, 2000; Cenoz et al., 2001). However, as yet there is little published work on the increasing number of children who are growing up multilingual at home as a result of increased mobility between countries and more mixed marriages, let alone studies of the children of communities that have been multilingual for centuries in Africa, Asia and elsewhere. The present book should go some way toward broadening our limited knowledge of trilingualism, by reducing the gap in studies on children in whom three languages are developing simultaneously. It examines the case of Jenny, a child in the Basque Country in Spain, who has been brought up with three languages (English, Basque and Spanish) from birth. The book looks at her general development in English, but more specifically examines the acquisition of questions in English from both a formal and functional perspective. These areas have received considerable interest in the past in monolingual studies but they have not been widely examined recently, nor from a bilingual and trilingual perspective. In order to understand trilingualism we need to imagine what is going on inside a brain where the three systems co-exist. In the view of Cook (1995: 94) ‘a single mind with more than one language has a totality that is very different from a mind with a single language’ and Cook terms the language capacities of such a mind ‘multi-competence’. The languages in a multi-competent individual may develop as separate systems, but the commonalities between them cannot be overlooked, and Cook suggests that such minds may have a flexible grammar rather than the single fixed

5 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Introduction

3

grammar suggested by the Universal Grammar approach to language acquisition (Cook, 1995: 96). Multilingual individuals should not therefore be seen as having two (or three) first languages added together but rather as the owners of their personal multi-competent knowledge, which is not measurable in terms of monolingual standards (Hoffmann, 1991). The present study begins from the theoretical standpoint of the autonomous development hypothesis (Meisel, 2001) that the child is developing her languages in separate systems. These systems will be ‘highly similar’ to those of monolinguals (De Houwer, 1995: 244) yet we should beware of seeing a bilingual child as two monolinguals, or a trilingual child as three monolinguals or even three bilinguals (Cook, 1992; Grosjean, 1989). It is simply a case of not comparing like with like. If a trilingual child is viewed holistically in terms of being a multi-competent user as described by Cook (1992, 1995, 1996), s/he is clearly more proficient than a monolingual or bilingual as s/he is processing three languages rather than one or two. However, if three languages are co-existing in close contact it is to be expected that there will be some cross-linguistic influence. Evidence of cross-linguistic influence provides the investigator with insights into the relationship between the three languages in the child’s mind. Likewise, comparison with development in other monolingual, bilingual and trilingual subjects helps to see the similarities and differences that exist (De Houwer, 1995). When we began recording the subject of our study, Jenny, at the age of one year and eleven months, she was making good progress in all three of her languages. However, for a number of reasons we decided to study only her English. Firstly, Jenny had never been to England and her knowledge of English had been acquired solely through her family (mother, siblings and visits from relations) and through exposure to media (books, television and videos). Her contact with English is therefore different from that found in other English studies and makes her case somewhat remarkable as it shows that it is possible to learn a language even if input is limited to the home context. Secondly, of the three languages, English is the one to which Jenny is not exposed in the community, and so its study provides an insight into how a language may be acquired, albeit in special conditions, with a minority language parent. Furthermore, English is of interest since it is a world language and the language most widely found in multilingual contexts. As a result it is also among the most widely reported languages in the literature on monolingual and bilingual acquisition, and so there are plenty of other studies available in English with which to draw comparisons in relation to bilingualism, trilingualism and the development of form and function in questions. By

6 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Early Trilingualism

4

comparing Jenny’s language development in English with that of other children, we may learn more about acquisition in general and trilingualism in particular. However, we should not overlook the fact that our view of Jenny’s development must necessarily be holistic since English, a Germanic language, is only one of three very different languages that the child is acquiring. Spanish is a Romance language, whilst Basque is a non-IndoEuropean language that is unrelated to any other language family. Questions were used by Socrates (469–399BC) to guide his discussions, and they have been widely studied for their powerful role in language and in thought. A review of the literature in the field of child language acquisition suggested that it was time to re-visit the subject of children’s acquisition of questions in English; previous studies on form and pragmatics in questions laid the foundations for the present study of the acquisition of questions within a new context, that of trilingualism. The study of questions is particularly interesting because questions can be analysed as units at both a formal and a pragmatic level. The development of linguistic form is related to the child’s development of thought and desire for knowledge; a child asks because s/he wants to know. Questions are easily identified because of their ‘distinct grammatical form’ (Dore, 1977: 145), yet at the same time they are complex because they are able to incorporate all the elements present in the grammar and lexicon of a language. Questions are a good point from which to observe language development since they constitute a relatively limited area in which progress can really be followed. As Labov and Labov point out in their study of questions: The notion that children learn syntactic regularities with great ease and rapidity does not find support in (the) data ... The problem then is not to ask, how does the child learn this syntactic rule so quickly, but rather: why does it take so long and so many trials? (Labov & Labov, 1978: 7) The fact that questions are formulated in other ways in Basque and Spanish highlights the variation that exists between Jenny’s three languages, and makes her ability to acquire three systems all the more interesting. It also helps us to identify when there is cross-linguistic influence at a syntactic level. Questions are also a rich field in which to find evidence of pragmatic development, since their use is closely connected with language function. A child may ask a question in order to obtain real information about the world; s/he may equally use a question, the answer to which s/he already knows, to initiate or maintain a conversation. Likewise, a question form may be used pragmatically to request an object or service, to ask permission

7 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Introduction

5

or to seek clarification. Furthermore, use of requests has been shown to be an area in which inter-linguistic transfer at a pragmatic level can take place (Blum-Kulka, 1991) and where influence from Jenny’s other two languages might be expected. Lastly, the study of questions is particularly suitable as a specific area through which to examine the English of a trilingual child interacting with her mother. Tyack and Ingram (1977: 214) have noted that ‘children probably direct more questions to a parent than to an intermittent outside visitor’ and Przetacznik-Gierowska and Ligeza (1990) consider that parent– child questioning interactions fulfil an important role in the development of sociocommunicative skills. Therefore methodological considerations relating to the use of a case study for longitudinal studies of child language development led us to choose interrogative behaviour as a focus for our work. This choice also permits us to study the development of Jenny’s communicative competence in relation to a specific area (Hymes 1972). The aim of the study is to examine whether the child is developing her English and interrogative behaviour in a way that is similar to other monolingual and bilingual subjects even though in her case it is a third language. Where her development differs from that of other monolingual children, we posit that this must be the result of either idiosyncratic development or cross-linguistic influence. By explaining those areas that are related to cross-linguistic influence, we learn more about bilingual and trilingual acquisition. The study also analyses the correspondence between the child’s questioning behaviour and that of her mother, in relation to the use of form, function and the feedback that is given to questions. In this way, the research addresses real communication since it examines not only linguistic competence but also communicative competence and sociolinguistic competence. The question and feedback data are initially classified according to function and speaker, hence it is from a pragmatic perspective that we carry out the subsequent analysis of question form. The description of the theoretical background to the investigation, its implementation and its findings are organised as follows. Part 1 deals with theoretical perspectives. Chapter 1 reviews the literature relating to the study of bilingual and trilingual acquisition. Chapter 2 describes the development of interrogative behaviour according to the literature available on the following: question form, the comprehension of questions and input, pragmatic aspects of question requests, interaction and the acquisition of questions in bilinguals, and question form in Basque and Spanish. The second part of the book deals with the study itself. Chapter 3 explains the methodology used in the study and the five research issues on

8 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Early Trilingualism

6

which it is based. Chapter 4 begins by describing Jenny’s general development in English, then outlines her acquisition of question form during the period under study before going on to discuss pragmatic development and to examine Jenny’s use of question functions and relate this to her communicative competence. It also describes the functions of her mother’s questions and examines the feedback of both the mother and child before dealing with cross-linguistic influence at a lexical level in Jenny’s English in general and at a syntactic level in her questions. Chapter 5 offers an interpretation of the findings in relation to each of the research issues together with some conclusions and future perspectives for further research

9 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Part 1

Theoretical Perspectives

10 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Chapter 1

Bilingual and Trilingual Acquisition Owing to lack of research findings, the phenomena of trilingual and multilingual language acquisition are often explained from the point of view of what is known about early bilingual acquisition. Hoffmann and Widdicombe (1999: 52) point out that ‘in the absence of any theoretical underpinning of our understanding of trilingualism as a distinct linguistic configuration, most of us are working within the theoretical framework of bilingualism’. In many cases trilingualism is explained as an extension of bilingualism (Hoffmann, 2001). On the other hand, Jessner (1997) views multilingualism holistically, proposing a framework (the Dynamic Model of Multilingualism) which stresses the multi-competence of a bilingual or multilingual speaker and in which languages may vary in proficiency according to an individual’s needs. Studies of children acquiring more than one language may cover different periods but a number of them commence at around the age of two (e.g. Deuchar & Quay, 2000: 10–11). Meisel (2001: 19) comments that he starts his analyses at ‘the point of development where children start using multiword utterances containing verbal elements, usually at around age 1;10,1 approximately at MLU2 1.75’. Such studies tend to be longitudinal case studies on a range of differing subjects and languages and employ varied methods of data collection and analysis. This variety has, in many cases, made it difficult to compare the results – see De Houwer (1990) for a detailed discussion of the problems. Studies that focus on acquisition before age 1;11 or beyond age 3;6 or on language combinations that are not relevant have therefore been omitted from the present discussion.

Definition of Early Bilingualism and Early Trilingualism McLaughlin (1978, 1984) makes a distinction between simultaneous and successive acquisition of two languages. Simultaneous acquisition occurs when a child has been exposed to two languages before the age of three, and successive acquisition occurs when exposure to the second language has taken place after the age of three. This is also the view of Taeschner (1983). Padilla and Lindholm (1984), however, consider that simultaneous acquisition refers to exposure to two languages from birth.

9 11 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Early Trilingualism

10

For a number of reasons, De Houwer (1990) rejects the term simultaneous acquisition and proposes Meisel’s (1989) term ‘bilingual first language acquisition’. According to De Houwer (1995: 222) bilingual language acquisition or infant bilingualism refers to ‘the result of the very early, simultaneous, regular and continued exposure to more than one language’. De Houwer distinguishes between bilingual first language acquisition (Meisel, 1989) – also termed ‘simultaneous development of bilingualism’ or 2L1 (Meisel, 2001) – in which a child is exposed to two languages from birth, and ‘bilingual second language acquisition’ De Houwer, 1987) in which exposure to a second language takes place between the age of one month and the age of two. De Houwer (1990: 4) considers that it is vital to know the exact time of first exposure to more than one language since ‘although it is quite possible that differences in ... exposure patterns are irrelevant to the acquisition process ... (they) might well be important’. Starting from birth also allows for full comparison with the monolingual acquisition process. De Houwer (1990) maintains that exposure to a second language at the age of one month is too late for acquisition to qualify as bilingual first language acquisition. Quay (2001: 179) in her study on a trilingual child argues that there is ‘no empirical support for the one-month-after-birth distinction’ for Bilingual First Language Acquisition proposed by De Houwer (1990, 1995). Quay claims that the results of her study provide evidence that exposure to three languages before first words are produced in them may also be termed ‘trilingual first language acquisition’. However, following McLaughlin (1978, 1984) and his interpretation of simultaneous bilingualism as occurring before age three, we suggest ‘early trilingualism’ as a broader term. Some children are exposed to three or more languages from birth, even though this does not occur as often as early bilingualism. Early multilingualism is not uncommon in some parts of the world (such as Asia and Africa), and is becoming increasingly frequent in Europe and elsewhere as a result of greater population mobility and international communications (Cenoz, 2000: 43). In early multilingualism, different patterns of language input and use can affect a child’s developing languages. Cenoz (2000: 40, 43) describes the acquisition order of the languages in early trilingualism in a formula Lx + Ly + Lz, in which contact with the different languages takes place simultaneously but in different situations. Numerous situational variations are possible (see Arnberg, 1987; Barron-Hauwaert, 2004; Cenoz, 2000; Harding & Riley, 1986; Hoffmann, 2001), but a common one is when each parent speaks just one language to the child following the principle of ‘one-parent-one-language’ and a third or fourth language is used in the community or in schooling. The language that the parents choose to speak

12 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

11

to each other is also crucial in determining the child’s language patterns. It may be a factor in the degree of trilingualism shown by the child since if, for example, the parents use a non-community language to each other this will be an additional source of input to the child. Other acquisition orders are possible if the child is exposed to the different languages at different times. A child might simultaneously acquire two languages at home with a third added later, for example on entering kindergarten; this can be represented in the following formula (Lx/Ly → L3) (Cenoz, 2000: 40). If four languages are involved, the complexity of possible combinations is even greater.

Language Separation in Bilingual Acquisition One system or two An area of much interest in the field of bilingual language acquisition is that of language separation, or whether the bilingual child develops one or more language systems for his or her languages. Volterra and Taeschner (1978) and Taeschner (1983) proposed a three-stage model of bilingual development that argues that initially the languages (in this case, German and Italian) are mixed and that gradually the languages form individual lexical and syntactic systems. The model states that: (1) The child has one lexical system that includes words from both languages. (2) The child distinguishes two different lexicons, but applies the same syntactic rules to both languages. (3) The child has two linguistic codes, differentiated in both lexicon and in syntax, but each language is exclusively associated with the person using that language. Several early studies support the one-system hypothesis. Redlinger and Park (1980) describe the language mixing of 4 two-year-old children, all of whom have German fathers and whose mothers are French, English and Spanish. They provide evidence that nouns are most frequently substituted, and that function words tend to be more frequently substituted than content words, and use this evidence in support of the one-system hypothesis. Vihman (1985) examined the acquisition of English and Estonian between the ages of 1;1 and 2;00, and focused on aspects of language mixing, particularly functors. She recognises that there is a relationship between early language mixing and code-switching in mature bilinguals and claims that the languages showed evidence of an initial single system. However, this evidence has been questioned by Pye (1986) and by Genesee (1989).

13 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Early Trilingualism

12

Criticism of the one-system hypothesis The ‘single system hypothesis’ has sparked academic debate particularly in relation to Stages I and II (De Houwer, 1990; Genesee, 1989; Meisel, 1989). One major criticism has been on methodological grounds, relating to data collection and the size of the corpus on which the hypotheses are based (De Houwer 1995: 231). De Houwer (1995: 231) argues that the claim (Volterra & Taeschner, 1978) that a single system exists depends on the ‘absence of certain forms rather than their presence’. This stage is also explained by Meisel (2001: 15) in terms of a fusion hypothesis whereby ‘children create a new system (randomly?)3 combining elements of the two or more systems’. It is claimed (Taeschner, 1983) that in Stage I the lexical items used by the child have no equivalence in the other language, so one undifferentiated system of words is used for both languages. Quay’s re-analysis (1995) of the data from Volterra and Taeschner (1978) and Taeschner (1983) seems to disprove the original theory by revealing instances of equivalent pairs. Quay (1995) describes the development of language choice and the bilingual lexicon in a Spanish/English infant bilingual from birth until age 1;10, and finds that translation equivalents make language choice possible from the beginning of speech. This contradicts the findings of Volterra and Taeschner (1978) and shows that in Quay’s own Spanish/English data a number of cross-linguistic equivalents were present. Jekat (1985) carried out a similar study to that of Taeschner (1983) using German/French subjects and was unable to decide whether or not equivalent pairs were present. Hayashi (1992) did not find conclusive evidence for the Stage I theory in Danish/English subjects, and suggests that the input may be responsible for the appearance or otherwise of equivalents. Arguments against Stage I have also been put forward by Genesee (1989), Mikès (1990), De Houwer (1990), Genesee et al. (1995), Nicoladis and Genesee (1996a) and Köppe (1997), at least for once grammatical differentiation has taken place. Meisel (2001: 23), despite also rejecting the hypothesis, suggests that a single system comprising elements that are from neither language may be an option at a pre-grammatical stage – if such a stage exists. Stage II of the single-system hypothesis (Volterra & Taeschner, 1978), which claims that morpho-syntactic rules also comprise a single system at their emergence, has also been criticised for inconsistencies and has become a source of controversy. Meisel and Mahlau (1988), like Quay (1995) above, consider that the evidence produced by Volterra and Taeschner (1978) contradicts their own arguments and in fact supports the idea of early bilingual separate development (see De Houwer, 1995: 233–234).

14 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

13

More recently, Deuchar and Quay (1998) re-examine the issue, arguing that there is a single rudimentary grammar in bilingual children before two years of age, the approximate time around which the linguistic systems become differentiated (Paradis & Genesee, 1996). Based on evidence taken from a Spanish/English bilingual infant (see also Deuchar, 1989) Deuchar and Quay argue that between the ages of 1;7 and 1;9 mixed utterances can ‘be attributed largely to limited lexical resources rather than to an initial syntactic system’ (Deuchar & Quay, 1998: 233). In their view, languagespecific utterances can be identified as soon as morphological marking appears and they find evidence of this in samples from the same child at 1;8 to 2;3. Deuchar (1999) reports on the development of the ability to make language choices in Spanish and English contexts. She examines mixed utterances in the same Spanish/English bilingual child between the ages of 1;7 and 1;9, which correspond to the child’s first two-word utterances. The study focuses specifically on function words, and finds that they are not language-specific in early bilingual two-word utterances. Content words, on the other hand, match the language context more frequently, and are therefore more language-specific. Deuchar and Quay (1998) and Deuchar (1999) claim that their results show that early bilingual children have one initial morpho-syntactic system prior to differentiation. Meisel (2001: 25) questions the findings of these two studies on some methodological grounds and the way rudimentary syntax is defined. In his view (Meisel, 2001: 26), more empirical and theoretical support is needed to support Deuchar and Quay’s position. Evidence of two systems In seeking evidence of two systems (Ingram, 1981; Meisel & Mahlau, 1988), it has been suggested that the best way to investigate how children acquire their language systems, and whether these systems are separate, is to focus on structural aspects where the adult versions of the two languages differ. Once these have been established, it is possible to follow the acquisition of these areas by monolingual children and compare it to the progress of the bilinguals in the languages concerned. By this means, potential differences in bilingual acquisition will show up. However, it may then be hard to distinguish non-adult-like forms as developmental or not (De Houwer, 1995: 236). De Houwer suggests that evidence for two systems may lie within the data from developing phonological systems such as those described by Ingram (1981), Deuchar (1989) and Deuchar and Clark (1996). De Houwer and Meisel (1996) propose using empirical evidence relating to ‘acquisitional problems’ that conform to specific requirements in order to decide whether or not child bilinguals have access to separate

15 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Early Trilingualism

14

linguistic systems. A number of criteria that this evidence must fulfil are specified (De Houwer & Meisel, 1996): (1) The phenomena investigated should be functionally equivalent in both target languages, but they should exhibit clear differences in their respective structural properties. (2) This requirement must also be met by the respective monolingual child languages, which may be structurally identical where the adult languages differ. (3) Data claimed to constitute evidence in favour of or against a onesystem phase must stem from an age period during which the children have access to grammatical knowledge. Meisel (2001) presents evidence from French/German and Spanish/ Basque bilinguals to support the Differentiation Hypothesis proposed by De Houwer and Meisel (1996) based on such an analysis. Bergman (1976) was amongst the first to suggest the possibility of two systems and to define an Independent Development Hypothesis in her own Spanish/English bilingual child to whom she usually spoke English. Bergman disagrees with the view that there is a stage of ‘mish-mash language’ in which the child may be trying to form one system out of two. Instead, Bergman produces evidence to show that the child is able to differentiate between the two languages which suggests that she is developing two systems (Bergman, 1976: 89). Drawing on evidence from mixing in the child’s developing possessive system which can be traced back to the input, Bergman posits that: each language will develop independently of the other, reflecting the acquisition of that language by monolingual children, unless it is the case that the lines between the two languages are not clearly drawn in the linguistic environment of the child. (Bergman, 1976: 94) Studies since Volterra and Taeschner (1978) have endeavoured to show that language development takes place through separate systems in children who are bilingual from birth and whose input conditions follow the principle of one-person-one-language (see De Houwer, 1995: 237 for a detailed list of these studies). De Houwer (1990) went further than Bergman (1976) in her study of a slightly older bilingual Dutch/English child (2;7–3;4) and proposed a Separate Development Hypothesis based on evidence of separate development in morpho-syntactic categories. There is little evidence of ‘mish-mash’ in De Houwer’s data since the child is already developing in a way similar to monolinguals in both her languages at the time of the study. De Houwer shows that there is little interaction between the two languages

16 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

15

at a morpho-syntactic level in her subject and that lexical mixing is also infrequent. Furthermore the child is able to make appropriate language choices with different interlocutors (monolingual Dutch or English) and can code switch with those interlocutors who are bilingual. Most studies carried out to show evidence of two systems have followed the development of specific morpho-syntactic areas in both the child’s languages. Meisel (1989) and Deuchar (1989) examined word order and found that children use language-specific word ordering at the stage of two-word utterances and complex sentences. Idiazabal (1988, 1991), De Houwer (1990) and Deuchar (1992) looked at different aspects of bound morphology and found that these develop in a language-specific manner. Subject–verb agreement was studied by Deuchar (1992) and Meisel (1989), who found that the subject–verb agreement rules of one language developed independently of the other. Other areas such as case-marking (Meisel, 1986), clitics (Meisel, 1990), tense and aspect (Schlyter, 1990) and indirect objects (De Houwer, 1990) have also been studied and confirm the Separate Development Hypothesis whereby ‘the morpho-syntactic development of a pre-school child regularly exposed to two languages from birth which are presented in a separate manner proceeds in a separate fashion for both languages’ (De Houwer, 1990: 339). Pérez-Vidal (1995) described the language development in English of a Catalan/English bilingual child between 3;2.28 and 4;2.10 who was also exposed to some Spanish. A three stage in-depth analysis based on a longitudinal study model by Fletcher (1985) was carried out on the child’s English. Pérez-Vidal found that, although the child’s production in English was slightly later to emerge than reported in other studies (e.g. De Houwer, 1990), it coincides with them in the way that it initially contains a large number of mixes later almost disappear (Pérez-Vidal, 1995: 365). JuanGarau (1996) studies the same child in Catalan and English contexts between the ages of 1;3 and 2;11. Paradis and Genesee (1996) sought evidence of transfer, acceleration and delay in three subjects with a French father and an English mother at the ages of 2;0, 2;6 and 3;0. They found that acquisition of finiteness, negation and pronominal subjects followed the same pattern as in monolinguals, and concluded that the subjects were acquiring English and French separately and autonomously, albeit at different rates. Juan-Garau and Pérez-Vidal (2000) report on use of the null subject in the same English/Catalan bilingual child. Catalan is a null subject language whilst English requires an overt subject.4 The child was shown to use the null subject in Catalan and practically no evidence was found of use of the

17 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:03

Color profile: Disabled Composite Default screen

16

Early Trilingualism

null subject in the child’s English, proving that the two languages are developing separately. It is now generally accepted that children can differentiate grammatical systems ‘as soon as language-specific word order properties and inflectional morphology emerge’ (Meisel, 2001: 17), though what happens prior to this needs further investigation. The possible theoretical options available have been summed up by De Houwer and Meisel (1996) in a series of hypotheses: (1) Fusion Hypothesis: the child creates a new system that combines elements of the two or more systems. (2) Differentiation Hypothesis: the children differentiate the two systems as soon as they have access to grammatical knowledge. (3) Interdependent Development Hypothesis: one of the languages serves as a developmental guide for the other. (4) Autonomous Development Hypothesis: the acquisition of each of the two languages by the bilingual individuals follows the same developmental logic that guides the acquisition of the respective languages by monolingual children (Meisel, 2001: 15–16). In general the Fusion and Interdependent Development Hypotheses are today rejected in favour of the Differentiation and Autonomous Development Hypotheses (Meisel, 2001).

Language Choice and Cross-linguistic Influence in Early Bilingualism Although the languages of a bilingual child are now widely considered to develop separately, the relationship between them provides evidence of the process of bilingual language acquisition. This relationship falls into two main areas, language choice and cross-linguistic influence, which are examined below. As in previous sections, most of the literature available reports on cases of bilingual acquisition. Specific aspects relating to trilingual acquisition will be explained later in the chapter. Language choice An issue that has come to the fore in recent years is that of language mode (Grosjean, 1994, 1997). Bilinguals find themselves at different points along a language mode continuum depending on whom they are talking to. At one end of the continuum, and if a bilingual is conversing with a monolingual, he or she is processing only one language and is therefore in monolingual mode. At the other end of the continuum and in a bilingual

18 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:06

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

17

context the bilingual may be processing two languages. The language that is being spoken at a given moment is referred to as the host or base language, while the other language, which is de-activated temporarily (e.g. until a code-switch appears), is referred to as the guest language. Factors that influence language mode in adult bilinguals include the participants and their language mixing habits, the situation, the form and content of the interaction and the function of the language. In the case of infant bilinguals, Lanza (1992, 1997) has found evidence of language mode where children adjust their language mixes according to whether the interlocutor demonstrates monolingual or bilingual language behaviour. Lanza (1997) examined the question of language mixing in the simultaneous acquisition of English and Norwegian in 2 two-year-old children from a sociolinguistic perspective and within the framework of the one-system and two-system hypotheses. She found that the children were able to mix languages in a context where it was socially appropriate and did not mix them when this was the most socially appropriate behaviour – thus showing that two systems were clearly in place (Lanza, 1997: 319). Furthermore Lanza supports the claims of the proponents of the two-system hypothesis that the type of input largely defines the amount and type of mixing that occurs. Lanza also argues that the discourse strategies used by the parents to negotiate a monolingual or bilingual context will affect the degree to which the child mixes languages. If the parent uses only one language to the child and only accepts or appears to understand utterances from the child in that language, a monolingual context is negotiated. If, on the other hand, the parent shows understanding, even without use of elements from the second language, then there is a bilingual context and mixing will occur. Grosjean (1998: 176) points out that in some research studies designed to describe bilingual children’s production in either of their languages the researchers, caregivers and assistants may themselves be bilinguals and that data is therefore collected, ‘however involuntarily’ in a bilingual rather than monolingual mode. It may be that bilinguals never totally de-activate the other language in monolingual mode (Hulk & Van der Linden, 1998: 177) and that claims that some bilingual children develop their two languages like monolinguals cannot be totally substantiated. Hulk and Van de Linden (1998: 177) ‘do not think that bilingual children are subject to direct transfer from one language into the other (but that) indirect influence from one language on the other plays a role’. They point out that in bilingual first language acquisition there is not a single first language on which to build, so the nature of transfer is necessarily different and should be termed not ‘transfer’ but

19 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Early Trilingualism

18

cross-linguistic influence. Such ‘influence’ might include the use of strategies to deal with a reduced input in either language in comparison to monolingual infants, exposure to a wider range of syntactic possibilities than the monolingual child and that areas acquired in one language may then be adapted into the other. In order to understand the child’s linguistic development, it is necessary to consider the input that the child has been exposed to in each language. Snow makes the point that: most children clearly receive more input than is strictly necessary to support normal language acquisition, as is shown by the fact that input can be distributed over two or three languages with the result that the child is a fully bilingual or trilingual speaker. (Snow, 1995: 187) The amount and frequency of input per language in the context of the child’s linguistic microcosm (De Houwer, 1990) of family, caregivers and friends are more important than the languages that are present in the community at large with which the child may have little contact (De Houwer, 1995: 224). Within the microcosm, individuals may vary in the way they separate the languages or not. The input condition known as one-person-one-language is seen as being the most successful at promoting the development of each language separately in the child (Bain & Yu 1980; Döpke, 1992; Lebrun & Paradis, 1984;). This principle is referred to as the rule of Grammont (Ronjat, 1913). Clyne (1987) terms it the ‘interlocutor principle’ according to which the languages produced by the child are predicted to be free of mixing and native-like. However, García (1983) has shown that Spanish/English speaking children who did not receive their input in this way were also able to use the languages as separate systems. De Houwer (1995: 225) describes an ‘input continuum’ along which some individuals may only ever address the child in one language, while others may always use more than one. It is claimed (Idiazabal, 1984; Goodz, 1989) that total separation of the languages spoken to the child is never possible. Goodz (1989: 41), in a study of four bilingual families with children aged between 14 and 28 months, finds that ‘despite strong avowals to the contrary, no parents were able to maintain a strict parent-language separation’. She puts this down to a number of factors. (1) Language switches are sometimes used, both by parents and children, either to attract attention or for emphasis. (2) As with monolingual first language acquisition, mothers in particular have a strong urge to communicate with their children, and may do so

20 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

19

by whatever means are available even if this means using words from the language of the other parent. (3) Adult expansions and repetitions of the child’s utterances take on ‘their own unique form’ in bilingual families where utterances may switch language in repetition and expansion on a mixed utterance typically leads to further language mix. According to Goodz: The motivation here seems to be to encourage language behaviour regardless of its form and is similar to the tolerance shown by monolingual parents for a variety of syntactic and phonological errors exhibited by their children at this age. (Goodz 1989: 42) The languages (and their mixes) through which the members of the linguistic microcosm address each other in the presence of the child may also be a factor to consider in the development of the child’s languages. Bergman (1976) and Lindholm and Padilla (1978) claim that language mixing in the input will produce language mixing in the child’s output, albeit through two different linguistic systems. Meisel (1989: 15) suggests that factors such as mixing in the linguistic environment, dominance in one language and social-psychological factors may all lead to mixing. According to Genesee (1989) if children are exposed to mixing in their input there seems to be no reason for them to know that their own speech ought not to be mixed. In the view of Quay (2001) the separation of input according to person is not absolutely necessary for bilingual development, and may even be impractical from a social and pragmatic point of view in situations where others are present who share only one of the child’s languages. This is consistent with the findings of Deuchar and Quay (2000) whose subject M receives mixed input in Spanish (at home) and English (on her mother’s university campus) from her mother. Her father mainly uses Spanish with the child, but also uses English. The child is ‘exposed to each parent as a speaker of two languages and as appropriate models of bilingual behaviour’ (Deuchar & Quay, 2000: 7). The child became bilingual in both languages. De Houwer (1990: 54) expresses the view that more evidence is needed to back up the question as to whether or not language presentation affects bilingual development. Research in the field of parental discourse strategies has shed some light on the matter. There is evidence that not only the amount and quality of the input, in terms of language mixing, are important, but also the discourse strategies used by parents and caregivers in

21 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Early Trilingualism

20

interaction. Döpke (1992) studied six German/English bilingual children (recorded at ages 2;4, 2;8 and 3;2) and their families in Australia and found that they were more likely to use German (the minority language), when the interaction between the German-speaking parent and the child was more child centred. Lanza (1997) examined the code-mixing behaviour of two two-year-old bilingual Norwegian/English children and concluded that within discourse the parental acceptance or not of utterances in both languages influenced the child’s use of mixing. This unconscious acceptance or rejection of language mixing in discourse by parents is referred to as negotiating a monolingual or bilingual context (Lanza, 1992, 1997, 2001). Kasuya (1998) in a study of four Japanese/English bilingual families found that when preference for the use of Japanese was made explicit to the child this influenced the child’s choice of Japanese for subsequent interactions. Nicoladis and Genesee (1998) carried out a similar qualitative study to those of Lanza (1992, 1997) on five French/English children (aged 23–29 months). Whilst accepting that children’s code-mixing may be affected by parental discourse strategies (Parental Discourse Hypothesis) their study found that there was no statistical evidence to support this claim. Nicoladis and Genesee (1998) suggest that differences in the sociolinguistic contexts of the two studies may account for the different results, and note that the failure of their subjects to respond to subtle parental strategies for language change (or not) was in line with findings that bilingual children do not see language as a potential source of communication breakdown until around the age of three (Comeau et al., 1997). It is not yet clear to what extent mixing in input and parental discourse strategies are responsible for mixing or lack of mixing in the language produced by bilingual children. However, it seems that sociolinguistic factors relating to the contexts of parental language use are involved, along with the pragmatic ability of bilingual children to make judgements about switching language based on subtle signals from the interlocutor (Nicoladis & Genesee, 1998). Adult bilinguals use code-switching (using different languages within a stretch of discourse) depending on different communicative and sociolinguistic contexts (Poplack, 1981). Children have also been observed to do the same depending on the interlocutor and the input that the child receives from the interlocutor. If a habitual interlocutor addresses a child in only one language, it is likely that the child will speak to that person in that language (De Houwer, 1990; Kielhöfer, 1987). However, if the interlocutor habitually uses mixed codes with the child, it is likely that the child will do the same (Bentahila & Davies, 1994; De Houwer, 1990; Genesee, 1989; Goodz, 1989; Nicoladis & Genesee 1996b; Saunders, 1988). It seems that

22 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

21

children are able to distinguish between those addressees they view as monolingual and those that are bilingual. It has also been shown that children may select a code or codes based on their knowledge of the amount of mixing to be expected in that code (Marcon & Coon, 1983; Pfaff, 1991). It has been suggested that it is the parental choice of codes that will encourage or discourage the use of one particular code or a mix of both (Lanza, 1990, 1992; Quay, 1995). Lanza (1992: 636) proposes that a ‘notion of bilingual awareness’ allows the child to know, depending on the context, the appropriateness of when to mix languages and when to keep them separate. This is not unlike the ‘language mode’ proposed by Grosjean (1994, 1997) whereby the bilingual’s languages and language-processing mechanisms will be in different states of activation at certain points in time. Grosjean (2001) suggests that for adults there are a number of factors that influence ‘language mode’ such as language proficiency, usual mode of interaction, kinship, physical location, the presence of monolinguals, and so on. A child’s ability to choose language according to interlocutor’s language is evidence that two separate systems exist, and has been reported in a number of different studies where the parents followed (by and large) the one-person-one-language principle. Genesee et al. (1995) found that five French/English bilingual children aged from 1;10 to 2;2 always used more French or English with their respective French- or English-speaking parent. However, Nicoladis and Genesee (1996a) in a study of four French/English bilingual children aged 1;7 to 3;0 found that this ability did not emerge until between 1;9 and 2;4. Likewise Köppe (1997) in a study of three French/ German bilingual children observed that mixing of the two languages decreased at around the same age. Deuchar and Quay (2000) describe how a mother varied language according to context (Spanish in the home context, and English outside) with a Spanish/English bilingual child aged 1;3 to 1;9 finding that the child was able to use the right language for the context. Cross-linguistic influence The utterance is the unit of analysis most frequently found in studies of language mixing in infant bilingualism. Code mixing or mixes at utterance level are also referred to as intrasentential code switches (Poplack, 1980) or mixed utterances (Deuchar, 1999; Lanza, 1992). Syntactic mixing in child bilinguals was studied by Volterra and Taeschner (1978) and Taeschner (1983), and lexical mixing by Swain and Wesche (1975), Redlinger and Park (1980) and Vihman (1985) to support the claim that the child initially has one linguistic system that will eventually form two languages. Lanza (1997:

23 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Early Trilingualism

22

3) suggests that the term ‘mixing’ pre-supposes that two separate systems are present and points out that mixing is a general term for all types of language contact. De Houwer outlines a number of factors that may result in mixed utterances: Relaxed speech production processes ... the existence of a lexical gap in B; the influence of a previous mention of a lexical item in A in the ongoing discourse; the initial learning of a concept in the context of A; the greater frequency of the inserted item in A; the possibility of making a semantic distinction in A that is not available in B; the greater saliency of the item in A as compared to B, and so on. (De Houwer, 1995: 24) It has been found that, with a monolingual speaker, children tend not to mix (Kielhöfer, 1987; Stavans, 1990; Vihman, 1981) except for occasional slips, mostly consisting of nouns. Lanza (1992: 638) describes the type of mixes that may occur in an utterance: • the co-occurrence of two languages within one word; • the co-occurrence of two languages within a group of words. She also describes the types of mixes within a conversational turn consisting of an utterance, or a group of utterances bounded by a pause or an utterance by another participant in the conversation: • • • •

a one-word utterance in the other language; a mixed utterance; a multi-word utterance in either language; a group of utterances (often involving repetition of the same utterance).

Redlinger and Park (1980), Vihman (1985) and Lanza (1992) report that grammatical words (functors) account for a large percentage of mixes in the mixed utterances of young bilingual children. Since these words are rarely mixed in the language of bilingual adults this suggests that this may be a characteristic of infant bilingualism (Lanza, 1992: 639). Lanza (1992) finds that the combinations of grammatical and lexical morphemes that co-occur in multi-word and single-word utterances in a bilingual Norwegian/ English child are indicative of language dominance. Norwegian grammatical morphemes co-occurred with Norwegian and English lexical morphemes, but English grammatical morphemes co-occurred only with English lexical morphemes, indicating dominance in Norwegian. In other words, evidence of grammatical morphemes from language A with lexical morphemes from language B implies that language A is dominant. However, Lanza (1992: 641) notes that ‘mixing as a result of dominance cannot be invoked as evidence for the child’s lack of language separation’.

24 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

23

Gawlitzek-Maiwald and Tracy (1996) have proposed that mixing is evidence of the bilingual child using her expertise in one language to solve problems in the other, and that mixing helps to bridge lexical and structural gaps. In their study of a German/English bilingual child, they show that the two languages develop separately from early on. Additionally they follow her mixed utterances closely as a third strand of development and find that certain types of mixes appear during specific stages of development in the two languages. They conclude that the mixes are not evidence of the child’s linguistic confusion, but rather that they reflect her competence in the languages (Gawlitzek-Maiwald & Tracy, 1996: 920). The process through which the two separate language systems work together complementarily is termed ‘bilingual bootstrapping’. Through it the child ‘pools her resources’ (Gawlitzek-Maiwald & Tracy, 1996: 920), ‘borrows what she needs’ (1996: 921) from either language and ‘language A fulfils a booster function for language B’ (1996: 903). Van der Linden (2000) examines lexical mixing in a bilingual French/ Dutch child between the ages of 2;3 and 3;11. The child has been brought up on a strict one-person-one-language principle. At the beginning of the study lexical insertions from one language into the other are frequent, but then decrease dramatically. Cross-linguistic equivalents are also noted in the same scripts. However this disappears after age 2;5.20 and Van der Linden (2000: 52) suggests that this is evidence of the ‘other’ language becoming ‘properly inhibited under the influence of the strictly separated language input’. Deuchar and Quay (2000) also find in their study of a younger bilingual child that mixing due to lexical gaps disappears once equivalents are available in both languages. Lanza (2000: 237) points out that, whereas in earlier research mixed utterances were seen as evidence of fusion or of a ‘dilemma’ for the bilingual child, they are now viewed as evidence of the child’s resourcefulness. Early differentiation between languages is now generally assumed, and recent work in the field of bilingual research has been concerned with the degree of separation and interaction between a bilingual’s two languages. According to Hulk and Müller: The development of two languages in a bilingual child may be largely autonomous, but this does not exclude the possibility that there can be influence from one language on the other. In such cases we will speak of cross-linguistic influence. (Hulk & Müller, 2000: 227) There has been a recent interest in cross-linguistic influence particularly at a syntactic level. Hulk (2000) looks at the acquisition of syntax in a French/German bilingual child and finds that input plays a crucial role.

25 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Early Trilingualism

24

Hulk (2000: 75) suggests that bilingual children may find input in two languages ambiguous, and that it may be harder for them to ‘disambiguate’ when adult mixing may be providing erroneous cues. Schelletter (2000) describes negation as a cross-linguistic structure in a German/English bilingual child. She finds that in both languages the child treats negation as an adverbial in a lexical category rather than as a syntactic category. Yip and Matthews (2000) examine the transfer of the Cantonese features of whinterrogatives, null objects and pronominal relatives into the English of an English/Cantonese bilingual boy. Their results show that language dominance, in this case by Cantonese, is the main factor in determining the direction of language transfer and that ambiguity in the input may also play a role. Juan-Garau and Pérez-Vidal (2000) find that a Catalan/English bilingual child acquires null subjects in Catalan but not in English. Döpke (2000: 5) makes the point that cross-linguistic structures are children’s own creations since ‘the children could not possibly have heard them from their interlocutors’. Tracy (2000: 16) focuses on some of the ambiguities found in examples of cross-linguistic influence. She points to the difficulty of assigning identical words to the correct language. Such words may even include proper names, since it is not clear from which language they come. Tracy also refers to the uncertainty of phonetically distinguishing words that are similar in sound or that may be subject to instability during the developmental period. Lanza (2000: 239) also mentions the problems of interpreting some novel verb forms. Several different methodological considerations have emerged in relation to the study of cross-linguistic influence. Following Myers-Scotton’s (1993) Matrix-Language Frame Model (MLF model) the two languages in a code-mixed utterance take different roles. The base language of the utterance is referred to as the ‘matrix language’ (the terms ‘host’ or ‘recipient’ are also used), whilst the language whose elements are inserted into it is termed the ‘embedded language’. The model makes a distinction between system and content morphemes that is close to the functional and lexical morphemes distinction of earlier studies. It describes an integrated set of constraints, and the child’s adherence (or not) to these constraints permits the examination of mixing behaviour over time. A further development is the comparison of the language of developing bilinguals with that of adult bilinguals. Paradis, Nicoladis and Genesee (2000) carried out a study using the MLF model in which they examine whether bilingual French/English children have the same code-switching constraints as adult bilinguals. They found that the children were able to adhere to adult-like structural constraints in their code-mixing. Tracy (2000) also relates child code-switching behaviour to that found in adults.

26 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

25

Developmental comparisons with monolingual children have always formed part of the methodology of bilingual acquisition research (De Houwer, 1990; Pérez-Vidal, 1995). Paradis (2000: 176) explains that ‘in order to investigate whether interaction is taking place in the bilingual acquisition of two languages, it is necessary to draw comparisons with monolingual children’. By comparison with adult and child monolinguals, it is possible to identify the extent to which the bilingual child shows cross-linguistic influence from either language. However, Paradis (2000: 176) also points out that ‘what may seem like cross-linguistic influence could be nothing more than a typical stage in monolingual development’. Furthermore, it should be noted that monolinguals also produce non-standard forms (Döpke, 2000: 78; Tracy, 2000: 24).

Metalinguistic Awareness in Bilinguals Metalinguistic awareness can be defined as ‘the ability to think about and reflect upon the nature and functions of language’ (Pratt & Grieve, 1984). Studies of school-age bilingual children have shown that bilinguals have more developed metalinguistic awareness than monolinguals (Ben Zeev, 1977; Cummins, 1976; Hakuta 1986; Lasagabaster, 1997). Clark (1978) describes a number of studies (e.g. Bates, 1976; Shatz & Gelman, 1973) on language awareness in monolingual English children as young as two that show that these children are capable of making judgements on language. With reference to his trilingual daughter, Slobin (1978: 45) lists aspects of language awareness that appear between the ages of two and six: • self-corrections and re-phrasings in the course of ongoing speech; • comments on the speech of others (pronunciation, dialect, language, meaning, appropriateness, style, volume, etc.); • explicit questions about speech and language; • comments on own speech and language; • response to direct questions about language. De Houwer (1990: 310), in her study of a bilingual child aged 2;7 to 3;4, argues that metalinguistic behaviour is related to general metacognitive development and therefore will accompany developmental changes in the child’s linguistic production. Furthermore she claims that these changes, since they are part of a general process, will not be specific to either of the child’s two languages. Signs of such metalinguistic behaviour include ‘spontaneous (or self-initiated) repairs, elicited (or other-initiated) repairs,

27 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Early Trilingualism

26

sound-play, hesitations, self-repetitions and explicit metalinguistic statements’ (De Houwer, 1990: 310). Lanza (1997: 65) suggests that ‘if we extend the notion of metalinguistic awareness to inexplicit reflection of language, we may be able to trace the young bilingual child’s evolving awareness of bilingualism’. General metalinguistic skills such as adjusting speech to the status of the listener (Clark, 1978) take on an added dimension for the bilingual child where they might be used in awareness of when it is appropriate to mix or not or in ‘deciding’ on language choice. In Lanza’s (1997: 67) view ‘the child’s bilingual awareness is part of his or her communicative competence (and) an increase in metalinguistic awareness over time will contribute to the bilingual child’s increasing pragmatic sophistication in various code-switching strategies’ as also noted by Vihman (1985: 317). In sum, monolingual children as young as two appear to be capable of reflecting on language and using skills such as adjusting language to the status of the listener. In the case of bilingual and trilingual children metalinguistic awareness involves extending such capacities in relation to bilingual awareness, for example in deciding which language to speak to whom and the degree of mixing that may used. Furthermore, bilingual children are able to make explicit comments on the nature of language.

Pragmatics in Bilinguals Studies that examine the area of developmental pragmatics with reference to bilingual children approach the issue from the point of view of pragmatic differentiation relating to language mixing and language choice and the perspective of developing communicative competence (Genesee et al., 1996; Meisel, 1994; Nicoladis 1998; Nicoladis & Genesee, 1996;). Meisel (1994) views the ability to code-switch as a pragmatic competence in young bilingual children. He explains that pragmatic differentiation precedes syntactic differentiation and finds evidence that by age 2;6, and possibly earlier, children’s mixing in discourse has similar underlying pragmatic and grammatical principles to those found in adult codeswitching. Nicoladis & Genesee (1996a) found an initial stage before the age of two when bilingual children do not differentiate their languages pragmatically in relation to language choice. They observed four French/English bilingual children between the ages of 1;7 and 3;0 to see if the children used their languages more with the respective parent speaker of that language in spite of themselves having a higher or lower proficiency in that language. The results show that, at the beginning of the period under investigation, none

28 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

27

of the subjects differentiated pragmatically, whereas by later sessions all were able to do so. Genesee, Boivin and Nicoladis (1996) find that children are also able to modify their language use (e.g. through reducing mixing), to suit the language of monolingual strangers. Nicoladis (1998) examined the way a Portuguese/English bilingual child aged 1;0 to 1;6 came to pragmatically differentiate between the two languages in the input he received. The child used more of the parent’s language with each particular parent than would be predicted on the basis of his own language dominance in each language, and before the child showed evidence of lexical differentiation. It is suggested that pragmatic differentiation (use of the appropriate languages for the context) precedes lexical differentiation, and therefore leads to an understanding that translation equivalents belong to two distinct languages. The area of sociopragmatics (Leech, 1983) is of interest in relation to the developing bilingual and trilingual child. Interlanguage pragmatics, for instance, examines the extent to which the pragmatics of one language encroaches on the other(s). The fact that differences exist is demonstrated by the interference across languages between adult learners of second languages when cross-cultural misunderstandings take place (Blum-Kulka et al., 1989). It is claimed that the strategies for conveying speech acts are basically the same across cultures and are therefore universal, although the appropriate uses of the strategies may vary in specific culture/language communities (Fraser, 1985). Wierbizcka (1985), however, argues that differences in speech act use are language-specific and based on cultural ethos. Wolfson, Marmor and Jones (1989: 180), on the other hand, say that ‘even in societies where it has been established that a named speech act can be translated, one cannot assume that what appears to be the same situation will result in the same speech act’. The bilingual or trilingual child may keep pragmatic knowledge of each language in separate systems together with linguistic knowledge. BlumKulka and House (1989) and Vazquez (1995) point out that Spanish is more direct than English, and Basque is thought to be even more direct (Barnes, 2001). A trilingual child might, for example, be more or less polite in the use of one language or another according to the norms s/he has acquired relating to speech acts in that language, or cross-cultural interference may occur in the form of inappropriate directness or over-politeness, in the formulation of requests, for example. This has been proposed for adults in the intercultural style hypothesis, which describes intercultural patterns reflecting bi-directional interaction between the languages (Blum-Kulka, 1991). For instance, requests made by immigrants to Israel who had English as a first language (L1) and were also competent speakers of Hebrew as a

29 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Early Trilingualism

28

second language (L2) presented a level of directness that was situated between the normal levels for those languages (Kasper & Blum-Kulka, 1993). Barnes (2001) reports that a five-year-old trilingual child (English/ Basque/Spanish) was more direct in English to a trilingual peer than he was to English monolingual peers. This finding suggests that children are not only able to differentiate pragmatically in relation to language mixing and language choice (Genesee et al., 1996; Meisel, 1994; Nicoladis, 1998; Nicoladis & Genesee, 1996a) but also in relation to appropriate levels of politeness.

Trilingual Acquisition In comparison with the field of bilingualism, there are relatively few studies of trilingualism, although in recent years interest in this subject has increased notably. Work in the field of trilingualism has focused mainly on acquisition in adults (e.g. Hammarberg & Williams, 1993; Clyne, 1997) and on areas related to education such as immersion, and particularly the learning of a third or fourth language (often English or French) in bilingual and multilingual school contexts (Cenoz, 1991; Cenoz & Genesee, 1998; Cenoz et al., 2000; Lasagabaster, 1997; Sagasta, 2000). However, studies on early trilingual acquisition are still infrequent. Hoffmann (2001: 3) describes five sets of social circumstances that may result in speakers of three languages: (1) trilingual children who are brought up with two home languages that are different from the one spoken in the wider community; (2) children who grow up in a bilingual community and whose home language (that of either one or both parents) is different from the community languages; (3) third language learners, that is, bilinguals who acquire a third language in the school context; (4) bilinguals who have become trilingual through immigration; (5) members of trilingual communities. Trilingual acquisition, learning and use are also considered to be multilingualism. For the purposes of the present research review, the term ‘trilingual’ refers to specific cases involving three languages, ‘quadrilingual’ refers to cases involving four languages, and ‘multilingual’ is a general term for language situations that are not bilingual and may involve three or more languages. Second language acquisition (SLA) has been widely used as a cover term for all types of language learning other than first language acquisition

30 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

29

(Sharwood Smith, 1994; Cenoz, 2000) and has much in common with trilingual and multilingual acquisition. Cenoz (2000: 47–48) explains that multilingual acquisition may be influenced by the individual and contextual factors related to second language acquisition as well as by the known outcomes of bilingualism such as ‘creativity, metalinguistic awareness and communicative sensitivity’. However Cenoz also points out ‘multilingual acquisition presents more diversity than second language acquisition and its study presents greater complexity’ (Cenoz, 2000: 40). This is due in part to the ‘interactions that are possible among the multiple languages being learned and the processes of learning them’ (Cenoz & Genesee, 1998a: 16). The typology of the languages involved and their linguistic distance is an important factor in considering the processes of multilingual acquisition. In the context of the Iberian Peninsula for example, the trilingual acquisition of Spanish, Catalan and English is likely to present different characteristics from the acquisition of Spanish, Basque and English. Spanish and Catalan are not only Indo-European languages like English, but are also both very close in linguistic distance. On the other hand, Basque is a non-Indo-European language unlike Spanish and English which, however, are relatively linguistically distant from each other. In multilingual acquisition it has been found that when the languages are similar there will be cross-linguistic transfer from one into the other (Cenoz & Genesee, 1998; Singleton, 1987). Research has shown that typological similarity between languages is a major factor in code-switching and crosslinguistic influence (Cenoz, 2001; Williams & Hammarberg, 1998). If the first language is non-Indo-European, learners will transfer between the languages that are Indo-European (Ringbom, 1987; Singh & Carroll, 1979). The Dynamic Model of Multilingualism states that ‘language competence in multilinguals is ... to be considered as consisting of dynamically interacting language subsystems which themselves do not necessarily represent any kind of constant but are subject to variation’ (Jessner, 1997: 27). Combined with factors such as the status of the languages in education or the community and the uses to which an individual may put the languages, it becomes clear that the potential for complexity is great and that multilingualism can give rise to unique situations of language acquisition (Cenoz, 2000: 42). Hoffmann (2001: 12) examines the competence of trilinguals in relation to speech modes (Grosjean, 1992). She explains that if this notion is applied to trilinguals three speech modes are possible: monolingual, bilingual and trilingual. From these three modes (A, B and C) it would be possible to produce seven different language ‘constellations’, namely:

31 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:04

Color profile: Disabled Composite Default screen

Early Trilingualism

30

• the monolingual modes A, B, and C; • the bilingual modes AB, AC, and BC; • the trilingual mode ABC. However, Hoffmann points out that although seven modes are possible, she has found only five to be used – the three monolingual codes and the two bilingual modes with the community language. She finds no evidence to support the existence of a trilingual speech mode. Hoffmann also comments that ‘trilingualism obviously shares a number of features with both bilingualism and multilingualism. At the same time, it retains characteristics of its own’ (Hoffmann, 2001:3).

Early Trilingualism Reports of early multilingualism tend to be of two types. First, there are reports that are descriptive, and in some cases anecdotal; these give details of children and their families, the languages they use and the contexts in which they are used. Second there are studies that, in addition to providing the information above, examine specific areas of the children’s language use and parental input such as evidence of cross-linguistic influence through code-switching and mixing, language dominance, language differentiation and topic relatedness. The following pages review the literature available on early multilingualism obtained from descriptive studies and the findings from studies that focus on specific linguistic areas. Descriptive studies Several sources describe children who appear to be trilingual and multilingual from birth or infancy. References to families using a variety of languages can be found in De Jong (1986), Harding and Riley (1986), and Arnberg (1987) and regularly feature in the Bilingual Family Newsletter. One such report is by Turrell (1997) who describes a Catalan/Spanish/English trilingual child. The child tends to mix between Catalan and Spanish because of their typological proximity, but does not mix between Catalan and English or Spanish and English. Tokuhama-Espinosa (2001) also describes cases of trilingualism and multilingualism in families. Barron-Hauwaert (2000) used questionnaires to investigate some of the issues surrounding language acquisition in ten trilingual families, mainly in Europe. She found a trend towards very young children using the mother’s language and displaying stages of monolingual initial language use, bilingual language use at pre-school age (often the father’s language) with full trilingualism (with the community language) developing at

32 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

31

school age. Barron-Hauwaert (2004) provides theoretical background and detailed advice to bilingual and multilingual families on using the oneperson-one-language approach, along with numerous case studies of bilingual, trilingual and multilingual family linguistic circumstances. Helot (1988) describes patterns of language use in Dublin in two families with older children who have been brought up in three languages. The mothers are French and the fathers are native speakers of Irish. In one family French was used in the home and English was the community language; Irish was acquired at primary school. In the second family French and Irish were used by the respective parents and English was the community language. The study concludes that it is possible to raise children trilingually in a family context and achieve competence in three languages, though this may be at various levels (native-like, competent or working knowledge) and may vary over time (see also Hoffmann 1985). Success is also seen to relate to the status of the languages concerned and to issues of additive and subtractive bilingualism (Hamers & Blanc, 1989; Lambert, 1975). Pérez-Vidal (1995) and Juan-Garau (1996) have conducted studies on the bilingualism of a child who has been exposed to Catalan (mother), English (father) and Spanish (relatives and community) since birth. The studies focus on the boy’s production and as he did not produce in Spanish until the age of five he is therefore bilingually productive although a passive trilingual (Pérez-Vidal, 1995: 85). Cenoz and Barnes (1997) describe the attempts of a trilingual English/ Basque/Spanish child to re-tell a tale in the three languages and describe aspects of his phonology, vocabulary, grammar and pragmatics in each language. They report little evidence of code-mixing in the data. Barnes (2001) examines the acquisition of English speech acts, specifically requests, within a trilingual context and with reference to a trilingual English, Basque, Spanish five-year-old and similar trilingual peers. The findings reflect those of Ervin-Tripp and Gordon (1986), Becker (1982, 1986) and Corsaro (1979), showing that the child tends to be more polite to adult speakers, especially in English, less polite to his mother (in English) and least polite to his peers. An interesting finding is that the child tends to be less polite when speaking English to trilingual speakers of English than when he speaks to monolingual speakers, suggesting transfer on a pragmatic level from Spanish and Basque which have higher levels of directness. Navracsics (1999) describes the development of siblings Nasim and Nabil in Canada who were exposed to English and Persian from their parents from birth. At age three and two respectively, they moved to

33 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Early Trilingualism

32

Hungary. The study focuses on the way Hungarian developed to become their dominant language. As well as descriptive studies, a growing number of studies of early trilingualism focus on specific linguistic areas, predominantly in relation to cross-linguistic influence and code-switching. As Hoffmann (2001: 2) points out, many of these studies have worked using data collection and analysis methods taken from the theoretical framework of bilingualism. Cross-linguistic influence and code-switching Oksaar (1978) reports on code-switching and interference in a child, Sven (aged 3;11–5;8) who acquired Estonian and Swedish at home from birth and German in the community from age 3;11. Like Stavans (1992) Oksaar finds that multilingual children will not switch if they know they are addressing a monolingual. However, utterances containing two or even three languages may be used for effect to a multilingual, for example in order to reinforce a request. Hoffmann (1985) reports on two children in the United Kingdom who acquired Spanish from their father and German from their mother. Cristina was exposed to direct and daily contact with English when she started to attend play school at age 2;9. Her younger brother Pascual began play school at 3;1 but this was not his first contact with English since he had indirectly been exposed to it through his sister. The study describes patterns of acquisition of the three languages, mixing and the results of vocabulary recall tests until the age of eight (Cristina) and five (Pascual). A changing order of competence in the languages is described until English, the language of schooling, became dominant. Code-switching tends to be in the direction of the language in which the children are most competent, and lexical resources may be more easily and quickly available in one language than in another. The study also examines social and psychological issues relating to a trilingual upbringing. Stavans (1992) describes code-switching in two children E (5;5–6;8) and M (2;6–3;9) in the United States. The input languages are Spanish from the father, Hebrew from the mother and English from the community, though Stavans does not mentioned at what age the children were exposed to English. The study finds that the linguistic background of the addressee (monolingual, bilingual or trilingual) influences the quantity and types of switches, it also finds that switches related to topic are culture bound, and that switches may be located in discourse routines such as requests, affirmations or negation. Stavans finds that with monolingual interlocutors switching is negligible. Cruz Ferreira (1999) examines prosodic mixes in three trilingual siblings

34 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

33

who have Portuguese and English from birth and English from schooling in Hong Kong. She finds that the children (aged 11, 9 and 7 years) extend prosodic mixes from their first two languages into the third in an attempt to ‘explore’ (1999: 20) and ‘in order to learn language specific constraints in the process of working their way up to linguistic competence in several languages’ in a dynamic multilingual acquisition process (Jessner, 1997). Quay (2001: 157) proposes that code-switching in multilingual children is affected by factors such as the participants in the interaction, topic, availability of lexical resources and emphatic function. Hoffmann (2001: 12) and Hoffmann and Widdicombe (1999) failed to find evidence in their data of situations in which all three languages were used, even if other trilinguals were present. Topic, context and person relatedness Hoffmann (1985: 489) suggests that in trilingual children there is a link between context and first occurrence of a linguistic item and its subsequent acquisition. Furthermore there may be a ‘crucial’ age for certain items to be acquired from particular individuals, which leads to the persistence of certain person specific forms. Of particular interest is the reference to ‘a sense in which language forms tend to be implanted in the brain in the way in which they were first acquired’. This may be due to the limited personspecific input that the child is exposed to in multilingual family contexts involving a minority language in contrast to acquisition in a wider community context. Döpke (2000: 4) also mentions this in relation to her own bilingual child, saying that when the sources of input of a language to a child are limited it is not difficult to trace when the child was first exposed to a particular linguistic item. Stavans (1992) reports that a topic (e.g. food items) may be culturally bound to one of the three languages and reference to those items would always be in that particular language regardless of language context. Oksaar (1978) has also found that some topics are bound to the location where one language is usually employed, for example topics referring to the activities carried out by the house helper or topics referring to play and playmates. Dominance Widdicombe (1997) and Hoffmann and Widdicombe (1999) describe code-switching, coining and interference in a child aged 4;4 to 4:5 who has acquired Italian from his father, English from his mother and French from the community. Although the child has been exposed to all three languages from birth, evidence from the quality and quantity of switches made in

35 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Early Trilingualism

34

French suggest that it is this language, which is also the language of his schooling, that is becoming dominant. As with Hoffman’s (1985) earlier study, Hoffman and Widdicombe find that it is the community language that becomes dominant and it is this language that is most evident in the code-switching, which may result from a lack of lexical resources in the weaker language(s). Input Quay (2001) examines input requirements and the importance of the age of exposure in a longitudinal study from 0;11 to 1;10 of a German/ English/ Japanese trilingual child in Japan. The child, Freddy, has been exposed to German from his father and English from his mother and is exposed to Japanese through day-care which commenced at 0;11. Until this time Freddy had some limited contact with Japanese since his parents both speak the language and use it with Japanese speakers, though Quay questions whether this a case of ‘later’ exposure to the third language, though prior to onset of speech (Quay 2001: 8). She finds that the parent’s knowledge of Japanese and their discourse strategies (Döpke 1992; Lanza 1997) play a key part in the child’s language development since, although he uses English or German with them, he has a preference for speaking Japanese. Through their acceptance of his utterances in Japanese and insertion of Japanese lexical items and code-switching, the parents have negotiated a trilingual context for the child (Lanza 1997). Although this is not the language they usually use with the child, sociolinguistic constraints have resulted in contexts where use of one, two, or three codes is acceptable. Quay (2001: 153) makes a point of distinguishing between those children who, like her subject, are exposed to three languages before the onset of speech, and those who are not. Faingold (1999) reports on a child, Noam, who has four languages, Spanish (father), Portuguese (mother) and Hebrew (parents, care-workers and community) from birth (in Israel), and English from age six (in the United States). Faingold describes the child’s development (including temporary loss and re-emergence) from age 0;6 to 14;3 in a series of six stages that are related to differing input and output patterns in the child’s multilingual environment. For example, the parents switched from using Hebrew to using English with the child as his English developed. Between the ages of 2:0 and 2;10 the child is reported to have code-switched both inter- and intra-sententially in all three languages and to both parents. Like Hoffmann (1985), Faingold addresses social and psychological aspects of a multilingual upbringing, and outlines the importance of peer contact in the non-community language(s) for later language growth as well as peer

36 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Bilingual and Trilingual Acquisition

35

approval towards multilingual language use in order to avoid language loss. Differentiation and separate systems Murrell (1966) reports on a child, Sandra, who is trilingual in Finnish, Swedish and English. Born in Finland to a Swedish-speaking mother and an English-speaking father, Sandra was exposed to Finnish through a caregiver at the age of six months. From 0;10 to 2;0 she attended a Finnish daynursery. Her parents use mostly English with each other and her father speaks partly English and partly Swedish with her. Genesee finds evidence in Murrell (1966) to support the differentiated-language-systems hypothesis in a trilingual child; ‘Contrary to evidencing a lack of linguistic differentiation, Sandra appears to have used Swedish and English differentially in contextually sensitive ways during the same conversation with an interlocutor who switches back and forth ...’ (Genesee, 1989: 173). When the child was 2;1 the family moved to England, and Murrell describes lexical, syntactic and phonological mixes in Sandra’s language between 2;0 and 2;8, and captures aspects of her language loss. There is enough data to show that the child was trilingual but ‘after the child had been nearly three months outside the Finnish environment, Finnish was rapidly on the way out’ (Murrell, 1966: 13). Mikès (1990) describes the acquisition of Hungarian, Serbo-Croatian and German by three children Vuk, Uva and Egon. Their input is provided by three individuals: their mother speaks Hungarian, their father speaks Serbo-Croat (which is also the language of the community) and their grandmother speaks German. The study is based on a daily diary and audio recordings, and describes lexical development and lexical differentiation. The findings suggest that early bilingual and trilingual children have two or three lexical systems from the beginning and that lexical resources in one topic area may be related to a particular language. However, Mikès (1990: 114) also reports instances of a trilingual child using utterances containing equivalent words in two or three different languages to reinforce communication, and this has also been described by Oksaar (1978). Clearly we are a long way from formulating a model for early multilingual language acquisition. The studies described here deal with a wide range of languages, ages and circumstances and have involved varied methodological procedures (see Quay 2001: 154–155). If progress is to be made in understanding early multilinguals, larger-scale studies in areas of Asia and Africa where childhood multilingualism is the norm are needed, in addition to more findings from individual case studies.

37 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Early Trilingualism

36

Notes 1. The use of the abbreviation year;month.day to show a child’s age is an established convention throughout the literature on child language acquisition. Thus 1;10 indicates that the child is 1 year and 10 months old, and 3;1.9 would indicate an age of 3 years, 1 month and 9 days. 2. Mean length of utterance is a standard measure of language growth in the field of child language development and describes the average length of utterance at a particular age. In this first example, it is 1.75 words per utterance at age 1 year and 10 months. 3. The parenthesis and interrogative are Meisel’s. 4. Null subject languages (such as Spanish, Catalan and Basque) do not require the subject of a verb to be specified; for example, tiene in Spanish means (he/she or it) has. By contrast, in overt subject languages (such as French or English) the subject must be specified with the verb; for example he has, il a.

38 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Chapter 2

The Development of Interrogative Behaviour Question Form The process of language acquisition leads towards the attainment of an adult model of the language, and examining the target structures helps us to understand the process through which they are achieved. According to Greenbaum and Quirk (1990: 232) questions can be divided into several major classes according to the type of reply they expect to elicit.

Yes/no questions These are questions that expect an affirmative or negative reply, as in Have you made the tea? They are also referred to as polar questions. In English, yes/no questions are formed by inverting the auxiliary be or have before the subject and are spoken with rising intonation: • Sally was late. Was Sally late? If the auxiliaries be and have are not present in the sentence, an appropriate form of the auxiliary verb do is inserted: • They live in Cardiff. Do they live in Cardiff? Placement of stress on a particular part of a yes/no question focuses the interrogation on an item of information that is unknown. Thus the focus falls on different places in otherwise identical questions: • Did Tom go to New Zealand in 1962? (I know Tom went to New Zealand – but was it in 1962 or some other time?) • Did Tom go to New Zealand in 1962? (I know Tom went somewhere in 1962 – but was it New Zealand?) Positive yes/no questions are not marked for negation, and expect either a positive or negative response. • Did you post the letter? Yes, I did / No, I didn’t.

37 39 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Early Trilingualism

38

Tag questions are formed by the addition of a question tag to a declarative statement. For the most common types of tag question, the question tag is negative if the statement is positive, and vice versa. • Valerie recognised you, didn’t she? (Surely Valerie recognised you.) • Valerie didn’t recognise you, did she? Declarative questions have the form of a declarative statement, except for a final rising intonation that invites the hearer’s verification: • You want something to eat?

Wh-questions Questions that expect a reply from an open range of replies, as in Where have you been? or How did you get here? are called open questions or wh-questions. Wh-questions are formed with the following interrogative words: who, whom, whose, what, which, when, where, how, why The wh-element is usually in initial position in the questioning clause. Unlike yes/no questions, wh-questions generally have falling intonation. The normal statement order is altered in wh-questions by the placement of the wh-element in initial position and also by the inversion of the subject and the operator. • • • • •

What did they eat? When will you be leaving? Where shall I put the equipment? Why didn’t you tell me? How did you break it?

This inversion does not occur when the wh-element in question refers to the subject. • Who came to the lesson? • Whose beautiful earrings are these? Subject–auxiliary inversion is the same in wh-questions as in yes/no questions; if there is no auxiliary be or have, an appropriate form of do is introduced into the question. • Where has he gone? (auxiliary ‘have’) • Where do you live?

Other questions A number of other question types fall outside the two main groups of yes/no and wh-questions. Alternative questions expect a reply of one of two

40 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

39

or more options presented in the question, as in Would you like to have a picnic or a barbecue? There are two main types of alternative questions. The first resembles a yes/no question and the second a wh-question: • Would you like chocolate or strawberry milkshake? • Which milkshake would you like? Chocolate, vanilla or strawberry? Rhetorical questions have the force of a strong assertion, not necessarily expecting an answer, for example: • Are you wearing that? Echo questions repeat part or all of what has been said. A: I’ll put it on the floor. B: On the floor? The formula What did you say? is described by Greenbaum and Quirk (1990: 241) as a generalised replicatory question, often shortened to What? Likewise, the alternative formula I beg your pardon? is reduced to Pardon? Other abbreviated requests for repetition in this category include Sorry? Uuuh? Hmm? and Eh?

Studies on the Acquisition of Question Form Studies into children’s questioning behaviour go back well over 100 years and range from the work of Darwin (1877) to an early study on the function of questions carried out by Sully (1896). There are a number of elements that need to be accounted for in a description of the acquisition of question form. These include length of question, proportion of questions to utterances, gender differences and, according to question type (yes/no or wh-), the time, order and rate of emergence and use of inversion and auxiliary insertion. In the 1930s a number of papers appeared giving accounts of children’s questions (Davis 1932; Lewis 1938; McCarthy 1930; Smith 1933) as a result of interest in Piaget’s (1926) work on language and cognitive development, specifically the logical functions of why? and other questions in a child from 6;3–7;1. Research, particularly in the area of syntax, was renewed in the 1960, 1970s and early 1980s, influenced by Chomsky’s (1957, 1965) theoretical framework and Brown’s developmental stages (1973). Klima and Bellugi (1966) proposed a model of three stages of question development that was incorporated into various studies (Bellugi, 1965, 1971; Brown, 1968, Brown et al., 1969; Brown & Hanlon, 1970, and Cazden, 1970) that examined the extent to which mature underlying grammatical structure is

41 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:06

Color profile: Disabled Composite Default screen

40

Early Trilingualism

present in early child questions. Labov and Labov (1978) provided further evidence of the three stages in a longitudinal study of the acquisition of their child’s wh-questions, with particular reference to why. These studies were also concerned with the inversion rule described in the third of Klima and Bellugi’s stages, where inversion does not occur in wh-questions but does appear in yes/no questions. Ingram and Tyack (1979), Erreich (1984) and Klee (1985) showed that wh-inversion does in fact occur. From the mid 1980s onwards, and within the framework of Government and Binding (Chomsky, 1981, 1982), the acquisition of wh- has been of major importance in syntax studies, since wh-questions appear to be universal to all languages (Ultan, 1978; Comrie 1981) and amenable to description within this framework (Stromswold, 1995). The acquisition, structure and movement of English wh-questions have been examined by many researchers, including Stromswold and Pinker (1986), Stromswold (1988), Thornton (1990), Radford (1990, 1994) and de Villiers (1991, 1995). Mean length of questions in words and proportion of questions to sentences The mean length of questions in words by age group measured by Davis (1932) revealed a slight but continuous increase in the length of questions and a trend for longer questions for older children. Davis found a similar distribution in length for children aged 3–8, children aged 8–12 and adults, and that the average length of English questions tends to be between four to six words. Meyer and Shane (1973) compared Davis’s (1932) data with their own, which had been collected by different procedures. They found that the average length of wh-question in younger children was identical to that found in Davis. Smith (1933) found that questions constituted 13% of children’s speech between the ages of 1;6 and 6;0. McCarthy (1930) reported 7–14%. Smith (1933) also found that the proportion of questions to sentences increased with age, but decreased between ages five and six when, she suggests, more sophisticated attention-getting devices became available. Gender differences Both Davis (1932) and Smith (1933) reported gender differences in interrogative behaviour. Davis (1932) found that girls tended to use more auxiliaries to start questions than boys did, especially as age increased. Girls ask slightly longer questions earlier (Davis, 1932), and Smith (1933) reported that at the age of two, girls asked more questions than boys do. Older girls use more auxiliary insertion and fewer wh-questions than younger girls do (Davis, 1932). Davis (1932) found that boys use wh-forms to start their ques-

42 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:06

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

41

tions more frequently than girls do. Both Meyer and Shane (1973) and Davis (1932) agree that girls are linguistically superior to boys in questioning behaviour. Emergence of interrogative behaviour and intonation In a longitudinal account (0;10,28–2;10,18) of one child’s acquisition of the form and function of questions Lewis (1938: 152) defined ‘securing of a response’ as the moment when questioning behaviour emerges, and placed this at 1;4. Smith (1933: 205) in a study of 219 children between the ages of 1;6 and 6;0 recognised the importance of ‘inflection of voice’ on otherwise unmarked utterances in the identification of early questions. Menyuk (1969) conducted experiments on the prosodic features of a child’s oneword utterances (e.g. the word door), and found that stress and intonation were used to designate ‘sentence’ types such as statement, question and emphatic even before these were syntactically marked. Smith (1933) pointed to the fact that 49% of her subjects asked questions at age 2;0, whereas by 6;0 93% of them were doing so. Tyack and Ingram (1977) also found a high frequency of yes/no, what and where questions by age 2;0. In a major longitudinal study on language acquisition, Wells (1985) presented findings compiled from 1350 observations of children aged from 15 months to 5 years. The recordings were made three-monthly over three and a half years with 128 subjects in Bristol, England, and Wells describes a clear sequence of emergence for interrogative sentences. According to Wells (1985) the interrogative is most generally formed by inversion of the subject and the first element of the verb phrase, with the addition of an auxiliary do when there is a lexical1 verb. Wh-questions with who, what or where followed by a copula2 verb have a structure that can be described as wh + cop + NP3 (e.g. Where is the ball?) in which it is the noun phrase (NP) that is being asked about (Wells, 1985: 198). Early questions are structurally simple in comparison with other types of interrogative such as wh-questions with lexical verbs or yes/no questions. Wells suggests that this is why what + cop + S (e.g. what’s that?) is the first type of question to emerge, closely followed by where + cop + S. He finds that the first yes/no question marked for inversion to emerge is aux + S + V + O/A, such as can you do it a car? (can you make a car?). This is followed by the first wh-question with a lexical verb and inversion wh + aux + S + V (e.g. what has Daddy done?) and by the first yes/no question in which the copula verb is inverted with the subject cop + S + X (e.g. is this pyjama?). Wells (1985: 198) also notes, like Brown (1968) and Cazden (1970), that the earliest and most frequent form of question, what dat? (what’s that?), may be learned as an unanalysed

43 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Early Trilingualism

42

whole since it is frequently addressed to children at around the age of two as part of naming games (McShane, 1980; Ninio & Bruner, 1978). From this point Wells claims that the sequence of emergence resembles that of declarative sentences (see Wells, 1985: 270, for a full description) with two exceptions: why interrogatives and who interrogatives containing a lexical verb. Why interrogatives are structurally more complex than other wh-interrogatives since they question the whole dependant clause (why + cause/ reason) rather than simply a constituent of it. Who interrogatives, on the other hand, are relatively infrequent. Emergence of yes/no questions Smith (1933) describes questions with inflected auxiliaries but without wh- as making up 61% of total questions collected in her subjects between the ages of 2 and 6, but she does not mention age of emergence. Tyack and Ingram (1977) used Davis’s (1932) parental data collection procedure and similarly found that slightly more than 60% of the total questions were of the yes/no type in the age group 2;0–3;11. James and Seebach (1982) studied the use of question forms in three different functional categories in 24 subjects aged between two and five years. Their results indicate that yes/no questions are present at around age two. Periods (or stages) were identified for question and negation acquisition in an analysis (Klima & Bellugi, 1966) of data collected from three children Adam, Eve and Sarah. The periods were based on the results of one child (Adam), but apply to the development of all three. Each period begins at approximately the MLU (mean length of utterance) and age shown and lasts approximately until the next MLU mentioned. The periods have been demarked by Cazden (1970) and describe development of yes/no questions as in Table 2.1. Early questions at Stage A consist of one or two words with rising intonation. At Stage B, as questions increase in length, they follow the word order of declarative sentences. At Stage C the child starts to invert yes/no questions, perhaps realising that question formation requires ‘fronting’ (movement towards the beginning of the sentence), before going on to correctly invert yes/no questions and to insert the auxiliary do where appropriate. At Stages D–F the child perfects inversion and insertion of do, and tag questions develop. Cazden (1970: 207) notes that, while the examples shown here were from Adam, the developmental sequence shown was the same for all three of the children studied although there were wide individual variations in the time of emergence. She also points out that each child in the study had his or her own idiosyncratic question forms, on which s/he relied extensively in periods A–C, and particularly in period B.

44 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

43

Table 2.1 The development of yes/no questions Bellugi's period

MLU

Examples

A 28 months

2.0

Expressed by intonation only: Sit chair?

B 35 months

2.5

No development of question forms except the appearance as routines of don’t and can’t: Dat black too? Mom pinch finger? You can't fix it?

C 38 months

3.5

Development of auxiliary verbs in entire grammatical system. Inversion of auxiliary and subject NP (noun phrase) in yes/no questions: Are you going to make it with me? Will you help me? Does the kitty stand up? Can I have a piece of paper?

4.0+

Development of tag questions from huh? Mature inversion of auxiliary and subject form: I have two turn, huh? We're playing, huh? That's funny, isn't it?

D–F 42–54 months

Based on Klima and Bellugi (1966) and Cazden (1970: 207)

Cazden gives the example of why not + pronoun + VP in the case of Adam, and suggests that such forms are ‘unanalysed chunks, unrelated to other forms in the child’s language system’ (Cazden, 1970: 208). In his study of an English child, Sophie, Fletcher (1985: 106) finds that large numbers of yes/no questions are present in the data. Sophie is recorded from 2;4 and Fletcher reports that at around 2;7 this type of question constitutes almost 50% of all her questions. He also reports that at approximately 2;5 a large number of inverted yes/no questions are can-initial (begin with can). Other frequent yes/no questions to appear early with inversion have the auxiliaries is or have in initial position. Whilst have appears with got, is may be found with other verbs. In an experimental study, Santelmann et al. (2002) examined inversion in yes/no questions in children aged 2–5 years, and found that inversion in questions with the auxiliary be or modals is present from an early age. They assessed the use of intonation questions and suggest that, although a child could avoid inversion by consistently using them, this is not the option chosen. Instead the child will produce verb movement to signify interrogation.

45 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:05

Color profile: Disabled Composite Default screen

Early Trilingualism

44

Emergence of wh-questions In relation to the acquisition of wh, Smith (1933) found that wh-questions were used to ask 36% of questions. At the age of 2, 49% of the children’s questions began with wh- and this decreased to 37% at 5 years. No cases of how, when and where were found at two years of age. Smith also reports that the initial main use of what and where decreased as how, when and why emerge. Who, whose and which showed no correlation between age and frequency of use. A chronological order of acquisition: what, where, why, how, who, when is described by Tyack and Ingram (1977), who note that why and how questions were present but infrequent at age two, and increased with age. Children in the under-four age group rarely asked who or when questions. On the whole Tyack and Ingram’s findings were similar to those of Smith (1933). Usage of what and where also compared favourably, with 15% and 12% respectively in Tyack and Ingram’s data and 15% and 11% in Smith’s results. In Tyack and Ingram why is more frequent across all age groups and shows an increase in use around 3;0. Several wh-words showed an increase in use across the age groups, suggesting that these words are in the process of being acquired. Where decreased in both sets of findings, possibly because new forms such as how and why compete with it (Tyack & Ingram, 1977) Tyack and Ingram relate the frequencies for the appearance of wh-words to the concepts the child needs to express at different ages (Piaget, 1926): What and where questions are closely tied to the child’s immediate environment. What will often be used to seek the names of new objects and where will be used to locate lost playthings. Later, as the child develops interest in concepts of cause, manner and time, the applicable question-words will appear. The infrequency of when questions can be traced to the fact that the child’s concept of time takes until the age of six to become developed. (Tyack & Ingram, 1977: 222) James and Seebach (1982) find the emergence of wh-questions at the different ages to be: Age 2 Age 3 Age 4 Age 5

what; what, where, why, who; what, where, why, who, when, how; what, where, why, who, when, how.

These findings reflect those of Wells (1985), who also reported an order of what, where, why and who.

46 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

45

With reference to syntactic development in wh-questions Cazden (1970) describes the following stages based on Klima and Bellugi (1966) shown in Table 2.2. At Stage A what appears as the first form in routines or formulaic chunks such as what ... doing? or whaddat? At Stage B more complex questions are asked but without the appearance of syntactic question forms and accompanied by the emergence of where, why and who. These questions follow the declarative syntactic forms of yes/no questions at this stage but the wh-form is placed in initial position. At Stage C, when inversion and do insertion are appearing elsewhere in the grammatical system, the child appears unable to perform the same movement when a wh-form is present (Cazden, 1970; Klima & Bellugi, 1966). The ability to do so appears during Stages D–F. The claim of non-inversion of wh-questions at this stage is subsequently tested (Erreich, 1980; Ingram & Tyack, 1979; Klee, 1985) and found to be non-substantiated. Overviews of frequency and developmental trends in wh-acquisition between 1;2 and 6;0 are also provided by Radford (1994) and Stromswold Table 2.2 The development of wh-questions Bellugi's period

MLU

Examples

A 28 months

2.0

Limited number of routines: What(s) that? Where NP go? What NP doing?

B 35 months

2.5

More complex sentences being asked but without development of question forms: What soldier marching? Where my mitten? Why you waking me up?

C 38 months

3.5

Development of auxiliary verbs in entire grammatical system. Inversion of auxiliary and subject NP in yes/no questions but not in whquestions: What I did yesterday? Why the Christmas tree going?

4.0+

Development of tag questions from huh? Mature inversion of auxiliary and subject form. Why are you thirsty? Why can’t we find the right one?

D–F 42–54 months

Based on Klima and Bellugi (1966) and Cazden (1970: 207)

47 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

Early Trilingualism

46

(1995). These studies bear out the findings of Smith (1933) and Tyack and Ingram (1977) on developmental order. Radford (1994) provides a contemporary reinterpretation of Klima and Bellugi’s (1966) work within a Government and Binding4 (GB) framework based on Chomsky (1981). Stromswold (1995) examines the acquisition of subject and object wh-questions by testing several linguistic descriptions from GB on question data from 12 subjects in the CHILDES database5 (MacWhinney, 1991). Wh-subject questions require no inversion or auxiliary and therefore would seem to be easier to acquire than wh-object questions, but Stromswold (1995) found that children produce wh-object questions correctly before wh-subject questions, and suggests that this is a result of acquisitional ordering. Fletcher (1985: 161) found how and when in a sample at age 3;5 whereas at 3;4 only who, what, where, why and which one were present. He also reports that, by age three, his subject Sophie appears to have acquired auxiliary verbs and mastered the rule of inversion for yes/no and wh-questions. Furthermore like Labov and Labov (1978) in their study of wh-acquisition in their child Jessie, Fletcher (1985) found that the child showed different rates of success depending on wh-type. In both studies inversion was most successful with how, followed by where, what and finally why. However, Fletcher (1985: 106) claims that the success of how is related to its use as a formulaic initial phrase how do you? Fletcher goes on to show that Sophie’s apparent success at auxiliary insertion and inversion with wh- was due to strategies of auxiliary insertion based on the use of sequences, such as what are and how do you rather than real auxiliary insertion and inversion. Capdevila and Serrat (1996: 21–22) present data in Spanish and Catalan and compare their findings with recent studies on wh-acquisition available in English in order to test Bellugi’s (1971) three stages across the three languages and conclude: (1) between about 19 and 24 months, children use very few wh-questions, principally with what and where and some examples of who; (2) between 24 and 26 months approximately, the questions are more complex but the basic structure of interrogatives is similar to the first stage and why starts to appear; (3) from about 26 to 36 months most interrogative pronouns are used productively apart from when, and auxiliary forms have appeared in both inverted and non-inverted forms. Rowland and Pine (2000) re-examine wh-inversion from a constructivist viewpoint, prompted by findings on early wh-questions that suggest that such questions can be explained in terms of formulaic wh-word and auxiliary combinations (Fletcher, 1985). Rowland and Pine examine the wh-

48 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

47

questions of Adam (Brown, 1973) between 2;3 and 4;10 to show that noninversion errors are not due to the failure to apply an inversion rule (de Villiers, 1991; Valian et al., 1992) but are related to the combinations of whwords and auxiliaries that the child receives in input. Inversion and auxiliary insertion in wh- and yes/no question form There seems an established order for the emergence of the question types. Yes/no questions are present from the beginning and wh-forms follow an order of emergence, what and where, followed by: (1) how, why, when and other (Smith, 1933); (2) why, how, who and when (Tyack & Ingram, 1977); (3) why, who, when and how (James & Seebach, 1982; Wells, 1985). However, both yes/no and wh-questions involve inversion and auxiliary insertion in order to become adult-like, and this appears to develop at around age three (or Period C) when errors that are evidence of the acquisition process are seen (Cazden, 1970; Klima & Bellugi, 1966). These errors may be of non-inversion, such as: • What he can ride in? where the wh-element is inserted in the correct position but the child has failed to invert the modal auxiliary can before the subject. Equally the child may generalise that questions are formed by fronting (putting the verb at the beginning of the sentence), and produce errors in yes/no questions of the type reported by Lightbown and Spada (1999: 8): • Is the teddy is tired? • Do I can have a cookie? Another frequent type of error is double tense marking (Ingram, 1989: 458), where the past tense is marked in both the auxiliary and the main verb., such as: • Did I caught it? A similar error is double auxiliary marking (Hurford, 1975): • What did you did? • Whose is that is? In the first example the subject and auxiliary are correctly inverted after the wh-element, but the auxiliary is replicated in the verb form, not unlike double tense marking. In the second example the copula is has been used as if it were an auxiliary requiring a main verb and is replicated in the usual

49 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:06

Color profile: Disabled Composite Default screen

Early Trilingualism

48

main verb position. Alternatively whose is or whose is that may have been learned as formulas. These findings have been explained in terms of performance factors and competence factors. In the performance factors approach the child is seen as having the adult grammar, but performance factors restrict the child from fully expressing grammatical knowledge (Bellugi, 1967; Brown, 1968; Brown et al., 1969). The competence factors approach (Erreich et al., 1980; Kuczaj & Brannick, 1979; Kuczaj & Maratsos, 1983) assumes ‘that the errors reflect the child’s grammatical system, that is, that the child has formed an incorrect grammar of English and needs to correct it’ (Ingram, 1989: 461). Labov and Labov (1978), Kuczaj and Brannick (1979) and Maratsos et al. (1979) have all reported the occurrence of non-inversion errors in whquestions relative to correctly inverted yes/no questions. Maratsos (1979) characterised the stages of auxiliary placement in wh-questions as omission (where we put that?), misplacement (where we should put that?) and correct placement (where should we put that?). He noted, however, that children might produce all three patterns to varying degrees at any given time. The findings of these studies suggested that the acquisition of yes/no questions differed from that of wh-questions in three ways summed up by Erreich (1984: 580): (1) auxiliaries appear in yes/no questions before wh-questions; (2) inversion is productive in yes/no questions before wh-questions; (3) inversion is productive in affirmative wh-questions before negative questions. Ingram and Tyack (1979), in a study of the acquisition of inversion in questions, contradict these earlier longitudinal findings and argue against the existence of a stage in which there is inversion in yes/no questions, but not in wh-questions. In a cross-sectional study of questions from 21 children aged 2;0 to 3;11 they found that all the children inverted in both yes/no and wh-questions. Only two children produced non-inversion errors. Erreich (1984) examined the questions of 18 children aged 2;5 to 3;0 and also found that inversion was not acquired in yes/no questions before wh-questions, concluding that children used an optional inversion rule (Erreich et al., 1980) in both question types, which accounts for the non-inversion errors that are a characteristic feature of the acquisition of wh-questions. Klee (1985) in a study of six children aged from 25 to 47 months, whilst disputing some of the methodological aspects of Erreich (1980) and Ingram and Tyack (1979), also found that children do not pass through a stage characterised by lack of inverted forms. Findings from studies on inversion and non-inversion in wh-questions

50 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

49

and yes/no questions and on the emergence and productive use of auxiliaries are summarised by Erreich (1984: 579–580): • Questions are formed by inverting a subject noun phrase and the auxiliary that follows it. • Inversion is optional in yes/no questions but it is obligatory in wh-questions.6 When inversion does not apply in yes/no questions, rising intonation may be used to indicate that the utterance is a question. Since inversion must apply in wh-object questions, failure to use it constitutes a non-inversion error. This is confirmed by the findings of Wells (1985) on order of emergence, in which he describes wh-questions with a lexical verb and inversion (wh + aux + S + V) as one of the first forms to appear. There seems to be a sequence of development in question formation in children (subject to some regression), the rate of emergence can vary from individual to individual and is related to cognitive development (Cazden, 1970). Wh-forms appear in a predictable order, beginning with what, which often appears as formulaic chunks such as whaddat? Soon to follow are where and who, reflecting children’s ability to answer such questions and their understanding of the here and now. Why appears a little later, at around the end of the second year and will be used frequently although the child may not correctly grasp its function of questioning cause and effect for a number of years. Nevertheless it has a ‘usefulness in ... getting adults to engage in conversation’ (Lightbown & Spada, 1999: 7). As children begin to understand manner and time, how and when emerge. The earliest syntactic forms of questions consist of one or two words with rising intonation and as they increase in length they follow the word order of declarative sentences. If the question has a wh-word this will simply be placed in the initial position without inversion. The child at this stage will then start to invert yes/no questions. This is followed by correct inversion of yes/no questions with do insertion where appropriate, while the child gradually learns to do the same with wh-questions.

Comprehension of Questions and Input Comprehension of form A child’s ability to recognise questions as such, to interpret them semantically, and to respond with appropriate semantic and syntactic forms tells us about the cognitive and psycholinguistic development of that child (Cairns & Hsu, 1978). Ervin-Tripp (1970) examined answers to questions in two groups of children. The first group was aged between 2;0 and 3;0 and

51 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

Early Trilingualism

50

consisted of five subjects observed longitudinally (in other words, they were observed on a regular basis over a long period of time). The second group consisted of 24 children aged between 2;6 and 3;1 who were given monthly interviews. Ervin-Tripp (1970: 95–96) proposed that the children employed four strategies in answering questions: (1) if you recognise the word, give a correct reply; (2) if there is a transitive verb, respond with the object of that verb; (3) if you are over three, and there is an animate subject and an intransitive verb, give a causal explanation; (4) for the remaining intransitives, give a locative or direction if it is missing. In the first group of children, Ervin-Tripp found that yes/no, what and where questions were the first to be understood. The larger sample revealed the following order: why, who-subject, how, where-from, when and who-object. Tyack and Ingram (1977) tested 100 children aged 3;0–5;5. on syntax, vocabulary and varied wh-words. The frequency of correct answers increased with the age of the children. When the children answered incorrectly, these answers were not random but appeared to follow question-answering strategies like those suggested by Ervin-Tripp (1970), including attention to the semantic features of verbs and the placement of verbs in the sentence. The work of Cairns and Hsu (1978) also supports the findings of Tyack and Ingram (1977). In this study 50 children between the ages of 3;0 and 5;6 were asked a range of wh-questions after viewing a videotaped sequence, and their answers were graded according to degree of difficulty. Results with who point to the difficulty of information retrieval and processing during discourse. The difficulty of why and when questions was related to the child’s ability to encode the relevant concepts linguistically. Difficulty with how questions was found to be due to their involvement with a number of unrelated skills. The difficulty of some questions has also been confirmed by Parnell, Paterson and Harding (1984), who studied the responses of 40 children aged 3 to 6 years to wh-question forms. These children were significantly less successful in giving appropriate and accurate responses when the questions referred to objects, people or events that were not represented in the immediate setting. These studies reflect that ‘the questions that are acquired late in the sequence, why and when, refer to more abstract, less tangible ideas than do what and where questions, which are learned first’ (Bloom et al., 1982: 1084).

52 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

51

Effect of input on production Syntactic question forms that children hear may be an important factor in the way their interrogative behaviour develops, along with semantic, pragmatic, discourse and non-linguistic conceptual factors. Until recently, very few studies addressed specific aspects relating to the interrogative syntactic forms that the child receives or to whether these forms are modified by the interlocutor. Cross (1977) shows that modification takes place in the question types used to the child. Bloom, Merkin and Wootten (1982:1088) found that whquestions with descriptive verbs (verbs that are not the pro-verbs, such as do, go, happen) were used by the child contingently (after any five previous discourse turns within which the adult had mentioned the same verb), for example: Adult: Let’s see what I can find. Child: What can you find? Why questions shared the adult verb the most frequently, and it was found that questions in which there was a descriptive verb could be traced back to the input in a prior adult utterance. Stromswold (1995) questioned the results of Ervin-Tripp (1970) and Tyack and Ingram (1977), and examined parental input to explain her finding that children produce whobject questions correctly before wh-subject questions and found that the frequency of such questions in input only partially accounted for the order of acquisition. Rowland and Pine (2000) looked at subject–auxiliary inversion in whquestions and, using data from Adam, found a period of non-inversion (Brown, 1973). They propose that the more frequently a wh-word and auxiliary collocation occurs in the input, the more likely it is to result in inversion, and they argue for the importance of input and the child’s interpretation of it. It may be that what the child externalises through comprehension (or not) of the question reflects only part of the process that is taking place internally. Van Valin (2002) proposes that the findings of Rowland and Pine (2000) cannot be explained by input frequency but rather in terms of Role and Reference Grammar, which accounts for both grammar-based and inputbased perspectives of language acquisition. Van Valin suggests three alternative patterns for different types of children, which apply not only to wh-questions but also to yes/no questions. In his view, this provides an explanation for the different patterns of wh- and yes/no inversion and noninversion that have appeared in the literature (see above). In a reply to this, Rowland and Pine (2003) argue that, although linguistic theories may be

53 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

Early Trilingualism

52

useful tools for describing adult phenomena, they prefer an ‘input- driven’ approach to explain children’s acquisition data, since ‘the speech that a child hears is perhaps the most direct, easily observable influence on language acquisition’ (Rowland & Pine, 2003: 211). Valian and Casey (2003) describe an intervention experiment to see exactly how the use of structured input can affect the acquisition of syntactic regularities, specifically wh-questions and in relation to auxiliary insertion and inversion. The experiment found that the children who received more practice were performed better at forming the questions correctly.

Pragmatic Aspects of Children’s Acquisition of Question Requests Early language and pragmatic competence From a very early age, children have the ability to make themselves understood to those closest to them, even though their language may be extremely limited. Early functional acts (Bates, 1976; Bruner, 1975; Dore, 1979) and similar notions such as Halliday’s (1975) ‘act of meaning’ have been proposed as the primary unit of pre-speech communication. The capacity of caregivers to make rich interpretations of the child’s utterances and gestures in relation to context, assists communication (Brown, 1973). Children’s telegraphic speech and gestures (Carter, 1974; Dore, 1973, 1979; Halliday, 1975) have characteristics such as proposition and illocutionary force that conform to the models proposed by Austin (1962) and Searle (1969). As more language forms emerge, these ‘primitive’ speech acts (Dore, 1973) increase in complexity and type (Ervin-Tripp, 1977) until, by the age of three, children’s ‘intentions and propositions become grammaticalised as full speech acts’ (Dore, 1975, 1978: 94). In Bialystok’s view ‘children are not aware of the words they use to express their intentions, and the categories they form are built out of the semantic properties of those intentions, not out of the formal means for expressing them’ (Bialystok, 1993: 49). Early pragmatic competence is based on the child’s understanding of meaning in contexts, and Macnamara (1972) sees the ability to understand utterances in context as the very basis of language acquisition. Linguistic approaches to language acquisition assume that language has a structure or grammar that is independent of language use (Chomsky, 1957, 1965, 1982). However, according to the interactionist approach, language development is the result of the interactive context that surrounds and includes the dialogue between caregiver and child. Vygotsky (1978) explores the social origins of cognitive processes and proposes that any function in

54 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

53

the child’s cultural development appears first as an inter-psychological category established between people. Interactional contexts associated with mother–infant game and book-reading routines have been widely studied (Bruner, 1983; Snow et al., 1981; Snow & Goldfield, 1983). Bruner (1978) terms such routines a Language Acquisition Support System and proposes (Bruner, 1978) that the caregiver plays a facilitative role via a number of strategies such as scaffolding which support the development of discourse (Döpke, 1992; Stubbs, 1983). Wells (1985: 404) refers to the way caregiver and child interact as ‘the collaborative construction of meaning’. Studies in the acquisition of question requests in children below the age of four reflect the importance of interaction as researchers focus on different aspects such as: (1) the production of question requests (Bates, 1976; Carter, 1974; Dore, 1973; Ervin-Tripp, 1977; Holzman, 1972; James & Seebach, 1982; Przetacnik & Ligeza, 1990; Ninio & Snow 1996); (2) the understanding of question requests (Bates, 1971, 1976; ErvinTripp, 1977; Gordon & Ervin-Tripp, 1984; Ninio & Snow, 1996; Przetacnik & Ligeza, 1990; Shatz, 1978; Steffensen, 1977); (3) caregiver’s question requests (Gordon & Ervin-Tripp, 1984; Holzman, 1972; Ninio & Snow, 1996; Olsen-Fulero & Conforti, 1983; Steffensen, 1977); (4) caregiver feedback (Corsaro, 1977; Olsen-Fulero & Conforti, 1983; Przetacnik & Ligeza, 1990); (5) repair of question requests (Marcos, 1991; Ninio & Snow, 1996; Sachs & Devin, 1976). Directives and requests In the literature on pragmatic development, reference is made to directives, instrumental moves and requests. Fraser (1976) elaborates on the work of Austin (1962) and Searle (1969) and provides an analysis of the class of speech acts known as directives and a taxonomy of requests. In Fraser’s (1976: 51) view a directive is ‘an act in which the speaker intends his utterance to count as an attempt to get something done’. Ervin-Tripp (1977: 172) defines directives as ‘an instrumental use of language to achieve other ends’. Directives may be expressed through gesture, imperatives, hints and requests. Garvey (1975: 62) describes requests as speech acts that ‘reflect(s) an intersection of social and linguistic competence’. Requests are a type of directive in which ‘the speaker attempts to get the hearer to act on the basis of a sense of mutual co-operation’ (Prinz, 1982: 142). Ervin–Tripp and Gordon (1986) explain that, within instrumental

55 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:06

Color profile: Disabled Composite Default screen

Early Trilingualism

54

language, there are different degrees of conventionality. Non-conventional instrumental moves are those that achieve the goal of a request without it being specifically stated. For example, ‘It’s freezing in here’ prompts the hearer indirectly to fulfil the conventional request of ‘Could you close the door?’ by responding to the information in the message rather than by imposing on the hearer directly (Ervin-Tripp & Gordon, 1986: 63). Requests can be described as conventional instrumental moves in which the utterance is clearly asking the hearer to perform an action. Most conventional requests take the form of imperatives and questions. Requests are early to emerge (Bruner et al., 1982; Ervin-Tripp, 1977; Snow et al., 1996) and have been widely studied since they are also frequent in child language and show sensitivity to social features (Corsaro, 1977; Ervin-Tripp & Gordon, 1986; Garvey, 1975; Marcos, 1991; Prinz, 1982). Ervin-Tripp and Gordon (1986: 68) suggest that in order to request successfully ‘a speaker must solve five problems: getting attention, being clear, being persuasive, maintaining desired social relations and making repairs’. Functional taxonomies for describing speech acts and the pragmatic use of language in children’s speech have been proposed by Dore (1974, 1977, 1978, 1979, 1986), Tough (1977), McShane (1980), Halliday (1975), Wells (1985), Ninio and Wheeler (1984) and others. A number of analyses of children’s directives and requests have also been carried out in functional frameworks using taxonomies designed specifically for categorisation of requests (Ervin-Tripp, 1970, 1977; Ervin-Tripp & Gordon, 1986; Gordon & Ervin-Tripp, 1984; Prinz, 1982). It is important to be clear about what is meant by function, in this context, and within the theory of speech acts. Speech acts simultaneously manifest the structure, content and function of language where the structure of the speech act is its grammar, the content consists of the conceptual substance of the proposition and (...) constitutes what is talked about, (and) its function is its illocutionary force which consists of the speaker’s intentions and expectations. (Dore, 1978: 108) In Dore’s view, function characterises the utterances of children at an early age and captures what they know about using language. He suggests that ‘perhaps the most significant development in terms of language acquisition is the three-year-old’s control of form and functions’ (Dore, 1979: 353). The analyses of the communicative acts and functional uses of children’s speech described in the literature are many and varied (Corsaro, 1977). Ninio and Snow describe the situation as one in which: There are no ‘natural’ criteria for discriminating among distinct types of

56 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:06

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

55

communicative acts. The variety of category systems used in research on pragmatic development clearly demonstrates that intuitions about how to categorise and distinguish communicative acts differ fundamentally. There is also no agreed-upon theory of communicative uses of speech from which similar category systems could have been derived by different research groups. (...) Since different researchers have used mutually incompatible typologies of communicative acts, comparing and synthesising their various findings is nearly impossible. (Ninio & Snow, 1996: 18) Further problems relate to generalising from the findings of studies carried out on small samples of very young children, and the usefulness of such findings for comparisons with slightly older children (for example, Bates, 1976; Dore, 1974; Halliday, 1975). Another point is that those studies that in Ninio and Snow’s (1996: 19) view ‘(use) the most detailed and theoretically founded systems’ have tended to look at subsets of communicative categories, such as requests (Corsaro, 1979; Ervin-Tripp & Gordon, 1986; Garvey, 1975, 1977) or repetitions (Keenan, 1977), and so their findings cannot be applied to general pragmatic abilities. Moreover, the comprehensive systems by Tough (1977), Dore (1978) and McShane (1980) which do code the speech acts of older and younger children are considered by Ninio and Snow (1996: 19) to be ‘totally inconsistent with one another’. Ninio and Snow (1996: 19) consider this lack of consensus to be due to ‘the absence of a firm theoretical basis for deriving a systematic category system’ resulting in a situation where ‘the level of detail adopted by a particular investigator may reflect personal preference rather than any theoretical stance’. Ninio and Snow (1996: 21) propose the Inventory of Communicative Acts system constructed by Ninio and Wheeler (1984), which is based on utterance in relation to context, as a viable solution to these difficulties. To classify communicative contexts at an upper level, the system identifies 63 types of named interchange . The lower level classifies as speech acts the utterances contained in these interchanges, and again there are 63 categories (Ninio & Wheeler, 1984). This taxonomy has been used reliably by Ninio and Snow (1996) for the analysis of Hebrew and English. The pragmatic studies by Holzman (1972), James and Seebach (1982) and Przetacznik and Ligeza (1990) that classify question requests within functional frameworks all propose taxonomies with varying quantities and types of categories. James and Seebach summarise the findings of earlier studies of function in children’s questions: The results of studies of both adults’ and children’s questions seem to indicate that questions serve different pragmatic functions. The func-

57 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

Early Trilingualism

56

tional categories used in these studies have varied, but three general pragmatic functions appear in every classification system: the information function, the directive function and the conversational function. Although there may be subcategories, these three functions appear to be the basic ones served by questions. (James & Seebach, 1982: 3) The support provided by the interactive context surrounding a young child enables him or her to develop not only linguistic but also pragmatic competence. Requests are among the earliest speech forms produced by young children, and they exhibit both of these competences. However, research into requests and their functions in children’s pragmatic development has not so far been able to describe a coherent pattern; this is due, in part, to a wide variation in research methods. The next section outlines studies using functional taxonomies to focus on question requests between the ages of two and four, together with relevant parts of general studies of directives and requests.

The Acquisition of the Functional Use of Questions Between the Ages of Two and Four The pragmatic function of children’s directives has been studied for infants younger than two (Bates, 1976; Carter, 1974; Dore, 1973, 1974; Halliday, 1975) but, since at that age the grammatical rules regarding questions are generally not yet formalised (Ervin-Tripp, 1977), such studies are mainly concerned with early vocalisations, cries and gestures that are not addressed in the present study. Adaptations from Piaget The function and categorisation of questions in child development has long interested psychologists and linguists (Hollingworth, 1928; Smith, 1926; Snyder, 1914; Stern, 1924; Sully, 1896; Trettien, 1904). Piaget’s (1926) study of child language also focuses on the functional uses of questions, principally those starting with why. This study, which also classifies the functions of questions not starting with why, seeks reasons for questioning behaviour in relation to child logic. The taxonomy was used on a child of six, but was subsequently adapted for younger children. In the 1930s, Piaget’s (1926) work prompted three linguistic studies on questions (Davis, 1932; Lewis, 1938; Smith, 1933) that examined both form and function. Davis (1932) conducted a functional analysis in which the questions were classified according to Piaget’s (1926) categories:

58 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

(1) (2) (3) (4) (5) (6) (7) (8) (9)

57

causal questions; reality and history; actions; name (questions relating to rules and usage); classifications and evaluation; rules; calculation; social relations; rhetorical questions.

Davis (1932) added the two last categories to Piaget’s original list of seven. She examined the distribution of questions among these categories by age and gender, showing that there are important differences between girls and boys. In her study, girls asked more about social relations, particularly relating to ‘permission’ and ‘subordination of self’ (Davis, 1932: 67). The boys asked more about cause, classification and definition. In general younger children asked more questions of cause than older children did. It would be interesting to compare the findings on gender if the same study were carried out today. The study also looked at the pragmatic functions of questions relating to addressee in relation to animacy, familiarity, age, sex and social status. It noted that a large percentage of questions were directed to the mother and a small percentage to the father, again possibly reflecting the societal norms of the day. Finally, Davis examined the possible origins of the questions in relation to the child’s experience and interests and looked at the way the questions were sequenced logically, concluding that there were no age or gender differences. In her study, Smith (1933) investigated age, gender, situation, form, frequency, function, child-adult interaction and child–peer interaction. Piaget’s (1926) 7-category taxonomy was again used, this time with the addition of a clarification category termed ‘What say’ and a category of directives with ‘the nature of imperatives’ (Smith, 1933: 207). In Piaget’s category of ‘causal questions’, Smith reported an increase with age of causal questions using why, and also that boys asked more of such questions than girls, confirming Davis’s (1932) results. In the ‘reality and history’ category, questions starting with where were most frequently asked by two-year-olds and this use decreased with age. According to Smith’s data ‘actions’ contains the largest proportion of questions, but questions of the type in the ‘name category’ were not produced until the children were four or five years old. Questions of ‘classification and evaluation’ were found to be the most nearly constant in proportion, and varied little over age. Questions of the type found in the ‘calculation’ category appeared at

59 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

Early Trilingualism

58

age four to five and were predominant in boys. The functional use of tag and negative yes/no question forms was described where the child sought confirmation or approbation of actions. Like Piaget (1926), Smith considered the development of why at some length. Both Davis (1932) and Smith (1933) reported that questions are more frequently asked of adults than of peers indicating, in Davis’s view, that adults are seen by the children as an authoritative source of information. Meyer and Shane (1973) in a study comparing Davis’s (1932) data with their own, reported that in both studies the functional category system developed by Piaget was sufficiently clear to permit objective categorising of the questions, and that the general trend of their question data within categories was quite similar to that found in Davis (1932). Taxonomies for the classification of functional use Lewis (1938) followed on from the work of Davis (1932) and Smith (1933), but did not use Piaget’s taxonomy. Lewis analysed question development from a functional perspective based on four developmental stages and pointed to the importance of social cooperation in the growth of various categories of questions, noting that these findings may be idiosyncratic with respect to the individual child under observation who was aged between 0;10 and 2;10. Lewis proposed that questions fulfilling the following functions emerge by stages: (1) (2) (3) (4)

the child demands linguistic intercourse; questions preliminary to action, within the present situation; questions with reference to an absent situation; questions independent of action.

The child’s question utterances were evaluated in terms of ‘speech-acts’ (Lewis, 1938: 152) and the paper focused throughout on their value to him and the way they develop in discourse with the adult, also examining his early spoken responses to questions. It suggested that around stages (3) and(4) more complex ‘ true information questions’ of the following types emerge: (1) (2) (3) (4)

where questions; naming questions; source or goal questions; why questions.

Holzman (1972) examined the relationship between question functions in the questions of both children and mothers, using three subjects (Adam, Eve and Sarah) from Brown’s (1968) data. Two samples were taken for each

60 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

59

child at Brown’s Stage III (Brown, 1973), averaging approximately 350 utterances per sample, and five functional categories were used to analyse them. (1) (2) (3) (4)

children’s requests for information; children’s requests for behaviour; examples of test questions; children’s interrogatives that suggest a communication difficulty, including Huh? What? Mm? and repetitions of part or all of the mother’s preceding utterance; (5) locutions that are interrogative in form but do not have the force of questions. Holzman (1972) provides many examples from the child data, as well as a section on use of the interrogative ‘what dat’ which she claims was used by all three children for purposes other than eliciting information. She also points out that the function of test questions when used by the child seems to differ from that used by the mothers, and the questions seem to serve as conversation initiators. Four 100-utterance samples from three mothers were also investigated and had five similar functional categories. Holzman found that interrogatives comprise 15–33% of mothers’ utterances, and that interrogatives whose purpose is not to seek information but to make suggestions, evaluations or reports of the child’s behaviour occur with all mothers. Sachs and Devin (1976) investigated the degree of politeness of fouryear-old children’s speech to listeners of different ages and categorised questions into eight functional categories: (1) (2) (3) (4) (5) (6) (7) (8)

external world; internal state; request; imperative; confirmation; attention getting; instruction; repetition.

The findings indicated that the children were able to adjust the form of their questions depending on whether or not they were speaking to adults, peers or baby dolls. Prinz’s (1982) study of pragmatic aspects of English requests in languagedisordered children aged 3;6–9 years old presents a detailed model for

61 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

Early Trilingualism

60

coding all types of requests. and comprises three main categories with subcategories: (1) purpose of request; (2) directness of request; (3) surface form of request (8 sections). Furthermore, a series of rules regarding use of intonation and paralinguistic clues such as eye contact and gesture were applied to the categories to obtain more accurate distinctions. Prinz found that the form and function of requests appeared in a parallel fashion in his disordered subjects, though their development was like that of normal pre-school children three or four years younger (Garvey, 1974). James and Seebach (1982) investigated pragmatic function in children’s questions in English in four groups comprising 24 children of mean age 2;4–5;7 who were observed during approximately 10 half-hour sessions. Three functional categories are analysed: (1) information seeking; (2) conversational; (3) directive. Few questions were produced in the two and three-year-old groups, where desire statements (Ervin-Tripp, 1977) were still frequently used. The distributional patterns for pragmatic function of questions were similar in both groups. Some 80 % of the questions served the information-seeking function and 15–20% served the conversational function. No two-year-olds and 5% of the three-year-olds used the directive function. The study also reports on the emergence of yes/no and wh-questions and relates them to use in the different functions. Yes/no questions (in the information seeking function only) and what questions are reported in the two-year-old children. The three-year-olds produced three new question types (where, why and who), but only in the information-seeking function. The study seems to confirm the evidence produced by Werner and Kaplan (1963) of the general developmental principle that new forms serve old functions and old forms serve new functions. Wells (1985) described child language between 15 and 60 months from the point of view of syntax, semantics and pragmatics. Pragmatic analysis was carried out at the level of clause, utterance, subsequence (or exchange) and sequence. In Wells’s (1985: 62) complex taxonomy, ‘sequences are classified according to the (four) sets of dominant (interpersonal) purposes they are taken to be designed to achieve’.. These dominant purpose groups are based on the function groups suggested by Searle (1977) and contain a

62 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

61

number of interpersonal functions. Function is defined as the smallest unit in the analysis of interpersonal purpose and 56 functions are available to convey the different purposes. Nine interpersonal functions related to interrogatives were found to be in use by 42 months (Wells, 1985: 366) and these are shown in Table 2.3. Wells also classified a mood system with five options, indicative, imperative, polar interrogative, wh-interrogative and moodless. Wells describes the relationship between mood and function as being one of probability: Sentences in the declarative and wh-interrogative mood have a high probability of association with, respectively, the giving and requesting of information; and polar interrogatives have a reasonably high probability of association with requests for confirmation ... (or) ... with the present or future actions of the participants. (Wells, 1985: 173) Wells (1985: 173) claims that choice of mood gives the cue to the functional interpretation that may be placed on content and form. Dore (1986: 39) provides a list of conversational act types used in coding the language of nursery school children aged 3;0 to 3;6. He divides the category of requestives into two sections: questions, where the speaker wants the hearer to provide information and requests where the speaker wants the

Table 2.3 Interpersonal functions in relation to interrogatives Interpersonal function

Purpose of sequence

Direct request

control

Content question

representational

Contingent query repetition

procedural

Yes/no question

representational

Request permission

control

Indirect request

control

Request justification

control

Question (state)

expressive

Contingent query

procedural

Source: Wells (1985: 366)

63 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

Early Trilingualism

62

hearer to perform an action. The questions section contains only interrogatives and codes three types: (1) choice questions seeking yes/no or either/or judgements; (2) product questions seeking information through the use of basic whforms; (3) process questions seeking explanations via more complex wh-forms such as why, how, what for, and how come. The requests section contains directive forms, including interrogatives such as: • action requests (e.g. Can you please do it?); • permission requests (e.g. May I go? Can I do it?). Dore (1986) used this taxonomy to examine the coherence of pre-school children in fantasy play and note increasing flexibility in the use of utterance functions. He speculated that this might be a result of greater independence and a change in values found in play outside the immediate home and family and in closer contact with peers. Przetacnik-Gierowska and Ligeza (1990) extracted question data from the speech journals of four Polish-speaking children ranging from 1;3–7;3 years of age and analysed it within functional categories. Their taxonomy comprises three main categories measuring a total of seven subcategories: (1) exchange of information; (2) exchange of emotions and attitudes; (3) initiating or maintaining interaction. Questions were analysed individually and as part of a question series. The first question in each series was noted in order to differentiate each series. Adult response was also studied along with the relationship between the speakers, and four types of information exchange were identified. Przetacnik-Gierowska and Ligeza (1990) found that informational questions increase with age in line with the children’s cognitive development. They claim that questions serve a dual purpose of developing both cognitive and linguistic growth, in addition to fulfilling social needs. In their view, as long as a child is in interaction with partners who stimulate these activities, informative questions will increase. However, Przetacnik-Gierowska and Ligeza point out that this type of interaction may not always be available – in day-centres for example. Snow et al. (1996) examined a number of pragmatic aspects of the development of children’s communicative abilities during the second and third year. They videotaped 52 children from English-speaking families with

64 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

63

their mothers in a semi-structured laboratory play environment at 14, 20 and 32 months. Using the Inventory of Communicative Acts-Abridged (INCA-A) (a shortened version of a system developed by Ninio and Wheeler 1984), the children’s communicative attempts were coded at interchange and speech act level to measure pragmatic flexibility (Snow et al., 1996: 62). Interrogative behaviour was measured in the following speech acts at utterance level in the category of ‘questions and responses’: • • • • • • • • • • • • •

wh-question (ask a product question); yes/no question (ask a yes/no question); restricted alternative; answer a wh-question with a statement; answer in the affirmative to a yes/no question; answer in the negative to a yes/no question; answer a question with a wh-question; answer a question with a yes/no question; non-satisfying answer to question; refuse to answer; eliciting question (e.g. hmmm?); aggravated question, expressing disapproval by restating a question; answer a limited alternative question.

The study reports that children begin asking wh- and yes/no questions at 32 months, although they have already been hearing and answering them at 14 and 20 months, perhaps because ‘questions are requests and as such demand a response’ (Snow et al., 1996: 94). Characteristics of functional use Garvey (1975) studied children’s ability to convey and respond to direct and indirect requests for action without the presence of adults. The requests analysed included direct and indirect question requests, and request directives. Some 36 English-speaking children aged between 3;6 and 5;7 were observed in peer dyads7 at play in a laboratory. Older dyads produced more indirect requests than younger ones, exhibiting a tendency with age towards more adult-like requesting strategies, as described by Ervin-Tripp (1977). Ervin-Tripp (1977) classified children’s directives taking into account the characteristics of those used by adults. Data collected from studies in languages such as Turkish (Aksu, 1973), Hungarian (MacWhinney, 1974), Italian (Bates, 1976) and English (Garvey, 1975; Holzman, 1972) was compared in order to find examples showing lack of explicitness present

65 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:07

Color profile: Disabled Composite Default screen

Early Trilingualism

64

in some adult forms. Five characteristics were found in the directives of children aged between two and four: (1) (2) (3) (4) (5)

telegraphic directives; limited routines; hints without explicit imperatives; elaborate oblique stratagems; embeddings and structural modifications.

Gordon and Ervin-Tripp (1984), and Ervin-Tripp and Gordon (1986) examined the acquisition of instrumental language and non-conventional instrumental moves in children. Their main focus was on the acquisition of requests and the pragmatics of politeness within a general framework based on the variables that may affect what the child says and the strategies employed by the child to express his or her wishes. Their findings seem to be summarised in Gordon and Ervin-Tripp who say: our discussion of the development of instrumental language has shown that by age five children attend to a wide range of variables, both social and situational, and can co-ordinate them with a considerable variety of instrumental strategies and moves. (Gordon & Ervin-Tripp, 1984: 310) They stress the importance of the child’s developing ability to apply causal reasoning and social understanding in assessing his or her linguistic choices. Newcombe and Zaslow (1981) carried out a study on 11 children aged 2;6 based on the criteria suggested by Ervin-Tripp (1977), in order to find out if children used hints and questions directives in a similar way to adults. They found that at age 2;6 children do not clearly hint, but rather use repetition of the same or a slightly stronger form if their request is not dealt with. The functions of children’s questions have been studied for over a century in a number of differing taxonomies, all of which confirm that several important speech acts are involved. Requests (which are among the earliest speech acts to emerge) soon take on the form of questions as language develops; these questions will become increasingly sophisticated in their level of politeness and explicitness as the child acquires pragmatic competence. However, it is not yet clear exactly how this developing competence is distributed to the different question forms available to the child, partly because of methodological differences between the different studies that have been carried out and partly because of a lack of consensus as to how speech acts and language functions should be classified.

66 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:07

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

65

Interaction and the Acquisition of Question Function Research into interaction in learners of English as second or foreign language has demonstrated the importance of interactional feedback to language development (Gass, 1997; Pica, 1994). In other studies of child second language (L2) learners, it has also been shown that feedback during interaction facilitates L2 development (Mackey & Oliver, 2002). According to Long’s interaction hypothesis, feedback obtained during conversational interaction promotes interlanguage development because interaction connects input, internal learner capacities and output in productive ways (Long, 1996). Interaction may also be beneficial because it provides learners with exposure to negative feedback (NF) in response to their non-target-like utterances (Mackey et al., 2003: 36). Such negative feedback involves repetition, confirmation checks, and clarification checks and re-casts from the interlocutor in order to overcome breakdowns in communication (Oliver, 2000: 120). Requests for clarification or confirmation may occur ‘in response to grammatical as well as ungrammatical utterances either as part of normal conversational routines or as a result of ... limited proficiency’ (Mackey et al., 2003: 38). Long (1996) describes such feedback for language learning in terms of negotiation for meaning. It may be that a similar type of process is taking place in caregiver–infant dyads at an earlier age and in natural contexts. Studies within the socialinteractionist approach attempt to account for language development in terms of adults mak(ing) the child’s task easier by pacing the complexity of the data ... children who are linguistically naive will receive grammatically simple input. As the child grows older the data provided by the environment will also increase in complexity. (Bohannon & Bonvillian, 2001: 290) Studies carried out in developmental pragmatics describe interaction in relation to the child’s acquisition of requests and interrogative behaviour. These studies fall into four main areas, with some overlap: (1) question request input via the caregiver’s questions and caregiver feedback to the child’s requests; (2) the child’s understanding of requests and feedback; (3) clarification requests; (4) repair, reformulation and repetition. Caregiver question requests and feedback Holzman (1972) suggests that adults use questions for a variety of pragmatic functions. She found that the function of ‘test’ questions (about

67 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

Early Trilingualism

66 S and H

Ia

Preparation of propositional content

S

Ib

Adjunct to request

II

Request

S H and S

(III)

V

H

IV

Acknowledgement of II

............................................................................................................................................... ............................................................................................................................................... S

(V)

Acknowledgement of IV

Figure 2.1 Structural domain of request Source: Garvey (1975: 49, Fig. 1). S = speaker, H = addressee. Optional elements are enclosed in parenthesis. Dotted lines indicate potential location of repeated request if IV is other than compliant acknowledgement.

known information) when used by adults does not seem to be the same in children’s questions since the adults use them to give the children an opportunity to display their knowledge. Garvey (1975: 45) presents a model to describe the structural domain of request interactions (see Figure 2.1) in a study of how children aged 3;6 to 5;7 make and respond to requests. Garvey later refers to the results of the 1975 study, which underline the significance of input to the children’s acquisition process: Responses to requests for action (RA) often invoked the same meaning factors (or conditions) that are postulated to underlie such requests. Furthermore, the same factors could be expressed to convey an RA indirectly ... 3½–5½year-old children were cognizant, in some way, of the relevant conditions that underlie the RA in the speech of adults. (Garvey, 1977: 65) Shatz (1979) analysed two group of mothers questions to an older and younger age group of children aged 18–34 months and found that mothers tended to use ‘paradigmatic frames’ (specific forms) for the following eleven functions.8 (1) (2) (3) (4)

testing: What does the cow say? directive: Can you make them slide? floor offer: Who’s gonna make the lunch? maintaining contact: You making the camel again?

68 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

67

(5) (6) (7) (8) (9)

requesting clarification: What? requesting information: Do you like the wagon? challenge: Does it go in there? calling attention: You know what this is? requesting confirmation: Maybe we better get the top back on the tent, do you think? (10) giving encouragement: Is that easier? (11) expressing empathy: That’s good isn’t it? The form–function pairings were not found to increase in complexity in relation to the child’s age and perceived understanding of them. Instead it was found that mothers would use form-function pairings irrespective of the age of the child. Shatz maintains that: differences in the frequency of functional uses observed in the two groups of mothers may depend as much on their need to recruit the child’s abilities for conversational participation as on their concern to simplify form-function relations for teaching purposes. (Shatz: 1979: 1098) In their study on children’s questions James and Seebach (1982: 3) note that the pragmatic functions mentioned by most investigators of adult questions to children include: (1) (2) (3) (4)

requesting new information and/or clarification; directing or requesting behaviour; maintaining conversations; testing (requesting information that the speaker already possesses).

Olsen-Fulero and Conforti (1983) studied 11 mothers and their children aged 2;5–3;0, and describe the mothers’ questioning behaviour and the children’s responsiveness within a functional framework of 7 categories: (1) (2) (3) (4) (5) (6) (7)

repairs: The doggie’s doing what? test: Is this tree red or green? real: Do you have to go to potty? verbal reflective: He does? Oh really? action reflective: Are you putting the dolly to bed? report questions: That car doesn’t fit? permission requests/offers of help: May I sit here?

Different types of question from the mother result in different types of feedback from the child. Olsen-Fulero and Conforti (1983: 495–496) examined the characteristics of mothers’ speech and the constraints on mothers’

69 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

Early Trilingualism

68

questioning behaviour, and point out that the predominance of questions reported in many studies reflect their importance in mother–child interaction (Remick, 1972; Sachs et al., 1972; Sachs & Devin, 1976; Snow, 1972). Olsen-Fulero and Conforti suggest that frequent use of questions by mothers may serve to make children feel the constraint of turn allocation as described by Blount (1977). By at least three years of age, children are much more likely to successfully take a conversational turn following a question than following a declarative sentence (Bloom et al., 1976). Olsen-Fulero and Conforti point out that a mother may use a question type too heavily, thus producing an idiosyncratic response style in her child (Olsen-Fulero & Conforti, 1983: 505). Furthermore, mothers were found to frequently usurp the children’s answers by beginning another conversational turn before the child had replied to the question. Olsen-Fulero and Conforti (1983: 499– 500) suggest various reasons for parental usurpation, including lack of interest in the child’s response. The amount of usurpation is found to correspond to the functional type of the question. Lucariello et al. (1986: 142) found significant differences in the function of maternal questions in different contexts. ‘Real questions’ that seek genuine information from the child via a linguistic response were used more in ‘event contexts’ than in other contexts. Event contexts are defined as ‘situations for which each member of the dyad has, and thereby shares, event knowledge ... to facilitate or enrich mother–child use of language’ (Lucariello et al., 1986: 140). Lucariello et al. (1986) distinguish between ‘real’ questions and ‘test’ questions for which the mother (and it is assumed the child) already has an answer and which were least used in the ‘event context’. Factors relating to sources of input other than the mother are not widely studied. There is a tendency for studies on child language acquisition to be carried out on dyads of a middle-class English speaking mother and first-born child, though Döpke’s (1992) study looked at aspects of father and sibling input in bilingual children. Indeed, in other cultures the primary caregivers may be siblings (Schieffelin, 1979). As far as caregiver feedback to the child’s questions is concerned, children are much better at maintaining conversation with adults than with their peers (Ninio & Snow, 1996: 151), since adults are willing and able to ‘engage in extensive repair’ and the continuation of the topic may depend on assistance from the adult that peers are not able to provide (Bloom et al., 1976). Furthermore the answers that parents give their children are important not only for the children’s linguistic development but also for their knowledge system (Przetacznik & Ligeza, 1990: 95). We must not overlook the fact that children ask questions, presumably, because they want to know some-

70 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

69

thing. Alternatively they may want to satisfy social or emotional needs. Both James and Seebach (1982) and Przetacznik and Ligeza (1990) have, we have seen, proposed functional taxonomies for the information that the children seek and suggest that it is up to the goodwill of parents and educators to supply these needs. The child’s understanding of requests and feedback Ervin-Tripp (1970) undertook one of the earliest studies of children’s understanding of questions and their capacity answer them. She suggests that discourse agreement predates the child’s second birthday, and that it is possible for such young children to distinguish between yes/no and whquestions. These findings are born out by Crosby (1976). Ervin-Tripp (1977) describes five basic types of requests, and hypothesises that the order of comprehension is as follows: (1) (2) (3) (4) (5)

imperatives; embedded imperatives; declarative statements; interrogatives; general statements of external condition.

Bates (1971) reports that children more often responded verbally to requests that were interrogative in form. Garvey (1975: 62) argues that the hearer’s ability to verbally express willingness to comply or not (thus affecting the speaker’s subsequent behaviour) shows that requests reflect an intersection of social and linguistic competence. Snow (1977) notes that mothers use questions to generate non-verbal and verbal responses in infants from a very early age. Steffensen (1977) examined verbal and nonverbal responses to yes/no questions in two children aged between 1;5 and 2;2 and found that, though the children did not give answers that were appropriate by the conventions of adult speech, the answers did seem to show that the children recognised the need to verbalise in response to the interrogative. Shatz (1978) describes evidence of children’s comprehension of direct and indirect request structure from a study of two-year-old children, and finds that the children gave appropriate responses most of the time (either verbally or through actions), generally when the requests were framed in the interrogative, imperative or embedded imperative. OlsenFulero and Conforti (1983) also report on the understanding of requests and find that not all types of mothers’ questions obtain an appropriate response. Children are most responsive to the questions that most benefit them, such as repairs and permission requests. Ledbetter and Dent (1988) report on two studies on listener compliance on the effects of adult request

71 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

Early Trilingualism

70

structure within a functional directive typology and find that responsiveness to requests relates to request structure, linguistic complexity and age (Ervin-Tripp, 1977; Ervin-Tripp & Gordon, 1986). Ninio and Snow (1996: 146) examined the extent to which children are ‘skilled conversationalists’ and consider that in some areas, such as turntaking, children are early developers, whereas in other area such as ‘ topic relevance or observing rules of timing and obligations to respond, they are considerably less precocious’. Questions are important in conversational exchange since they must be either acknowledged or answered if the rules for turn exchange are not to be violated (Ninio & Snow, 1996: 145). Children produce answers to questions before they produce questions themselves whereas responses to statements emerge later (Ninio & Snow, 1996: 94). Babelot and Marcos (1999) relate the capacity of children aged 1;7–2;5 to understand requests with their ability to reply to them, and show that context of utterance had an influence on replies and that by 2;6 children are able to relate the context of an utterance to its form. Clarification requests An area that has received considerable attention in the study of child– adult questioning behaviour is that of the clarification request, and its subgrouping of confirmation. Corsaro (1977) analysed the form and function of the clarification request used in the interactive behaviour of the parents, other adults and peers with two brothers aged 2;6 to 3;8 and their elder sister, and comments on the possible effects of interactive style. He concludes that the adults’ consistent use of the clarification request obliges the child to provide ‘subtle background cues’ to keep the interaction going, and this process serves to socialise the child into more abstract cultural norms (Corsaro, 1977: 205). Garvey (1977) produced a typology for the description of clarification requests or ‘contingent queries’ – this term is used to describe clarification requests that are close to or ‘contingent’ to the unclear utterance. The requests are coded as follows: • non-specific request for repetition, e.g. Y: Father, can I come? X: What? • specific request for repetition, e.g. Y: We found a parrot in our house. X: A what?

72 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

71

• specific request for confirmation, e.g. Y: If you’re not a scaredy-cat. X: Scaredy-cat? Y: Uh-huh. • specific request for specification, e.g. Y: Joe Glick knows a friend of yours. X: Which one? Solicited and unsolicited requests for clarification or confirmation formed the focus of Garvey’s work on contingent queries (1977). Solicited queries promote mutual attention or rapport, whilst unsolicited queries serve the purpose of maintaining mutual understanding (Garvey, 1977: 91). The forms and use of contingent queries are reported to have been learned by age 3–3;6. Garvey emphasises the importance of clarification as a speech act, and it is found to have implications for the development of discourse and communicative competence. Wells (1985: 407) in his discussion of aspects of parental–child interaction leading to the ‘collaborative construction of meaning’ also makes use of Garvey’s typology. Shatz and Watson (1990) used an adaptation of Garvey’s (1977) typology to examine the clarification requests and responses of children aged 2;6 and their parents. They found that by age three children do not have a full understanding of the use of clarification requests since they often failed to use them appropriately. However, they suggest that the children’s efforts to do so demonstrate that they understand the importance of maintaining a conversation. Ninio and Snow (1996: 167) describe 7 types of clarification demands used by mothers: (1) (2) (3) (4) (5)

rerun requests; yes/no questions that were full repetitions of the original utterance; yes/no questions that were partial repetitions of the original utterance; direct queries of the meaning of a word the child used; yes/no questions that substituted new forms for all or part of the original utterance; (6) yes/no questions that were repetitions, either partial or full, of the original utterance, but that also added some new material to the original utterance; (7) wh-questions that demanded product completion concerning a missing component of the utterance. Ninio and Snow found that the clarification demands were fine-tuned to

73 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:07

Color profile: Disabled Composite Default screen

Early Trilingualism

72

the children’s linguistic level and that use of the different types was adjusted with age. Repair, reformulation and repetition We have seen that apart from promoting turn-taking, many adult questions to children serve to clarify or confirm the utterance that the child has just made and that the children learn to do this too. However, there is also evidence of the way adults repair and reformulate questions when communication has failed to obtain a reply. Question reformulation (reiteration) and repetition occur when the mother is motivated to obtain an answer, and therefore intensifies the questioning (Olsen-Fulero & Conforti, 1983: 507). Sachs and Devin (1976: 84) point out that ‘Mothers often repeat utterances to young children, either exactly or maintaining the same general semantic content while changing the linguistic form’ (Ervin-Tripp, 1970; Kobashigawa, 1969; Snow, 1972). Wells (1985: 404) reports that parents imitate and expand children’s answers through the use of confirmation questions. It seems that children also realise that their request has been misunderstood or has not been acknowledged (Golinkoff, 1986; Shatz, 1983), and reformulate the request (Marcos, 1991) and retry (Ervin-Tripp, 1988) in order to repair the conversational breakdown or achieve compliance. Bates (1976) found that children of two and a half will add politeness markers, such as question intonation to requests in order to make them more polite if their initial request is not successful. However, Newcombe and Zazlow (1981) found in their study of two-and-a-half-year-olds that, in spontaneous speech, only 7% of second tries were more polite. Ervin-Tripp and Gordon (1986) report that increased politeness does not seem to affect compliance with the request. Ervin-Tripp (1988) examined retries in children aged between two and six years, and found that children changed tactics with age and depending on whether the retry followed refusal or ignoral. Repetition also has an important, though underestimated, role to play (Ochs-Keenan, 1977). Children also sometimes use repetition of the adult utterance, for the purposes of clarification or otherwise. Ochs-Keanen (1977: 131) found that a child may imitate selectively in order to ‘satisfy his obligations as a conversational partner’ and notes that ‘the child’s response (for instance, partial imitation) shows that the child treats the investigator’s utterance not as a model to be imitated, but as a question to be answered’ (Ochs-Keenan, 1977: 130). In Ochs-Keenan’s view, through imitation the child is learning ‘ the human uses of language, what Dell Hymes (1972) has called communicative competence’ (Ochs-Keenan, 1977:133).

74 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

73

Overall, the findings of the studies outlined above show that the interaction between child and interlocutor reveals complex relationships in the area of question requests. The known importance of parental questioning behaviour is confirmed, and it has been seen that parental questions serve different functions in different contexts. Through the use of clarification techniques, parental questioning behaviour guides the child towards the development of discourse frameworks. Children are able to recognise forms and functions of adult questioning behaviour before the age of two, and respond to those that are most beneficial to them.

Interrogative Behaviour in Bilinguals Question form in bilingual children Several studies of the syntax of bilingual children who have English as one of their languages describe the acquisition of English question form although there is not, to our knowledge, any longitudinal study that looks expressly at the questioning behaviour of bilingual and trilingual children. An early reference to questions in bilinguals is found in Smith (1933: 204), who collected data from 46 bilingual children between the ages of two and six years, though the other language(s) are not specified. Except at age three, the bilingual children asked a lower proportion of questions to sentences than the monolingual children. At age three the bilingual children produced far more utterances than at any other age, in other words they began to say a lot. Even so, the amount of questions they produced did not increase proportionally. A cross-sectional study by Swain (1972) looked at two French/English bilingual children, one aged from 3;2 to 3;9, the other 4;0 to 4;5, and compared them with monolingual children. Swain found the acquisition of structures was similar in both the bilingual and monolingual children, but the first use of yes/no questions appears later in the bilingual children than in the monolingual children. De Houwer (1990) extracted all wh-questions and yes/no questions that contained a subject and a finite verb from her English corpus of the Dutch/ English bilingual child, Kate, between the ages of 2;7 and 3;4. There were 98 questions that met these criteria, and all except 2 of them were adult-like in terms of inversion and auxiliary insertion. The data for order and frequency of wh-forms coincided with those reported in monolingual data (Tyack & Ingram, 1977; Wells 1985). Most auxiliaries used in yes/no questions matched those described by Fletcher (1985) in his monolingual subject Sophie at age 3;5. Kate’s word order in questions ‘reflects the usage patterns produced by ... some ... monolingual English-speaking children’ (De Houwer, 1990: 272).

75 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

Early Trilingualism

74

When De Houwer (1990) compared findings on interrogatives in Dutch and English, a number of interesting points arose. Firstly, although the total proportion of interrogatives is the same in both languages, yes/no questions appeared with more frequency in English than in Dutch, and wh-questions were more frequent in Dutch. De Houwer (1990: 284) suggested that these tendencies are due to maternal speech patterns in questioning behaviour, since it has been shown that Dutch mothers use wh-questions almost three times as often as they use yes/no questions (Snow et al., 1976). Furthermore, English-speaking mothers have been shown to use slightly more yes/no questions than wh-questions (Furrow et al., 1979). Secondly, De Houwer noted that the child commenced questioning behaviour (with wh-questions first, and then yes/no questions) earlier in Dutch than she did in English (where yes/no questions appear first, followed by wh-questions). However, De Houwer (1990: 287) did not see this as evidence of Dutch question formation going ahead of English, but rather as the child using ‘quite distinct means employed in question formation available in the separate input systems’. Lastly she compared frequencies of declaratives, yes/no questions and wh-questions in the child’s English with similar data from Wells (1985: Table A17) on monolingual English children and finds they are identical at 82% (declaratives), 10% (yes/no) and 8% (wh-questions). She tentatively suggested (De Houwer, 1990: 290) that this similarity, along with the corresponding input findings for mothers mentioned above, may be explained by ‘some general phenomenon pertinent to the organisation of discourse interaction’. Overall, it seems clear from De Houwer’s study that this bilingual child had acquired question form in a way that reflects that found in monolingual peers. A study by Pérez-Vidal (1995) of a Catalan/English bilingual child between the ages of 3;2 and 4;2 also described the acquisition of questions along with the child’s general linguistic development in English. The child, who was exposed to Catalan approximately two-thirds of the time (JuanGarau & Pérez-Vidal, 2001: 63), did not show a very developed interrogative system in English during the period under study. The speech sample recorded at age 3;2,28, includes 14 interrogatives, but only 3 of these (all of which are wh-questions, 2 with what and 1 with where) are considered by the researcher to be ‘complete, intelligible and spontaneous’ (Pérez-Vidal, 1995: 225). There are no yes/no questions, although there are 4 affirmative statements with question intonation. In the next sample examined, at 3;4;12, no yes/no questions appear either, though there are 7 affirmative statements with intonation. There are 10 wh-questions using what (4), where (5) and who (1). By the third sample, at age 4;0;14, the child had not progressed further in interrogation, since only 15 examples are produced,

76 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:08

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

75

predominantly of the affirmative-with-intonation type. The researcher saw this as ‘really strange’ (‘verdaderamente curioso’) and explained it as possible transfer from Catalan, where intonation is used on the affirmative to mark interrogation (Pérez-Vidal, 1995: 338). Like De Houwer’s (1990) work, this study also suggests that parental input is important in the acquisition of question form, but from a rather different perspective. The researcher (Pérez-Vidal, 1995: 113, 299) noted that the father, the principal source of English input to the child, frequently contracted the question auxiliary do, which may have been imperceptible to the child and she suggests that this might explain the child’s failure to produce auxiliaries in question forms. Furthermore the father was often observed to ask questions consisting of utterance with rising intonation of the type most frequently produced by the child. The parental input the child receives in English may therefore be partly responsible for the style of questioning the child uses. The two studies described above therefore show differing patterns of development. In De Houwer’s (1990) study, the way the child developed the interrogative systems of her two languages was similar to that of monolingual English and Dutch children. The child described by Pérez-Vidal (1995) showed some delay in the acquisition of questions; few questions were produced and there was little evidence of auxiliary use. Questions were predominantly formed by the affirmative with intonation. Both studies recognised the effect of parental input on questioning behaviour as important. Clarification requests in bilingual contexts Several studies on bilingual children examine interrogation from a pragmatic perspective, particularly in relation to the ability to understand requests for clarification and conversational repair (Comeau & Genesee, 2001; Juan-Garau & Pérez-Vidal, 2001; Lanza, 1992, 1997, 2001). Lanza (1992: 650) explains that in the following five parental discourse strategies the minimal grasp category and the expressed guess strategy (proposed by Ochs, 1988) correspond to clarification request types found elsewhere in developmental literature as e.g. ‘requests for repetition’ and ‘requests for confirmation’. These strategies are placed along a continuum that stretches from the negotiation of a monolingual context at one end to a bilingual context at the other (Figure 2.2). (1) Adult requests clarification: minimal grasp strategy (Ochs, 1988). (2) Adult requests clarification: expressed guess strategy (Ochs, 1988). (3) Adult repetition of the content of the child’s utterance using the other language. (4) Move on strategy: the conversation merely continues.

77 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

Color profile: Disabled Composite Default screen

Early Trilingualism

76 Monolingual

Bilingual

context

context Minimal grasp

Expressed guess

Adult repetition

Move on strategy

Codeswitching

Figure 2.2 Parental strategies towards child language mixes Source: Lanza (1992: 649, Fig. II).

(5) Adult code switches (Lanza, 1992: 649). In the minimal grasp strategy the parent attempts, through the use of whquestions and expressions (such as ‘I don’t understand’ or ‘say that again’) to get the child to repair and re-say the utterance in a clearer way. In the expressed guess strategy it is the parent who reformulates the utterance, most usually in a yes/no question that the child may then confirm or not. When a parent is bilingual, yet wishes to promote the use of only one of the languages with the child, it is the degree of mixing that the parent is prepared to accept that will define the parent as monolingual or bilingual. In other words, the more the parent accepts both languages, the more the child will do the same and the parent will not be ‘negotiating a monolingual context’ ( Lanza, 1992). Lanza (1992: 647) notes that the two-year-old Norwegian/English bilingual girl in her study responded to a request for clarification not only by repairing the utterance but by recognising that the error in question was related to language choice. The child, who was speaking to her Englishspeaking mother, inserted a Norwegian word, yet changed it to English when her mother requested clarification. By employing this minimal grasp strategy, in Lanza’s (1992: 648) view ‘the mother has conveyed to her child a meta-communicative message, namely, “this is a context in which to speak English only”’. Siri: Mama lope (run) Mother: What do you want Mama to do? Siri: Run. However, with the expressed guess strategy the parent may demonstrate understanding of the non-parent language word by translating it into the parent language for use in the expressed guess: Siri: Tiss (pee) Mother: Aw, is he peeing? Siri: Yeah.

78 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:07

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

77

In cases where this happens, the context is more bilingual than monolingual (Grosjean, 1982, 1985, 1988). Lanza (1992: 652) also describes the use of the request for clarification hm? by the mother in different exchanges and notes that the child is able to interpret it differently as: (1) a cue to expand her utterance; (2) a request to repeat her utterance more loudly; (3) a request to repair her utterance by replacing a Norwegian verb with its English equivalent. Interestingly, the child is reported not to repair language mixes in clarification requests made by her Norwegian father. It seems that the child does not perceive the problem he wants clarifying as being related to language mixing, since he does not use strategies that insist on a monolingual context as her mother does. Lanza (1997, 2001) links such strategies with language socialisation: If the parent normally responds with negative sanctioning to the child’s mixing and negotiates a monolingual context with the child, the child is socialised into language separation. If on the other hand, the parent negotiates a bilingual context with the child, the child is socialised into language mixing, or code-switching. (Lanza, 1997:269) Ochs (1988) also sees clarification sequences as tools for socialisation in Western and other societies. Döpke (1992: 56) mentions the importance of the clarification request or ‘the “what” strategy’ as a way to instil the need to speak the second language in bilingual children who might otherwise become receptive bilinguals (able to understand a second language, but not to speak it). She also describes insisting strategies, which are constraints requiring the children to answer in the appropriate language. Through a system of translations and questions the parents display different degrees of understanding that put constraints on the child’s response. A correlation is found between the level of insisting strategy used by the parents and the child’s degree of bilingualism. Comeau and Genesee (2001) focused on the way children’s communicative competence develops through their ability to make repairs in response to clarification requests. In order to repair, bilingual children must determine whether the request for clarification is related to a change in language or some other aspect. Comeau and Genesee examined recordings of twelve 3-year-old and six 5-year–old French/English bilingual children in play with an experimenter in the child’s non-dominant language. The experimenters were bilingual, but made every effort to appear not to know the

79 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

Color profile: Disabled Composite Default screen

Early Trilingualism

78

dominant language; it was deemed preferable to use bilingual speakers rather than monolingual speakers as experimenters, since bilinguals could more easily determine the language of the child’s utterance and adjust the clarification requests accordingly. The study aimed to see if children could identify language choice as the cause of communication breakdown by making appropriate repairs. They found that most of the children were capable of maintaining use of their non-dominant language during the play sessions. This confirms the findings of Genesee, Boivin and Nicoladis (1996) that, with an unfamiliar interlocutor, young bilingual children are able to select language appropriately. The older children had fewer languagebased breakdowns because they were better at keeping to one language with the interlocutor. Although the younger children were evidently capable of identifying language-based and non-language-based breakdowns, they used inappropriate strategies when attempting to repair language-based breakdowns. This may have been due to factors relating to performance (such as lack of vocabulary in the interlocutor’s language), and the researchers conclude that ‘it appears that young bilingual children possess the same ability to repair as monolingual children as well as abilities that are specific to bilinguals’ (Comeau & Genesee, 2001: 254). Juan-Garau and Pérez-Vidal (2001) examine parental input strategies in a family of a Catalan/English bilingual child. Following Lanza’s (1992, 1997) categorisation for mixing, the Catalan-speaking mother is described as negotiating a bilingual context of interaction with the child, whereas the English-speaking father uses strategies to negotiate a monolingual context. The parents use Catalan to each other. Juan-Garau and Pérez-Vidal (2001: 74) note that ‘in bilingual child–parent interactions, requests for clarification can function as a cue to switch languages.’ It is claimed that through the consistent use of the minimal grasp and expressed guess strategies (Lanza, 1992, 1997) from the child’s age of three onwards in his requests for clarification, the father helps his son to recognise both himself and his father as monolingual speakers, which boosts the child’s linguistic competence. In this way the father, who is a minority language parent in this particular linguistic context, was able to make his son a productive bilingual able to both understand and speak in two languages. The research described above examines parental strategies and the way these are interpreted by the child when making a response to interlocutor feedback. However there have been no studies of how bilingual and trilingual children set about making their own requests (clarification or otherwise), through the use of their linguistic and pragmatic knowledge. The present study deals with the acquisition of English questions but, since the subject is trilingual, it is possible that there may be some cross-

80 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

79

linguistic influence from questioning behaviour in Basque and Spanish. For this reason a brief description of the acquisition of questions in Spanish and Basque follows.

The Acquisition of Questions in Basque Basque is non-Indo-European, highly inflected SOV (subject object verb) language that allows variation in word order and exhibits head-final order (in other words the most important information comes towards the end of the sentence, not at the beginning as it does in English). It has 16 morphological cases and is ergative and agglutinative9 (Saltarelli, 1988). Except for some modal particles and the interrogative particle al, no element can intervene between the main verb and its auxiliary in affirmative sentences. Verbs agree with their subjects, direct objects and indirect objects and all three arguments can be left lexically empty. In Basque the nucleus of the sentence is placed to the right of the sentence. Question formation in adult Basque Like the rest of the Basque language, wh-forms are marked for case. Our description here will be limited to questions containing the interrogative pronouns that form the basis for the case system and are also used in isolation in adult language. These are also the first forms to emerge in child language interrogatives. Simple yes/no questions will also be described. Wh-questions

When wh-questions are formed the wh-element is fronted (moved to the left) and the subject is inverted to final position. The verb is placed immediately after the wh-word (Barreña, 2001). The following examples are from Zubiri and Zubiri (1995: 589): • Zer ekarri du gizonak? (What brought has man the) What has the man brought? • Zer da hori? (What is that?) What is that? • Non dago (Where is Where is Miren?

Miren? Miren?)

81 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

Color profile: Disabled Composite Default screen

Early Trilingualism

80

• Noiz hasiko da ikastaroa? (When begin (future) is the course?) When will the course start? • Nor da mutil (Who is boy Who is that tall boy? • Nork (Who (ergative) Who hasn’t paid?

altu hori? tall that?)

ez du ordaindu? not has paid?)

• Zergatik ez zinen afarira (Why not you+past the dinner Why didn’t you come to the dinner?

etorri? to come?)

Yes/no questions

Yes/no questions can be formed either by adding intonation to the affirmative: • Mikelekin joango (Mikel with go (future) Will you go with Mikel?

zara? are you?)

or by introducing the particle al: • Etorriko al zarete bihar? (Come (future) particle you (plural) tomorrow?) Will you come tomorrow? Question formation in child Basque The following examples of the early questions of a monolingual Basque child Oitz are from a selection in Barreña (1994) who describes the development of word order in relation to the positioning of wh-. (In each example, Oitz’s age is shown on the right, in the usual form of years;months,days.) Wh-question

• Non? Where?

(1;08.29)

NP wh-question

Here the wh-question refers to the noun phrase: • Hori norena? (That whose?) Whose is that?

(1;08.29)

82 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

• Ni non? (Me where?) Me where?

81

(1;10.12)

Barreña (1994) notes that in these questions the wh-element is placed in the final position and that a similar type of question is sometimes used by adults. • Bere aita nun? (His father where?) Where his father?

(2;00.26)

Adv Wh

The child learns that the wh-question can be placed after an adverbial: • Hor nor? (There who?) Who’s there?

(2;00.05)

(X) wh V

The child learns that wh- may be placed in the medial (middle) position before a verb. • Hori non (That where Where paint that?

pintatu? to paint?)

(2;00.05)

• Nik hau nora bota? (I this where to throw?) Where (shall) I throw this?

(2;01.19)

• Tibis nor (Tibis who Who is Tibis?

(2;00.26)

da? is?)

Wh NP

The child learns that wh- may be placed in the initial position in front of a noun phrase: • Non lehioak? (Where lions?) Where the lions?

(2;07.05)

Wh V- NP

The child learns that wh- may be placed in the initial position in front of a noun phrase:

83 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

Color profile: Disabled Composite Default screen

Early Trilingualism

82

• Nork izetu argia? (Who switch on light the?) Who switched on the light?

(2;04.11)

(X) Wh V+Y

The child learns that further information (that is not necessarily a noun phrase) may be added to the end of a question: • Honek zer esaten dio mutiliari? (2;08.13) (This (person) what say aux boy the to?) What does s/he (this person) say to the boy? It can be seen that as sentence complexity increases, wh- moves across to the left into the initial position.

The Acquisition of Questions in Spanish Spanish is an inflected, SVO (subject verb object) Romance language in which the nucleus is placed to the left of the sentence. Verbs agree with their subjects in person and number, and double negation is permitted. Determiners, nouns and adjectives are marked for gender and number. Spanish can have null subjects, where the subject position is left lexically empty. Question formation in adult Spanish Wh-questions

As in English, the interrogative pronoun is placed in sentence initial position in Spanish. When wh-questions are formed, the wh-element moves to initial position on the left and there is inversion of the subject to final position. In wh-questions wh-movement is obligatory. The first example shows how inversion is necessary to form a correct wh-question with the subject, Jorge, moving to final position in three steps. • Jorge (Jorge

ha has

dicho said

algo something)

*¿Qué Jorge ha *What Jorge has said?

dicho?10

¿Qué ha dicho (What has said What did Jorge say?

Jorge? Jorge?)

As in Basque, the verb is positioned after the wh-word, e.g.

84 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

83

• ¿Qué ha traído el hombre? (What has brought the man?) What has the man brought? Yes/no questions

In yes/no questions, the marker of interrogative is rising intonation on any declarative sentence (Fleta, 1999: 140). Question formation in child Spanish Hernandez Pina (1984: 242–244) describes the order of emergence of interrogative pronouns in a Spanish monolingual child (again, the child’s age is shown on the right): • ¿Qué? What?

(1;10)

• ¿Qué es? (What is?) What’s that?

(1;10)

• ¿Qué haces? (What do you?) What are you doing?

(1;10)

• ¿Qué tal? How are you? (formula)

(1;10)

• ¿Por qué? Why?

(1;10)

• ¿(es)tá taza (Is (location – where is) cup Where’s baby’s cup?

nene? baby?)

(2;0–2;4)

The verb form allows omission of the wh-word without misunderstanding: • ¿(es)tán (Are (location: where are) Where are the slippers?

zapatillas? slippers?)

(2;0–2;4)

The verb form again compensates for the wh-omission: • ¿Dón(de) (es)tá(n) las papas? (Where are (location) the potatoes?) Where are the potatoes?

85 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:09

(2;0–2;4)

Color profile: Disabled Composite Default screen

Early Trilingualism

84

• ¿Dón(de) pusiste (Where put+you+past Where did you put the purse? • ¿Dónde (es)tá (Where is(location) Where is the bird?

la cartera? the purse?) pájaro? bird?

(2;0–2;4)

(2;0–2;4)

• ¿Cómo dice el pato? (How goes the duck?) How does the duck go?

(2;4)

• ¿Cuándo pinta(d)o papeles? (When painted papers?) When did you colour the papers?

(2;5)

• ¿De quién (Of who Whose is that cat?

(2;6)

es is

ese that

gato? cat?)

A study has been carried out on the acquisition of wh-movement in two Spanish monolingual children, one Basque/Spanish bilingual child and a monolingual Basque child (Barreña, 2001). The study finds that in the bilingual child Spanish wh- questions follow the adult model. Basque wh- questions develop in a similar way in both bilingual and monolingual Basque children and also follow the adult model, though through a slightly slower process, possibly due to the complexity of Basque grammar. Notes 1. A lexical verb is one that carries meaning in the clause, and is neither an auxiliary verb nor a modal verb. 2. A copula verb refers to the verb ‘to be’ when used in clauses with a complement, noun or adjectival phrase, e.g. he is a friendly man. 3. The abbreviations used here are as follows: cop = copula verb, NP = noun phrase, S = subject, aux = auxiliary, X = complement, V = verb, A = adverb. 4. Government and Binding is a syntactic theory in the tradition of transformational grammar proposed by Chomsky in the 1980s. 5. The CHILDES database is a collection of language transcriptions and other data collected from children that can be accessed to obtain samples for research purposes. 6. Studies on inversion with wh- may examine questions involving: (1) wh-movement of the type: I gave him a gun. You gave him what? What did you give him? (2) wh-object questions such as Who did he see? Why did he do it? Where did he go? (3) wh-subject questions such as Who ate the cake? What broke the window? In this case inversion is non-obligatory (Stromswold, 1995).

86 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

The Development of Interrogative Behaviour

85

7. A peer dyad is two people of the same age (in this case children) interacting with each other. 8. Examples are taken from Shatz (1979: 1095), where they are contextualised in discourse exchanges. For reasons of space, these exchanges are not included here. 9. An ergative language is one in which the use of a transitive or intransitive verb can be seen in the subject of the verb. An agglutinative language is one in which the words are formed by adding morphemes together. 10.* is used to indicate that the sentence or phrase that follows is ungrammatical.

87 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Part 2

The Acquisition of English Question Form and Function in a Trilingual Child

88 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Chapter 3

Research Questions and Design Specific Research Issues The present study focuses on the acquisition of interrogative behaviour in English in a trilingual child who has received limited exposure to English. In order to understand and comment on the child’s interrogative behaviour, we must first ascertain her general proficiency in English. Various research studies on the acquisition of English as a first language have described the changes that take place in the language of monolingual children between their first words and the age of five (Brown, 1973; Fletcher, 1985; Wells, 1985). Other studies such as De Houwer (1990) and Pérez-Vidal (1995) have described the acquisition of English in children who are bilingual from birth. Standard measures of language growth in the field of child language acquisition include mean length of utterance (MLU), which describes development in length of utterance, and type token ratio (TTR), which describes development in language richness. So the first research issue for our study is: • how the child’s general acquisition of English takes place as compared to other children in monolingual or bilingual environments in terms of mean length of utterance (MLU) and type token ratio (TTR). The acquisition of question form involves a number of features. General growth in questioning behaviour is measured by counting the number of words in questions and the number of questions to utterances (Davis, 1932; Meyer & Shane, 1973; Smith, 1933). In the case of yes/no questions, development involves going from a declarative with intonation to an adult-like inverted question with appropriate auxiliary insertion (Cazden, 1970; Fletcher, 1985; Klima & Bellugi, 1966; Smith, 1933). The acquisition of adultlike wh-questions shows a developmental sequence for using the different wh-forms and also requires inversion and auxiliary insertion (Erreich, 1984; Ingram & Tyack, 1979; Tyack & Ingram, 1977; Wells, 1985). Apart from these major classes of question forms, other types (such as tag questions) are also found (Greenbaum & Quirk, 1990). The second issue investigated in this study concerns:

89 89 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Early Trilingualism

90

• the way the development of the child’s question form takes place in terms of length of question, number of questions, and the type of question produced. Several studies have described the development of pragmatic competence in monolingual children with reference to question functions (Davis, 1932; Dore, 1979, Holzman, 1972; James & Seebach, 1982; Smith, 1933; Snow et al., 1996; Wells, 1985). So the third issue for our study relates to: • how a trilingual child uses her questions functionally at different ages and as part of her communicative competence. Other studies have described the characteristics of the interlocutor’s interrogative behaviour and its relation to the ability of monolingual children to comprehend and use the pragmatic functions of questions (Babelot & Marcos, 1999; Ervin-Tripp, 1970, 1977; Garvey, 1975; Holzman, 1972; James & Seebach, 1982; Ninio & Snow, 1996; Shatz, 1979; Snow et al., 1983), and feedback to the questions asked by the interlocutor has also been examined (Bates 1971; Ervin-Tripp, 1970; Ninio & Snow, 1996; Olsen-Fulero & Conforti, 1983; Steffensen, 1977). Comeau and Genesee (2001) have demonstrated that bilingual children are able to select their language in response to clarification requests. Taking into account research studies on pragmatic competence in relation to input and feedback, another research issue addressed by the present study is: • whether the child’s functional use of questions and her ability to give feedback to questions is related to the input and feedback she has received. Research studies on bilingual and trilingual acquisition report that various amounts of mixing occur across the languages. Mixing has been used as evidence both that the child’s languages are developing as a single system (Redlinger & Park, 1980; Taeschner, 1983; Volterra & Taeschner, 1978; Vihman, 1985) and that they are developing as separate systems (De Houwer, 1990; Genesee, 1989; Meisel, 1989, 2001; Murrell, 1966; Quay, 1995). Cross-linguistic influence such as lexical and syntactic mixing and code switching have also been seen as evidence of the child’s developing language competence (Hoffmann, 1985). The amount of mixing in the child’s language(s) is also viewed as an indication of the child’s perception of the interlocutor as bilingual or not (Grosjean, 1994, 1997; Juan-Garau & Pérez-Vidal 2001; Lanza 1997; Nicoladis & Genesee, 1998). Taking into account the different theories on language separation and mixing in early bilingual acquisition the final research issue concerns: • the extent to which there is evidence of cross-linguistic influence in the child’s production.

90 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Research Questions and Design

91

The Subject The subject of our study is a trilingual girl, Jenny, who was recorded between the ages of 1;11.23 and 3;6;17. She is the youngest of three children in a trilingual family and has two elder brothers, Mikel and Jon Ander, who were aged 5;10–7;4 and 3;11–5;5 respectively during the period of recording. The family lives in a small industrial town in the Basque Country of Spain, where both Basque and Spanish are spoken. As is the case throughout the Basque Country, Basque is the minority language. However, the family lives in the province of Guipúzcoa in an area where Basque is widely used. Jenny’s father has Basque as his first language and his family are Basque speakers though, like all adult Basque speakers, he is bilingual in Spanish. He has Cambridge First Certificate level English, and has a degree in economics and works in the finance department of a local company. Jenny’s mother was born in England, has a degree in Spanish and has lived in Spain since 1975. She has studied Basque to lower intermediate level, and is a teacher trainer at university. Jenny’s parents spoke Spanish together for six years prior to the birth of their children and then changed (not without great effort), to speaking English in the presence of the children. Harding and Riley (1986: 79) report that they know of no cases where language change between spouses has been achieved. However, each of Jenny’s parents now uses his or her native language with the children, and Spanish has been virtually excluded from the home except for occasional television programmes and visits from Spanish speakers. The decision to do this was taken in order to promote the two minority languages, Basque and English. The aim was to make the children trilingual as it was certain that they, like all other Basque speakers, would not fail to acquire Spanish through exposure to it outside the home. In the process of promoting English as the language of the family, except when Basque is used between the father and children, the parents have achieved a high level of passive knowledge of each other’s mother tongue. As a result each parent is able to get the gist of what was said to the children by the other parent without showing overt understanding of the message. This is of particular value to the mother in enabling her to maintain her ‘English only’ status while at the same time not feeling excluded. Jenny has been exposed to English, Spanish and Basque since birth and her relationship with each of these languages will be examined in turn. Her main source of English has been the family environment, principally her mother and to a lesser extent her brothers and the interaction of the four of them. English is the language used by all family members in their activities; only the father and the children use Basque. Jenny has never been abroad,

91 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Early Trilingualism

92

but has had frequent opportunities to be in touch with English through various visits lasting between four days and a week from English-speaking family and friends (at ages 0;2, 0;3, 0;10, 0;11, 1;9, 1;11, 2;1, 2;3, 3;1, 3;2 and 3;3). The family also have a number of local friends with a similar linguistic background with whom they spend time and speak English. Indeed, Jenny has a school playmate from a trilingual background and the teacher has reported that the girls sometimes use English together at school. Since she was baby Jenny has also been exposed to English through educational and entertainment videos and television programmes, games, songs and rhymes, and has been read stories in English by her mother. Television viewing for the family consists mainly of English language videos and satellite programmes. Jenny was introduced to her Spanish-speaking caregiver within hours of being born, and therefore can be considered trilingual from birth (De Houwer, 1990: 11). The caregiver had been the family’s domestic helper for six months before Jenny was born, and continued to be present throughout Jenny’s mother’s maternity leave. She is a monolingual speaker of Spanish from the south of Spain, and has two teenage daughters. Jenny was cared for all day in her own surroundings by the Spanish caregiver until the age of 2;5, when she started Basque-medium nursery school. We can estimate that, excluding holidays and weekends, Jenny received at least a third of her language input in Spanish between the ages of 0;5 (when her mother returned to work) and 2;5 (when she began nursery). The input was related to feeding, dressing, the daily routine and domestic tasks, and did not involve play and storytelling like the parental input. The caregiver reported this herself when she was asked to ‘play’ with the child for the purposes of data collection in Spanish, saying that she found this very hard to do. Jenny went on holiday for two weeks to the south of Spain with her family at 1;1 and for one week at 2;3 and again at 3;3. These holidays involved plenty of contact with monolingual Spanish speakers. Jenny has been exposed to Basque since birth through her father, his family and the local Basque-speaking community within which the parents tend to operate. Jenny has been widely exposed to Basque videos, television programmes, games, songs and storybooks, both at school and at home. At the age of 2;5 Jenny started going to Basque-medium nursery school in the mornings. This meant that her exposure to Basque was dramatically increased and her contact with Spanish was reduced by three hours a day, although her caretaker continued to give her lunch and to look after her until her mother came at 5.30 – though part of this time included Jenny’s afternoon nap. At the age of 3;5, Jenny began to stay at nursery school all day until 4:30 and for lunch, which meant that the time she was

92 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Research Questions and Design

93

exposed to Spanish was reduced to an hour a day, that is to say 5 hours a week apart from those times she had contact with Spanish outside the home and with her Spanish-speaking playmates. From her birth until the age of 2;5, when she started nursery school, Jenny had been living in a trilingual environment where she received input from her three languages in similar proportions. From 2;5, and then 3;5 onwards, her exposure to Spanish and Basque began to change and we would expect that from then on her Basque would continue to develop ahead of Spanish as is suggested in Cenoz and Barnes (1997) and Barnes (2001). The period of data collection in English for this study (from 1;11–3;6) coincides with this period of change in exposure to Basque and Spanish, but these changes do not affect the conditions under which Jenny was exposed to English, which remained constant. She has no history of ear infections that might have affected her hearing. Several recordings unrelated to this study have been made of Jenny using Spanish and Basque. MLU and TTR have been calculated for these recordings and give an indication of the child’s balanced proficiency in all three languages, and in relation to peers. Sibling effects Jenny’s linguistic microcosm (De Houwer, 1995: 224) is complex, not least because she is the youngest of three children and receives input not only from her parents but also from her siblings. Siblings have been seen to serve as a bridge into different communities of speakers (Mannle et al., 1991). For this reason it is of interest to describe some general characteristics of the children’s language behaviour together. They have no problem with relating one parent to one language. There is little language mixing and when it occurs it tends to be into Basque where English loanwords relating to the home are used between the father and the children. All three children always use English to their mother and occasionally to their father. No instances have been observed where the children address the mother in Basque, except for occasional loanwords relating to school. The children’s loyalty to the use of English with their mother has sometimes proved problematic. Older Basque family members have expressed their feeling of exclusion (see also Romaine, 1989) when English is used at gatherings of the wider family where predominantly Basque, but also some Spanish, is spoken. The eldest boy will occasionally use a little Spanish to his mother when monolingual Spanish speakers are present, a practice that is common among Basque speakers, and he clearly realises that she is proficient in Spanish. The two younger children have never used Spanish with her. The children all attend Basque-medium schools and therefore it might

93 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Early Trilingualism

94

be expected that Basque, or even some Spanish through contact at playtime with children from Spanish-speaking home backgrounds, would be the languages used in conversation and play together. However, this is not the case, as the siblings use either English or Basque in the presence of the mother, the father or both parents. Basque tends to be the language chosen for construction games, boisterous play and sports whereas English is used in more creative types of play such as artwork and make-believe. This tendency reflects the type of activity carried out in the company of either parent. In all types of play a small amount of code-switching between English and Basque takes place. Use of Spanish by the children in the home is restricted to a few lexical items. The amount of each language used at school must remain an unknown quantity. The enthusiastic use of English in play by the children is largely due to the role of the eldest child in leading play and establishing English for that purpose among the siblings. He is the subject of other studies (Cenoz & Barnes, 1997; Barnes, 2001) that examine some features of his trilingualism.

Data Collection Case studies In the introduction to their case study of a Spanish/English bilingual child, Deuchar and Quay (2000: 2–4) point out some of the advantages and shortcomings of the methodology. One shortcoming that is mentioned is that generalisations cannot be made on the basis of case studies. However case studies reveal features that surely must form part of any generalisation, they may produce data that can always refute an existing generalisation and moreover case studies are a useful source of hypotheses (Platt, 1988). Case studies are seen as particularly suited to the study of language acquisition since they are able to capture the constant changes of the early linguistic period (Dromi, 1987). Bilingual and trilingual case studies like ours have been carried out by several other parent-researchers on one or two children (Deuchar, 1989; Faingold, 1999; Fantini, 1985; Hoffmann, 1985; Leopold, 1970; Pérez-Vidal, 1995; Ronjat, 1913; Saunders, 1982; Vihman, 1985), but the mother- researcher is not always the interlocutor. Deuchar(1989) observed her child interacting in English with the maternal grandmother and the Spanish- speaking father. In Pérez-Vidal’s study (1995), the mother-researcher recorded the Englishspeaking father with his child. Deuchar and Quay (2000) consider that there are advantages to being a parent-researcher such as having access to a potentially much larger sample of utterances in a wider range of situations. Parents also have the advantage of superior understanding of their chil-

94 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Research Questions and Design

95

dren’s utterances. Wells (1985: 147) points out that parents make the assumption that ‘children mean more than they are able to say, and what they are judged to mean is usually relevant to the context of utterance’ as is set out in Brown (1973). Davis (1932), in her study on children’s questions, chose mothers to collect the data as, in her view, mothers were better able to understand their children’s speech than outsiders and because children directed more questions to a parent, particularly the mother. Recordings The data for the analysis of Jenny’s acquisition of English were collected approximately fortnightly from the age of 1;11.24 until 3;6.18. Jenny was recorded in conversation with her mother on audio and videotape and 32 sets of tapes were produced. For the audio recordings a Sony TCM-359 pocket-sized cassette recorder was used with a clip-attachable separate mini-microphone with a long flex for ease of movement. The microphone was known to Jenny as ‘Birdy’ and, until she got used to wearing it, it was a source of some problems. The video recordings were made on a portable Hitachi camera with tripod. The recording sessions lasted between 45 minutes and an hour, and the first 30 minutes were transcribed. The recordings all took place in Jenny’s bedroom. In this way Jenny was able to interact with her mother in a similar way in each recording, and the context was the same. Most recordings were made in the late afternoon before the evening meal. A typical recording included activities such as playing together with bricks or beakers, talking about toys, family and clothes, looking at storybooks and photos, colouring, bathing and toilet. Jenny was permitted to do whatever she fancied. A number of topics have been identified during each conversation and these are listed in Table 3.1. Tapes were transcribed following the transcription conventions of the CHILDES Chat system (MacWhinney, 1995).

Analysis of Language Development English Mean Length of Utterance (MLU) and Type Token Ratio (TTR) were calculated for all 32 scripts using the Clan programme from CHILDES (MacWhinney, 1995) to obtain an impression of Jenny’s general development in English. MLU in words was calculated on all Jenny’s utterances in each script. TTR was calculated on the first 285 utterances in each script in order to standardise the results since this was the least number of utterances Jenny produced in any script (Script 3, age 2;1.2). All scripts were also examined manually for other developmental points of interest.

95 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Early Trilingualism

96

Table 3.1 The recordings: Jenny’s age, length of transcript and topics covered Script

Age

Pages Topic

1

1;11.24

17

Noddy, teddy, duck, balloon, toy eggs, dress, handkerchief, shoes

2

2;0.6

12

bricks, ears, cooking, feeding dolly

3

2;1.2

10

Birdy, puppet, kisses, toilet

4

2;1.26

17

Birdy, doll, changing nappy, feeding bottle

5

2;2.24

15

Muzzy, nap, pyjamas, toilet

6

2;3.14

16

clothes, paper, birthdays, teddies, jumping, injuries

7

2;4.2

13

books, swimming, helicopter, clothes, Birdy

8

2;4.17

10

books, doll, brothers, father, clothes

9

2;5.5

15

books, animals, keys, abacus,

10

2;5.19

13

toilet, bath toys, toy piano.

11

2;6.3

14

books, doll, toilet, bath toys, school teachers

12

2;6.19

14

bath, frog, pyjamas, lost teddy, book

13

2;7.1

12

hats, shoes, fitting toys in box, pyjamas

14

2;7.9

13

books, toybox, eggs, kitchen

15

2;7.19

12

toy farm, books, toybox, football

16

2;8.3

15

Noddy blocks and pictures, schoolwork, school, friends, toilet

17

2;8.19

11

bed, toilet, clothes, bricks

18

2;9.5

17

colours, books, dolls, feeding

19

2;9.22

12

teddy’s name, teddy book, jungle animals

20

2;10.2

15

bed, toy padlock, colours, eggs, books

21

2;10.14

13

duck, Birdy, names, colours

22

2;10.30

12

colours, doing puzzles

23

2;11.17

11

cooking, filling pots

24

3;0.5

14

Birdy, cooking, foods, birthday presents, friends, clothes, colours

25

3;0.19

8

general toys

26

3;0.30

9

teddy cloth book

27

3;1.9

10

puppets, zoo animals,

28

3;1.22

14

books, dolls, holiday, puppets, farm animals

29

3;3.8

13

the tea-set

30

3;4.7

10

opening the sack of old baby toys

31

3;5.18

15

tea party, family photos, animals, bath

32

3;6.18

11

bicycle, babies, Christmas, colours,

96 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Research Questions and Design

97

Basque and Spanish Jenny’s MLU in Basque and Spanish was calculated in order to demonstrate her degree of trilingualism. A large proportion of infants born in the Basque country with at least one Basque-speaking parent tend to be bilingual from birth since, although the Basque-speaking parent may use Basque to the child, Spanish will be used by the other parent. Spanish is also likely to be the language used between the parents because of its dominance and because all adult Basque speakers are also bilingual in Spanish. Furthermore, unless the child is in a Basque-speaking linguistic microcosm (such as in a small rural community), s/he will probably be exposed to a good deal of Spanish in the community and will develop bilingually to a greater or lesser degree depending on the languages used at home (two Basque-speaking parents, or one Basque-speaking and one Spanishspeaking parent). Ezeizabarrena (1996) describes the acquisition of verbal morphology in two bilingual Spanish/Basque subjects (Mikel aged 1;7–3;11 and Jurgi aged 1;10–4;0), and concludes that the children learn the two languages in a way that is similar to monolinguals and possibly parallel. The MLU in Basque and Spanish of the children in her study were used for comparison with those of Jenny, which were calculated on the same criteria.

Question Analysis Choice of taxonomy From the literature on the acquisition of questions, it can be seen that there are a number of perspectives from which the study of the acquisition of questions may be approached: (1) the syntactic perspective, which is principally concerned with acquisition of wh- within a framework of Government and Binding (Radford, 1994; de Villiers, 1995; Stromswold, 1995); (2) the examination of the order of acquisition of question forms in relation to developmental stages (Brown, 1968; Cazden, 1970; Tyack & Ingram, 1977); (3) the pragmatic perspective where the focus is on the development of the functional aspects of questions and the child’s growing ability to understand these functions, and use them appropriately (Holzman, 1972; Ervin-Tripp, 1977; Dore, 1986; Przetacnik-Gierowska & Ligeza, 1990). As Ninio and Snow (1996: 15) point out, it is impossible to separate form and function since a child’s capacity to use the language appropriately will

97 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:10

Color profile: Disabled Composite Default screen

Early Trilingualism

98

necessarily grow in relation to the amount and complexity of language she has at her disposal. For this reason it was important to measure Jenny’s question development in terms of both form and function. However, the primary classification of our data is into functional categories, focusing specifically on the way Jenny uses her questions for pragmatic purposes, then proceeding to the analysis of the forms she uses. The role of the prime interlocutor, in this case the mother, in providing the scaffolding for interaction was also considered vital, hence the decision to include data on the mother’s questions and her feedback to the child, as well as data on the feedback of the child to her mother. A taxonomy was designed to assemble this data, which needed to have validity. Dollaghan and Miller address the problem of validity in observational studies of communicative competence and refer specifically to the identification of requests for information: The perceived validity of a taxonomy or measurement procedure depends in large part on its comprehensiveness, or on the degree to which it addresses all of the events relevant to a particular theory. The test of the content validity of an observational instrument lies not in objective criterion but in the degree to which its users perceive it to be an accurate, logically and theoretically motivated reflection of the phenomenon of interest. (Dollaghan & Miller, 1986: 104–105) Dollaghan and Miller suggest that explicit recognition rules for the pragmatic phenomena under study should be agreed on by two observers. At the start of the present study, this was carried out by the researcher and a second linguist and forms the basis for the designation of the categories and the identification of the items to be included in each category. Nonetheless, use of the taxonomy necessitated minor adjustments as more and richer data become available with the passage of time. Analyses of interrogative behaviour were carried out on all 32 scripts, each of which was examined by hand by the researcher according to the following criteria and in the following order. Exchange

The first step in the analysis of the data was to identify the discourse units containing one or more questions from either participant or both. Olsen-Fulero and Conforti (1983: 499) refer to this discourse unit as a question sequence whilst Przetacnik Gierowska and Ligeza (1990: 86) use the term question series. Garvey (1977) calls them episodes. In the present study, these units are referred to as exchanges, a term used by Döpke (1992: 88) and

98 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Research Questions and Design

99

adapted from Stubbs’ (1983) model based on the basic category of the move proposed by Sinclair and Coulthard (1975): Verbal interaction consists of chains of moves which are related to each other. These chains are called exchanges. An exchange is a series of subsequent moves, each of which is conditioned by the move which immediately precedes it ... A minimal exchange consists of an initiation and a response or an initiation and a follow up. (Döpke 1992: 88) Interlocutor

The questions identified in the exchanges were classified according to the questioner (either the child or the mother) and then placed in a functional category. Functional category

The six functional categories available for the classification of the questions were: (1) (2) (3) (4) (5) (6)

real information; known information; confirmation; clarification; interactive; affective.

These are described in detail later in the ‘Functional categories’ section later in the chapter. Length

Once the question had been assigned a functional category, the researcher counted the number of words it contained, then examined it for form. Formal category

The next step was to classify the question according to its grammatical form as described by Greenbaum and Quirk (1990). The typology and examples can be found in ‘Formal categories’ section later in the chapter. Type of feedback

Feedback was classified as: (1) (2) (3) (4) (5)

none; acceptable; unacceptable; repetition; reformulation.

99 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Early Trilingualism

100

Examples are given in ‘Feedback categories’ section later in the chapter. A protocol was designed for classifying this information. For every script, 12 identical protocols were available, 6 for each interlocutor. Each protocol consisted of two main sections, one for recording details of the questions based on the formal categories for questions, and the other for recording feedback. Functional categories The taxonomy has four major categories plus two minor categories. The four major categories are: (1) (2) (3) (4)

request for real information; request for known information; request for confirmation; request for clarification.

These categories can be recognized in taxonomies such as those of Holzman (1972), James and Seebach (1982) and Shatz (1979). Confirmation and Clarification are designated two categories for the purposes of this study since these categories coincide with Lanza’s (1992) expressed guess and minimal grasp respectively, but elsewhere have been collapsed into one, as in Corsaro (1977: 185). There is a fifth category for those questions that are not as clearly definable as the others, and we have called this ‘interactive’. A unifying characteristic of interactive questions is they do not form part of the other four categories, and this is borne in mind when analysing the data. The real information, known information, confirmation and clarification categories are standard, and can be compared with each other and across studies whilst the interactive category cannot be contrasted except within itself. A sixth category, affective questions, was originally included in the present taxonomy. In the event, few examples appeared, and where they did, they were in the mother’s data. Nonetheless, since affective questions are clearly definable as such, it did not seem right to collapse this category into ‘interactive’. Request for real information

This category is used for questions considered to be seeking information the questioner does not have and which she presumably needs or wants to know. Once again, in this and the later examples, the script number (S) and Jenny’s age (in the form years;months,days) are shown on the right. CHI: What’s this colour? (S32. 3;6.18) MOT: Hmm? CHI: This?

100 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Research Questions and Design

MOT:

101

What’s that colour # it’s turquoise.1

Offers, suggestions including ‘we’ and ‘do you want...’ and requests for permission that required a response were also included in this category. The respondent’s answer to such questions (either in the affirmative or negative) will affect the asker in that she will take action on the basis of this information. Therefore asking the question serves the purpose of providing real information. Request for known information

This category is used for questions where the asker is presumed to already know the answer. The communicative purpose of the question is to pretend to seek novel information in order to initiate or maintain conversation or elicit answers. Uses of such questions include: CHI: Mummy. (S30. 3;4.7) CHI: Mummy. CHI: Where’s my thing for Leire? (showing an object in her hand) MOT: Would you like to give her xxx? CHI: That blue xxx. Request for confirmation (checking)

This category is used for questions referring back to something that has been said before by one of the participants and the original utterance has clearly been understood. These questions do not therefore have the full illocutionary force of a request for information but rather that of checking information already received. MOT: Right? (S25. 3;0.19) CHI: Ah here it is hairbrush. CHI: What is # hairbrush? MOT: Hairbrush yes that’s right. The aim of such questions is to maintain the conversation or to show interest in the interlocutor’s words. This category also includes expressions such as alright? and okay? for checking approval that has already been shown or is taken for granted (see also the interactive category below). Requests for clarification (repair)

This category is used when a previous utterance or utterances have not been understood and further explanation or repetition is required by the hearer. CHI: Sandelak. (S7. 2;4.2) MOT: You want to wear your sandals?

101 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Early Trilingualism

102

CHI: MOT: MOT: CHI: MOT:

Yes. I don’t know where they are. Do you know where they are? What? Do you know where your sandals are?

This category also contains requests for further explanation and requests for repetition such as: Hmm? Pardon? eh? and uhh? Interactive

This umbrella category contains a variety of question functions that do not fit into the previous four categories. They are termed ‘interactive’ since their use covers a variety of functions that will keep the interaction going rather than seek information. CHI: Er what’s (S14. 2;7.9) CHI: Oh what’s that? CHI: Mum what’s that ## book? MOT: What darling? CHI: The book. Jenny’s own questions to herself were also coded as interactive since they lack the illocutionary force of ‘request’ of the four previous categories above. Affective

This category contains questions that do not seek information, maintain the conversation, or direct attention, but which are markers of empathy. MOT: What’s the matter with it? (S32. 3;6.18) CHI: It was it was two on (wheels). CHI: Here two. CHI: And here two. MOT: Hmm. MOT: It’s sad isn’t it? The overall corpus was also examined for examples of the child’s use of the question functions in context as part of her communicative competence. Twenty samples were extracted and described, in line with the work of Wells (1985: 405–415), for the collaborative construction of meaning. Formal categories For the purpose of examining the development of form, questions were classified into various categories.

102 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Research Questions and Design

103

Wh-questions

Utterances with what, where, who, when, how (much/many) and why in initial position were classified as wh-questions. CHI: What’s it get in this here? (S16. 2;8.3) MOT: Where? CHI: This is hard. Non wh-questions

These questions were sub-divided into three categories: declarative questions, full yes/no questions and intonation questions. • Declarative questions MOT: CHI: MOT:

xxx xxx. In here? That’s right.

(S32. 3;6.18)

• Full yes/no questions These questions contain inversion of one or more elements as markers of question form, and may contain auxiliary insertion. CHI: Can we take this chair off? (S31. 3;5.18) MOT: Oh dear, careful now, mind your foot. • Intonation questions These are utterances that were partly incomprehensible. However, intonation marking enabled them to be understood by the interlocutor as questions and appropriate responses were given. CHI: xxx a play? (S18. 2;9.5) MOT: Are you going to play? Other questions

This category contains questions that cannot be classified into the above two categories. It includes tag questions, echo questions and abbreviated forms of repetition such as eh? and hmm? MOT: Oh we want those chickens. (S16. 2;8.3) MOT: Where’s those chickens? CHI: Eh? MOT: Look # here. MOT: They go in here don’t they? Feedback categories Feedback to the mother or child questions was calculated for only nine

103 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Early Trilingualism

104

scripts (Scripts 1, 4, 8, 12, 16, 20, 24, 28, 32) in order to have a more balanced sample, since to analyse all the feedback given by the mother to Jenny’s questions over 32 scripts would have changed the focus of the study. The feedback classification was somewhat similar to that of Olsen-Fulero and Conforti (1983) and as follows: None

Here, there are two types. First, no uptake to the question, possibly followed by the repetition or reformulation of a previous utterance by the questioner, an unrelated new utterance or the initiation of a new exchange. MOT: And then we’ll put birdy’s flex. (S24. 3;0.5) MOT: Where shall we put birdy’s flex? MOT: In the other pocket shall we? MOT: ##### There like that. Alternatively, there may be no uptake, followed by silence before either of the participants begins a new exchange. Acceptable responses

The most obvious is when the interlocutor provides a comprehensible and appropriate answer. CHI: Mummy I want the leg ... lego. (S32. 3;6.18) MOT: You want the lego? MOT: The lego or this thing? MOT: This? CHI: No the lego one. On other occasions, the child provides an utterance in response to the question which seems to be the answer but the meaning of which is only known to her (e.g. babble with appropriate intonation). This is also categorised as acceptable, for example: CHI: Er flower er er figwam. (S1. 1;11.24) MOT: What love? CHI: Er fefe figwam. Also considered ‘acceptable’ is an answer that is ‘incorrect’ e.g. blue, when the correct answer is red. For example: CHI: What’s this? (S31. 3;5.18) MOT: That’s a squirrel. CHI: Rabbit cat # this? MOT: That’s a hedgehog. CHI: Rabbit.

104 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Research Questions and Design

MOT: CHI: MOT:

105

And that? Rabbit. Hmm?

Unacceptable responses

This category includes the inappropriate answer. The answerer takes up the turn (uptake) but does not respond to the question or initiates a new exchange. CHI: Mummy. (S30. 3;4.7) CHI: Mummy. CHI: Where’s my thing for Leire? MOT: Would you like to give her xxx? CHI: That blue xxx. Also in this category are incomprehensible answers that are either phonetically unintelligible or are babbles that have an inappropriate intonation. CHI: What’s this? (S24. 3;0.5) CHI: Is this keys? MOT: xxx CHI: And this? Repetition

Repetition is classified but not codified, and is noted only when the question is identical to the previous utterance, not when it may have taken place in a previous turn in the same exchange. Reformulation

Instances of reformulation are not classified but, in the case of Jenny, have all been noted in the reports on each recording.

Mother’s Data Exactly the same data was collected for Jenny’s mother following the procedure and using the taxonomy described above. However, since the main focus of the present study is Jenny, only nine of her mother’s transcripts were analysed (Scripts 1, 4, 8, 12, 16, 20, 24, 28, and 32). The number of questions per functional category and the length of question was calculated and compared with the number of child questions produced in the same category. The type of question forms used in the functional categories was also worked out to see if this was reflected in the child’s use of particular forms in a functional category. The aim of this analysis was to see to what extent the child’s use of functions and forms in questioning reflected the input Jenny has received from her mother.

105 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Early Trilingualism

106

The same procedure was carried out for feedback, and the analysis that was done on Jenny’s scripts was also carried out on her mother’s scripts. This involved the calculation of the amounts of acceptable, unacceptable and no feedback that were received from Jenny in reply to the mother’s questions. Again, the aim was to see if Jenny’s use of feedback reflected the feedback she had received from her mother.

Cross-linguistic Influence Jenny’s general language was analysed to identify areas of cross-linguistic influence at the lexical level, and her questions were further analysed for cross-linguistic influence at the syntactic level. Lexical analysis All 32 scripts were examined for evidence of cross-linguistic influence in Jenny’s English at the lexical level since this is the area in which most work has been done in the field of bilingual and trilingual acquisition. It was hoped that the present study might benefit from methodological aspects of previous research in this area as well as making a contribution to this growing field. The purpose of this analysis is to ascertain the extent to which the child’s English has developed as a separate system. A sample of the mother’s discourse strategies in relation to language choice was also examined to see if Jenny was pragmatically sensitive to making repairs. Syntactic analysis of questions Jenny’s questions were examined for evidence of cross-linguistic influence at two levels, the general and the specific. The methodology was based on that used in previous studies of syntax in bilingual children, Schelletter’s (2000) study of negation in German and English, and Yip and Matthews’s (2000) work on interrogatives, null objects and relatives in Cantonese and English. These studies identified areas where adult syntactic forms in one language differ from those found in the other. When the differences have been isolated, evidence of their presence is then sought in child language development in the languages concerned. Their presence of such forms is seen as evidence that cross-linguistic influence is taking place. The examination of Jenny’s questions on a general level was based on descriptions of adult questions in Basque and Spanish and on developmental findings on the acquisition of questions in Basque-speaking and Spanishspeaking children. The following potential areas of cross-linguistic influence were identified: (A) Declarative questions. In both Basque and Spanish yes/no questions are

106 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Research Questions and Design

107

formed by placing rising intonation on the declarative form. This form also exists in English, but the most usual form of yes/no question in English involves inversion and auxiliary insertion. (B) Null subject. Both Basque and Spanish permit a null subject, and Basque permits a null lexical object.2 English does not permit a null subject. (C) Late wh-. In child Basque wh- is positioned late in the question prior to the development of fronting.3 Spanish has wh-fronting the same as English. So the appearance of wh- in non-fronted position may indicate cross-linguistic influence from Basque. (D) Subject on right. Both Spanish and Basque allow the subject to be on the right, whilst in English the subject must go on the left. During the analysis of the question data, Jenny’s questions were examined in general for evidence of cross-linguistic influence in areas A–D. It was expected that the data might yield other questions that deviated from the norm described for monolingual acquisition (taking into consideration the idiosyncratic nature of question acquisition), and these would be examined for cross-linguistic influence from Basque and Spanish. Such questions might equally exhibit characteristics of Basque, Spanish and English or be a hybrid result of the contact between the languages. These types were called E–K, and turned out to be: (E) (F) (G) (H) (I) (J) (K)

dependence on is + 0/it/this/that; dependence on wh- + it/this/that; subject repetition; word order; wh-confusion; hybrids; translation.

On a specific level, five samples of Jenny’s questions, at each of Brown’s (1973) five stages, were compared with samples from two monolingual children Sarah (Brown, 1973) and Sophie (Fletcher, 1985) with a similar MLU to see how question development differed. Sarah is a speaker of American English and Sophie is a speaker of British English, the variety spoken by Jenny’s mother. Of the three children, Adam, Eve and Sarah, studied by Brown (1973), Sarah was chosen because Eve did not correspond on age range (her samples finished at age 2;2). An examination of the questions of the subjects Adam and Sophie indicated that they produced more well-formed questions than Jenny did at a similar MLU. Sarah’s data showed that she clearly produced fewer and less complex questions than

107 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Early Trilingualism

108

Jenny at the same MLU. It was therefore decided that two monolingual subjects, one behind Jenny and the other ahead of Jenny in question production, would give a clearer picture of how Jenny was progressing at comparable MLUs, so Sarah and Sophie were chosen. A script from Jenny at each of Brown’s Stages I–V of development was examined for examples of non-standard question forms. The same was done for Sophie at Brown’s Stages III and V. The non-standard forms were analysed for evidence of areas A–K: and for other areas identified during the data analyses as anomalous to normal monolingual question development. If anomalies were found in the monolingual children’s data, as well as in Jenny’s data, they were considered developmental or idiosyncratic, whilst if they appeared in Jenny’s data they were considered to be evidence of cross-linguistic influence. Notes 1. According to the standard transcription system of the CHILDES database, # signifies a pause, ## signifies a longer pause (and so on), and xxx signifies unintelligible speech. 2. Null subject: in Basque and Spanish it is not necessary to specify the subject of a sentence with a pronoun. Null lexical object: in Basque it not necessary to specify the object pronoun of a sentence, as this is included in the verb form. 3. Fronting means that wh- words are positioned at the front (or beginning) of the question.

108 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Chapter 4

The Findings General Language Development In order to address the issue of how the child’s general acquisition of English takes place as compared to other children in monolingual or bilingual environments, MLU in words and TTR were calculated using the CHILDES Clan programme (MacWhinney, 1995). All Jenny’s utterances on all scripts were used to calculate the MLU, as it was considered important to analyse all the data and not to cut the scripts. TTR was calculated on 285 utterances (the length in child utterances of the shortest script in the corpus) in order to obtain the richest possible sample at a standard length over all the scripts. The results give a general indication of Jenny’s linguistic development and provide a benchmark with which to compare her with other monolingual and bilingual subjects. Although MLU is widely used as an index of development and comparison, and TTR is used to show vocabulary diversity, it should be noted that the reliability of these measures has been debated (Crystal, 1974; Richards, 1987). Likewise, different researchers may adapt the basic rules first laid down for calculating MLU by Brown (1973), along with his corresponding five developmental stages. With reference to the subjects mentioned below for comparison with Jenny, De Houwer (1990), for instance, calculates MLU on the basis of the first 80 utterances rather than Brown’s suggested 100 utterances starting from the 51st in the script, whilst Fletcher (1985) follows Brown’s rules with adaptations from Miller (1981). Table 4.1 shows that Jenny’s MLU in English clearly progresses from Brown’s (1973) Stage I in the early scripts to beyond Brown’s Stage V at age 3;1.9. In spite of some regression there is a general upward trend. The TTR also increases with age, which suggests that her vocabulary is growing. In Tables 4.2–4.5, Jenny’s progress is compared with that of other monolingual subjects, using samples from Sophie (Fletcher, 1985), Sarah (Brown, 1973) and Eve (Brown 1973) at ages as near as possible to Jenny’s samples. When data is not available for a particular child, this is marked NA. It must be emphasised that the aim of these comparisons is to give a general impression of Jenny’s progress in relation to that of other children since MLU and TTR, though widely accepted as useful tools for comparisons, may have been reached by slightly different calculations.

109 109 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:11

Color profile: Disabled Composite Default screen

Early Trilingualism

110

Table 4.1 Jenny’s scripts in terms of age, MLU, TTR and Brown’s stages (1973) Script

Age

MLU

TTTR

Brown’s stage

1

1;11.24

1.526

0.279

I

2

2;0.6

1.824

0.255

I

3

2;1.2

1.917

0.300

I

4

2;1.26

2.204

0.292

II

5

2;2.24

1.767

0.284

I

6

2;3.14

2.168

0.328

II

7

2;4.2

2.285

0.381

II

8

2;4.17

2.322

0.335

II

9

2;5.5

2.564

0.326

III

10

2;5.19

2.724

0.331

III

11

2;6.3

2.993

0.342

III

12

2;6.19

2.698

0.320

III

13

2;7.1

2.602

0.390

III

14

2;7.9

2.815

0.376

III

15

2;7.19

2.732

0.403

III

16

2;8.3

2.803

0.372

III

17

2;8.19

2.976

0.356

III

18

2;9.5

3.164

0.309

IV

19

2;9.22

2.835

0.362

III

20

2;10.2

2.672

0.361

III

21

2;10.14

3.562

0.327

V

22

2;10.30

2.695

0.392

III

23

2;11.17

3.069

0.319

IV

24

3;0.5

2.889

0.346

III

25

3;0.19

3.243

0.394

IV

26

3;0.30

3.332

0.371

IV

27

3;1.9

4.425

0.358

V+

28

3;1.22

3.045

0.399

IV

29

3;3.8

3.549

0.377

V

30

3;4.7

3.139

0.285

IV

31

3;5.18

2.985

0.423

III

32

3;6.18

2.985

0.427

III

110 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:12

Color profile: Disabled Composite Default screen

The Findings

111

Table 4.2 Jenny’s age, MLU and TTR for comparison with monolingual and bilingual subjects Age

MLU

TTR

1;11.24

1.526

0.279

2;4.17

2.332

0.335

2;10.14

3.562

0.327

3;0.5

2.889

0.346

3;1.9

4.425

0.358

3;3.8

3.549

0.377

3;4.7

3.139

0.285

3;5.18

2.985

0.423

At 2;4 Jenny and Sophie (Table 4.3) have similar MLUs at 2.33 and 2.53 respectively. However, Sophie then appears to progress more rapidly. Jenny’s vocabulary seems to as rich as Sophie’s, as demonstrated by her TTR. Table 4.3 MLU and TTR for English-speaking monolingual, Sophie Age

MLU

TTR

2;4.28

2.53

0.34

3;0.4

3.82

0.31

3;5.15

4.47

0.26

Source: Fletcher (1985)

Jenny and Sarah have a similar TTR at age 2;4 (Table 4.4). However Jenny’s TTR shows a greater increase than Sarah’s at subsequent ages. Sarah’s MLU is also considerably lower than Jenny’s at the same age in all the samples studied. Table 4.4 MLU and TTR for English-speaking monolingual, Sarah Age

MLU

TTR

2;4.26

1.721

0.327

2;11.30

1.628

0.375

3;5.13

2.378

0.273

Source: Brown (1973)

111 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:12

Color profile: Disabled Composite Default screen

Early Trilingualism

112

Table 4.5 MLU for English-speaking monolingual, Eve Age

MLU

1;11

3.141

2;3

2.627

Source: Brown (1973)

Eve is described as a ‘linguistically precocious child’ (MacWhinney, 1995: 294) and it can be seen (Table 4.5) that at 1;11 she has a much higher level of English than Jenny has. The last sample available from her at 2;3, though lower, confirms this impression. These comparisons suggest that Jenny’s development in English falls within a range similar to that of monolingual children. Tables 4.6 and 4.7 show MLUs and TTRs of children who are bilingual in English and another language. I am not aware of MLUs and TTRs available in English for trilingual children of a similar age range. The data for Kate (Table 4.6) suggest that her English is developing well in advance of Jenny and also of the monolingual Sophie at comparable ages, for example, at 3;0. Table 4.6 MLU for bilingual, Kate Age

MLU

2;10.5

4.04

3;2.28

5.11

3;1.13

4.86

Source: De Houwer (1990)

At around 3;3 Jenny has a higher MLU than Andreu (Table 4.7). By the time they are both 3;4 Andreu appears to have caught up with her. Again Jenny has a similar TTR in comparison with this child. Table 4.7 MLU and TTR for bilingual, Andreu Age

MLU

TTR

3;2.28

2.24

0.41

3;4.12

3.21

0.24

Source: Perez-Vidal (1995: 203, 257)

112 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:12

Color profile: Disabled Composite Default screen

The Findings

113

Table 4.8 Jenny’s MLU in Basque and Spanish Age

MLU: Basque

MLU: Spanish

2:02

NA

1.9

2;8

2.24

NA

3;2

3.88

2.9

3;8

4.48

2.64

Data on the MLU of Jenny in Basque and Spanish (Table 4.8) give us some idea of her proficiency in these languages and suggest that she is developing in a similar way to other bilingual children, such as Mikel and Jurgi (Ezeizabarrena, 1996). These MLU were calculated in morphemes as described by Ezeizabarrena (1996), and it should be pointed out that crosslinguistic comparisons of MLU can be misleading, especially in languages as different as Basque and Spanish. The limited data available suggest that Jenny is stronger in Basque than in Spanish (Table 4.8). The decline in Spanish at 3;8 coincides with her starting kindergarten in Basque full-time at 3;5, as a result of which she began to spend much less time with her monolingual Spanish childminder. Comparing Tables 4.8 and 4.9, we can see that at age 2;2 Jenny has a MLU in Spanish that is lower than Mikel but higher than that of Jurgi. If we compare Jenny’s Basque at 2;8 with Mikel’s, we find that it is considerably lower, but it is only slightly lower than that of Jurgi. At 3;2 Jenny is higher in both Basque and Spanish than Jurgi yet lower in both languages than Mikel. She also has a lower level than Mikel at age 3;8 in both languages. Overall she seems to have a level in Basque and Spanish that falls somewhere between that of the two boys. Table 4.9 MLU of Spanish/Basque bilinguals Mikel

Jurgi

Age

MLU: Basque

MLU: Spanish

MLU: Basque

MLU: Spanish

2;2

2.41

2.93

1.07

1.14

2;8

3.41

3.28

2.34

1.84

3;2

4.53

5.58

3.44

2.3

3:08

5.09

5.28

NA

NA

Source: Ezeizabarrena (1996)

113 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:12

Color profile: Disabled Composite Default screen

Early Trilingualism

114

Formal Development of Interrogative Behaviour The length, number and type of questions produced by Jenny were examined in order to find out how the development of the child’s question form took place in terms of length of question, number of questions and the type of question produced. An overall picture of Jenny’s development in question form in numbers and total percentages over time can be seen in Table 4.10, which clearly shows that the area of most growth is that of wh-questions, which form one area of our study. The development of yes/no, intonation and declarative questions are closely related to each other and form a separate group. Figure 4.1 shows the percentage made up by the different question forms over the study as a whole. The first stage of question formation without wh- is the intonation question, which is an utterance containing one or more child-invented forms that are marked by intonation. CHI: Fu foh Mummy? (S3. 2;1.2) MOT: Hmm? CHI: Fu foh Mummy? MOT: Are you going to go and put it on? As the data show (Table 4.10) intonation questions decrease as the child’s ability to produce more adult-like words and inverted question forms increases, and they will not be further discussed here. Declarative questions consisting of the declarative form of the utterance with rising intonation are structurally indistinguishable from non-inverted yes/no questions (Fletcher, 1985: 105). So it is not surprising that, as shown in Yes/no 5%

Any other 12%

Declarative 37%

Intonation 2%

Figure 4.1 Total percentage of each question form

114 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:13

Wh-questions 44%

Color profile: Disabled Composite Default screen

The Findings

115

Table 4.10 Development of question form in numbers Tape

Age

Yes/No

Wh

Intonation Declarative

Other

Total

1

1;11.24

0

3

1

2

2;0.6

0

3

4

10

1

15

38

3

3

2;1.2

0

1

48

3

11

7

4

2;1.26

0

22

3

1

12

8

5

2;2.24

24

0

0

0

5

0

5

6 7

2;3.14

3

26

1

21

21

72

2;4.2

0

7

0

15

6

28

8

2;4.17

0

3

1

19

9

32

9

2;5.5

0

13

2

8

15

38

10

2;5.19

0

12

1

19

4

36

11

2;6.3

1

20

0

10

0

31

12

2;6.19

3

27

4

16

6

56

13

2;7.1

1

19

1

27

0

48

14

2;7.9

0

24

1

9

3

37

15

2;7.19

2

13

0

7

8

30

16

2;8.3

5

24

2

18

3

52

17

2;8.19

0

26

0

16

6

48

18

2;9.5

0

11

1

21

0

33

19

2;9.22

1

17

0

7

3

28

20

2;10.2

1

15

0

13

1

30

21

2;10.14

0

15

0

9

1

25

22

2;10.30

3

18

1

18

2

42

23

2;11.17

1

12

0

9

0

22

24

3;0.5

4

14

0

10

3

31

25

3;0.19

2

25

0

11

3

41

26

3;0.30

1

24

0

4

10

39

27

3;1.9

2

15

0

16

1

34

28

3;1.22

5

22

2

13

11

53

29

3;3.8

1

17

1

16

4

39

30

3;4.7

16

64

0

33

4

117

31

3;5.18

7

29

0

7

2

45

32

3;6.18

6

23

0

2

0

31 1232

Total

62

545

27

453

146

Total %

5%

44.2%

2.2%

36.7%

11.8%

115 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:13

Color profile: Disabled Composite Default screen

Early Trilingualism

116

Table 4.10, they decrease as the child becomes more proficient at producing the inversion that defines a question as a yes/no type. Nevertheless even though they decrease, declarative questions remain prevalent throughout the period of sampling, an effect that is not seen in the monolingual subjects we have studied. With Jenny, yes/no questions appear at age 2;3.14 and, although they are not as frequent as wh-questions and declarative questions, their development increases steadily. Other questions, many of which are based on the declarative form, are present throughout the period under study. Length of questions In order to find out if Jenny’s questions increased in length between the ages of 1;11.23 and 3;06.18 the total number of words produced in questions in a recording was divided by the number of questions, to give the average length of question in words per recording (Figure 4.2). Figure 4.2 indicates that there is growth in the number of words included in a question, reaching a peak at age 2;11.7 (Script 23) with an average of around 3.5 words per question. After this, growth in question length begins to drop off. Growth does not seem to be steady as the amount of words per question found can vary considerably from one script to another.

3.5

Average number of words

3.0 2.5 2.0 1.5 1.0 0.5 0 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 Scripts

Figure 4.2 Average length of question in words per recording

116 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:17

Color profile: Disabled Composite Default screen

The Findings

117

Number of questions The percentage of questions to utterances (in other words, what percentage of Jenny’s utterances were questions) was calculated to see if questioning behaviour increased between the ages of 1;11.23 and 3;06.18. The number of questions per recording and the number of utterances per recording are shown in Table 4.11 along with the percentage of questions to utterances. The data indicate that there is a small increase in the percentage of questions to utterances over the period of the study. In the first recordings the percentage is relatively low but then remains steady at around 10–15%. Script 30 (3;4.7) is an exception to this. In this script Jenny produces 399 utterances including questions, making it one of her most talkative scripts. The topic, a sack of old baby toys, seems to have been of particular interest for her as 117 (41%) of the utterances are questions such as: • • • • • •

What’s this? Who’s bought this? What’s this for doing? Where’s the other rabbit of this? What’s this for? Mummy what’s this boy on the horsey?

The number of questions to utterances is usually somewhere between 10% and 20%, but there is a slight overall increase over time (Figure 4.3).

45% 40% 35% 30% 25% 20% 15% 10% 5% 0%

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 Scripts

Figure 4.3 Percentage growth in number of questions per recording

117 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:21

Color profile: Disabled Composite Default screen

Early Trilingualism

118

Table 4.11 Percentage of questions to utterances % Questions/ Utterances

Tape

Age

Questions

Utterances

1

1;11.24

15

428

3.50%

2

2;0.6

48

294

16.33%

3

2;1.2

22

208

10.58%

4

2;1.26

24

273

8.79%

5

2;2.24

5

241

2.07%

6

2;3.14

72

344

20.93%

7

2;4.2

28

256

10.94%

8

2;4.17

32

204

15.69%

9

2;5.5

38

402

9.45%

10

2;5.19

36

353

10.20%

11

2;6.3

31

366

8.47%

12

2;6.19

56

430

13.02%

13

2;7.1

48

251

19.12%

14

2;7.9

37

341

10.85%

15

2;7.19

30

294

10.20%

16

2;8.3

52

379

13.72%

17

2;8.19

48

332

14.46%

18

2;9.5

33

253

13.04%

19

2;9.22

28

253

11.07%

20

2;10.2

30

273

10.99%

21

2;10.14

25

324

7.72%

22

2;10.30

42

268

15.67%

23

2;11.17

22

226

9.73%

24

3;0.5

31

272

11.40%

25

3;0.19

41

244

16.80%

26

3;0.30

39

274

14.23%

27

3;1.9

34

298

11.41%

28

3;1.22

53

349

15.19%

29

3;3.8

39

302

12.91%

30

3;4.7

117

282

41.49%

31

3;5.18

45

308

14.61%

32

3;6.18

Total

31

206

15.05%

1.232

9.528

12.93%

118 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:21

Color profile: Disabled Composite Default screen

The Findings

119

Yes/no Questions Declarative questions Any discussion of the development of yes/no questions must first deal with declarative questions, since non-inverted declarative questions may be considered as potential yes/no questions (Fletcher 1985: 105). Declarative questions are formed by placing rising intonation on an affirmative utterance; no further syntactic modification is required. Syntactic modification in the form of inversion and auxiliary insertion converts them into yes/no questions. Declarative questions make up 36.7% of the overall corpus and are present in all of Jenny’s scripts, sometimes in large numbers. There are 38 declarative questions at 2;0.6 (S2) whilst at 3;4.7 (S30), 16 months later, she still produces 33. In other scripts at similar ages she produces as few as 5 (S5. 2;2.24) or 7 (S31.3;5.18) declarative questions. So the presence of declarative questions is constant, although their actual numbers seem to be variable throughout the period under study. However if we examine the proportion of questions that are declarative in relation to numbers of whquestions and yes/no questions, we see a clear decrease over time in declarative questions (Figure 4.4) and a clear increase in the other two types (Figures 4.5 and 4.6).

100%

80%

60%

40%

20%

0%

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 Scripts

Figure 4.4 Decrease in percentage declarative questions per recording

119 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:25

Color profile: Disabled Composite Default screen

Early Trilingualism

120

We will now look at some examples of declarative questions from across the corpus. Here the mother is offering the child a handkerchief: MOT: Take a hankie # go on. (S1. 1;11.24) MOT: Hmm? CHI: This one? MOT: Okay. In the next example, Jenny uses a declarative to ask where her father (Aitor) and her brother (Jon Ander) have gone. CHI: Mummy Aitor? (S4. 2;1.26) CHI: Mummy Aitor a Jon Ander? MOT: Oh they’ve gone out. These examples show that, from an early age, Jenny is able to use declarative question effectively as a tool for requesting information.

Yes/no questions The percentage presence of yes/no questions in scripts increases steadily over time as is shown in Figure 4.5. However, only 62 questions showing the yes/no question characteristics of inversion and auxiliary insertion were identified in the data. We will now look at some of them.

20%

15%

10%

5%

0%

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

-5%

Scripts

Figure 4.5 Increase in percentage yes/no questions per recording

120 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:30

Color profile: Disabled Composite Default screen

The Findings

121

Early inversion: Can

Up until age 2;3.14 (S6), Jenny’s non-wh questions consist of intonation, declarative and any other questions. The first three questions classified as yes/ no questions because of the presence of inversion are all can-initial (in other words, they begin with can). In the following example Jenny requests her mother’s participation with reading a book. CHI: Can we do it this? (S6. 2;3.14) CHI: This? MOT: What do you want, love? CHI: This. There are 19 examples of can-initial questions in the data. At 2;8.3 Jenny enlists her mother’s help in taking off her shoes: CHI: I want to take it woosh. (gloss: shoes) (S16. 2;8.3) CHI: Can you take? MOT: (ignores child) CHI: Mummy Mummy I taking shoes. MOT: Okay take your shoes off then. The following example, taken from the period around age three and coinciding with the earliest evidence of Jenny’s do and does insertion, suggests that can-initial may have acted as a kind of rudimentary aux-insertion. CHI: Mummy runner beans is this. (S24. 3;0.5) CHI: Can I you like runner beans? MOT: I love runner beans. Can does not appear again until the final three scripts 3;4.7 (S30) to 3;6.18. (S32) in which there are many cases, nearly all well-formed. The questions are longer, and there are many other questions and question forms. • Mummy can you give me that teddy? • Can I put this in here? • Can you take this all out?

(S30. 3;4.7) (S30. 3;4.7) (S32. 3;6.18)

Early use of can-initial yes/no questions may be explained as having been learnt as a routine. In the input the child hears can placed in initial position and followed by rising intonation, thus indicating an interrogative. It has no inflections and thus may provide an easy model of question formation for the child to use. Early inversion: Is (are)

Apart from are you stupid? produced by Jenny at 2;6.3 (S11) and which, due to its well-formedness and isolation, is assumed to be a formula rather

121 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:07

Color profile: Disabled Composite Default screen

Early Trilingualism

122

than an example of inversion, there are no further examples of yes/no questions until 2;6.19 (S12). From then onwards Jenny builds on is in a variety of combinations that emerge over time to form the backbone of her yes/no questioning. At 2;6.19 three combinations appear: is + this, is + 0 and is + it. • is + this In this example, Jenny and her mother are sorting out a drawer: CHI: Is this pijamoa? [pyjamas] (S12. 2;6.19) MOT: No that’s a tee-shirt. CHI: Tee-shirt this one? • is + 0 Later Jenny is having difficulty putting on a tee shirt, and asks if she has done it right: CHI: Is no number (S12. 2;6.19) CHI: I can’t do it no. CHI: Mummy I can’t. CHI: Oh dear oh dear. CHI: Is only on mine? (gloss: Is it on me right?) MOT: No you’ve got it the wrong way round now. • is + it In the final example, from 2;6.19 (S12), mother and daughter are playing at taking photos of each other. It of interest not only for the is + it yes/no question, but also for the various other developmental forms produced by Jenny at this stage which we cannot go into here: MOT: Take me a photo. (S12. 2;6.19) CHI: Here you. CHI: Agained. CHI: Is it doed it? (gloss: Has the photo been taken?) CHI: I can’t do(dled) that. CHI: I do doed it. (gloss: I took the photo) At age 2;7.1 (S13) the only yes/no question produced is of the is + it type, but the inversion is placed at the end of the question. Mother and child are naming Jenny’s dolls: MOT: Who’s that? (S13. 2;7.1) CHI: Humpty Dumpty is it? MOT: No it’s not Humpty Dumpty. MOT: What’s that dolly called?

122 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:30

Color profile: Disabled Composite Default screen

The Findings

123

The next and only yes/no question to appear at 2;7;19 (S15) is also of the is + it type, but in this case it is placed in the medial position. Jenny tries to explain that her leg hurts when jumping. CHI: Hurt. (S15. 2;7.19) CHI: xxx is it jump? MOT: Jenny jump? MOT: It hurts does it? Further is + it questions appear later in the data, and show the non-initial position for the is + it inversion. In the following example, mother and child are looking for a lost apron: MOT: Oh I can’t here I can’t see it. (S22. 2;10.30) CHI: Up there is it? MOT: Is it? MOT: I can’t see it up here either. At 2;8.3 two more examples of the type is + 0 appear. In the first one, Jenny wants to know if one of the characters in a book is Noddy’s mother. CHI: Is Noddy Mummy? (S16. 2;8.3) MOT: Is that Noddy? (Clarification) CHI: No Noddy Mummy. MOT: Oh that’s Noddy’s Mummy. MOT: So it is. In the next example, Jenny produces a mix of a is + 0, yes/no and wh-question, which may be an attempt at a subordinate clause. MOT: Now look shall I show you something in this box? CHI: Is a open the box who is it? (gloss: If we open the box will we see who it is?) MOT: (opens box) CHI: Ooh it’s a something. Is + this continues to be used frequently during the rest of the period under study, especially as is + this for at the following ages: • • • •

Is this is green? Is this alright? Is this the watcha? Is this for doing in the floor?

(S23. 2;11.17) (S24. 3;0–5) (S25. 3;0.19) (S30. 3;4.7)

123 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:30

Color profile: Disabled Composite Default screen

124

Early Trilingualism

• is + he A new combination consisting of is + he appears at age 2;9.22 (S19). In the following example mother and child are looking for hairclips inside a teddy pyjama case: CHI: Is he got the teddy with the hole in his back? (S19. 2;9.22) (gloss: has the pyjama case teddy got it?) CHI: Look is got it, is got it look. In this example the child uses the subject pronoun he in the correct position for inversion but then also includes the subject the teddy. No more examples of is + he appear in the data. The last is combination to appear is is + that. • is + that In the next example, Jenny is pretending to give a tea party to her school friends: CHI: Is that my friend is it yet? (S31. 3;5.18) MOT: Whose friend’s that? CHI: Sinead. (her friend) Finally at age 3;4.7 the first evidence of the productive use of are in a yes/ no question appears, though not in a well-formed question. Jenny wants her mother to show her the old toy bag: CHI: Are you do that? (S30. 3;4.7) MOT: You want to do that, do you? The examples show that Jenny progresses from using is + 0/it/this, which emerges at around 2;6.19, to gradually incorporating is + he at age 2;9.22, and is + that at age 3;5.18. Finally, towards the end of this period at 3;4.7, she begins to use are productively. Early inversion: Will, shall, have

Apart from the use of the modal auxiliary can and the auxiliary is which we have just outlined, there is also limited evidence of other auxiliaries. The first of these to appear is will, which is found at age 2;8.22. Here Jenny is asking her mother to dry her back at bath-time: CHI: Oh I got it wet. (S16. 2;8.22) CHI: Will you clean, Mum? CHI: Got it on the back. There is only one example in the study of the use of shall in a yes/no ques-

124 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:07

Color profile: Disabled Composite Default screen

The Findings

125

tion. This is in the last recording at age 3;6.18 (S32) and is not in a wellformed question. It occurs when Jenny is asking her mother to lift her up. CHI: Mummy shall I you, you put me up? (S32. 3;6.18) MOT: You want to go up there? CHI: Yes. There is also only one example to be found of have in a yes/no question. Jenny is playing with a tea-set and talking to herself. CHI: Where’s the top? (S31. 3;5.18) CHI: Top of from this. CHI: Have you lost it? CHI: Yes. So the first evidence of the development of the use of other auxiliaries such as will, shall and have is found slightly later than the appearance of the use of is to front questions, that is to say from 2;8.22 onwards. Nevertheless the period prior to age three and after seems to be important in Jenny’s development for the emergence of the use of auxiliaries in questions, as at 2;10;30 does also appears. The emergence of auxiliary insertion: Do, does, don’t

There are only six cases in the data where a form of the auxiliary do is used with inversion and they appear in the following order: does, do, don’t. The only example of does appears at age 2;10.30 in the following exchange, where Jenny and her mother are fitting stacking beakers together: CHI: This is? (S 22. 2;10.30) MOT: That’s the top. CHI: Top. CHI: What’s that there? CHI: Does the this in here? (gloss: does this go in here?) MOT: Ahah. (Affirmation) It could be argued that this is not a case of do-insertion owing to the absence of a main verb, and that does is functioning as a lexical verb that has been placed in initial position as a type of inversion. The child’s first utterance is also a question (this is?) but it lacks inversion. This invites us to ask to what extent the child is able to correctly apply inversion at this stage, and to which types of auxiliaries or verbs. There are four examples of use of do in yes/no questions. In the first case Jenny asks a well-formed question about her mother’s likes. It comes slighter later in the same script (S24) mentioned above, when Jenny attempts the question can I you like runner beans? This time she is able to ask properly:

125 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:30

Color profile: Disabled Composite Default screen

Early Trilingualism

126

CHI: MOT:

D’you like runner beans? Yum yes I do.

(S24. 3:0.5)

These examples are evidence of the child not merely learning a formula do you like but of the strategies she applies when attempting to use it. The following example, as Jenny and her mother are playing with a set of farmyard animals. CHI: Do you want to help me with other ones? (S28. 3;1.22) MOT: Let’s put the animals on the farm then. Here the child’s polite request may be the result of learning a formula heard from the mother. The next example is very similar, and could also be a formula. The child is playing with her tea-set: CHI: Now do you want anon(ther) cup of tea? (S29. 3;3.8) MOT: Yes please. The last example with do comes at age 3;6.18 (S32). Jenny and her mother are identifying characters on a football: MOT: What else has it got on it? (S32.3;6.18) CHI: Duck. MOT: Duck. CHI: Do you know him? MOT: Minnie Mouse. As she points at the character Jenny manages to ask an adult-like yes/no question with a lexical verb. The next example shows a fully-formed yes/no question with the auxiliary have, with a developmental error on the form of the past participle. Jenny wants to know if her mother will be cooking for her baby cousin: CHI: Have you make purée for Leire? (S27. 3;1.9) MOT: Purée for Leire? CHI: For eat. There is only one example where the child uses a negative yes/no question. In this situation she is misbehaving and defying her mother to touch something: CHI: Mum I’m touching this Mum! (S27. 3;1.9) CHI: I touching. MOT: Okay there’s a good girl. (wearily) CHI: Mum what(’re) you doing with that thing? CHI: Eee wewewe wewewe! CHI: Don’t you like it?

126 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:30

Color profile: Disabled Composite Default screen

The Findings

127

Given the generally low level of Jenny’s do auxiliary development, it is quite surprising to find this well-formed question here. The context in which it is used to annoy her mother suggests that it is an example of accurate functional use of a formulaic expression and not evidence of emergence of initial don’t. Summary of yes/no questions So far we have examined yes/no questions according to the types that are most frequently found and the age at which each type appears. However if we look at Jenny’s yes/no question development as a whole other patterns emerge. In spite of acquisition errors not specifically related to questioning, a good number of well-formed yes/no questions appear from around 2;8.3 (S16) onwards although correct do insertion is limited to five cases. Jenny’s progress in yes/no question development will now be related to that described by Klima and Bellugi (1966) and Cazden (1970) whereby the child passes through periods: Bellugi’s Period A

This period corresponds to a MLU of 2.0–2.5, when Jenny was aged 1;11.24–2;4.17 (scripts 1–8). In this period ‘questions are expressed by intonation only’ (declarative questions) (Cazden, 1970). We find this in these examples from Jenny’s data, though some of her questions are longer than the two-word examples given by Cazden. • • • • •

Okay here? Another one? Kang(ar)oo a baby? It here # on the floor? Put in here?

(S1. 1;11.24) (S2. 2;0.6) (S6. 2;3.13) (S7. 2;4.2) (S8. 2;4.17)

Bellugi’s Period B

This period corresponds to a MLU of 2.5–3.5, when Jenny was aged 2;5.5–3;0.5 (scripts 9–24). According to Cazden (1970), in this period there is ‘no development of question forms except the appearance as routines of don’t and can’t.’ However, in Jenny’s case there are already examples with can used productively at 2;3.14 (S6), and a number of other instances of can appear in this period. There are single uses of do, and does and two uses of will. This period is also rich in examples of inversion with the copula is + 0, is + it, and is + this: • Can I get it xxx? • Does the this in here? • Will you clean Mum?

(S16. 2;8.3) (S22. 2;10.30) (S16. 2;8.3)

127 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:30

Color profile: Disabled Composite Default screen

Early Trilingualism

128

• Is putting in here? • Is it doed it?

(S20. 2;10.2) (S12. 2;6.19)

Bellugi’s Period C

This period corresponds to a MLU of 3.5–4.5, when Jenny was aged 3;0.19–3;6.18 (scripts 25–32). Cazden (1970) reports that this period shows ‘development of auxiliary verbs in entire grammatical system. Inversion of AUX and subject NP in yes/no questions.’ In Jenny’s data for this period we find examples of the use of shall, do, don’t and is + that as well as features relating to can and the copula present from the previous stage: • • • • •

Shall we do something? Do you want to help me with the other ones? Don’t you like it? Is that my friend is it yet? Mummy can you give me that teddy?

(S28. 3;1.22) (S28. 3;1.22) (S27. 3;1.9) (S31. 3;5.18) (S30. 3;4.7)

We find no evidence in Jenny’s data of did or to be+-ing as described in the lists of yes/no questions in Klima and Bellugi (1966) for these periods. On the other hand, their data does not include instances of is +0/it/this/that like those reported for Jenny. If we look at the development described by Fletcher (1985) we also see some similarities and differences. Can-initial questions, for instance, form a large proportion of the many inverted yes/no questions produced by Sophie from age 2;4.8 onwards. Some of the examples cited by Fletcher (1985: 107) for Sophie are quite similar to those produced by Jenny: • Can me cut in this side? • Can you put that in my hair? However, Fletcher reports that other yes/no questions to appear include have/has. In our data on yes/no questions we have just one example with have/ has (though some are present in wh-questions). Fletcher makes the point that that these forms are ‘successors to is-initial questions in earlier samples’. It may be, therefore, that Jenny’s use of is + 0/it/this/that will develop into have forms.

Wh-questions Time and order of emergence Wh-questions are without doubt the aspect of interrogative behaviour that undergoes greatest growth and change in the period under study. They also form the largest section of the corpus, when examined from the

128 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:30

Color profile: Disabled Composite Default screen

The Findings

129

perspective of form, with 545 examples. In order to understand their development they must be examined for (1) time and order of emergence of whforms, (2) inversion and auxiliary insertion. After a general summary of Jenny’s progress, we will describe and comment on examples from the data that illustrate key moments in her wh-acquisition in relation to (1) and (2). Labov and Labov (1978) and Fletcher (1985) report sizeable differences between the development of the different wh-words. With this in mind we decided to trace the development of inversion and auxiliary insertion as it appears in each wh-form. Table 4.12 shows the actual numbers of wh-questions to be found in each script. Except for at age 2;2.24 (S5) they are present throughout the sample, initially in small numbers but then stabilising at between 10 and 20 per recording. Figure 4.6 shows the percentage of the total questions in each recording that are wh-questions. It can be seen that this increases steadily over time. Compare this with Figure 4.7, which shows the overall decrease in the percentage of all other (non-wh) questions found in the recordings. Figure 4.8 compares the progress of the two classes of questions that show most growth, wh-questions and yes/no questions. It is of interest to

80% 70% 60% 50% 40% 30% 20% 10% 0%

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

Scripts

Figure 4.6 Percentage of wh-questions per recording

129 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:35

Color profile: Disabled Composite Default screen

Early Trilingualism

130

Table 4.12 The development of wh-questions Script

Age

Number of wh-questions

1

1;11.24

3

2

2;0.6

3

3

2;1.2

1

4

2;1.26

3

5

2;2.24

0

6

2;3.14

26

7

2;4.2

7

8

2;4.17

3

9

2;5.5

13

10

2;5.19

12

11

2;6.3

20

12

2;6.19

27

13

2;7.1

19

14

2;7.9

24

15

2;7.19

13

16

2;8.3

24

17

2;8.19

26

18

2;9.5

11

19

2;9.22

17

20

2;10.2

15

21

2;10.14

15

22

2;10.30

18

23

2;11.17

12

24

3;0.5

14

25

3;0.19

25

26

3;0.30

24

27

3;1.9

15

28

3;1.22

22

29

3;3.8

17

30

3;4.7

69

31

3;5.18

24

32

3;6.18

23

130 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:35

Color profile: Disabled Composite Default screen

The Findings

131

120%

100%

80%

60%

40%

20%

0%

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

Scripts

Figure 4.7 Percentage of non-wh-questions per recording 70 Wh-questions

Yes/no questions

60

Number of questions

50 40 30 20 10 0

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 Scripts

Figure 4.8 The growth of wh-questions and yes/no questions

131 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:44

Color profile: Disabled Composite Default screen

Early Trilingualism

132

note that, although fewer yes/no questions are produced, their pattern of growth closely follows that of wh-questions. Table 4.13 shows the time of emergence of the different wh-forms and how many of each wh-form were produced in each recording. The order of emergence for Jenny’s wh-forms is what, where, who, how, why and when. This reflects, by and large, the order reported in other studies (Smith, 1933; Tyack & Ingram, 1977; James & Seebach, 1982) except for who, which in Jenny’s case appears earlier and is used more frequently than reported elsewhere. The earliest examples of each form are as follows: what wha da? (S1. 1;11.24) where where you go a? (S4. 2;1.26) who who? (S6. 2;3.14) how how are you? (S10. 2;5.19) why why why? (S11. 2;6.3) when when it comes in your book? (S15. 2;7.9) The emergence of wh-forms: Inversion and auxiliary insertion Examination of the wh-question data revealed that development was different from that found in yes/no questions. Once wh-fronting had taken place, the question forms used by the child in the rest of the question bore little relation to the yes/no questions outlined above, and so the classification that follows is different from the one previously described. As there are developmental differences in the different types of wh- question (Fletcher, 1985; Labov & Labov 1978), we examined each wh-type separately. All instances of how, why and when have been described in full. Within the larger groups, what, where, and who, we selected the questions that most closely matched fully-formed questions in terms of inversion and auxiliary insertion in order to see when and in which environments such characteristics appeared. We have also described a number of non-fully formed questions we consider to be of interest in tracing development in each wh-type. At the end of the analysis the findings will be related to Klima and Bellugi (1966) and Cazden (1970), as we did for yes/no questions. Fully-formed what questions Fully-formed what questions should show the form of wh + aux + NP (noun phrase) + VP (verb phrase) and be similar to forms used by adults. Most of Jenny’s early what questions are of the type whaddat? what that? reported in the literature (Brown, 1968; Cazden; 1970; Holzman, 1972). At age 2;0.6 (S2) Jenny correctly produces the formula what that for? Apart from these formulaic what questions the first almost fully-formed what

132 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:44

Color profile: Disabled Composite Default screen

The Findings

133

Table 4.13 The emergence of Jenny’s wh-forms Script

Age

MLU

What

1

1;11.24

1.526

3

3

2

2;0.6

1.824

3

3

3

2;1.2

1.917

1

4

2;1.26

2.204

2

1

5

2;2.24

1.767

6

2;3.14

2.168

21

1

7

2;4.2

2.285

7

8

2;4.17

2.322

3

3

6

2;5.5

2.564

13

13

10

2;5.19

2.724

6

4

11

2;6.3

2.993

11

9

12

2;6.19

2.698

17

5

4

13

2;7.1

2.602

2

13

4

14

2;7.9

2.815

18

4

1

15

2;7.19

2.732

7

4

1

16

2;8.3

2.803

16

3

5

24

17

2;8.19

2.976

16

1

9

26

18

2;9.5

3.164

7

3

1

11

19

2;9.22

2.835

15

1

1

17

20

2;10.2

2.672

13

1

1

21

2;10.14

3.562

6

3

5

22

2;10.30

2.695

11

5

2

18

23

2;11.17

3.069

8

3

1

12

24

3;0.5

2.889

8

5

1

25

3;0.19

3.243

24

26

3;0.30

3.332

14

7

1

1

23

27

3;1.9

4.425

10

2

2

1

15

28

3;1.22

3.045

16

1

4

1

22

29

3;3.8

3.549

11

3

3

30

3;4.7

3.139

51

5

6

2

64

31

3;5.18

2.985

14

4

5

32

3;6.18

2.957

17

1

1

371

89

63

Total

Where

Who

How

Why

Total

1 3 4

26 7

1

1

12 2

22

1

27 19

1 1

24 13

15 1

15

14 1

133 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:44

When

25

17 1 2

1 8

3

28

1

20

9

542

Color profile: Disabled Composite Default screen

Early Trilingualism

134

question appears at 2;6.3 (S11) although the auxiliary is has been omitted. In it, Jenny wants her mother to look at a teddy bear: CHI: Mummy look. (S11. 2;6.3) CHI: Teddy hat. MOT: Teddy hat so it is. CHI: Hey what the teddy doing? The next fully-formed question appears as part of a reformulation sequence in which Jenny is attracting her mother’s attention. She improves on the initial question to produce a fully-formed question with inverted is. CHI: What on here? (S11. 2;6.3) CHI: What’s it on here? CHI: Hey what’s it on here? MOT: What’s it on here? The next example is not well formed as a complete utterance, but the question form is correct. The ’s in what’s may be the appearance of has as an auxiliary. So far Jenny has produced questions starting with what’s and it has been assumed that ’s stands for is. The meaning of the question is something like ‘who’s got a hat like this?’ CHI: What this? (S16. 2;8.3) MOT: That’s Mr Plod the policeman, isn’t it? CHI: What’s he got a like hat? MOT: Got a funny hat, hasn’t he? At 2;10.14 Jenny produces two well-formed questions as a result of her mother’s prompt. The second question in slightly different from the first. MOT: What’s that called do you remember? (S21. 2;10.14) CHI: What’s that called? CHI: What’s it called? MOT: It’s called a padlock isn’t it? So far all the well-formed questions with what have included ’s, either as is or has. At 2;10.30 we find the first example in the data of a what question without ’s and with can. The following exchange is taken from a jigsaw puzzle extract. Jenny is reflecting to herself on where to put the pieces. CHI: Now. (looking at pieces) (S22. 2;10.30) CHI: What can I do? CHI: Who it who are you?

134 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:44

Color profile: Disabled Composite Default screen

The Findings

135

The next example is the first well-formed what question to contain are and you. Jenny is provoking her mother: CHI: Mum what (re) you doing with that thing? (S27. 3;1.9) CHI: Eee wewewe wewewe. CHI: Don’t you like it? At around this age, Jenny asks simple well-formed questions based on what + (’s) + it/this/that + doing as well as non-fully-formed questions showing a variety of characteristics that we will examine in the next section. However, we now see the first evidence of a more mature what question form emerging with correct inversion and auxiliary insertion The pronoun we is also used appropriately for the first time in the next example. Jenny and her mother are playing with the tea set and Jenny has an idea: CHI: Ah what’ll we do? (S29. 3;3.8) CHI: Hah! (has an idea) CHI: Thing some milky on it. A final example of a well-formed what question comes when Jenny sees a photo that she has not looked at yet. Apart from the question form, the correct deixis1 is of interest: CHI: What’s that on there? (S31. 3;5.18) MOT: On where? CHI: Under the hat that. In summary we can say that the child’s early well-formed what questions are generally based on what(’s) + it/this/that + (for) ... ing. From age 2;8.3 (S16) onwards other forms such as ’s got, and the auxiliaries can, will and do start to appear correctly placed with inversion in occasional examples. Non-fully-formed what questions Early forms

The earliest what question forms to appear, apart from the formulaic whaddat, are with doing as in the next example. Note that the subject is in the final position, unlike monolingual examples at this stage, where it is usually placed in a non-inverted position after what, e.g. what + NP + doing? CHI: What doing Jenny Mummy? (S6. 2;3.14) CHI: What doing Jenny? MOT: Jumping isn’t she? After this initial stage of dependence on what’s (what + ’s), at around 2;6.19 (S12) other auxiliaries and verbs start to appear with what(’s) and in

135 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:44

Color profile: Disabled Composite Default screen

Early Trilingualism

136

non + ing forms. Some contain a wide variety of developmental errors and possible cross-linguistic influence. Nevertheless certain patterns emerge that are of interest, and these will now be described. What(’s) + it/this/that?

The what’s + it/this /that phrase fronts the question and seems to act in place of auxiliary insertion in marking a question form as will be seen in the next example. Here Jenny and her mother are ‘treating’ an ‘injured’ doll: CHI: I’ve got it dolly (S14. 2;7.9) CHI: Mummy what’s that get dolly there? MOT: What’s she got? CHI: It’s got it bowen water. (broken) In the next two examples, the auxiliary (either is or has) is included in the what’s + it/this/that phrase. In the following example mother and child are dressing a toy bear: MOT: Oh blue knickers (S16. 2;8.3) CHI: What’s it get in this here? MOT: Where? CHI: This is hard. MOT: That’s right. In the next example, Jenny and her mother are playing with a toy containing coloured balls. In the question there is double auxiliary, ’s and is. It is possible that one of these may be a shortened form of has, the auxiliary required by done. MOT: What’s this? (S.17.2;8.19) MOT: What’s what? CHI: What’s this is done it? MOT: It’s a little ball isn’t it? As her what questions develop, we find that Jenny is still very dependent on what, what’s, what’s this and what’s that although the rest of the question may be more complex (with get, for instance), and with inversion. The presence of it, this, or that also provides a convenient filler for the subject position. It is possible that until Jenny learns to separate the formula what’s + it/ this /that she will not be able to correctly insert the auxiliary. Word order and subject repetition

In the next example the subject this has not been inverted to the NP (noun-phrase) position to form correctly wh + aux +NP+ verb. Mother and child are talking about the actions and sounds of different animals:

136 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:44

Color profile: Disabled Composite Default screen

The Findings

CHI: MOT: CHI:

137

Mummy what’s doing this? Gallopy gallopy gallopy .... That’s right Mummy.

(S17. 2;8.19)

In the following exchanges, Jenny’s questions become more complex. What is fronted to initial position in the question yet Jenny fails to invert the full subject into noun-phrase position. Instead the subject pronoun is placed after the auxiliary and the subject is repeated at the end of the question. For instance, Jenny is pointing to animals in a book and asks: CHI: What’s it doing this dzirrup? [giraffe] (S19. 2;9.22) MOT: Show me all your animals ## giraffe right. In the next example, in which Jenny is pointing out that her brother Jon Ander has some peanuts, the auxiliary is missing. CHI: What he doing Jon? (S23. 2;11.17) CHI: Hey where # Jon where he doing? CHI: ### There nair sitting the chair. MOT: He’s sitting on the chair is he? CHI: No. CHI: That he eat cacahuetes. (peanuts) From around 2;8.19 (S17) onwards there seems to be movement away from the what’s + it/this/that formula towards a more mature form with other pronouns (e.g. what’s it/he/this + doing) with repetition of the noun phrase. Other types of question development

This section looks at different what-questions that exhibit a mix of characteristics from different stages of question development along with those from ordinary development. Their complexity may also be as a result of cross-linguistic influence. The next examples shows correct use of have auxiliary insertion with done but without correct inversion. Instead, the subject (Daddy) is in final position. CHI: What has done Daddy? (S12. 2;6.19) MOT: Daddy’s gone to church. In the next example, Jenny combines what’s her name with got to produce the following question meaning something like an emphatic what IS her name. The ’s on what could be either is or has. MOT: What’s your dolly’s name? (S12. 2;6.19) MOT: Hmm?

137 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

Early Trilingualism

138

MOT: CHI: CHI: MOT: CHI:

What’s her name? ### Her name I don’t know. What’s her name a got it? Pardon? I don’t know.

The next example has correct insertion of does, although the subject it is omitted. It contains developmental errors such as the use of said, and may be influenced by the mother’s what do you say? in the third line of the exchange. Jenny and her mother are looking at a book together: CHI: Mummy book! (S14. 2;7.9) MOT: Oh the Noddy book, okay. MOT: What do you say? CHI: Thank you. CHI: What does here said? At age 3;0.3 the ungrammatical auxiliary *sham, a possible mixture of shall and am, is correctly inserted in an inverted question: CHI: Mummy what sham I do? (S26. 3;0.30) MOT: xxx the lambs. CHI: What sham I do? MOT: What shall you do? MOT: Watch a video. By 3;1.22 (S28) Jenny seems to understand the rules of inversion but not the rules of auxiliary insertion and, in general, questions become more complex. Here she is talking to a teddy bear: CHI: What you doing with my basket? (S29. 3;3.8) CHI: Mister! The examples shown here illustrate some of the variety and complexity of the question forms produced by a child in the process of acquisition. It is possible that such forms are the result of the developmental process as the child applies hypotheses to the language system she is acquiring – in this case interrogation. They may also be idiosyncratic or the result of crosslinguistic influence, or they may be a mixture of all these factors. Wh-confusion

The analysis of wh-questions revealed several instances where Jenny uses a wh-form with an unclear meaning or with the meaning of another wh-form. These cases have been termed wh-confusion. The next examples contain what questions that are similar in meaning to why.

138 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

The Findings

139

In the first example, Jenny is taking off her shoes and giving them to her mother. The meaning of her question as she hands them to her mother is not clear. Wh-fronting takes place, but the object of the question (the shoes) is not deleted; this structure is often found with why questions. CHI: Mummy Mummy I taking shoes. (S16. 2;8.3) MOT: Okay take your shoes off then. CHI: Mummy what you take the shoes? MOT: #### Thank you. In the second example Jenny and her mother are working in the kitchen: CHI: Mummy what are you doing runner beans like that? (S24. 3;0.5) MOT: That’s right. To summarise, after an initial dependence on formulas such as what that and what that for Jenny passes on to a stage in which what-questions begin with the phrase what’s + it/this/that with the later addition of doing. Inversion is seen to develop over time, but correct auxiliary inversion does not appear until around age three, and then in isolated cases. The child seems to depend on the phrase what’s + it/this/that (which includes an auxiliary that can be is or has) and, until this dependence decreases, she seems unable to move on to insertion of other auxiliaries such as can, will, and do. Our study includes evidence of question forms not reported in monolingual data, and these will be examined more closely later, in the section on cross-linguistic influence. However such uses may also be idiosyncratic since Bellugi (1965) found that each of her three monolingual subjects regularly used forms not used by the others. In Jenny’s case these include widespread use of the phrase what’s + it/this/that, non-standard word order for early questions (e.g. what doing Jenny?) and repetition of the subject (e.g. what he doing Jon?). The use of do as a main verb and as an auxiliary seems to present particular problems. Fully-formed where questions The simple nature of where-questions with the form where + is + NP, makes it easy for Jenny to make well-formed questions like these: CHI: Where’s Eeyore? (S10. 2;5.19) MOT: Where’s Eeyore? MOT: Here he is. CHI: MOT: CHI:

Hey where is it? Where is it? Where is what? Eh my dolly ### here.

139 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

(S11. 2;6.3)

Color profile: Disabled Composite Default screen

Early Trilingualism

140

Likewise, the phonology of where in certain environments facilitates the sound of contractions with are and you although these may have been only partially uttered or not at all, resulting in a more adult-sounding question. In this example, Jenny calls out after her mother: CHI: Hey where(y) going? (S11. 2;6.3) In the next two examples, it is unclear from the child’s pronunciation whether the contracted are has been joined onto where or not. Nevertheless in both cases the resulting question sounds fully formed. CHI: Where(’re) you going? (S24. 3;0.5) CHI: xxx. In the following question Jenny has lost her page, and means ‘where are we (in the book)?’ CHI: Now where(’re) we going? (S27. 3;1.9) CHI: Is got a ice-cream and is a fall off and is get a crying. Next we see the first instance of correct inversion with do insertion and a main verb put. Jenny is playing in the bath with a tea set, and is looking around for a coffee cup she has lost, but she does not know the word for cup. CHI: Where do you put the coffee here please? (S31. 3;5.18) MOT: Where did I put what? The examples show that the simple form where + is, along with the phonology of contractions in environments in which where is found, enable Jenny to make adult-like questions from the age of 2;5.19. Non-fully-formed where-questions Evidence of categories similar to those found in what-questions was found amongst the non-fully-formed where-questions studied, including dependence on a where’s it/this/that phrase and non-standard word order such as repetition of the subject. Wh-confusion was also found. Where’s it/this/that?

In the following example, Jenny is having trouble fixing the strap on her rucksack and asks her mother to help: CHI: Mummy this is not put it. (S16. 2;8.3) MOT: What? CHI: Where’s it put it? CHI: Done you you.

140 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

The Findings

141

In the next example, Jenny is looking for an apron with a cat on it: CHI: Where’s that my pussy cat lovely? (S22. 2;10.30) MOT: Which one? CHI: Yeah I said that. Subject repetition and lack of subject

In the following example we see how the child may repeat the subject as a pronoun or may use no subject at all. CHI: Where’s he Mikel? (S10. 2;5.19) CHI: Where go? Other types of question development

In the next example, Jenny suggests that they go and play outside. Her mother is not sure what this means: MOT: We’re going outside? (S4. 2;1.26) CHI: Yes goodbye. MOT: Goodbye. CHI: Where you go a? Here Jenny is demanding a box of toys and uses a where question as a strategy to achieve this. The question lacks auxiliary insertion. CHI: Mum! (S14. 2;7.9) CHI: I want it my box. CHI: Where you put it? Wh-confusion

In the next two examples, Jenny asks questions with a main verb put, but in both cases she omits the auxiliary. In the first one, it is not clear whether she is saying where or why; her mother understands that she has said why. CHI: Where (why) put it # ## where (why) put it? (S13. 2;7.1) CHI: Where (why) put it my? MOT: Why (where) put it my what? In the second example, Jenny and her mother are examining the video recorder and her mother interprets the meaning of where as what. CHI: Mum there where put? (S28. 3;1.22) MOT: That’s the video recorder. The development of where does not seem to present any special difficulties for Jenny apart from auxiliary insertion with main verbs, a problem that is found throughout question acquisition. Put appears frequently as a main verb in combination with where and Jenny uses plenty of well-formed

141 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

Early Trilingualism

142

questions. Her non fully-formed where-questions are not especially complex, and fall into categories similar to those found in her what-questions (such as where’s it, where’s this, where’s that dependence, and subject and complement repetition). Although aspects of Jenny’s phonological development are beyond the scope of this study, she seems to have difficulty distinguishing the different wh-words from each other phonetically. Fully–formed who-questions Who-questions are different from what-questions and where-questions in that they usually enquire about the subject rather than about the complement of the verb. The formation of who-questions about the subject requires wh-substitution rather than wh-fronting, and no further transformations are necessary for simple questions to be correctly formed. The following description of Jenny’s who-question acquisition, whilst discussing traits in fully-formed and non-fully-formed questions, is therefore different from the previous descriptions of what and where. Who’s + correct verb/complement

In the example Jenny is picking out clothes for her teddy bear: CHI: Too big too big. (S13. 2;7.1) CHI: Who’s done it? CHI: One too big. Jenny asks the following questions about toys she has found. Both are well-formed and used in context. CHI: Who’s bought this? (S30. 3;4.7) MOT: I don’t know who bought that, but it’s very nice. CHI: MOT: CHI: MOT:

Oh it’s got broken It’s a broken dolly, isn’t it? Who’s broken it? I don’t know.

(S31. 3;5.18)

Both questions use the present perfect tense where the simple past (who bought it? and who broke it?) would be more appropriate. This may be crosslinguistic influence from Basque and Spanish, where the present perfect is widely used in relation to present circumstances relating to the past. Who + correct verb

In the next example, Jenny is serving drinks with her tea set: CHI: Drink it! (S29. 3;3.8)

142 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

The Findings

MOT: CHI: CHI:

143

Mmm. Mmm delicious. Who wants sugar with with with milk?

In these examples, Jenny is able to make fully-formed questions by combining who’s with different verbs and complements. Non-fully-formed who-questions Who’s + incorrect verb

In the next example, Jenny combines who’s with an incorrect verb form. Her mother is showing her a plasticine model her brother Jon Ander has made: CHI: Who’s do that small one? (S30. 3;4.7) MOT: Jon Ander did it for me. Who + incorrect verb

The following is an example of the use of who with an incorrect verb form. CHI: Who’s in my bed? (S13.2;7.1) CHI: Who done it? Who + no verb

In the following example Jenny is putting away some pictures to help with a puzzle. She does not place a verb with who, and may be trying to say whose or whose is. CHI: Who this picture? (S22. 2;10.30) CHI: Let’s see if you can do this. Other types of question development

The next example shows a failure to insert the auxiliary. Jenny is looking for a doll. CHI: Who got the dolly ### Noddy? (S12. 2;6.19) CHI: Hmm? CHI: Dolly. In the next example, Jenny and her mother are talking about the animals they saw on a visit to the zoo. Jenny seems to be trying to use the past tense as well as asking a who-question about an object, which requires inversion and auxiliary insertion. There also appears to be wh-confusion. CHI: Who we sawed it this? (S27. 3;1.9) CHI: We sawn this. MOT: Oh you saw that? MOT: What is it?

143 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

Early Trilingualism

144

CHI:

Hmm elephant.

Wh-confusion

Jenny and mother are talking about a jungle story: CHI: Here a elephant. MOT: Elephant yes # here you are elephant. CHI: Who’s done the elephant? MOT: What’s he doing? CHI: Poo!

(S19. 2;9.22)

The question is well formed, but Jenny has confused the meaning of who and what. In general Jenny finds it easy to produce well-formed who questions because of the flexibility of who’s and the lack of complexity involved with asking questions about the subject. As in the case of what and where questions, there is evidence of wh-confusion.

How questions How first appears as part of a routine in which Jenny is talking to one of her toys: CHI: Hello bear xxx. (S10. 2;5.19) CHI: Hey how are you? CHI: I’m fine thanks. How is not found again in the recordings until five months later, when Jenny and her mother are looking at a book and Jenny asks what a character is doing on a seesaw: CHI: How is it do Goofy? (S21. 2;10.14) MOT: That’s right. MOT: There he is on the other end of the seesaw. Although a main verb do is used, here it is preceded by how + is + it, similar to the what’s + it, where’s + it question formula used widely by Jenny elsewhere. In the next example Jenny and her mother are talking about their visit to the zoo, and Jenny asks the following question using how + many: CHI: And how many teddies we see there? (S27. 3;1.9) MOT: How many teddies? MOT: Where in Cabarceno? Apart from the absence of auxiliary insertion the question is otherwise fully formed.

144 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

The Findings

145

The next two examples both lack auxiliary insertion. In the first one, Jenny wants to put on a necklace: CHI: How you put this on? (S30. 3;4.7) MOT: Like that. In the second example, the meaning of how is not clear, though the context suggests that it means why don’t you? Jenny may have used how since such language is not available to her yet. It may also be an example of wh-word confusion as where, why or when would make more sense. CHI: Mummy Mummy. (S30. 3;4.7) CHI: Mummy how you buy one of these for Leire? MOT: Buy one of those for Leire? In the final use of how in the data, Jenny uses the wh-word correctly to ask about the way her cousins play the violin, an instrument she is not familiar with. MOT: Kepa and Uxue play the violin. (S31. 3;5.18) MOT: That’s nice isn’t it? CHI: How? MOT: That’s how they go. (mimes violin)

Why-questions Nine uses of why were found in the data. The first instance appears as Jenny is talking to a bath toy that is leaking: CHI: No # in a towel. (S11. 2;6.3) CHI: Oh # why why? CHI: Silly man. The next case involves wh-word confusion. Jenny is trying unsuccessfully to put on a T-shirt. Finally, in exasperation she asks her mother a question, which appears to mean how. CHI: Oh dear. (S12. 2;6.19) CHI: Oh dear. CHI: Why is do then? There is also evidence of a question formula, in this case why + is, used with a main verb do that has been seen elsewhere in the data in Jenny’s whquestions. In the next example Jenny uses a main verb fall with a negative, though the auxiliary is omitted. Her doll has fallen behind the cot, so the meaning of the negative in this question is unclear.

145 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:45

Color profile: Disabled Composite Default screen

Early Trilingualism

146

CHI: CHI: CHI:

Why not dolly fall? Mummy? Dolly’s gone here.

(S14. 2;7.9)

In this example, Jenny is looking at some clothes for her doll: CHI: Look dolly’s clothes. (S25. 3;0.19) CHI: Mum. CHI: Dolly’s clothes more. CHI: Why this is doom (doing) so small Mum? In this case, Jenny does not use her common strategy of wh-question formula wh + ’s + it/this/that. She fronts wh- (why) but leaves the rest of the question in the affirmative. This order is frequently found with why questions (Labov & Labov, 1978). The following exchange contains three instances of the insistent use of causal why often reported in monolingual studies (Piaget, 1926; Fletcher, 1985). Mother and child are looking at a broken doll: CHI: Who is broken it? (S31. 3;5.18) MOT: I don’t know. MOT: Perhaps she fell on the floor. CHI: Why? CHI: Why? CHI: Why? MOT: Why? MOT: She must have fallen off the picture. The last case appears in an exchange in which Jenny and her mother are discussing a sibling’s behaviour with a blanket. Apart from Jenny’s use of why as a question, we can also see how it is used without rising intonation to mean because. In Spanish, why and because are expressed by the same word porque, and the same sometimes happens (incorrectly as a result of transfer from Spanish) with the Basque zergatik. MOT: Why’s he got a blanket there? (S32. 3;6.18) CHI: Why xxx. (meaning = because) MOT: Why did Jon Ander put a blanket there? CHI: Why he it did? CHI: ### We got here slippers! Jenny’s why question does not show inversion or auxiliary insertion in question form, but the object is located next to the subject. This may be an attempt at inversion, the result of cross-linguistic influence, or an idiosyncratic developmental error.

146 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:46

Color profile: Disabled Composite Default screen

The Findings

147

Why is found in our data from 2;6 onwards and (excluding repetitions) is used by Jenny on only six separate occasions. In two of these, the meaning she gives to the question particle is not entirely clear. These results are clearly in contrast to monolingual data where use of why is more widespread by this age.

When-questions Only two cases of when appear in the data. Both of them have a full verb, but neither has inversion or auxiliary insertion. The meaning of the question in the first example is not clear, and the mother misunderstands. However, from the context it seems that Jenny wants to know when her mother will put the books in her rucksack (motxila). CHI: Mummy? (S15. 2;7.19) MOT: Yes. CHI: When it comes in your book? MOT: Come and show me your books again. MOT: Come and show me your book. CHI: No no! CHI: I want it them motxil in here. In the second use of when in a question, Jenny asks a rhetorical question about herself that could be interpreted as a what-question. She is playing in the bath and she and her mother are being playful: MOT: Jenny! (laughing) (S31. 3;5.18) CHI: When I do this? CHI: (sings) CHI: Look I jumping. Summary of wh-questions Jenny’s progress in wh-question development will now be related to that described by Klima and Bellugi (1966) and Cazden (1970) whereby the child passes through periods. The summary of progress at each period will refer in the main to what-, where- and who-questions, since these make up most of the data. Bellugi’s Period A

According to Cazden (1970), in this period2 wh-questions are limited to a series of routines. This is the case in Jenny’s data, though one of the routines described by Cazden (1970) during this period (what + NP + doing) does not

147 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:46

Color profile: Disabled Composite Default screen

Early Trilingualism

148

appear until Period B. Instead we find the appearance of what + doing + NP. The following are examples from Period A: • • • •

Wha da? What that for? What’s that Mummy? Who’s that Mummy?

(S1. 1;11.24) (S2. 2;0.6) (S4. 2;1.26) (S6. 2;3.13)

Bellugi’s Period B

Cazden (1970: 207) describes this period3 as one in which ‘more complex sentences (are) being asked but without the development of question forms.’ In Jenny’s case, the period is marked by the use of the phrase what’s/ where’s + it/this/that with a main verb, often get or do, or with verb + ing, particularly in the later part of the period. No such examples appear in Klima and Bellugi’s (1966: 202) list of sample questions at this stage. Who-questions appear with and without the auxiliary ´s and tend to be well-formed. There are occasional attempts at questions without what’s/where’s + it/this/that. • What’s put in here? (S10. 2;5.19) • What the teddy doing? (S11. 2;6.3) • Who got the dolly # noddy? (S12. 2;6.19) • Who’s in my bed? (S13. 2;7.1) • What does here said? (S14. 2;7.9) • What’s this is done it? (S17. 2;8.19) • What’s it doing this giraffe? (S19. 2;9.22) • Mummy where the my box? (S22. 2;10.30) • What you eating? (S23. 2;11.17) Several examples of why and when appear in this period, some of which partially use the phrase wh-(’s) + it/this/that? • Why is do then? • When it comes in your book?

(S12. 2;6.19) (S15. 2;7.19)

Bellugi’s Period C

Cazden (1970: 207) describes this period4 as being one in which auxiliary verbs develop throughout the entire grammatical system and in yes/no questions, though not in wh- questions. We have already seen that, in Jenny’s yes/no questions, inversion and auxiliary insertion are present during this period along with other forms, particularly is + 0/this/that? In the case of wh-questions, the picture is somewhat different. Jenny continues to use wh-(’s) + it/this/that + (ing) but shows less dependency on it, as questions containing other combinations of verb forms with and without auxiliary insertion and inversion begin to be more frequent, although they are

148 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:46

Color profile: Disabled Composite Default screen

The Findings

149

rarely fully-formed. Fully-formed questions tend to consist of to be + -ing rather than do + verb. Only three examples are found containing a full verb with correct do insertion and inversion: what do you say? (at age 3;1.22) and what do you want? and where do you put the coffee here please?(at age 3;5.18). Questions with subject repetition (such as, what’s he doing Jon?) are also found from age 2;9 and onwards into this period: • What’s this doing here this button? (S25. 3;0.19) • What’s this eggy doing over here? (S26. 3;0.30) • Who’s your name? (S28. 3;1.22) • What you doing with my basket? (S29. 3;3.8) • Where’s the other rabbit of this? (S30. 3;4.7) • Who did this? (S31. 3;5.18) • How many teddies we see there? (S27. 3;1.9) If we compare the findings for yes/no questions with those for wh-questions we see some differences. Jenny’s development in yes/no questions is characterised by declarative questions and the use of is + 0/it/this/that, along with a progressive increase in inversion and auxiliary insertion from Period B. Jenny’s wh- questions develop from early routines to widespread use of wh+ (’s) + it/this/that (+ doing). From around age 2;9 examples are found of subject repetition and appearance of wh- in non-fronted position. There is movement away from use of the wh+ (’s) + it/this/that phrase and towards use of full verbs with or without a correctly placed auxiliary, though the first wh-questions showing complete auxiliary insertion and inversion appear in Bellugi’s Period C, later than in yes/no questions.

Other Questions The acquisition of other questions The category other questions contains all those questions that do not fit into either the wh-questions or the yes/no questions (including intonation and declarative questions) classification. In our data, 146 other questions have been identified, and 126 of these are of the type eh? used as a request for clarification or confirmation – such questions will not be considered here as they do not show aspects of form. However there are several examples of alternative questions and echo questions that are particularly interesting because they show how the child is able to use questions functionally to maintain the conversation Echo-questions

In the following example Jenny and her mother are looking through some old baby things:

149 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:46

Color profile: Disabled Composite Default screen

Early Trilingualism

150

MOT: CHI:

Do you know where Mummy put the bottles? Where?

(S18. 2;9.5)

In the next example, mother and child are getting the lunch ready: MOT: xxx do a few carrots as well. (S24. 3;0.5) CHI: Do carrots? MOT: You can carry on doing the beans. From age 2;9 onwards Jenny is able to pick out words (such as carrot in the example above) and wh-words (such as where above) in her mother’s utterances and use them correctly in echo-questions to continue the discourse. Alternative questions

In this example, mother and child are playing with bricks: CHI: Mummy this or this? (S16. 2;8.3) MOT: Have you got all the bits? Here, they are looking at family photos: MOT: That’s Aunty Sally, isn’t it? CHI: Where in wour house or in him house? (gloss: Where in our house or in her house?) CHI: Him house. (child answers own question)

(S31. 3;5.18)

40% 35% 30% 25% 20% 15% 10% 5% 0%

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 Scripts

Figure 4.9 Percentage of other questions in total interrogatives produced

150 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:50

Color profile: Disabled Composite Default screen

The Findings

151

We compared the frequency of other questions with that of those in the other four formal categories. As Figure 4.9 shows, the percentage of other questions in the child’s interrogative behaviour decreases over time, as was seen in the case of intonation questions and declarative questions. Presumably as the child has more language at her disposal she has less need to resort to expressions such as eh? for confirmation and clarification.

Pragmatic Development The child’s and the mother’s questions have been examined not only from a formal perspective, but also as part of five functional categories: real information, known information, confirmation, clarification, and interactive. The analyses described in this section were performed in order to find out how the child uses her questions functionally at different ages and as part of her communicative competence. The functions of the child’s questions will now be considered at different stages to see the numbers of questions used in each function, the length of the questions used and what forms are used in each function. Then we will analyse some extracts of discourse in which Jenny can be observed using the question functions as part of her communicative competence. Next, the mother’s questions are examined from a functional perspective to see if this has any relation to the trends seen in Jenny’s question functions. Finally both the child’s and the mother’s feedback to questions will be examined according to the function to see if there is a relationship between the two. The child’s question functions Number of questions

Figures 4.10–4.14 show the number of questions used in each recording in each functional category. Upward and downward trends have been marked. Figure 4.10 shows that real information questions were asked frequently from the beginning of the sample and their number increased steadily between the ages of 1;11.24 (S1) and 3;6.18 (S32). Figure 4.11 shows that fewer known information questions were asked and growth in this category does not follow a regular pattern. However, in the earliest recordings – between 2;0.6 (S2) and 2;2.24 (S5) – Jenny asks no known information questions at all, though over time there is limited growth in the number of known information questions asked by her. Figure 4.12 shows a similarly erratic distribution in the numbers of confirmation questions Jenny asks, varying between 0 and 12 questions irrespective of age. There is evidence that Jenny is able to use confirmation

151 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:08

Color profile: Disabled Composite Default screen

Early Trilingualism

152 90 80

Number of questions

70 60 50 40 30 20 10 0

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 Scripts

Figure 4.10 Number of real information questions asked by Jenny

25

Number of questions

20

15

10

5

0

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 Scripts

Figure 4.11 Number of known information questions asked by Jenny

152 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:53:58

Color profile: Disabled Composite Default screen

The Findings

153

14

Number of questions

12 10 8 6 4 2 0

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 Scripts

Figure 4.12 Number of confirmation questions asked by Jenny questions from age 2;0.6 (S2), when 9 are asked, and that she is able to use them throughout the sample. Thus there is no evidence of growth in the number of confirmation questions, rather a slight decrease is shown. Clarification questions are also used from the beginning of the period under study (Figure 4.13) and Jenny appears to use more of them than confirmation questions. There is a clear decrease in the number of clarification questions used over time, which suggests that there may be less need for them as the child has more language at her disposal and becomes more proficient at understanding. Figure 4.14 shows Jenny’s development of interactive questions. Not many questions are found, the maximum being 11 at age 2;7.1 (S13), but they are present throughout the sample and show a tendency towards growth. Of all the functional categories, real information questions are present in the largest numbers and continue to increase, which may reflect the child’s growing desire for knowledge. The numbers of confirmation and clarification questions decrease, whilst known information and interactive questions become more frequent, possibly in line with Jenny’s known information from the beginning of the sample, at around the age of 2.

153 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:03

Color profile: Disabled Composite Default screen

Early Trilingualism

154 30

Number of questions

25

20

15

10

5

0

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 Scripts

Figure 4.13 Number of clarification questions asked by Jenny

12

Number of questions

10

8

6

4

2

0 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 Scripts

Figure 4.14 Number of interactive questions asked by Jenny

154 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:13

Color profile: Disabled Composite Default screen

The Findings

155

Length of questions

Figures 4.15–4.19 show the average number of words used by Jenny in questions in each category. It is not surprising that length of question in all the categories increases, along with the child’s MLU, over time. However it is of interest to note that some categories tend to use longer questions than others. The growth in length of real information questions is shown in Figure 4.15. A steady development can be seen from an average of 1.25 words in the first sample at 1;11.24 (S1) to 4.13 in the last one at 3;6.18 (S32). Figure 4.16 shows that known information questions are in general slightly longer than real information questions and range from 2.5–5.0 words, and sometimes longer. The tendency for growth is almost identical to that found in real information questions. Confirmation and clarification questions are shorter than those in the other functional categories. Figure 4.17 shows that confirmation questions rarely reach more than 2.5 words in length, and many are around 1.0 and 1.5. Clarification questions shown in Figure 4.18 are even shorter, most usually with a length of between 1.0 and 1.5 words. There is an exception at age 2;10.14 (S21) when Jenny asks who cleaning his teeth? Many of the one-

4.5 4.0

Number of words

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0

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 Scripts

Figure 4.15 Length of real information questions

155 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:17

Color profile: Disabled Composite Default screen

Early Trilingualism

156 8 7

Number of words

6 5 4 3 2 1 0 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 Scripts

Figure 4.16 Length of known information questions

3.5

Number of words

3.0 2.5 2.0 1.5 1.0 0.5 0

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 Scripts

Figure 4.17 Length of confirmation questions

156 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:24

Color profile: Disabled Composite Default screen

The Findings

157

4.5 4.0

Number of words

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0

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 Scripts

Figure 4.18 Length of clarification questions word clarification questions are expressions such as eh? hmm? what? and pardon? The interactive questions show more variety in length (Figure 4.19), and there is steady growth from the early questions at around age 3 (S1–S3, S5 and S6), most of which are 1–2 words long – with the exception of 5 words at 2;1.26 (S4). An average of approximately 5 words is reached at around age 3 (S22, S23, S25, S26). After age 3;1.9 (S27) growth appears to drop off. To sum up, question length varies between functional categories. The longest questions are found in the interactive category and the shortest in clarification and then in confirmation. Questions in the real information and known information categories are approximately the same length over time. Type of question form used in each function

In Figures 4.20–4.24 each functional category is described in terms of the distribution of the question forms found in it over the whole corpus. Considerable differences can be seen from one category to another. Figure 4.20 shows that 53% of the questions used in the real information category are wh-questions, 36% are declarative questions, 7% are yes/no questions, and the remainder are intonation and other questions. Wh-questions (71%) and declarative questions (22%) are also the most used

157 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:28

Color profile: Disabled Composite Default screen

Early Trilingualism

158 7

Number of words

6 5 4 3 2 1 0

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 Scripts

Figure 4.19 Length of interactive questions in known information questions (Figure 4.21). Yes/no questions and intonation questions both make up 3% and the remaining 1% are other questions. Confirmation questions (Figure 4.22) show an altogether different distribution. Here 87% of questions are declarative, 6% are wh-questions, 5% are other questions, while intonation and yes/no questions are both 1%.

Any other 1% Declarative 36%

Yes/no 7%

Wh-questions 53%

Intonation 3%

Figure 4.20 Question forms used in real information questions

158 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:33

Color profile: Disabled Composite Default screen

The Findings

159

Any other Yes/no 1% 3%

Declarative 22%

Intonation 3% Wh-questions 71%

Figure 4.21 Question forms used in known information questions Clarification questions are different again (Figure 4.23) In this category other questions make up 71%., while both wh-questions and declarative questions make up 14% each. Intonation questions make up 1%. No yes/no questions are found. Interactive questions (Figure 4.24) show a more balanced picture in which declarative questions make up 49%, wh-questions make up 43% and yes/no questions and intonation questions make up 4% and 3% respectively. Other questions make up the remaining 1%.

Wh-questions 6%

Any other 5%

Intonation 1% Yes/no 1%

Declarative 87%

Figure 4.22 Question forms used in confirmation questions

159 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:35

Color profile: Disabled Composite Default screen

Early Trilingualism

160

Intonation 1%

Declarative 14% Wh-questions 14%

Any other 71%

Figure 4.23 Question forms used in clarification questions Overall, the results show that there is a clear tendency for question forms to be used in different proportions for certain functions. The forms most used are wh-questions and declarative questions, except in the case of clarification questions where other questions make up by far the largest proportion. This variation suggests that the child is aware that different question forms can be used to serve different functional purposes.

Any other 1%

Intonation 3%

Yes/no 4%

Wh-questions 43%

Declarative 49%

Figure 4.24 Question forms used in interactive questions

160 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:37

Color profile: Disabled Composite Default screen

The Findings

161

Question Functions as Part of the Child’s Communicative Competence In the course of analysing the data from both a formal and a functional perspective, it became apparent that the child was involved in negotiating meaning with her mother through the use of a range of questioning strategies. The child asked many questions pragmatically to initiate topics, to ask about new words, to avoid the truth and even to display a sense of humour. Some of these questions involved repetition and reformulation, others used known questions to initiate an exchange. It seemed that the child had in some way learned what was expected of her as a participant in the discourse and was eager to take the lead. In the light of what has been presented in relation to question functions, we will now examine a number of exchanges in which the child displays such communicative strategies. The extracts will be described chronologically so that the growth in the child’s communicative competence can be observed and, as before, the extracts have been demarked into the question periods described by Klima and Bellugi (1966) and Cazden (1970). Bellugi’s Period A

Although the following interaction includes only two Jenny questions, one known (line 4) and one confirmation (line 20), they are pivotal to her strategy of denying responsibility for the damaged tube, and for this reason all three exchanges are included. In line 4 the relatively complex who did? is probably formulaic and possibly picked up from the mother telling off Jenny’s brothers with who did this? As a speech act it is a lie, since Jenny later confesses to having damaged the tube. By the same token the confirmation question here? is deliberately misleading so that Jenny’s mother does not find the missing tube top. • Avoiding the truth 1 2 3 4 5 6 7 8 9 10 11

CHI: MOT: CHI: CHI: MOT: MOT: MOT: CHI: MOT: MOT: CHI:

It broken. Oh dear, it’s broken is it? Yes. Who did? Who broke it? Who broke it then? Who broke it? Here. (evading answer) No no no. Who broke this? (SILENCE)

161 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:37

(S3. 2;1.2)

Color profile: Disabled Composite Default screen

Early Trilingualism

162

12 13 14 15 16 17 18 19 20 21 22 23

MOT: MOT: MOT: MOT: MOT: MOT: MOT: MOT: CHI: MOT: MOT: CHI:

Hmm. Where’s the bit? Where’s the bit? Go and pick it up please. I can see it. It’s under your bed. Under your bed. Go and get # it’s under your bed please. Here? No that’s not under your bed is it? Where’s your bed? Here.

Bellugi’s Period B

In the next example, Jenny is identifying animals in a picture book and has so far identified a dog and a duck but does not know the words for tortoise (she transfers from Basque, line 1) or mouse. It may be that she does not know the word for mouse in Basque and therefore does not attempt transfer as a relief strategy, although she does introduce an idiosyncratic word laiya at this stage. Instead, she uses two real information questions (lines 5 and 6) to ask her mother directly about the animal’s name, although she already ‘knows’ what the animal is, as is made clear in lines 9 and 10. • Seeking a new lexical item 1 2 3 4 5 6 7 8 9 10

CHI: CHI: CHI: CHI: CHI: CHI: MOT: CHI: MOT: CHI:

Mum look tortoka ata orea. Look # # here. Eh # look #this one. This # a laiya. What is it? It’s what’s that? A mouse. A mouse. Is it like in Tom and Jerry? Yes in Tom and Jerry.

(S9. 2;5.5)

The following example consists of two extracts. In the first, Jenny initiates the exchange with a real information question (line 1) about a character in a book. The mother replies (line 2) and goes on to extend the exchange by means of known information questions (3, 5, 9 and 12) in different reformulations in a playful attempt to elicit the name of a child, William, who had hat a similar to that of the character.

162 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:37

Color profile: Disabled Composite Default screen

The Findings

163

• Referring back to make a joke 1 2 3 4 5 6 7 8 9 10 11 12 13 14

CHI: What this? (S16. 2;8.3) MOT: That’s Mr Plod the policeman. MOT: Who has a hat like that? MOT: Do you remember? MOT: Who’s got a hat like that? CHI: Bet. CHI: Noddy. MOT: Noddy’s got a hat like that. MOT: Who had a hat like that on the other day? CHI: Like Jonander. MOT: Jonander? MOT: I thought William had a hat like that, didn’t he? CHI: Wee. CHI: Put it. In the second extract, later in the recording after several exchanges relating to the same story, Jenny again initiates the exchange in line 57, this time with a known information question (she already knows the answer from line 2). The known information questions she asks in lines 59 and 61, although not fully-formed, mirror the type of questions asked by the mother in lines 3, 5, 9 and 12, including the word like, and have the same intent of eliciting the name William. Jenny seems to be trying to copy her mother’s questioning strategies and to involve her in a ‘joke,’ recalling their earlier comments. 57 CHI: What this? 58 MOT: That’s Mr Plod the policeman isn’t it? 59 CHI: What’s he got a like hat? 60 MOT: Got a funny hat, hasn’t he? 61 CHI: What’s he got a like hat? 62 CHI: Like William. 63 MOT: Like William’s hat wasn’t it? 64 CHI: Yes. Bellugi’s Period C

In the next example Jenny shows that she is able to use interactive questions functionally to express feigned surprise that she has wet herself. • Showing surprise 1 2 3 4

MOT: CHI: CHI: MOT:

All the toys back in the box. What a my weewee? What a my weewee doing here? Ooh.

163 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:37

(S26. 3;0.30)

Color profile: Disabled Composite Default screen

Early Trilingualism

164

In the next exchange we see how Jenny’s questioning behaviour has become more sophisticated as she seeks real information about how the video works. Although the forms are simple declarative questions (lines 6, 8, 19 and 22) and wh-questions (lines 1, 4, 11, 13 and 14) 7 out of the 9 real information questions that Jenny asks about the video camera are reformulations of a basic theme. • Seeking information 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

CHI: MOT: MOT: CHI: MOT: CHI: MOT: CHI: MOT: MOT: CHI: MOT: CHI: CHI: MOT: MOT: MOT: CHI: MOT: CHI: MOT: CHI:

Mumum there where put? That’s the video. There we are. Mum what’s this? A button to switch it on and off. And this? That’s the button to take the cassette in now. And this? Don’t fiddle with that Jenny. No no no don’t do that! Mummy what’s that? You mustn’t touch that. Mummy what’s that? What’s this? That’s for when I’m making a copy. You see that little door? A fairy goes in there. There. A fairy goes all the way down through there. This? Whoops. Like this?

(S28. 3;1.22)

In the next example, Jenny is looking in a bag of old baby toys. As before, her varied and inquisitive real information questions reflect her linguistic and psychological development. From a total of 8 questions that Jenny asks in the exchange (lines 1, 5, 8, 9, 11, 14, 18 and 20), only two are repeated (9 and 14). Lines 18 and 20 are confirmation questions. • Showing interest and enthusiasm 1 2 3

CHI: MOT: CHI:

What for is open with this keys? Whose keys are they? I think those Leire’s keys.

164 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:08

(S30. 3;4.7)

Color profile: Disabled Composite Default screen

The Findings

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

CHI: CHI: MOT: MOT: CHI: CHI: MOT: CHI: MOT: CHI: CHI: MOT: CHI: MOT: CHI: MOT: CHI: MOT:

165

And this is. ### Mummy what’s this for doing? This? Look it makes a squeak. And this? What’s this for doing? For making a squeak as well. Who’s buy this duck? Mummy bought that. Hey go round! What’s this for doing? It rattles if you touch it. xxx. You put it on your pram. Pram here? And it goes rattle rattle rattle. Like this? That’s it.

In the next example, Jenny uses confirmation questions to herself while doing a jigsaw puzzle. MOT: Noddy’s hats. (S22. 2;10.30) CHI: This is? MOT: Ahah. CHI: No. CHI: This is? CHI: Oh nothing. CHI: This is? MOT: That’s the top! CHI: Top. CHI: What’s that there? CHI: Does the this in here? MOT: Ahah. CHI: This is the here. The description presented here demonstrates the application of the taxonomy to practical examples. Such examples illustrate Jenny’s developing pragmatic awareness. The extracts show that she is able to use questions pragmatically in strategies as part of her communicative competence, and it may be that such routines are a type of training for strategies she will later apply in conversation. One of Jenny’s most common uses of questions

165 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:37

Color profile: Disabled Composite Default screen

Early Trilingualism

166

is using question strategies to obtain lexical items. She also shows evidence of metalinguistic awareness in her strategies to find out how to express concepts in English for which she may already have the word(s) in her other languages. She is also able to initiate and extend play, manipulate the truth, attract attention, express surprise and initiate and maintain discourse.

The Mother’s Question Functions The results of the analysis of a sample of nine recordings from the child’s mother will now be described. The purpose of the analysis is to see if the child’s functional use of questions and her ability to give feedback to questions is related to the input and feedback she has received from her mother. As in the case of the child the number of questions, length of question and type of question form used in each functional category has been measured. The scripts have been chosen at the following intervals of the child’s development: Script 1 Script 4 Script 8 Script 12 Script 16 Script 20 Script 24 Script 28 Script 32

1;11.24 2;1.26 2;4.17 2;6.19 2;8.3 2;10.2 3;0.5 3;1.22 3;6.18

The classification of the mother’s questions included a sixth functional category, affective questions. However, she did not produce such questions frequently and the child did not produce any affective questions at all, so they are not considered here. Number of questions Figures 4.25–4.29 show the number of questions that Jenny’s mother asked in each of the functional categories; the lines indicate trends. The numbers of real information questions produced are shown in Figure 4.25. Although the numbers when Jenny was aged 1;11.24 (S1), 2;11.26 (S4) and 2;4.17 (S8) display large differences, the numbers are generally stable at around 50 questions. The first recording at 1;11.24 (S1) and the last at 3;6.18 (S32) have almost the same number of questions (40 and 43 respectively). Furthermore the scripts between age 2;6.3 (S12) and 3;6.18 (S32) are fairly stable in numbers (around 50). This is reflected in the trend marker line.

166 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:37

Color profile: Disabled Composite Default screen

The Findings

167

100 90

Number of questions

80 70 60 50 40 30 20 10 0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.25 Numbers of real information questions asked by the mother

120

Number of questions

100

80

60

40

20

0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.26 Number of known information questions asked by the mother

167 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:41

Color profile: Disabled Composite Default screen

Early Trilingualism

168

Numbers of known information questions (Figure 4.26) show a somewhat similar picture of stability, as indicated by the trend marker line. Apart from a jump at age 2;10.2 (S20) the numbers of questions asked remain at around 40–60 from the earliest recordings to the last ones. Fewer confirmation questions are produced than in previous categories, ranging from 6 questions at age 2;4.17 (S8) to 14 at age 2;10.2 (S20). However, there is upwards growth (Figure 4.27). The first recording at 1;11.24 (S1) has 7 questions whilst the last at 3;6.18 S32) has 13 questions. Clarification questions are not frequent either (Figure 4.28), and there is a slight tendency for them to decrease over time. There were 11 questions in the first script at 1;11.24 (S1) and only 13 in the final one at 3;6.18 (S32), though there is a dramatic increase in use at ages 2;6.19 (S12) and 2;8.3 (S16) following a drop to 7 questions at age 2;4.17 (S8). Figure 4.29 shows that the category with fewest questions is interactive, with just four questions at ages 2;1.26 (S4), 2;4.17 (S8) and 3;0.5 (S24), and five at ages 2;6.19 (S12), 2;8.3 (S16), 2;10.2 (S20) and 3;6.18 (S32). There is a tendency towards a slight decrease in these questions. So, Figures 4.25–4.29 show that, except in the case of confirmation questions, the mother uses approximately the same number of questions in all functional categories at the beginning of the study when Jenny is aged 16 14

Number of questions

12 10 8 6 4 2 0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.27 Number of confirmation questions asked by the mother

168 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:43

Color profile: Disabled Composite Default screen

The Findings

169

35

Number of questions

30 25 20 15 10 5 0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.28 Number of clarification questions asked by the mother

10 9

Number of questions

8 7 6 5 4 3 2 1 0

1

4

8

12

16

20

24

28

Scripts

Figure 4.29 Numbers of interactive questions asked by the mother

169 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:47

32

Color profile: Disabled Composite Default screen

Early Trilingualism

170

1;11.24 (S1) and at the end when Jenny is 3;6.18 (S32). The trend marker lines show insignificant movement in the case of real and known information questions and a slight downward trend in the case of clarification and interactive questions. Length of questions Figures 4.30–4.34 show the average length of the mother’s questions in each functional category. The mother’s questions in the real information category (Figure 4.30) remain at a constant average length of around 5 words per question throughout the study. This is surprising when contrasted with Figure 4.31, which shows a clear increase in the length of known information questions over time. At the beginning of the study, when Jenny was 1;11.24 (S1) her mother uses questions with an average length of just under 1.5 words per question. By Jenny’s age 2;10.2 (S20) this has increased to an average of over 2.5 words. Question length reaches a peak of more than 2.5 words between ages 2;8.3 (S16) and 3;0.5 (S20), followed by a drop to less than 2 words per question before another rise. By the end of the study when Jenny is 3;6.18 (S32), her mother is using questions with an average length of just under 2.5 words. 6

5

Number of words

4

3

2

1

0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.30 Length of real information questions asked by the mother

170 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:49

Color profile: Disabled Composite Default screen

The Findings

171

Figure 4.32 shows that in this study confirmation questions start at an average of around 3 words per question when the child is 1;11.24 (S1) and by the end of the study the average number of words has increased to around 3.5. The trend marker line indicates that there is an overall downward tendency in average length of question. In spite of this, a rise to an average of four words and over can be seen from age 2;1.26 (S4) until 2;10.2 (S20), reaching a peak at age 2;8.3 (S16). Figure 4.33 shows that the mother’s clarification questions tend to be short and remain at an average length of 1.5–2.5 words throughout the sample. However, like known information and confirmation questions, their length reaches a peak at around Jenny’s age 2;8.3 (S16). A slight downward tendency can be seen in the trend marker line. The average length of interactive questions (Figure 4.34) varies considerably, but the overall tendency in this category is towards a slight reduction in the length of questions. To sum up, except in the case of known information questions there is a tendency for the mother’s questions to reduce slightly in length over time. Real information and interactive questions are the longest, with an average of around 5 words whilst known information and clarification are the

3.0

Number of words

2.5

2.0

1.5

1.0

0.5

0.0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.31 Length of known information questions asked by the mother

171 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:51

Color profile: Disabled Composite Default screen

Early Trilingualism

172 5.0 4.5 4.0

Number of words

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.32 Length of confirmation questions asked by Jenny’s mother

3.0

Number of words

2.5

2.0

1.5

1.0

0.5

0.0

1

4

8

12

16

20

24

28

Scripts

Figure 4.33 Length of clarification questions asked by the mother

172 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:55

32

Color profile: Disabled Composite Default screen

The Findings

173

7 6

Number of words

5 4 3 2 1 0

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.34 Length of interactive questions asked by the mother shortest with 1.5–2.5 words. Known information, confirmation and clarification questions reach a peak in length when Jenny was around the ages of 2;8.3 (S16) and 2;10.2 (S20). Type of question form used in each function category Figures 4.35–4.39 show the question forms used by Jenny’s mother in the functional categories. The mother’s data is obviously richer than the child’s because her linguistic system is fully developed. In order to compare like with like, we should remember that according to Greenbaum and Quirk (1990) tag questions and declarative questions form part of a yes/no type formal category of questions. Whilst a child may be limited to using mainly declarative questions since yes/no questions are as yet largely undeveloped, an adult has a much wider related repertoire at her disposal and may be using, for example, tag questions, positive yes/no questions, negative yes/no questions and declarative questions. For the purposes of comparison, these four types of adult questions will be collapsed and considered as equal to the child’s production in yes/no and declarative questions. Figure 4.35 shows that wh-questions make up approximately half (52%) of the mother’s questions in the real information category. Yes/no questions also make up a considerable part (32%), along with declarative questions (10%), tag questions (2%) and negative yes/no questions (1%). Total yes/no

173 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:57

Color profile: Disabled Composite Default screen

Early Trilingualism

174 Any other 3%

+Yes/no 32%

Wh-questions 52%

Declarative 10%

Tag 2%

Yes/no 1%

Figure 4.35 Question forms used by Jenny’s mother in real information questions type formal category questions therefore comprise 45% of questions in this category. The remaining 3% is made up of other questions. Wh-questions are also the largest group (76%) in the category of known information questions (Figure 4.36). Yes/no type formal category questions comprise 23%, comprising yes/no questions, 7% declarative questions and 7% tag questions. Other questions make up the remaining 1%. The forms used in the confirmation questions category are shown in Figure 4.37. Here the picture is quite different, as questions from the yes/no type formal category make up the majority with a total of 85% of the ques+Yes/no 9%

Any other 1%

Tag 7% Declarative 7%

Wh-questions 76%

Figure 4.36 Question forms used by Jenny’s mother in known information questions

174 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:54:59

Color profile: Disabled Composite Default screen

The Findings

Wh-questions 8%

175 Any other 7%

+Yes/no 11%

-Yes/no 1%

Tag 32%

Declarative 41%

Figure 4.37 Question forms used by Jenny’s mother in confirmation questions tions; 41% are declarative questions, 32% are tag questions, 11% are yes/no questions and 1% are negative yes/no questions. Wh-questions make up only 8% and other questions the remaining 7%. The picture is different again for the distribution of question forms for clarification questions (Figure 4.38). Wh-questions make up just over onethird (37%), another third (34%) is made up of other questions, while the remaining and slightly smaller third (29%) comprises the yes/no type formal category, made up of 26% declarative questions, 2% yes/no questions and 1% tag questions.

Any other 34%

+Yes/no 2%

Declarative 26%

Tag 1%

Wh-questions 37%

Figure 4.38 Question forms used by the mother in clarification questions

175 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:02

Color profile: Disabled Composite Default screen

Early Trilingualism

176 +Yes/no 10%

Any other 14% Wh-questions 21%

-Yes/no 5%

Tag 30% Declarative 20%

Figure 4.39 Question forms used by the mother in interactive questions In interactive questions the yes/no type formal category makes up 65% of as can be seen in Figure 4.39, where 30% are tag questions, 20% are declarative questions, 10% are yes/no questions and 5% are negative yes/no questions. Wh-questions make up 21% and the remaining 14% are other questions. Overall (Figures 4.35–4.39) there is considerable variation in the forms used by Jenny’s mother from one functional category to another. She uses all the forms, but in differing amounts depending on the function of the question. Certain forms are used more with certain functions. In known information questions, for example, wh-questions clearly dominate, though these are hardly used in confirmation questions where tag questions are important, as they also are in the interactive category. Other questions, which are rarely used in general, play an important role in clarification questions. If we compare the mother’s and the child’s use of form in each function, taking into consideration the yes/no type formal category, we find the results are similar, especially in the functions of real information, known information and confirmation. This will be further discussed in the next section and in Chapter 5. Comparison of mother and child’s use of functions We can now compare the child’s and the mother’s use of question functions with regard to number of questions, length of question and type of question form used in each function. Over the course of the study, the child increased the number of questions she used in the real information, known information and interactive cate-

176 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:03

Color profile: Disabled Composite Default screen

The Findings

177

Table 4.14 Comparison of percentage use of mother (M) and child’s (C) question forms in different function categories Real information Wh- Y/N

O

Known information Wh- Y/N

O

Confirmation Wh- Y/N

O

Clarification Wh- Y/N

O

Interactive Wh- Y/N

O

C

53

46

1

71

28

1

6

89

5

14

15

71

43

56

1

M

52

43

3

76

23

1

8

85

7

37

29

34

21

65

14

Note: O = Other questions

gories. Overall, the number of confirmation questions she asks showed a very slight decrease, whilst clarification questions showed a clear decrease, possibly because, as Jenny’s language skills develop, she has less need to request clarification. Jenny’s mother, however, was found to use approximately the same number of questions in each functional category at the beginning of the study as she did at the end, except in the case of confirmation questions which show an increase in number. In relation to average length of question Jenny shows an increase in all the functional categories, whilst her mother shows a tendency towards a slight reduction in length in all except the known information category. Examination of the findings for the percentages of forms used by mother and child in functions revealed some interesting tendencies, especially when the all the yes/no type questions (yes/no questions, declarative questions, tag questions) were collapsed into one category. The results (Table 4.14) show that in the real information, known information and confirmation categories mother and child use almost identical proportions of question forms. In the real information category, wh-questions (MOT 52%, CHI 53%) are used slightly more than yes/no questions (MOT 43%, CHI 46%), while other questions (MOT 3%, CHI 1%) are rarely used. In contrast, in the known information category, wh-questions (MOT 76%, CHI 71%) are much more widely used than yes/no questions (MOT 23%, CHI 28%) and other questions (MOT 1%, CHI 1%). In the confirmation category yes/no questions (MOT 85%, CHI 89%) are dominant in comparison to wh-questions (MOT 8%, CHI 6%) and other questions (MOT 7%, CHI 5%). In the categories of clarification and interactive questions patterns in the use of certain forms by both mother and child can also be observed, but they are not as closely matched as in the other three categories. Clarification, for instance, favours use of other questions whilst yes/no questions dominate for both mother and child in the interactive category.

177 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:03

Color profile: Disabled Composite Default screen

178

Early Trilingualism

Mother and Child Feedback to Questions The system for classifying feedback was described in the previous chapter, and three types of feedback to questions were identified: (1) None. There is no immediate uptake to the question, so the questioner either repeats or reformulates the question or discards it for a new utterance, or initiates a new exchange. (2) Acceptable. The answer given is comprehensible and appropriate. (3) Unacceptable. The interlocutor takes up the turn but does not answer the question, either giving an incomprehensible or inappropriate answer or initiating a new exchange. We will now look at the numbers of replies provided to the questions asked by the child and her mother in the different functional categories in nine scripts (S1, S4, S8, S12, S16, S20, S24, S28, and S32). When the data were classified it was found that a large amount of feedback was classified as ‘none’, but it should be noted this is related to the behaviour of the questioner. In other words the person to whom the question is directed is in some way prevented from giving an answer because the questioner does not wait for a response to her own question. Olsen-Fulero and Conforti (1983) refer to this behaviour as usurpation. So, in interpreting the figures relating to feedback data we should keep in mind that ‘none’ means that the person being asked the question was unable to reply. This may be due to usurpation or a new exchange being initiated, and says more about the questioning behaviour of the interlocutor than the person referred to in the figure. Feedback classified as unacceptable, however, is related to the behaviour of the person to whom the question is asked, who has taken up the turn but is unable to respond appropriately. The child’s feedback

Figure 4.40 shows that the amount of acceptable feedback given by Jenny in response to her mother’s real information questions increases over time. From age 2;6.19 (S12) until the end of the recordings, Jenny always answers more than 55% of the questions asked. Unacceptable feedback is found across the samples in varying amounts. If we look at the questions to which feedback is ‘none’, we see that the mother uses more usurpation and exchange changing when the child is younger than in later recordings. Jenny’s feedback to known information questions can be seen in Figure 4.41. She answers a large percentage of these questions, and this increases slightly over time. In most cases, except at age 2;6.19 (S12), the percentage of questions with acceptable feedback exceeds that of unacceptable or no feedback. Unacceptable feedback is found at all ages ranging from 26% at

178 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:08

Color profile: Disabled Composite Default screen

The Findings

179

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

32

28

32

Scripts

Figure 4.40 Jenny’s feedback to real information questions

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

Scripts

Figure 4.41 Jenny’s feedback to known information questions

179 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:07

Color profile: Disabled Composite Default screen

Early Trilingualism

180

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

32

28

32

Scripts

Figure 4.42 Jenny’s feedback to confirmation questions

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

Scripts

Figure 4.43 Jenny’s feedback to clarification questions

180 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:11

Color profile: Disabled Composite Default screen

The Findings

181

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.44 Jenny’s feedback to interactive questions 2;5.5 (S9) to 3% at 3;6.18 (S32). The mother’s uses fewer questions that elicit no response in this category than in the real information category. In the category of confirmation questions (Figure 4.42), many more questions are unanswered than answered by Jenny either because she gives unacceptable feedback or because she gives no feedback owing to her mother’s usurpation or initiation of a new exchange. It may be that Jenny’s mother uses confirmation questions for purposes other than getting an answer, such as maintaining the interaction. Jenny’s feedback to her mother’s clarification questions is shown in Figure 4.43. Here the numbers of questions answered vary considerably but on the whole, and in comparison with confirmation questions, Jenny replies to a large percentage of clarification questions. This makes sense if we consider that the mother must ask clarification questions in order to clarify a misunderstanding and it is in both the child’s and the mother’s interest that feedback is acceptable. Nevertheles, the numbers of questions with no feedback show that the mother also asks clarification questions to which she does not allow the child time to give feedback. No consistent pattern can be seen in relation to the percentage of interactive questions answered with acceptable feedback, unacceptable feedback or no feedback (Figure 4.44). However there are some surprising results, for

181 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:15

Color profile: Disabled Composite Default screen

Early Trilingualism

182

instance at age 3;1.22 (S28) where the mother does not ask questions that prevent feedback. Only one other case of this is found in the data, at age 3;0.5 (S24) in clarification questions. At 2;1.26 only 17% of answers are acceptable in contrast to 78% at 3;1.22 (S28). There is evidence that Jenny gives unacceptable answers across the samples and in all five categories, with the exception of confirmation questions at ages 1;11.24 (S1), 2;4.17 (S8) and 3;0.5 (S24), clarification questions at ages 1;11.24 (S1), 3;1.22 (S28) and 3;6.18 (S32) and interactive questions at ages 2;8.3 (S16) and 2;10.2 (S20). When compared to the feedback given by her mother, Jenny has considerably greater difficulty in producing acceptable feedback. The mother’s feedback

Figure 4.45 shows that the mother answers at least 50% of Jenny’s real information questions except at age 2;8.3 (S16), when only 43% of questions receive acceptable feedback. Overall Jenny’s mother responds acceptably to a higher percentage of real information questions than Jenny does (Figure 4.40). She seems to make most effort to answer real information questions from Jenny’s age 2;10.2 (S20) onwards, whilst from Jenny’s age 2;4.17 (S8) to 2;8.3 (S16) she gives less acceptable feedback. It can also be seen that Jenny asks questions of the type that elicit no feedback – she may have learned to do this in imitation of her mother’s behaviour. Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.45 The mother’s feedback to Jenny’s real information questions

182 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:17

Color profile: Disabled Composite Default screen

The Findings

183

The mother’s feedback to Jenny’s known information questions can be seen in Figure 4.46. Apart from when Jenny is aged 3;0.5 (S24), her mother gives little feedback to the questions she asks. This suggests that Jenny has asked the questions in such a way that they do not require feedback. As the questions are about known information, there is no obligation to supply an answer, and Jenny’s mother recognises this. However if we look back at Jenny’s feedback in Figure 4.41, we see that she has given a high percentage of acceptable answers to the known information questions. This suggests that Jenny does not yet recognise the questions as low constraint. Furthermore in Figure 4.41 the relatively low percentages of questions with no feedback show that Jenny’s mother has not used usurpation to the extent she has in other categories, which leaves Jenny more leeway to respond to the questions. We would like to suggest that the mother is using known information questions as a type of scaffolding for Jenny’ s discourse. Figure 4.47 shows that the mother has given acceptable feedback to the vast majority of Jenny’s confirmation questions. The fact that questions with no feedback are found only at ages 2;8.3 (S16), 3;0.5 (S24) and 3;1.22 (S28) suggests that in the rest of her questions Jenny has not used usurpation and that the questions have been marked as requiring an answer – as

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.46 The mother’s feedback to Jenny’s known information questions

183 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:19

Color profile: Disabled Composite Default screen

Early Trilingualism

184

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.47 The mother’s feedback to Jenny’s confirmation questions

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

32

Scripts

Figure 4.48 The mother’s feedback to Jenny’s clarification questions

184 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:23

Color profile: Disabled Composite Default screen

The Findings

185

high constraint. Furthermore her mother does not produce any unacceptable feedback, which suggests that she is making an effort to be clear. This makes sense if confirmation questions are viewed as requests for checking or reiteration. A somewhat similar picture is found in Figure 4.48 where, in response to Jenny’s clarification questions, her mother supplies a high percentage of acceptable feedback with minimal unacceptable feedback. However, more evidence of no feedback can be seen here than in the confirmation category, suggesting that Jenny has more often not waited for her mother to reply before going on to her next question or utterance. Jenny asked a total of 12 interactive questions (Figure 4.49) to which her mother’s feedback was exactly divided between none and acceptable. So, except in the category of known information questions, Jenny’s mother gives a greater percentage of acceptable feedback to Jenny than Jenny does to her mother. A number of factors must be considered here. First the mother asks a far greater number of questions than the child does, and it seems logical that it is easier to attend to a limited number of questions (Jenny’s) properly. The mother’s questions seem to serve functions other than information seeking and this is shown in the amount of usurpation that takes place. It seems somehow strange that the mother should ask

Unacceptable

Acceptable

None

100%

Utterances

80%

60%

40%

20%

0%

1

4

8

12

16

20

24

28

Scripts

Figure 4.49 The mother’s feedback to Jenny’s interactive questions

185 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

32

Color profile: Disabled Composite Default screen

Early Trilingualism

186

so many questions and then not wait for an answer, jumping in with another question or initiating a new exchange. In Jenny’s case, the number of no feedback responses shows that she too is capable of asking questions without waiting for a reply – in many cases she answers her own question. Comparison of mother and child’s feedback patterns.

In order to compare the findings for mother and child feedback, the total percentage of feedback that was ‘acceptable’, ‘unacceptable’ or ‘none’ was calculated for each functional category. ‘None’ corresponds to those questions that are asked, but to which a reply is not given. In such cases the speaker continues to take the speaking turn after asking a question and this is termed ‘usurpation’ (Olsen-Fulero & Conforti, 1983). The results for child and mother are shown in Table 4.15. Jenny’s feedback results show similar patterns to those of her mother, with some interesting contrasts. Overall, Jenny’s mother provides more acceptable feedback and gives fewer unacceptable feedback responses than Jenny does. The interactive category has one of the lowest percentages for question feedback and one of the highest percentages for ‘No feedback’ for both participants, whilst feedback to clarification questions is high for both child and mother. However, confirmation questions, which are the ones Jenny answers least, are the ones most answered by her mother. Both Jenny and her mother carry out usurpation of answers (as shown in the None category) to approximately the same degree within each functional category – except in the case of confirmation questions where again results are polarised as mother usurps 52.6 % of Jenny’s answers while Jenny usurps only 17.8%. The findings on child feedback are similar to those found by OlsenFulero and Conforti (1983). Table 4.16 compares Jenny’s results with the results from the 11 subjects in the study by Olsen-Fulero and Conforti (1983) on the giving of feedback by children. The names of the functional categories used in Olsen-Fulero and Conforti’s study correspond to those Table 4.15 Comparison of mother and child’s use of feedback

C

Real information

Known information

Confirmation

Clarification

A

N

U

A

A

A

53

37

10

56.8 31.8 13.2 40.7 52.6

32

3.6

58.3 33.3

M 64.4

N

U

6.5

82

N

17.8

U

N

U

5.8

58.8 32.4

8.5

0

69.7 29.5

1

Interactive A

N

45.2 42.2 13.4 50

50

Note: A = acceptable feedback, N = no feedback, U = unacceptable feedback

186 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

U

0

Color profile: Disabled Composite Default screen

The Findings

187

Table 4.16 Comparing Jenny’s feedback (%) with that of other children

J

Real information

Known information

Confirmation

Clarification

A

N

U

A

A

A

53

37

10

OC 61.4 15.7

N

U

N

56.8 31.8 13.2 40.7 52.6 62.8 33.3

48.6 28.3

U 5.8

N

58.8 32.4 67.1 12.9

U 8.5

Interactive A

N

U

45.2 42.2 13.4 46.5 23.3

Adapted from Olsen-Fulero and Conforti (1983) Note: A = acceptable feedback, N = no feedback, U = unacceptable feedback, OC = Children from Olsen-Fulero and Conforti’s study

used in the present study. Data about unacceptable feedback is not clearly described by Olsen-Fulero and Conforti (1983), so it has not been included in our analysis. Across the results there are clear similarities to be seen although, overall, Jenny’s percentages for acceptable feedback are lower than those of the children in the Olsen-Fulero and Conforti (1983) study, and the amount of ‘No feedback’ is higher. In both sets of results the children give the greatest number of acceptable responses to clarification questions, followed by known information then real information questions. Interactive and confirmation questions are those that receive least feedback from the children. In cases of ‘No feedback’ from the child (because of usurpation of turn by the mother), there are also similarities between the results for Jenny and the children in the other study. In both cases the most usurpation takes place when mothers ask confirmation questions. The next most frequent group is interactive questions and real information questions are again in the third and middle position. The questions that are least usurped by mothers are those in the clarification and known information categories. The results suggest that clarification and known information questions have an important function since they are the categories most answered by children and least usurped by mothers. We suggest that clarification is important if conversation not to be impeded and break down, and known information questions serve the purpose of initiating and maintaining discourse topics. It does not seem so important to give feedback to questions with the real information function as these consistently remain in third place for amounts of acceptable or no feedback given. However, real information is more important than the confirmation and interactive functions, which are answered less by the children and usurped more by the mothers. Results for the type of feedback given to children’s questions are not available in the study by Olsen-Fulero and Conforti (1983).

187 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

Early Trilingualism

188

Table 4.17 Total percentage of feedback type Mother Jenny OC

Acceptable

None

Unacceptable

59.8

37.5

2.2

50

39

10.1

57.28

18.44

-

Finally, the total amounts of acceptable, unacceptable and ‘no feedback’ were calculated for Jenny, her mother and the 11 children from OlsenFulero and Conforti’s (1983) study. The results in Table 4.17 show that acceptable feedback given to questions lies at 50–60%. Even taking into account those answers that are unacceptable, in which the percentages are relatively low, this finding suggests that questions serve a purpose other than simply to be answered and that the explanation for this lies in the use of usurped questions.

Cross-linguistic Influence in the Child’s General English The rest of this chapter looks at the two sets of analyses that were carried out in order to find out the extent to which there is evidence of cross-linguistic influence in the child’s production. With regard to Jenny’s general English the analysis examined lexical mixing, whilst a syntactic analysis was carried out for the questions. Evidence of cross-linguistic influence in the corpus was sought by extracting all examples of transfer from Spanish and Basque. Every instance of transfer was treated as an individual case, even if the same words were repeated in contingent utterances, contingent exchanges or elsewhere in the recording, During the study, Jenny used some idiosyncratic words that are clearly not English and which show characteristics of Spanish, Basque or both. Dewaele (1998: 473) calls such words in adult speech lexical inventions, but in the present study they are called invented words since they may be due to developmental factors as well as cross-linguistic ones. These words are classified as belonging to the language they most resemble. One such word used by the child is pamelo [handkerchief], which closely resembles the Spanish pañuelo (the Basque word is zapi). Words that cannot be identified as such in any of the three languages have not been included. Another category that is not counted as language mixing is that of words from either Spanish or Basque or both that are used across the three languages in relation to the home context. These are termed homewords and

188 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

The Findings

189

are not regarded as transfer since they are used by the mother in the home context as if they were English words. • Homewords from Basque: momo [monster], lolo [sleep] and osaba [uncle]. • Homewords from Spanish: colonia [cologne], bañito [bath] and pegatina [sticker]. • Homewords that are Basque/Spanish or cognates: motxila/mochila [rucksack], txankleta/chancleta [flip-flop], mantía [blanket] and lolitos [sleep] Six categories of transfer were identified: B complete utterances in Basque; S complete utterances in Spanish; EB mixed utterances containing English and Basque; BS mixed utterances containing Basque and Spanish; EBS mixed utterances containing English, Basque and Spanish, Table 4.18 shows instances of transfer in the six categories, and the total number of transfers per recording. The percentage of transfer clearly reduces over time, and only 1.71% of utterances are mixed. There were 24 Basque-only utterances compared with 35 Spanish-only utterances. When mixing occurred with English in an utterance there were also slightly more mixes in Spanish (48 mixes) than in Basque (42 mixes). Complete utterances in Basque (B) These are usually multi-word utterances, though there are occasional one-word utterances. (S1.1;11.24) CHI: Agur. (Goodbye) (spoken to a doll) MOT: CHI:

MOT: MOT:

What’s it called that thing? Martinsaltua (grasshopper + determiner) (‘A grasshopper’) Mar chitatua? (asking for clarification) I think you’re calling it a martinsaltua aren’t you?

There are 24 transfers from Basque at utterance level.

189 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

(S9.2;5.5)

Color profile: Disabled Composite Default screen

Early Trilingualism

190

Table 4.18 Total transfers per recording in all six categories Script

Age

MLU

Utterances per script

B

S

1

1;11.24

1.526

428

1

8

2

2;0.6

1.824

294

1

7

3

2;1.2

1.917

208

4

4

2;1.26

2.204

273

5

2;2.24

1.767

241

6

2;3.14

2.168

344

1

7

2;4.2

2.285

256

5

1

8

2;4.17

2.322

204

3

3

2

2

9

2;5.5

2.564

402

2

1

6

4

10

2;5.19

2.724

353

1

1

1

11

2;6.3

2.993

366

2

12

2;6.19

2.698

430

1

13

2;7.1

2.602

251

14

2;7.9

2.815

341

15

2;7.19

2.732

294

16

2;8.3

2.803

379

17

2;8.19

2.976

332

18

2;9.5

3.164

253

19

2;9.22

2.835

253

20

2;10.2

2.672

273

21

2;10.14

3.562

324

22

2;10.30

2.695

268

23

2;11.17

3.069

226

24

3;0.5

2.889

272

25

3;0.19

3.243

244

26

3;0.30

3.332

274

27

3;1.9

4.425

298

28

3;1.22

3.045

349

29

3;3.8

3.549

302

30

3;4.7

3.139

282

1

31

3;5.18

2.985

308

1

32

3;6.18

2.957

206

Total

9,528

3

1

ES

BS

5

1

EBS

2

1

3.5

11

3.74

5

2.4

2

5

1

13

4.76

1

4

1

14

5.8

1

1 1 1

1

3.23

5

1.42

7

1.91 0.93

1

0.39

1

0.29

1

0.34

1

0.26

1

5

1.5

2

8

3.16

2

0.79

1

0.37

2

0.62

5

1.86

2

0.88

1

0.37

2

1 1

1 1

1

2

1

1

1

1 1 2

1

4

1.64

2

1

3

1.09

1

4

6

2.01

2

3

5

1.43

3

0.99

2

0.71

2

3

0.97

1

1

0.49

163

1.71

3

35

5.39

13

4

1

1

11

3

4

1

2.73

1

1

1

0.3

7

2

1

1

1

1 1

24

15

5

1

2

Total % transfers Transfers per to utterscript ances

5

190 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

EB

42

1

48

5

9

Color profile: Disabled Composite Default screen

The Findings

191

Complete utterances in Spanish (S) As in Basque, one-word utterances and multi-word utterances were counted. In the following example, Jenny child wants the mother to look at the video microphone: MOT: Oh what’s that? (S4. 2;1.26) CHI: Allí. (‘There.’) MOT: Eh what is it? CHI: Allí. MOT: What is it? CHI: A bedi (‘Birdy’ – the child’s name for the microphone) In the next example, Jenny wants her mother to put a hat on her. CHI: No up ## here. (points to hat) S2. 2;0.6) CHI: Pónmelo pónmelo. (Put me it put me it.) (‘Put it on me, put it on me.’) MOT: Oh look at this. MOT: What a lovely hat. MOT: Put this on love. MOT: Here you are CHI: On hat. In the two examples above the child has used Spanish pragmatically to fulfil the function of directing attention. Other expressions from Spanish, often two-word utterances, were also found to have similar emphatic uses. A total of 35 transfers from Spanish were identified at utterance level. Mixed utterances containing English and Basque (EB) The data included 42 mixed utterances containing English and Basque. In the following exchange, where a doll has fallen and ‘hurt’ herself, Jenny mixes as a relief strategy for a word she does not know in English. MOT: Oh dear what’s happened? (S26. 3;0.30) CHI: Put it got odola. (Put it got blood + article) (‘She’s bleeding.’) MOT: Oh dear. MOT: What’s happened to dolly? CHI: ’S got the odola in my hair. (sic) (has got the blood + art. in her hair) (‘She’s got blood in her hair.’)

191 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

Early Trilingualism

192

The child inserts odol [blood] but with the determiner suffix –a, which is the norm in Basque child language. However in the second instance she also uses the determiner the in English. She omits the subject she; this is typical of Basque, which allows for a null subject. Mixed utterances containing English and Spanish (ES) Altogether 48 mixed utterances containing English and Spanish were identified. In the following example Jenny is playing at tea parties. CHI: Bigears is wanting a cup of tea. (S21. 2;10.14) MOT: Does he want a cup of tea? CHI: I give him that cup of tea and you too. CHI: He’s got it babero. (He’s got it bib.) (‘He’s wearing a bib.’) MOT: Yeah he’s got his bib on hasn’t he? Mixed utterances containing Basque and Spanish (BS) This category contains 5 possible Basque and Spanish mixes and demonstrates some of the complexity of teasing apart the different languages in multilingual mixes where developmental factors are also present. Several types of mixes have been identified. The first type is a complete utterance consisting of Basque and Spanish without English. In the example below the mix is preceded by an utterance in Basque with an incomprehensible word (shown as xxx) that may be in any of the languages. The child may mix with Spanish to draw attention to the fact (see emphatic uses of Spanish above) that a water bottle is broken. (S8.2;4.17) CHI: BABBLE Apurtuta dago xxx. (BABBLE broken is xxx.) (‘xxx is broken’) CHI: Mira (Spanish) apurtu (Basque). (look broke-n) (‘Look it’s broken.’) Some words that are used in both Basque and Spanish are identical except for their written forms e.g. txuleta/chuleta [steak], dutxa/ducha [shower], pailaso/payaso [clown] and saltxitxa/salchicha [sausage]. These words have been classified as Basque/Spanish when they are used in isolation since it is not clear which is the source language. When they appear as mixes in an English utterance they are classed as English/Basque/Spanish (EBS).

192 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

The Findings

193

Mixed utterances containing all three languages (EBS) There are nine utterances that could be argued to contain all three languages. Several types of mix have been identified, again bearing in mind the difficulty of identifying the language source of cognates. In the following example saltxitxa/salchicha [sausage] is both Spanish and Basque, and could come from either language. Hau da is Basque and Mummy is English, so at least two (Basque and English) and possibly all three languages are present in this utterance in which Jenny prepares breakfast for her dolls. (S4. 2;1;24) CHI: ### Mummy hau da (B) saltxitxa/salchicha (BS) (Mummy that is sausage) (‘Mummy here’s a sausage.’) CHI: Saltxitxa/salchicha (BS). MOT: Oh are you going to give them a saltxitxa/salchicha(BS)? MOT: A sausage. MOT: Say sausage. CHI: Sausage saltxitxa/salchicha (BS). Unusual word orders in English are evidence of syntactic mixing in the next example: (S5. 2;2.24) CHI: Pamelo one. (Handkerchief one) (‘One handkerchief’) Pamelo (handkerchief) is an invented word similar to the word for handkerchief (pañuelo) in Spanish. In this utterance it is used with the English one, but the word order is transferred from Basque where it is obligatory for the number one (bat) to come after the noun. This is not possible in Spanish or English where number must precede the noun. There is evidence of possible trilingual mixing at a morphological level in the following example. As in previous examples, a cognate is involved and in this case it is the same word in all three languages sandal (E) and sandalia (B/S). MOT: Which shoes are you going to wear? (S7. 2;4.2) CHI: Sandalak. (sandal (with stress on first syllable, as in English)+ plur. (B)) CHI: Sandalak. MOT: You want to wear your sandals? CHI: Yes. The child’s English pronunciation of sandal suggests that this is the

193 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

Early Trilingualism

194

source language and that the Basque plural suffix has been added. If this is the case, the utterance is a Basque/English mix, but the presence of Spanish cannot be ruled out. Lastly we find an instance where an English verb ending –ing is added to a root made up of a Basque/Spanish cognate, in this case dutxa (B)/ducha (S) meaning shower in English. Jenny is looking at an old ring she used to keep herself upright in the bath. CHI: Here what’s for doing? (S30. 3;4.7) MOT: That doesn’t make a squeaky noise it just rattles. CHI: Mummy is this for dutx-ing me? MOT: That was yours when you were a little baby. To sum up, if we look at those categories where transfer is clearly identifiable as proceeding from Basque or Spanish (categories B, S, EB, and ES), we see that there are more instances of transfer from Spanish (83) than from Basque (66) The majority of mixes (13 out of 14 instances) involving Basque and Spanish together (BS), or Basque and Spanish together with English (EBS), contain words that are cognates in Basque and Spanish that result from the close contact that exists between the two languages. Otherwise there is some ambiguity about the source language of the mix, which makes it impossible to identify the linguistic source of the Basque and Spanish mixes. If this is taken into account we can speculate that: (1) The three languages do not co-occur in one utterance. (2) The child mixes the cognates in a similar way to the mixes she makes in categories EB and ES. (3) English can occur with either Basque or Spanish depending on the source of the cognate. More work is needed to identify the sources in such cases. In order to ascertain if trilingual mixing can take place, it is necessary to work with three languages in which there are no cognates. Cross-linguistic equivalents During the analysis we noted that the child sometimes used a Basque or Spanish word in the same context as its English equivalent. This might be in the same utterance, the same exchange or in previous or subsequent exchanges or utterances. Occasionally the same word was found to occur elsewhere in another language in the same recording. Of the 163 cases of transfer noted in the corpus, 20 (12.3%) have cross-linguistic equivalents. Two types were identified:

194 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

The Findings

195

• Cases where the mother uses strategies to elicit the English word when Jenny has used a Basque or Spanish equivalent. • Cases where the child, in an unprompted switch, uses equivalents from Basque or Spanish in the same environment as the English word. Mother’s strategies

We identified 11 cases in which the mother used one of the five parental discourse strategies described by Lanza (1992, 1997). In the following examples the strategy from Lanza’s typology used by the parent is shown in parentheses. In the first example Jenny finds a ball in her cot. Her mother repeats the word pelota (S) and then models the English equivalent ball (Repetition, with modelling). The child corrects herself by saying ball but repeats the word pelota (S) in the same utterance. However in the following utterance the mother code switches and uses the word pelota (S) herself and does so again four utterances later. Finally the child also introduces the equivalent pilota (B) whilst continuing to use pelota (S). CHI: Pelota (S1. 1;11.23) MOT: Pelota a ball yes. (Repetition, with modelling) CHI: Ball ## pelota. MOT: Pelota. (Code switch) MOT: What else? MOT: What else have you got? CHI: Uh pelota. MOT: Pelota # what else has Jenny got? (Code-switch) CHI: Pilota # pelota. CHI: (BABBLE) pelota. In the next example Jenny’s utterances after using kapelo (B) contain hat, after her mother has repeated the Basque word and modelled the English word in a question. (Repetition, with expressed guess). (S3. 2;1.2) CHI: Mummy hello I got it a here kapelo. CHI: Here yes # CHI: Kapelo. MOT: Kapela a hat is it? (Repetition, with Expressed Guess) MOT: Come on. CHI: Hat! CHI: Look hat! In the next example Jenny shows that she knows the difference between shoes and sandals, although she does not know the word in English, and

195 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

Early Trilingualism

196

uses a word that is a mixture of English, Basque and Spanish. Her mother uses the Expressed Guess to provide her with the word in English. MOT: What are you doing now? (S7. 2;4.2) CHI: A blue one. CHI: Blue one shoes on. CHI: (BABBLE) MOT: No don’t take it off. MOT: What are you going to do now? CHI: Blue one. MOT: Are you getting different shoes? CHI: Eh? MOT: Which shoes are you going to wear? CHI: Sandalak. CHI: Sandalak. MOT: You want to wear your sandals? (Expressed Guess) CHI: Yes. The mother points out the mix by repeating with questioning in the next example, but she does not provide the word in English. Five utterances after using pianoa (B), Jenny provides the word piano in English herself. MOT: Eh what’s that? (S10. 2;5.19) CHI: Pianoa. MOT: Pianoa? (Repetition, with questioning) CHI: Yes. CHI: Mummy this one is dolly. CHI: This one. CHI: This one is piano this one. MOT: This one is piano. Here the child utterance following olloa (B) contains the word chicken, after her mother models the English word through an Expressed Guess: CHI: What’s that? (S11. 2;6.3) MOT: You know what that is. MOT: What is it? CHI: Olloa. MOT: A chicken # a chicken. (Expressed Guess) CHI: A chicken. CHI: Hello chicken. Same-script cross-linguistic equivalents

Jenny occasionally switched languages without prompting from her

196 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

The Findings

197

mother, and the equivalents often appeared in the same context. In the next example, Jenny says elephant and elefantea (B) in the same utterance. She implies that the English noun is correct and the Basque one is incorrect: CHI: What’s that? (S7. 2;4.2) MOT: Show me. MOT: An elephant is it? CHI: Yes an elephant, no he a elefantea. In the following example, four utterances before saying cinco (S) Jenny uses the word five in English: MOT: How many? (S8. 2;4.17) CHI: One. MOT: And this? CHI: Hmm three # four five. MOT: What are these? CHI: Ei ei rabbit. MOT: How many? CHI: Rabbit rabbit rabbit rabbit rabbit rabbit. MOT: Rabbit good girl. MOT: And how many rabbits? CHI: Erm. MOT: Right. CHI: Cinco. CHI: Look look. In the following exchange, it can be seen that the child uses both the English word here, then later its Basque equivalent hemen. The child makes the language adjustment automatically. MOT: Who’s got a headache? (S9. 2;5.5) CHI: A rabbit. CHI: Hey du deim here # here. MOT: Oh what have you got? MOT: Show me. CHI: (BABBLE) CHI: I’m a headache # headache. MOT: Headache? CHI: Yes. MOT: Headache? CHI: Yes hemen [here] look. In the next example the utterance before busti (B) contains the word wet. Jenny makes the language adjustment automatically.

197 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

Early Trilingualism

198

CHI: CHI: CHI: CHI: CHI:

Oh Jenny chair’s wet. Busti [wet]. Upsi. Oh this. Mum!

(S28. 3;1.22)

Other points of interest In the corpus, Jenny applies English morphology to a Basque/Spanish root to create the mix dutx-ing. There are two other cases of foreignisation in the data, both involving Basque. In the first one Jenny comes across a toy windmill: CHI: What’s this doing here? (S26. 3;0.30) CHI: Mum what’s this? CHI: Mum you got it? (blows windmill) CHI: I’m putz-ing Mum. (‘I’m blowing Mum.’) MOT: xxx. In the next example, Jenny is fitting shaped toys into a box. MOT: What are you doing, Jenny? (S17. 2;8.19) CHI: I’m putting in a ### I’m putting a. CHI: Don’t txist-en the toys. MOT: Twisting the toys are you? Jenny puts an English verb ending on an idiosyncratic verb root of Basque appearance, but follows the phonetic rule of English verb endings on verbs with /st/ as in listen or christen. The results regarding cross-linguistic influence show that mixes from Spanish occurred more frequently than from Basque. The data described above show that most mixes were borrowings of content words from the other languages without morphological adaptation, though there a few cases of foreignising and morphological transfer, and all of them involve Basque. It is worth noting that some of the transfers are preceded by a pause (indicated by ##) denoting hesitation in word choice for some reason.

Cross-linguistic Influence in Questions In Chapter 3 the following areas were identified for examination in the child’s questions as possible sources of cross-linguistic influence from Spanish and Basque.

198 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

The Findings

199

(A) Basque and Spanish yes/no questions are formed by adding rising intonation to a declarative utterance. (B) Both languages permit a null subject and Basque permits a null lexical object.5 (C) In child Basque wh- is positioned late in the question prior to the development of fronting. Spanish has wh-fronting the same as English.6 (D) Both Spanish and Basque allow the subject to be on the right. During the course of the analysis of the question data, other areas appeared to be anomalous to the normal acquisition of questions by monolinguals, taking into consideration the idiosyncratic nature of question acquisition noted in Cazden (1970: 208). These will be also be discussed in relation to cross-linguistic influence and are as follows: (E) (F) (G) (H) (I) (J) (K)

dependence on is + 0/it/this/that; dependence on wh- + it/this/that; subject repetition; word order; wh-confusion; hybrids; translation.

These 11 areas will now be considered with examples from within the general context of Jenny’s question data as a whole. Then five samples of Jenny’s question data taken from one of each of Brown’s five (1973) stages will be contrasted with five samples from a monolingual child Sarah (Brown, 1973), and two samples from a monolingual child Sophie (Fletcher, 1985). General analysis of cross-linguistic influence in the child’s questions Possible cases of cross-linguistic influence from Basque and Spanish can be said to exist in Jenny’s questions on the basis of the following evidence: (A) Pervasive use of declarative questions

Declarative questions make up 36.7% of the total questions produced by Jenny during the period under study, making them the second most frequently used group of questions after wh-questions (44.2%). Although the percentage of declarative questions clearly decreases over time (see Figure 4.4) as the percentage of yes/no and wh-questions grows, even after the age of 3 they still comprise between 20–40% of questions (Scripts 27–30). This is not found in monolingual data, where the use of declarative questions disappears as the use of yes/no questions develops. We have seen that

199 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

Early Trilingualism

200

the emergence of Jenny’s yes/no questions follows a normal route and is even somewhat in advance of the periods described for monolinguals by Cazden (1970) and Klima and Bellugi (1966). However Jenny is well behind monolinguals in the numbers of yes/no questions she produces. Since the use of intonation on declarative utterances to form a yes/no question is common to both Basque and Spanish, it is likely that Jenny’s pervasive use of declarative questions is the result of cross-linguistic influence. (B) Null subject

In the following example, neither the auxiliary ’s nor the subject is present. Jenny repeats the question a number of times to herself as she climbs out of her cot: CHI: Eh what doing here? (S9. 2;5.5) CHI: What doing here? CHI: What doing here? CHI: What doing here? MOT: What’s doing here? Klima and Bellugi (1966) report wh- + NP + doing as one of the basic whroutines for Period A wh-questions. However, the routine does not appear in Jenny’s early questions until What the teddy doing? at age 2;6.3 (S11). As both Basque and Spanish permit a null subject, the absence of a subject pronoun in Jenny’s early questions may signify cross-linguistic influence. (C) Late wh-fronting

Wh-fronting (putting the wh-word at the beginning of the sentence) is required by all three of Jenny’s languages, and in most of her wh-questions Jenny correctly fronts wh- to sentence initial position. However Barreña (1994) has noted that in child Basque (and occasionally in adult Basque) the wh-element may be placed in final position, as in • Bere aita non? (His father where?) (Where’s his father?’)

(2; 0.26)

The wh-element may precede a verb that is in final position: • Nik hau nora bota? (I this where to throw?) (‘Where (shall) I throw this?’)

(2;1.19)

(D) Subject on the right

As seen in B. above, Jenny sometimes omits the subject in her questions and did not produce what + NP + doing until age 2;6.3 (S11) when she asked

200 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

The Findings

201

What the teddy doing? However, on three occasions prior to this she used a related question form what + doing + NP, not found in monolingual reports. For example: CHI: What’s doing Panda? (S9. 2;5.5) CHI: Mummy what’s doing? Since Spanish and Basque allow the subject of a question to be on the right, this question form could be the result of cross-linguistic influence. (E) Is + 0/it/this/that dependence

Many of Jenny’s yes/no questions from 2; 6.19 (S12) onwards are introduced by the question phrase is + 0/it/this/that which results in both fullyformed and non-fully-formed questions. This raises two issues of possible significance in relation to cross-linguistic influence. Firstly the use of is + 0: CHI: Is putting in here? (S20. 2;10.2) This will not produce any fully-formed questions, since the dummy it or some other subject must be inserted to make the question fully-formed. The omission of the subject may be because Spanish and Basque permit a null subject as outlined in B. above. Also, the question phrase is + 0/it/this/that appears exclusively with is at the front. Both Spanish and Basque have two separate forms of the verb to be, one of which is permanent (ser, izan) and the other temporary or expressing location (estar, egon), and neither of these requires a subject pronoun. As English has just one form, in the third person is, it is virtually outnumbered four to one by the other two languages since Basque has da and dago and Spanish has es and está – all of which may be in the initial position of a yes/no question. A possible explanation for the pervasive use of is + 0/it/this/that is that Jenny may realise that is can be used for da, dago, es and está and may over-generalise. Sometimes she does produce fully–formed questions: CHI: Is this a dinosaur? (S32. 3;6.19) MOT: That’s right. However, overuse of the is + 0/it/this/that phrase also results in ill-formed questions, as in the following example: CHI: Is that my friend is it yet? (S31. 3;5.18) MOT: Whose friend’s that? CHI: Sinead. (her best friend) (F) Wh-(’s) + it/this/that dependence

Jenny depends on wh-(’s) + it/this/that in much the same way as she uses

201 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:25

Color profile: Disabled Composite Default screen

Early Trilingualism

202

the is + 0/it/this/that phrase discussed in section E. In the following example she is looking through her toy box: CHI: What’s this do it here? (S25. 3;0.19) MOT: That’s your hair drier isn’t it? However, Jenny’s dependence on wh-(´s) + it/this/that is unlikely to be directly related to the verb to be in the three languages as we have just suggested, as ’s is not in initial position and appears only as contraction attached to wh-. Nevertheless, we would like to speculate that the two phrases are related and somehow an indirect result of cross-linguistic influence since they are not common in monolingual data, both phrases have a similar structure and they seem to serve the common purpose of a strategy. Our findings for yes/no and wh-questions suggest that Jenny adopted is + 0/it/this/that and wh-(´s) + it/this/that to act as a basis on which to form questions until she had the linguistic resources to attempt inversion and auxiliary insertion. (G) Subject repetition

In the course of analysing the what questions, we saw that Jenny sometimes used non-standard word orders in her questions often involving two subjects, one as a full noun and the other as a pronoun. The pronoun tended to precede the full noun reflecting a word order that is common to both Basque and Spanish, although in both languages there is a null subject on the verb, as in the following example: CHI: What’s it doing this dzirrup? (giraffe) (S19. 2;9.22) MOT: Show me all your animals ## giraffe right. Note how Jenny realises that in English she must insert the subject pronoun, which is not necessary in Spanish or Basque. (H) Word order

Non-standard word orders may also indicate cross-linguistic influence. In the following example the preposition precedes the verb, which is typical of Basque. Jenny is talking to herself as she ‘repairs’ the bedroom with a toy tool kit: CHI: Mum this I wan(t)it doing here. (S26. 3;0.30) CHI: Oh how is the under the door makes? CHI: Oh yes. (I) Wh-confusion

There are a number of instances in which either it is not clear which whword Jenny is saying or she uses a wh-word inappropriately. To my knowl-

202 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

The Findings

203

edge this has not been reported in other monolingual or bilingual studies of questions and may be the result of cross-linguistic influence in a trilingual context. In the following example it is not clear whether Jenny is saying what or where as part of a reformulation. The question word sounds like a mixture of the two, wha. The verb going indicates that she probably means where. CHI: Mummy wha (where) is going? (S12. 2;6.19) CHI: Mummy wha (what) is doing? CHI: Mummy wha (what) is doing? MOT: Who? (J) Hybrids

The complexity of some questions makes it difficult to tease apart the possible reasons for their structure. Developmentally they might contain performance errors or be the result of competence factors or simply be idiosyncratic. On the other hand they may be the result of cross-linguistic interaction between the child’s different languages or even the result of a mixture of any or all of these factors. In the following example, Jenny and her mother are looking at some pictures provided to help with a jigsaw puzzle: CHI: What’s this? (S22. 2;10.30) CHI: The papers? CHI: Ooh yes. CHI: This is the papers. MOT: And what’s on the papers? CHI: What my a putting? (K) Translation

In contrast to hybrids there are other questions that are not English in structure and which clearly follow the syntactic rules of one of the other languages. In the next example mother and daughter are looking at farmyard animals and making their sounds: CHI: How do a donkey? (S28. 3;1.22) (How does a donkey go?) MOT: Ee aw ee aw. Jenny’s question shows transfer from both Basque and Spanish syntax, as can be seen from the following outlines of how it would be said in adult Basque and Spanish: Nola egiten du astoak? (BASQUE) how do + aux a donkey?

203 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Early Trilingualism

204

Cómo How

hace does

el burro the donkey?

(SPANISH)

Comparison with monolingual subjects In order to examine cross-linguistic influence at a syntactic level in Jenny’s questions, samples from 5 of her scripts were compared with 5 scripts from the monolingual child Sarah (Brown, 1973) and two from Sophie (Fletcher, 1985). The scripts were selected to correspond to each of Brown’s (1973) five stages and to a similar MLU in each child. This information can be seen in Table 4.19. The scripts for Sarah were taken from the CHILDES (MacWhinney, 1995) database. In Sophie’s case, the two scripts that corresponded to Jenny’s MLU were taken from Fletcher (1985). General comparison of Jenny and two monolingual subjects

Table 4.20 shows the percentages of question types used by each child in the sample examined at each stage. At Brown’s Stage I both Sarah and Jenny use more declarative questions than other types, and neither produces any yes/no questions. Both are producing intonation questions and wh-questions, though the latter consist primarily of whaddat questions (Cazden, 1970). At Brown’s Stage II Sarah and Jenny have both increased the amount of wh-questions they use and decreased the amount of declarative questions. Intonation questions have also decreased. Jenny produces 4 yes/no questions and 29 other questions, whilst Sarah does not produce any questions in these categories. Stage III does not indicate any clear trends in relation to other stages, except in the case of Sarah where wh-questions and other questions increase dramatically, whilst declarative questions decrease. Sophie also uses a large amount of wh-questions (87%), while declaratives make up the rest of her questions (13%). Table 4.19 Age and MLU of the subjects Sample

Brown’s stage

Jenny

Sarah

Script

Age

MLU

Age

MLU

1

I

S1

1;11.23

1.526

2;3.19

1.573

2

II

S6

2; 3.13

2.168

2;10.11

2.055

3

III

S9

2;5.4

2.564

3;3.7

2.547

4

IV

S23

2;10.12

3.069

3;6.23

3.061

5

V

S29

3;3.7

3;5.49

4;0.28

3.701

204 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Sophie Age

MLU

2;4.28

2.53

3;0.4

3.82

Color profile: Disabled Composite Default screen

The Findings

205

Table 4.20 Percentage of questions per type used at each stage Brown’s stage

Yes/No Je

So

WhSa

Je

So

Intonation Sa

Je

So

Declarative

Sa

Je

So

Other

Sa

Je

So

Sa

I

0

0

20

5

6

14

68

79

6

1

II

4

0

36

42

1.3

1.3

30

54

29

0

III

0

0

34

66

5.2

0

22

5

39

IV

4.5

16

55

22

0

0

40

55

0

V

2.5

17

43

37

2.5

0

42

18

10

0

20

87

74

0

0

13

5.1

0

29 5.5

0

30

Note: Je = Jenny, So = Sophie, Sa = Sarah

By Stage IV Sarah is showing a trend towards an increasing amount of yes/no questions, which by Stage V are well in place for both her and Sophie. In Sophie’s case, 20% yes/no questions have now appeared where there were previously none. Sophie’s wh-questions and declarative questions have both dropped accordingly. By Stage V all the girls produce more wh-questions and yes/no questions than at the beginning of the period under study, whilst both declarative and intonation questions have reduced, though Jenny continues to use them. Sarah has stopped using intonation questions altogether, while Jenny is still using them at Stage V. Sarah produces many more other questions than Jenny, but the questions that she produces are of the declarative type with huh. For example: SAR: Playground # take the whole thing # huh? (SAR 89. 4; 0.28) This type of question is more typical of the American English that Sarah hears than the British variety that Jenny is exposed to. Jenny’s questions comprise mainly wh- and declarative questions, whereas Sarah frequently uses a wider variety of types. Analysis of potential areas of cross-linguistic influence in questions The selected samples of questions from Jenny, Sarah and Sophie were subsequently examined for evidence of 10 of the 11 categories identified above. Translation was not included, as the monolinguals were not expected to show this. If the monolingual girls produced a question that could be considered to fit any of the categories, this was taken as evidence of the form being developmental rather than the result of cross-linguistic influence. Table 4.21 shows the overall results for the stages. Where a child has

205 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Early Trilingualism

206

Table 4.21 Summary of findings for Brown’s stages 1–V for comparison between a trilingual child and monolingual children Jenny

Sarah

Sophie

A declaratives

These do not reduce to the same extent in Jenny as they do in the other subjects (see Table 4.20)

B null-subject

What are saying?

What doing?

0

C late wh-

It’s what that?

My hat what I’m doing?

0

D subject on right

What’s doing Jenny? What’s doing Panda?

E is + 0/it/this/that dependence

Is this is green?

F wh-(‘s) + it/this/that What’s that a fish? dependence Where is it is things?

0 0 What’s a you?

Where do this bit?

0 0

G subject repetition

What he doing Jon?

0

0

H word order

Evidence in other scripts

0

Where’s those two (a)nother things broke?

I

wh-confusion

Evidence in other

0

0

J

hybrids

Evidence in other scripts

What my doing? This are you can make an elephant?

0

used a question from group A–J, the question has been written in the column under her name next to the corresponding group. If a 0 appears, this means that no examples of this type were found in the data for that child. In Jenny’s case, examples from all 10 groups have, of course, been identified since this was the basis for the analysis. However in the particular scripts used for comparison with monolingual children, occasionally no examples were produced by Jenny, and such cases have been noted (e.g. H, I and J). At Stage I there was no evidence of the 10 categories and as language developed examples of various forms appeared in the questions of all three girls. Jenny’s scripts show more evidence of non-fully-formed questions than those of Sarah and Sophie, which may be a result of factors including idiosyncrasy and cross-linguistic influence. However only 4 of the 10 potential cross-linguistic categories do not appear at all in the monolingual data (these appear as shaded sections on Table 4.21), confirming them as likely to be the result of cross-linguistic influence. They are:

206 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

The Findings

207

(A) numbers of declarative questions; (E) is + 0/it/this /that dependence; (G) subject repetition; (I) wh-confusion. In sum, with regard to potential areas of cross-linguistic influence at a syntactic level in questions, the monolingual subjects in fact produce nonfully-formed questions of the following types also produced by Jenny: (B) null subject; (C) late wh-; (D) subject on right; (F) wh- (’s) + it/this/that dependence; (H) word order; (J) hybrids. Therefore such questions must be considered to be the result of developmental and idiosyncratic factors. Only types A, E, G, and I are identified as being cross-linguistic in nature and in each case the source of the influence is related to syntactic aspects of question formation common to both Basque and Spanish. This seems to reflect the findings on lexical mixing, where it was difficult to tease apart Spanish and Basque in utterances containing mixes of all three languages. Notes 1. Deixis refers to the distinction in location between ‘here, this, these’ and ‘there, that and those’. Here it is of interest that Jenny correctly uses ‘that’ and ‘there’ together. 2. Bellugi’s Period A corresponds to a MLU of 2.0–2.5, when Jenny was aged 1;11.24–2;4.17 (scripts 1–8). 3. Bellugi’s Period B corresponds to a MLU of 2.5–3.5, when Jenny was aged 2;5.5– 3;0.5 (scripts 9–24). 4. Bellugi’s Period C corresponds to a MLU of 3.5–4.5, when Jenny was aged 3;0.19– 3;6.18 (scripts 25–32). 5. In Basque and Spanish it is not necessary to specify the subject of a sentence with a pronoun and in Basque it is not necessary to specify the object pronoun of a sentence, as this is included in the verb form. • Basque: non dago (mutikoa)? where is (+ null subject) he (the boy)? • Spanish: dónde está (el chico)? where is (+ null subject) he (the boy)? 6. In other words, wh-words are positioned at the front (or beginning) of the question.

207 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Chapter 5

Interpretation of the Findings General Language Development The subject of this study, Jenny, had only limited exposure to English in the home and had never been in an English-speaking country. It was therefore of interest to measure her proficiency in English to see: • how her general acquisition of English took place as compared to other children in monolingual or bilingual environments with regards to mean length of utterance (MLU) and type token ratio (TTR)? Mean length of utterance Jenny’s mean length of utterance (MLU) increased steadily with age. At age 1;11.24 she had an MLU of 1.526 and was at Brown’s (1973) Stage I. Her highest MLU was 4.425 at age 3;1.9 at Stage V+. Jenny’s MLU in English at different ages compares well to that of three monolingual subjects. With regard to English-speaking bilingual subjects, Jenny’s MLU is slightly higher than that of Andreu (Pérez-Vidal, 1995) and lower than Kate (De Houwer, 1990). MLU in Jenny’s other languages was found to be similar to that of bilingual Basque/Spanish children of the same age. Type token ratio Jenny’s type token ratio (TTR) also increased during the period under study, from 0.279 at 1;11.24 to 0.427 at 3;6.18. A comparison of TTR was also carried out on the same monolingual and bilingual subjects where it was available. Again, Jenny showed a similar development to other children. On the basis of these results it seems that during the period under study Jenny’s proficiency in English was like that of monolingual and bilingual children at comparable ages. These findings are consistent with those of De Houwer (1990) in relation to the separate development of languages. Separate development is closely related to the quality of input in the linguistic microcosm (De Houwer, 1990, 1995) and caregiver discourse strategies (Döpke, 1992; Lanza, 1992, 1997). In Jenny’s case, the languages were clearly demarked into monolingual contexts (Lanza, 1992, 1997) relating to her mother (English), her father (Basque) and her caregiver (Spanish), and rules for their use were strictly adhered to. We believe that these factors

208 208 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Interpretation of the Findings

209

enabled Jenny to develop her English to an extent where her development, in terms of MLU and TTR, is comparable to that of monolinguals.

Formal Development of Interrogative Behaviour The second issue addressed in the study was: • the way the development of the child’s question form took place in terms of length of question, number of questions and the type of question produced. Length of questions The results on mean length of question show that there was steady growth to a peak average length of around 3.5 words per question at age 2;11.7. Thus partially reflects the results of Davis (1932) who found that after a continuous increase, the average length of question from age three until adulthood stabilises at 4–6 words per question. After this point Jenny’s growth in question length drops off, coinciding with the period around age three when we see some instability in the growth of her MLU. Number of questions The percentage of questions to utterances was calculated across the sample. Apart from the earliest recordings, the percentage of questions asked comprised approximately 10–15% of the language produced by Jenny. The percentage of questions to utterances increased slightly over time and the overall percentage was 12.93%. This finding is consistent with that of Smith (1933) who found that the percentage of questions to utterances grew and that questions constituted 13% of children’s utterances between the ages of 1;6 and 6;0; it was also consistent with the findings of McCarthy (1930), who reports 6–14%. De Houwer (1990) calculated that, for her bilingual subject Kate, 18% of the English utterances were questions (10% yes/no and 8% wh-questions) and 82% were declarative utterances. This matched exactly with Wells’ (1985) findings for monolingual children. In Jenny’s case, we find somewhat similar proportions in that non-wh-questions comprise 7.21%, whquestions 5.72% and declarative utterances 87.07%. This finding in a trilingual child would appear to add some support to the tentative hypothesis made by De Houwer (1990: 290) that there may be some common phenomenon at work, at least in relation to the acquisition of questions.

Yes/no questions Fletcher (1985) describes declarative questions as potential yes/no questions. Smith (1933) and Lewis (1938) report declarative questions as being

209 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Early Trilingualism

210

present well before age two and they are present throughout Jenny’s data from the start of the sample at 1;11. They constitute, with 37%, the second largest group of questions produced overall. Although declarative questions show a decrease over time this does not take place as early as in monolingual subjects. When we examined data from Wells (1985) and from the monolingual subjects Sophie (Fletcher, 1985) and Sarah (Brown, 1973) we observed that declarative questions decrease as yes/no questions rapidly increase. This is not so in Jenny’s case where numbers of declarative questions, in spite of decreasing, remain high and numbers of yes/no questions increase but remain low. This discrepancy can, we believe, be explained by cross-linguistic influence from Spanish and Basque, as will be seen below. Pérez-Vidal (1995) also reports a predominance of declarative questions in her Catalan/English bilingual subject and suggests that this is transfer from Catalan in which, like Spanish (and to a lesser extent Basque), intonation on the affirmative is the principal marker of yes/no questions. Tyack and Ingram (1977) found that yes/no questions constituted 60% of questions used by children in the 2;0–3;11 age group they studied. This finding corresponds to that of 61% found by Smith (1933) for children aged 2–6 years. According to our findings yes/no questions form only 5% of the sample between 1;11.24 and 3;6.18. However closer examination reveals that both Tyack and Ingram (1977) and Smith (1933) are referring to non whquestions. If the total number of non wh-questions (including intonation, declarative, other and yes/no questions) is calculated for Jenny we obtain a total of 56% which clearly matches their findings. Correspondingly whquestions comprise around 40% in the findings of Tyack and Ingram (1977) and Smith (1933) and 44% in Jenny’s data. With regard to inversion and auxiliary insertion the earliest inverted yes/ no questions found in Jenny’s data appear at 2;3.14. There are three examples and they are all can-initial (in other words, they begin with can). Fletcher (1985) also finds that can-initial questions are a common early inverted form at 2;5. This is consistent with the findings of Wells (1985), who reports that yes/no questions of the type aux + S + V + O/Aare the first to appear. It is also an example of the early inversion described by Santelmann et al. (2002). The next auxiliary to appear in initial position in Jenny’s data is will at 2;8.22 and 2;10.30. Do insertion in aux + S + V + O/A does not take place until 2;10.30, when it appears as does. Only five other cases involving do subsequently appear in the data. The last novel auxiliaries to appear in initial position in the data are have at 3;1.9 and shall at 3;6.18. In general, if Jenny’s progress is compared with that of the children from Klima and Bellugi’s (1966) study described in Cazden (1970), she compares well and is even slightly in advance of Period C (age 3;2, MLU 3.5) when inversion and

210 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Interpretation of the Findings

211

auxiliary insertion is described as taking place. If we consider Jenny to have some characteristics of a bilingual, this would contradict the finding of Swain (1972) that first use of yes/no questions appears later in bilingual children than monolinguals. After the emergence of the can-initial aux + S + V + O/A type question in Jenny’s data at 2;3.4 as described above, the next type of yes/no question to emerge is copula + S + X at 2;6.19. This is also consistent with Well’s (1985) findings on the order of acquisition for yes/no questions. However in Jenny’s data we find a tendency to use is with a pronoun it, this or that or simply a null subject (is + 0), instead of the use of a full subject. Moreover, there is persistent use of sequences containing is + 0/it/ this/that in initial position and this is not reported elsewhere. This overuse of the copula may be related to cross-linguistic factors that will be outlined below.

Wh-questions Jenny’s earliest wh-questions are of the what’s that/whaddat? type that are described as the first to emerge in the monolingual literature (Brown, 1968; Cazden, 1970; Holzman, 1972; Klima & Bellugi, 1966; Wells, 1985). The order of emergence for Jenny’s wh-forms is what, where, who, how, when and why, which corresponds in part with the findings of other researchers: • • • •

what, where, how, why, when and other (Smith, 1933); what, where, why, how, who and when (Tyack & Ingram, 1977); what, where, why, who, when and how (James & Seebach, 1982); what, where, why and who (Wells, 1985).

It is of interest that in Jenny’s case why, which appears early in the order in the above studies, is the last wh-word to appear in the data (age 2;6.3). On the other hand, who appears earlier (age 2;3.14) in Jenny’s order of acquisition than reported elsewhere. We have mentioned that wh-questions make up 44% of Jenny’s questions overall and that this corresponds to the findings of Tyack and Ingram (1977) and Smith (1933). However, there are differences in relation to the amounts of wh-question types produced. Firstly what-questions, which comprise 68%, account for a far larger proportion of Jenny’s questions than the 15% reported by Tyack and Ingram (1977) and Smith (1933). Secondly, whoquestions are reported as rarely appearing in children under four in Tyack and Ingram’s study and Smith (1933) also finds that who is rarely used (only 2% of questions overall). This contradicts our finding that who-questions increase steadily from 2;5.19 onwards and make up the third largest group of wh-words used by Jenny (11.6%). Thirdly why-questions are reported as frequent across all age groups by Tyack and Ingram (1977) and they are said

211 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Early Trilingualism

212

to increase from age 3;0. In Jenny’s data why-questions are rarely found and they make up 1.65% of total wh-questions. Where-questions make up 16% of Jenny’s questions which is more similar to the findings of other studies, such as 12% found by Tyack and Ingram (1977) and 11% found by Smith (1933). It is of note that the order of emergence of wh-forms reported by Smith (1933), Tyack and Ingram (1977), James and Seebach (1982) and Wells (1985) shows some variation between studies after the initial appearance and growth of what- and where-questions. So it is not surprising that Jenny also shows some variation. What is of more interest is the degree to which Jenny’s use of the question forms what, who and why varies from monolingual studies, possibly owing to semantic factors related to order of whacquisition in Jenny’s other languages. Alternatively, it has been suggested that the frequency of wh-words is related to the concepts that the child needs to express at different ages (Tyack & Ingram, 1977). It is certainly true that within the child-rearing context of Basque society Jenny has been exposed to many more people than is typical in the upbringing of a British toddler where a setting of mother and child with immediate family and friends tends to be the norm. This may explain the frequent and early use of who in her questions. With reference to the issue of inversion in wh-questions and yes/no questions we find that Jenny’s first fully formed wh-question what’s this funny noise? of the type wh + copula, appears at 2;8.19. The first wh-question to show correct inversion is what can I do? at 2;10.30. If we consider the number of correctly inverted yes/no questions and wh-questions that appear with can, we see that inversion appears earlier in yes/no questions (at 2;3.13) than in wh-questions (2;10.30). Insertion of the auxiliary do with inversion is found at 3:1.22 in what do you say? and there is evidence of do-insertion in yes/no questions at age 2;10.30. It therefore seems that do insertion appears in both types of questions at around age three, though slightly earlier in yes/no questions. This is consistent with Period C as described by Klima and Bellugi (1966) and Cazden (1970), when auxiliary verbs appear throughout the whole grammatical system. However it should be noted that full inversion and auxiliary insertion with do appear only occasionally in Jenny’s data (there are only 10 instances, 6 yes/no questions and 4 wh-questions) and that, in most questions requiring auxiliary insertion and inversion, one or both features are absent. Jenny’s development follows that described in the longitudinal studies by Klima and Bellugi (1966), Brown (1968) and Labov and Labov (1978) where yes/no questions are found to have full inversion and auxiliary

212 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Interpretation of the Findings

213

insertion before wh-questions. This contradicts Ingram and Tyack (1979) and Erreich (1984), who argue against a stage in which there is inversion and auxiliary insertion in yes/no questions but not in wh-questions. Fletcher (1985: 106) describes the widespread use of formulaic wh-initial phrases in his subject’s questions. He explains that these formulae, which appear to be inversion, are in fact used by the child as a strategy for auxiliary insertion. We find something very similar in Jenny’s data in her dependency on sequences consisting of wh+(’s)+ it/this/that for wh-questions. When used by Jenny, these sequences precede a variable predicate just as reported by Fletcher (1985). Rowland and Pine (2000) find that such formulae are made available in maternal input and that their use facilitates the production of correctly-inverted wh-questions. Jenny also persistently uses a yes/no question formula is + 0/it/this/that which does not appear in monolingual descriptions. This may be reliance on an idiosyncratic question form (Cazden, 1970; Labov & Labov, 1978) or evidence of crosslinguistic influence.

Other questions Other questions constitute the third largest group (11.8%) of questions in the corpus. Most of them are one-word utterances of the type eh? or huh? requesting clarification. If we look at the British subject Sophie (Fletcher, 1985) we find no examples of other questions in the British-English sample. This is not surprising since such question forms are considered impolite. On the other hand it is acceptable to use these expressions in American English, and we find that Sarah (Brown, 1973) does indeed use them. It is of interest to note that both Jenny and her British-English-speaking mother use this type of question. This may be as a result of transfer from Spanish and Basque, where their use is acceptable. In sum, the results show that, in general, Jenny’s development in question form follows a monolingual route in terms of time of emergence for whforms and auxiliary insertion and inversion. However, Jenny differs in the amount of wh-forms used and in the small number of questions that show correct inversion and auxiliary insertion. It may be that the few examples of productive auxiliary insertion and inversion result from her persistent use of the declarative for yes/no questions and of is + 0/it/this/that and wh+(’s) it/ this/that sequences.

213 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Early Trilingualism

214

Pragmatic Development The child’s question functions The third issue in the study related to • how this trilingual child used her questions functionally at different ages and as part of her communicative competence Jenny’s questions in each functional category all get longer over time, though there are differences between the categories. The longest questions occur in the interactive category and the shortest in clarification. In terms of numbers, real information questions formed the largest group of questions, and showed the greatest increase over the period of the study. Known information questions appear at age 1;11 and again at 2;2.24 and are used sporadically but with some increase in numbers. Interactive questions are also infrequent but do increase over the study. Confirmation and clarification questions decrease, presumably as Jenny’s command and understanding of the language gets stronger. Our findings are consistent with those of James and Seebach (1982) for numbers of questions in their informative category (our real information category).1 However, James and Seebach report a decrease in these questions as children get older, whilst in our study we find an increase. Holzman (1972) also finds growth in real information questions, as do PrzetacnikGierowska and Ligeza (1990) who relate this to cognitive development and quality of questioning interaction. James and Seebach (1982) find that their directive (interactive) function is not used until the age of 3, whilst in our findings it is present from age 1;11. In James and Seebach’s study conversational (known) questions are found to be present from age two and to decrease in the age three group; this is not in agreement with our findings, which show a slight increase. Results relating to children’s questions in the confirmation and clarification functions are not available in other studies for comparison. Our study showed a clear tendency for different forms to be used with different functions. This was also found by James and Seebach (1982), who noted, for instance, that yes/no questions are used more for directive (interactive) questions than for questions in their other informative (real) and conversational (known) categories. This is also true of Jenny’s data (if declarative questions are collapsed into the yes/no category). Ervin-Tripp (1977) and Garvey (1975) also report use of yes/no questions for the directive (interactive) function. James and Seebach (1982) found that wh-questions are used for informative (real) and conversational (known) functions, and this is confirmed by our data. Our findings also reflect those of Wells (1985)

214 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Interpretation of the Findings

215

in that wh-questions are used to request information whilst yes/no type questions (including declarative questions) are associated with requests for confirmation. Prinz’s (1982) findings that form and function appear in a parallel fashion are not in agreement with ours. Instead we find that, although Jenny may not have many forms available in the early recordings, she is able to use all the functions except known information (although this appears in the first recording, it is not found again until three months later). However this is not to say that Jenny does not make more sophisticated use of the functions as her language develops. The findings on form and function are consistent with those of Dore (1977: 148) who reports that ‘children interpret forms as having a variety of functions and functions as having a variety of forms’. The fact that Jenny is able to use questions (rather than statements) in all the functional categories by age 1;11 and shows pragmatic growth equivalent to or greater than that reported in other studies on groups of children can be taken as evidence of considerable pragmatic flexibility. This may be explained by a combination of factors. Firstly, Przetacnik-Gierowska and Ligeza (1990) relate question use and pragmatic function to the quality of questioning interaction in which the child is a participant and go on to suggest that it is superior in home dyads2 than in kindergartens with teachers. As most of Jenny’s questioning interaction in English has taken place with her mother this may help to explain her apparent flexibility. The same applies to Jenny’s use of Spanish with her caregiver and Basque with her father. On the other hand, Dore (1986) has found that flexibility in function can be a result of fantasy play with peers. As well as playing at home with her brothers and friends, from the age of 2;5 Jenny has attended kindergarten (in Basque) where this type of play took place. It is also possible that any resulting pragmatic growth in Basque will also be seen in English since pragmatic skills may be transferred across languages (Barnes, 2001; Blum-Kulka & House, 1989). From the point of view of developing pragmatic flexibility it would therefore seem that Jenny has had the best of both worlds both at home and in kindergarten. If we tentatively include as a factor the possibility that her three languages might influence one another, we have further justification for the pragmatic flexibility revealed both by the taxonomy and in the examples of her communicative competence. Communicative competence The findings on Jenny’s use of question functions as part of her communicative competence (Hymes, 1972) consist of 7 representative examples of her in the collaborative construction of meaning with her mother (Wells,

215 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Early Trilingualism

216

1985). Many other examples could have been included. The examples demonstrate that Jenny is able to use questions for functional purposes from an early age and to express a variety of communicative intents. She is able to attract attention, avoid the truth, initiate and maintain conversation, ask about new words, initiate play, express humour, show surprise, interest and enthusiasm, and to use strategies to obtain information she needs through the pragmatic use of questions. We note that Jenny becomes increasingly proficient at using known information questions to initiate and maintain conversations. This is consistent with the findings of Holzman (1972) who reports that test (known) questions seem to serve as conversation initiators, and that her subject Adam enjoys using test (known) questions back to his mother to maintain the conversation. Jenny’s use of question functions gets more sophisticated over time as is described by Gordon and Ervin-Tripp (1984) in the development of instrumental language. Her use of questions also seems to fulfil the criteria of solving the five problems of getting attention, clarity, persuasiveness, maintaining social relations and performing repairs (Ervin-Tripp & Gordon, 1986). Jenny makes use of reformulation and repetition to modify the function of the questions in order to achieve her communicative aim; this has also been reported by others (Ervin-Tripp, 1988; Gordon and Ervin-Tripp, 1984; Newcombe & Zaslow, 1981; Ochs-Keenan, 1977). Jenny also copies her mother’s functional use, for example when she makes a ‘joke’ by repeating questions from part of her mother’s earlier exchange. Retracing at a lexical level (in which the child picks up on words that have been used earlier in the conversation by the interlocutor and then uses them in his/her own speech) has been reported by Bloom, Merkin and Wootten (1982). It has not, however, as far we know, been described at a functional level. In sum, Jenny demonstrates communicative competence in that she not only has knowledge of the language but also has the ability to use the language appropriately. Mother’s functions The fourth research issue related to whether: • the child’s functional use of questions and her ability to give feedback to questions is related to the input and feedback she has received. The study showed that Jenny’s mother, except in the case of confirmation questions, used approximately the same number of questions in each functional category at the beginning of the study as she did at the end. This confirms the findings of Shatz (1979) that mothers do not use fewer func-

216 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Interpretation of the Findings

217

tions with fewer forms when their children are younger. Furthermore, Holzman (1972: 317) finds that the numbers of test (known) and clarification questions that mothers ask their infants decline during the period when the children have an utterance length of 2–4 morphemes (MLU 2–4), and (unlike the present study) Holzman reports a decline in confirmation questions. Jenny’s questions grow longer as her language develops. However, it is of interest that her mother keeps her own question length within the limit of 4–6 words described by Davis (1932) as the natural length for questions from age three to adulthood. For both mother and child, real information questions are the longest and clarification questions the shortest. Close similarities are to be found between mother and child in the distribution patterns of forms to functions. In order to make the comparison, the percentages of all yes/no type questions (yes/no questions, tag questions and declarative questions) were collapsed into one category. In the real information, known information and confirmation categories, Jenny uses proportions of wh- and yes/no question forms that are almost identical to those of her mother. In the confirmation questions category, both mother and daughter make use of other questions in similar proportions not seen elsewhere. Clarification questions are distinguished by the use of a higher proportion of other questions by both participants, and lower numbers of wh- and yes/no questions. The results suggest that, irrespective of the child’s age, the mother uses steady numbers of questions and length of question with a variety of forms and functions, as has also been described by Shatz (1979) and Olsen-Fulero and Conforti (1983). Although we have limited our examination of form– function pairings to general patterns of yes/no, wh- and other questions, we observed a relationship between the mother and child’s use of them. This suggests that there is a relation between the input the child has received and her own production. Mother and child feedback The results show that 50% of the feedback Jenny gave to her mother was acceptable. This is fairly consistent with findings from Olsen-Fulero and Conforti’s (1983) study of 11 children which showed that 57.28% of their feedback to their mothers was acceptable. Of the feedback given by Jenny’s mother, 59.8% was acceptable. Jenny’s mother gave very little unacceptable feedback (2.2%) to her, suggesting that she made an effort to reply to the questions (although she had fewer of them to answer than Jenny had). Jenny gave a somewhat greater amount of unacceptable feedback (10.1%), perhaps because she was

217 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Early Trilingualism

218

less proficient in her use of language and may have failed to understand the questions or had linguistic difficulties in replying. Additionally she may not have known what to reply. This is consistent with Shatz’s (1979: 1096) finding that children often failed to reply meaningfully to parental questioning. Likewise Dore (1977: 153) reports that, for children, failure to answer a question does not yet seem constitute a violation of social obligation. Taking into consideration the amounts of acceptable and unacceptable feedback that have been described, there appears to be a sizeable quantity of questions that receive no feedback as they are usurped by the speaker before the interlocutor is able to reply. In the case of the mother’s feedback this is 37.5% of questions, and Jenny’s case it is 39%. The proportions for mother and child are remarkably similar, which suggests that there is a relationship between the amount of usurpation performed by the mother and the amount that Jenny performs. In the absence of data from other studies with which to make a similar comparison, we cannot take this speculation further. If they are taken overall and not by individual functional category, the mother and child’s feedback results show similar patterns. When the responses are divided into functional categories some interesting contrasts between mother and child appear. The results for Jenny by functional category reflect those found by Olsen-Fulero and Conforti (1983) for children, but there are no results on mothers’ feedback available with which to compare the findings on Jenny’s mother. Salient contrasts include the finding that, when ranked, real information feedback consistently appears in a mid-position for both mother and child. This differs from the other categories where the feedback results from mother and child take up polarised positions of varying degrees. This is particularly true of feedback to confirmation questions where we find that Jenny’s mother gives the greatest amount of acceptable feedback (82%) and Jenny gives her lowest (40.7%). In addition cases of no feedback are lowest for the mother in this category (17.8%) and highest for Jenny (52.6%). The findings for mother and child feedback may now be related to those for question function. There is least variation between mother and child in the real information and interactive question categories – these are also the two categories in which most and least questions are asked respectively. Confirmation and clarification questions again show the most variation. For the mother, confirmation is the only category that shows an increase in the number of questions used whilst in Jenny’s case, along with clarification, it is the only category that shows a decrease and polarisation can again be observed. It is tempting to relate the increase in Jenny’s mother’s use of

218 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:26

Color profile: Disabled Composite Default screen

Interpretation of the Findings

219

confirmation questions to discourse factors. As Jenny’s utterances increase, her mother may feel motivated to maintain the conversation by using confirmation questions since these provide a response whilst maintaining the discourse as the turn is given back to the child (Garvey, 1977; OlsenFulero & Conforti, 1983). However our results on feedback reveal that Jenny supplies answers to only about 40% of these questions and that they are usurped by her mother 52% of the time. The usurpation of confirmation questions is also reported by Olsen-Fulero and Conforti (1983), though it is not clear why mothers should do this. The patterns that we have observed in the use of question functions and their feedback by the mother and the child relating to similarities and polarisation suggest that there is a relationship between input and feedback. The most striking polarisations (uses of the same function or feedback by the mother and child that diverge so widely as to almost seem opposite) appear in the categories of confirmation and clarification. This is consistent with findings from previous studies that describe the importance of contingent queries and clarification requests (including the confirmation category) in repair, turn allocation and the development of discourse and communicative competence (Corsaro, 1977; Garvey, 1977; Wells, 1985). In terms of the pragmatic development of interrogative behaviour, our results indicate that a trilingual child can develop her knowledge of the functional use of questions in a similar way to monolingual children. Compared with data available from other studies, the mother’s behaviour, by and large, reflects that of mothers in monolingual contexts in terms of the maintenance of question length and the use of similar numbers of questions per function irrespective of the child’s age. Relationships between the mother’s and the child’s questioning and feedback behaviour are suggested by the findings in the area of form–function pairings, overall feedback patterns and polarisation patterns in individual functional categories.

Cross-linguistic Influence In previous sections we have explained Jenny’s development in English and in interrogative behaviour as if it were monolingual. This is in line with research findings on language development in bilingual children, which conclude that each language develops separately (De Houwer, 1990, 1995; Genesee, 1989; Meisel, 1989, 2001; Murrell, 1966; Quay, 1995) and takes into account findings on early trilingualism. Jenny’s language acquisition was found to follow, by and large, a monolingual route and any variations as a possible consequence of her trilingualism have been noted. With reference

219 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

220

to Jenny’s development in English, her MLU and TTR are similar to those found in a sample of monolingual and bilingual subjects. Lexical mixing involves fewer than 2% of utterances in the whole corpus. Given that Jenny’s input in English is from extremely limited sources, we therefore concluded that these sources were sufficient for successful language development to occur. Nevertheless, since Jenny is trilingual, the data were examined for specific evidence of cross-linguistic influence from Basque and Spanish and the final research issue concerned the extent to which there is evidence of cross-linguistic influence in the child’s production. The findings will now be summarised and explained in relation to mixing in Jenny’s general English and her syntactic influence in the acquisition of questions. Cross-linguistic influence in general English There is minimal evidence in our study of the presence of Basque and Spanish in Jenny’s utterances (1.71%), and more mixing appears in Jenny’s early recordings at around age 2 than at around age 3. This confirms the findings of De Houwer (1990) that lexical mixing at age three is infrequent and of Deuchar and Quay (1998, 2000) and Van der Linden (2000) that evidence of the other language becomes inhibited with time. Our search for examples of cross-linguistic influence was not straightforward. An analysis of the early data for mixes revealed some words could not be identified as language specific. Such words could be the result of developmental forms from any of the three languages or possibly a mixture of two or three of them. This is also reported by Quay (2001). Proper names and words that were Spanish/Basque cognates complicated the identification of the base language of the utterance, as also described by Tracy (2000: 16–17). On the other hand, since there are only limited possibilities of input available to Jenny, it is relatively easy to trace the source of a word in a trilingual child as was also found by Hoffmann (1985) and, in a bilingual, by Döpke (2000). Language-specific homewords were identified (Stavans, 1992) and there was evidence of language being related to topic (Mikes, 1990). There were more instances of Spanish-only (35) or Spanish/English (48) utterances than Basque-only (24) and Basque/English (42) utterances, showing that influence from Spanish was slightly greater. However, mixes from Spanish usually involved lexical items and expressions, whereas Basque mixes involved lexical items and some morphology. Lanza (1992) has suggested that, for bilingual children, evidence of grammatical morphemes from language A with lexical items from language B implies that language A is dominant. In the present study we would not go so far as to suggest that Basque is dominant on the basis of only a few examples, espe-

220 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Interpretation of the Findings

221

cially in view of the fact that Spanish is the language that most often appears in mixes. According to Hoffmann (1985) and Hoffmann and Widdicombe (1999) for trilinguals, it is the community language, and the language of schooling that will show dominance. In Jenny’s case, the status of Basque and Spanish is arguable since the child lives in a bilingual community. However, up until the time she started schooling in Basque at age 2;5, Jenny’s exposure to all three languages in her linguistic microcosm (De Houwer, 1995) was deemed to be equal. The only way we could say more about dominance, if at all pertinent, would be to search for evidence of mixing in English in the recordings we have of Jenny speaking Spanish and Basque. The presence of slightly more mixing from Spanish than from Basque might also be explained by the linguistic typology of the languages. It has been shown that in multilingual contexts languages that are similar are more likely to show transfer into each other than are those that are different (Cenoz, 2001; Cenoz & Genesee, 1998; Singleton, 1987). Although English and Spanish are not closely related, they are both Indo-European languages whilst Basque is not. For this reason we might expect to see more influence from Spanish in Jenny’s English than from Basque. As the child used some words that are cognates existing in both Spanish and Basque, it was difficult to address the issue of whether the three languages were ever mixed in one utterance, since no examples were found in which all three languages could be clearly identified. It has been noted that trilinguals do not code-switch with monolingual speakers (Stavans, 1992), yet may use all their languages with trilinguals or multilinguals (Faingold, 1999; Mikès, 1990; Oksaar, 1978; Quay, 2001). A tendency has also been reported for trilingual children to code-switch in only two languages even if the interlocutor is trilingual (Stavans, 1992; Hoffmann & Widdicombe 1999; Hoffmann 2001). Jenny has not heard her mother speaking Basque (though she may realise that her mother understands it in family contexts). However she does know that her mother is a fluent speaker of Spanish, having seen her use it with the child minder and others, so she may perceive of her mother as a bilingual (Grosjean, 1994, 1997; Lanza, 1997; Nicoladis & Genesee, 1998). This might explain why Jenny uses more Spanish than Basque mixes with her mother, and the appearance of two-language mixes (with Basque/Spanish cognates) rather than mixes in which three languages are clearly present. The small number of mixes that appear in the corpus show that the child’s base language with her mother is clearly English, and the mother has negotiated a monolingual context (Lanza, 1992). This is proved by the examination of a sample of 11 of the mother’s clarification requests in

221 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

222

relation to the repair of the child’s mixes. As the sample was selected on the basis of the mother’s involvement in repairing the mix, it must be presumed that, as described in Döpke (1992), the mother ignored most other mixes. When required to make a repair relating to language choice, the child usually did so, as is also reported in Comeau and Genesee (2001). If bilingual behaviour is viewed as extending along a continuum (Lanza, 1992, 1997) or scale (Grosjean, 1992) from monolingual mode to bilingual mode, in the sample Jenny’s mother uses a mixture of strategies from the monolingual end and the middle of the continuum, such as minimal grasp, expressed guess and repetition (Lanza, 1992, 1997). She makes it clear through her responses to the child’s mixes that their discourse is in monolingual mode and rarely uses strategies from the bilingual end of the continuum (Juan-Garau & Pérez-Vidal, 2001). English, the minority language, is further supported by being used between the parents and between the mother and Jenny’s siblings, although different degrees of mixing are probably taking place (Goodz, 1989). Various studies describe the presence of cross-linguistic equivalents in recordings from bilingual children’s different languages at the same age (Deuchar & Quay, 2000; Quay, 1995). This is used as evidence of the child’s differentiation of the lexicon. In our study, however, we find a number of cross-linguistic equivalents in the same utterance (1 case), in the same exchange (7 cases) or the same script (1 case). The words are not related to the maternal clarification strategies outlined above, but seem to appear unprompted. Slightly more of these equivalents involved Basque (5 cases) than Spanish (3 cases). One other word was a Basque/Spanish cognate. Such words tended to appear after the English word (6 times) rather than before it (3 times), suggesting a sudden switch out of the base language. It is not clear what may trigger such switches. Van der Linden (2000) also describes the use of equivalents in the same recording of the child, though it is not clear if they occur after or before the same word in the base language. Van der Linden notes that they diminish at around 2;5, along with evidence of the other language. However in Jenny’s case, they turn up from age 2;4 until the end of data collection at 3;6. Cross-linguistic influence in questions Cross-linguistic influence in questions was analysed at both a general and a specific level. In the general analysis, a number of areas in which cross-linguistic influence might be expected to appear were predicted on the basis of adult and developmental forms in Spanish and Basque that differed from English. These were:

222 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Interpretation of the Findings

(A) (B) (C) (D)

223

use of declaratives with intonation for questions; use of null-subjects; wh- positioned late in the question; subject placed on the right.

Evidence of all four areas was found in Jenny’s questions. During the analysis of question form a number of other areas were identified as non-standard and the possible result of cross-linguistic influence. These were: (E) (F) (G) (H) (I) (J) (K)

use of is + 0/it/this/that? use of wh + it/this/that? subject repetition; word order; wh-confusion; hybrids; translation.

It was concluded on the basis of these findings that 11 potential areas of cross-linguistic influence were present. Cross-linguistic errors may also be confused with developmental errors or might be explained either as performance errors or by competence factors. They may simply be idiosyncratic to the child (Döpke, 2000: 5), or as a result of trilingual contact (Paradis, 2000: 176). In order to identify those errors that were probably cross-linguistic in origin, a specific analysis was carried out on samples of questions from two monolingual children, Sophie (Fletcher, 1985) and Sarah (Brown, 1973) at Brown’s (1973) Stages I– V. The results were compared with Jenny’s questions at the same stages. If examples of categories A–K above were found in the monolingual children’s question data, they were considered to be developmental or idiosyncratic forms rather than a result of cross-linguistic influence. Evidence of the following forms were identified in the question samples taken from the two monolingual children: (B) (C) (D) (E) (H) (J)

use of null subjects; wh- positioned late in the question; subject placed on the right. use of is + 0/it/this/that? word order; wh-confusion.

Five areas were found only in Jenny’s question data, and might therefore be claimed to be the result of cross-linguistic influence:

223 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

224

(A) Numbers of declarative questions with intonation

Our finding that Jenny used large numbers of declarative questions is consistent with that of Pérez-Vidal (1995), who found that this type of question predominated in her Catalan/English bilingual subject. Catalan is a language close to Spanish and also uses declarative utterances with intonation for yes/no questions. Basque also permits this type of questions, but has additional yes/no question forms based on word order. Jenny’s persistent use of the declarative question form in English and the possible relation of this to the reduced numbers of fully-formed yes/no questions she produces may be explained as cross-linguistic influence from her other languages, particularly Spanish. (E) Use of is + 0/it/this/that?

The copula sequence is + 0/it/this/that? is found at the beginning of many well-formed yes/no questions in the data, but it is also used to produce non fully-formed yes/no questions such as Is that is my friend is it yet? or Is this is green? Fletcher (1985) and Rowland and Pine (2000) describe the formulaic use of wh-questions such as what’s + 0/it/this/that as a type of scaffolding prior to inversions and insertion of an auxiliary. We propose that Jenny makes a similar use of is + 0/it/this/that expressions. She produced few fullyformed yes/no questions during the period under study though questions of the is + 0/it/this/that type are frequent and seem to be used as interrogative markers in a stage between declarative questions and fully-formed yes/ no questions. There is no evidence of this type of non-fully-formed yes/no question in the samples taken from two monolingual subjects, which suggests that it is the result of cross-linguistic influence. Furthermore its frequency may be related to the dependence on declarative questions attributed to cross-linguistic factors above. The whole process suggests that the systems of interrogative formation related to use of the declarative from Spanish and (to a lesser extent) Basque are working together complementarily with the English system and that bootstrapping is taking place (Gawlitzek-Maiwald & Tracy, 1996). The frequent appearance of is + 0/it/this/that in fully-formed and nonfully-formed questions may also be a result of cross-linguistic influence for a different reason. Basque and Spanish each have two copula verbs, one of which expresses permanence and the other expresses temporariness and location, whilst English has one (to be) that expresses all three concepts. None of the four copula verbs available for Basque and Spanish requires a subject and thus they outweigh English usage (which requires a subject) four to one. Null subjects have been excluded from the list of possible crosslinguistic influence as they are found in the questions of the monolingual

224 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Interpretation of the Findings

225

child Sarah (e.g. what doing?) and have also been reported as an occasional feature of early child English. Furthermore, Juan-Garau and Pérez-Vidal (2001) find that their bilingual child Andreu is able to correctly place the subject in English affirmatives and leave a null subject in Catalan. An informal check revealed that Jenny correctly uses subjects in her affirmatives. However, we suspect that the null subject in the interrogative expression is + 0? is the result of cross-linguistic influence, since no examples of questions beginning is + 0? were found in the monolingual data and they are quite common in Jenny’s questions. We tentatively propose that Jenny may be over-extending to English the rule related to the use of the nullsubject copula verbs in initial position for yes/no questions in Spanish, and to a lesser extent Basque. If this is so, it exemplifies the complexity that may result from multilingualism (Cenoz, 2000; Hoffmann, 2001). (G) Subject repetition

The process just described seems to be reversed in the case of subject repetition. The child produces questions such as what’s it doing this giraffe? or what he doing Jon? in which a subject pronoun is correctly inserted after wh-movement, but then the subject is repeated at the end of the question, resulting in double use of the subject. This structure is not normally found in adult questions and does not appear in the two monolingual samples studied. However, it has also been noted in slightly older bilingual Basque/ Spanish children who are learning English as a third language (Cenoz, personal communication; Cenoz, 2000, 2001). A possible explanation for this type of question again lies with the null-subject found in Spanish and Basque as Jenny seems to overcompensate for it in English. In cases of subject repetition she actually follows the English rule whereby (unlike Spanish and Basque), a subject pronoun must be inserted. She places a pronoun after wh- and the auxiliary or copula to be, if they are present, but also adds the full subject at the end of the question. This would be the normal position for a full subject in Spanish questions and sometimes in Basque. The findings lead us to conclude that Jenny has two rules relating to the subject that result in non-fully-formed questions in English. If the question contains only a subject pronoun, she applies the null-subject rule from Basque and Spanish and omits it, e.g. is only on mine? However if the question contains a full subject, this will be placed in final position (as in Spanish and Basque) and a subject pronoun will also be inserted, e.g. what’s this doing here this button? (I) Wh-confusion

Cases of wh-confusion were found in questions beginning with what, where, who and sometimes involved why as a possible meaning. It has

225 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

226

already been noted that Jenny used why considerably less than her monolingual peers did. She either misused the interrogative pronoun, e.g. who’s done the elephant? (meaning what), or pronounced it in an unclear way as if she realised she had to use a word beginning with wh- but was unsure of which wh-word to use, e.g. Mummy wha (where) is going? Wh-confusion has not, to our knowledge, been reported in the literature on questions in monolingual children or bilingual children (see Gut, 2001 for a report on the acquisition of question intonation in an English/German bilingual child). It may be evidence of semantic and phonological confusion resulting from the number of different (and somewhat similar) wh-words Jenny that has available to her if all three languages are taken into consideration. (K) Translation

Swain and Wesche (1975) have reported the direct syntactic copying of question forms from one language into the other in a French/English bilingual child. In the present study, few examples of this type of transfer were found, either because they were simply not there or because developmental forms and the interaction of the three languages made them complex and difficult to identify, and such cases have been examined above. One example of direct translation is how do a donkey? which is a syntactic and lexical copy of the way Basque and Spanish ask ‘how does a donkey go?’ The five categories (A, E, G, I, K) identified above as evidence of crosslinguistic influence in Jenny’s questions are related to syntactic patterns common to both Spanish and Basque. As in the case of lexical mixing, where it was found that mixes in three language tended to involve Basque/ Spanish cognates, it is difficult to say exactly which of the two languages is revealed, and how, in the evidence of cross-linguistic influence. This is an area that needs further study.

Conclusions and Implications It has been the aim of this book to explore different aspects of the acquisition of interrogative behaviour in English in a trilingual child. This might be considered over-ambitious, and in fact during the course of the study a number of areas were revealed which could have been investigated in greater detail. The relation between the child’s questions and her mother’s questions, for instance, turned out to be an area of research worthy of a book in itself. Some limitations with relation to the child, her questions and her trilingualism were noted. Firstly, this is a case study, and so looks at the development of a single child, and there is no way of knowing if these results will prove to be true of other children, trilingual or otherwise. In

226 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Interpretation of the Findings

227

relation to individual differences we might investigate if Jenny has developed as a referential or expressive child (Nelson, 1981) as the result of being the youngest of three children. In the area of questions we found that Jenny produced only nine examples of why, and it would be interesting to investigate why she produced so few. If she had produced more, we could have compared her results with work by Piaget (1926), Labov and Labov (1978) and Fletcher (1985). Also, in comparison with monolingual children, Jenny made more use of who and less use of why and we should investigate if this is the result of crosslinguistic influence at a semantic level. We have also examined the way whforms as a whole have been used in the different functional categories, yet it might be interesting to see if there is a tendency for specific wh-forms to be used in certain categories, and to note when these have emerged. By doing so further comparisons could be made with the findings of James and Seebach (1982). Since we examined only one of Jenny’s languages, we are somewhat limited in the assumptions we may make about her as a trilingual. More detailed comparisons with the questions of English monolinguals, and bilingual and trilingual children who have English as one of their languages along with Spanish and Basque, should be carried out to confirm the areas identified in this study as possible evidence of cross-linguistic influence. With reference to Jenny’s question acquisition and her mother’s linguistic behaviour, more work could be done in various areas. In connection with form, we might calculate the frequency of wh+ it/this/that in Jenny’s data and in the mother’s data to find if there is a relation between them as described in Rowland and Pine (2000, 2003). It would also be of interest to examine how the mother’s form and style have affected Jenny’s acquisition of questions (and other aspects of her acquisition). In the area of function our findings have shown that there is a link between the mother’s and the child’s use of forms and functions in questions and that this is related to the development of discourse skills. However there is a need to look more closely at the use of clarification and confirmation questions by Jenny and her mother as contingent queries in discourse development. More, too, could be said about feedback, which has proved to be a very rich area for examining the interaction between mother and child. Repetitions, reformulations and the length of replies could examined. The study of the mother’s influence at cross-linguistic level should be further developed. Firstly, it would be useful to know the amount of mixing that is found in her English. All 163 examples of mixes used by Jenny in the

227 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

228

data should then be examined to see how her mother responds to them according to Lanza’s (1992) categorisation of parental discourse strategies. A number of exciting areas remain to be investigated, as the corpus is a resource for future research on mother and daughter. Firstly, the findings from this study could be compared with an analysis of the three recordings available in Basque and Spanish to see the extent of cross-linguistic influence from English. Secondly, it could be of interest to look at other areas of syntactic development in Jenny’s English. At a more general level, Jenny’s pronunciation in English as one of three languages in contact is an area of great interest to the researcher and may be the subject of another study. Lastly a complete analysis of all the mother’s 32 recordings could be carried out in order to examine in more depth the way the input has affected the child’s acquisition of questions, question function, and her trilingualism. In our study of a trilingual child’s acquisition of English questions from both a formal and functional perspective, in relation to the mother’s questioning behaviour and in comparison to studies of monolingual children, a number of noteworthy points were reported. In the area of question form, we found that Jenny produced proportions of questions and utterances similar to those reported elsewhere and, moreover, that she produced whand yes/no questions in similar amounts to those reported in other studies of monolinguals and bilinguals. This confirms the tentative hypothesis expressed by De Houwer (1990) in relation to the monolingual and bilingual acquisition of questions. This needs further research as it may have important implications for the understanding of language acquisition. A second finding has been that order of emergence and rate of use may be affected by contextual and societal constraints. In our subject the interrogative pronoun who appears earlier and is more frequent than why when compared with monolingual acquisition. We have speculated that this may be due to societal differences in child-rearing practices between the Basque Country, Britain and North America. In relation to function we have demonstrated that from the age of 1;11 the trilingual child Jenny is capable of conveying a wide variety of communicative intents with limited language by making pragmatic use of questions, thus participating in the collaborative construction of meaning (Wells, 1985). Jenny also develops in advance of monolingual peers in the use of question function (James & Seebach, 1982). Pragmatic flexibility has been attributed to the quality of questioning interaction (Przetacnik-Gierowska & Ligeza, 1990) and fantasy play with peers (Dore, 1986). We suggest that Jenny’s pragmatic abilities have been enhanced by her upbringing in which she has had access to quality question interaction in three languages plus play experience with siblings and at kindergarten.

228 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Interpretation of the Findings

229

Some interesting facts emerge when we examine the data from Jenny’s mother. We find that she does not increase the number of questions that she uses in each functional category over time, which confirms the findings of Holzman (1972) and Shatz (1979). Furthermore our study confirms and relates previous research on form-function pairings in the mother (Shatz, 1979) and child (James & Seebach, 1982) to show that there is a clear relationship between the input Jenny receives and her own choice of forms to express questions. The descriptions of child feedback patterns obtained from the data reflect those found in a study of monolingual children (Olsen-Fulero & Conforti, 1983). To our knowledge there are no other studies relating to maternal feedback to children’s question functions. Our data on maternal feedback show that the mother makes an effort to reply to a larger number of questions than the child does to her and also that the child performs usurpation of answer to the same degree as her mother (Olsen-Fulero & Conforti, 1983). These findings reveal the importance of usurpation as a feature of mother and child pragmatic behaviour and suggest that feedback patterns have an important role to play in the development of questioning behaviour and of discourse. An overall comparison of the mother and child’s questions and feedback reveals a number of similar patterns and polarised contrasts which suggest that question interactions are intricately woven together and serve a purpose in discourse construction. This is an area that needs further study. Jenny’s results in English MLU and TTR and the small amount of mixing that takes place at a lexical level seem to prove that, given the right linguistic environment, albeit with limited input, a third language can develop to a proficiency equivalent to that of monolinguals. However, the considerable cross-linguistic influence we find in relation to her question development lead us to conclude that, although Jenny may appear to be like a monolingual at first glance, at a syntactic level things are not so straightforward. We discover that she largely follows a monolingual route in her acquisition of questions, but shows some differences that are attributable to cross-linguistic influence, principally the persistent use of declarative questions. As a result of this, the appearance of inversion and auxiliary insertion as found in fully-formed English questions is limited, though not delayed. Moreover we observe her dependence on formulas based on is + =/it/ this/that? and wh- + it/this/that? as bootstrapping on to inversion and auxiliary insertion, and the use of over-extension and under-extension of rules relating to null subjects in Spanish and Basque. Our research shows that the examination of cross-linguistic structures in bilinguals and trilinguals is crucial to an understanding of the issue of how languages develop (Paradis

229 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

230

& Genesee, 1996) in a field where the discussion of lexical mixing has long prevailed. We therefore welcome the recent interest (Genesee, 2000, Döpke, 2000) in cross-linguistic syntactic aspects. Lastly, with regard to mixing it was found that Jenny spontaneously produced equivalents in either Spanish or Basque in the same utterance or exchange that she used an English lexical item. She did not produce any utterances in which she clearly used three languages but rather made use of words that were Basque and Spanish cognates, or that could not be clearly identified as coming from Basque or Spanish. In order to find out more about trilingual mixing we need to work with languages in which there are no cognates. Notes 1. The functional category name used in the present study is placed in parenthesis after the name used in other studies. As pointed out in Chapter 2, the degree of correspondence between functional categories across studies is open to discussion. 2. A pair consisting of the caregiver (e.g. mother) and a child interacting in a homelike context rather than an institutional one.

230 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

References Aksu, A. (1973) The development of request forms in Turkish children. Unpublished manuscript, University of California. Arnberg, L. (1987) Raising Children Bilingually: The Pre-school Years. Clevedon: Multilingual Matters. Austin, J. (1962) How to Do Things With Words. New York: Oxford University Press. Babelot, G. and Marcos, H. (1999) Comprehension of directives in young children: Influence of social situation and linguistic form. First Language 19, 165–186. Bain, B. and Yu, A. (1980) Cognitive consequences of raising children bilingually: ‘One parent, one language’. Canadian Journal of Psychology 34, 304–313. Barnes, J. (2001) Politeness in English, Basque and Spanish: Evidence from a trilingual child. Jakingarriak 45, 40–45 Barreña, A. (1994) Sobre la adquisición de la categoría funcional COMP por niños vascos [On the acquisition of the functional category COMP by Basque children]. In J.M. Meisel (ed.) La adquisición del vasco y del castellano en niños bilingües [The Acquisition of Basque and Spanish by Bilingual Children] (pp. 231–284). Madrid: Iberoamericana. Barreña, A. (2001) La adquisición de COMP por un niño bilingüe vasco-español [The acquisition of COMP by a Basque-Spanish bilingual child]. Revista Española de Lingüística. Barron-Hauwaert, S. (2000) Issues surrounding trilingual families: Children with simultaneous exposure to three languages. Zeitschrift für Interkulturellen Fremdsprachenunterricht 5 (1). On WWW at http://www.ualberta.ca/~german/ ejournal/barron.htm. Accessed 01.5.00. Barron-Hauwert, S. (2004) Language Strategies for Bilingual Families: The One-parent One-language Approach. Clevedon: Multilingual Matters. Bates, E. (1971) The development of conversational skill in 2, 3 and 4-year olds. Masters dissertation, University of Chicago. Bates, E. (1976) Language and Context: The Acquisition of Pragmatics. New York: Academic Press. Becker, J. (1982) Children’s strategic use of requests to mark and manipulate social status. In S. Kuczaj (ed.) Language Development: Language, Thought and Culture (Vol. 2). Hillsdale, NJ: Erlbaum. Becker, J. (1986) Bossy and nice requests: Children’s production and interpretation. Merrill Palmer Quarterly 32, 393–413. Bellugi, U. (1965) The development of interrogative structures in children’s speech. In K. Reigel (ed.) The Development of Language Functions. Ann Arbor: Michigan Language Development Program Report No. 8. Bellugi, U. (1967) The acquisition of the system of negation in children’s speech. Unpublished doctoral thesis, Harvard University.

231 231 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

232

Bellugi, U. (1971) Simplification in children’s language. In R. Huxley and E. Ingram (eds) Methods and Models in Language Acquisition. New York: Academic Press. Ben-Zeev, S. (1977) The influence of bilingualism on cognitive strategy and cognitive development. Child Development 48, 1009–1018. Bentahila, A. and Davies, E. (1994) Two languages, three varieties: Code switching patterns of bilingual children. In G. Extra and L. Verhoeven (eds) The Crosslinguistic Study of Bilingual Development (pp. 75–97). Amsterdam: Royal Netherlands Academy of Arts and Sciences. Bergman, C. (1976) Interference vs. independent development in infant bilingualism. In D. Keller, R. Teschner and S. Viera (eds) Bilingualism in the Bicentennial and Beyond (pp. 86–96). New York: Bilingual Press/Editorial Bilingüe. Bialystok, E. (1993) Symbolic representation and attentional control in pragmatic competence. In G. Kasper and S. Blum-Kulka (eds) Interlanguage Pragmatics (pp. 43–57). Oxford: Oxford University Press. Bloom, L., Merkin, S. and Wootten, J. (1982) Wh-questions: Linguistic factors that contribute to the sequence of acquisition. Child Development 53, 1084–1092. Bloom, L., Rocissano, L. and Hood, L. (1976) Adult–child discourse: Developmental interaction between information processing and linguistic knowledge. Cognitive Psychology 8, 521–552. Blount, B. (1977) Ethnography and caretaker–child interaction. In C.E. Snow and C.A. Ferguson (eds) Talking to Children: Language Input and Acquisition. Cambridge: Cambridge University Press. Blum-Kulka, S. (1991) Interlanguage pragmatics: The case of requests. In R. Phillipson, E. Kellermann, L. Selinker, M. Sharwood Smith and M. Swain (eds) Foreign/Second Language Pedagogy Research (pp. 255–272). Clevedon: Multilingual Matters. Blum-Kulka, S. and House, J. (1989) Variation in requesting behaviour. In S. BlumKulka, J. House and G. Kasper (eds) Cross-cultural Pragmatics: Requests and Apologies (pp. 123–153). New Jersey: Ablex. Bohannon III, J. and Bonvillian, J. (2001) Theoretical approaches to language acquisition. In J. Berko Gleason, The Development of Language (5th edn; pp. 254–315). Needham, MA: Allyn & Bacon. Brown, R. (1968) The development of wh-questions in child speech. Journal of Verbal Learning and Verbal Behaviour 7, 279–290. Brown, R. (1973) A First Language: The Early Stages. London: Allen and Unwin. Brown, R., Cazden, C. and Bellugi, U. (1969) The child’s grammar from I to III. In J.P. Hill (ed.) The 1967 Minnesota Symposium on Child Psychology (Vol. 2). Minneapolis: University of Minnesota Press. Brown, R. and Hanlon, C. (1970) Derivational complexity and order of acquisition in child speech. In R. Brown (ed.) Psycholinguistics. New York: Free Press. Bruner, J. (1975) The ontogenesis of speech acts. Journal of Child Language 2, 1–19. Bruner, J. (1978) From communication to language: A psychological perspective. In I. Markova (ed.) The Social Context of Language. London: Wiley. Bruner, J. (1983) Child’s Talk. New York: Norton. Bruner, J., Roy, C. and Ratner, N. (1982) The beginnings of request. In K.E. Nelson, (ed.) Children’s Language (Vol. 3; pp. 91–137). Hillsdale, NJ: Erlbaum. Byram, M. and Leman, J. (1989) (eds) Bicultural and Trilingual Education: The Foyer Model in Brussels. Clevedon: Multilingual Matters.

232 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

References

233

Cairns, H.S. and Hsu, J.R. (1978) Who, why, when and how: A developmental study. Journal of Child Language 5, 477–488. Capdevila I Batet, M. and Serrat I Sellabona, E. (1996) Adquisición de las Interrogativas QU [The acquisition of Wh-questions]. Sintagma 8, 17–31. Carter, A. (1974) Communication in the sensorimotor period. Doctoral dissertation, University of California, Berkeley. Cazden, C. (1970) Children’s questions: Their forms, functions and roles in education. Young Children March, 202–219. Cenoz, J. (1991) Enseñanza-aprendizaje del inglés como L2 o L3 [The Teaching and Learning of English as an L2 or L3]. Leioa: University of the Basque Country. Cenoz, J. (2000) Research on multilingual acquisition. In J. Cenoz and U. Jessner (eds) English in Europe: The Acquisition of a Third Language (pp. 39–53). Clevedon: Multilingual Matters. Cenoz, J. (2001) The effect of linguistic distance, L2 status and age on crosslinguistic influence in third language acquisition. In J. Cenoz, B. Hufeisen and U. Jessner (eds) Cross-linguistic Influence in Third Language Acquisition (pp. 12–33). Clevedon: Multilingual Matters. Cenoz, J. and Barnes, J. (1997) Early trilingualism: Basque, English and Spanish. ERIC Document Reproduction Service No. ED413761. Cenoz, J. and Genesee, F. (eds) (1998a) Psycholinguistic perspectives on multilingualism and multilingual education. In J. Cenoz and F. Genesee (eds) Beyond Bilingualism: Multilingualism and Multilingual Education. Clevedon: Multilingual Matters. Cenoz, J. and Genesee, F. (eds) (1998b) Beyond Bilingualism: Multilingualism and Multilingual Education. Clevedon: Multilingual Matters. Cenoz, J., Hufeisen, B. and Jessner, U. (2000) Towards trilingual education. International Journal of Bilingual Education 1–2, 1–19. Cenoz, J. and Jessner, U. (2000) English in Europe: The Acquisition of a Third Language. Clevedon: Multilingual Matters. Chomsky, N. (1957) Syntactic Structures. The Hague: Mouton. Chomsky, N. (1965) Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Chomsky, N. (1981) Lectures on Government and Binding. Dordrecht: Foris. Chomsky, N. (1982) Some Concepts and Consequences of the Theory of Government and Binding. Cambridge, MA: MIT Press. Chomsky, N. (1988) Some Notes on Economy of Derivation and Representation. Cambridge, MA: MIT Press. Clark, E. (1978) Awareness of language: Some evidence from what children say and do. In A. Sinclair, R. Jarvella and W.J. Levelt (eds) The Child’s Conception of Language (pp. 17–43). Berlin: Springer. Clyne, M. (1987) Grammatical constraints on code-switching: How universal are they? Linguistics 25, 739–764. Clyne, M. (1997) Some of the things trilinguals do. International Journal of Bilingualism 1 (2), 95–116. Comeau, L. and Genesee, F. (2001) Bilingual children’s repair strategies during dyadic communication. In J. Cenoz, and F. Genesee (eds) Trends in Bilingual Acquisition (pp. 231–256). Amsterdam: John Benjamins.

233 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

Early Trilingualism

234

Comeau, L., Genesee, F., Nicoladis, E. and Vrakas, G. (1997) Can young bilingual children identify their language choice as a cause of breakdown in communication? In E. Hughes, M. Hughes and A. Greenhill (eds) Proceedings of the 21st Annual Boston University Conference on Language Development (pp. 79–90). Somerville, MA: Cascadilla Press. Comrie, B. (1981) Language Universals and Linguistic Typology. Oxford: Blackwell. Cook, V. (1992) Evidence for multi-competence. Language Learning 42 (4), 557–591. Cook, V. (1995) Multi-competence and the learning of many languages. Language, Culture and Curriculum 8, 93–98. Cook, V. (1996) Competence and multi-competence. In G. Brown, K. Malmkjaer and J. Williams (eds) Performance and Competence in Second Language Acquisition (pp. 57–69). New York: Cambridge University Press. Corsaro, W. (1977) The clarification request as a feature of adult interactive styles with young children. Language in Society 6, 183–207. Corsaro, W. (1979) We’re friends, right? Children’s use of access rituals in a nursery school. Language in Society 8, 315–336. Crosby, F. (1976) Early discourse agreement. Journal of Child Language 3, 125–126. Cross, T. (1977) Mother’s speech adjustments: The contributions of selected child listener variables. In C. Ferguson and C. Snow (eds) Talking to Children: Language Input and Acquisition (pp. 151–188). Cambridge: Cambridge University Press. Cruz Ferreira, M. (1999) Prosodic mixes: Strategies in multilingual language acquisition. The International Journal of Bilingualism, 1 (3), 1–21. Crystal, D. (1974) Review of Roger Brown’s, ‘A First Language’. Journal of Child Language 1, 289–334. Cummins, J. (1976) The influence of bilingualism on cognitive growth: A synthesis of research findings and explanatory hypotheses. Working Papers on Bilingualism 9, 1–43. Darwin, C. (1877) A biographical sketch of an infant. Mind 2, 285–294. Davis, E.A. (1932) The form and function of children’s questions. Child Development 3, 57–74. De Houwer, A. (1987) Two at a time: An exploration of how children acquire two languages from birth. Doctoral dissertation, Vrije Universiteit Brussels. De Houwer, A. (1990) The Acquisition of Two Languages from Birth: A Case Study. Cambridge: Cambridge University Press. De Houwer, A. (1995) Bilingual language acquisition. In P. Fletcher and B. MacWhinney (eds) Handbook of Child Language (pp. 219–250). Cambridge, MA: Blackwell. De Houwer, A. and Meisel, J. (1996) Analyzing the relationship between two developing languages in bilingual first language acquisition: Methodology, data, findings. Paper presented at the Workshop on Language Contact: Linking Different Levels of Analysis. Wassenaar: Netherlands Institute for Advanced Study. De Jong, E. (1986) The Bilingual Experience: A Book for Parents. Cambridge: Cambridge University Press. De Villiers, J. (1991) Why questions? In T. Maxfield and B. Plunkett (eds) The Acquisition of Wh (pp. 155–174). Occasional Papers in Linguistics. University of Massachusetts.

234 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

References

235

De Villiers, J. (1995) Empty categories and complex sentences: The case of whquestions. In P. Fletcher and B. MacWhinney (eds) Handbook of Child Language (pp. 508–540). Cambridge, MA: Blackwell. Deuchar, M. (1989) Infant Bilingualism: One System or Two? (ESRC project no. C00232393). Deuchar, M. (1992) Can Government and Binding theory account for language acquisition? In C. Vide (ed.) Lenguajes Naturales y Lenguajes Formales [Natural Languages and Formal Languages] VIII (pp. 273–9). Barcelona: Universitat de Barcelona. Deuchar, M. (1999) Are function words non-language-specific in early bilingual two-word utterances? Bilingualism: Language and Cognition 2 (1), 23–34. Deuchar, M. and Clark, A. (1996) Early bilingual acquisition of the voicing contrast in English and Spanish. Journal of Phonetics 24, 351–365. Deuchar, M. and Quay, S. (1998) How early is it possible to establish one vs.two systems in bilingual syntactic development? Bilingualism: Language and Cognition 1 (3), 231–243. Deuchar, M. and Quay, S. (2000) Bilingual Acquisition: Theoretical Implications of a Case Study. Oxford: Oxford University Press. Dewaele, J. (1998) Lexical inventions: French interlanguage as L2 versus L3. Applied Linguistics 19 (4), 471–490. Dollaghan, C. and Miller, J. (1986) Observational methods in the study of communicative competence. In R.L. Schiefelbusch (ed.) Language Competence: Assessment and Intervention (pp. 99–129). London: Taylor and Francis. Döpke, S. (1992) One Parent–One Language: An Interactional Approach. Amsterdam: Benjamins. Döpke, S. (2000) Cross-linguistic Structures in Simultaneous Bilngualism. Amsterdam: John Benjamins. Dore, J. (1973) The development of speech acts. Doctoral dissertation, City University of New York. Dore, J. (1974) A pragmatic description of early language development. Journal of Psycholinguistic Research 3, 343–350. Dore, J. (1975) Holophrases, speech acts and language universals. Journal of Child Language 2, 21–40. Dore, J. (1977) Oh them sheriff: A pragmatic analysis of children’s responses to questions. In S. Ervin-Tripp and C. Mitchell-Kernan (eds) Child Discourse (pp. 139–163). New York: Academic Press. Dore, J. (1978) Conditions for the acquisition of speech acts. In I. Markova (ed.) The Social Context of Language (pp. 87–111). Chichester: Wiley. Dore, J. (1979) Conversational acts and the acquisition of language. In E. Ochs and B. Schieffelin (eds.) Developmental Pragmatics (pp. 339-361). NewYork: Academic Press. Dore, J. (1986) The development of conversational competence. In R.L. Schiefelbusch (ed.) Language Competence: Assessment and Intervention (pp. 3–60). London: Taylor and Francis. Dromi, E. (1987) Early Lexical Development. Cambridge: Cambridge University Press. Erreich, A. (1980) Why you won’t play with me? Non-inversion errors in wh-questions. Paper presented at the Fifth Annual Boston University Conference on Language Development, Boston.

235 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:27

Color profile: Disabled Composite Default screen

236

Early Trilingualism

Erreich, A. (1984) Learning how to ask: Patterns of inversion in yes-no and wh-questions. Journal of Child Language 11, 579–592. Erreich, A., Valian, V. and Winzemer, J. (1980) Aspects of a theory of language acquisition. Journal of Child Language 7, 157–179. Ervin-Tripp, S. (1970) Discourse agreement: How children answer questions. In J.R. Hayes (ed.) Cognition and Development of Language (pp. 79–107). New York: Wiley. Ervin-Tripp, S. (1977) Wait for me, roller skate. In S. Ervin-Tripp and C. MitchellKernan (eds) Child Discourse (pp. 165–188). New York: Academic Press. Ervin-Tripp, S. (1988) Request retries. Lenguas Modernas 15, 25–34. Universidad de Chile. Ervin-Tripp, S. and Gordon, D. (1986) The development of requests. In R.L. Schiefelbusch (ed.) Language Competence: Assessment and Intervention (pp. 61–95). London: Taylor and Francis. Ezeizabarrena, M.J. (1996) Adquisición de la morfología verbal en euskera y castellano por niños bilingües [Acquisition of Verbal Morphology in Basque and Spanish by Bilingual Children]. Leioa: University of the Basque Country. Faingold, E. (1999) The re-emergence of Spanish and Hebrew in a multilingual adolescent. International Journal of Bilingual Education and Bilingualism 4 (2), 283– 296. Fantini, A. (1985) Language Acquisition of a Bilingual Child. Clevedon: Multilingual Matters. Fletcher, P. (1985) A Child’s Learning of English. Oxford: Basil Blackwell. Fleta, M.T. (1999) La adquisición del inglés no nativo por niños: el desarrollo de la cláusula [The acquisition of non-native English by children: Clausal development]. Doctoral dissertation, Universidad Complutense de Madrid. Fraser, B. (1976) On requesting: An essay in pragmatics. Unpublished manuscript, Boston University. Fraser, B. (1985) On the universality of speech act strategies. In S.George (ed.) From the Linguistic to the Social Context (pp. 43–49). Bologna, Italy: CLUEB. Furrow, D., Nelson, K., and Benedict, H. (1979) Mothers’ speech to children and syntactic development: Some simple relationships. Journal of Child Language 6, 423–442. García, E. (1983) Early Childhood Bilingualism. Albuquerque: University of New Mexico Press. Garvey, C. (1974) Some properties of social play. Merrill-Palmer Quarterly 20, 163– 180. Garvey, C. (1975) Requests and responses in children’s speech. Journal of Child Language 2, 41–59. Garvey, C. (1977) The contingent query: A dependent act in conversation. In M. Lewis and L. Rosenblum (eds) Interaction, Conversation and the Development of Language: the Origin of Behaviour (Vol. 5). New York: Wiley. Gass, S. (1997) Input, Interaction and the Second Language Learner. Mahwah, NJ: Lawrence Erlbaum. Gawlitzek-Maiwald, I. and Tracy, R. (1996) Bilingual bootstrapping. Linguistics 34, 901–926. Genesee, F. (1989) Early bilingual development: One language or two? Journal of Child Language 16, 161–79.

236 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:08

Color profile: Disabled Composite Default screen

References

237

Genesee, F., Boivin, I. and Nicoladis, E. (1996) Talking with strangers: A study of bilingual children´s communicative competence. Applied Psycholinguistics 17, 427–442. Genesee, F., Nicoladis, E. and Paradis, J. (1995) Language differentiation in early bilingual development. Journal of Child Language 22, 611–631. Golinkoff, R. (1986) ‘I beg your pardon?’ The preverbal negotiation of failed messages. Journal of Child Language 13, 455–476. Goodz, N. (1989) Parental language mixing in bilingual families. Infant Mental Health Journal 1 (10), 25–43. Gordon, D. and Ervin-Tripp, S. (1984) The structure of children’s requests. In R.L. Schiefelbusch and J. Pickar (eds) The Acquisition of Communicative Competence (pp. 295–321). Baltimore, MD: University Park Press. Greenbaum, S. and Quirk, R. (1990) A Students’ Grammar of the English Language. Harlow: Longman. Grosjean, F. (1982) Life with Two Languages: An Introduction to Bilingualism. Cambridge, MA: Harvard University Press. Grosjean, F. (1985) The bilingual as a competent but specific speaker-hearer. Journal of Multilingual and Multicultural Development 6, 467–477. Grosjean, F. (1988) Exploring the recognition of guest words in bilingual speech. Language and Cognitive Processes 3 (3), 233–274. Grosjean, F. (1989) Neurolinguists, beware! The bilingual is not two monolinguals in one person. Brain and Language 36, 3–15. Grosjean, F. (1992) Another view of bilingualism. In R.J. Harris (ed.) Cognitive Processing in Bilinguals (pp. 51–52). Amsterdam: North Holland. Grosjean, F. (1994) Individual bilingualism. The Encyclopedia of Language and Linguistics (pp. 1656–1660). Oxford: Pergamon. Grosjean, F. (1997) Processing mixed language: Issues, findings, and models. In A. de Groot and J. Kroll (eds) Tutorials in Bilingualism: Psycholinguistic Perspectives (pp. 225–54). Mahwah, NJ: Erlbaum. Grosjean, F. (1998) Studying bilinguals: Methodological and conceptual issues. Bilingualism: Language and Cognition 2 (1), 131–49. Grosjean, F. (2001) The bilingual´s language modes. In J.L. Nicol (ed.) Language Processing in the Bilingual. Oxford: Blackwell. Gut, Ulrike. (2000) Cross-linguistic structures in the acquisition of intonational phonology by German-English bilingual children. In S. Döpke (ed.) Crosslinguistic Structures in Simultaneous Bilingualism (pp. 201–226). Amsterdam: John Benjamins. Hakuta, K. (1986) Mirror of Language: The Debate on Bilingualism. New York: Basic Books. Halliday, M.A.K. (1975) Learning How to Mean. London: Edward Arnold. Hamers, J. and Blanc, M. (1989) Bilinguality and Bilingualism. Cambridge: Cambridge University Press. Hammarberg, B. and Williams, S. (1993) A study of third language acquisition. In B. Hammarberg (ed.) Problem, Process, Product in Language Learning (pp. 60–70). Stockholm: Stockholm University. Harding, E. and Riley, P. (1986) The Bilingual Family: A Handbook for Parents. Cambridge: Cambridge University Press. Hayashi, M. (1992) Lexical development and early translation equivalents in bilingual children. Poster presented at the Child Language Seminar, Glasgow.

237 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:08

Color profile: Disabled Composite Default screen

238

Early Trilingualism

Helot, C. (1988) Bringing up children in English, French and Irish: Two case studies. Language, Culture and Curriculum 1 (3), 281–287. Hernández Pina, F. (1984) Teorías psicolingüísticas y su aplicación al español como lengua materna [Psychocolinguistic Theories and their Application to Spanish as a Mother Tongue]. Madrid: Siglo XXI. Hoffmann, C. (1985) Language acquisition in two trilingual children. Journal of Multilingual and Multicultural Development 6 (6), 479–495. Hoffmann, C. (1991) An Introduction to Bilingualism. London: Longman. Hoffmann, C. (2001) Towards a description of trilingual competence. The International Journal of Bilingualism 5 (1), 1–17. Hoffmann, C. and Widdicombe, S. (1999) Code-switching and language dominance in the trilingual child. AILE Proceedings of 8th EUROSLA Conference Paris 1 (Special Issue) (pp. 51–62). Hollingsworth, H. (1928) Mental Growth and Decline. New York: Appleton and Co. Holzman, M. (1972) The use of interrogative forms in the verbal interaction of three mothers and their children. Journal of Psycholinguistic Research 1 (4), 311–336. Hulk, A. (2000) Non-selective access and activation in child bilingualism: The syntax. In S. Döpke (ed.) Cross-linguistic Structures in Simultaneous Bilingualism (pp. 57–78). Amsterdam: John Benjamins. Hulk, A. and Müller, N. (2000) Bilingual first language acquisition at the interface between syntax and pragmatics. Bilingualism: Language and Cognition 3 (3), 227– 244. Hulk, A. and Van der Linden, E. (1998) Evidence for transfer in bilingual children? Bilingualism: Language and Cognition 3 (1), 177–180. Hurford, J. (1975) A child and the English question formation rule. Journal of Child Language 2, 299–301. Hymes, D. (1972) Towards Communicative Competence. Philadelphia: University of Pennsylvania Press. Idiazábal, I. (1984) Conciencia bilingüe del niño bilingüe [Bilingual awareness of the bilingual child]. In M. Siguan (ed.) Adquisición precoz de una segunda lengua [The Early Acquisition of a Second Language] (pp. 55–63). Barcelona: Publicacions i edicions de la Universitat de Barcelona. Idiazábal, I. (1988) First verbal productions of a bilingual child learning Basque and Spanish simultaneously: Analysis of the noun phrase. Unpublished manuscript. Idiazábal, I. (1991) Evolución de la determinación nominal en un niño bilingüe vasco-hispanófono [Development of the nominal determiner in a BasqueSpanish bilingual child]. Paper presented at the 1st International Conference on Spanish in Contact with Other Languages, University of Southern California, Los Angeles. Ingram, D. (1981) The emerging phonological system of an Italian-English bilingual child. Journal of Italian Linguistics 2, 95–113. Ingram, D. (1989) First Language Acquisition. Cambridge: Cambridge University Press. Ingram, D. and Tyack, D. (1979) Inversion of subject NP and Aux in children’s questions. Journal of Psycholinguistic Research 8 (4), 333–341. James, S. and Seebach, M. (1982). The pragmatic function of children’s questions. Journal of Speech and Hearing Research 25, 2–11.

238 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2 changes a.vp 20 November 2005 22:21:08

Color profile: Disabled Composite Default screen

References

239

Jekat, S. (1985) Die Entwicklung des Wortschatzes bei bilingualen Kindern (Frz-Dt.) in den ersten vier Lebensjahren. [The development of the lexicon in bilingual children (French-German) in the first four years of life.] Masters dissertation, University of Hamburg. Jessner, U. (1997) Towards a dynamic view of multilingualism. In M. Pütz (ed.) Language Choices, Conditions, Constraints and Consequences (pp. 17–30). Amsterdam: John Benjamins. Juan-Garau, M. (1996) Language Development in a Catalan–English Bilingual Between the Ages of 1 and 3. Barcelona: PPU. Juan-Garau, M. and Pérez-Vidal, C. (2000) Subject realisation in the syntactic development of a bilingual child. Bilingualism: Language and Cognition 3 (3), 173–192. Juan-Garau, M. and Pérez-Vidal, C. (2001) Mixing and pragmatic parental strategies in early bilingual acquisition. Journal of Child Language 28, 59–86. Kasper, G. and Blum-Kulka, S. (1993) Interlanguage Pragmatics. Oxford: Oxford University Press. Kasuya, H. (1998) Determinants of language choice in bilingual children: The role of input. International Journal of Bilingualism 2 (3), 327–346. Keenan, E. (1977) Making it last: Repetition in children’s discourse. In S. ErvinTripp and C. Mitchell-Kernan (eds) Child Discourse (pp. 125–138). New York: Academic Press. Kielhöfer, B. (1987) Le ‘bon’ changement de langue et le ‘mauvais’ mélange de langues [Good code-switching and bad code-mixing]. In G. Lüdi (ed.) Devenir bilingüe-parler bilingüe [Becoming Belingual, Speaking Bilingual] (pp. 135–55). Tübingen: Niemeyer. Klee, M. (1985) Role of inversion in children’s question development. Journal of Speech and Hearing Research 28, 225–232. Klima, E. and Bellugi, U. (1966) Syntactic regularities in the speech of children. In J. Lyons and J. Wales (eds) Psycholinguistic Papers (pp. 183–208). Edinburgh: Edinburgh University Press. Kobashigawa, B. (1969) Repetitions in a mother’s speech to her child. Working paper, University of California, Berkeley. Köppe, R. (1997) Sprachentrennung im Frühen bilingualen Erstsprachenerwerb Französisch/Deustsch [Language Separation in Early Bilingual First Language Acquisition: French and German]. Tübingen: Verlag. Kuczaj, S. and Brannick, N. (1979) Children’s use of the wh-question modal auxiliary placement rule. Journal of Experimental Child Psychology 6, 563–579. Kuczaj, S. and Maratsos, M. (1983) Initial verbs of yes-no questions: A different kind of grammatical category. Developmental Psychology, 19. Labov, W. and Labov T. (1978) Learning the syntax of questions. In R.N. Campbell and P.T. Smith (eds) Recent Advances in the Psychology of Language (Vol.1; pp. 1– 44). London: Plenum Press. Lambert, W. (1975) Culture and language as factors in learning and education. In A. Wolfgang (ed.) Education of Immigrant Students. Toronto: Ontario Institute for Studies in Education. Lanza, E. (1990) Language mixing in infant bilingualism: A sociolinguistic perspective. Doctoral dissertation, Georgetown University. Lanza, E. (1992) Can bilingual two-year-olds code-switch? Journal of Child Language 19, 633–58.

239 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

Early Trilingualism

240

Lanza, E. (1997) Language Mixing in Infant Bilingualism: A Sociolinguistic Perspective. Oxford: Clarendon Press. Lanza, E. (2000) Language contact: A dilemma for the bilingual child or for the linguist? In S. Döpke (ed.) Cross-linguistic Structures in Simultaneous Bilingualism (pp. 227–246). Amsterdam: John Benjamins. Lanza, E. (2001) Bilingual first language acquisition: A discourse perspective on language contact in parent–child interaction. In J.Cenoz and F. Genesee (eds) Trends in Bilingual Acquisition (pp. 201–230). Amsterdam: John Benjamins. Lasagabaster, D. (1997) Creatividad y conciencia metalingüística: Incidencia en el aprendizaje del inglés como L3 [Creativity and Metalinguistic Awareness: The Learning of English as an L3]. Leioa: University of the Basque Country. Lebrun, Y. and Paradis, M. (1984) To be or not to be an early bilingual? In Y. Lebrun and M. Paradis (eds) Early Bilingualism and Child Development (pp. 9–18). Amsterdam: Swets and Zeitlinger. Ledbetter, P. and Dent, C. (1988) Young children’s sensitivity to direct and indirect request structure. First Language 8, 227–246. Leech, G. (1983) Principles of Pragmatics. London: Longman. Leopold, W. (1970) Speech Development of a Bilingual Child: A Linguist’s Record (4 vols). New York: AMS Press. (Original work published in 1939–1949.) Lewis, M.M. (1938) The beginning and early function of questions in a child’s speech. British Journal of Educational Psychology 8, 150–171. Lightbown, P. and Spada, N. (1999) How Languages are Learned (2nd edn). Oxford: Oxford University Press. Lindholm, K. and Padilla, A. (1978) Language mixing in bilingual children. Journal of Child Language 5, 327–335. Long, M.H. (1996) The role of the linguistic environment in second language acquisition. In C. Ritchie and T.K. Bhatia (eds) Handbook of Language Acquisition (Vol. 2): Second Language Acquisition (pp. 413–468). New York: Academic Press, Lucariello, J., Kyratis, A. and Engel, S. (1986) Event representations, context and knowledge. In K. Nelson (ed.) Event Knowledge (pp. 137–159). Hillsdale, NJ: Erlbaum. Mackey, A. and Oliver, R. (2002) Instructional feedback and children’s L2 development. System, 30 (4), 460–477. Mackey, A., Oliver, R. and Leeman, J. (2003) Interactional input and the incorporation of feedback. An exploration of NS-NNS and NNN-NNS adult and child dyads. Language Learning 53 (1), 35–66. Macnamara, J. (1972) The cognitive basis of language learning in infants. Psychological Review 79, 1–13. MacWhinney, B. (1974) How Hungarian children learn to speak. Doctoral dissertation, University of California, Berkeley. MacWhinney, B. (1991) The CHILDES Project: Computational Tools for Analyzing Talk. Hillsdale, NJ: Erlbaum. MacWhinney, B. (1995) The CHILDES Project: Computational Tools for Analyzing Talk (2nd edn). Hillsdale, NJ: Erlbaum. Mannle, S., Barton, M. and Tomasello, M. (1991) Two-year-olds’ conversations with their mothers and preschool-aged siblings. First Language 12, 57–71. Maratsos, M.P. (1979) How to get from words to sentences. In D. Aaronson and R. Reiber (eds) Perspectives in Psycholinguistics. Hillsdale, NJ: Erlbaum.

240 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

References

241

Maratsos, M., Kuczaj, S., Fox, D. and Chalkley, M. (1979) Some empirical studies in the acquisition of transformational relations: Passives, negatives and the past tense. In W.A. Collins (ed.) The Minnesota Symposium on Child Psychology (Vol. 12). Minneapolis: University of Minnesota Press. Marcon, R. and Coon, R. (1983) Communication styles of bilingual preschoolers in preferred and non-preferred languages. Journal of Genetic Psychology 142, 189– 202. Marcos, H. (1991) Reformulating requests at 18 months: Gestures, vocalizations and words. First Language 11, 361–375. McCarthy, D. (1930) The language development of the pre-school child. Institute of Child Welfare Monograph Series, 4. Minneapolis: University of Minnesota Press. McLaughlin, B. (1978) Second Language Acquisition in Childhood. Hillsdale, NJ: Erlbaum. McLaughlin, B. (1984) Early bilingualism: Methodological and theoretical issues. In M. Paradis and Y. Lebrun (eds) Early Bilingualism and Child Development (pp. 19– 45). Lisse: Swets and Zeitlinger. McShane, J. (1980) Learning to Talk. Cambridge: Cambridge University Press. Meisel, J. M. (1986) Word order and case marking in early child language: Evidence from simultaneous acquisition of two first languages: French and German. Linguistics 24, 123–183. Meisel, J. (1989) Early differentiation of languages in bilingual children. In K. Hyltenstam and L. Obler (eds) Bilingualism Across the Lifespan: Aspects of Acquisition, Maturity and Loss (pp. 13–40). Cambridge: Cambridge University Press. Meisel, J. (1990) INFL-ection: Subjects and subject verb agreement. In J. Meisel (ed.) Two First Languages: Early Grammatical Development in Bilingual Children (pp. 237– 298). Dordrecht: Foris. Meisel, J. (1994) Code-switching in young bilingual children. The acquisition of grammatical constraints. Studies in Second Language Acquisition 16, 413–439. Meisel, J. (2001) The simultaneous acquisition of two first languages: Early differentiation and subsequent development of grammars. In J. Cenoz and F. Genesee (eds) Trends in Bilingual Acquisition (pp. 11–41). Amsterdam: John Benjamins. Meisel, M. and Mahlau, A. (1988) La adquisición simultanea de dos primeras lenguas. Discusión general e implicaciones para el estuudio del bilingüísmo en euzkadi [The simultaneous acquisition of two first languages. General discussion and implications for the study of bilingualism in the Basque Country – Euskadi]. Actas del II Congreso Mundial Vasco: Congreso sobre la Lengua Vasca, Tomo III. Vitoria: Servicio de publicaciones del Gobierno Vasco. Menyuk, P. (1969) Sentences Children Use. Cambridge, MA.: MIT Press. Meyer, W.J. and Shane, J. (1973) The form and functions of children’s questions. Journal of Genetic Psychology 123, 285–296. Mikès, M. (1990) Some issues of lexical development in early bi-and trilinguals. In G. Conti-Ramsden and C. Snow (eds) Children’s Language (Vol. 7; (pp. 103–20). Hillsdale, NJ: Erlbaum. Miller, J. (1981) Assessing Language Production in Children: Experimental Procedures. London: Edward Arnold. Miller, J. and Chapman, R. (1985) System Analysis of Language Transcripts. University of Wisconsin. Murrell, M. (1966) Language acquisition in a trilingual environment: Notes from a case study. Studia Lingüística 20, 9–35.

241 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

Early Trilingualism

242

Myers-Scotton, C. (1993) Social Motivations for Codeswitching. Oxford: Oxford University Press. Navracsics, J. (1999) The acquisition of Hungarian by trilingual children. Doctoral dissertation, University of Vesprém, Vesprém-Pécs. Nelson, K. (1981) Individual differences in language development: Implications for development and language. Developmental Psychology 17 (2), 170–187. Newcombe, N. and Zaslow, M. (1981) Do 2½-year-olds hint? A study of directive forms in the speech of 2½-year-old children to adults. Discourse Processes 4, 239– 252. Nicoladis, E. (1998) First clues to the existence of two input languages: Pragmatic and lexical differentiation in a bilingual child. Bilingualism: Language and Cognition 1, 105–16. Nicoladis, E. and Genesee, F. (1996a) A longitudinal study of pragmatic differentiation in young bilingual children. Language Learning 46 (3), 439–64. Nicoladis, E. and Genesee, F. (1996b) The role of parental input and language dominance in bilingual children’s codemixing. In E. Hughes, M. Hughes and A. Greenhill (eds) Proceedings of the 22nd Annual Boston University Conference on Language Development (pp. 422–32). Somerville, MA: Cascadilla Press. Nicoladis, E. and Genesee, F. (1998) Parental discourse and codemixing in bilingual children. International Journal of Bilingualism 1 (2), 85–99. Ninio, A. and Bruner, J. (1978) The achievement and antecedents of labeling. Journal of Child Language 5, 1–15. Ninio, A. and Snow, C. (1996) Pragmatic Development. Oxford: Westview Press. Ninio, A. and Wheeler, P. (1984) A manual for classifying verbal communicative acts in mother–infant interaction. Working Papers in Developmental Psychology, number 1. Jerusalem: The Martin and Vivian Levin Center, Hebrew University. (Reprinted in Transcript Analysis (1986) 3, 1–82.) Ochs, E. (1988) Culture and Language Development: Language Acquisition and Language Socialization in a Samoan Village. Cambridge: Cambridge University Press. Ochs-Keenan, E. (1977) Making it last: Repetition in children’s discourse. In S. Ervin-Tripp and C. Mitchell Kernan (eds) Child Discourse (pp. 125–131). New York: Academic Press. Oksaar, E. (1978) Preschool trilingualism: A case study. In F.C. Peng and W. RafflerEngel (eds) Language Acquisition and Developmental Kinesics (pp. 129–38). Hiroshima: Bunka Hyoron Publishing. Oliver, R. (2000) Age effects in negotiation and feedback in classroom and pairwork. Language Learning 50 (1), 119–151. Olsen-Fulero, L. and Conforti, J. (1983) Child responsiveness to mother questions of varying type and presentation. Journal of Child Language 10, 495–520. Padilla, A. and Lindholm, K. (1984) Child bilingualism: The same old issues revisited. In J.L. Martínez and R. Mendoza (eds) Chicano Psychology (2nd edn; pp. 369– 408). New York: Academic Press. Paradis, J. (2000) Beyond ‘one system or two?’: Degrees of separation between the languages of French-English bilingual children. In S. Döpke (ed.) Cross-linguistic Structures in Simultaneous Bilingualism (pp. 175–200). Amsterdam: John Benjamins. Paradis, J. and Genesee, F. (1996) Syntactic acquisition in bilingual children: Autonomous or interdependent? Studies in Second Language Acquisition 18, 1–25.

242 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

References

243

Paradis, J., Nicoladis, E. and Genesee, F. (2000) Early emergence of structural constraints on code-mixing: Evidence from French-English bilingual children. Bilingualism: Language and Cognition 3 (3), 245–262. Parnell, M., Paterson, S. and Harding, M. (1984) Answers to wh-questions: A developmental study. Journal of Speech and Hearing Research 27, 297–305. Pérez-Vidal, C. (1995) La adquisición del inglés de un niño bilingüe catalán-inglés [The acquisition of English by a Catalan-English bilingual child]. Doctoral dissertation, University of Barcelona. Pfaff, C. (1991) Turkish in contact with German: Language maintenance and loss among immigrant children in Berlin (West). International Journal for the Sociology of Language 90, 97–129. Piaget, J. (1926) The Language and Thought of the Child. London: Routledge and Kegan Paul. Pica, T. (1994) Research on negotiation: What does it reveal about second language learning conditions, processes and outcomes? Language Learning 44, 493–527. Platt, J. (1988) What can case studies do? Studies in Qualitative Methodology 1, 1–23. Poplack, S. (1980) Sometimes I’ll start a sentence in Spanish y termino en español: Toward a typology of codeswitching. Linguistics 18, 581–618. Poplack, S. (1981) Syntactic structure and social function of code-switching. In R. Durán (ed.) Latino Language and Communicative Behaviour (pp. 169–84). Norwood, NJ: Ablex. Pratt, C. and Grieve, R. (1984) The development of metalinguistic awareness: An introduction. In W.E. Tunmer, C. Pratt and M.L. Herriman (eds) Metalinguistic Awareness in Children: Theory, Research and Implications. Berlin: Springer-Verlag. Prinz, P. (1982) ‘Requesting’ in normal and language-disordered children. In K. Nelson (ed.) Children’s Language (Vol.3; pp. 139–203). Hillsdale, NJ: Erlbaum. Przetacnik-Gierowska, M. and Ligeza, M. (1990) Cognitive and interpersonal functions of children’s questions. In G. Conti-Ramsden and C. Snow (eds) Children’s Language (Vol. 7; pp. 69–101). Hillsdale, NJ: Erlbaum. Quay, S. (1995) The bilingual lexicon, implications for studies of language choice. Journal of Child Language 22, 369–387. Quay, S. (2001) Managing linguistic boundaries in early trilingual development. J. Cenoz and F. Genesee (eds) Trends in Bilingual Acquisition (pp. 149–200). Amsterdam: John Benjamins. Radford, A. (1990) Syntactic Theory and the Acquisition of English Syntax: The Nature of Early Child Grammars of English. Oxford: Blackwell. Radford, A. (1994) The syntax of questions in child English. Journal of Child Language 21, 211–236. Redlinger, W. and Park, T. (1980) Language mixing in young bilinguals. Journal of Child Language 7, 337–352. Remick, H. (1972) Maternal speech to children during language acquisition. Paper read at the International Symposium on First Language Acquisition, Florence. Richards, B. (1987) Type/Token Ratios: What do they really tell us? Journal of Child Language 14, 201–209. Ringbom, H. (1987) The Role of the First Language in Foreign Language Learning. Clevedon: Multilingual Matters. Romaine, S. (1989) Bilingualism. Oxford: Blackwell. Ronjat, J. (1913) Le développement du langage observé chez un enfant bilingüe [Language Development in a Bilingual Child]. Paris: Champion.

243 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

Early Trilingualism

244

Rowland, C. and Pine, J. (2000) Subject-auxiliary inversion errors and wh-question acquisition: ‘What children do know?’ Journal of Child Language 27, 157–181. Rowland, C. and Pine, J. (2003) The development of inversion in wh-questions: A reply to Van Valin. Journal of Child Language 30, 197–212. Sachs, J., Brown, R. and Salerno, P. (1972) Adults’ speech to children. Paper read at the International Symposium on First Language Acquisition, Florence. Sachs J. and Devin J. (1976) Young children’s use of age-appropriate speech styles in social interaction and role-playing. Journal of Child Language 3, 81–98. Sagasta, P. (2000) La producción escrita en euskara, castellano e inglés en el modelo D y en el modelo de inmersión [Writing in Basque, Spanish and English in D model and the immersion model]. Doctoral dissertation, University of the Basque Country. Saltarelli, M. (1988) Basque. London: Routledge. Santelmann, L., Berk, S., Austin, J., Somashekar, S. and Lust, B. (2002) Continuity and development in the acquisition of inversion in yes-no questions: Dissociating movement and inflection. Journal of Child Language 29, 813–842. Saunders, G. (1982) Bilingual Children: Guidance for the Family. Clevedon: Multilingual Matters. Saunders, G. (1988) Bilingual Children: From Birth to Teens. Clevedon: Multilingual Matters. Schelletter, C. (2000) Negation as a cross-linguistic structure in a German-English bilingual child. In S. Döpke (ed.) Cross-linguistic Structures in Simultaneous Bilingualism (pp. 105–122). Amsterdam: John Benjamins. Schieffelin, B. (1979) Getting it together: An ethnographic approach to the study of the development of communicative competence. In E. Ochs and B. Schieffelin (eds) Developmental Pragmatics. New York: Academic Press. Schlyter, S. (1990) The acquisition of tense and aspect. In J. Meisel (ed.) Two First Languages: Early Grammatical Development in Bilingual Children (pp. 87–122). Dordrecht: Foris Publications. Searle, J.R. (1969) Speech Acts. Cambridge: Cambridge University Press. Seale, J.R. (1977) A classification of illocutionary acts. Language in Society 5, 1–23. Sharwood Smith, M. (1994) Second Language Learning: Theoretical Foundations. London: Longman. Shatz, M. (1978) Children’s comprehension of their mothers’ question directives. Journal of Child Language 5, 39–46. Shatz, M. (1979) How to do things by asking: Form–function pairings in mothers’ questions and their relationship to children’s responses. Child Development 50, 1093–1099. Shatz, M. (1983) Communication. In J. Flavell and E. Markman (eds) Handbook of Child Psychology (Vol. 3): Cognitive Development (4th edn; pp. 841–889). New York: Wiley. Shatz, M. and Gelman, R. (1973) The development of communication skills: Modification in the speech of young children as a function of listeners. Monographs of the Society for Research in Child Development 38 (5) no. 152. Shatz, M. and Watson, (1990) Conversational or communicative skill? A reassessment of two-year-olds’ behaviour in miscommunication episodes. Journal of Child Language 17, 131–146. Sinclair, J. and Coulthard, R. (1975) Towards an Analysis of Discourse: The English Used by Teachers and Pupils. London: Oxford University Press.

244 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

References

245

Singh, R. and Carroll, S. (1979) L1, L2 and L3. Indian Journal of Applied Linguistics 5, 51–63. Singleton, D. (1987) Mother and other tongue influence on learner French: A case study. Studies in Second Language Acquisition 3 (9), 327–45. Slobin, D. (1985) (ed.) The Crosslinguistic Study of Language Acquisition (Vols 1–2). Hillsdale, NJ: Erlbaum. Smith, M. (1926) An investigation of the development of the sentence and the extent of vocabulary in young children. University of Iowa Studies in Child Welfare 3, (5). Smith, M. (1933) The influence of age, sex and situation on the frequency, form and function of questions asked by pre-school children. Child Development 4, (3) 201– 213. Snow, C. (1972) Mother’s speech to children learning languages. Child Development 43, 549–565. Snow, C. (1977) The development of conversation between mothers and babies. Journal of Child Language 4, 1–22. Snow, C. (1995) Issues in the study of input: Fine-tuning, universality, individual and developmental differences, and necessary causes. In P. Fletcher and B. MacWhinney (eds) Handbook of Child Language. Cambridge, MA: Blackwell. Snow, C., Arlmann-Rupp, A., Hassing, Y., Jobse, J., Joosten, J. and Vorster, J. (1976) Mother’s speech in three social classes. Journal of Psycholinguistic Research 5, 1–20. Snow, C., Dubber, C. and De Blauw, A. (1981) Routines in parent–child interaction. In L. Feagans and D. Farran (eds) The Language of Children Reared in Poverty: Implications for Evaluation and Intervention. New York: Academic Press. Snow, C. and Goldfield, B. (1983) Turn the page please: Situation-specific language learning. Journal of Child Language 10, 551–570. Snow, C., Pan, B., Imbens-Bailey, A. and Herman, J. (1996) Learning how to say what one means: A longitudinal study of children’s speech act use. Social Development 5, 56–84) Snyder, A. (1914) Notes on the talk of a two-and-one-half year old boy. Pedagogic Seminar 11, 412–424. Stavans, A. (1990) Code-switching in children acquiring English, Spanish and Hebrew: A case study. Doctoral dissertation. University of Pittsburgh. Stavans, A. (1992) Sociolinguistic factors affecting codeswitches produced by trilingual children. Language, Culture and Curriculum 5 (1), 41–53. Steffensen, M. (1977) Satisfying inquisitive adults: Some simple methods of answering yes-no questions. Journal of Child Language 5, 221–236. Stern, W. (1924) Psychology of Early Childhood. New York: Henry Holt. Stromswold, K. (1988) Linguistic representations of children’s wh-questions. Papers and Reports on Child Language (Stanford University) 27, 107–114. Stromswold, K. (1995) The acquisition of subject and object wh-questions. Language Acquisition 4 (1–2), 5–48. Stromswold, K. and Pinker, S. (1986) Questions about wh-questions. Paper presented at the 11th Annual Boston University Conference on Language Development, Boston. Stubbs, M. (1983) Discourse Analysis: The Sociolinguistic Analysis of Natural Language. London: Blackwell. Sully, J. (1896) Studies of Childhood. New York: Appleton. Swain, M. (1972) Bilingualism as a first language. Doctoral dissertation, University of California, Irvine.

245 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

Early Trilingualism

246

Swain, M. and Wesche, M. (1975) Linguistic interaction: Case study of a bilingual child. Language Sciences 37, 17–22. Taeschner, T. (1983) The Sun is Feminine: A Study on Language Acquisition in Bilingual Children. Berlin/Heidelberg: Springer. Thornton, R. (1990) Adventures in long-distance moving: The acquisition of complex wh-questions. Doctoral dissertation, University of Connecticut, Storrs. Tokuhama-Espinosa, T. (2001) Raising Multilingual Children: Foreign Language Acquisition and Children. Westport, CT: Bergin & Garvey Tough, J. (1977) The Development of Meaning. New York: Halsted Press. Tracy, R. (2000) Language mixing as a challenge for linguistics. In S. Döpke (ed.) Cross-linguistic Structures in Simultaneous Bilingualism (pp. 11–36). Amsterdam: John Benjamins. Trettien, A. (1904) Psychology of language interest of children. Pedagogic Seminar 11, 13–177. Turrell, M. T. (1997) Our trilingual experience: A Catalan/English/Spanish trilingual family in Catalonia. The Bilingual Family Newsletter 14 (4), 1, 2 and 7. Tyack, D. and Ingram D. (1977) Children’s production and comprehension of questions. Journal of Child Language 4, 211–223. Ultan, R. (1978) Some general characteristics of interrogative systems. In J. Greenberg (ed.) Universals of Human Language (Vol. 4). Stanford, California: Stanford University Press. Valian, V. and Casey, L. (2003) Young children’s acquisition of wh-questions: The role of structured input. Journal of Child Language 30, 117–143. Valian, V., Lasser, I. and Mandelbaum, D. (1992) Competing analyses of children’s early questions. Paper presented at the 17th Annual Boston University Conference on Language Development. Van der Linden, E. (2000) Non-selective access and activation in child bilingualism: The lexicon. In S. Döpke (ed.) Cross-linguistic Structures in Simultaneous Bilingualism (pp. 37–56). Amsterdam: John Benjamins. Van Valin, R. (2002) The development of subject auxiliary inversión in English whquestions: An alternative analysis. Journal of Child Language 29, 161–175. Vazquez, I. (1995) A Contrastive Study of Politeness Phenomena in England and Spain. University of Duisberg, Series B, no. 267. Vihman, M. (1981) Language mixing and code-switching: Conversations between a pair of bilingual siblings. Paper presented at the Second International Congress for the Study of Child Language, Vancouver. Vihman, M. (1985) Language differentiation by the bilingual infant. Journal of Child Language 12, 297–324. Volterra, V. and Taeschner, T. (1978) The acquisition and development of language by bilingual children. Journal of Child Language 5, 311–26. Vygotsky, L. (1978) Mind in Society. Cambridge: Harvard University Press. Wells, G. (1985) Language Development in the Pre-school Years. Cambridge: Cambridge University Press. Werner, H. and Kaplan, B. (1963) Symbol Formation. New York: Wiley. Widdicome, S. (1997). Code-switching, coining and interference in trilingual first language acquisition: A case study. MSc dissertation, University of Aston. Wierbizcka, A. (1985) Different cultures, different languages, different speech acts. Journal of Pragmatics 9, 145–178.

246 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Color profile: Disabled Composite Default screen

References

247

Williams, S. and Hammarberg, B. (1998) Language switches in L3 production: Implications for a polyglot speaking model. Applied Linguistics 19, 295–333. Wolfson, N., Marmor, T. and Jones, S. (1989) Problems in the comparison of speech acts across cultures. In S. Blum-Kulka, J. House and G. Kasper (eds) Crosscultural Pragmatics: Requests and Apologies (pp. 174–196). New Jersey: Ablex. Yip, V. and Matthews, S. (2000) Syntactic transfer in a Cantonese-English bilingual child. Bilingualism: Language and Cognition 3 (3), 193–208. Zubiri, I. and Zubiri, E. (1995) Euskal gramatika osoa [The Complete Basque Grammar]. Bilbao: Didaktiker.

247 W:\Clients\MLM\Barnes\Proof2\Barnes Proof 2.vp 04 October 2005 18:55:28

Author Index Aksu, A. 63 Arnberg, L. 30 Austin, J. 52, 53 Babelot, G. 70, 90 Bain, B. 18 Barnes, J. 28, 31, 93, 94, 215 Barreña, A. 79, 80, 81, 84 Barron-Hauwaert, S. 30, 31 Bates, E. 25, 52, 53, 55, 56, 63, 69 Becker, J. 31 Bellugi, U. 39, 40, 42, 45, 46, 48, 89, 127-128, 139, 147, 148, 161, 207, 210, 211, 212 Ben Zeev, S. 25 Bentahila, A. 20 Bergman, C. 14, 19 Bialystok, E. 52 Blanc, M. 31 Bloom, L. 50, 51, 216 Blum-Kulka, S. 5, 27, 28, 215 Bohannon III, J. 65 Boivin, I. 27, 78 Bonvillian, J. 65 Brannick 48 Brown, R. 1, 39, 42, 47, 48, 51, 59, 89, 95, 97, 107, 109, 111, 112, 132, 204, 208, 210, 211, 212, 213, 223 Bruner, J. 42, 52, 53, 54 Cairns, H.S. 49, 50 Capdevila I Batet, M. 46 Carroll, S. 29 Carter, A. 52, 53, 56 Casey, L. 52 Cazden, C. 39, 41, 42, 45, 49, 89, 97, 127-128, 132, 147, 148, 161, 199, 204, 210, 211, 213 Cenoz, J. 2, 3, 10, 11, 28, 29, 31, 93, 94, 221, 225 Chomsky, N 39, 40, 46, 52 Clark, A. 13 Clark, E. 25, 26 Clyne, M. 28 Comeau, L. 20, 75, 77, 78, 90, 222 Conforti, J. 53, 67, 68, 69, 72, 90, 98, 104, 186-188, 217, 218, 219, 229

Cook, V. 2 Coon, R. 21 Corsaro, W. 31, 53, 54, 55, 70, 100, 219 Coulthard, R. 99 Crosby, F. 69 Cross, T. 51 Cruz Ferreira, M. 32 Crystal, D. 109 Cummins, J. 25 Darwin, C. 39 Davies, E. 20 Davis, E.A 39, 40, 41, 42, 56, 57, 58, 89, 90, 95, 209, 217 De Houwer, A. 1, 3, 9, 10, 12, 13, 14, 15, 18, 20, 22, 25, 26, 73-74, 75, 89, 90, 92, 93, 112, 208, 209, 219, 220, 221, 228 De Jong, E. 30 De Villiers, J. 40, 47, 97 Dent, C. 69 Deuchar, M. 13, 15, 19, 21, 23, 94, 220, 222 Devin, J. 53, 59, 68, 72 Dewaele, J. 188 Dollaghan, C. 98 Döpke , S. 18, 19, 24, 25, 33, 53, 68, 77, 98, 99, 208, 220 , 222, 223 Dore, J. 4, 52, 53, 55, 56, 61-62, 90, 97, 215, 218, 228 Dromi, E. 94 Erreich, A. 40, 45, 48, 49, 213 Ervin-Tripp, S. 31, 49, 50, 51, 52, 53, 54, 55, 56, 60, 63, 64, 69, 70, 72, 90, 97, 214, 216 Ezeizabarrena, M.J. 97, 113 Faingold, E. 34, 94, 221 Fantini, A. 94 Fleta, M.T. 83 Fletcher, P. 15, 43, 46, 73, 89, 107, 109, 111, 119, 128, 129, 132, 146, 204, 209, 213, 223, 224, 227 Fraser, B. 27, 53 Garcia, O. 18 Garvey, C. 53, 54, 55, 60, 63, 66, 69, 70, 71,

248

Early Trilingualism Index

Index 90, 98, 214, 219 Gass, S. 65 Gawlitzek-Maiwald, I . 23, 224 Gelman, R. 25 Genesee, F. 2, 11, 12, 15, 20, 21, 26, 27, 28, 35, 75, 77, 78, 90, 219, 221, 222, 230 Goldfield, B. 53 Golinkoff, R. 72 Goodz, N. 18, 19 Gordon, D. 31 , 53, 54, 55, 64, 70, 72, 216 Greenbaum, S. 37, 89, 173 Grieve, R. 25 Grosjean, F. 3, 6, 17, 21, 29, 77, 90, 221, 222 Gut, U. 226 Hakuta, K. 25 Halliday, M.A.K. 52, 55, 56 Hamers, J. 31 Hammarberg, B. 28, 29 Hanlon, C. 39 Harding, E. 30, 91 Harding, M. 50 Helot, C. 31 Hernández Pina, F. 83 Hoffmann, C. 2, 3, 10, 28, 29, 30, 32, 33, 34, 90, 94, 220, 221, 225 Hollingsworth, H. 56 Holzman, M. 53, 55, 58-59, 63, 65, 90, 97, 100, 132, 211, 214, 216, 217, 229 House, G. 27, 215 Hsu, J.R 49, 50 Hulk, A. 17, 23, 24 Hurford, J. 47 Hymes, D. 5, 72, 215 Idiazábal, I. 15, 18 Ingram, D. 13, 40, 42, 44-50, 73, 132, 210, 211, 212, 213 James, S. 42, 44, 47, 53, 55, 60, 67, 69, 90, 100, 132, 211, 214, 228 Jekat, S. 12 Jessner, U. 9, 29, 33 Jones, S. 27 Juan-Garau M. 24, 31, 75, 78, 90, 222, 225 Kaplan, B. 60 Kasper, G. 28 Kasuya, H. 20 Kielhöfer, B. 20, 22 Klee, M. 40, 45, 48 Klima, E. 39, 40, 42, 45, 46, 89, 127-128, 147, 148, 161, 210, 211, 212 Kobashigawa, B. 72

249 Köppe, R. 12, 21 Kuczaj, S. 48 Labov, W. & T. 4, 46, 129, 132, 146, 212, 213, 227 Lambert, W 31 Lanza, E. 17, 19, 20, 21, 22, 23, 24, 26, 75-77, 78, 90, 100, 195, 208, 220, 221, 222, 228 Lasagabaster, D. 25, 28 Lebrun, Y. 18 Ledbetter, P 69 Leech, G. 27 Leopold, W. 1, 94 Lewis, M. 39, 41, 42, 56, 58, 209 Ligeza, M. 5, 53, 55, 62, 68, 69, 97, 98, 214, 215, 228 Lightbown, P. 47, 49 Lindholm, K. 19 Long, M.H. 65 Lucariello, J. 68 Mackey, A. 65 Macnamara, J. 52 MacWhinney, B. 1, 46, 63, 95, 109, 112 Mahlau, A. 12 Mannle, S. 92 Maratsos, M 48 Marcon, R. 21 Marcos, H. 53, 54, 70, 72, 90 Marmor, T. 27 Matthews, S. 24, 106 McCarthy, D. 39, 40, 209 McLaughlin, B. 9, 10 McShane, J. 42, 44 Meisel, J. 2, 3, 9, 10, 12, 13, 14, 15, 19, 26, 28, 90, 219 Merkin, S. 51, 216 Meyer, W.J. 40, 41, 89 Mikès, M. 12, 35, 220, 221 Miller, J. 98, 109 Müller, N. 23 Murrell, M. 35, 90, 219 Myers-Scotton, C. 24 Navracsics, J. 31 Nelson, K. 227 Newcombe, N. 64, 72, 216 Nicoladis , E. 12, 20, 21, 24, 26, 27, 28, 78, 90, 221 Ninio, A. 42, 53, 54, 55, 63, 68, 70, 71, 90, 97 Ochs, E. 75, 77 Ochs-Keenan, E. 72, 216

Early Trilingualism

250 Oksaar, E. 32, 33, 35, 221 Oliver, R. 65 Olsen-Fulero, L. 53, 67, 68, 69, 72, 90 , 98, 104, 186-188, 217, 218, 219, 229 Padilla, A. 19 Paradis , J 13, 15, 18, 24, 25, 223, 230 Park, T. 11, 21, 22, 90 Parnell, M. 50 Paterson, S. 50 Pérez-Vidal, C. 15, 24, 25, 31, 74-75, 78, 89, 90, 94, 112, 208, 210, 222, 224, 225 Pfaff, C. 21 Piaget, J. 44, 56-58, 146, 227 Pica, T. 65 Pine, J. 46, 51, 52, 213, 224, 227 Pinker, S. 40 Platt, J. 94 Poplack, S. 20 Pratt, C. 25 Prinz, P. 59-60 Przetacnik-Gierowska, M. 5 , 53, 55, 62, 68, 69, 97, 98, 214, 215, 228 Quay, S. 10, 12, 19, 21, 23, 33, 34, 35, 90, 94, 219, 220, 221, 222 Quirk, R. 37, 89, 173 Radford, A. 40, 45, 46, 97 Redlinger, W. 11, 21, 22, 90 Remick, H. 68 Richards, B. 109 Riley, P. 30, 91 Romaine, S. 93 Ronjat, J. 1, 18, 94 Rowland, C. 46, 51, 52, 213, 224, 227 Sachs, J. 53, 59, 68, 72 Sagasta, P. 28 Santelmann, L. 43, 210 Saunders, G. 20, 94 Schelletter, C. 24, 106 Schieffelin, B. 68 Schlyter, S. 15 Searle, J.R. 52, 53, 60 Seebach, M. 42, 44, 47, 53, 55, 60, 67, 69, 90, 100, 132, 211, 214, 228, 229 Serrat I Sellabona, E. 46 Shane, J. 40, 41, 89 Sharwood Smith, M 29. Shatz, M. 25, 53, 66, 67, 69, 71, 72, 90, 100, 216, 217, 218, 229

Sinclair, J. 99 Singh, R. 29 Singleton, D. 29, 221 Slobin, D. 25 Smith, M. 39, 40, 41, 42, 44, 46, 47, 56, 57, 58, 73, 89, 90, 132, 209, 210, 211 Snow, C. 18, 53, 54, 55, 62, 63, 68, 70, 71, 72, 90, 97 Snyder, A. 56 Spada, N. 47, 49 Stavans, A. 22, 32, 33, 220, 221 Steffensen, M. 53, 69, 90 Stern, W. 56 Stromswold, K. 40, 45, 46, 51, 97 Stubbs, M. 53, 99 Sully, J. 56 Swain, M. 21, 73, 226 Taeschner, T. 1, 9, 12, 21, 90 Thornton, R. 40 Tokuhama-Espinosa, T. 30 Tough, J. 54 Tracy, R. 23, 24, 25, 224 Trettien, A. 56 Turrell, M. T. 30 Tyack, D. 40, 42, 44-50, 73, 97, 132, 210, 211, 212, 213 Valian, V. 47, 52 Van de Linden, E. 17, 23, 220, 222 Van Valin, R. 51 Vazquez, I. 27 Vihman, M. 11, 21, 22, 26, 90, 94 Volterra, V. 1, 12, 21, 90 Vygotsky, L. 52 Wells, G. 42, 44, 47, 48, 53, 60, 61, 72, 73, 89, 90, 95, 102, 210, 211, 214, 219, 228 Werner, H. 60 Wesche, M. 21, 226 Wheeler, P. 54, 55, 63 Widdicome, S. 33, 34, 221 Wierbizcka, A. 27 Williams, S. 28, 29 Wolfson, N. 27 Wootten, J. 51, 216 Yip, V. 24, 106 Yu, A.18 Zaslow, M. 64, 72, 216 Zubiri, I. and Zubiri, E. 79

Subject Index Acquisition order 10, 11 Act of meaning 52 Adam, Eve and Sarah 42, 58 Additive and subtractive bilingualism 31 Advice to bilingual and multilingual families 31 Affective category 99, 102, 214-215 Age of exposure 34 Agglutinative language 85 Alternative questions 38, 107, 205, 213 American English 107, 213 Answering questions 50 Autonomous development hypothesis 3 Auxiliaries – emergence of 49, 52, 212, 46 – productive use of 49 Base language 17 Basque 14, 27, 28, 29, 79-82, 91-94, 97, 106, 107, 113, 146, 188-207, 208, 210, 212, 213, 215, 220, 222-226, 230 – non-Indo-European language 4 Bellugi’s periods/ stages 42, 45, 161-163, 212 Bilingual awareness 21, 26 Bilingual context 17 Bilingual Family Newsletter 30 Bilingual first language acquisition – definition of 10 Bilingual mode 16-17 Bilinguals 228, see also Early Bilingualism Bootstrapping 22, 224, 229 Bristol 41 British English 107, 205, 213 Brown’s stages 108, 109, 205, 208, 223 Can 47 Can initial questions 43, 210 Cantonese 24, 106 Caregiver feedback 53 Caregiver questions, requests, feedback 65-72 Case marking 15 Case study as a research methodology 94 Catalan 15, 29, 30, 31, 46, 74, 78, 210, 224, 225

Causal questions 57 Child’s question functions 214-215 CHILDES ( Child Language Data Exchange System) 1, 46, 84 CHILDES Chat 95 CHILDES Clan 95, 109 Child-rearing context 212 Clarification category 99, 101-102, 214-215 Clarification requests 70 Clarification requests in bilingual contexts 75-79 Clitics 15 Code mixing 21-24 Code-switching 15, 34, 76, 77 Cognates 220, 222, 230 Collaborative construction of meaning 53 Communicative acts 54, 55 Communicative competence 5, 26, 72, 215-216; see also Jenny Communicative intent 228 Community language 34 Competence factors 48 Comprehension – of feedback 69 – of form 49-50 – of requests 69 Confirmation 70 Confirmation category 99, 101, 214-215 Content words 13 Contingent queries 70, 71 Conventionality 54 Conversational exchange 70 Cross-cultural interference 27 Cross-linguistic equivalents 22, 222 Cross-linguistic influence analysis 106, 188-207 Cross-linguistic influence 3, 21-25, 219-226 – in questions 222-226 Cross-linguistic influence in the child’s general English 188-199 – cognates 193, 194 – equivalents 194-197 – expressed guess 196 – foreignisation 198 – hesitation 198

251

Early Trilingualism

252 – homewords 188, 189 – invented words 188 – lexical analysis 106, 188-199 – lexical inventions 188 – mixed utterances 189, see also crosslinguistic influence mixed utterances containing English, Basque and Spanish 193 utterances in Basque 189 utterances in Basque and Spanish 192 utterances in English and Basque 191 utterances in English and Spanish 192 utterances in Spanish 191 – mother’s strategies 195 – phonetic rule 198 – repetition 195, 196 – transfers 198 – trilingual mixing 193 Crucial age 33 Cultural ethos 27 Culture bound 32, 33 Day-centres 62 Declarative questions 38, 103 Declarative utterances 209-210 Developmental order of acquisition 46 Developmental pragmatics 65 Developmental stages 97 Differentiated-language-systems hypothesis 35 Differentiation Hypothesis 14, 16 Directives 53, 64 Discourse agreement 69 Discourse strategies 17, 34; see also parental discourse strategies Do-insertion 49 Double auxiliary marking 47 Double tense marking 47 Dutch 15, 23, 73, 74 Dynamic Model of Multilingualism 9, 29, 33 Early bilingualism 9-28 – definition of 9-11 – cross-linguistic influence in 21-26 – language choice in 16-21 – language separation in 11-16 – pragmatics in 26-28 Early inversion 210 Early trilingualism 9 – complexity of 11 – definition of 28-36 – descriptive studies of 30-32 – differentiations and separate systems 35 – dominance 33-34 – input 34-35

– topic, context and person relatedness 33 Echo questions 39 Emergence of interrogative behaviour 41-49 – of intonation 41-42 – of wh questions 44-47 – of yes-no questions 42-43 Emphatic function 33 English as a second or foreign language 65 English 11, 12, 13, 15, 18 – Germanic language 4 Equivalents 229-230 Equivalent pairs 12 Equivalent words 35 Ergative language 85 Estonian 11 Exchange 98 Expressed guess 75, 76, 78 Fathers 68 Feedback 65 Feedback categories 99, 178, 217-219 – acceptable feedback 99, 104, 178 – child’s feedback to mother 178-182 – comparison of Jenny’s feedback with other children 187 – comparison of mother and child’s feedback patterns 186-188 – definition of 103 – mother’s feedback to child 182-186 – none (no feedback ) 99, 104, 178 – unacceptable feedback 99 105, 178 Feedback patterns 229 Finiteness 15 Finnish 35 Formal categories of questions 102-103 Form-function pairings 67, 229 Formulaic combinations 46, 49 – phrases 213 Formulas 229 French 11, 14, 15, 20, 21, 23, 24, 26, 31, 33, 34, 73, 77, 226 Frequency of appearance of wh-forms 130, 135 Fronting 47, 108 Function words 13 Functional categories of questions 59-65 Functional category of question; see also real information, known information, confirmation, clarification, interactive, affective 99, 100-102 Functional taxonomies 54, 55, 56-69 Functional use 63 Fusion hypothesis 12, 16 Gender 40, 57

Index German 11, 14, 20, 21, 23, 24, 32, 34, 35, 106, 226 Gestures 52 Government and Binding 40, 46, 84, 97 Guest language 15 Hebrew 27, 32, 34, 55 Hints 64 Homeword 220 Host language 17 Hungarian 32, 35, 63 Idiosyncracy 206, 207, 213 Idiosyncratic question forms 42 Immersion 28 Independent development hypothesis 14 Indirect objects 15 Indo-European languages 29, 221 Infants under age two 56 Initial single system 11 Initial syntactic system 13 Input 18 – effect of on production 51-52 – input driven approach 52 – parental 17 Insisting strategies 77 Interaction 5, 65 – hypothesis 65 Interactional contexts 53, 56 Interactionist approach 52 Interactive category 99, 102, 214-215 Interchanges 55 Intercultural style hypothesis 27 Interdependent development hypothesis 16 Interlanguage pragmatics 27 Inter-linguistic transfer 5 Interlocutor 99 Interpersonal functions 60 Interrogative behaviour 98; see also question Intonation 37, 38, 41 – intonation questions 43 – rising 49 Inventory of Communicative Acts 55, 63 Inversion 43, 46, 48, 52, 212-213 – of subject 38 – of subject and auxiliary 51 Irish 31 Italian 33 Japanese 20, 34 Jenny – educational background 93 – family background 3, 91 – patterns of exposure 91-93 Jenny Communicative competence 161-166

253 – avoiding the truth 161 – making a joke 163 – seeking a new lexical item 162 – seeking information 164 – showing interest and enthusiasm 164 – showing surprise 163 Jenny: General language development – average length of questions produced 116 – comparison with monolingual and bilingual subjects 111-113 – MLU and TTR 110 – MLU in Basque and Spanish 113 – number of questions produced 115 – percentage of each question form produced 114 – percentage of questions to utterances 117, 118 Jenny: Other questions 149-151 – Alternative questions 150-151 – Echo-questions 149-150 Jenny: Pragmatic development 151-166 – clarification category comparison with mother 177 feedback to mother’s questions 180 form type 160 length 157 number 154 – confirmation category comparison with mother 177 feedback to mother’s questions 180 form type 159 length 156 number 153 – interactive category comparison with mother 177 feedback to mother’s questions 181 form type 160 length 158 number 154 – known information category comparison with mother 177 feedback to mother’s questions 179 form type 159 length 156 number 152 – length of questions 155-157 – number of questions 151-154 – real information category comparison with mother 177 feedback to mother’s questions 179 form type 158 length 155 number 152 – type of question form used in each

254 function 157-160 Jenny: Wh-questions 128-149, 199-207 – Bellugi’s periods for wh-questions 147-149 – earliest examples of wh-forms 132 – emergence of inversion and auxiliary insertion 132-135 – emergence of wh-forms 132 – how questions 144-145 – number of wh-questions per recording 130 – percentage of non wh-questions per recording 131 – percentage of wh-questions per recording 129 – summary of wh-questions 147-149 – the growth of wh-questions and yes-no questions 131 – time and order of emergence 128-132 – what questions early forms 135 fully formed what questions 132-135 non-fully formed what questions 135-137 other types of question development 137 What (‘s) + it/this/that/? 136 Wh-confusion 138 word order and subject repetition 136 – when questions 147 – where questions fully formed where questions 139 -140 non-fully formed where questions 140-142 other types of question development 141 subject repetition and lack of subject 141 Wh-confusion 141 Where (‘s) + it/this/that/? 140 – who questions fully formed who questions 142-144 other types of question development 143 Wh-confusion 144 Who + incorrect verb 143 Who + no verb 143 – why questions 145-147 Jenny: Yes-no questions 119-128, 199-207 – Bellugi’s periods for yes-no questions 127-128 – declarative questions 119 – early inversion can 121 is ( are) 121 is + 0 122 is + he 124 is + it 122 is + this 122 is + that 124 will, shall, have 124-125 – emergence of auxiliary insertion – do,

Early Trilingualism does, don’t, 125-127 – summary of yes-no questions 127-128 – Yes-no questions with inversion and auxiliary insertion 121-127 Kindergartens 215 Known information category 99, 101, 214-216 Language Acquisition Support System 53 Language choice 16-21, 26, 28 Language constellation 29 Language dominance 22, 27 Language mixing 28 Language mode 16-17, 21 Language of schooling 34 Language separation 11-16 – evidence of two systems 13-16 – one system or two 11 Language status 29 Language typology 29 Learning languages at school 28 Length of questions 209 Lexical development 35 Lexical differentiation 35 Lexical mixing 22, 220 Linguistic microcosm 18, 19, 93, 97 Maternal input 213 Maternal speech patterns 74 Matrix Language Frame model (MLF) 24 Mean Length of Utterance (MLU) 9, see also MLU Metacognitive development 25 Metalinguistic awareness 25-26 Minimal grasp 75, 76, 78 Mish-mash language 14 Mixed utterances 13 Mixing 17, 20, 26, 32, 35, 229 – in input 19 – in the home 93 MLU 42, 89, 95, 107, 204, 207, 208, 229 – in Basque 113 – in Spanish 113 – reliability as a measure 109 Modification of input forms 51 Monolingual – acquisition process 10 – children 26, 227 – context 17, 75, 77, 208, 221 – mode 16-17 – subjects 204 – studies of monolinguals and bilinguals 227, 228 Morpho-syntactic development 15

Index Mother and child feedback 217 Mother as a researcher 94 Mother’s data 105 Mother’s functions 217 Mother’s question functions 166-177 – clarification category comparison with child 177 feedback to child’s questions 184 form type 175 length 172 number 169 – confirmation category comparison with child 177 feedback to child’s questions 184 form type 175 length 172 number 168 – interactive category comparison with child 177 feedback to child’s questions 185 form type 176 length 173 number 169 – known information category comparison with child 177 feedback to child’s questions 183 form type 174 length 171 number 167 – length of questions 170-173 – number of questions 166-170 – real information category comparison with child 177 feedback to child’s questions 183 form type 174 length 170 number 167 – type of question form used in each function 173-176 Mothers’speech 67 Mother-child interaction 68 – questions 68 Move on strategy 75 Move 98 Multicompetence 2-3, 9 Multilingual acquisition 29 Multilingual upbringing 34 Multilingualism 2, 28 Negation 15 Negotiate a trilingual context 34 Negotiation of meaning 65 Non wh-questions 210 Non-community language 11

255 Non-inversion errors 47, 48 Norwegian 17, 21, 22, 76-77 Null subject 15, 16, 108 Number of questions 209 One system hypothesis 12-13 – criticism of 12-13 One-parent-one language 10 One-person–one language 14, 18, 31 – principle 21 One-system phase 14 One-word utterances 213 Order of acquisition for wh forms 44 Order of emergence 212, 228 – for question types 47, 49 Other questions 38, 213; see also Jenny Overt subject 15 Parental discourse strategies 75, 228 Parental discourse strategies, questioning behaviour 19, 20, 73 Parental input 74-75, 78 Parental repetition strategy 75 Passive trilingual 31 Peer dyad 85 Percentage of wh questions 44 Percentage of yes-no questions 42 Performance factors 48 Persian 31 Person-specific input 33 Phonological mixes 35, systems 13 Phonology of contractions 140 Piaget – adaptations from 56-58 – 7-category taxonomy 57, 58 Play 62, 63, 94, 215 Polarisation 219 Polish 62 Politeness 28, 31, 60, 64, 72 Portuguese 27, 33, 34 Pragmatic differentiation 26, 27 – competence 52 – flexibility 215 Pragmatics – in bilinguals 26 – in early language 52 Pronominal subjects 15 Prosodic features 41 – mixes 33 Quadrilingual 28 Question form 37-39 – in bilingual children 73-75 see also Jenny

256 Question formation – in adult Basque 79-80 – in adult Spanish 82-83 – in child Basque 80-82 – in child Spanish 83-84 Question function 54-69, 97; see also Pragmatic(s), Jenny Question sequence/ series 98 Questions – characteristics of 37-39 – mean length of 40, 209 – percentage/proportion of questions to sentences 40 Real information category 99, 100, 214-215 Recordings 95 Reformulation, reiteration 72, 99, 105 Repair 25, 72 Repetition 72, 99, 105 Requests 56, 61, 64 – questions 53, 54, 55 Rhetorical questions 39 Rich interpretation 52 Role and Reference Grammar 51 Rule of Grammont 18 SALT ( Systematic Analysis of language transcripts) 1 Scaffolding 53, 183 Second language acquisition as a term 28 Separate Development Hypothesis 14 Separate development 12 Separation of input according to person 19 Sequence of emergence 42 Serbo-Croatian 35 Siblings 68 – sibling effects 93-94 Simultaneous acquisition 9, 10 Single system hypothesis 12 Social-interactionist approach 65 Socialisation 77 Socio-communicative skills 5 Sociolinguistic competence 5 Sociopragmatics 27 Spanish 11, 12, 14, 18, 19, 21, 27, 28, 29, 30, 31, 32, 46, 82-84, 91-94, 97, 106, 107, 113, 146, 188-207, 208, 210, 213, 215, 220, 222-226, 230 – Romance language 4 Speech acts 31, 52, 54, 58, 64 Speech modes 29-30; see also language mode Studies of children’s questions 39-40 Successive acquisition 9 Swedish 35 Syntactic analysis of questions 106, 199-207

Early Trilingualism – declarative questions 107, 199-200, 206-7, 223, 224 – hybrids 107, 199, 203, 206-7, 223 – is +0/it/this/that dependence 107, 199, 201, 206-7, 223, 224-225 – null subject 107, 199, 200, 206-7, 223 – late wh 107, 199, 200, 206-7, 223 – subject on right 107, 199, 200, 206-7, 223 – subject repetition 107, 199, 202, 206-7, 223, 225 – translation 107, 199, 203, 206-7, 223, 226 – wh + it/this/that dependence 107, 199, 201, 206-7, 223 – wh-confusion 107, 199, 203, 206-7, 223, 225-226 – word order 107, 199, 202-203, 206-7, 223 Tag questions 38 Taxonomy of question analysis 97 Telegraphic speech 52 Tense and aspect 15 Time of first exposure 10 Topic 33 Transfer 17 Trilingual acquisition 28-30 Trilingual children 3, 227 Trilingualism 228; see also Early Trilingualism TTR (Type Token Ratio) 89, 95, 208, 229 – reliability as a measure 109 Turkish 63 Two systems 13-16 Unanalysed chunks 43 Understanding questions 50 Usurpation 185-187, 219 Validity of the taxonomy 98 Verb-features required for understanding 50 – copula 84 – lexical 84 Wh-questions 103, 211-213 – formation of 38 – inversion and auxiliary insertion 47-49 see also Jenny – non-inversion of 45 – syntactic development in 45 Why 39, 40, 42, 49, 56, 226, 228 Yes-no questions 209-211 – formation of 37-38 – inversion and auxiliary insertion 47-4 – positive 37 – stress on 37 see also Jenny