Language Development Over the Lifespan

cover

next page >

Cover

title: author: publisher: isbn10 | asin: print isbn13: ebook isbn13: language: subject publication date: lcc: ddc: subject:

Language Development Over the Lifespan De Bot, Kees.; Schrauf, Robert W. Taylor & Francis Routledge 0415998530 9780415998536 9780203880937 English Language acquisition. 2009 P118.L3634 2009eb 401/.93 Language acquisition.

cover

next page >

< previous page

page_i

next page >

Page i LANGUAGE DEVELOPMENT OVER THE LIFESPAN Language Development Over the Lifespan is intended as a reference resource for those conducting research on language development and the aging process, and as a supplementary textbook for MA courses in applied linguistics/ bilingualism programs (in schools of Education and Liberal Arts) that focus on language attrition/aging and adult literacy development in second languages. It offers an integrative approach to language development that examines changes in language over a lifetime, organized by different theoretical perspectives, which are presented by well-known international scholars. Kees de Bot is Chair of Applied Linguistics at the University of Groningen in the Netherlands. His research interests include foreign language attrition, language and dementia in multilingual settings, and more recently the application of Dynamic Systems Theory in SLA and multilingualism. He is coauthor of Second Language Acquisition: An Advanced Resource Book (Routledge 2005) and has published widely in the field of Applied Linguistics. Robert W. Schrauf is Associate Professor of Applied Linguistics at Pennsylvania State University in the United States. He conducts research in three main areas: bilingual autobiographical memory; language, culture, and cognitive aging; and applied linguistics and health sciences. He is president of the Association for Anthropology and Gerontology and associate editor of The Journal of Cross-Cultural Gerontology.

< previous page

page_i

next page >

< previous page

page_ii

next page >

page_ii

next page >

Page ii This page intentionally left blank.

< previous page

< previous page

page_iii

next page >

Page iii LANGUAGE DEVELOPMENT OVER THE LIFESPAN Edited by Kees de Bot and Robert W. Schrauf NEW YORK AND LONDON

< previous page

page_iii

next page >

< previous page

page_iv

next page >

Page iv First published 2009 by Routledge 270 Madison Ave, New York, NY 10016 Simultaneously published in the UK by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business This edition published in the Taylor & Francis e-Library, 2009. To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk. © 2009 Taylor & Francis All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging in Publication Data Language development over the lifespan / edited by Kees de Bot and Robert W. Schrauf. p. cm. Includes bibliographical references and index. 1. Language acquisition. P118 .L3634 2009 401/.93 22 2008032796 ISBN 0-203-88093-5 Master e-book ISBN ISBN10: 0-415-99853-0 (hbk) ISBN10: 0-8058-6460-1 (pbk) ISBN10: 0-203-88093-5 (ebk) ISBN13: 978-0-415-99853-6 (hbk) ISBN13: 978-0-8058-6460-1 (pbk) ISBN13: 978-0-203-88093-7 (ebk)

< previous page

page_iv

next page >

< previous page

page_v

next page >

Page v CONTENTS Introduction KEES DE BOT & ROBERT W. SCHRAUF

11

PART I Theoretical Approaches to Language Development

17

1 First and Second Language Development from a UG Perspective SHARON UNSWORTH 2 First Language Acquisition from a Usage-based Perspective HEIKE BEHRENS 3 A Comprehensive Dynamic Systems Theory of Language Development PAUL VAN GEERT

19 40 60

PART II Second Language Development

105

4 Lifespan Development of the L2 as an Intellectualization Process: An Ontogenetic Sociocultural Theory Perspective MARIA C. M. DE GUERRERO 5 A Dynamic View of Second Language Development Across the Lifespan WANDER LOWIE, MARJOLIJN VERSPOOR, & KEES DE BOT

107

< previous page

page_v

125

next page >

< previous page

page_vi

Page vi 6 Bilinguals with Primary Language Impairment KATHRYN KOHNERT 7 L1 Attrition Across the Lifespan MONIKA S. SCHMID

next page > 146 171

PART III Non-verbal Aspects of Language Development 8 The Development of Gesture MARION TELLIER 9 The Development of Sign Language DEBORAH CHEN PICHLER

189 191 217

PART IV Methodological and Neurolinguistic Aspects of Development

243

10 Longitudinal Designs in Studies of Multilingualism ROBERT W. SCHRAUF 11 The Role of Working Memory in Language Development Over the Lifespan SUSAN KEMPER 12 The Development of Neural Substrates of Language Over the Lifespan ARTURO HERNANDEZ, MERRILL HISCOCK, & ELIZABETH A. BATES

245

Index

< previous page

271 288 309

page_vi

next page >

< previous page

page_1

next page >

Page 1 INTRODUCTION Kees de Bot and Robert W. Schrauf Until recently research on language development has focused almost entirely on childhood and early adulthood, dealing largely with first and second language acquisition and syndromes of language impairment, and more recently on the late life issues of language attrition, and (again) syndromes of language impairment. It is as if normal language development leveled off at some early point and remained perfectly stable for some 40–50 years. Nevertheless, we know, both intuitively and empirically, that both first and second languages are dynamic and changing throughout the lifespan—affected by myriad cognitive, psychological, social and cultural shifts, reversals, and advances. This book seeks to address this rather curious lacuna in the literature by bringing together a series of researchers who review key areas of language development from a lifespan perspective. The collection is representative but not exhaustive, and we offer it as a stimulus to further research. In some sense, a lifespan approach to language is a new area of research. In handbooks on lifespan developmental psychology there is some interest in language, but the focus is generally narrow. Bernicot’s chapter entitled “Communication and lifespan development” in the edited volume Lifespan developmental psychology (Demetriou, Doise, & van Lieshout, 1998) is limited to first language development in well-known stages such as the prelinguistic period, the one-word period, the two-word period, and the sentence period for 3-year olds and older. There are short sections on second language development in childhood, and a section on communication and language in the elderly that starts with the sentence, “Communication and language during adulthood and old age have not yet been studied in a systematic way” (p. 162). However, as the contributions to the present volume show, this not really the case. Research in language and aging has boomed over the last decade. Klein’s “Language acquisition at different ages” in the edited volume The lifespan development of individuals (Magnusson, 1996) focused on the comparison between first language development in children and second language development in adults, in particular migrants in Europe, but paid little attention to the large body of research on language in the elderly. His work shows that

< previous page

page_1

next page >

< previous page

page_2

next page >

Page 2 “the age-factor” has been studied intensively in second language acquisition research. The main question concerned (then and now) whether and to what extent second language learners can reach a native level of proficiency (see Birdsong 2006 for a recent in-depth overview). Here “age” is basically limited to the time from birth till puberty. The same picture emerges in handbooks on language development, where the focus is on the early stages of development. This is logical. The foundations of language are laid in those early years. However, there seems to be an inverse relation between the age of the child and the number of studies devoted to language development at that age. Fewer studies address older children. After the development of oral skills, scientific interest seems to shift to literacy (the development of reading and writing skills) and developmental disorders such as dyslexia. Berman’s (2004) edited volume Language development across childhood and adolescence is unusual in terms of the age range studied; nevertheless, there are no studies on children older than 16. Surprisingly, there is hardly any research on language development in the age range from 18 till 55. Although there is research on specific issues in different age groups, such as age-grading in sociolinguistics and the development of ethnolects and youth languages in adolescents, no research looks systematically at how language develops over this long span between ages 18 and 55. Interestingly, however, there may actually be more information about adults than we are generally aware of—in the sense that adults often serve as controls in studies on either elderly or young/adolescent language users. While a systematic analysis may provide us with useful information, the investigation of control groups is surely an odd way to conduct a research program. In the following paragraphs, we lay out the rudiments of a lifespan approach to language development, taken largely from the advances made in lifespan development by our sister discipline of psychology. We begin by considering the variety of time-courses germane to a truly developmental approach and move from there to a summary of the key assumptions of the lifespan perspective. Lifespans and Timespans In developmental psychology there is a long-standing interest in changes over the lifespan (e.g. Baltes, Reese, & Lipsett, 1980; Demetriou, Doise, & van Lieshout 1998). Levinson (1978) distinguishes the following stages in lifespan development: 1. Early childhood (age 0–3) 2. Late childhood (age 3–12) 3. Adolescence (age 12–17) 4. Early adulthood (age 17–45)

< previous page

page_2

next page >

< previous page

page_3

next page >

Page 3 5. Middle adulthood (age 45–65) 6. Late adulthood (age 65 and older) Two groups of factors define the different stages in Levinson’s list: maturational factors, particularly in early childhood and possibly in the “old-old” stage (age 85 and older); and psychosocial factors: for example, participating in education, occupation-related changes, and retirement. Physical changes occur throughout the lifespan and interact with the other factors in a dynamic fashion: for example, a busy job may lead to a decrease in physical exercise, which may lead to health problems due to overweight which may have an effect on energy levels and job satisfaction. Baltes (2003) argues for a further differentiation of the last stage to include a separate period for the old-old, since this has its own characteristics that are different from those in earlier stages. The list presented by Levinson is only one timescale out of many that play a role in development. In fact, it is useful to think of a range of timescales laid out on a continuum. For instance, although the lifespan forms the most comprehensive timescale on the individual level, it could be argued that individuals are part of a larger history that they carry with them and which extends over a longer period than the individual lifespan. Thus, language development can also be studied as a process of change over many lifespans, and this is what historical linguistics does. At the other end of the continuum, within the individual, there are many micro-level scales. If we take the example of second language development, changes are seen to occur over several years: for example, when a person migrates to another country and starts using a new language as the language of daily communication (see Schmid, this volume); but also over months and weeks, as in instruction-based language development. Even at the micro-level of days , there may be differences. For instance, research has shown that vocabulary knowledge is fairly unstable in the sense that words may be remembered one day and yet be inaccessible the next day. More finegrained timescales have not been studied extensively. For instance, diurnal variation in the availability of linguistics resources is a common experience, but systematic study is still lacking. The interaction of timescales is a key issue in the lifespan view taken in this volume. Long-term changes are the result of the accumulation of small changes at finer timescales, and vice versa. For example, linguistic innovations by adolescents may be very local and short-lived, but some of them are diffused rather widely (e.g. via the media or the internet) and become part of the official language system through their inclusion in dictionaries. The Developmental Perspective In the 1980 edition of the Annual Review of Psychology, Baltes, Reese, and Lipsitt defined the lifespan perspective as “concerned with the description,

< previous page

page_3

next page >

< previous page

page_4

next page >

Page 4 explanation, and modification (optimization) of developmental processes in the human life course from conception to death” (1980, p. 86). They laid out several major assumptions that characterize lifespan studies to this day. Development is a Lifelong Process Baltes et al. (1980) argue that development is not limited to early periods but rather that it is in fact a lifelong process. This stands in contrast to a biological-maturational orientation that sees early development terminating in “maturity,” after which: “change, as a corollary, is often seen as decline or aging rather than development (Birren, 1964; Strehler, 1977)” (p. 70). Perhaps for scholars in language acquisition these sorts of distinctions call up currently debated notions of “critical periods” or “sensitive periods” in first and second language development, and certainly there may be age-related individual differences in the cognitive mechanisms available for language learning at different points in the lifespan. But there is a larger point at issue here. As the chapters in this book show, when the focus is widened to include the entire lifespan of language learners and language users, different sorts of questions come to the fore. Development is Subject to both Ontogeny and History A second feature of the lifespan perspective is that long-term development takes place within an overall framework of “biocultural (historical) change” (Baltes, Reese, & Lipsitt, 1980). This runs counter to an over-emphasis on the changing individual. In another context, Baltes, Reese, and Nesselroade (1977, p. 1) note that: “Developmental psychology recognizes that the individual is changing in a changing world, and that this changing context of development can affect the nature of individual change.” Thus, psychological development is not simply the workingout-over-time of internal processes according to an internal, individual, ontogenetic logic; rather, the course of development is affected by external, social, historical and environmental factors. “Historical factors” evoke changes at a macro-level that can affect the life course of many individuals simultaneously. As applied linguists interested in multilingualism, we might think of societal situations of language contact as a result of wars, major migration, shifting national language policies, and the like, in which whole populations undergo language shifts. From a developmental perspective, the timing of such changes is quite variable across the age range of individuals who are affected by them, and this creates cohort effects in development. For example, it is likely that Franco’s imposition of “Castilian” Spanish over all other language varieties in the Spain of the 1940s would predict large individual differences in the subsequent multilingual development of children in the middle years of development (e.g. ages 5– 7) vs. adults at mid-career in the workforce (e.g. ages

< previous page

page_4

next page >

< previous page

page_5

next page >

Page 5 35–40) vs. older adults (e.g. ages 60–70). These are cohort effects that affect age groups of language users and language learners. At a more granular level, we would also expect idiosyncratic “environmental factors” to play a role in language development at the individual level. A major topic in lifespan developmental psychology is the role of major life events that significantly alter the lines of development. Negative events have been shown to play a major role in the development of stress and depression (Hankin & Abramson 2001), and it is to be expected that such events, positive or negative, will also play a role in language development. In a multilingual society, for instance, multilingual (or monolingual!) educational programs, marriage, divorce, work opportunities, emigration, etc., can all affect language development at different points across the lifespan. Events such as going to school, or taking part in international student exchanges, or even a short holiday love-affair may have a significant impact on the development of an individual’s language. Clearly, as Baltes et al. (1980) point out, ontogeny does not proceed according to its own logic but is subject to myriad “external” influences. Developmental Processes are Not Necessarily Unilinear and Cumulative Baltes et al. note that “lifespan changes have been found to take many forms in terms of time extension, directionality, degree of interindividual variability, and plasticity” (p. 72). Given these realities, researchers are best served by a pluralistic notion of development. For instance, Baltes et al. distinguish between a continuity-oriented approach and a discontinuous approach. In the continuity-oriented approach, models that have been developed for a specific period of life are applied to other phases of the lifespan. The application of Piaget’s model of early childhood to later ages is an apt example. In the discontinuous approach, qualitatively different models are applied for different age spans. A pluralistic approach to development implies the acknowledgement that a wide range of processes is involved, each with its own logic, yet integrally interrelated: cognitive development, perceptual and motor development, social development, personality development and developmental psychopathology (see Magnusson, 1996, and Demetriou et al., 1998, for overviews). In the past, developments in these areas have been treated as specialized sub-fields that had few connections. However, in the last two decades awareness has grown that developments on different levels and between different sub-systems are interconnected. In particular, the booming research on brain functions has made it clear that cognitive processes are interacting with neurological processes, but also with physical changes and changes in the environment. Not all functions and sub-systems develop in parallel; rather, some undergird mutual growth or decline, while others are

< previous page

page_5

next page >

< previous page

page_6

next page >

Page 6 compensatory in nature. In addition, there is evidence that how individuals differ in development may be taskdependent. One key distinction concerns context-dependent vs. context-free tasks: The stability of differences in cognitive achievements on complex tasks between individuals is surprisingly low during childhood as well as during adulthood…In contrast, individual differences for highly aggregated measures of content free abilities (IQ) tend to be very stable. (Weinert & Perner, 1996, p. 214) Organization of the Volume In this book a number of issues of language development are discussed. We do not claim that this is a complete overview of what happens with language over the lifespan, nor do we organize the material by traditional distinctions in language and applied linguistics. We do not, for example, include separate chapters on traditional language components, such as phonology, syntax, discourse, and narrative development. Nor have we followed a psycholinguistic script—with attention paid to issues such as lexical retrieval or self-correction in speech. Rather, we present the lineaments of a new way of looking at language development fully cognizant that there are gaping lacunae in the evidence base. (For instance, there simply are not many studies of language development in mid-life.) We seek to review and summarize the existing developmental literature in language science for the purpose of creating a new perspective on our field—a lifespan perspective —and we expect to raise as many questions as we answer. The book is organized in four parts: (I) theoretical approaches to language development, (II) second language development, (III) non-verbal aspects of language development, and (IV) methodological and neurolinguistic aspects of development. Part I: Theoretical Approaches to Language Development Unsworth argues for the role of Universal Grammar (UG) in first and second language development. Compelling evidence for UG derives from what has been labeled the logical problem of language acquisition. This states that the input that children receive is insufficient to build the kind of complex language system normally seen among adults. Thus, language development must result from the activation of some innate knowledge (UG) by input from the environment. Broadly, this activated, innate knowledge is termed the Language Acquisition Device, and UG is basically a set of constraints that describe which hypotheses about possible language structures are universal and which are untenable in structuring particular grammars. These constraints limit the variation in development but do leave enough room for the natural variation observed.

< previous page

page_6

next page >

< previous page

page_7

next page >

Page 7 For second language (L2) acquisition, the big question is whether the Language Acquisition Device still works at later ages—particularly in adulthood. The chapter reviews various positions on this issue and draws the conclusion that the interlanguage of L2 learners is a natural language, and therefore constrained in similar ways as other natural languages, but that it is often difficult to disentangle the role of UG from the first language (L1). In its strictest form, the logical problem of language acquisition no longer applies to L2 learners since presumably they have already acquired a complete language system and can use elements from that fully developed L1 to develop the L2. As Unsworth shows, the role of L1 in L2 acquisition has been neglected for quite some time, because it weakened the UG assumption. In current research, the focus is more on structures for which the “poverty of input” argument still holds. That is, some structures are rare and will not be part of the input learners receive (be it in normal interaction or through instruction), and therefore when such constructions are acquired, some form of innate knowledge has to be assumed. Research within the UG paradigm has little to say about the role of the social context or on linguistic aspects other than syntax. Versions of the theory vary on the role they assign to the lexicon, but lexical acquisition as such is not the focus of such studies, as is the acquisition of pragmalinguistic and socio-pragmatic knowledge. The strength of the UG approach is in the formality of its arguments and the testability of its assumptions. Behrens argues for a usage-based approach to language development that differs in significant ways from the UGbased approach presented by Unsworth. First, usage-based approaches assume that language is acquired through a (possibly uniquely human) combination of social cognition, pattern recognition, and general learning mechanisms, and this eliminates the need for innate knowledge or a dedicated Language Acquisition Device. The second is that children go beyond the “poverty of input” by using analogy formation and abstraction from experience. “Poverty of input” describes the fact that children use linguistic structures that they have never heard before and that are more complex than the simplified input to which they are exposed. Such “precocious” usage has long been taken as clear support for the existence of an innate grammar (but see Pullum & Scholz, 2002). In the usage-based approach, it is assumed that complexity emerges from the interaction of more simple mechanisms. Computer simulations have shown that on a purely logical level, fairly simple mechanisms can develop considerable complexity. The third argument is that the non-linguistic environment contains much more linguistically relevant information than has been assumed previously. At the simplest level, for instance, the environment provides cues about the links between linguistic signs and entities to which they refer. The fourth argument is that in language development, children are conservative in an almost literal sense: they stick with what they have until

< previous page

page_7

next page >

< previous page

page_8

next page >

Page 8 something new and useful becomes available. Simple constructions are used in a “filler-and-slot” manner to express intentions, and the complexity of the structures develops much more slowly than was thought previously. Thus, children do not appear as creative as the proponents of the UG-based approach suggest. From a usage-based perspective, Behrens argues for research in which dense data are collected from a limited number of children, preferably consisting of interactional data and videos to allow for maximally rich interpretations of the children’s utterances. Because no specific innate mechanisms are assumed, the same sociocognitive mechanisms that are at work in early child development continue to play a role in development of the first language over the lifespan and in second language development. In sum, changes in the linguistic environment and in language-related resources, such as working memory, will have an impact on language development over the lifespan. Van Geert considers first language acquisition from the viewpoint of dynamic systems theory. This framework involves the careful dismantling of the representationalist stance (“mapping form to meaning”) of information processing theory, so that words and speech acts cannot be dissolved into pairs of phonological or orthographic forms linked to mental meanings. Rather, acts of thinking, speaking, deciding, rejecting, hearing, etc., are thoroughly embodied interactions of brain and environment that change both brain and environment over short-term time intervals, and bring about development (through more complex interlinkings with many other brain environment interactions) over longer intervals. This is not to say that more elemental components cannot be distinguished and studied at simpler levels of language acquisition. Indeed, van Geert describes the results from a previous study (Bassano and van Geert, 2007) of the developmental emergence of successively more complex “generators” starting from the child’s one-word grammar (holophrastic generator), simple combinations of words (combinatorial generator) to combinations of words governed by more complex syntactic rules (syntactic generator). Thus, movement from more elemental to more complex levels of language development, or a more fine-grained focus on highly specified components of language growth (e.g. development of aspects of phonology or syntax or semantics), can be modeled as dynamic fields in which long-term, multidimensional processes have simple or multiple outcomes with associated probabilities. As van Geert points out, such dynamic fields of developing competencies are driven in part by stabilizing conservative forces and unsettling progressive forces. Critically, development does not take place according to a preplanned, genetically grounded design. Rather, development is emergent and results from the interaction of multiple components, each with its own timescale, that self-organize in interaction with a changing environment. As with other researchers addressing the infant/child’s first language acquisition, van Geert looks to the embodied processes of joint attention and the inten-

< previous page

page_8

next page >

< previous page

page_9

next page >

Page 9 tional nature of social actors as primary factors. However, beyond child acquisition of language, the principles of dynamic systems theory describe an overarching framework for lifespan language research. Throughout the lifespan, language users are essentially adaptive systems in interaction with a changing environment, and throughout the lifespan, various language components are in flux, growth, and decline. Part II: Second Language Development De Guerrero takes up processes of both first and second language acquisition from within the framework of sociocultural theory (SCT). The chapter traces the Vygotskian theory of child development from early social interaction, mediated by language, through the internalization of those interactions as mental processes, again mediated by language. This process of internalization crucially involves developing “speech for thinking,” initially as vocalized private speech and later as subvocalized thinking. These developmental steps ground subsequent processes of concept formation that continue development throughout life. Although these processes are linked to first language acquisition in childhood, De Guerrero marshals considerable evidence from the second language acquisition literature to show that essentially the same processes occur in second language acquisition as well, which of course may also occur at later times during the lifespan. Thus, second language learners move through the same stages as do children learning their first language: use of private speech, the internalization of private speech as inner speech, concept formation, and metalinguistic, mental management of speech. From the viewpoint of lifespan psychology, the notion that any language acquisition seems to involve these four developmental moments raises the question of what kinds of variability we might expect in the quality, duration, and timing of these learning moments as learners grow older. Indeed, De Guerrero articulates exactly this question by appealing to Luria’s work (1973, 1981) on cerebral lesions and other brain disorders that in fact correlate with aging and affect inner speech. Thus, lifespan language scientists might profitably investigate the effects of diminished cognitive resources, stress, and chronic illness (often associated with aging) on inner speech and ultimately second language acquisition at later ages. The chapter by Lowie, Verspoor, and de Bot represents the reframing of existing literature and findings about second language acquisition, as these are currently articulated from within the multivariate, linear framework of information processing theory, and re-organizes these findings within a dynamic systems perspective (DST). In this sense, the chapter builds on the explanations and summaries of DST offered in earlier chapters by Behrens and by van Geert. But where Behrens and van Geert were concerned to reframe first language acquisition in terms of dynamic systems theory, this

< previous page

page_9

next page >

< previous page

page_10

next page >

Page 10 chapter seeks to reframe second language acquisition in terms of DST. A first section provides a brief summary of a DST view of language development, and this is followed by sections that review factors previously identified in the SLA literature that likely interact during second language development in a dynamic and non-linear way. These include sociocultural factors (e.g. instructed vs. non-instructed, foreign vs. second language learning), affective variables (e.g. motivation, attitude, anxieties, personality factors), cognitive variables (e.g. aptitude in various cognitive systems, working memory, and processing speed). Of course, the SLA literature has also targeted a number of age-related variables involved in acquiring a second language, and Lowie et al. review these as well (e.g. physiological changes and the critical period hypothesis). A final section places the previous material into a lifespan perspective, which embeds second language development in the biography of individuals. To wit: “Major life events can explain why a system changes, while a DST approach may help to understand how the system changes, but the why and how cannot always be clearly separated.” Embedding DST in a lifespan perspective opens a window on the interaction, simultaneity, and serendipity of these processes. The articulations in the conclusion are fully resonant with the key assumptions of the lifespan perspective as set out by Baltes, Reese, and Lipsitt (1980). Further, they demonstrate the deep compatibility between dynamic systems theory and the lifespan perspective. For example, Lowie et al. point to the multiplicity of interdependent systems (“systems tend to be nested. The language system is within the cognitive system which is itself nested in the physical systems of the body”) that develop in an interrelated way over time. They highlight interindividual and intraindividual variability over time: “there is not only variation in the acquisition process between individuals, some learn fast, others slow, but also within individuals there is a great deal of variation.” In her contribution, Kohnert focuses on the impact of primary language impairment in bilinguals at different ages. “Primary” refers to problems that are for the most part linguistic and not directly due to sensory, cognitive, or neurological irregularities or damage. As Kohnert points out, primary language impairments have been most often investigated in monolinguals, and the chapter extends this focus to bilinguals and how language is affected. Language learning, use, and maintenance depend on three groups of factors: means, opportunities and motive (the “MOM” framework). Means include linguistic knowledge, but also perception and memory capacity. Opportunities refer to possible contacts with, and use of, the language. Motive refers to socio-emotional aspects of language use and language learning. Kohnert’s main argument is that these factors dynamically interact during successive phases of the lifespan, and language proficiency is the ultimate outcome of these interactions. This is an important point for the assessment of language disorders in bilinguals. While the means (language capabilities,

< previous page

page_10

next page >

< previous page

page_11

next page >

Page 11 cognitive skills, and memory) available to bilinguals may be similar to monolinguals, their opportunities to use each language will by definition be more limited than for monolinguals. Thus, delays in language development in bilingual children may be caused by Primary Language Impairment, but may also be the result of lack of contact with one or the other language. While in most cases of simultaneous bilingualism both languages develop apace or “catch up” fairly smoothly, for some languages the contact may be below threshold and errors may become structural. The interaction of these variables makes the diagnosis of bilingual children problematic, since the means, opportunities, and motives (particularly this latter in the case of low status, minority languages) cannot be separated clearly. In later stages of life, the same factors likely play a role, and similar measurement problems arise. For instance, elderly bilinguals may show language decline, but again it needs to be established whether such shifts are caused by a decline in linguistic and cognitive means, a decline in opportunities to use languages in a meaningful way, or a decline in the motivation to communicate and use the language. Primary language impairment can have serious effects on various aspects of life. Academic success and level of job will be affected by it, but interestingly self-reports are not as negative as might be expected: despite lower educational and employment status, adults with primary language impairment appear to rate themselves not as less happy than matched controls. In her contribution, Schmid shows that language attrition, the loss of language proficiency within an individual over time, is a normal phenomenon in many language contact situations. Over the years, research on language attrition has shifted from an emphasis on the assessment of declining skills in the L1 to an emphasis on phases of growth and decline involved in language development in general. In essence, the extreme case, in which migration is forced and traumatic, and in which migrants’ or refugees’ first language is not spoken, brings into relief many processes of decline that are paralleled in subtler ways in less dramatic situations across all phases of the lifespan. For example, children may lose their mother tongue when they are no longer in contact with that language, as in the case of the Korean adoptees reported on by Pallier and his colleagues (2003). Or, for instance, older migrants may lose the language they have spoken most of their lives—though current research places in question complete loss. Of the many substantive findings in this research, one striking fact is that even people who believe they have undergone significant language losses may not in fact have lost as much as they believe. Why some people show more attrition than others is still poorly understood. For instance, preliterate children are particularly prone to attrition, but chronological age is surely a proxy for other social, psychological, and physiological variables here. Further, at early ages, it is difficult to separate out attrition from non-acquisition, largely because language skills depend critically on the development of other

< previous page

page_11

next page >

< previous page

page_12

next page >

Page 12 cognitive and motor skills, which develop at their own pace and not always in uniform fashion. Similarly, at the other end of the lifespan, an age-related decline in memory performance has an impact on language performance, but we may debate whether this is tantamount to language attrition. As Schmid indicates, normal variation in development and the onset of dementia cannot be distinguished. Again, longitudinal designs will play a key role in disentangling these factors (see Schrauf, this volume). Attrition can be a slow process and may happen after a few years or after many years, depending on sets of variables that interact over time. Too few points of measurement may lead to missing the crucial stages; too many may lead to relearning due to testing. As the study of attrition converges with the study of acquisition these issues will only grow in importance in future research. Part III: Non-verbal Aspects of Language Development For interesting sociohistorical reasons, the acquisition of sign language as a first language is distributed over childhood and young adulthood in a pattern that contrasts with the acquisition of spoken language by hearing individuals. In her chapter on the development of sign language Chen Pichler notes that only a small portion of deaf children are exposed to sign language from birth, because most deaf children are born to hearing parents who do not know sign language. Rather, the majority of deaf children begin learning language later in childhood than would hearing children. The chapter reviews four burgeoning areas of research in childhood sign language acquisition: phonology, nonmanual markers, fingerspelling, and the spatial components of narratives. As Chen Pichler points out, the acquisition of sign language by adults has attracted less research attention and has focused largely on phonology. Beyond adulthood, almost no literature has addressed either the acquisition of sign language or the development and decline of sign language in later life. Studies of the effects of Parkinson’s disease on signing have drawn some attention, but clearly a lifespan approach to sign language remains a research objective for the future. The study of gestures is a fairly recent development in linguistic research, even though some of the studies in the use of gestures in multilingual settings like Efron’s 1941 (1972) study on Jews and Italians in New York City date back many decades. Tellier discusses the research on the acquisition of gesture in first and second language development. Gestures come in various forms, ranging from more conventionalized gesticulation via pantomime and emblems to sign languages. Most of the work on language and gesture has focused on gesticulation since that is explicitly speech related and according to some theoretical models of gesturing coming from the same conceptual source. One of the intriguing questions in this field is why we gesture. Studies have shown that speech is less comprehensible when gesture information is

< previous page

page_12

next page >

< previous page

page_13

next page >

Page 13 removed; speakers become less fluent when they cannot use their hands, and even people that are born blind gesture, also when conversing with another blind person, as do speakers without a visual handicap when using the telephone. Apparently, gestures are an essential part of our communicative system. Gestures appear to play a role from the very first stages of language development and in all stages the development of speech and gestures goes hand in hand, though the two systems are not used at the same time till 18 months or later. After this period the child begins to synchronize gestures and speech. Pointing is typically a first sign of symbolic behavior and of the ability to decontextualize. While there are some studies on children between 7 and 11 years, there is hardly any research on later stages. An exception is the study by Cohen and Borzoi (1996) who looked at verbal description in younger and older females. It was expected that the older informants would use more gestures to compensate for word-finding difficulties but this was not born out by the data. There is a rapidly growing body of research on gesturing in multilinguals. Though the empirical support for this assumption is still limited, there is a general feeling that speakers of different languages, the Italians being the typical example, gesture differently, and that becoming fluent in a second language involves the acquisition of the right gestures. The Efron study mentioned earlier showed that immigrants adapt their gesturing to their new linguistic environment. Several other studies have shown that there is cross-linguistic transfer of gestures, both from the first to the second language and from the second to the first. As Tellier argues, this is a field in which a lot of very basic research still needs to be done before we get a proper idea of development of gesture over the lifespan. Part IV: Methodological and Neurolinguistic Aspects of Development Schrauf provides an overview of methodological issues related to the study of language development over the lifespan. He summarizes a rich literature on different types of designs for the study of cognitive aging in monolinguals and draws out implications for the study of multilingualism. Currently, longitudinal studies of multilingualism are rare, and cross-sectional studies show that overall age differences are similar for monolinguals and multilinguals, as are patterns of development. The crucial design problem is how to disentangle age, time-of-measurement, and cohort effects in the two traditional designs: crosssectional , in which data is collected at one time on groups differing in age, and longitudinal, in which data is collected on the same sample across a number of time points. Cross-sectional studies, by far more common, have typically used an “extreme-age” groups design in which groups of young and

< previous page

page_13

next page >

< previous page

page_14

next page >

Page 14 old adults are compared. In such a design the effects of age differences and cohort effects are confounded. An interesting conclusion from the literature is that many of the age differences found in cross-sectional studies have not been replicated in longitudinal studies. Thus, results taken to reflect age-effects more likely reflect either cohort effects or time-of-measurement effects. This is particularly true for studies on lexical development over the lifespan. Where cross-sectional studies are inconclusive and often contradictory, longitudinal studies typically show a late-life decline in lexical skills. Leading researchers in the field of lifespan developmental psychology, such as Paul Baltes and Werner Schaie, have proposed designs that combine aspects of both longitudinal and cross-sectional approaches, and the chapter provides a careful overview of these sequential designs that focus on data collected over time from multiple groups differing in age. Such designs make it possible to disentangle the effects of age, cohort and time-of-measurement. Although longitudinal designs have been considered superior for the study of cognitive aging, such designs are prey to several threats to validity. These include cohort effects, test–retest effects, and selection and drop-out effects. Moreover, there are specific threats that are peculiar to studies of multilingualism. These include: the dynamics of language proficiency and linguistic environments, and problems related to the measurement of language contact. One aspect that no design can really compensate for is the issue of selection and drop-out. It turns out that the more motivated, more intelligent, more healthy and better educated people are, the more likely they are to take part in research studies and show a lower drop-out rate. This means that in most studies age-related changes are probably underestimated. One aspect of these more advanced research designs needs to be mentioned here: they typically take many years to be completed, and in the current 3- to 4-year project cycles of most funding agencies there is little room for studies that may take considerably longer. Kemper provides a useful review of the concept of working memory and executive function, and a brief description of 22 instruments currently used to measure these capacities (15 of executive function, five of verbal working memory, and two of visual working memory). Indeed, this listing and bibliographic references are an invaluable guide to the operationalization of both cognitive processes. Further, insofar as the literature demonstrates both individual differences in working memory capacity and clear declines with age, a lifespan view of language development must necessarily include these processes. The chapter by Hernandez, Hiscock, and Bates asks two questions: (1) how does language processing in children lead to the neural organization seen in adulthood? and (2) how does this system change through the adult years? In addressing each of these questions, the chapter summarizes the relevant experimental/neuroimaging literature and the neuropsychological/clinical literature in separate sections. Childhood is of course the scene of much

< previous page

page_14

next page >

< previous page

page_15

next page >

Page 15 change at both macro-levels (e.g. differential rates of regional maturation favoring the right vs. the left hemisphere, anterior to posterior cortices, etc.) and micro-levels (e.g. synaptogenesis, synaptic pruning, myelination). The chapter concludes that significant reorganization does not take place in adulthood, but that the frontotemporal network subserving language remains basically stable once the native language is acquired. Interestingly, however, Hernandez et al. cite evidence that older adults recruit wider areas of the brain than do young adults in tasks requiring complex language processing. Thus, although it falls outside the scope of the chapter, the complex task of second language acquisition could well be the site of many fruitful research questions in brain and language science. References Baltes, P. B. (2003). Die Zuknuft des Alterns . Berlin: Max Planck Institut. Baltes, P. B., Reese, H., & Lipsett, L. (1980). Lifespan developmental psychology, Annual Review of Psychology, 31, 65–110. Baltes, P. B., Reese, H., & Nesselroade, J. (1977). Life-span developmental psychology: Introduction to research methods . Monterey: Brooks/Cole Publishing. Bassano, D., & van Geert, P. (2007). Modeling continuity and discontinuity in utterance length: A quantitative approach to changes, transitions and intra-individual variability in early grammatical development. Developmental Science , 10(5), 588–612. Berman, R. (2004). Language development across childhood and adolescence. Amsterdam/Philadelphia: John Benjamins. Bernicot, J. (1998). Communication and language development. In A. Demetriou, W. Doise, & C. van Lieshout (Eds.), Life-span developmental psychology (pp. 137–178). Chichester: John Wiley. Birdsong, D. (2006). Age and second language acquisition and processing: A selective overview. In M. Gullberg & P. Indefrey (Eds.), The cognitive neuroscience of second language acquisition (pp. 9–49). Malden: Blackwell. Birren, J. E. (1964). The psychology of aging. Englewood Cliffs, NJ: Prentice-Hall. Cohen, L. R., & Borsoi, D. (1996). The role of gestures in description-communication: A cross-sectional study of aging. Journal of Nonverbal Behavior, 20(1), 45–63. Demetriou, A., Doise, W., & van Lieshout, C. (Eds.). (1998). Life-span developmental psychology. Chichester: John Wiley. Efron, D. (1972 [1941]). Gesture and environment: A tentative study of some of the spatio-temporal and linguistic aspects of the gestural behavior of eastern Jews and southern Italians in New York City. The Hague/Paris: Mouton de Gruyter. Hankin, B., & Abramson, L. (2001). Development of gender differences in depression: An elaborated cognitive vulnerability-transactional stress theory. Psychological Bulletin, 127 , 773–796. Klein, W. (1996). Language acquisition at different ages. In D. Magnusson (Ed.), The lifespan development of individuals: Behavioral, neurobiological and psychosocial perspectives (pp. 244–264). Cambridge: Cambridge University Press.

< previous page

page_15

next page >

< previous page

page_16

next page >

Page 16 Levinson, D. (1978). The seasons of a man’s life. New York: Knopf. Luria, A. R. (1973). The working brain. An introduction to neuropsychology (B. High, Trans.). New York: Basic Books. Luria, A. R. (1981). Language and cognition (J. V. Wertsch, Ed.). New York: John Wiley. Magnusson, D. (Ed.). (1996). The lifespan development of individuals: Behavioral, neurobiological and psychosocial perspectives. Cambridge: Cambridge University Press. Pallier, C., Dehaene, S., Poline, J.-B., LeBihan, D., Argenti, A.-M., Dupoux, E., & Mehler, J. (2003). Brain imaging of language plasticity in adopted adults: Can a second language replace the first? Cerebral Cortex , 13, 155–161. Pullum, G. K., & Scholz, B. C. (2002). Empirical assessment of stimulus poverty arguments. The Linguistic Review, 19, 9–50. Strehler, B. L. (1977). Time, cells, and aging. New York: Academic Press. Weinert, F., & Perner, J. (1996). Cognitive development. In D. Magnusson (Ed.), The lifespan development of individuals: Behavioral, neurobiological, and psychosocial perspectives (pp. 207–222). Cambridge: Cambridge University Press.

< previous page

page_16

next page >

< previous page

page_17

next page >

Page 17 Part I THEORETICAL APPROACHES TO LANGUAGE DEVELOPMENT

< previous page

page_17

next page >

< previous page

page_18

next page >

page_18

next page >

Page 18 This page intentionally left blank.

< previous page

< previous page

page_19

next page >

Page 19 1 FIRST AND SECOND LANGUAGE DEVELOPMENT FROM A UG PERSPECTIVE Sharon Unsworth 1 Motivating UG: L1 Acquisition The central tenet of the generative approach (Chomsky 1959, 1965, 1975, 1981, 1986, 1995, 1999) is that humans are predetermined to acquire language; this biological endowment ensures that all (non-impaired) children acquire the language (or languages) to which they are exposed in their environment, and it distinguishes humans from all other species. The argument for such a language acquisition device is derived from the observation that the linguistic knowledge which children acquire goes far beyond the input to which they are exposed. Before considering what this predetermined knowledge involves, let us examine the argument for its existence in more detail. Children’s linguistic development progresses at an astonishing rate. At the age of 6–8 months already, when children are only just starting to crawl, their linguistic abilities are so sensitive that they are able to discriminate between the sounds of what will become their native language and between sounds in other languages (e.g. Werker & Tees, 1984). At around age 1 year, they start to produce their first words, and approximately 6 months later they start to combine words with each other, adding grammatical elements such as articles, plural forms and prepositions, to form increasingly complex sentences such that at around age when they might be learning how to ride a tricycle, they start to produce complex sentences such as relative clauses.1 What is more remarkable is that these stages are consistent across children and that they have been observed for various languages, including sign languages (see e.g. Lillo-Martin, 1999). The rapidity and uniformity of language acquisition across different learning conditions also suggests (though of course does not necessarily entail) that children must come equipped for the task with innate knowledge about language. Perhaps the most crucial argument for the existence of an innate language faculty comes from the observation that children go beyond the input they

< previous page

page_19

next page >

< previous page

page_20

next page >

Page 20 hear, that is, they come to know more than the totality of that to which they are exposed. How children progress beyond the input, a problem which has come to be known as the logical problem of language acquisition or the poverty of the stimulus , has to be addressed by any account of language acquisition. The generativist’s solution to this problem is to posit that certain parts of linguistic knowledge are innate. It is claimed that the child cannot solely rely on the input because not only does this vary from child to child, but more importantly, it is also degenerate, finite and restricted in scope (Chomsky, 1965). It is degenerate because it includes the incomplete/unfinished sentences which are characteristic of everyday speech. It also only contains a finite number of sentences, yet the knowledge which children acquire allows them to produce a potentially infinite number of sentences. Finally, the input is restricted in scope in that it only contains positive evidence, that is, evidence of what is possible in the language in question. Negative evidence, that is, information about what is ungrammatical in a particular language, is not available and when it is (e.g. through correction by caretakers), it is not used (Marcus, 1993; Brown & Hanlon, 1970). Nevertheless, children still come to know which strings in their native language(s) are ungrammatical. This innate knowledge or language faculty constitutes the initial state for the language-learning child. It provides information about the range of possible languages, including a set of abstract universal principles common to all languages, and a set of parameters capturing the variation observed between different languages. Together, these are known as Universal Grammar (UG). UG is often likened to a pre-wired box with a set of switches (Chomsky, 1988, analogy attributed to J. Higginbotham). The wiring specifies the possible options and the switch settings determine the available choices. Language acquisition thus involves—metaphorically speaking—flicking the appropriate switches on the basis of the input which the child hears. The abstract universal principles of UG take the form of constraints. Constraints restrict the grammar in such a way that certain combinations, for example, of words or of sounds and meanings, are prohibited. These constraints hold for all languages and they may not be violated. The sentences in (1) and (2) illustrate a well-known constraint on form which prevents certain phonological processes from applying after wh-movement has taken place. More specifically, these examples concern the optional contraction between want and to in English; as we shall see, this phonological process is only possible under certain syntactic conditions. Let us start by briefly considering the derivation of wh-questions. The question in (1)-a is formed by moving the wh-word, what , from its base position as the object of eat , to sentence-initial position, leaving behind what is called a trace in the process; this trace, considered to be psychologically real, marks the wh-word’s original position, as illustrated in (1)-b, where t signifies trace. In (1)-c, want and to are contracted to wanna , whereas in the wh-question in (2) this is ungrammatical, as indicated by the asterisk in (2)-b. This rather subtle difference results from the fact that the

< previous page

page_20

next page >

< previous page

page_21

next page >

Page 21 questioned constituent in the infinitival clause is different in each of the two sentences: in (1), the questioned constituent is the object of eat , and in (2) the subject of go. More specifically, the trace left by the wh-word in (2)-b intervenes between want and to, as shown in (2)-c, and consequently, wanna -contraction is not possible. Native speakers of English have implicit knowledge of this constraint, which manifests itself in other languages, too, albeit in different ways; knowledge of this constraint entails that (2)-b is ungrammatical (see Crain & Thornton, 1998, Chapter 23, for more details). (1) a. What do you want to eat? b. Whati do you want to eat ti? c. What do you wanna eat? (2) a. Who do you want to go home? b. *Who do you wanna go home? c. Whoi do you want ti to go home? The sentence in (3) illustrates a constraint on meaning. (3) Roy loves Angelai, but Traceyj hates heri/*j The pronoun her refers to a singular antecedent who is female. The sentence in (3) contains two such potential antecedents, namely Tracey and Angela. As the co-indexing indicates, however, there is only one antecedent which is grammatical, namely Angela. Tracey cannot function as the antecedent for her because—and this is the relevant constraint—anaphoric interpretations are prohibited in certain structural configurations. (To be specific, the relevant constraint here is Principle B of the Binding Theory—Chomsky, 1981.) Variation is instantiated in parameters linked to universal principles. For example, there is a principle which states that all phrasal constructions have “heads.” A noun phrase (NP) thus has a noun as its head, and a verb phrase (VP) a verb, etc. The position of this head relative to its complement is, however, subject to variation, and this is captured in a so-called Headedness Parameter, which has two options, head-initial and head-final. Languages such as English and French are head-initial, for example, whereas Japanese and Turkish are head-final. In recent versions of generative theory, the locus of variation has shifted to the lexicon, and more specifically, to functional categories.2 For example, the functional category Tense is said to be weak in English and strong in French. This difference in “feature strength” accounts for the word order variation with respect to negation illustrated in (4). (4) a. I do not eat peas very often b. Je ne mange pas de petits pois très souvent I eat not of peas very often “I do not eat peas very often” c. *Je ne pas mange de petits pois très souvent

< previous page

page_21

next page >

< previous page

page_22

next page >

Page 22 In (4)-a, the main verb, eat , appears to the right of the negator, not , whereas in (4)-b, the main verb has “raised” over the negator, pas, and appears to its left. This is because the tense feature in French is strong and therefore requires the verb to move to that projection; if it does not, ungrammaticality results, as illustrated in (4)-c. One area of debate in L1 acquisition research has been whether functional categories and their respective projections, for example Tense Phrase (TP), Determiner Phrase (DP) and Complementizer Phrase (CP), are available in the earliest stages of linguistic development. Proponents of so-called weak continuity claim that functional projections emerge one by one: children initially start with a VP and they subsequently add a TP, followed by a CP (e.g. Clahsen, Eisenbeiss, & Vainikka, 1994; Vainikka, 1993/1994). When children produce non-nominative subjects as in (5), it is claimed that this is because the position where nominative case is assigned, TP, is not yet available. This is also why verbs in children’s early utterances are often non-finite, as in (6). (5) My make a house (Nina 2;0) (Vainikka, 1993/1994, p. 273, ex. 276) (6) a. die helemaal kapot maken (Dutch, Niek 3;2) that altogether broken make-INF “completely break that one” (Wijnen, 1997, p. 1, ex. 1) b. dormir là Michel (French, P 2;2) sleep-INF there Michel “Michael sleep there” (Deprez & Pierce, 1994) Proponents of strong continuity (or the Full Competence Hypothesis—Poeppel & Wexler, 1993) argue that children do have access to all functional categories from the start, however (Weissenborn, 1990; Boser, Lust, Santelmann, & Whitman, 1992; Hyams, 1992). On this view, although not identical, child and adult grammars include the same grammatical objects and mechanisms, namely those made available by UG (see Guasti, 2002 for introduction to relevant issues). Within this approach, it has been argued that the development of syntax follows a predetermined biological schedule (Borer & Wexler, 1992); thus, children’s use of non-finite verbs in matrix clauses (so-called Root Infinitives) is due to immature principles of UG rather than the lack of functional categories (Rizzi, 1993/1994; Wexler, 1998). To sum up, according to the generative approach, the initial state of L1 acquisition consists of a set of universal constraints, Universal Grammar, which restricts the hypotheses children (subconsciously) formulate when

< previous page

page_22

next page >

< previous page

page_23

next page >

Page 23 constructing their grammars. The variation observed in natural language is captured in a predetermined set of options from which children make a selection on the basis of the linguistic input to which they are exposed, be that from one or more languages. By around age 5, all of the correct options will have been selected.3 2 UG in Adult L2 Acquisition The central question driving generative non-native (L2) acquisition research is whether L2 acquisition is constrained in the same way as L1 acquisition, that is, whether L2 grammars obey the same set of universal constraints as native-speaker grammars. Following early work by Corder (1967) and Selinker (1972), the non-native acquirer’s (L2er) grammar, or interlanguage (IL), is viewed as a system in its own right, rather than in terms of the L1 or the target language (TL). The questions which generative L2 research seeks to address include: What are the properties of the L2er’s interlanguage system? Are they properties which are characteristic of natural language grammar? How exactly are these properties acquired? As in most studies, the focus in this section will be on adult L2ers. The study of child L2 acquisition from a generative perspective is dealt with in the following section. 2.1 UG and/or L1 Transfer Adult L2ers and L1 children differ by definition because L2ers already know another language. The extent to which L2ers use this knowledge in the L2 acquisition process, and whether and how the role of the L1 can be teased apart from the role of UG, has been subject to considerable debate. In early generative work, the question of whether L2 grammars are UG-constrained was framed in terms of access to UG. L2ers were claimed either to have no access (e.g. Clahsen & Muysken, 1986), direct or full access (e.g. duPlessis, Solin, Travis & White, 1987; Schwartz & Tomaselli, 1990; Epstein, Flynn & Martohardjono, 1996) or indirect (or partial ) access to UG (e.g. Clahsen & Muysken, 1989). On the no access view, L2ers are claimed to make use of general learning mechanisms, such as “linear sequencing strategies which apply to surface strings” (Meisel, 1997, p. 258). In the approaches proposing access to UG, the existence of L1 influence was generally left implicit or denied, often because attributing a role to the L1 was considered to weaken the case for UG in L2 acquisition (White, 2003b, p. 27). For those espousing the indirect access view, UG is only available via the L1. If L2ers demonstrate unconscious knowledge of UG principles which could not be acquired as a result of L2 input, these are claimed to derive from the L1, which is viewed as a particular instantiation of UG (Bley-Vroman, 1990; Schachter, 1989). On this latter approach, UG is no longer available as a separate entity; rather, it can only be accessed in terms

< previous page

page_23

next page >

< previous page

page_24

next page >

Page 24 of the language-specific grammar that is the L1. Any parameter settings which are not instantiated in the L1 are lost. As White (2003b) notes, in early work the terms direct and partial access were used inconsistently across the research community. Furthermore, the dichotomy which was often adopted between UG access and L1 influence was “overly simplistic and misleading” (White 2003b, p. 27). When L2ers demonstrate knowledge of UG principles, it may be impossible, for some L1/L2 combinations at least, to determine whether this knowledge stems directly from UG or from the L1 (Hale, 1996). In light of this problem, it is argued by some that the role of UG in L2 acquisition is best established by determining whether, as in L1 acquisition, the learner faces a logical problem and whether this problem can be overcome (White 1985, 1989; Schwartz & Sprouse, 2000). This research is reviewed in the following section. The role of the L1 has been explored more explicitly as part of recent approaches focusing on the initial state . The initial state refers to the earliest stage of L2 development, that is, the knowledge which the L2er brings to the acquisitional task. There are three main approaches: Schwartz and Sprouse’s (1994, 1996) Full Transfer/Full Access, Vainikka and Young-Scholten’s (1994, 1996a, 1996b) Minimal Trees, and Eubank’s (1993/94, 1994, 1996) Valueless Features Hypothesis. Let us briefly examine each of these approaches in turn. As its name suggests, Schwartz and Sprouse’s (1994, 1996) Full Transfer/Full Access model claims that L2ers initially transfer all of their L1 grammar (excluding the phonetic matrices of lexical/morphological items) to their interlanguage grammar. In other words, the L2 initial state is the L1 final state. When this L1-based analysis of the TL fails, the IL grammar is restructured. During this restructuring process, learners have “access” to properties of UG, including those which are not instantiated in their L1 grammar. On this approach, then, L2 development is determined by L1 transfer (the initial state), UG and L2 input, and a learner’s IL grammar—even if it is like neither the L1 nor the target language—remains within the confines of UG at all stages of development (cf. Epstein et al.’s 1996 Full Access approach, which claims that L2ers have full access to UG but that the initial state is not the L1). For example, in their analysis of the development of German by a native speaker of Turkish called Cevdet, Schwartz and Sprouse observe that at a certain point in his acquisition of word order, Cevdet’s IL grammar is restricted in such a way that it resembles French rather than Turkish or German. In German, finite verbs in matrix clauses always appear in second constituent position, a phenomenon known as Verb Second (V2). Thus, when a non-subject constituent appears sentence-initially, the finite verb and subject invert so that the verb appears in second position and the subject in third position. This is illustrated in (7)-b; the canonical order for matrix clauses is given in (7)-a.

< previous page

page_24

next page >

< previous page

page_25

next page >

Page 25 (7) a. Er hat gestern meine Tante gesehen he has yesterday my aunt seen “He saw my aunt yesterday” b. Gestern hat er meine Tante gesehen c. dann trinken wir bis neun Uhr then drink we until nine o’clock “Then we drink until nine o’clock” d. in der Türkei der Lehrer kann den Schüler schlagen in the Turkey the teacher can the pupil beat “In Turkey, the teacher can hit the pupil” (Schwartz & Sprouse, 1994, pp. 336–337, exs. 21-c and 22-c) Schwartz and Sprouse (1994) observe that when Cevdet initially starts to use V2, it is restricted to utterances where there is a pronominal subject. When the subject is a full NP, no subject–verb inversion takes place, as in (7)-d. This restriction is not consistent with the target language, German, but it is consistent with an option present in French. In question formation in French, full NP subjects cannot invert with the verb, but pronominal subjects can. Thus, Schwartz and Sprouse argue, although Cevdet’s grammar resembles neither the L1 nor the target language, it is still a natural language grammar, that is, it is still constrained by UG (see Finer & Broselow, 1986; Hirakawa, 1990, for similar results). Evidence supporting the existence of full transfer at the morphological level comes from work by Montrul (2000, 2001). Vainikka and Young-Scholten (1994, 1996a, 1996b) also claim that IL grammars are constrained by UG but, in contrast to Schwartz and Sprouse, these authors adopt a weak continuity approach, claiming that the initial state comprises the Verb Phrase (VP) projection only. Furthermore, according to Vainikka and Young-Scholten, this is the only category which is transferred from the L2er’s L1. Evidence for this claim comes from data on the acquisition of German word order by native speakers of Turkish, Korean, Spanish, and Italian. The VP in German is head-final, which means that in matrix clauses, the non-finite verb (V) appears in sentence-final position, to the right of the object (O), as in (7)-a above. Turkish and Korean are also OV languages, but Italian and Spanish, like English, are VO languages. Vainikka and Young-Scholten (1994, 1996a, 1996b) observe that in the initial stages of their L2 development, the Turkish and Korean learners in their study produced right-headed VPs, as in (8), whereas the Italian and Spanish learners produced left-headed VPs, as in (9). (8) Oya Zigarette trinken (Aysel, L1 Turkish) Oya cigarette drink-INF “Oya drinks cigarettes” (Vainikka & Young-Scholten, 1996a, p. 16, ex. 16-a)

< previous page

page_25

next page >

< previous page

page_26

next page >

Page 26 (9) Ich sprechen die meine Firma (Salvatore, L1 Italian) I talk-INF the my firm “I speak (to/at) my firm”

(Vainikka & Young-Scholten, 1996b, p. 156, ex. 159-b) At the same time, the learners’ production is characterized by a marked absence of properties relating to functional projections, such as verb raising, auxiliaries and modals, and inflectional morphology, which, according to Vainikka and Young-Scholten, suggests that functional categories are absent at this stage of development. These categories are subsequently added in a stepwise fashion and, unlike the VP, they are not subject to L1 transfer. Rather, the IL grammar develops on the basis of the L2 input interacting with UG. On Eubank’s (1993/94, 1994, 1996) Valueless Features account, L2ers transfer all of the L1 grammar except the “strength” of the features of functional categories. Such features are unvalued or “inert” at the initial state and they are set to “weak” or “strong” during the course of development, once the L2er has acquired the overt inflectional morphology to which they are related. Thus, for example, the strength of the feature in Tense in the IL grammar of an English speaker acquiring French would be set to strong once the L2er had acquired verbal inflectional morphology. 2.2 Logical Problem of L2 Acquisition? The motivation for an innate language acquisition device driving L1 acquisition comes from the logical problem of language acquisition: children go beyond the input to which they are exposed. In recent years, researchers have argued that there is also a logical problem of L2 acquisition. Like L1 children, L2ers develop knowledge of complex and subtle properties of the target language which are underdetermined by the input (White, 1985; Schwartz & Sprouse, 2000). Note, however, that this does not necessitate that they make use of the same innate language acquisition device as L1 children. Unlike L1 children, L2ers know another language, and according to some researchers, this is the source of the complex and subtle L2 knowledge which they acquire (e.g. Bley-Vroman, 1990). Another important difference between L1 children and L2ers is that L2ers often receive language instruction. This could potentially also account for the development of sophisticated L2 knowledge. In order to demonstrate that there is a logical problem of L2 acquisition and that L2ers must use UG to overcome this problem, it is necessary to demonstrate that L2ers develop knowledge of complex and subtle properties of the target language which could not be derived from either (i) the L2 input, (ii) the L1, or (iii) instruction (White, 1989, 1990; Schwartz & Sprouse, 2000). It should also not be possible to acquire these properties using general learning mechanisms, such

< previous page

page_26

next page >

< previous page

page_27

next page >

Page 27 as analogy, frequency or the linear sequencing strategies mentioned above (White, 2003a, p. 23). In one of a series of studies on the L2 acquisition of French by native speakers of English, Dekydtspotter, Sprouse, and Thyre (1999/2000) demonstrate that adult L2ers are indeed capable of acquiring target language knowledge which must be derived from UG. The property of French under investigation is the interpretation of so-called discontinuous and continuous combien-questions. French allows the interrogative numeral determiner combien “how many” to appear either together with or separated from its nominal complement, as in (10)-a and (10)-b, respectively. (10) a. Combien de livres est-ce que tous les étudiants lisent? how many of books is-it that all the students read “How many books do all the students read?” b. Combien est-ce que tous les étudiants lisent de livres? how many is-it that all the students read of books “How many books do all the students read?” These so-called continuous and discontinuous constituents differ in interpretation. Take the following situation: one student, Graham, reads The Tax Inspector, Oscar and Lucinda and Jack Maggs , and another student, Lorna, reads Oscar and Lucinda , Jack Maggs and Illywhacker. In this situation, the answer to the question in (10)-a is either two, because there are two books which both Graham and Lorna read, namely Oscar and Lucinda and Jack Maggs (the so-called “wide-scope” reading), or three, the number of books per student (the “narrow-scope” reading). The only answer to question in (10)-b, on the other hand, is three.4 For the English-speaking L2er, the acquisition of this particular property of French represents a poverty of the stimulus problem (as it does for the L1 child acquiring French). There is nothing in the TL input or the L1 to prevent the L2er from assuming that the (much less frequent) discontinuous pattern is simply a rewrite of the continuous pattern because (i) the L1 equivalent, the continuous form given in the gloss in (10), allows both narrow- and widescope readings, and (ii) whenever the continuous combien interrogative is true, the discontinuous combien interrogative is true. Thus, Dekydtspotter et al. (1999/2000) argue, if L2ers demonstrate knowledge of this property of French, which furthermore is not covered in the L2 French classroom, this must result from the L2 hypothesis space being severely constrained and in a similar fashion to L1 acquisition. In a truth–value judgment task, Dekydtspotter et al. found that, like native speakers, advanced English-speaking adult L2ers of French make the relevant distinction between continuous and discontinuous questions. Such a distinction, Dekydtspotter et al. claim, “could not feasibly be acquired without a restricted relation between levels of syntactic and conceptual structure representations” (Dekydtspotter et al., 1999/2000, p. 170).

< previous page

page_27

next page >

< previous page

page_28

next page >

Page 28 Similar results are observed in several other studies focusing on the logical problem of L2 acquisition (e.g. Dekydtspotter, Sprouse, & Anderson, 1997; Kanno, 1998; Marsden, 2004). One advantage of this type of approach to L2 acquisition is that it is not deeply embedded within a particular linguistic theory, as are some generative studies of L2 acquisition (Schwartz & Sprouse, 2000). This means that when the linguistic theory in question is revised, this revision will not have potentially fatal consequences for the conclusion that, for example, L2 acquisition is constrained by UG. 2.3 Ultimate Attainment Most (adult) L2ers fail to reach the same level of knowledge as native speakers. This observation is often used as an argument against UG involvement in L2 acquisition: if L2ers make use of the same innate language acquisition device as children, who—under “normal” circumstances—are always successful, then, the argument goes, L2ers should be equally successful. In other words, whether or not L2ers are successful is measured in terms of how close they come to the native-speaker standard. Bley-Vroman (1983) dubbed this line of reasoning the comparative fallacy. He pointed out that “the learner’s system is worthy of study in its own right, not just as a degenerate form of the target system” (Bley-Vroman, 1983, p. 4). The question which ought to be asked in order to determine UG involvement in adult L2 acquisition is thus not whether IL grammars resemble native-speaker grammars but whether they are constrained by UG. In other words, the focus should be on whether IL grammars fall within the realms of natural language grammars, rather than whether they are identical to the particular natural language grammar which is the target of acquisition (e.g. Martohardjono, 1993; duPlessis, Solin, Travis, & White, 1987; Schwartz & Sprouse, 1994). Another objection to such a teleological approach to L2 acquisition is put forward by Schwartz (1990). She draws an analogy with L1 children who, in situations of language change, do not necessarily exactly reconstruct the grammar of the input providers, and with the grammar of English, which has changed over time (to wit, Old English, Middle English, and Modern English). Both cases involve different grammars, but no-one would claim that these grammars are epistemologically non-equivalent. Thus, Schwartz argues, if IL grammars fail to converge on the target, they, too, cannot necessarily be deemed epistemologically non-equivalent. L2ers may look different from L1 children at the end state, quite simply, because their initial state is different (Schwartz, 1998; Schwartz & Sprouse, 1994). Failure to reach a nativelike level of ultimate attainment cannot in and of itself be used against the claim that UG is involved in L2 acquisition, but it still requires explanation, of course. Various explanations for non-targetlike end states have been proposed within the generative paradigm. For example, in accounting for the consistently nontargetlike verbal inflectional mor-

< previous page

page_28

next page >

< previous page

page_29

next page >

Page 29 phology in a Chinese-speaking L2er of English clearly at her end state, Lardiere (1998) suggests that L2ers may experience problems in the specific mapping from syntax to morphology (see also e.g. Prévost & White, 2000). While their underlying syntactic representations may be targetlike, they may, for example, for processing reasons fail to consistently produce certain morphemes. Sorace (2003) accounts for the residual/persistent optionality she observes in end-state grammars as the L2ers’ failure to eradicate the non-targetlike parameter setting which they transferred from their L1, whereas Beck (1998), in an extension of Eubank’s (1993/94, 1994, 1996) Valueless Features hypothesis, proposes that feature strength in IL grammars is permanently impaired and that this accounts for optionality in, for example, verb raising (cf. (4)-b). In their Failed Functional Features Hypothesis, Hawkins and Chan (1997) argue that L2 adults cannot acquire features which differ from those in their L1; even when learners are relatively successful in acquiring a TL property which is not instantiated in their L1, their interlanguage grammars are, according to these authors, still derived from an L1-based analysis of the TL input. 2.4 Summary The generative approach to L2 acquisition seeks to determine whether IL grammars are constrained in the same way as native-speaker grammars. It has been claimed that, like L1 children, L2ers go beyond the input to which they are exposed, that is, there is a logical problem of L2 acquisition and in order to overcome this problem, L2ers must use UG. Most researchers agree that the initial state of L2 acquisition is characterized by some form of L1 transfer, although what exactly is transferred is subject to debate. With respect to the end state, it is important to note that, in and of itself, failure to achieve a nativelike level of ultimate attainment does not constitute an argument against UG involvement in L2 acquisition. 3 UG in Child L2 acquisition After having been the focus of an active research program in the 1970s (e.g. Hakuta, 1976; Felix, 1977; Cancino, Rosansky, & Schumann, 1978; Ervin-Tripp, 1978; Ravem, 1978; Wode, 1978, among others), child L2 acquisition had until recently somewhat taken the back seat in the field of generative L2 acquisition. As the previous section illustrates, in the 1980s and early 1990s the focus was mainly on adult L2ers. In recent years, however, child L2 acquisition has been the subject of renewed interest in the field (see Lakshmanan, 1995; Paradis, in press; Unsworth, 2005, for reviews). The term child L2 acquisition is generally used to denote L2ers whose first exposure to the second language occurs early in childhood, at a point at which the bulk of their first (native) language is already in place, say around age 4.

< previous page

page_29

next page >

< previous page

page_30

next page >

Page 30 The age at which child L2 acquisition becomes adult L2 acquisition is subject to considerable debate; it is an issue which is intrinsically linked to the debate about whether there is a critical period in L2 acquisition, that is, whether “there is a limited developmental period during which it is possible to acquire a language, be it L1 or L2, to normal, nativelike levels” (Birdsong, 1999, p. 1; see Hyltenstam & Abrahamsson, 2003a, for a recent review of the literature). This debate is far from resolved and various ages have been put forward as the start of a decline in the ability to reach nativelike levels of L2 proficiency: 3 years (Meisel, 2007), 5 years (Krashen, 1973), 7 years (Johnson & Newport, 1989; DeKeyser, 2000), 8 years (Bialystok & Miller, 1999), 9 years (Penfield & Roberts, 1959), puberty (Lenneberg. 1967), and 15 years (Long, 1990). Others have argued that there is no cut-off point, that is, that there is no critical period for L2 acquisition (e.g. Birdsong, 1992; Bongaerts, 1999; White & Genesee, 1996; Slabakova, 2006). Studies investigating the critical period hypothesis use child L2 acquisition to say something about adult L2 acquisition. By comparing L2 children and L2 adults, it is possible to determine whether L2 acquisition proceeds in the same fashion for both groups, that is, whether, as a result of biological and/or cognitive and/or sociological factors, the acquisition of an L2 as an adult is fundamentally different (Bley-Vroman, 1989) from the acquisition of an L2 as a child. The standard assumption made throughout the literature is that child L2 acquisition is constrained by UG. The evidence for this assumption comes from the critical period studies mentioned above which observe that, on the whole, L2 children are more successful than L2 adults (e.g. DeKeyser, 2000; Johnson & Newport, 1989, 1991). Recently, however, this assumption has been questioned: Hyltenstam and Abrahamsson (2003b) present data showing non-nativelike levels of ultimate attainment in the child L2 acquisition of Swedish. Meisel (2007) claims that only those L2 children who are exposed to the L2 before the age of 3 or 4 will pattern similarly to L1 children. Virtually all critical period studies focus on ultimate attainment, and in order to ensure that the L2 children have indeed reached their end state, they are usually tested when they are adults. Not only is this focus on ultimate attainment problematic for the reasons outlined in the preceding section, it also means that nothing can be said about child L2 development . Irrespective of the level of ultimate attainment which L2 children reach, it is possible that the route they took on their way to this (non-)nativelike end state differs from L1 acquisition (Schwartz, to appear). To obtain more insight into all three types of acquisition—child L2, adult L2, and L1—we need to investigate child L2 development and systematically compare this, on the one hand, with L1 development, and on the other, with adult L2 development (Schwartz, 1992, 2003, 2004; Unsworth, 2005; van de Craats, 2000). In Unsworth (2005), I compare English-speaking L2 children and adults with L1 Dutch children in their acquisition of direct object scrambling in

< previous page

page_30

next page >

< previous page

page_31

next page >

Page 31 Dutch. I observe that in their production of scrambled objects, the L2 children and adults in this study pass through the same developmental sequence and that this sequence subsumes the one which L1 children are observed to follow. The major difference between the L2 and L1 sequences is the presence of an initial stage of L1 transfer for the L2ers. Thus, like L2 adult development, child L2 development is also characterized by L1 transfer (Haberzettl, 1999; Whong-Barr & Schwartz, 2002; Haznedar, 1997; but cf. Zdorenko & Paradis, 2008). The comparison between L1 and L2 children has been particularly fruitful in assessing approaches to L1 acquisition which involve some element of maturation. Specifically, several researchers (e.g. Haznedar & Schwartz, 1997; Ionin & Wexler, 2002; Prévost, 2003; Schwartz, 2004) have argued that given that L2 children are cognitively more mature than L1 children, developmental patterns in L1 acquisition which are explained by making recourse to maturation should not be found in child L2 acquisition when the L2 children are beyond the relevant maturational point. The area which has generated the most research using this logic is the phenomenon of Root Infinitives (cf. (6) above). L1 children (both monolingual and bilingual), acquiring languages such as Dutch, French, German, and English, have been observed to pass through a stage where their declarative matrix clauses regularly contain nonfinite forms where a finite verb is required. Two of the accounts put forward to explain this phenomenon, Rizzi’s (1993/1994) Truncation Hypothesis and Wexler’s (1998) Very Early Parameter Setting account, are both couched in terms of maturation. Both propose that the RI stage in L1 acquisition is maturationally driven and that by about the age of 3 children leave this developmental stage and, consequently, finite verb forms replace the non-finite forms. Such approaches make very clear-cut predictions for child (and adult) L2 acquisition: L2 children should not pass through an RI stage of the same type as L1 children because (on the definition of child L2 acquisition adopted here), they are beyond the relevant maturational point. Several studies have investigated whether this is the case, but as yet the results are mixed (Ionin & Wexler, 2002; Haznedar, 1997; Prévost, 2003; Grondin & White, 1996). To summarize, child L2 acquisition, the study of which is witnessing resurgence in generative L2 acquisition, is usually assumed to be constrained by UG. By adopting a comparative approach, child L2 acquisition can be employed to increase our understanding of adult L2 acquisition, on the one hand, and L1 acquisition, on the other. Investigating child L2 development—in addition to ultimate attainment, the focus of studies investigating the critical period hypothesis—will also allow us to find out more about child L2 acquisition itself. In order to achieve this goal, more data from L2 children are needed.

< previous page

page_31

next page >

< previous page

page_32

next page >

Page 32 4 What the UG Approach Doesn’t Do, What It Does Do, and What It Still Needs To Do The generative approach to language acquisition is often criticized because it is rather limited. For example, it has nothing to say about how acquisition is affected by social factors, or about which factors affect performance. Rather, it seeks to “describe and explain the nature of interlanguage [or the child’s developing—SU] competence, defined in a technical and limited sense” (White 2003a, p. xi). It is thus concerned with grammar in its narrowest sense rather than language in a broad sense. Most of the research in this paradigm—and consequently this chapter—concentrates on linguistic representations. In addition to describing and explaining what a learner’s developing grammar looks like at a given moment in time, it is also necessary to describe how learners’ developing grammars develop over time. In other words, in addition to a “property theory” of language acquisition, a “transition theory” is also needed (Gregg, 1996), that is, a theory which will explain how development comes about and what its causes are. This aspect of the acquisition process, dubbed by Felix (1984) as the “developmental problem of language acquisition,” has on the whole been given short shrift in the generative paradigm, especially with respect to L2 acquisition (Gregg, 1996; Carroll, 2001). Carroll’s (1999, 2001) work constitutes one of the few attempts to address this problem for L2 acquisition. One of the advantages of the generative approach to L2 acquisition is its formality. Formalisms are needed to characterize the learner systems which are the object of study (Gregg, 1989); and unlike other formal approaches to language, the generative approach is also concerned with acquisition (Gregg, 1989, p. 31, fn. 18). Furthermore, the generative approach to language acquisition can be used to make clear and falsifiable predictions. The linguistic principles which it uses are also independently motivated; the task of the acquisition researcher is to determine whether and how they are acquired. Notes 1. For more on the milestones which L1 English children pass through, see, for example, the pioneering work of Brown (1973) and De Villiers and De Villiers (1985). 2. Functional categories are contrasted with lexical categories (or contentives). Lexical categories have meaning or content, e.g. nouns, verbs, adjectives, whereas functional categories, such as tense, inflection, determiners, do not. 3. This does not mean that language acquisition is complete at this age: some aspects of language, e.g. relating to the more complex properties of discourse, are acquired later. 4. According to Dekydtspotter et al. (1999/2000), the unavailability of the wide-scope reading in (10)-b results from the interaction of various syntactic and interface constraints (Obenauer, 1984/1985; Diesing, 1992), the details of which need not concern us here.

< previous page

page_32

next page >

< previous page

page_33

next page >

Page 33 References Beck, M. L. (1998). L2 acquisition and obligatory head movement. Studies in Second Language Acquisition , 20, 311– 348. Bialystok, E., & Miller, B. (1999). The problem of age in second-language acquisition: Influences from language, structure and task. Bilingualism: Language and Cognition, 2 , 127–145. Birdsong, D. (1992). Ultimate attainment in second language acquisition. Language, 68, 706–755. Birdsong, D. (1999). Introduction: Whys and why nots of the Critical Period Hypothesis for second language acquisition. In D. Birdsong (Ed.), Second language acquisition and the critical period hypothesis (pp. 1–22). Mahwah, NJ: Lawrence Erlbaum. Bley-Vroman, R. (1983). The comparative fallacy in interlanguage studies: The case of systematicity. Language Learning , 33, 1–17. Bley-Vroman, R. (1989). What is the logical problem of foreign language learning? In S. M. Gass & J. Schachter (Eds.), Linguistic perspectives on second language acquisition (pp. 42–68). Cambridge: Cambridge University Press. Bley-Vroman, R. (1990). The logical problem of foreign language learning. Linguistic Analysis , 20, 3–49. Bongaerts, T. (1999). Ultimate attainment in foreign language pronunciation: The case of very advanced late foreign language learners. In D. Birdsong (Ed.), Second language acquisition and the critical period hypothesis (pp. 133– 159). Mahwah, NJ: Erlbaum. Borer, H., & Wexler, K. (1992). Bi-unique relations and the maturation of grammatical principles. Natural Language and Linguistic Theory , 10, 147–189. Boser, K., Lust, B., Santelmann, L., & Whitman, J. (1992). The syntax of CP and V2 in early child German (ECG). The Strong Continuity Hypothesis. NELS , 22, 51–65. Brown, R. (1973). A first language. Cambridge, MA: Harvard University Press. Brown, R., & Hanlon, C. (1970). Derivational complexity and order of acquisition in child speech. In J. Hayes (Ed.), Cognition and the development of language (pp. 11–53). New York: John Wiley. Cancino, H., Rosansky, E. J., & Schumann, J. H. (1978). The acquisition of English negatives and interrogatives by native Spanish speakers. In E. M. Hatch (Ed.), Second language acquisition. A book of readings (pp. 207–230). Rowley, MA: Newbury House. Carroll, S. E. (1999). Putting “input” in its proper place. Second Language Research, 15, 337–388. Carroll, S. E. (2001). Input and evidence: The raw material of second language acquisition. Amsterdam: John Benjamins. Chomsky, N. (1959). A review of B. F. Skinner’s Verbal Behaviour . Language, 35, 26–58. Chomsky, N. (1965). Aspects of the theory of syntax . Cambridge, MA: MIT Press. Chomsky, N. (1975). Reflections on language. New York: Pantheon Books. Chomsky, N. (1981). Lectures on government and binding . Dordrecht: Foris. Chomsky, N. (1986). Knowledge of language: Its nature, origin, and use . New York: Praeger.

< previous page

page_33

next page >

< previous page

page_34

next page >

Page 34 Chomsky, N. (1988). Language and problems of knowledge. The Managua lectures . Cambridge, MA: MIT Press. Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press. Chomsky, N. (1999). On the nature, use and acquisition of language. In T. K. Bhatia & W. C. Ritchie (Eds.), Handbook of child language acquisition (pp. 33–54). San Diego: Academic Press. Clahsen, H., & Muysken, P. (1986). The availability of Universal Grammar to adult and child learners—a study of the acquisition of German word order. Second Language Research, 2 , 93–119. Clahsen, H., & Muysken, P. (1989). The UG paradox in L2 acquisition. Second Language Research, 5 , 1–29. Clahsen, H., Eisenbeiss, S., & Vainikka, A. (1994). The seeds of structure: A syntactic analysis of the acquisition of case marking. In T. Hoekstra & B. D. Schwartz (Eds.), Language acquisition studies in Generative Grammar (pp. 85– 118). Amsterdam: John Benjamins. Corder, S. P. (1967). The significance of learners’ errors. International Review of Applied Linguistics , 5 , 161–170. Crain, S., & Thornton, R. (1998). Investigations in Universal Grammar. A guide to experiments on the acquisition of syntax and semantics. Cambridge, MA: MIT Press. De Villiers, J. G., & De Villiers, P. (1985). The acquisition of English. In D. I. Slobin (Ed.), The crosslinguistic study of language acquisition (pp. 27–139). Hillsdale, NJ: Lawrence Erlbaum Associates. DeKeyser, R. M. (2000). The robustness of critical period effects in second language acquisition. Studies in Second Language Acquisition , 22, 499–533. Dekydtspotter, L., Sprouse, R. A., & Anderson, B. (1997). The interpretive interface in L2 acquisition: The process– result distinction in English–French interlanguage grammars. Language Acquisition , 6 , 297–332. Dekydtspotter, L., Sprouse, R. A., & Thyre, R. (1999/2000). The interpretation of quantification at a distance in English–French interlanguage: Domain specificity and second-language acquisition. Language Acquisition , 8 , 265– 320. Deprez, V., & Pierce, A. (1994). Crosslinguistic evidence for functional projections in early child grammar. In T. Hoekstra & B. D. Schwartz (Eds.), Language acquisition studies in Generative Grammar (pp. 57–84). Amsterdam: John Benjamins. Diesing, M. (1992). Indefinites . Cambridge, MA: MIT Press. duPlessis, J., Solin, D., Travis, L., & White, L. (1987). UG or not UG, that is the question: A reply to Clahsen and Muysken. Second Language Research, 3 , 56–75. Epstein, S., Flynn, S., & Martohardjono, G. (1996). Second language acquisition: theoretical and experimental issues in contemporary research. Brain and Behavioral Sciences, 19, 677–758. Ervin-Tripp, S. M. (1978). Is second language learning like the first? In E. M. Hatch (Ed.), Second language acquisition. A book of readings (pp. 190–206). Rowley, MA: Newbury House. Eubank, L. (1993/1994). On the transfer of parametric values in L2 development. Language Acquisition , 3 , 183–208. Eubank, L. (1994). Optionality and the initial state in L2 development. In T. Hoekstra & B. D. Schwartz (Eds.), Language acquisition studies in Generative Grammar . Amsterdam: John Benjamins.

< previous page

page_34

next page >

< previous page

page_35

next page >

Page 35 Eubank, L. (1996). Negation in early German–English interlanguage: More valueless features in the L2 initial state. Second Language Research, 12, 73–106. Felix, S. W. (1977). Interference, interlanguage and related issues. In C. Molony, H. Zobl, & W. Stölting (Eds.), Deutsch im Kontakt mit Anderen Sprachen [German in contact with other languages] (pp. 237–258). Kronberg: Scriptor. Felix, S. W. (1984). Maturational aspects of Universal Grammar. In A. Davies, C. Criper, & A. P. R. Howatt (Eds.), Interlanguage (pp. 133–161). Edinburgh: Edinburgh University Press. Finer, D., & Broselow, E. (1986). Second language acquisition of reflexive-binding. In S. Berman, J.-W. Choe, & J. McDonough (Eds.), Proceedings of NELS 16 (pp. 154–168). Amherst: UMass at Amherst, Graduate Linguistics Students Association. Gregg, K. R. (1989). Second language acquisition theory: The case for a generative perspective. In S. M. Gass & J. Schachter (Eds.), Linguistic perspectives on second language acquisition (pp. 15–40). Cambridge: Cambridge University Press. Gregg, K. R. (1996). The logical and developmental problems of second language acquisition. In W. C. Ritchie & T. K. Bhatia (Eds.), Handbook of second language acquisition (pp. 49–81). San Diego: Academic Press. Grondin, N., & White, L. (1996). Functional categories in child L2 acquisition of French. Language Acquisition , 5 , 1– 34. Guasti, M.-T. (2002). The growth of grammar . Cambridge, MA: MIT Press. Haberzettl, S. (1999). Katze Maus essen vs. Katze essen Maus: Die L1 als Königs- oder Holzwegbereiter zur L2? Zum Einfluss des L1-Wissens im Erwerb der deutschen Verbstellung durch türkische und russische Kinder. In H. O. Spillmann & I. Warnke (Eds.), Internationale Tendenzen der Syntaktik, Semantik und Pragmatik [International trends in syntax, semantics, and pragmatics] (pp. 157–165). Frankfurt: Peter Lan. Hakuta, K. (1976). A case study of a Japanese child learning English as a second language. Language Learning , 26, 321–351. Hale, K. (1996). Can UG and the L1 be distinguished in L2 acquisition? Brain and Behavioral Sciences, 19, 728–730. Hawkins, R., & Chan, C. Y.-H. (1997). The partial availability of Universal Grammar in second language acquisition: The “failed functional features hypothesis.” Second Language Research, 13, 187–226. Haznedar, B. (1997). L2 acquisition by a Turkish-speaking child: Evidence of L1 influence. In E. Hughes, M. Hughes, & A. Greenhill (Eds.), Proceedings of the 21st Boston University Conference on Language Development . Somerville, MA: Cascadilla Press. Haznedar, B., & Schwartz, B. D. (1997). Are there Optional Infinitives in child L2 acquisition? In E. Hughes, M. Hughes, & A. Greenhill (Eds.), Proceedings of the 21st Boston University Conference on Language Development (pp. 257–268). Somerville, MA: Cascadilla Press. Hirakawa, M. (1990). A study of the L2 acquisition of English reflexives. Second Language Research, 6 , 60–85. Hyams, N. (1992). The genesis of clause structure. In J. M. Meisel (Ed.), The acquisition of verb placement (pp. 371–400). Dordrecht: Kluwer. Hyltenstam, K., & Abrahamsson, N. (2003a). Maturational constraints in SLA. In C. Doughty & M. H. Long (Eds.), The handbook of second language acquisition (pp. 539–588). Oxford: Blackwell.

< previous page

page_35

next page >

< previous page

page_36

next page >

Page 36 Hyltenstam, K., & Abrahamsson, N. (2003b). Age of onset and ultimate attainment in near-native speakers of Swedish. In K. Fraurud & K. Hyltenstam (Eds.), Multilingualism in global and local perspective. Selected papers from the 8th Nordic Conference on Bilingualism (pp. 319–340). Stockholm: Centre for Research on Bilingualism and Rinkeby Institute of Multilingual Research. Ionin, T., & Wexler, K. (2002). Why is “is” easier than ’-s? Acquisition of tense/agreement morphology by child second language learners of English. Second Language Research, 18, 95–136. Johnson, J. S., & Newport, E. L. (1989). Critical period effects in second language learning: The influence of maturational state on the acquisition of English as a second language. Cognitive Psychology, 21, 60–99. Johnson, J. S., & Newport, E. L. (1991). Critical period effects on universal properties of language: The status of subjacency in the acquisition of a second language. Cognition, 39, 215–258. Kanno, K. (1998). The stability of UG principles in second-language acquisition: Evidence from Japanese. Linguistics , 36, 1125–1146. Krashen, S. (1973). Lateralization, language learning and the critical period: Some new evidence. Language Learning , 23, 63–74. Lakshmanan, U. (1995). Child second language acquisition of syntax. Studies in Second Language Acquisition , 17, 301–329. Lardiere, D. (1998). Case and tense in the “fossilized” steady state. Second Language Research, 14, 1–26. Lenneberg, E. (1967). Biological foundations of language. New York: John Wiley. Lillo-Martin, D. (1999). Modality effects and modularity in language acquisition: The acquisition of American Sign Language. In T. K. Bhatia & W. C. Ritchie (Eds.), Handbook of language acquisition (pp. 531–567). San Diego: Academic Press. Long, M. H. (1990). Maturational constraints on language development. Studies in Second Language Acquisition , 12, 251–285. Marcus, G. F. (1993). Negative evidence in language acquisition. Cognition, 46, 53–85. Marsden, H. (2004). Quantifier scope in non-native Japanese: A comparative interlanguage study of Chinese, English and Korean-speaking learners . Unpublished PhD dissertation, University of Durham. Martohardjono, G. (1993). Wh-movement in the acquisition of a second language: A crosslinguistic study of three languages with and without movement. Unpublished PhD dissertation, Cornell University. Meisel, J. M. (1997). The acquisition of the syntax of negation in French and German: Contrasting first and second language development. Second Language Research, 13, 227–263. Meisel, J. M. (2007). Child second language acquisition or successive first language acquisition? Hamburg Working Papers in Multilingualism, 80, 33–64. Montrul, S. A. (2000). Transitivity alternations in L2 acquisition: toward a modular view of transfer. Studies in Second Language Acquisition , 22, 229–273. Montrul, S. A. (2001). First-language-constrained variability in the second-language acquisition of argumentstructure-changing morphology with causative verbs. Second Language Research, 17, 144–194. Obenauer, H.-G. (1984/1985). On the identification of empty categories. The Linguistic Review, 4 , 153–202.

< previous page

page_36

next page >

< previous page

page_37

next page >

Page 37 Paradis, J. (in press). Child second language acquisition. In E. Hoff & M. Shatz (Eds.), Handbook on language acquisition. Oxford: Blackwell. Penfield, W., & Roberts, L. (1959). Speech and Brain Mechanisms. Princeton: Princeton University Press. Poeppel, D., & Wexler, K. (1993). The full competence hypothesis of clause structure in early German. Language, 69, 1–33. Prévost, P. (2003). Truncation and missing inflection in initial child L2 German. Studies in Second Language Acquisition , 25, 65–97. Prévost, P., & White, L. (2000). Missing surface inflection or impairment in second language acquisition? Evidence from tense and agreement. Second Language Research, 16, 103–133. Ravem, R. (1978). Two Norwegian children’s acquisition of English syntax. In E. M. Hatch (Ed.), Second language acquisition: A book of readings (pp. 148–154). Rowley, MA: Newbury House. Rizzi, L. (1993/1994). Some notes on linguistic theory and language development: The case of root infinitives. Language Acquisition , 3 , 371–393. Schachter, J. (1989). Testing a proposed universal. In S. M. Gass & J. Schachter (Eds.), Linguistic perspectives on second language acquisition (pp. 73–88). Cambridge: Cambridge University Press. Schwartz, B. D. (1990). Un-motivating the motivation for the Fundamental Difference Hypothesis. In H. Burmeister & P. Rounds (Eds.), Variability in second language acquisition (pp. 667–684). Eugene, OR: University of Oregon. Schwartz, B. D. (1992). Testing between UG-based and problem-solving models of L2A: Developmental sequence data. Language Acquisition , 2 , 1–19. Schwartz, B. D. (1998). The second language instinct. Lingua, 106 , 133–160. Schwartz, B. D. (2003). Child L2 acquisition: Paving the way. In B. Beachley, A. Brown, & F. Conlin (Eds.), Proceedings of the 27th Boston University Conference on Language Development (pp. 26–50). Somerville, MA: Cascadilla Press. Schwartz, B. D. (2004). Why child L2 acquisition? In J. Van Kampen & S. Baauw (Eds.), Proceedings of Generative Approaches to Language Acquisition 2003 (pp. 47–66). Utrecht, Netherlands: LOT Occasional Series. Schwartz, B. D. (to appear). Unraveling inflection in child L2 development. Morphology. Schwartz, B. D., & Sprouse, R. A. (1994). Word order and nominative case in non-native language acquisition: A longitudinal study of L1 Turkish–German interlanguage. In T. Hoekstra & B. D. Schwartz (Eds.), Language acquisition studies in Generative Grammar (pp. 317–368). Amsterdam: John Benjamins. Schwartz, B. D., & Sprouse, R. A. (1996). L2 cognitive states and the Full Transfer/Full Access model. Second Language Research, 12, 40–72. Schwartz, B. D., & Sprouse, R. A. (2000). When syntactic theories evolve: Consequences for L2 acquisition research. In J. Archibald (Ed.), Second language acquisition and linguistic theory (pp. 156–186). Oxford: Blackwell. Schwartz, B. D., & Tomaselli, A. (1990). Some implications from an analysis of German word order. In W. Abraham, W. Kosmeijer, & E. Reuland (Eds.), Issues in Germanic syntax (pp. 252–274). Berlin: Mouton de Gruyter. Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics , 10, 209–231. Slabakova, R. (2006). Is there a critical period for semantics? Second Language Research, 22(3), 302–338.

< previous page

page_37

next page >

< previous page

page_38

next page >

Page 38 Sorace, A. (2003). Near-nativeness. In C. Doughty & M. H. Long (Eds.), The handbook of second language acquisition (pp. 130–151). Oxford: Blackwell. Unsworth, S. (2005). Child L2, adult L2, child L1: Differences and similarities. A study on the acquisition of direct object scrambling in Dutch . Unpublished PhD thesis, University of Utrecht. Vainikka, A. (1993/1994). Case in the development of English syntax. Language Acquisition , 3 , 257–325. Vainikka, A., & Young-Scholten, M. (1994). Direct access to X-bar theory: Evidence from Korean and Turkish adults learning German. In T. Hoekstra & B. D. Schwartz (Eds.), Language acquisition studies in Generative Grammar (pp. 265–316). Amsterdam: John Benjamins. Vainikka, A., & Young-Scholten, M. (1996a). Gradual development of L2 phrase structure. Second Language Research, 23, 7–39. Vainikka, A., & Young-Scholten, M. (1996b). The early stages in adult L2 syntax: Additional evidence from Romance speakers. Second Language Research, 12, 140–176. van de Craats, I. (2000). Conservation in the acquisition of possessive constructions. Unpublished PhD dissertation, Catholic University of Brabant. Weissenborn, J. (1990). Functional categories and verb movement: The acquisition of German syntax reconsidered. In M. Rothweiler (Ed.), Spracherwerb und Grammatik. Linguistische Untersuchungen zum Erwerb von Syntax und Morphologie [Language acquisition and grammar. Linguistic investigations into the acquisition of syntax and morphology] (pp. 190–223). Opladen: Westdeutscher Verlag. Werker, J. F., & Tees, R. C. (1984). Cross-language speech perception: Evidence for perceptual reoganization during the first year of life. Infant Behavior and Development , 7 , 49–63. Wexler, K. (1998). Very early parameter setting and the Unique Checking Constraint: A new explanation for the Optional Infinitive stage. Lingua, 106 , 23–79. White, L. (1985). Is there a logical problem of second language acquisition? TESL Canada, 2 , 29–41. White, L. (1989). Universal Grammar and second language acquisition. Amsterdam: John Benjamins. White, L. (1990). Second language acquisition and Universal Grammar. Studies in Second Language Acquisition , 12, 121–133. White, L. (2003a). Second language acquisition and Universal Grammar . Cambridge: Cambridge University Press. White, L. (2003b). On the nature of interlanguage representation: Universal Grammar in the second language. In C. Doughty & M. H. Long (Eds.), The handbook of second language acquisition (pp. 19–42). Oxford: Blackwell. White, L., & Genesee, F. (1996). How native is near-native? The issue of ultimate attainment in adult second language acquisition. Second Language Research, 12, 238–265. Whong-Barr, M., & Schwartz, B. D. (2002). Morphological and syntactic transfer in child L2 acquisition of the English dative alternation. Studies in Second Language Acquisition , 22, 579–616. Wijnen, F. (1997). Temporal reference and eventivity in root infinitives. MIT Occasional Papers in Linguistics , 12, 1– 25.

< previous page

page_38

next page >

< previous page

page_39

next page >

Page 39 Wode, H. (1978). Developmental sequences in naturalistic L2 acquisition. In E. M. Hatch (Ed.), Second language acquisition: A book of readings (pp. 101–117). Rowley, MA: Newbury House. Zdorenko, T., & Paradis, J. (2008). The acquisition of articles in child L2 English: Fluctuation, transfer or both? Second Language Research, 24, 227–250.

< previous page

page_39

next page >

< previous page

page_40

next page >

Page 40 2 FIRST LANGUAGE ACQUISITION FROM A USAGE-BASED PERSPECTIVE Heike Behrens 1 Introduction Chomsky’s theory of language acquisition shed doubt on the assumption that language can be learned on the basis of experience (see Unsworth, this volume). But currently, input-driven or constructivist theories of language acquisition gain ground quickly, as both corpus and experimental methods allow us to investigate children’s learning processes in more detail. There are different labels for such theories, for instance, usage-based theory of acquisition (Tomasello, 2000; 2003; Goldberg, 2006), or emergentism (Elman et al., 1996; MacWhinney, 1999; O’Grady, 2005). Usage-based theories emphasize that language learning in children is the result of their experience with language use, and consequently, adult linguistic representations are usage-based, not innate. We build linguistic categories through repeated exposure to similar or related structures (e.g. Bybee, 2006; Ellis, 2008). The term “emergentism” emphasizes the progressive nature of the process, namely that qualitatively new and more complex representations can emerge on the basis of knowing its simpler component parts (O’Grady, 2005). These approaches are constructivist because they assume that children, when learning the language, “reconstruct” the system by generalizing over usage-events, that is, utterances they have heard or produced, rather than activating an innate representation of grammar. Proponents of usage-based theories assume that language, although it seems to be a very complex capacity that is unique to the human species, can be explained by the same cognitive processes that are attested in other domains that have to be learned. Language learning is based on a combination of social cognition, pattern recognition, and efficient general learning mechanisms. These skills supposedly suffice to acquire more and more complex features of language within a couple of years. This approach differs from Generative Grammar (see Unsworth, this

< previous page

page_40

next page >

< previous page

page_41

next page >

Page 41 volume) in a number of ways. It explicitly rejects the assumptions that humans are genetically equipped with linguistic representations in the form of an innate “Universal Grammar” (Elman et al., 1996; Tomasello, 1995, 2003). Furthermore, usage-based models of grammar do not stipulate principles or constraints that are special to language. Rather, linguistic categories can be derived by abstraction over experience (see Section 6, “Generalization”). This difference leads to different assumptions regarding language learning mechanisms. If the general principles of language are innate, the child only has to activate the particular settings relevant for the target language (e.g. triggers, maturation, linking rules or other purely linguistic learning mechanisms by which innate features of Universal Grammar are activated). But if children are born with specific linguistic representations, they need to acquire increasingly abstract linguistic schemas by noticing direct and relational analogical information (see below). In order to do so, children can make use of their social cognition for cultural learning. Language is a means of communication, and to communicate well requires that we understand what our discourse partners think or want. To solve this task, children need an understanding of how people behave in certain situations, as well as an understanding that the state of mind of the other person might be different from their own, the so-called theory of mind (Premack & Woodruff, 1978; see Tomasello & Rakoczy, 2003, for a recent survey). Hence, social cognition and experience with other people’s behavior and expectations in social situations can help children to figure out what people want when they talk, that is, it provides a resource for the semantic and pragmatic interpretation of the speech signal. But children also have to identify the rule system underlying the utterances they hear. In order to do this, they can make use of powerful generalization skills like pattern recognition and stochastic or statistical learning. Both aspects of cognition combined with extended exposure to language use or input explain how language learning can be usage-based. In the next section I will first focus on the skills children show in the prelinguistic phase, before they start to produce language themselves, in order to demonstrate how these general social and cognitive skills contribute to language learning. I will then briefly discuss the early phases of language acquisition when children gradually build up their formal and semantic repertoire. Finally, I will focus on more complex aspects of grammar in order to demonstrate how usage-based and input-driven approaches explain the acquisition of more abstract and less “obvious” rules. The article concludes with a summary of the major differences between generative and usage-based approaches to language acquisition.

< previous page

page_41

next page >

< previous page

page_42

next page >

Page 42 2 Prelinguistic Development 2.1 Development of Social Cognition Before and while infants learn to talk, they develop aspects of social cognition which are instrumental for their success in communication. Eight to 9-month-old children start to show a number of social behaviors like gaze following, social referencing, and joint attention. They become actively interested in what others are interested in (gaze following) and they start to direct the attention of others to what they themselves are interested in (social referencing, e.g. pointing to exciting things or events). Joint attention refers to the awareness that the communicative partners are engaged in the same activity, like looking at a bug or having a conversation. Skills like these are a crucial prerequisite for communication because communication only works when the partners coordinate their attention. By 1 year of age, children have developed an understanding that other persons are intentional agents, that is, they understand that people’s behavior is goal-directed and start to pay more attention to the goal of the action than the action itself (Gergely, Nadasdy, Csibra, & Biro, 1995; Meltzoff, 1995). Tomasello and Rakoczy (2003, p. 122) argue that these ontogenetic steps in social cognition may be unique to humans because it enables them “to pool their cognitive resources, that is, to create and participate in collective cultural activities and products.” Critically, this cognitive difference leads to different types of social learning. The first is emulation learning where one learns about objects. For example, chimpanzee babies may realize that termite hills contain tasty termites and by coincidence figure out how to fish these termites by sticking a thin branch into a hole of the termite hill. They may have observed that adult chimpanzees stick branches into termite hills, but they have not realized that the adults did so in order to fish for termites. In a certain sense, they reinvent the wheel by making discoveries of the species anew. The second type of social learning is imitative learning where one copies goal-directed strategies of others (Tomasello & Rakoczy, 2003, pp. 129–130). So far, this form of learning has been observed in human children, but not in the Great Apes. Crucially, in imitative learning one does not simply mimic the exact execution of an action. Rather, it requires an understanding of the intention of the other such that one can achieve what they did. For example, I can observe somebody ringing the bell with his head because he holds a very heavy box in his hand. If I understand the intention behind the action (= ringing the bell) I can do this more easily with my hands rather than mimic-king the action (= ringing the bell with my head). Only imitative learning leads to the cultural transmission of knowledge which is a key factor in the cognitive development of the human species. This ability distinguishes competitive from cooperative actions. Several species are able to take the

< previous page

page_42

next page >

< previous page

page_43

next page >

Page 43 knowledge of conspecifics into account when pursuing certain goals, like getting food when one knows that the higher ranking conspecific cannot see it. Human children, in addition, are able to cooperate with others in competitive and non-competitive situations because they share a certain goal like playing a game together (see Moll & Tomasello, 2007). Tomasello and colleagues argue that the shared intentionality shown even by prelinguistic children provides the social-cognitive basis for language use, an activity in which they share information and establish common goals. 3 Linguistic Development The first milestone of language acquisition seems to be the production of the first word, usually sometime around the child’s first birthday. However, research over the past three decades has shown that this is just the accumulation of developments. Apart from the social cognitive skills discussed above, children develop perceptual and pattern recognition skills. 3.1 Categorical Perception and Pattern Recognition One of the most exciting recent findings in acquisition research is that babies and even newborns have amazing abilities regarding speech recognition. Experiments showed that newborns recognize not only their mother’s voice, but also their mother tongue (e.g. French versus Russian), and even passages read aloud during pregnancy (Gomez & Gerken, 2000). “To recognize” in this context means that they show a different reaction to familiar than to unfamiliar sound sequences, not that they “understand” the stories in a semantic sense. New experimental techniques lead to insights in children’s perceptual abilities even in the womb. It is well attested that infants are able to discriminate a wide range of phonemic contrasts. For example, Dutch children perceive the distinction between certain phoneme contrasts in Hindi, a distinction that adult Dutch speakers are oblivious to. This ability diminishes towards the end of the first year of life as the children become more attuned to the ambient language(s). Thus, it seems that children come to the language learning task with a relatively wide predisposition for sound discrimination that enables them to learn any language, and subsequently lose this advantage but gain proficiency in native language skills (Clark, 2003, pp. 55–71). While the ability to discriminate sound contrasts is very crucial for the ability to acquire a spoken language, it is probably not a language-specific or human-only ability. Subsequent research demonstrated that this ability is not language-specific because infants also discriminate artificial sound contrasts, not species-specific in that some animals succeed in these tasks as well (cf. Clark, 2003; Jusczyk, 1997; Newport, Hauser, Spaepen, & Aslin, 2004). Similar results were observed in the categorization of semantic relations.

< previous page

page_43

next page >

< previous page

page_44

next page >

Page 44 Some languages focus on the topological (neighborhood) relationship between objects like in, on, or under. Other languages discriminate between the form of the container (e.g. some Mayan languages), or the nature of the contact between the objects, as in Korean. Again, it seems that children are able to learn the different types of spatial categorization quite easily. They show language-specific differences in spatial categorizations as young as age 2. As infants, however, they seem to be open to various forms of classification (cf. Bowerman & Choi, 2003). These findings suggest that even infants are able to recognize similarities across acoustic and visual stimuli. But how do they go about this? First, this requires memory in order to register repeated perceptual events. Second, this requires some level of abstraction if the stimuli are not completely identical (which is the typical case because in our natural environment something changes from situation to situation: shade of the light, position, background noise). What is required, in other words, is categorical perception. Categorical perception is a very efficient process because it allows us to ignore irrelevant detail and focus on the relevant or distinctive information. For example, the distinction between voiced and voiceless consonants (e.g. [b] vs. [p]) seems to be categorical: regardless of the exact onset of voicing, the feature that causes this contrast, we perceive a sound as either [b] or [p], not a sound that is somewhere in between. To categorize on a more abstract level helps to solve the “invariance problem” (Clark, 2003, p. 56): Children will need to recognize the sounds and units of a language even if, for example, the concrete pronunciation will differ from speaker to speaker and even situation to situation. Third, children need to be able to recognize recurrent patterns in order to identify the systematic properties of the language they are acquiring. That infants are good at recognizing patterns has been demonstrated by research on stochastic or statistical learning. 3.2 Stochastic Learning One of the first tasks in language learning is the segmentation problem involved in word recognition, since the sound stream we hear is usually continuous and does not provide pauses at word boundaries. A solution to this problem is stochastic learning (Saffran, 2003). Imagine an infant who frequently hears the sound sequence prettybaby. How would he or she figure out whether there is a word boundary and where it is? Prosodic cues are not fully reliable when trying to find out where a new word starts, but transitional probabilities between syllables turn out to be more reliable. Transitional probabilities measure the likelihood that syllable A is followed by syllable B. An analysis of input directed to English-speaking children showed that there is an 80% likelihood that pre- precedes -tty , but only a 0.03% likelihood that ty- precedes -ba. Therefore, pre-tty is a probable word; ty-ba is not (Saffran, 2003).

< previous page

page_44

next page >

< previous page

page_45

next page >

Page 45 In order to test whether humans (and non-humans) are able to derive such knowledge online, numerous experiments have been carried out. For example, the subjects were exposed to two minutes of input consisting of nonsense syllables (e.g. bidakupadotigolabubidakutupiropadoti), read without pauses and in a flat intonation. Then it was tested whether they recognized “words” consisting of three syllables. Sometimes these “words” had occurred in that order in the input (e.g. bidaku or golabu ), sometimes not (e.g. dapiku or tiladu ). It turned out that 8-month-old infants (as well as cotton-top tamarin monkeys) recognize the differences between “legal” and “illegal” three-syllable sequences. In order to do so, they must have registered the transitional probabilities between words because, obviously, they did not know these sequences before, nor did prosody or semantics help to solve the task (Saffran, Newport, & Aslin, 1996). There is a growing body of evidence that children are very efficient pattern recognizers (Newport & Aslin, 2004). Children even manage to repair inconsistencies in the input (Hudson-Kam & Newport, 2005). Also, children younger than 1 make use of distributional information to extract high frequency words from the input (Höhle & Weissenborn, 2000). There is ongoing research on whether these abilities are language- or species-specific (Newport et al., 2004) and it seems that this ability is not limited to the human species. This implies that this ability is not specifically “designed” for language but rather one of several cognitive prerequisites. 3.3 Structure Mapping and “Starting Small” Stochastic learning may be relevant for language learning in the following way. The point is not that children learn statistics; rather, they use statistical properties of the input to derive generalizations about language structure. In order to do so, they can generalize based on physical similarity, for example when they perceive the same sound sequence again and again. But in order to acquire more abstract linguistic categories, they have to succeed in structure mapping (Gentner, 2003) by noticing relational similarities between the same or similar items being used in different constellations. Consider the three sentences in (1). They share some aspects (the names of the people involved, the – s at the end of the second word), but to recognize the syntactic similarities between the sentences requires finding similarities that go beyond identifying shared material. In addition, children have to recognize that John relates to Anna in sentence (1a) like Anna relates to Jack in sentence (1b) and Jack to John in (1c). (1) a. John loves Anna b. Anna kisses Jack c. Jack hits John

< previous page

page_45

next page >

< previous page

page_46

next page >

Page 46 Depending on the language they are acquiring, learners will have to find out whether such relational analogies between “subjects” and “objects” are encoded by word order, as in English, or by case marking (in addition to word order), as for example in German. But how do children come to notice such relational similarities? Surprisingly, they may be helped by their limited working memory. Originally, Generative Grammarians made the point that powerful generalization mechanisms may be an impediment to language learning because children could keep track of all input data and discover hundreds of possible relations between linguistic units, and thus come up with many meaningless generalizations (Pinker, 1984). In the nativist view, Universal Grammar would reduce the hypothesis space because the child has innate constraints regarding relevant linguistic structures. The usage-based stance differs in two respects: First, children take communicative intent into account as well as the feedback they receive in interactions (Tomasello, 2003; Chouinard & Clark, 2003). Second, children are actually helped by their small working memory. A small working memory automatically focuses children’s attention to relations between adjacent elements, which is where most linguistic relations are encoded. Thus, children’s limited working memory capacity may help them to learn language successfully (Newport, 1990), because it prevents them from being able to carry out too many irrelevant computations. The starting small hypothesis was tested with computational networks. It turned out that networks could process complex sentences more efficiently when they were trained with shorter rather than complex sequences (Elman, 1993). Freudenthal, Pine, and Gobet (2002, 2005) trained a computer model to learn verbal inflection in English and Dutch. Initially, the model had only a limited working memory and was programmed to pay attention to the end of utterances. Gradually, the model’s memory was extended. With these manipulations that correspond to the focus of attention of young children, the researchers could simulate the course of acquisition in these languages. Typically, Dutch (and German) children start out with infinitives like maken “make,” followed by utterances with a subject or object preceding the infinitive (e.g. grapje maken “make a joke”) . This order of acquisition reflects right to left processing in transitive sentences like (2) because the rightmost element is best remembered. (2) Papa zal grapje maken Daddy shall joke make-INF Thus, acquisition can be modeled successfully if we limit generalization to those aspects of language children are able to grasp because they are perceptually salient and/or within easy reach of the working memory (see Gobet, Freudenthal, & Pine, 2007, for a refined version of the model).

< previous page

page_46

next page >

< previous page

page_47

next page >

Page 47 3.4 Summary on Early Linguistic Development From an emergentist perspective it seems that what children bring to the language learning task is a general predisposition for categorical perception, pattern recognition, statistical learning, as well as generalization skills that will help them to discover even the most complex rules of linguistic structure. In evolution, an organism that is a quick and efficient generalizer is far more flexible and adaptable than an organism with highly specialized genetic predispositions. This is how Gentner (2003) explains “why we are so smart.” According to her, the human flexibility in recognizing even very abstract commonalities or analogies (e.g. picking the middle one from two structured arrays of objects, even if the objects share no perceptual attributes) enables us to learn almost everything, even across perceptual domains. This ability can also act as a constraint on possible grammars. A language cannot survive if its grammar cannot pass through the learning bottleneck (Hurford, 2002). 4 Developmental Processes A theory of language development has to answer two questions. What is the nature of children’s linguistic representations in comparison to those of adults, and what are the developmental processes by which they change over time? Regarding the first question, usage-based theories assume that children start out with unanalyzed knowledge in the form of chunks of speech, and analyze the underlying structure in a bottom-up fashion without the help of top-down processes (e.g. a Universal Grammar that helps them to make sense of their linguistic environment). Regarding the second question, usage-based theories assume that children build up linguistic knowledge by gradual generalization. They thus refine their linguistic categories on a piecemeal basis. A classic view of language acquisition is that it proceeds from simple to complex. Children start out with holophrases (one-word stage), proceed to the two-word stage and from there to multi-word utterances. Likewise, simple phrases or sentences precede complex phrases or sentences. Much of the research in first language acquisition focused on establishing the order of acquisition of a number of phenomena in a number of languages (see, for example, Brown, 1973, and Slobin’s volumes on the cross-linguistic acquisition of language structure; Slobin, 1985–1995). Studies from a usage-based perspective tend to focus on another aspect that may be crucial for language acquisition, namely the communicative factor. If one sees communicative needs as the driving force, it follows that children should be sensitive to communicative acts, the typical unit of which is the utterance, not the isolated word or morpheme (Tomasello, 2003). Consequently, children may pick up larger units with a specific function, for example a chunk like wheresthe. They can use such chunks to ask for the

< previous page

page_47

next page >

< previous page

page_48

next page >

Page 48 location of people or objects without having analyzed the internal structure of the phrase into words ( where is the) or grammatical categories ( wh-pronoun + auxiliary-3rd person singular + definite article). In the course of development they will realize that this chunk is composed of parts that are shared by other constructions (e.g. where’s X or is the X ). In this view, language learning is not just a linear process from simple to complex structures, but also one of dissecting the structure of larger units that they have acquired in a holistic fashion and analyzing the relationships that hold between the component parts within the utterance as well as to related utterances. The claim that children use formulas or formulaic sequences early on in their language development is not a new one. Several researchers stressed the item-based nature of children’s early word combinations. Braine (1963) proposed that early multiword utterances are built around a “pivot” or anchor point, and that later structures can be described as slot and frame patterns. Take an utterance like wanna ride horsie. If children are credited with an adult-like mental representation of grammar (continuity assumption, see Unsworth, this volume), this sentence would have a mental representation as “infinitival complement with omitted subject.” However, it could well be that the child represents this phrase like an idiom to encode wannaridehorsie without being aware of any relations of this constructions to other constructions. The mental representation is like that of a holophrase; the child has no or little knowledge about the internal structure of this phrase. This is similar to a second language learner who knows the phrase for ordering water but is unaware of the internal structure of that phrase. Once the child realizes that horsie or ride can be exchanged with other nouns or verbs, the mental representation is a slot and frame pattern (see below). Here, the child has a frame with one or more open slots into which new elements can be inserted ( wanna ride X , or wanna X Y). Tomasello (1992) suggested that many of those frames may be centered on individual verbs in so-called “verb-island constructions.” He argued that a characteristic feature of early child language is that the child does not draw links between individual verb islands, such that knowing something about one transitive verb like kiss does not entail that the child treats another transitive verb like bake the same way. He thus hypothesizes that the best predictor for a given verb’s use is the previous usage pattern of the same word. Children learn only gradually to generalize over such concrete item-based knowledge and to establish more abstract schemas. Subsequent research showed that children acquire complex structures in a rather piecemeal fashion (e.g. Tomasello, 2000). When children start to generalize, one finds limited scope formulae (Braine, 1963) or slot and frame patterns (e.g. Lieven, Pine, & Baldwin, 1997), that is, partially frozen word combinations with an open slot ( where’s the X , he’s Xing). For example, a child may be able to use different nouns in a given slot (e.g. where’s the car? where’s the milk? where’s the cat?) but only gradually draw a relation between this construction and other wh-constructions (e.g. where’s

< previous page

page_48

next page >

< previous page

page_49

next page >

Page 49 Daddy? where are my shoes? what’s that?). Often, just a few of such combinations account for a large percentage of children’s overall production of that structure. In a longitudinal study of 12 British children between 2 and 3, Pine, Lieven, and Rowland (1998) found that the five most frequent subject–verb combinations accounted for 66–90% of these children’s total production of subject–verb combinations. If children’s knowledge is lexically specific and item-based rather than adult-like and abstract, three predictions follow. First, children’s utterances should be quite conservative, that is, they rely on what they hear in the input. Second, they build up their own repertoire rather gradually rather than generating each new utterance from scratch. Third, children’s early language should be low in productivity; they are not able to generalize in an abstract fashion. There are only a few studies that compare children’s language with the input they have. Behrens (2003, 2006) found that by age 2;6 to 3 children’s language matches that of the adult in terms of the distribution of part-of-speech categories and noun phrase types, for example. This suggests that children pay close attention to the distributional properties of the ambient language (see also Cameron-Faulkner, Lieven, & Tomasello, 2003). The second prediction was tested, for example, by a method called “traceback” in which one tries to compute the proportion of “pre-fabricated units” compared to “new” utterances (Lieven, Behrens, Speares, & Tomasello, 2003). An English-speaking girl called Jessie was recorded 5 hours per week for six weeks, between age 2;0 and 2;1.29. In addition, her mother took diary notes throughout the day to sample the most complex utterances produced by Jessie. The authors looked at the last transcript of this period and compared the utterances in this transcript with the previous transcripts in order to find all utterances that were “new” in the sense of not having been said in exactly that form before. Of the 295 utterances of the child, only 37% were new, and 74% of the new utterances differed in just one aspect from utterances said before. In most cases, a word was substituted, or a word was added to an old construction. Fewer than 10% of all utterances differed in two or more aspects from what the child had said before. These figures demonstrate that a lot of young children’s language production is based on existing knowledge, for example well-rehearsed formulas or slot and frame patterns where one element in a construction can be substituted. But frequent repetition of structures that were acquired earlier does not constitute proof that the utterances are not productive, that is, produced on the basis of rules that the child has acquired. The third prediction is that early child language shows only limited productivity. Several methods have been developed to assess the underlying productivity of children’s utterances and will be discussed in more detail in the next section.

< previous page

page_49

next page >

< previous page

page_50

next page >

Page 50 5 Productivity The term productivity refers to the state of mastery. A structure is productive if the learner can use it in an adult-like fashion. Several criteria might be used to define productivity. The most striking case is overgeneralization where a rule is applied to situations where it should not apply, like irregular words. This affects morphology, for example, past tense or plural overgeneralization ( go-ed, mouse-s), as well as argument structure patterns, for example, the overgeneralization of causative constructions ( he disappeared the ball , he fell the cup). Another method for assessing productivity is to test children with nonce words, that is, made-up words that they cannot have heard in the input and must have learned by rote. In 1958, Jean Berko conducted a number of tests on children’s understanding of past tense or plural marking. She showed them pictures with new objects or actions and asked them to produce another form of the nonce word (e.g. Look, here’s a wug. See, there’s another one! Now there are two …?). If the children correspond correctly ( wugs) it shows that they have developed an understanding of the plural formation rules in English. It is important to note that correct (= error-free) usage does not imply productivity per se. In the beginning, verbal knowledge might be frozen or formulaic. The child’s utterance seems to be correct and error-free, but inflexible and not related to the full paradigm. Again, this is similar to cases in which L2 learners know particular phrases in the L2 without having a full understanding of the component structures and rules. Limited productivity can show up in a number of ways, for example, as undergeneralization when a term or structure is used only in a subset of possible contexts. For instance, the child applies the word dog only to the neighbor’s dog, or uses past tense morphology only with telic verbs to encode the resultant state. Undergeneralization is discovered by distributional analyses when we find that the child produces a more limited range than adults or older children. In the following section I will outline the generalization process as conceived of in usage-based theories, and show how children use the same learning mechanisms throughout development. 6 Generalization 6.1 From Concrete to Abstract Usage in Early Language Acquisition In usage-based theory, generalization is a bottom-up process (cf. Langacker, 2000, p. 4f.). Usage-events (instances) are registered, and repeated encounter leads to entrenchment when the memory traces become more robust and the item can be retrieved faster and more reliably. Also, stored

< previous page

page_50

next page >

< previous page

page_51

next page >

Page 51 items are compared to one another. In this comparison, children may detect analogies between related aspects of constructions and build low-level schemas. Thus, new mental representations emerge through reinforcement of the commonalities. For example, the child repeatedly hears past tense forms like knew , loved, made, faked, hated , swam, borrowed , kneeled, typed, and realizes that many forms have a dental stop as suffix. They may then generalize that the dental stop serves as marker for past tense and apply this to other verbs as well. Subsequently, children may hear more –ed forms and categorize the unit as “verb” and “past tense marker” accordingly. The result of such an abstraction is a schema that is based on the commonalities between structures on a higher level. Schemas can be more or less abstract. In low-level schemas, there are a number of surface similarities (e.g. the where’s the schema); in abstract schemas, all surface elements may differ (e.g. different instantiations of transitive sentences). Linguistic rules, then, are nothing but very general schemas. Tomasello (2003, p. 174) summarizes the cognitive processes necessary for generalization as follows: 1. On the level of an expression, children have to identify communicative intentions in the input (e.g. I wanna see it), and they have to be able to reproduce these expressions. 2. Children form a pivot schema by forming a schema on the one hand, and a slot-filler category on the other (e.g. throw ball , throw can , throw pillow). 3. Next, children form item-based constructions as second-order symbols (e.g. Mary hugs John, John hugs Mary). In this example, they know something about the different participant roles “hugging,” but may not be able to transfer this knowledge to other transitive verbs. 4. Then, children form abstract constructions like syntactic roles based on analogy ( A hugs B , X kisses Y). 5. And they form paradigmatic categories (e.g. part-of-speech categories or inflectional paradigms) based on categorization by distributional analysis. It is often difficult to find evidence for limited productivity in corpora of spontaneous speech, simply because these corpora are often too small to allow for detailed distributional analyses. Experiments with nonce words are used to test whether children can generalize their linguistic knowledge to words they have not heard before (Tomasello, 2000). This survey of results on experimental studies on children’s command of the transitive construction shows that it takes children some time to transfer their knowledge from one construction to verbs they have not heard before. For example, if they have only heard the nonce word meek in an objectless construction ( Look! He’s meeking) they will have problems producing it with a direct object ( He’s meeking the car) although they produce direct objects with other verbs . While the limited productivity of early child language is well attested, one may ask whether the same processes of generalization over their linguistic

< previous page

page_51

next page >

< previous page

page_52

next page >

Page 52 experience is still at stake with older children who learn more complex constructions. 6.2 The Acquisition of Complex Sentences Diessel and Tomasello (2000) analyzed four English-speaking children’s development of the relative clauses. The children were between 1;9 and 5;11 years of age, and the corpora contained a total of 324 relative clauses. They found that presentational constructions were the starting point for relative clauses, and like structures observed earlier in development, they show characteristics of slot and frame patterns in that they have a restricted set of onset sequences, for example, Here’s the X that or That’s the X that as in (3). (3) Here’s the toy that goes around That’s the sugar that fell out A limited number of such onsets account for a large portion of the relative clauses produced in the earliest transcripts. For example, 75% of the earliest relative clauses were presentational ( This is the girl that turns around), 17.5% were object relatives ( She has a bathtub that goes with it), and the remaining 7.5% were noun phrase (NP) only relatives ( The girl that came with us ). The later transcripts show the same tendency, but NP-nominals became more frequent and presentationals a bit less frequent. In addition, 6% of the relative clauses modified the oblique argument ( I am going to the zoo that has snakes). Similar restrictions are found in children’s early complement clauses (Diessel & Tomasello, 2001) where the subordinate clauses functions as an argument of a predicate. For example, the complement clause can refer to the subject of superordinate clause ( That Bill wasn’t in class annoyed the teacher), or to the object of the superordinate clause ( The teacher noticed that Bill wasn’t in class ). In sum, the acquisition of complex clauses seems to be a piecemeal, stepwise and rather slow process. Initially, we find only a very restricted set of relative clauses or complement clauses. Semantically, these constructions do not encode a genuine subordination where the matrix clause makes a proposition over the subordinate clause ( Mary hypothesizes that Socrates’ contribution to philosophy is less important than commonly thought ) but, for example, highlight an object in a presentational construction ( The girl that came with us ) or encode our epistemic stance on affairs when the superordinate clause I think is used like the adverb maybe to mean “I am not sure.” Similar constraints hold in other languages as well, for example in German where the equivalent of I think is often inserted in the middle of a sentence like an adverbial ( Das ist , glaube ich, ein Huhn: “This is, I think, a chicken”; cf. Brandt, Diessel, & Tomasello, in press). Early relative clauses in

< previous page

page_52

next page >

< previous page

page_53

next page >

Page 53 German only display a limited set of the morphological paradigms (e.g. only few person–number combinations and tenses are used), and syntactic frames. Typically, some arguments are omitted, and often the children seem to use sentence frames. As in English, early relative clauses are mono-propositional in that the subordinate phrase merely provides a specification of the subject or object in the superordinate phrase. While these general learning mechanisms seem to suffice to explain children’s early language production, can they also explain very complex constructions, in particular those which may be very low frequency or even absent in the input? In other words, can learning mechanisms work where the input is poor (the poverty of the stimulus condition; see Unsworth, this volume)? 6.3 Generalization over Indirect Positive Evidence: Inducing the Missing Link Researchers from a usage-based perspective do not agree with the generativist claim that the input is underspecified. Instead, the underlying assumption is that input is rich and children seem to observe the patterns they find in the input language (e.g. Behrens, 2006. See Pullum & Scholz, 2002, for further discussion on “poverty of the stimulus” argument). Moreover, fine-grained studies on adult–child interaction show that a large number of incorrect utterances of the child receive corrective feedback (Chouinard & Clark, 2003). This suggests that children get direct and indirect negative evidence. Finally, even if the child never hears a particular structure, indirect positive evidence could suffice to learn the structure (Elman, 2003). Indirect positive evidence means that the child does not need to hear the complex structure itself, but could induce it from hearing related structures that provide evidence for all the component parts. Consider a sentence like (4a). In order to form a yes/no -question, the child has to extract the second is (4b), not the first (4c), as the child might suspect from inversion in simple clauses ( The boy is crazy → Is the boy crazy? or The girl can smoke → Can the girl smoke?). In order to avoid such errors in sentences with embedded relative clauses, the child has to know that the extraction of the auxiliary is structure-dependent because the extraction has to be from the main clause as in (4c), not the embedded clause in (4b). (4) a. The boy who is smoking is crazy b. *Is the boy who_ smoking is crazy? c. Is the boy who is smoking _ crazy? Elman (1993) conducted a computer simulation and trained a so-called Simple Recurrent Network with simple sentences of the following types (5a–h):

< previous page

page_53

next page >

< previous page

page_54

next page >

Page 54 (5) a. The women smoke cigarettes b. The boy is funny c. Can you read this? d. Is she friendly? e. The girl with the cat is nice f. The toy with the wheels is mine g. The man who came late is your uncle h. I like the teacher who speaks Spanish What can we learn from such simple sentences? Sentences (5a) and (5b) show person and number agreement as well as verb-specific subcategorization frames and selectional restrictions (i.e. the arguments a given verb may occur with). Sentence (5c) and (5d) show question formation in simple sentences. Sentences (5e–h), finally, demonstrate basic noun phrase constituency and simple complex sentences. Thus, these simple sentences provide evidence for all the ingredients of the extraction from embedded clauses, as in (4c) above. It turned out that after being trained with simple structures, the network was able to generate complex structures it had never seen before in this combination. In terms of learning, the network acquired abstract linguistic knowledge based on direct positive evidence. The application of knowledge derived from direct evidence (simple structures) then served as indirect positive evidence for the production of a previously unattested complex structure. Of course, the open issue is whether we can transfer the mechanics of machine learning to children’s learning. Controlled input experiments of that size are impossible with humans. But as our corpora of input and child language become better and bigger, we can test whether children make use of such cues. 7 How Constraints are Learned So far, it seems that the powerful learning mechanisms children have at their disposal are a big advantage. However, they also could be a disadvantage, namely if a child ends up with a grammar that is overly general (Bowerman, 1988). For example, a child learns that the past tense of English verbs is formed by adding the suffix – ed to the stem, and consequently produces goed instead of went . How could a child ever “unlearn” such an overgeneralization? In this case, there is a correct form ( went ) that fills the position in the paradigm that goed could take. Therefore, goed is blocked because its position in the system is occupied by went . But “unlearning” is less straightforward with more complex and abstract rules. For example, in English many transitive verbs undergo the socalled causative alternation. Here, a transitive verb gets a causative reading in the intransitive (cf. I broke the stick and The stick broke ). In the intransitive case the reading is that something or somebody made the stick break. A child who has heard a couple of such causative alter-

< previous page

page_54

next page >

< previous page

page_55

next page >

Page 55 nations might reasonably conclude that the causative intransitive construction The dog disappeared has a transitive counterpart ( I disappeared the dog ), and indeed, children make such errors (Bowerman, 1988). This error cannot be unlearned by positive evidence; because there is no transitive construction with “disappear,” the child has filled a gap in the system (unlike in the goed case where the child has filled the paradigm with the wrong form, and this form gets kicked out by hearing the correct form). In such cases, children could make use of statistical pre-emption. When they never hear other people use I disappeared the dog they will conclude that the common way to encode this notion is I made the dog disappear (cf. Goldberg, 2006, pp. 93–102). 8 Summary: Differences between Nativist and Usage-based Accounts of Acquisition As has been demonstrated above, the usage-based approach differs in a number of respects from nativist and generativist accounts of language learning. In the summary I spell out the different assumptions regarding the nature of language and how they lead to different hypotheses of how we learn language. 8.1 Representation at the Level of the Brain The two approaches differ in their assumptions regarding the foundations of language acquisition. UG is a mentalistic approach that explains our linguistic competence by an innate core grammar. The assumption of an innate Universal Grammar presupposes that the representation in the brain is localistic and domain-specific or modular. Modularity means that different regions in our brain are responsible for the various tasks involved in language processing (Fodor, 1983). One of these modules would be core grammar, that is, those aspects of language that are controlled or governed by the supposed universal features of our language capacity. Usage-based and emergentist approaches assume that the brain shows great plasticity. There are preferences for the localization of particular cognitive functions under normal developmental conditions, but in cases of prenatal or very early lesions, those functions can often be localized elsewhere (Elman et al., 1996). Another important concept is that of developmental modularity (Karmiloff-Smith, 1992), which explains how modularity emerges as a product of development. That is, the fact that we find modularity in adult language processing does not provide evidence that this modularity is an innate property of the human brain. The notions of plasticity and developmental modularity are also associated with representational redescription, that is, the notion that representations may change in the course of development, depending on our experiences (Karmiloff-Smith, 1992).

< previous page

page_55

next page >

< previous page

page_56

next page >

Page 56 8.2 Nature of Linguistic Representations Most researchers taking a UG stance assume that children’s underlying representations are adult-like since core syntax is innate (see Unsworth, this volume). In usage-based learning theory it is assumed that child language is very conservative in the early stages: children rely on what they have heard in the input. Not all phenomena are productive; much linguistic knowledge is initially lexically specific (tied to certain words, not general). Therefore, children’s language can look adult-like and errorfree, but may in fact be still unproductive because the child has not fully learned the underlying rules. It is as yet an open issue how to conceive of the intermediate stages of children’s grammars or graded representations (AbbotSmith & Tomasello, 2006). 8.3 Role of the Input In UG approaches the input is deficient and underdetermined (“poverty of the stimulus,” “no negative evidence problem”). UG provides the child with linguistic representations to overcome the induction problem. The role of the input is needed to “trigger” or activate the language-specific settings of UG. In an emergentist perspective all linguistic structures, even very rare and complex ones, can be generalized from the input. Induction mechanisms are pattern recognition, generalization based on analogy, intervening (generalizations from indirect evidence). 8.4 Mechanisms of Acquisition In UG, core grammar is supposed to be innate but one needs mechanisms to activate it (e.g. maturation, triggering, lexical learning or linking rules). Under an emergentist approach, children will have to work out the linguistic system as a whole by making use of (a) adaptive social-communicative behavior like goal-oriented imitation that allows for cultural transmission of knowledge, (b) pattern recognition skills, and (c) the ability to generalize based on concrete or abstract analogies. 9 Outlook Research so far has focused on these processes in the early stages of multiword speech and, to a lesser extent, on the acquisition of complex syntactic constructions. Since the basic assumptions of usage-based linguistics is that linguistic structure is derived from language use, more detailed research on the language input is needed in order to shed light on exactly how children

< previous page

page_56

next page >

< previous page

page_57

next page >

Page 57 generalize over what they hear. Also, fine-grained studies on the concrete interactional processes can show how the child takes up cues from the input, especially regarding feedback on incorrect utterances (see Chouinard & Clark, 2003). Finally, insights from the learning processes in early childhood should be applied to language-learning strategies in later stages of development and to bilingualism or second language learning. Since it is assumed that learners draw relations between new information and the information they have already acquired and categorized, the exact nature of the generalization processes should vary in dependence of the individual learner’s state of knowledge (see van Geert, this volume, for related assumptions in dynamic systems theory). References Abbot-Smith, K., & Tomasello, M. (2006). Exemplar versus prototype models in syntactic acquisition. The Linguistic Review, 23, 275–290. Behrens, H. (2003). Verbal prefixation in German child and adult language. Acta Linguistica Hungarica , 50, 37–55. Behrens, H. (2006). The input–output relationship in first language acquisition. Language and Cognitive Processes, 21, 2–24. Bowerman, M. (1988). The “no negative evidence” problem: How do children avoid constructing an overly general grammar? In J. A. Hawkins (Ed.), Explaining language universals (pp. 73–101). Oxford: Blackwell. Bowerman, M., & Choi, S. (2003). Space under construction: Language specific spatial categorization in first language acquisition. In D. Gentner & S. Goldin-Meadow (Eds.), Language in mind: Advances in the study of language and cognition (pp. 387–427). Cambridge, MA: MIT Press. Braine, M. D. S. (1963). The ontology of English phrase structure: The first phase. Language, 39, 1–13. Brandt, S., Diessel, H., & Tomasello M. (in press). The acquisition of German relative clauses: A case study. Journal of Child Language. Brown, R. (1973). A first language: The early stages. Cambridge, MA: Harvard University Press. Bybee, J. L. (2006). From usage to grammar: The mind’s response to repetition. Language, 82, 711–733. Cameron-Faulkner, T., Lieven, E., & Tomasello, M. (2003). A construction based analysis of child directed speech. Cognitive Science , 27, 843–873. Chouinard, M. M., & Clark, E. V. (2003). Adult reformulations of child errors as negative evidence. Journal of Child Language, 30, 617–669. Clark, E. V. (2003). First language acquisition. Cambridge: Cambridge University Press. Diessel, H., & Tomasello, M. (2000). The development of relative clauses in spontaneous child speech. Cognitive Linguistics , 11, 131–151. Diessel, H., & Tomasello, M. (2001). The development of finite complement clauses in English: A corpus-based analysis. Cognitive Linguistics , 12, 1–45. Ellis, N. C. (2008). The dynamics of second language emergence: Cycles of language

< previous page

page_57

next page >

< previous page

page_58

next page >

Page 58 use, language change, and language acquisition. Modern Language Journal, 92(2), 232–249. Elman, J. L. (1993). Learning and development in neural networks: The importance of starting small. Cognition, 48, 71–99. Elman, J. L. (2003). Generalization from sparse input. In Mary Andronis, Erin Debenport, Anne Pycha, & Keiko Yoshimura (Eds.), Proceedings of the 38th Meeting of the Chicago Linguistics Society (pp. 175–200). Chicago: University of Chicago Press. Elman, J. L., Bates, E. A., Johnson, M. H., Karmiloff-Smith, A., Parisi, D., & Plunkett, K. (1996). Rethinking innateness: A connectionist perspective on development. Cambridge, MA.: MIT Press. Fodor, J. (1983). The modularity of mind. Cambridge, MA: MIT Press. Freudenthal, D., Pine, J., & Gobet, F. (2002). Modelling the development of Dutch optional infinitives in MOSAIC. In W. D. Gray & C. D. Schunn (Eds.), Proceedings of the 24th Meeting of the Cognitive Science Society (pp. 328–333). Mahwah, NJ: Erlbaum. Freudenthal, D., Pine, J., & Gobet, F. (2005). Simulating optional infinitive errors in child speech through the omission of sentence-internal elements. In B. G. Bara, L. Barsalou, & M. Buchiarelli (Eds.), Proceedings of the 27th Annual Meeting of the Cognitive Science Society (pp. 708–713). Mahwah, NJ: Erlbaum. Gentner, D. (2003). Why we are so smart. In D. Gentner & S. Goldin-Meadow (Eds.), Language in mind: Advances in the study of language and cognition (pp. 195–235). Cambridge, MA: MIT Press. Gergely, G., Nadasdy, Z., Csibra, G., & Biro, S. (1995). Taking the intentional stance at 12 months of age. Cognition, 56, 165–193. Gobet, F., Freudenthal, D., & Pine, J. (2007). Towards a unified model of language acquisition. In S. Vosniadu, D. Kayser, & A. Protopapas (Eds.), Proceedings of the European Cognitive Science Conference 2007 (pp. 602–607). New York: Taylor & Francis. Goldberg, A. E. (2006). Constructions at work: The nature of generalization in language. Oxford: Oxford University Press. Gomez, R. L., & Gerken, L. (2000). Infant artificial language learning and language acquisition. Trends in Cognitive Science , 4 , 178–186. Höhle, B., & Weissenborn, J. (2000). Discovering grammar: Prosodic and morphosyntactic aspects of rule formation in first language acquisition. In A. D. Friederici & R. Menzel (Eds.), Learning: Rule extraction and representation (pp. 37–69). Berlin/New York: de Gruyter. Hudson Kam, C. J., & Newport, E. L. (2005). Regularizing unpredictable variation: The roles of adult and child learners in language formation and change. Language Learning and Development, 1 , 151–195. Hurford, J. R. (2002). Expression/induction models of language evolution: Dimensions and issues. In T. Briscoe (Ed.), Linguistic evolution through language acquisition (pp. 301–344). Cambridge: Cambridge University Press. Jusczyk, P. (1997). The discovery of spoken language. Cambridge, MA: MIT Press. Karmiloff-Smith, A. (1992). Beyond modularity: A developmental perspective on cognitive science. Cambridge, MA: MIT Press. Langacker, R. W. (2000). A dynamic usage-based model. In M. Barlow & S. Kemmer (Eds.), Usage-based models of language (pp. 1–63). Stanford: CSLI Publications.

< previous page

page_58

next page >

< previous page

page_59

next page >

Page 59 Lieven, E., Behrens, H., Speares, J., & Tomasello, M. (2003). Early syntactic creativity: A usage-based approach. Journal of Child Language, 30, 333–370. Lieven, E., Pine, J. M., & Baldwin, G. (1997). Lexically-based learning and early grammatical development. Journal of Child Language, 24, 187–219. MacWhinney, B. (Ed.). (1999). The emergence of language. Mahwah, NJ: Erlbaum. Meltzoff, A. N. (1995). Understanding the intentions of others: Re-enactment of intended acts by 18-month-old children. Developmental Psychology, 31, 838–850. Moll, H., & Tomasello, M. (2007). Cooperation and human cognition: The Vygotskian intelligence hypothesis. Philosophical Transactions of the Royal Society B , 1–10. Newport, E. L. (1990). Maturational constraints on language learning. Cognitive Science , 14, 11–28. Newport, E. L., & Aslin, R. N. (2004). Learning at a distance: I. Statistical learning of non-adjacent dependencies. Cognitive Psychology, 48, 127–162. Newport, E. N., Hauser, M. D., Spaepen, G., & Aslin, R. N. (2004). Learning at a distance II. Statistical learning of non-adjacent dependencies in a non-human primate. Cognitive Psychology, 49, 85–117. O’Grady, W. (2005). The syntactic carpentry: An emergentist approach to syntax . Mahwah, NJ: Erlbaum. Pine, J., Lieven, E., & Rowland, C. (1998). Comparing different models of the English verb category. Linguistics , 36, 807–830. Pinker, S. (1984). Language learnability and language development . Cambridge, MA: Harvard University Press. Premack, D. G., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 1 , 515–526. Pullum, G. K., & Scholz, B. C. (2002). Empirical assessment of stimulus poverty arguments. The Linguistic Review, 19, 9–50. Saffran, J. R. (2003). Statistical language learning: Mechanisms and constraints. Current Directions in Psychological Science , 12, 110–114. Saffran, J. R., Newport, E. L., & Aslin, R. N. (1996). Word segmentation: The role of distributional cues. Journal of Memory and Language, 35, 606–621. Slobin, D. I. (Ed.). (1985–1995). The crosslinguistic study of language acquisition. Vols. 1–5 . Hillsdale, NJ and Mahwah, NJ: Erlbaum. Tomasello, M. (1992). First verbs: A case study of early grammatical development . Cambridge: Cambridge University Press. Tomasello, M. (1995). Language is not an instinct. Book review of Steven Pinker: The language instinct: How the mind creates language. Cognitive Development , 10, 131–156. Tomasello, M. (2000). Do you children have adult syntactic competence? Cognition, 74, 209–253. Tomasello, M. (2003). Constructing a language: A usage-based account of language acquisition. Cambridge, MA: Harvard University Press. Tomasello, M., & Rakoczy, H. (2003). What makes human cognition unique? From individual to shared to collective intentionality. Mind & Language, 18, 121–147.

< previous page

page_59

next page >

< previous page

page_60

next page >

Page 60 3 A COMPREHENSIVE DYNAMIC SYSTEMS THEORY OF LANGUAGE DEVELOPMENT Paul van Geert 1 A Comprehensive Dynamic Systems Theory of Language Development The current literature on dynamic systems theory of development comprises a number of different perspectives or approaches. Before making an attempt at formulating an integrated dynamic systems theory of language acquisition, I shall discuss various approaches that can be used as starting points for a comprehensive framework of language development along the lines of complex dynamic systems theory. 1.1 An Overview of Dynamic Systems Theories 1.1.1 Development and the Theory of Embodied Action According to Thelen and Smith (1994), current psychological theories tend to invoke “ghostly” things to explain behavior, namely internal representations and concepts. The representationalist stance, also known as the computational–representational understanding of mind, or information-processing theory, states that thinking can best be understood in terms of representational structures in the mind and of computational procedures that operate on these structures (Thagard, 1996). According to this view, in order to explain thinking you need internal entities. For instance, if a toddler says he wants a cookie, the meaning of what he says is internally represented as a string of internal, mental concepts, such as “want” and “cookie.” According to Smith, Thelen, Titzer, & McLin (1999; see also Smith, 2005), invoking the notion of “concept” to explain a particular behavior or action related to that concept is a categorical error, as if the meaning of cookie is explained by referring to the underlying factor the-meaning-ofcookie. It is this categorical mistake that Thelen and others

< previous page

page_60

next page >

< previous page

page_61

next page >

Page 61 seek to repair by postulating a theory of situated action, in which an embodied subject acts with the help and under the constraints of a physical world that includes the external environment and the physical properties of the body (Thelen & Smith, 1994, 1998). In essence, cognition, thinking and action are explained as dynamic patterns unfolding from the continuous, “hereand-now” interaction between the person and the immediate environment, according to what is called the embodiment hypothesis: The embodiment hypothesis is the idea that intelligence emerges in the interaction of an organism with an environment and as a result of sensory-motor activity. The continual coupling of cognition to the world through the body both adapts cognition to the idiosyncrasies of the here and now, makes it relevant, and provides the mechanism for developmental change. (Smith, 2005, p. 205) The dynamic system at issue is the continuous coupling between the organism and its environment. This system shows a short-term time-evolution that takes the form of intelligent action, which changes the body and the brain through processes of learning and adaptation, thus giving rise to a long-term evolution we call “development.” The theory of embedded and embodied action has successfully explained the development of intelligent action that has a strong sensori-motor component. Its applicability to language—especially the typically linguistic aspects involving syntax, for instance—may be less obvious, but can, nevertheless, be formulated (see, for instance, Section 3.1.1). 1.1.2 Development and Resource-dependent Competition-support Systems As any complex system, a developmental system can be viewed as a collection of elements or components. These components are related through functional relationships, implying that one component can change another, and vice versa. The system is embedded in an environment with which it is also functionally related (it can affect the environment and can be affected by it). Minimally, a developmental system can be defined as a single variable or component, for instance a particular child’s lexical knowledge. This single-variable system then defines an environment in the more abstract sense of the word, consisting of any other component that functionally affects the component at issue (lexical knowledge), for instance, the child’s syntactic knowledge and conceptual knowledge, in addition to environmental components such as the language of the child’s family members. To give a somewhat more complex example, a system consisting of two

< previous page

page_61

next page >

< previous page

page_62

next page >

Page 62 variables, for instance a child’s lexical knowledge as one component and the language addressed to the child by the mother as the second component, is embedded in an environment that consists of all sorts of internal and external phenomena that are related to the two variables at issue (such as the cognitive systems of the child and of the mother, their emotions, but also the material and cultural artifacts of their homes, other family members, etc.). If the system’s components can be described by means of numerical variables that specify the level of the component, the system dynamics amount to a limited set of relationships among the components that affect the components in terms of their numerical values. For instance, a child’s lexical knowledge can be described as a particular number of words actually known or understood by the child. The child’s syntactic knowledge can be described by a numerical “ruler” that quantifies the child’s syntactic knowledge (see Fischer’s notion of developmental rulers, van Geert & Fischer, 2008; Fischer & Rose, 1999; Rose & Fischer, 1998). This representation by means of numerical variables forms part of the abstract description of the system, and does not necessarily need to correspond with a homologous empirical measure of the variable in question (for instance, it is not intended that syntax is “really” a single numerical variable, the only thing that counts is whether or to what extent such a description is a fruitful simplification, needed to describe and understand typical properties of the developmental dynamics of language; see van Geert & Fischer, 2008, for discussion). The description of the developmental system as a dynamic ecological system makes use of the following general assumptions (for thorough discussions of these principles, see van Geert, 1991, 1994, 2003). First, development is defined as the growth or increase in level of more developmentally advanced or complex variables and the decline or decrease of less developmentally advanced variables. The growth of a variable (e.g. a child’s lexicon, a child’s syntactic level) is an auto-catalytic process, in that it depends on the level already attained. Thus, if l is the current level of some developmental variable (e.g. a child’s level of lexical knowledge), the growth or change of l over some time duration is expressed as ∆l = rl for r any rate or change parameter. In an ecological system, growth or change depends on the availability of resources, which are limited. For instance, a resource factor for lexical growth is the language spoken in the environment, but also the child’s auditory system that helps it pick up acoustic signals that form the physical basis of spoken and heard language. Given that resources are limited, the growth parameter r approaches a zero limit as l approaches the level that is sustained by the available resources (a simple example is the resource “lexicon of the environment” which consists of a very big but limited number of words, and which thus limits the number of lexical items

< previous page

page_62

next page >

< previous page

page_63

next page >

Page 63 that can be learned by a child living in that environment). The effect of the limited resources is expressed mathematically as follows ∆l = r( 1 — l/K) l for K the limit level of l under the given resource conditions. This equation corresponds to the well-known and basic logistic or Verhulst equation (see van Geert, 1991, for further explanation; see also de Bot, Lowie, & Verspoor, 2007, for an application to L2-learning). If applied to knowledge-related variables, it basically states that the growth of knowledge depends on what one already knows and on what one does not know yet, given what there is to know in the current context of the particular person in a particular environment. For a system description to be really explanatorily interesting, it should contain various coupled components. For instance, it is relatively obvious that the child’s acquiring new words is related to his acquiring new syntactic rules, and the other way round (see van Geert, 1991, 1994, for examples; Robinson & Mervis, 1998). Robinson and Mervis (1998) presented a dynamic model of the relationships between lexical growth and the growth of plural use as follows: “Lexical and plural development shared the following characteristics: Plural growth began only after a threshold was reached in vocabulary size; lexical growth slowed as plural growth increased. As plural use reached full mastery, lexical growth began again to increase” (p. 363). A mathematical model of these dynamic relationships based on the logistic growth curve model (van Geert, 1991) provided an exceptionally good fit of the data. The model is an example of the ecological relationships that can hold for any couple of “growers,” that is, relatively autonomous components of a developmental system. These relationships can be symmetrical supportive, symmetrical competitive, asymmetrical competitive and supportive (as in a predator–prey relationship in a biological model) and, finally, conditional (if a particular, minimal level of a component is a necessary precondition for another component to start growing). The relationships are represented in Figure 3.1. A developmental system is characterized by relationships between any of its components and forms a web of relationships, formally similar to the food-webs described in biological models (Ripa & Ives, 2003). Language development provides an exceptionally good field of application of these general dynamic principles. By conceiving of language as a developmental system, consisting of components such as semantics, syntax, phonetics, etc., or components on a lower level of organization, such as prepositions, adjectives, verbs, nouns, etc., ecological network models can be specified simulating the growth of linguistic variables in a single child (see, for instance, van Geert, 1991, 1994; Robinson & Mervis, 1998, on lexical

< previous page

page_63

next page >

< previous page

page_64

next page >

Page 64

Figure 3.1 Four types of dynamic relationships. development in relation to the growth of plurals; Ruhland & van Geert, 1998, on the growth of closed-class words; and Bassano & van Geert, 2007, on the pattern of growth and decline of sentences of various sentence length). A study by Bassano and van Geert (2007) illustrates the process of the emergence of three hypothesized, developmentally successive syntactic generators, the holophrastic, combinatorial and syntactic generators. The holophrastic generator is basically a “one-word grammar,” that is, the set of early grammatical principles that generate utterances with a characteristic word length of one. The combinatorial generator is the developmentally more advanced set of principles that generate combinations of words, typically two per utterance. The syntactic generator is the set of principles that use the syntactic rules of sentence formation typical of mature language. Bassano and van Geert (2007) assume a series of asymmetric relationships between a less and a more advanced developmental structure, for instance the holophrastic and the combinatorial generator. The less advanced structure has a conditional and supportive relationship with the more advanced structure. For instance, a minimum level of one-word productions is needed for the combinatorial generator to emerge, and, in addition, the level of one-word production supports the growth of the combinatorial generator, that is, the production of two-to-three-word sentences based on a simple combinatorial principle. The more advanced structure, on the other hand, has a competitive relationship with the less advanced structure. For instance, the use of two-to-three-word sentences negatively affects the use of one-word sentences (see Figure 3.2). A mathematical model of these relationships with three connected growers (the holophrastic, combinatorial and syntactic grower) provides a

< previous page

page_64

next page >

< previous page

page_65

next page >

Page 65

Figure 3.2 Hypothesized grammatical generators, with functional relationships (source: After Bassano & van Geert, 2007). good fit of the empirical data (model is shown in comparison with the smoothed data; see Figure 3.3). 1.1.3 Development and Dynamic Field Theory Thelen, Smith and co-workers explain the short-term dynamics of action, thinking and knowledge and the long-term dynamics of development, by means of dynamic fields. Dynamic fields are formal descriptive tools that the researcher uses to describe a child’s perception of an environment in terms of action potentialities and expectations with regard to objects present in that environment. A dynamic field is defined by an abstract metric dimension (or a space consisting of various such dimensions) that describes the main variable (or variables) of an action or thinking process. For instance, a major variable in an object search task in which babies try to retrieve a hidden object is the spatial position of the hiding objects and hiding places. This position also defines the major variable of the child’s action, which is the place toward which the infant will reach in order to retrieve the hidden object. The child perceives the space before him as a field of action

< previous page

page_65

next page >

< previous page

page_66

next page >

Page 66

Figure 3.3 Smoothed dataset of Pauline, with fitted dynamic growth model (source: after Bassano & van Geert, 2007). potentials, that is, some parts of the field (where the objects were hidden) have more potential to attract a reaching and grasping action than other places. All these action potentials distributed over the space that the child perceives, form a field in the mathematical sense of the word. Formally speaking, for each point of the metric variable, which in this case is the child’s perceived space before him, there exists a particular activation value, which, in the case of the current example, would mean a likelihood that the child reaches at a particular location (Schutte & Spencer, 2002; Schutte, Spencer, & Schöner, 2003; Thelen, Schöner, Scheier, & Smith, 2001). Dynamic fields are represented by neural networks in the brain and change as a result of short- and long-term processes, including perceptions, experiences, maturation and self-organizing processes (Lewis, 2005; Lewis & Todd, 2007). The inputs to a dynamic field are not just linearly superposed: they show cooperative and competitive interactions, leading to self-stabilization of activation patterns (Erlhagen & Schöner, 2002). In addition to the development of the object concept, dynamic field theory has been applied to development of habituation (Schöner & Thelen, 2006) and development of working memory (Schutte, Spencer, & Schöner, 2003). Dynamic fields have been invoked as representations of the short-term dynamics of meaning production during language use by scholars using the formalism of catastrophe theory. Catastrophe theory is a mathematical and geometrical theory of discontinuities (Thom, 1972; Zeeman, 1976). The word “catastrophe” was coined by mathematicians to refer to any kind of

< previous page

page_66

next page >

< previous page

page_67

next page >

Page 67 sudden change of a phenomenon, i.e. to discontinuities in general, and has nothing to do with the colloquial meaning of a catastrophe as something violent or tragic. Discontinuities are related to categorical perception (e.g. one tends to hear a particular word as either “bout” or “pout”), whereas the physical variables upon which the categorical distinction is based are continuous. A similar problem relates to categorical aspects of meaning, that is, conceptual distinctions in a continuous world, marked by distinct words. Various scholars have attempted to link catastrophe theory of differences and distinctions to a theory of meaning and linguistic concepts (e.g. Petitot, 1979, 1985; Visetti, 2004; Wildgen, 1981, 1982, 2002, 2006). Although catastrophe theory has been linked to issues of language evolution (Wildgen, 2006), it has not yet been used in a theory of first language acquisition, although the same kind of problems as in language evolution occur at the level of ontogeny (how to map discontinuous, i.e. categorical, distinctions onto a continuous world). Catastrophe-theoretic descriptions have been successfully applied to categorical speech perception (e.g. the aforementioned bout–pout distinction) and are thus in principle also suited to describe and conceptualize the development of phoneme perception (e.g. Case, Tuller, Mingzhou, & Kelso, 1995). Dynamic fields can also be specified for abstract properties of the developmental state space in order to model longterm changes and mechanisms of development (van Geert, 1998). A state space is nothing other than the collection of all the possible states of the system. Take, for instance, a simple dynamic system defined by the child’s lexical knowledge, that is, the number of words known by a child (we shall assume that we can decide whether a word is actually “known” or not, to simplify the exercise). The state space is formed by a single dimension, namely number of words known. A state of the system is the actual number of words known, and this state changes over time. We call it a state space because the states are ordered from zero words known to one, to two, etc., to any possible number of words known. In order to give a more complete account of a child’s linguistic knowledge, we need to add more dimensions to the state space, each dimension referring to an aspect of linguistic knowledge. For the purpose of the current explanation, we need not be concerned with how such dimensions should be defined: it suffices to say that there are probably many, that they form a multidimensional space, and that the state of a child’s linguistic knowledge is a point in that space. A point in the space is defined by spatial coordinates, which are the values on each of the distinct dimensions (e.g. the number of words in the lexicon). More formally stated, a developing system can be described as a manifold of dimensions or variables, describing all of its relevant developmental properties. Since all those dimensions can be ordered along a scale of developmental progress (a developmental “ruler”), the developmental state space is thus characterized by a main direction, which corresponds to the statistical

< previous page

page_67

next page >

< previous page

page_68

next page >

Page 68 notion of a principal component. Thus, change in a multidimensional space that is too complicated to study can be replaced by change in a one-dimensional (or in general low-dimensional) space, defining developmental progress or succession. Any point on this developmental distance dimension has a certain likelihood of being “visited” by the developing system. The actual likelihood, that is, the actual activation field, is determined by inputs from the child’s momentary experience of the context, retrieved memories, actions from other persons, and so forth. These likelihoods can be represented as a scalar field, that is, distribution of probabilities, or a distribution of likelihoods of activation. The scalar field can specify a single peak, in which case the developmental state of the individual is crisp and uni-modal (the classical ideal), or by a landscape of peaks, in which case the developmental state of the individual is multi-modal, fluctuating and fuzzy (which is more like reality; see Figure 3.4). For instance, if a child alternates between using a (correct) determiner before a particular word and an (incorrect) determinerless use, it occupies two regions in the abstract linguistic space between which it shifts randomly. Development can then be represented as the change of the vector field over time, beginning with a dominant mode in the lower and ending with a dominant mode in the upper regions. The short-term dynamics of development consist of the individual’s actions, experiences and interactions in real time. These real-time actions have a lasting effect on the structure of the vector field. The effect is moderated through two mechanisms that have already been described in the work of Piaget and Vygotsky. They see development as the result of conservative and progressive forces (see van Geert, 1998, 2000, for an explanation of the model). A conservative force is one that tends to hold the system in a particular state (e.g. the child tends to be relatively stable in its use of one-word sentences, or in the determiner-less use of nouns, and so forth). A progressive force is one that determines whether a property that is new to the child’s current level of language acquisition becomes a target of attention and learning. In Section 2.1, I shall employ the notion of dynamic fields to explain the concept of epigenetic developmental models. 1.1.4 Developmental Agent and Robotics Models Although development is a prime example of a long-term adaptive process, there is relatively little fundamental work on the dynamics of complex adaptive systems in the field of development. An adaptive system can be defined as a collection of agents and artifacts that interact with one another and through that interaction are aiming at reaching their goals, concerns or interests (Axelrod, 1997; Bankes & Lempert, 2004; Elliott & Kiel, 2004; Smith & Conrey, 2007). Agents pursue their goals by means of their action repertoires, their knowledge of the world and the information they obtain through acting.

< previous page

page_68

next page >

< previous page

page_69

next page >

Page 69

Figure 3.4 A dynamic field representation of changes in attractor states of a developing language. Note The figure shows a transition from a unimodal to a bimodal state; the bimodal state will probably further evolve towards a new unimodal state. In a complex adaptive system, agents are interdependent, yet autonomous in achieving their goals. From a developmental viewpoint, agents learn from their experiences and show long-term change in their goals, knowledge, action repertoire and skills, and environment, a long-term process we call development (Schlesinger & Parisi, 2001; Steenbeek & van Geert, 2008; van Geert & Steenbeek, 2005). In a situated and embodied agent, development not only concerns the change in the person, but also changes in the person’s niches or preferred environments (Clark, 1997, 2006).

< previous page

page_69

next page >

< previous page

page_70

next page >

Page 70 A characteristic property of agent-based models is that they conceive of agents as being equipped with relatively simple rules (Smith & Conrey, 2007). Developmental psychology, on the other hand, emphasizes the complexity and the richness of a child’s emerging knowledge and skills. However, the two approaches need not be in conflict with one another, in that very simple rules of agent interaction can in fact emerge on the basis of complex, selforganizing processes requiring the interplay of the environment, experience, knowledge, and so forth. An example of how agent-based modeling and development come together is Steenbeek and van Geert’s work on the development of social status, social power and social skills in young children (Steenbeek & van Geert, 2007a, 2007b, 2008). One possible drawback of agent models for development is that the nonlinear and complex patterns that typically result from interactions between agents do so if the number of interacting agents is high (see Ball, 2005, for many examples). The typical number of agents in developmental models is small, for example, two in dyadic interaction, or 10 to 20 in small group interactions (e.g. a group of friends, a peer group). In spite of this limitation, agent models are typically suited for simulating short-term temporal patterns of interaction among persons of various developmental levels, and it is from these patterns that the interacting persons learn, adapt and develop. Agent models have been used to study and simulate language development in connection with language evolution, by modeling generations of language learning, teaching and communicating agents. Such models explain, among others, why language tends to evolve towards a capacity that is optimally adapted to biological and social learning principles, and thus tends to become an increasingly “innate” type of capacity (examples of such studies are Christiansen & Kirby, 2003; Kello, 2004; Kirby, 2002; Kirby, Smith, & Brighton, 2004; and Smith, Kirby, & Brighton, 2003). A different type of agent models in developmental studies concerns the epigenetic robotics models. Metta and Berthouze (2006) define epigenetic robotics as the study of how a realistically embedded and embodied robot-model of a person, including a brain, sensory and effector organs, changes and develops in interaction with a real world (for introductions and reviews, see Berthouze, Metta, & Sun, 2005; Metta & Berthouze, 2006; Schlesinger, 2003). Epigenetic robotics serves two purposes. The first is of a technical nature, and is aimed at designing self-organizing and learning robots that serve practical goals, in cases where direct programming of the robot is too complicated (e.g. Yavuz, 2007). The second is to understand developmental processes in humans by studying simulated but embodied (robotic) agents in real environments. Epigenetic robotics approaches have so far dealt with the following aspects of development. The first refers to the observation that human learning and development very strongly depends on social scaffolding and socially situated processes of cognition and perception. An important issue is joint attention, which

< previous page

page_70

next page >

< previous page

page_71

next page >

Page 71 occurs at a very early age (approximately around 9 months) and involves the infant’s capacity to infer an object or event of attention to an adult by using the adult’s gaze, pointing, etc., and to share that topic of attention with the adult. Joint attention is a typical human capacity, and greatly facilitates the process of cognitive and social learning. A host of robotics studies have been carried out, showing that processes of joint attention or those closely related to it can be implemented in a robotic system and greatly enhance the processes of cognitive, linguistic and social learning (Kaplan & Hafner, 2006; Nagai, Hosoda, Morita, & Asada, 2003; Schlesinger & Casey, 2003; Striano, Henning, & Stahl, 2006; Striano, Henning, & Vaish, 2006). Joint attention is closely related to empathy, that is, the ability to intuitively understand the minds of others by picking up their intentions, an ability which is related to specific neurophysiological structures, the so-called mirror-neurons, in human and primate brains (for studies using principles of epigenetic robotics, see Borenstein & Ruppin, 2005; Metta, Sandini, Natale, Craighero, & Fadiga, 2006; and for applications to impaired attentional processes, Bjorne & Balkenius, 2005). Socially situated learning rests heavily on imitation learning, or learning through emulation, which means transforming the perceived actions into one’s own action repertoires, aiming at the inferred goal or intention of the perceived action (for studies on imitation learning in epigenetic robots, see Breazeal & Scassellati, 2002; Demiris & Johnson, 2003; Gergely, 2003). Epigenetic robot studies have tested embodied and socially situated processes of cognitive and language development (Dominey, 2007; Dominey & Boucher, 2005; Lindblom & Ziemke, 2006; Lopes & Chauhan, 2007; Yoshikawa, Asada, Hosoda, & Koga, 2003) and emotional development and communication (Breazeal, 2003; Breazeal & Brooks, 2005; Breazeal & Scassellati, 2000; Cañamero & Gaussier, 2005). An important problem that these epigenetic robotics studies are trying to solve, as far as language development is concerned, is the problem of the emergence of socially grounded symbols (Cangelosi, 2006). Language involves the production of categorizable sounds (or gestures, visual symbols…) that are shared by the members of a linguistic community in the form of a grounding of the symbols in a practical relationship between the symbols and structures of language to things in the world. This grounding and sharing of linguistic symbols may be the fundamental problem of language acquisition, from which the more classical problems of language acquisition—how is grammar acquired, how does a child learn the morphology of his mother tongue, and so forth—are but mere derivatives. Joint attention, topic sharing and symbol grounding are all closely related to the intentionality of language, that is, the “aboutness” of linguistic signs, which relates to the problem of intentionality as a general property of human action and experience. Intentionality is not a pre-existing property of the human mind, but an emergent phenomenon arising out of a person’s actions and interactions with others (Shaw, 2001; Steenbeek & van Geert, 2007a, 2007b).

< previous page

page_71

next page >

< previous page

page_72

next page >

Page 72 1.2 A Comprehensive Model: Language Development as a Complex Dynamic System The starting point of a comprehensive developmental model is that development, and by implication language as a developmental process, is an emergent phenomenon, that is, that it does not arise on the basis of representative design. An example of a representative design explanation is an explanation that sees a developmental trajectory as one based on a number of instructions represented in a design-like format, for instance genetically represented instructions for a sequence of developmental steps. Emergent phenomena are phenomena that self-organize on the basis of the interactions between a great many components or influences characterized by their own particular timing and intensity. These components and influences form a complex dynamic system. Complex systems consist of a great many components and connections among such components and undergo continuous change. Stable states are dynamic in the sense that they are self-sustaining or self-recreating, and not in the sense that change has come to a halt. Complex systems are characterized by a high level of order or structure, and these orders or structures are typical examples of self-organizing and self-sustaining features. Complex systems can easily and naturally be divided into sub-systems (for instance language and cognition, or semantics and syntax in language itself), but these divisions are intrinsically fuzzy and often ambiguous, because of the interconnectedness among constituents that is essential of such systems. The natural subdivision of a system into sub-systems has a dynamic counterpart in the natural subdivision of time into timescales. Events occurring in complex systems have a characteristic timescale. Examples are the short-term timescale of the production of language in real time and in concrete contexts, and the long-term timescale of the development of language that comprises the change in the composition and parameters of the language system in an individual, and the even longer-term timescale of the historical and evolutionary change of language. Timescales are interconnected: events at the short-term timescale (e.g. the formation of a single utterance, holding a conversation, giving an answer to a question …) are directly governed by the components and parameters of the language that the person has developed so far, and these components and parameters are directly governed by the historical or evolutionary state of the language that the person has mastered or is mastering. Relationships between timescales are symmetrical in the sense that if a relationship holds from scale a to b, there is also a relationship (but a different one) that holds from b to a. This is an important feature from a developmental point of view. It is almost trivially so that the actual language use of a child at a particular time is directly affected by the child’s level of language development at that time, and also that the development of the child’s language is governed and determined by the actual use of language in

< previous page

page_72

next page >

< previous page

page_73

next page >

Page 73 the context of social and communicative interaction, at the short-term timescale. In a design- or instruction-oriented theory of language development, this principle of symmetrically occurring relationships between timescales need not hold. In such an approach, the long-term timescale of language development can in principle be determined by its own “laws” or hidden instructions, for example a genetic program, with the short-term productions of actual language utterances as uni-directional “offspring” or product of the underlying principles. The input from short-term processes into the long-term processes of change can in principle be of any kind and need not correspond with the actual short-term expressions of the language competence in the form of utterances, conversations, verbalizations, etc. (see Figure 3.5). In a complex dynamic system, short- and long-term processes are intertwined, implying that the short-term dynamics of language production directly affects the long-term dynamics of language development, and the other way around. A major theoretical insight from dynamic systems is that the patterns of action, thinking, or development in the long term, do not result from any

Figure 3.5 A simplified representation of relationships between the short-term processes of language use (production and understanding in communicative settings) and the long-term dynamics of language development. Note The upper figure represents a more traditional input–output view, the lower figure represents an intertwined shortand long-term dynamics.

< previous page

page_73

next page >

< previous page

page_74

next page >

Page 74 single factor (plus some additional “noise”), but from the local and temporal confluence of many factors, operating on many timescales (Barsalou, Brezeal, & Smith, 2007; Schoner & Dineva, 2007). A consequence of this view is that there exists no single or main “factor” or process that causally explains the development of language or, for that matter, the actual production of a one-word sentence by a -year-old, in a particular action context. The causal explanation of language use and language development thus requires models that reconstruct the dynamics of many interacting components or factors, with no single factor being the major player. In that sense, cognitive aspects (relating to the way the person experiences his or her psychological life space here-and-now through perceptions, actions and appraisals) are dynamically intertwined with intra-linguistic aspects (pertaining to morphosyntactic properties of the language), with functional aspects (relating to limits on short-term memory and perception) and with the time-dependent aspects of the language production itself (the language utterances actually produced by the speaker and his communicative partner). The difficulty for an explanatory theory is to disentangle the dynamic relationships between these factors: which one retrieves which other one, and what does the other one then cause to happen, and so on. The starting point of any process of development is concretely embedded and embodied action. In the case of language development, this means the actual use and understanding of language in action and interaction contexts, in which language has function and meaning in terms of the child’s action goals, intentions and evaluations, in the context of the goals, intentions and evaluations of other persons with whom the child is interacting. Embedded and embodied action is a typical short-term phenomenon, the understanding of which requires a fundamental model of the agent and of its actions, in terms of concerns, interests and goals, action means, contexts, interaction cycles with the context, appraisals and evaluations, shared intentions with others, limited resources, and so forth (Barsalou et al., 2007; Smith, 2005; van Geert, 1991, 1994; van Geert & Steenbeek, 2005). In a developmental framework, understanding embedded and embodied action involving language also requires an understanding or hypotheses about the long-term process of language development, as specified above. 2 An Attempt Towards a Theory of Language Development Based Upon a Dynamics of Embedded and Embodied Action 2.1 The Dynamics of Grounded Meaning and Physical Symbols in a Developmental Context A more or less classic way of phrasing the problem of language acquisition is that the child must learn to map sounds onto meanings, and vice versa.

< previous page

page_74

next page >

< previous page

page_75

next page >

Page 75 Associated with this description (but not necessarily logically implied by it) is that there are two separable “realms” or categories, one of sound and one of meaning, which each have their proper realization (the way they exist in reality). The mapping problem then amounts to relating some structure or sequence in one domain to some structure or sequence in another domain. If the mapping occurs from sound to meaning, the psychological operation is one of understanding; if it occurs from meaning to sound, the psychological operation is one of production of language. The crucial issue, from a developmental theoretical point of view, is how sound and meaning are defined or conceptualized here. It is, of course, very well possible to define both in an entirely decontextualized meaning, for instance a particular sound (e.g. the sound that corresponds to what someone utters if he reads the word “doll”) and a particular meaning (the meaning of doll). However, it is likely that a decontextualized approach to the problem is not in any way related to the actual problem of development, because—and this is the viewpoint according to an embedded and embodied action approach—there is no such thing as a decontextualized sound and/or meaning in development (one could argue that the ability to decontextualize is a late acquisition, more like an end point than a starting point). For instance, if we carefully observe a toddler utter a particular sequence of sounds we see that the sounds are far from being independent of what happens at that particular moment, or what the child does at a particular moment. The sounds (that the child and other people hear as words) accompany or conclude certain action sequences, follow other sounds (words and sentences), are followed by still other sounds, and so forth. That is, they are well positioned in time and space, in the course of the child’s action and experiences that the child shares with other persons. Hence, we are not dealing with a loose collection of separate elements, sounds, actions, perceptions, but with a pattern, more precisely a pattern with a temporal dynamic. Although the sound can be physically and descriptively isolated from the rest of the context, its psychological function or meaning in the most general sense of the word is a matter of its being a temporal and spatial node in a dynamic pattern. This dynamic pattern is meaningful for those who share it. During an event in which a young child partly crawls and partly walks towards a toy bike, meets a sibling on his way and a father who is videotaping the little family scene, the child’s utterances—understandable words, not-understandable ones, sequences that a linguist would describe as a two-word-sentence—the pattern is meaningful to and shared by all the participants, although they have different roles in and perspectives on its creation and continuation. With regard to those participants, the term meaningful implies that they know how to participate in the event, extend it, share interpretations and evaluations, smile if something happens that is unexpected but still consistent with the event, and so on. Thus, meaning is not an internal representation, but a patterned event in which

< previous page

page_75

next page >

< previous page

page_76

next page >

Page 76 those who “understand” it can participate (in different ways, one child can, for its own reasons, try to disrupt it, grab the toy bike and disappear with it, leaving his younger brother in tears and his father angry). The type of meaningful events and action patterns that are shared among people—children, siblings, parents …—has in itself nothing particularly “linguistic.” One can have such patterns without any language at all (a comparable scene, though with different intentions and objects, could have occurred with a bonobo family …). However, in the above example, language is part of the event, the action patterns are not just motor actions like pushing the bike, or emotional expressions like smiling or getting angry. Language utterances like the child saying “bike-bike” and the parent saying “yes, go and get your bike,” are functional components of the event sequence, determining its structure, that is, how the participants perceive and evaluate it, continue it and so forth. The objects that feature in the scene are similarly involved in determining the structure of the event. The toy bike is not a mere uninterpreted physical object, it is an object with particular affordances (in the Gibsonian sense) consistent with the physical abilities of the participants and with the cultural interpretations and values that the object at issue is embedded in (imagine a similar scene with a chainsaw instead of a plastic toy bike). The cultural meaning of an object—the toy bike or the chainsaw—is something that the child has to learn through its actions with the objects (or the fact that the child is not yet allowed to act with them), the actions of others with the objects, and the emotional evaluations that actions with these objects bring about. As such, objects derive their meaning, that is, the way they are put to use, evaluated and shared, from the network of objects, actions and evaluations in which they are embedded, with “network” meaning the possibilities of temporal assemblies of objects into actions and evaluations that the child can recognize and contribute to. From this developmentally early, functional point of view, material objects, cultural artifacts and words (and sentences, and word forms, utterances, etc.) are essentially similar kinds. They occur in the form of temporal dynamic patterns, they are based on temporal assembly rules that depend on their affordances, and they can be broken down in manipulable units and subunits—implying that they can occur on various scales of aggregation, that is, that they are hierarchical—with which new temporal dynamic assemblies can be constructed. And, finally, they are shared among individuals, that is, they are structures of participation, which are deeply communal and mutual. For young children the developmental problem is not so much that they have to learn that meanings, intentions, and so on can be shared with others and that they are common and transparent to all participants in an event. On the contrary, the problem is rather that they have to learn that information, knowledge and intentions are sometimes not shared: studies on the development of theory-of-mind illustrate this observation. Hence, it is highly plausible that a sound pattern produced by a child that is perceived as equal

< previous page

page_76

next page >

< previous page

page_77

next page >

Page 77 to a sound pattern produced by an adult, with a timing that grounds the sound pattern at issue in a shared event or action sequence, is immediately perceived as a sound–meaning unit that gives concrete, observable shape to a joint intention. In this sense, meaning and signification are constructed through very elementary social and material acts, involving a joint interest and attention of various participants, including the young child who utters the sound and finds himself in an event in which that sound is repeated or in any other way reciprocated by other persons. The sharing of contexts, objects and values is further enhanced by the contagiousness of behavior and emotion, that is, the automatic assimilation of another person’s actions or evaluations of an event in the form of imitations or emulations of the other person’s actions or emotional expressions. This tendency to imitate in an intentional sense of the word (co-regulating one’s actions) is a very early acquisition and is present in babies and infants (Callaghan, Rochat, & MacGillivray, 2004; Carpenter, Call, & Tomasello, 2005; Gergely, Bekkering, & Király, 2002; Meltzoff, 1988; Meltzoff & Moore, 1983; Thompson & Russell, 2004). Imitation in infants is not trivial but extends to appropriating relatively abstract aspects of the other person’s actions (Call & Carpenter, 2002; Thompson & Russell, 2004). However, words, utterances and sounds are also different from objects such as toy bikes, chairs and spoons. The sequencing and temporal dynamics of linguistic components is different from that of the sequencing of events and actions involving physical objects, although the underlying general principles may be similar. The structure and sequencing of events involving material objects is based on the affordances of the objects and the way these affordances interact with the agent’s abilities and actions. An affordance is an action potentiality of a material object (in the broadest sense of the word), relating to specific possibilities of incorporating that object in an action sequence involving other objects. If the objects are linguistic, the general principle of “affordance” still holds, but the affordances themselves are typically linguistic, though embedded in a context of concrete action. The “affordances” of a particular word in terms of its position and function in a sentence depend on the word’s grammatical properties, the grammatical category to which it belongs, its meaning, etc. The structure of affordances of the objects or units of language constitutes what is linguistically called “grammar,” including the rules of syntax (the hierarchical sequencing of words) and semantics (the grounding of the linguistic utterances in physical action or conscious imagery). These affordance structures are the result of a historical (evolutionary) process of the emergence of language in the human species. This evolution has been a self-organizing “very long”-time process, which was governed by principles of learnability under the constraints of an individual ontogenetic process in a social context. The principle of the historical evolution of affordances and values (or valences) of cultural artifacts is not in essence different from that of physical

< previous page

page_77

next page >

< previous page

page_78

next page >

Page 78 tools and objects, and requires a developmental process of appropriation of those affordances through joint action and “imitation” of other people’s actions. Cultural artifacts obtain their culturally specific affordances for a person because the young person imitates (emulates, participates in) the actions of more competent and usually older persons and thus appropriates the affordances associated with the objects in question. If the “object” is a word, the principle of the emergence of such affordances and of their appropriation by young members of the species is not essentially different from that occurring with other cultural artifacts (with the exception that physical artifacts also carry purely material affordances, to a far greater extent than objects consisting of sound). The problem of what syntax means in terms of action in real time (i.e. the actual production of utterances in communicative or self-directed speech and understanding) is not solved by the above speculations, but whatever the answer to that question it must be framed in a way that does justice to the actual forms, properties and constraints on real development. Thus, the main problem to solve in order to arrive at an understanding of how language develops in an emergent fashion is the problem of real-time integration of shared intentional events, real-time dynamics of language production, and real-time dynamics of language comprehension on the short term and long term. 2.2 Shared, Structured and Meaningful Events as the Starting Point of Language The preceding thoughts on the emergence of language are consistent with a number of recent developments in linguistics and approaches to the nature of language. To begin with, the idea that from a developmental point of view language does not amount to a problem of mapping sound to meaning and vice versa is supported by views that language is not a code (e.g. Bickhard, 2007; Carr, 2007; Kravchenko, 2007; Love, 2007). From a linguistic as well as a developmental point of view, the starting point for language lies in the structure of meaningful, joint action and shared perceptions and evaluations of the context by infants and older or more competent speakers. For instance, infants develop an understanding of other people as intentional agents around the age of 1 year (Tomasello, 1995; Tomasello & Rakoczy, 2003; Woodward, Sommerville, & Guajardo, 2001), and understanding of other people’s goal-directed action occurs already around the age of 6–9 months of age (Gergely & Csibra, 2003; Király, Jovanovic, Prinz, Aschersleben, & Gergely, 2003; Meltzoff, 1995; Woodward, 1998, 1999). That is, intentions of other people—in the sense of other people being intentionally directed towards some particular aspect or value of the common context—are automatically incorporated in the intentional actions of the child him or herself, which is expressed in the form of joint attention and the emulation

< previous page

page_78

next page >

< previous page

page_79

next page >

Page 79 of other people’s goals by very young children. This understanding of intentionality is a property of the action itself, and does not require explicit, conscious representations of the goals to be achieved (Roberts & Lee, 2002). The almost automatic transparency of another person’s goals—in the sense of intentions, directedness-towards or referencing—lies at the basis of the infant’s joint attention (Tomasello, 1995) and the infant’s early capacity to share topics of perception and action with others in a meaningful content and the capacity to functionally understand signs of shared attention such as pointing and looking direction (Bigelow, MacLean, & Proctor, 2004; Cleveland, Schugg, & Striano, 2007; Eilan, Hoerl, McCormack, & Roessler, 2005; Mundy et al., 2007; Striano, Chen, Cleveland, and Bradshaw, 2006; Woodward, 2005). Language is of course not identical to joint attention, shared intentions, structures of actions, evaluations and events, but it seems clear that essential features of language—joint reference and structured meaning in the experienced world—can find their developmental starting points in these early abilities. In short, linguistic acts are at their very start already embedded in a structure of meaning, which is a structure of relationships among objects, values, actions, space and time that is shared among the participants in an event that brings them together. The developmental task is to learn how to increase the complexity of the linguistic components of the shared event in accordance with the rules maintained by the language community and to later decontextualize the use of language such that its binding onto concrete here-and-now shared events can be reduced. The traditional way to view the action aspects of words and phrases is via pragmatics, that is, how to do things with words, how to accomplish things in the world through the use of language. However, a major aspect of the action dynamics of language is to establish a sequential sound pattern in time, that is, to speak and to understand speech, which occurs in real time, according to the rules of the language in question. Dynamic systems theory has dealt with the issue of establishing action sequences with physical objects and spaces by analyzing the constraints and possibilities that emerge from the “affordances” that objects, space and acting bodies offer. These affordances can be defined as properties of the actor-environment as a unified system (Chemero, 2003; Chemero & Turvey, 2007; Smith, 2005). “[A]ffordances are properties of the environment that have some significance to some animal’s behavior” (Chemero, 2003: p.184). Turvey (1992) defines affordances as dispositional properties, that is, the tendency to show a particular property in particular circumstances. As such, affordances are important for prospective control of action; they help determine what comes next or what to do next. Recently, a number of authors have argued in favor of an application of the concept of affordance to language, language evolution and language development (Atkinson, Churchill, Nishino, & Okada, 2007; Cangelosi, 2007; Hodges, 2007a, 2007b). If applied to language production, or language comprehension for that matter, it is important to ask what the affordances of

< previous page

page_79

next page >

< previous page

page_80

next page >

Page 80 a word like “door” or “the” or “closed” are with regard to linguistic action, namely the production of a particular utterance (like “the door is closed” in a particular situation, which could be descriptive, or a metaphorical saying during a conversation, and so forth). The affordances of a word are the semantic and syntactic prospects it offers for completing the utterance in question during the process of real-time language production. These affordances are semantic, syntactic, pragmatic and phonetic and it is likely that during the process of language production itself, they are not separated as if they were separate modules contributing to the production (Bickhard, 2007; Butler, 2008). The problem of language acquisition for the child is to master the ability to produce utterances as his or her more competent conspecifics do, under the constraints of dynamic language production and social grounding (Yu & Ballard, 2007). The notion of dynamic language production implies that syntactic, semantic and phonetic properties of the language produced by a speaker in real time are fundamentally afforded and constrained by the temporal characteristics (sequence and speed) of language production. An increasing number of studies illustrates how the form of language is shaped by the fact that language production is an action of producing sound sequences in real time (Auer, 2007; Butler, 2008; Diessel, 2007; Hawkins, 2007; Port, 2007; Raçzaszek, Tuller, Shapiro, Case, & Kelso, 1999; Raçzaszek-Leonardi, Shapiro, Tuller, & Kelso, 2008). In this light, a good metaphor for the learning of language is the learning of juggling skills, which involves the ability to control actions on multiple timescales, the embedding of sub-systems into dynamic higher-order units and the construction and maintenance of stable but flexible patterns over time, through a process of learning that takes place in the form of practice (Dessing, Daffertshofer, Peper, and Beek, 2007; Huys, Daffertshofer, & Beek, 2004). However, in addition to acquiring the juggling skills of language production in real time, the child must also learn the social grounding of the language, that is, the sharing of meaning with others. We have seen that this social grounding is in fact not an end point of language learning, but the starting point, in the form of shared intentional action, perception and evaluation of events. Development requires a process of (relative) dissociation of the language aspect from the events in which it is embedded, such that language becomes a flexible and interchangeable form of shared and socially distributed action through symbolic means (Bickhard, 2007; Cangelosi, 2007; Clark, 2006; Cowley, 2006, 2007; Cowley, Moodley, & FioriCowley, 2004; Linell, 2007; Vogt & Divina, 2007). Various studies of language development and language learning are now conducted that illustrate the developmental relationship between the short-term temporal dynamics of language production and comprehension with the longterm dynamics of language development (see, for instance, Lowie, Verspoor, & de Bot, this volume and Verspoor, Lowie, & van Dijk, 2008; Behrens, this volume; Atkinson, Churchill, Nishino, & Okada, 2007; Estes, Evans,

< previous page

page_80

next page >

< previous page

page_81

next page >

Page 81 Alibali, & Saffran, 2007; Goldberg, Casenhiser, & White, 2007; Greenspan & Shanker, 2007; Jones & Smith, 2002; Kersten & Smith, 2002; Tuller, Jatzen, & Jirsa, 2008). The fact that actions (including language actions) can be taken out of their functional context and transformed into flexible action formats that can be assembled into new action sequences and that become relatively independent compartments of the person’s psychological life space is a typical human property, and is probably the major species-specific property that is characteristic of but not limited to language (Carr, 2007; Clark, 2006; Hauser, 2008; Hauser, Chomsky, & Fitch, 2002). This is probably the kind of property that in Chomsky’s terms is called “competence.” Is there any species-specific aspect to acquiring language? The starting point in the dynamics of joint action is not necessarily solely human, except maybe for the intensity with which it occurs. Is the onset of true language development uniquely explained by the fact that the child is embedded in a linguistic environment from the start, an environment that is by definition adapted to its learning abilities (otherwise, it could not have evolved; see Deacon, 1997; Kirby & Hurford, 2002, and other authors on language evolution). Is language development dependent on the fact that the language learner shares a social world with his conspecifics that use the language (chimp experiments on language learning miss this latter aspect; it is not transmitted or used by conspecifics of the chimp baby). Or is it so that human babies have something special, that enables them to get language acquisition off the ground? The ingredients so far are thus: (1) joint meaningful (intentional, goal-oriented, emotionally evaluated) action which is perceived as a structure of nested components and relationships; (2) the fact that language is part of the child’s natural niche (the language as spoken by the child’s caretakers and conspecifics); and (3) the fact that human infants are capable of decomposing structures into constituents that can be used in other contexts, that is, that can be reassembled to new structures; this applies first to the ability to decompose action streams and easily experiment with means (see Piaget’s observations on circular reactions), to the ease with which “symbolic” objects in the actions (words, signs) are decomposed and put in a sequence (see Senghas, Kita, & Özyürek, 2004), and to the ease with which infants apply what Morton once called “predicate raising” (Morton, 1986), and finally to what has recently been called the “floodlight” instead of “laserbeam” nature of human intelligence (Hauser, 2008); numerosity experiments showing the ability of seeing groups of individual objects as units (see also Leslie & Chen, 2007).

< previous page

page_81

next page >

< previous page

page_82

next page >

Page 82 3 Representing the Long-term Dynamics of Language Development 3.1 An Epigenetic Model of the Long-term Dynamics A convenient way for representing the dynamics of language development over the developmental timescale is by means of a one-dimensional state space representation, which leads to an epigenetic landscape representation. The assumptions are as follows. The first assumption has already been treated under the description of the dynamic field approach (see Section 1.1.3). It holds that any dynamic system can be represented in the form of a state space, that is, a collection of all the dimensions, factors or constituents one needs to describe the dynamic system at issue. Since each constituent can be treated as a dimension (a familiar numerical dimension, or a present–absent dimension, or anything in between), the state of a system, for instance the current state of the system of language development in a particular child, can be represented as a single point in the space of all relevant dimensions. Since the state of the system changes continuously, the position of the point in the space changes over time, thus tracing a trajectory across the space. It is not necessary—and in fact impossible—to specify all relevant dimensions, but essential features of the dynamics of the system are conserved in any well-chosen simplification of the state space in the form of one or a few “indicator variables” (with “well-chosen” dependent on the phenomenon under study). These indicator variables can be anything, for example the lexicon, the frequency of use of a particular grammatical form, etc. The choice of such indicator variables must be inspired by what one considers the major or order variables of the system (Haken, 2006; Tschacher & Haken, 2007), that is, those properties that specify the characteristic features of the system at issue, in this particular case language. In this sense, the notion of order variables comes close to the criterion of descriptive adequacy known from generative linguistics. The simplification of the state space into eventually a single dimension can also be done by collapsing or compressing the entire state space onto a single, abstract dimension that is used to represent the major position differences in the state space that are relevant for description. For any possible point in the space, that is, for any possible state of the system, there is a vector of motion, that is, a rate and direction of change. What these vectors basically represent is this: if you are “here now” you will be “there later.” These vectors of motion are defined by the system’s “evolution term,” that is, the rule or set of rules that specify the basic form of change, or the basic “law” of change in the system. This law or evolution term can be arbitrarily complex—and for complex systems it is almost by definition very complex—but it is often so that its central features can be captured by a highly simplified representation. For instance, the evolution law or

< previous page

page_82

next page >

< previous page

page_83

next page >

Page 83 term governing the change in language development is no doubt of incredible complexity, but what it basically does can often be captured in the form of a simple rule of increase or decrease over one or a few indicator variables (van Geert, 1991). This is basically what happens in dynamic growth models of development, based on simple rules of increase and decrease coupled to the quantitative level of some chosen indicator variables (e.g. the size of the lexicon, the syntactically correct use of verbs, etc.). The laws or rules governing the evolution of the system are not necessarily fixed, but change as a result of the states and dynamics they create in the system. For instance, once the language use of a child changes into the regular production of two-word sentences (this is clearly a simplification, but it will do for the current purposes) as a consequence of a great variety of interdependent and interacting factors, the existence and sequences of two-word sentences itself becomes a dynamic factor of the system, thus changing the system’s rules or laws of change. The dynamic nature of the rules of change itself is a common and important feature of a complex developing system, but it makes it virtually impossible to describe the actual rules in any detail (except for the highly simplified and general rules hinted at above). A convenient possibility for representing the dynamics of the evolution laws of the system is to describe them in the form of a so-called potential landscape. For each point in the space there is a certain quantity or scalar, also called a potential. These potentials form a landscape of hills and valleys, as is shown in Figure 3.6. The state of the system is represented metaphorically by a ball in the landscape that rolls towards the nearest lowest point. The potential difference is a quantity related to the amount of energy required to move the ball from one place to another. In the case of language acquisition, this energy is equivalent to information, examples, corrections, practice, and so forth. Figure 3.6 metaphorically shows how the rules governing the dynamics, that is, the rules governing the potential changes of the system and thus of the potential places the ball can occupy in the landscape, can change over time, thus leading to changes in the potential landscape. The most important points in the landscape are the possible points of stability (the wells), and since these points “attract” the ball (the ball rolls towards the locally lowest point), this type of landscape is also called an attractor landscape . Since the changes in the attracting valleys or wells and the repelling hill sides and tops are caused not only by external forces, but primarily by the state of the system itself, the resulting landscape is often called an epigenetic landscape. The observation that the landscape is shaped by the state of the system itself implies, metaphorically speaking, that the form of the landscape is co-determined by the position, weight and speed of the ball. These metaphorical properties can stand for various real properties of the developing system. Another consequence of the observation that the landscape is shaped by the properties of the system itself (its rate of change, its position in

< previous page

page_83

next page >

< previous page

page_84

next page >

Page 84

Figure 3.6 An epigenetic landscape representation of attractor states in language development. Note An imaginary situation of attractors is shown; epigenetic landscapes are comparable with dynamic field representations, but the attractor is reversed (the valley instead of the summit). the landscape …) is that the landscapes corresponding with the developmental trajectories of distinct individuals tend to diversify, in that small differences between persons can be amplified and can lead to increasing between-person variability. In line with the non-linearity of the process, such amplifications occur only at particular places in the landscape, if the valleys or wells (the local minima) are shallow, allowing for more fluctuation and variability than in places where the valleys are deep. The scientific importance of the epigenetic landscape metaphor lies in the fact that it allows for a comprehensive and coherent description of the main features of developmental processes as they occur in complex dynamic systems. Although the landscape itself is a metaphor, it has a specific mathematical and formal interpretation that links it to the theory of dynamic systems (Newell, Liu, & Mayer-Kress, 2003; Striedter, 1998). Case et al. (1995) use the notion of potential-landscape to represent attractors in the space of speech perception (see also Porter & Hogue, 1998). Results like those in Figure 3.4 have been found for many of the acoustic

< previous page

page_84

next page >

< previous page

page_85

next page >

Page 85 features of speech. These include temporal characteristics like voice onset time, formant transition durations, and turbulent noise durations, as well (Porter & Hogue, 1998). 3.2 Grammar as a Complex Ecological System In preceding sections, I described language as an embodied and embedded human and social activity and I have tried to explain some of the dynamics of language change and development resulting from this framework. However, in linguistics it is customary to view language as a more or less closed system, that is, a system that can be fruitfully studied without taking too much of the non-linguistic world into account. That is, language is conceived of as a system governed by its internal principles and rules. The fact that it can be fruitfully studied in (conceptual) isolation from the rest of the world is a property that it owes from its character as a complex system itself, that even if conceived in relative isolation conserves its relationships with the physical and human world in which it is embedded. From this point of view, the short-term dynamics of language concerns the question of how the actual unfolding of linguistic utterances—informal or formal—should be understood over the timescale of natural speech production or understanding over various media (spoken or written). The long-term question regards the change in the rules or knowledge governing this short-term unfolding, that is, governing the actual production and understanding of language, in light of its formal principles. The viewpoint that I shall defend is that observable properties of developing language—for instance, the mean length of utterance, the number of utterances containing a verb, the increase of the lexicon, the number of so-called function words (e.g. prepositions, modals, articles …) and so forth are all indicators of the underlying long-term dynamics of language. They are like properties referring to internal aspects of the complex dynamics of language, but their reference to these internal properties is not trivial. For instance, if we take the number of two-word sentences as an indicator of what goes on in the early linguistic system of a child, we can say that the two-word sentences are produced by a two-word generator, a type of combinatorial principle. However, in line with the view on dynamic systems as systems determined by the interplay of all operational factors, we cannot conclude that somewhere inside the linguistic system there is a dedicated two-word generator, which contains a number of combinatorial rules. What we are confronted with is a complex dynamic system that acts as a producer of combinations of two words, and this is a characterization of that system that may be relatively correct for some limited developmental time. The system as a whole is, in some fuzzy sense of the word, the two-word generator. The same sort of observation holds for the production of articles (or anything comparable, for that matter). A child’s language may be characterized

< previous page

page_85

next page >

< previous page

page_86

next page >

Page 86 by a certain level of correct use of articles, but this is not to say that in some literal sense of the word the child’s linguistic system is acquiring an article generator (or a rule that specifies how and where articles should be used, which is basically the same thing). From a dynamic systems point of view, we can take a number of observable indicators of the underlying dynamics of the language, associate them with some particular property of the system under study (and this property basically entails the fact that it is a system capable of generating the observed phenomena) and try to understand how these observable indicators are dynamically related over developmental time. The art of theory formation entails that we choose indicators (observable variables) that make theoretical sense, that is, that are related to some of the fundamental properties of the phenomenon under study, which is developing language in this particular case. In the study with Bassano (Bassano & van Geert, 2007) that I mentioned earlier, we took the number of one-word, two- to three-word and 4+-word sentences as three such indicators. We spoke about the developing linguistic system as being at first a holophrastic generator, then transforming into a combinatorial generator, then finally into a syntactic generator. This notion of generator served as a shorthand reference to the underlying complex processes of language production and learning and to the processes of transformation and transition that characterize the observed developmental sequence. In this sense, one should not think of the young child’s holophrastic linguistic system as a little engine that picks single words from a sample of known words, and the child’s combinatorial system as one that picks two words at a time from the sample. Such descriptions, even if they are phrased in considerably more complicated terms involving rule manipulation and so on, are mainly metaphors referring to the complex system of an embodied and socially embedded human brain, the rules and principles of which are only barely known. However, the choice of the metaphor or of the observable indicators is warranted if we can understand or reconstruct its dynamics in ways that correspond with defendable intuitions or justifiable knowledge about the field of inquiry. In the case of the division into hypothesized holophrastic, combinatorial and syntactic generator structures, for instance, we found that the data—the change in proportion of one-word, two- to three-word and 4+word sentences in a child’s spontaneous utterances—could be reconstructed by a dynamic growth model based on some simple principles described earlier in this chapter. Additional support for the relevance of the three-generator model was found by inspecting the intraindividual variability in the production of the three types of utterances. Complex systems can have various forms of organization (i.e. characteristic types of how it operates, what it does …) and these forms of organization can exist simultaneously (in which case they can be selected or “soft-assembled” as required by the current circumstances), but

< previous page

page_86

next page >

< previous page

page_87

next page >

Page 87 they can also exist sequentially (in which case one will follow the other). Complex developing systems often show simultaneity as well as sequentiality. In this particular case, different generators (e.g. the holophrastic and the combinatorial) can exist simultaneously in the form of a temporary period of overlap. This is exactly what is represented through the notion of epigenetic landscapes, which can have several attractors at the same time. At this time, fluctuations between using one or the other generator are maximal, which is observable in the form of a temporary peak in variability (differences between the proportion of one-word and two- to three-word sentences in sequential language samples). This peak in variability was also found in the data, and was found to occur twice, namely once at the transition between holophrastic and combinatorial language use, and once during the transition between combinatorial and syntactic language use. A second example of a dynamic model of language change in the form of a web of dynamic relationships is based on a dataset collected by Ruhland, Wijnen and van Geert (1995) from a Dutch boy, Peter. Ruhland and van Geert (1998) focused in particular on so-called function words, a closed class of words that contain modal verbs, articles and pronomina. The developmental curves of these three classes were almost identical and they were taken together to form the variable function words . In order to compare the emergence of function words with other aspects of syntactic development, I shall also present the data on the changes in proportion of utterances containing verbs over the total number of utterances, and the growth of the lexicon. It is important to note that the growth of the lexicon does not refer to the size of the lexicon (the number of words known), but to the rate of increase of the number of words (number of new words learned, or expressed, during a particular period of observation). Whereas the proportion of utterances containing verbs is an indicator of early grammaticalization of the child’s language, the growth of function words appears to represent a later linguistic development. The growth of the lexicon concerns a different aspect of language, which is related to but clearly different from the development of syntax. In order to prepare the data for validation of the dynamic model, the raw data were first smoothed and then normalized. Several types of growth models can be fitted to the data. The default type of model is a completely connected growth model. That is, each variable is assumed to change because of an inherent growth process (based on learning, practice, or any other conceivable mechanism of increase of the linguistic variable in question). In addition, all variables affected each other. That is, each variable has a supportive or competitive relationship to each other variable, which depends on the level of the variable and on the change in the variable (for details, see van Geert, 1994, 2003; Fischer & Bidell, 2006). Dependence on the level of a variable means that the growth in the dependent variable is proportional to the level of the variable on which it depends.

< previous page

page_87

next page >

< previous page

page_88

next page >

Page 88 Relationships can be positive (supportive) or negative (competitive), in which case the relationship is inverse (e.g. the higher the level of an independent variable, the smaller the growth in the dependent variable). Competition through change often occurs if the growth in one variable (e.g. syntactic growth) depends on limited resources (e.g. time, attention, effort …) that must be shared among variables (e.g. syntactic and lexical growth), implying that the growth of one variable (e.g. syntax) goes at the cost of growth in the other (e.g. lexicon). Competition based on change in a related variable often leads to temporary regressions in the growth trajectories of the variables that suffer from the competition. Adding parameters relating to the effect of rate of change in the variables often results in nearly perfect fits of the data (simply because the number of free parameters in the model is almost doubled by adding them to the parameters relating to the effect of level of a variable on the growth of others). The parameters should only be added if there is a theoretical justification for them, that is, if we can think of some plausible reason why the rate of change in a variable would affect the growth of another one. There don’t seem to exist any such plausible reasons as far as the three current variables are concerned (an example from an entirely different field is from a study by Lichtwarck-Aschoff, van Geert, Kunnen, and Bosma (2008), who modeled the growth of autonomy in adolescents relative to their level of connectedness with their parents; the growth in autonomy corresponded with the number of fights and conflicts the adolescents had with their parents, and these conflicts negatively affected the connectedness they felt with the parents; as the fights were over, connectedness rapidly restored itself to its equilibrium value, which was based on the family’s carrying capacity with regard to mutual; connectedness among its members). In addition to being affected via support or competition by other linguistic variables, every variable has its own equilibrium level, which it would finally reach if it were not affected by any other variable. This unaffected equilibrium level, which is also known as the carrying capacity of the system, is in a sense a latent property of the system, since the observed equilibrium level is a function of the unaffected carrying capacity on the one hand and all the additional effects on the variable on the other hand. If the growth pattern shows a stepwise increase (which is clearly the case in the variable lexical growth, and considerably less so in the variables utterances with verbs and function words, where the first step occupies a very low level), the level of the first developmental step often corresponds to the variables’ original carrying capacity, which is a function of a host of connected background variables that are not explicitly distinguished in the form of separate variables (examples of such background variables are cognitive understanding, ability to imitate recognizable words without knowing the exact syntactic function or meaning, and so on). Given that the stepwise increase is an obvious property of the current dataset (in particular as concerns the lexical

< previous page

page_88

next page >

< previous page

page_89

next page >

Page 89 growth rate), the carrying capacities of the three variables will be added as free parameters to the model. The growth model that was fitted to the smoothed data contained the following free parameters: the growth rates of the three variables, the carrying capacity of the three variables, and finally for each variable a parameter affecting the remaining two variables. Initial values were set to very low values for the variables utterances with verbs and function words, indicating that such variables are virtually absent during the first observations (for mathematical reasons, the values cannot be set to 0). For the variable growth of lexicon, the initial level was set to a value approximately equal to the observed initial value. The growth model was then fitted to the smoothed data and resulted in a very good fit (see Figure 3.7). The high quality of the fit is not entirely unexpected, since connected growth models can in general achieve very good fits, due to the fact that they contain a considerable number of parameters (growth rates, and support and competition parameters for each possible combination of variables).

Figure 3.7 Smoothed dataset of Peter’s proportion utterances with verbs, lexical change and function words (top), compared with fitted dynamic model (bottom). Note Smoothed curves are based on normalized datasets (ranging from 0 to 1).

< previous page

page_89

next page >

< previous page

page_90

next page >

Page 90 The major question is whether the parameters resulting from the fitting procedure make any theoretical sense, and also whether they can teach us something about the principles of the underlying dynamics of long-term language development. Another important thing while fitting models to data is to look for possible multiple solutions. That is, different combinations of parameter values can result in comparably good fits (the different solutions correspond with locally optimal solutions, and often depend on the values of the initial guesses of the parameters). The dataset is about equally well fitted by the following two sets of parameters (see Table 3.1). Model 1 specifies relatively high growth rates for the three variables, and a carrying capacity of lexical growth which is moderate (0.56), one for function words which is low (0.20) and one for utterances with verbs, which is maximal (1). That is, the increase of lexical growth to its maximal value of 1 is partly due to the supportive effect of the level of function words, which is very strong (0.95). That is, the late growth of function words is indicative of a developmental transition that has a strong positive effect on the acceleration of lexical growth and is responsible for the two-step character of the curve of lexical growth rate (i.e. it is not literally intended that the growth of function words themselves causes an increase in the rate of word learning). On the other hand, however, the effect of the level of utterances with verbs on the acceleration of lexical growth is negative, hence the proportion of utterances with verbs has a damping effect on the acceleration of lexical growth, a phenomenon that is less easy to explain. The growth of the Table 3.1 Growth rates for Models 1 and 2 Model 1 Model 2 Growth rate % utterances with verbs 0.55 0.08 Growth rate lexical growth 0.80 0.81 Growth rate function words 0.51 0.04 Effect of level of utterances with verbs on change in lexical growth −0.38 −0.05 Effect of level of utterances with verbs on change in function words 1.96 0.92 Effect of level of growth of lexicon on change in utterances with verbs 0.00 1.27 Effect of level of growth of lexicon on change in function words −0.58 0.93 Effect of level of function words on growth in utterances with verbs 0.00 −0.69 Effect of level of function words on growth in change in lexical growth 0.95 0.58 Carrying capacity of lexical growth 0.56 0.54 Carrying capacity of function words 0.20 0.02 Carrying capacity of utterances with verbs 1.00 0.15

< previous page

page_90

next page >

< previous page

page_91

next page >

Page 91 number of utterances with verbs is virtually unrelated to the level of lexical growth and the level of function words (the latter is easy to understand, since the growth of utterances containing verbs is basically finished when the growth of function words begins). Finally, the growth in the number of function words is strongly positively affected by the growth in utterances with verbs (both are related to typically “syntactic” aspects of the language), but is negatively related to the rate of lexical growth (rate of lexical growth dampens the rate of growth in function words). These damping effects could be developmentally functional, in the sense that they refer to a certain incompatibility between rapid growth in lexical learning on the one hand and rapid growth in function words on the other hand. The damping effects could also be mathematical artifacts, in the sense that they mathematically compensate for the strong positive effect of the third variable. The choice between these options depends on theoretical arguments, specifying the functional plausibility of such negative influences. Model 2 makes a distinction between lexical growth on the one hand, with a high growth rate of its own, and utterances with verbs and function words on the other hand, which have very low growth rates, and whose growth must thus depend primarily on the influence of other variables that support them. This difference in growth rate is also consistent with the low intrinsic carrying capacities for utterances with verbs and function words (0.15 and 0.02 respectively). Hence, without the direct support of other variables, function words would never get off the ground and the percentage utterances with verbs would remain low. Lexical growth rate has a strong positive effect on the growth of utterances with verbs as well as on the growth of function words (1.27 and 0.93 respectively; during times of rapid lexical growth there is also high growth in utterances with verbs and function words). High levels of function words cause high lexical growth (0.58). The level of utterances with verbs has only little (negative) effect on lexical growth rate. High levels of utterances with verbs cause high growth in function words, but high levels of function words negatively affect the growth of utterances with verbs (−0.69). The structure of relationships in Model 2 is shown in Figure 3.8. In short, the estimated growth parameters specify a system in which the growth in utterances with verbs and in function words strongly depends on external parameters (external to the variables at issue, that is), which in the current model are canalized via the variable lexical growth (which probably reflects a host of factors not explicitly specified in the current model). The growth in proportion of utterances with verbs supports the growth in function words, whereas the growth in function words competes with the growth in proportion utterances with verbs. Finally, the growth of function words positively supports the rate of lexical growth. It is important to note that the current model should not be taken too literally, and the interpretation of the relationships could easily lead to

< previous page

page_91

next page >

< previous page

page_92

next page >

Page 92

Figure 3.8 Graph of relationships between variables in the dynamic growth model. Note Magnitude of influence relationships is shown by the thickness of the connectors, negative (competitive) relationships are shown by dashed lines; VU: proportion utterances with verbs, FW: Function words, GLex: growth rate of lexicon. confusion. The three variables are chosen because they are assumed to be good indicators of an underlying complex dynamic system, which consists of a great many interacting and interdependent components. A characteristic feature of a dynamic system is that its observable behavior is not caused by a single “factor” responsible for that behavior, but that the behavior results from the intertwined and interdependent actions of the many system components, most of which are probably not or only partially known (and this is what makes the system complex and dynamic). Hence, in order to obtain some qualitative understanding of the nature of the dynamics, one can take a number of indicators of the system’s behavior. These indicators are to a certain extent arbitrary (but it is also not so that one can take any possible indicator). In fact, each indicator stands for a hypothetical general and characteristic property of the system, and by taking indicators from lexical learning, composition of utterances and words with particular syntactic functions, I have hoped to tap sufficient diversity in the complex system’s behavior. If we conceive of the change of the system as a process of quantitative growth (which is clearly only one way of seeing the change—linguists will probably focus more on the qualitative, structural changes in the language of the child), we can apply a dynamic growth model to fit the data. This fit, and in particular the parameters resulting from the fit, can hopefully teach us something about the underlying dynamics of the system, which in this case turns out to be a dynamics of mainly supportive relationships. Finally, the variables are not supposed to refer to underlying variables in a transparent way: what they refer to is a largely unknown structure of relationships among components of the system that, in some way, are expressed by the variable to a

< previous page

page_92

next page >

< previous page

page_93

next page >

Page 93 degree that is sufficient for capturing the nature of the dynamics. More specifically, using the variable lexical growth in the dynamic model does not imply that an underlying, identifiable lexical growth factor is assumed. Observable lexical growth refers to a host of interdependent dynamic components, which are responsible for the observed lexical growth and which, in other “soft-assembly” constellations, relate to many other observable aspects of the language system. That this is indeed so is not a proven thing, but an assumption that is consistent with the main assumptions of a complex dynamic systems approach. 3.3 Change and Variability The preceding growth model referred to the smoothed dataset, that is to an estimation of the local averages of the observed levels. The use of such local averages is just a matter of convenience, of reducing the amount of information present in the actual observations, to make it easier to follow the trajectory of growth. It does not entail any assumption about local averages referring to alleged “true” levels. However, the information reduction has also disadvantages, in that the observed levels can contain information that the smoothed curves do not reveal. The study of the meaning of changes in the observed data brings us to the topic of intraindividual variability (see, for instance, de Weerth & van Geert, 2002; van Dijk & van Geert, 2007; van Geert & van Dijk, 2002). Intraindividual variability played an important role in the aforementioned study on developmental transitions by Bassano and van Geert (2007), where peaks of increased fluctuations over short time intervals were interpreted as indicators of underlying transitions in the nature of the linguistic generators. Another example of the use of intraindividual variability in the data in order to uncover the possible underlying dynamics, comes from a study of van Dijk and van Geert (2007) on spatial prepositions. Spatial prepositions were chosen as an indicator variable for possible underlying transitions in the use of typical syntactic categories such as prepositions that refer to relationships among objects (see, for instance, Brown, 1973, suggesting that prepositions might follow a transitional development). In the study by van Dijk and van Geert (2007), four children (Heleen, Jessica, Berend, and Lisa) were followed over the course of a year, from around age 1 year and 6 months to age 2 years and 6 months. The study is based on videotaped observations of spontaneous speech under naturalistic circumstances (the child’s home; for further details on the data collection and measurement design, see van Geert & van Dijk, 2002). Figure 3.9 shows the observed values for Heleen and Jessica. Heleen’s data clearly show a transitional jump, Jessica’s data show a more gradual change, and one would be inclined to represent the data by means of a simple linear regression. However, inspection of the intraindividual variability in the data shows a different pattern. A simple way to show

< previous page

page_93

next page >

< previous page

page_94

next page >

Page 94

Figure 3.9 Use of prepositions (frequencies of occurrence during a 60-minute observation episode) of Jessica and Heleen. Note Squared difference scores refer to intra-individual variability (dotted line, left axis). intraindividual variability is to plot the distance between two consecutive data points (the distance is the absolute value of the difference between the points, e.g. a level of proposition use at observation 5 and one at observation 6); another way is to take the squared value of the difference, which magnifies the variability. Figure 3.9, which shows the levels of spatial prepositions and the squared differences, suggests two variability peaks in Jessica’s data, and one in Heleen’s data. The variability peaks help us discern developmental steps in the data (Heleen’s data show one very clear step; those of Jessica suggest two such steps). In order to better show the eventual stepwise character of the data, the extreme values can be used to make a socalled progressive maximum–regressive minimum plot (the technique and rational are described in van Geert

< previous page

page_94

next page >

< previous page

page_95

next page >

Page 95 and van Dijk, 2002; see also Verspoor, Lowie and van Dijk, 2008 for applications in second language development). This plot represents the “envelope” around the data, thus specifying the data as a range of varying width, instead of a range characterized solely by its central tendency (e.g. a regression model, moving average, etc.). Figure 3.10 shows the plot of Heleen’s and Jessica’s data and suggests a one-step transition in Heleen and a two-step transition in Jessica. In order to statistically test whether these transitions are real discontinuities, van Dijk and van Geert (2007) developed a technique for determining whether an observation that one assumes to correspond with a transition is indeed discontinuous on the intraindividual variability range of the preceding observations. The results show at least one discontinuity in the data of the four children under study (Heleen and Jessica, and Lisa and Berend, who are not discussed in the current chapter). In short, the intraindividual variability data show transitions in the four children, although visual inspection of the levels of preposition use suggests that only one of the four children—Heleen— shows such a transition.

Figure 3.10 Progressive maximum–regressive minimum representations of spatial preposition use in Jessica and Heleen. Note In combination with the intra-individual variability peaks, Heleen’s graph suggests one transition, Jessica’s suggests two transitions or stepwise changes.

< previous page

page_95

next page >

< previous page

page_96

next page >

Page 96 4 Conclusion Complex dynamic systems provide a general approach to language development, defining language development as a dynamic system consisting of a great number of interdependent components and properties, operating on various interconnected timescales. Observable variables are viewed as ways to obtain insight into the underlying dynamics of the system, although there is no direct way of relating the observed variables to clearly discernible causal components or to any simple and manageable system that underlies the observed phenomena. Language is a complex system operating not only on different timescales but also on different levels of organization. It is a socially shared and physically grounded pattern of activity, allowing for a particular type of symbolic, that is linguistic, action. Young children have to grow into the structure of this shared symbolic action, and they are biologically and socially well prepared for this developmental process. Meanwhile, language can be conceived of as a complex dynamic system of its own, as a network of interdependent dynamic components that find their expression in real-time processes of language production and understanding. The timing issues that young children are confronted with require a coordination of the short-term timescale of actual language use and the long-term timescale of language development. The principles of language construction and acquisition discussed in this chapter focused on early and first language acquisition. The general underlying principles, however, are universal, in the sense that they are applicable to processes of change, adaptation and learning throughout the human lifespan. Their “modus operandi,” however, will depend on the contexts in which they feature, and such contexts tend to change with development and aging (consider, for instance, the effect of an aging nervous system on the processes of learning, or the effect of increasing knowledge, for instance in L2 acquisition as opposed to L1). The methods discussed in this chapter—dynamic growth models, epigenetic and dynamic field models, the study of intraindividual fluctuation—are not intended as replacements for the formal and categorical descriptions and analyses that are part and parcel of the linguistic understanding of language and language development. A more complete understanding of language development will hopefully show how these formal linguistic analyses can be reconciled with the dynamic, more quantitatively oriented approach presented in this chapter and will help us go beyond all too simple conceptions of the underlying structures and mechanisms of developing and evolving language, which is probably one of the most complicated structures in the universe of phenomena that are available to scientific inquiry.

< previous page

page_96

next page >

< previous page

page_97

next page >

Page 97 References Atkinson, D., Churchill, E., Nishino, T., & Okada, H. (2007). Alignment and interaction in a sociocognitive approach to second language acquisition. Modern Language Journal, 91(2), 169–188. Auer, P. (2007). On-line syntax: Thoughts on the temporality of spoken language. Language Sciences, in press, corrected proof. Axelrod, R. (1997). The complexity of cooperation: Agent-based models of competition and collaboration. Princeton, NJ: Princeton University Press. Ball, P. (2005). Critical mass: How one thing leads to another. London: Arrow Books/Random House. Bankes, S., & Lempert, R. (2004). Robust reasoning with agent-based modeling. Nonlinear Dynamics, Psychology, and Life Sciences, 8 (2), 259–277. Barsalou, L., Breazeal, C., & Smith, L. (2007). Cognition as coordinated non-cognition. Cognitive Processing, 8 (2), 79–91. Bassano, D., & van Geert, P. (2007). Modeling continuity and discontinuity in utterance length: A quantitative approach to changes, transitions and intra-individual variability in early grammatical development. Developmental Science , 10(5), 588–612. Berthouze, L., Metta, G., & Sun, R. (2005). Editorial: Epigenetic robotics: Modelling cognitive development in robotic systems. Cognitive Systems Research , 6 (3), 189–192. Bickhard, M. (2007). Language as an interaction system. New Ideas in Psychology, 25(2), 171–187. Bigelow, A., MacLean, K., & Proctor, J. (2004, November). The role of joint attention in the development of infants’ play with objects. Developmental Science , 7 (5), 518–526. Bjorne, P., & Balkenius, C. (2005). A model of attentional impairments in autism: First steps toward a computational theory. Cognitive Systems Research , 6 , 193–204. Borenstein, E., & Ruppin, E. (2005). The evolution of imitation and mirror neurons in adaptive agents. Cognitive Systems Research , 6 , 229–242. Breazeal, C. (2003). Emotion and sociable humanoid robots. International Journal of Human–Computer Studies, 59(1), 119–155. Breazeal, C., & Brooks, R. (2005). Robot emotion: A functional perspective. In J.-M. Fellous & M. A. Arbib (Eds.), Who needs emotions? The brain meets the robot (pp. 271–310). New York: Oxford University Press. Breazeal, C., & Scassellati, B. (2002). Robots that imitate humans. Trends in Cognitive Sciences, 6 , 481–487. Brown, R. (1973). A first language: The early stages. London: Allen & Unwin. Butler, C. S. (2008). Cognitive adequacy in structural–functional theories of language. Language Sciences, 30, 1–30. Call, J., & Carpenter, M. (2002). Three sources of information in social learning. In K. Dautenhahn & C. L. Nehaniv (Eds.), Imitation in animals and artifacts (pp. 211–228). Cambridge, MA: MIT Press. Callaghan, T. C., Rochat, P., & MacGillivray, T. (2004). Modeling referential actions in 6- to 18-month-old infants: A precursor to symbolic understanding. Child Developmen t, 75, 11733–11744. Cañamero, L., & Gaussier, P. (2005). Emotion understanding: Robots as tools and

< previous page

page_97

next page >

< previous page

page_98

next page >

Page 98 models. In J. Nadel & D. Muir (Eds.), Emotional development: Recent research advances (pp. 235–258). New York: Oxford University Press. Cangelosi, A. (2006). The grounding and sharing of symbols. Pragmatics and Cognition, 14, 275–285. Cangelosi, A. (2007). Adaptive agent modeling of distributed language: Investigations on the effects of cultural variation and internal action representations. Language Sciences, 29(5), 633–649. Carpenter, M., Call, J., & Tomasello, M. (2005). Twelve- and 18-month-olds copy actions in terms of goals. Developmental Science , 8 , 13–20. Carr, P. (2007). Internalism, externalism and coding, Language Sciences, 29, 672–689. Case, P., Tuller, B., Mingzhou, D., & Kelso, J. A. S. (1995). Evaluation of a dynamical model of speech perception. Perception & Psychophysics, 57, 977–988. Chemero, A. (2003). An outline of a theory of affordances. Ecological Psychology, 15(2), 181–195. Chemero, A., & Turvey, M. (2007). Gibsonian affordances for roboticists. Adaptive Behavior, 15(4), 473–480. Christiansen, M. H., & Kirby, S. (2003). Language evolution: Consensus and controversies. Trends in Cognitive Sciences, 7 , 300–307. Clark, A. (1997). Being there: Putting brain, body and world together again . Cambridge, MA: MIT Press. Clark, A. (2006). Language, embodiment, and the cognitive niche. Trends in Cognitive Sciences, 10, 370–374. Cleveland, A., Schug, M., & Striano, T. (2007, May). Joint attention and object learning in 5- and 7-month-old infants. Infant and Child Development , 16(3), 295–306. Cowley, S. J. (2006). Language and biosemiosis: A necessary unity? Semiotica, 162 (1/4), 417–444. Cowley, S. J. (2007). The cognitive dynamics of distributed language. Language Sciences, 29, 575–583. Cowley, S. J., Moodley, S., & Fiori-Cowley, A. (2004). Grounding signs of culture: Primary intersubjectivity in social semiosis. Mind, Culture and Activity, 11(2), 109–132. de Bot, K., Lowie, W., & Verspoor, M. (2007). A dynamic systems theory approach to second language acquisition. Bilingualism: Language and Cognition, 10(1), 7–21. de Weerth, C., & van Geert, P. (2002). Changing patterns of infant behavior and mother–infant interaction: Intraand interindividual variability. Infant Behavior and Development , 24(4): 347–371. Deacon, T. W. (1997). The symbolic species: The co-evolution of language and the human brain. London: Penguin Books. Demiris, Y., & Johnson, M. (2003). Distributed, predictive perception of actions: A biologically inspired robotics architecture for imitation and learning. Connection Science , 15(4), 231–243. Dessing, J., Daffertshofer, A., Peper, C., & Beek, P. (2007). Pattern stability and error correction during in-phase and antiphase four-ball juggling. Journal of Motor Behavior, 39(5), 433–446. Diessel, H. (2007, August). Frequency effects in language acquisition, language use, and diachronic change. New Ideas in Psychology, 25, 108–127. Dominey, P. (2007). Towards a construction-based framework for development of

< previous page

page_98

next page >

< previous page

page_99

next page >

Page 99 \ language, event perception and social cognition: Insights from grounded robotics and simulation. Neurocomputing: An International Journal, 70, 2288–2302. Dominey, P., & Boucher, J. (2005). Developmental stages of perception and language acquisition in a perceptually grounded robot. Cognitive Systems Research , 6 , 243–259. Eilan, N., Hoerl, C., McCormack, T., & Roessler, J. (2005). Joint attention: Communication and other minds: Issues in philosophy and psychology. New York: Clarendon Press/Oxford University Press. Elliott, E., & Kiel, L. (2004). Agent-based modeling in the social and behavioral sciences. Nonlinear Dynamics, Psychology, and Life Sciences, 8 , 121–130. Erlhagen, W., & Schöner, G. (2002). Dynamic field theory of movement preparation. Psychological Review, 109 , 545–572. Estes, K., Evans, J., Alibali, M., & Saffran, J. (2007). Can infants map meaning to newly segmented words? Statistical segmentation and word learning. Psychological Science , 18, 254–260. Fischer, K. W., & Bidell, T. R. (2006). Dynamic development of action, thought, and emotion. In W. Damon & R. M. Lerner (Eds.), Theoretical models of human development. Handbook of child psychology (6th ed., Vol. 1, pp. 313– 399). New York: John Wiley. Fischer, K. W., & Rose, S. P. (1999). Rulers, models, and nonlinear dynamics: Measurement and method in developmental research. In G. Savelsbergh, H. van der Maas, & P. van Geert (Eds.), Nonlinear developmental processes (pp. 197–212). Amsterdam: Royal Netherlands Academy of Arts and Sciences. Gergely, G. (2003). What should a robot learn from an infant? Mechanisms of action interpretation and observational learning in infancy. Connection Science , 15, 191–209. Gergely, G., & Csibra, G. (2003). Teleological reasoning in infancy: The naive theory of rational action. Trends in Cognitive Sciences, 7 , 287–292. Gergely, G., Bekkering, H., & Király, I. (2002). Rational imitation in preverbal infants. Nature, 415 , 755. Goldberg, A., Casenhiser, D., & White, T. (2007). Constructions as categories of language. New Ideas in Psychology, 25(2), 70–86. Greenspan, S., & Shanker, S. (2007). The developmental pathways leading to pattern recognition, joint attention, language and cognition. New Ideas in Psychology, 25, 128–142. Haken, H. (2006, May). Synergetics of brain function. International Journal of Psychophysiology , 60(2), 110–124. Hauser, M., Chomsky, N., & Fitch, W. (2002). The faculty of language: What is it, who has it, and how did it evolve? Science , 298 , 1569–1579. Hawkins, J. (2007). Processing typology and why psychologists need to know about it. New Ideas in Psychology, 25(2), 87–107. Hodges, B. (2007a). Values define fields: The intentional dynamics of driving, carrying, leading, negotiating, and conversing. Ecological Psychology, 19(2), 153–178. Hodges, B. (2007b). Good prospects: ecological and social perspectives on conforming, creating, and caring in conversation. Language Sciences, 29, 584–604. Huys, R., Daffertshofer, A., & Beek, P. (2004). Multiple time scales and subsystem embedding in the learning of juggling. Human Movement Science , 23(3), 315–336.

< previous page

page_99

next page >

< previous page

page_100

next page >

Page 100 Jones, S. S. J., & Smith, L. B. (2002). How children know the relevant properties for generalizing object names. Developmental Science , 5 , 219–232. Kaplan, F., & Hafner, V. (2006). The challenges of joint attention. Interaction Studies, 7 (2), 135–169. Kello, C. T. (2004). Characterizing the evolutionary dynamics of language. Trends in Cognitive Sciences, 8 , 392–394. Kersten, A. W., & Smith, L. B. (2002). Attention to novel objects during verb learning. Child Development , 73, 93– 109. Király, I., Jovanovic, B., Prinz, W., Aschersleben, G., & Gergely, G. (2003). The early origins of goal attribution in infancy. Consciousness and Cognition: An International Journal, 12, 752–769. Kirby, S. (2002). Natural language from artificial life. Artificial Life, 8 , 185–215. Kirby, S., & Hurford, J. R. (2002). The emergence of linguistic structure: An overview of the iterated learning model. In A. Cangelosi & D. Parisi (Eds.), Simulating the evolution of language (pp. 121–148). London: Springer. Kirby, S., Smith, K., & Brighton, H. (2004). From UG to universals: Linguistic adaptation through iterated learning. Studies in Language, 28, 587–607. Kravchenko, A. V. (2007). Essential properties of language, or, why language is not a code. Language Sciences, 29, 650–671. Lavoie, A. (2008, 14 February). Hauser presents theory of “humaniqueness”. Retrieved 6 January, 2009 from http://harvardscience.harvard.edu/foundations/ articles/hauser-presents-theory-humaniqueness. Leslie, A., & Chen, M. (2007). Individuation of pairs of objects in infancy. Developmental Science , 10(4), 423–430. Lewis, M. (2005). Self-organizing individual differences in brain development. Developmental Review, 25, 252–277. Lewis, M., & Todd, R. (2007). The self-regulating brain: Cortical–subcortical feedback and the development of intelligent action. Cognitive Development , 22(4), 406–430. Lichtwarck-Aschoff, A., van Geert, P., Kunnen, S., & Bosma, H. (2008). Time and identity: A framework for research and theory formation. Developmental Review, 28, in press. Lindblom, J., & Ziemke, T. (2006). The social body in motion: Cognitive development in infants and androids. Connection Science , 18(4), 333–346. Linell, P. (2007). Dialogicality in languages, minds and brains: Is there a convergence between dialogism and neurobiology? Language Sciences, 29, 605–620. Lopes, L., & Chauhan, A. (2007). How many words can my robot learn? An approach and experiments with one-class learning. Interaction Studies, 8 (1), 53–81. Love, N. (2007) Are languages digital codes? Language sciences. Cognitive Dynamics in Language, 29, 690–709. Meltzoff, A. N. (1988). Infant imitation and memory: Nine-month-olds in immediate and deferred tests. Child Development , 59, 217–225. Meltzoff, A. N. (1995). Understanding the intentions of others: Re-enactment of intended acts by 18-month-old children. Developmental Psychology, 31, 838–850. Meltzoff, A. N., & Moore, M. K. (1983). Newborn infants imitate adult facial gestures. Child Development , 54, 702– 709. Metta, G., & Berthouze, L. (2006). Epigenetic robotics: Modelling cognitive development in robotic systems. Interaction Studies, 7 (2), 129–134.

< previous page

page_100

next page >

< previous page

page_101

next page >

Page 101 Metta, G., Sandini, G., Natale, L., Craighero, L., & Fadiga, L. (2006). Understanding mirror neurons: A bio-robotic approach. Interaction Studies, 7 (2), 197–232. Morton, J. (1986). Developmental contingency modelling: A framework for discussing the processes of change and the consequences of deficiency. In P. L. C. van Geert (Ed.), Theory building in developmental psychology (pp. 141– 165). New York: Elsevier. Mundy, P., Block, J., Delgado, C., Pomares, Y., van Hecke, A., & Parlade, M. (2007, May). Individual differences and the development of joint attention in infancy. Child Development , 78(3), 938–954. Nagai, Y., Hosoda, K., Morita, A., & Asada, M. (2003). A constructive model for the development of joint attention. Connection Science , 15(4), 211–229. Newell, K., Liu, Y., & Mayer-Kress, G. (2003). A dynamical systems interpretation of epigenetic landscapes for infant motor development. Infant Behavior & Development , 26, 449–472. Petitot, J. (1979). Hypothèse localiste et théorie des catastrophes. Note sur le débat. In M. Piatelli Palmerini (Ed.), Théories linguistiques et théorie de l’apprentissage: Le débat entre Jean Piaget et Noam Chomsky. Paris: Editions du Seuil (pp. 516–24). Petitot, J. (1985). Morphogenèse du sens. Paris: PUF. Port, R. (2007). How are words stored in memory? Beyond phones and phonemes. New Ideas in Psychology, 25, 143–170. Porter, R., & Hogue, D. (1998). Nonlinear dynamical systems in speech perception and production. Nonlinear Dynamics, Psychology, and Life Sciences, 2 (2), 95–131. Raczaszek, J., Tuller, B., Shapiro, L., Case, P., & Kelso, S. (1999). Categorization of ambiguous sentences as a function of a changing prosodic parameter: A dynamical approach. Journal of Psycholinguistic Research , 28, 367– 393. Raczaszek-Leonardi, J., Shapiro, L., Tuller, B., & Kelso, J. (2008). Activating basic category exemplars in sentence contexts: A dynamical account. Journal of Psycholinguistic Research , 37, 87–113. Ripa, J., & Ives, A. R. (2003). Food web dynamics in correlated and autocorrelated environments. Theoretical Population Biology, 64, 369–384. Roberts, L. D., & Lee, C. (2002). Problems about young children’s knowledge of the theory of mind and of intentionality. Journal for the Theory of Social Behaviour , 32, 295–310. Robinson, B. F., & Mervis, C. B. (1998). Disentangling early language development: Modeling lexical and acquisition using and extension of case-study methodology. Developmental Psychology, 34, 363–375. Rose, S. P., & Fischer, K. W. (1998). Models and rulers in dynamical development. British Journal of Developmental Psychology, 16, 123–131. Ruhland, R., & van Geert, P. (1998). Jumping into syntax. British Journal of Developmental Psychology, 16, 65–95. Ruhland, R., Wijnen, F., & van Geert, P. (1995). An exploration into the application of dynamic systems modelling to language acquisition. In M. Verrips & F. Wijnen (Eds.), Approaches to parameter setting (Amsterdam Series in Child Language Acquisition 4) (pp. 107–134). Institute for General Linguistics, Amsterdam. Sandhofer, C. M., Smith, L. B., & Luo, J. (2000). Counting nouns and verbs in the input: Differential frequencies, different kinds of learning? Journal of Child Language, 27(33), 561–585.

< previous page

page_101

next page >

< previous page

page_102

next page >

Page 102 Schlesinger, M. (2003). A lesson from robotics: Modeling infants as autonomous agents. Adaptive Behavior, 11(2), 97–107. Schlesinger, M., & Casey, P. (2003). Where infants look when impossible things happen: Simulating and testing a gaze-direction model. Connection Science , 15, 271–280. Schlesinger, M., & Parisi, D. (2001). The agent-based approach: A new direction for computational models of development. Developmental Review, 21, 121–146. Schöner G., & Dineva, E. (2007). Dynamic instabilities as mechanisms for emergence. Developmental Science , 10, 1 69. Schöner, G., & Thelen, E. (2006). Using dynamic field theory to rethink infant habituation. Psychological Review, 113 (2), 273–299. Schutte, A. R., & Spencer, J. P. (2002). Generalizing the dynamic field theory of the A-not-B error beyond infancy: Three-year-olds’ delay- and experience-dependent location memory biases. Child Development , 73, 377–404. Schutte, A. R., Spencer, J. P., & Schöner, G. (2003). Testing the dynamic field theory: Working memory for locations becomes more spatially precise over development. Child Development , 74, 1393–1417. Senghas, A., Kita, S., & Özyürek, A. (2004, September). Children creating core properties of language: Evidence from an emerging sign language in Nicaragua. Science , 305 (5691), 1779–1782. Shaw, R. (2001). Processes, acts, and experiences: Three stances on the problem of intentionality. Ecological Psychology, 13(4), 275–314. Smith, E., & Conrey, F. (2007). Agent-based modeling: A new approach for theory building in social psychology. Personality and Social Psychology Review, 11(1), 1–18. Smith, K., Kirby, S., & Brighton, H. (2003). Iterated learning: A framework for the emergence of language. Artificial Life, 9 , 371–386. Smith, L. B. (2005). Cognition as a dynamic system: Principles from embodiment. Developmental Review, 25, 278– 298. Smith, L. B., Thelen, E., Titzer, R., & McLin, D. (1999). Knowing in the context of acting: The task dynamics of the A-not-B error. Psychological Review, 106 , 235–260. Steenbeek, H., & van Geert, P. (2007a). “Do you still like to play with him?” Variability and the dynamic nature of children’s sociometric ratings. Netherlands Journal of Psychology, 63, 86–101. Steenbeek. H., & van Geert, P. (2007b). The empirical validation of a dynamic systems model of interaction: Do children of different sociometric statuses differ in their dyadic play interactions? Developmental Science , 11, 253– 281. Steenbeek. H., & van Geert, P. (2008). A dynamic systems approach to dyadic interaction in children: Emotional expression, action, dyadic play, and sociometric status. Developmental Review, 27, 1–40. Striano, T., Chen, X., Cleveland, A., & Bradshaw, S. (2006). Joint attention social cues influence infant learning. European Journal of Developmental Psychology, 3 (3), 289–299. Striano, T., Henning, A., & Stahl, D. (2006). Sensitivity to interpersonal timing at 3 and 6 months of age. Interaction Studies, 7 (2), 251–271. Striano, T., Henning, A., & Vaish, A. (2006). Selective looking by 12-month-olds to a temporally contingent partner. Interaction Studies, 7 (2), 233–250. Striedter, G. (1998). Stepping into the same river twice: Homologues as recurring

< previous page

page_102

next page >

< previous page

page_103

next page >

Page 103 attractors in epigenetic landscapes. Brain, Behavior and Evolution, 52(4), 218–231. Thagard, R. (1996) Mind: Introduction to cognitive science. Cambridge, MA: MIT Press. Thelen, E., & Smith, L. (1994). A dynamic systems approach to the development of cognition and action . Cambridge, MA: MIT Press. Thelen, E., & Smith, L. (1998). Dynamic systems theories. Handbook of child psychology: Volume 1: Theorectical models of human development (5th ed., pp. 563–634). Hoboken, NJ: John Wiley. Thelen, E., Schöner, G., Scheier, C., & Smith, L. (2001). The dynamics of embodiment: A field theory of infant perseverative reaching. Behavioral and Brain Sciences, 24, 1–86. Thom, R. (1972). Stabilité structurelle et morphogenèse . New York: Benjamin (also Paris: Intereditions, 1977). Thompson, D. E., & Russell, J. (2004). The ghost condition: Imitation versus emulation in young children’s observational learning. Developmental Psychology, 40, 882–889. Tomasello, M. (1995). Joint attention as social cognition. In C. Moore & P. J. Dunham (Eds.), Joint attention: Its origins and role in development (pp. 103–130). Hillsdale, NJ: Lawrence Erlbaum Associates. Tomasello, M., & Rakoczy, H. (2003). What makes human cognition unique? From individual to shared to collective intentionality. Mind and Language, 18, 121–147. Tschacher, W., & Haken, H. (2007). Intentionality in non-equilibrium systems? The functional aspects of selforganized pattern formation. New Ideas in Psychology, 25(1), 1–15. Tuller, B., Jantzen, G., & Jirsa, V. K. (2008). A dynamical approach to speech categorization: Two routes to learning. New Ideas in Psychology. In press, corrected proof. Turvey, M. (1992). Affordances and prospective control: An outline of the ontology. Ecological Psychology, 4 , 173– 187. van Dijk, M., & van Geert, P. (2007). Wobbles, humps and sudden jumps: A case study of continuity, discontinuity and variability in early language development. Infant and Child Development , 16(1), 7–33. van Geert, P. (1991). A dynamic systems model of cognitive and language growth. Psychological Review, 98, 3–53. van Geert, P. (1994). Dynamic systems of development. Change between complexity and chaos . New York: Harvester. van Geert, P. (1998). A dynamic systems model of basic developmental mechanisms: Piaget, Vygotsky and beyond. Psychological Review, 105 (5, 4), 634–677. van Geert, P. (2000). The dynamics of general developmental mechanisms: From Piaget and Vygotsky to dynamic systems models. Current Directions in Psychological Science , 9 (2), 64–68. van Geert, P. (2003). Dynamic systems approaches and modeling of developmental processes. In J. Valsiner & K. J. Conolly (Eds.), Handbook of developmental psychology (pp. 640–672). London: Sage. van Geert, P., & Fischer, K. W. (2008; in press). Dynamic systems and the quest for individual-based models of change and development. In J. P. Spencer, M. S. C. Thomas, & J. McClelland (Eds.), Toward a new grand theory of development? Connectionism and dynamic systems theory re-considered . Oxford: Oxford University Press.

< previous page

page_103

next page >

< previous page

page_104

next page >

Page 104 van Geert, P., & Steenbeek, H. (2005). Explaining after by before. Basic aspects of a dynamic systems approach to the study of development. Developmental Review, 25(3–4), 408–442. van Geert, P., & van Dijk, M. (2002). Focus on variability: New tools to study intra-individual variability in developmental data. Infant Behavior & Development , 25, 340–374. Verspoor, M., Lowie, W., & van Dijk, M. (2008). Variability in second language development from a dynamic systems perspective. Modern Language Journal, 92, 214–231. Visetti, Y.-M. (2004). Language, space and the theory of semantic forms. In A. Carsetti (Ed.), Seeing, thinking and knowing (pp. 245–275). Dordrecht: Kluwer. Vogt, P., & Divina, F. (2007). Social symbol grounding and language evolution . Interaction Studies: Social Behaviour and Communication in Biological and Artificial Systems , 8 (1), 31–52. Wildgen W. (1981). Archetypal dynamics in word semantics: An application of catastrophe theory. In H. J. Eikemeyer & H. Raiser (Eds.), Words, worlds and contexts (pp. 234–296). New York: de Gruyter. Wildgen W. (1982) . Catastrophe theoretic semantics. An elaboration and application of René Thom’s theory. Amsterdam: John Benjamins. Wildgen, W. (2002). Natural ontologies and semantic roles in sentences. Axiomathes, 12, 171–193. Wildgen, W. (2006). Emergence of semantic and syntactic complexity in language. Paper presented at the Language Culture and Mind Conference: Integrating Perspectives and Methodologies in the Study of Language, Paris, July 17– 20, 2006. Woodward, A. L. (1998). Infants selectively encode the goal object of an actor’s reach. Cognition, 69, 1–34. Woodward, A. L. (1999). Infants’ ability to distinguish between purposeful and non-purposeful behaviors. Infant Behavior and Development , 22, 145–160. Woodward, A. L. (2005). Infants’ understanding of the actions involved in joint attention. In N. Eilan, C. Hoerl, T. McCormack, & J. Roessler (Eds.), Joint attention: Communication and other minds: Issues in philosophy and psychology (pp. 110–128). New York: Clarendon Press/Oxford University Press. Woodward, A. L., Sommerville, J. A., & Guajardo, J. J. (2001). How infants make sense of intentional action. In B. F. Malle, L. J. Moses, & D. A. Baldwin (Eds.), Intentions and intentionality: Foundations of social cognition (pp. 149– 169). Cambridge, MA: MIT Press. Yavuz, H. (2007). An integrated approach to the conceptual design and development of an intelligent autonomous mobile robot. Robotics and Autonomous Systems , 55(6), 498–512. Yoshikawa, Y., Asada, M., Hosoda, K., & Koga, J. (2003). A constructivist approach to infants’ vowel acquisition through mother–infant interaction. Connection Science , 15, 245–258. Yu, C., & Ballard, D. H. (2007). A unified model of early word learning: Integrating statistical and social cues. Neurocomputing , 70(13–15), 2149–2165. Zeeman, E. C. (1976). Catastrophe theory. Scientific American, 236 (4), 65–83.

< previous page

page_104

next page >

< previous page

page_105

next page >

page_105

next page >

Page 105 Part II SECOND LANGUAGE DEVELOPMENT

< previous page

< previous page

page_106

next page >

page_106

next page >

Page 106 This page intentionally left blank.

< previous page

< previous page

page_107

next page >

Page 107 4 LIFESPAN DEVELOPMENT OF THE L2 AS AN INTELLECTUALIZATION PROCESS An Ontogenetic Sociocultural Theory Perspective Maria C. M. de Guerrero Introduction In this chapter second language (L2) learning is treated as an intellectualization process characterized by the internalization of L2 social speech and its potential culmination in the capacity to think verbally in the L2. This developmental process is seen through the lens of Vygotskyan sociocultural theory (SCT), a set of principles and hypotheses inspired in the work of Vygotsky and followers which allows for the observation of psycholinguistic development in ontogenesis (development over the lifespan). The focus will thus be on those processes leading to the development of the capacity to think through a new language which are related to the age of the learner—not as a biologically governed maturational factor, but as a function of radical psychological changes resulting from the interweaving of nature and culture in human development. In particular, the impact of learning a new language on the intellectual development of the individual at various stages throughout the lifespan will be addressed. This chapter will first lay out some basic general principles of SCT pertaining to cognitive and linguistic development and will then focus on what research has to say about the development of the faculty to think verbally in the L2, with particular reference to the age factor. The chapter will end with recommendations for research on insufficiently attended age-related issues of L2 development and pedagogical implications for the development of the L2 as a thinking tool among children and adults.

< previous page

page_107

next page >

< previous page

page_108

next page >

Page 108 Sociocultural Theory and Lifespan Development As a theory of mind which places emphasis on the developmental history of higher psychological functions, SCT is an ideal framework for the study of cognitive and linguistic development across the lifetime. From the start, it should be made clear that SCT does not treat cognitive development as a series of gradual cumulative changes in psychological functions associated mainly with neurophysiological maturational factors. Rather, Vygotskyan theory views intellectual development as a revolutionary process characterized by moments of crisis (referred to as “critical periods” by Vygotsky, 1998; Mahn, 2003) at which a fundamental transformation in modes of thinking occurs as a result of the introduction of new forms of mediation—culturally based, indirect “instruments” of behavior (Vygotsky, 1978, p. 73)—in the child’s interaction with the environment. Vygotsky rejected approaching the analysis of cognitive development on the basis of a single explanatory criterion that could serve for all ages; such an approach would not take into account the multiplicity of factors that are at play in development at any given moment (Vygotsky, 1998, p. 188; Wertsch, 1985, p. 20). Rather, he insisted on taking into account the “social situation of development” (Vygotsky, 1998, p. 198; Karpov, 2003, p. 139) as the particular and unique relationship that occurs between the child and his/her environment at the beginning of each critical age period. Specifically, Vygotsky rejected attributing psychological changes solely to maturational factors. Although he recognized the role of the biological endowment, processes, and constraints, Vygotsky viewed the introduction of socially instituted forms of mediation within a culture —language in particular—as crucial to the radical transformation of psychological functions. The Genetic Method For Vygotsky, the best approach to understanding human cognition lies in genetic analysis. According to Vygotsky (1978), to truly appreciate the nature of higher mental functions, they must be studied in their formation, as processes that change over time. Vygotsky included several genetic domains as necessary for the study of mental development: the phylogenetic (biological evolution of the species), the sociocultural (society’s cultural history), the ontogenetic (individual development over a lifetime), and the microgenetic (directly observable, short-term changes in behavior) (Wertsch, 1985). Vygotsky devoted much attention to ontogenesis, as a process resulting from multiple causal forces—biological, cultural, and historical—operating simultaneously (Wertsch, 1985, p. 41), and as a process that could particularly explain how the growth of a child into a fully articulated intellectual being could be distinguished from that of other anthropoids (Vygotsky, 1986).

< previous page

page_108

next page >

< previous page

page_109

next page >

Page 109 The ontogenesis of human mental functioning, unlike that of apes and other higher species, is crucially affected by the intersection at some moment in the child’s life (around the age of 2, according to Vygotsky, 1986, p. 82) of two separate lines of development: thinking and speech. It is at this point that elementary thinking functions, such as memory, attention, and perception, representing the “natural” line of thinking, are drastically altered by the use of language (the “cultural” line of development). Through speech—the active use of language—the child starts to see the world differently, not just in “color and shape” but “as a world with sense and meaning” (Vygotsky, 1978, p. 33). At this critical juncture, the child enters the stage of symbolic thinking (see Tomasello, 2003, pp. 27–28). Internalization Vygotsky argued that higher psychological processes, such as voluntary attention, complex (indirect) perception, logical memory, and rational thought, are mediated principally not only by language but also by other psychological artifacts (signs or tools of a psychological nature, such as mathematical symbols, mnemonic devices, and gestures), which are the products of particular cultural and historical social groups. The process by which signs come to mediate mental processes is known as internalization. This process entails the transformation of external signs and sign-mediated activity into internal forms of mediation. Internalization does not mean the incorporation of external forms of mediation into an already existing mental plane but the creation of a new, culturally mediated, intrapersonal sphere of intellectual action (Galperin, 1967). The internalization of cultural signs and activities by necessity implies a radical alteration of their form and functions. In the case of speech signs and activity, the resulting intramental outcome is inner speech (Vygotsky, 1978, p. 57). From Social to Inner Speech In child ontogenesis, speech internalization occurs roughly over a span of 4 years (“between three and seven years” of age, Vygotsky, 1986, p. 230). Vygotsky identified four phases in the child’s development of speech and thought. The first is the prelinguistic and preintellectual phase of development. At this point, thought is non-verbal and speech is non-intellectual. Prelinguistic thought is typical of the practical, concrete operations of the child; preintellectual speech (babblings, first words) has a purely social purpose. At about the age of 2, the child begins to realize that things have a name. This revolutionary discovery in the child’s life signals the onset of the intellectualization of speech: “Speech begins to serve the intellect” (p. 82). Thought and language begin to interconnect, but the relationship is limited. The child may use correct grammar but does not fully grasp the logic behind

< previous page

page_109

next page >

< previous page

page_110

next page >

Page 110 it: The child “masters the syntax of speech before the syntax of thought” (p. 87). Although the intellectualization of speech has started, speech continues to be eminently social. Gradually, the child moves on to the third phase, the stage of external operations and external speech. Mental operations are conducted externally (e.g. counting with the fingers), and speech is used externally for both social and thinking purposes. Speech for thinking is external but selfdirected. The child thinks aloud through what is called “egocentric speech.” At about school age, the child enters the fourth stage. External operations can be carried out internally, and egocentric speech turns inward. The child begins to operate intellectually on the basis of inner signs and logic. Thought and language finally merge as verbal thought and inner speech.1 This final stage signals the child’s ability to “think words” (Vygotsky, 1986, p. 230) rather than to merely pronounce them. Concept Development Cognitive/linguistic development does not stop with the internalization of the speech function. As we have seen, the relationship between words and meanings begins to be developed in infancy, when children discover that each thing has a name; however, merely connecting words to objects is not enough to produce concepts. According to Vygotsky, “real concept formation and abstract reasoning appear only in adolescents” (1986, p. 98).2 Concept formation does not result from a quantitative increase in associative links between words and objects (or from an increment in other functions, such as attention, imagery, and judgment), but entails a qualitative change in the nature of signs, a process which ripens in puberty (Vygotsky, 1986, pp. 106–109).3 An integral part of the process of concept formation is relating concepts to word meanings. Just as concept formation never ends, so do word meanings continually evolve. “When a new word has been learned by the child, its development is barely starting” (Vygotsky, 1986, p. 149). This implies that, as “dynamic rather than static formations” (p. 217), word meanings and senses continue to be acquired and modified throughout life. Instruction plays a powerful role in the development of concepts and word meanings, particularly by introducing scientific concepts. When learning scientific concepts in school, the child does not have to reach back to the world of objects and recapitulate the progressive evolution of conceptual development; the acquisition of scientific concepts is mediated by spontaneous concepts already acquired and by the (adult) teacher’s intervention (pp. 148–161). Similarly, Vygotsky contends, “while learning a foreign language, we use word meanings that are already well developed in the native language, and only translate them” (p. 159). The native language and spontaneous concepts play a mediative role in the acquisition of a foreign language (FL) and scientific concepts, respectively; in turn, development of spontaneous con-

< previous page

page_110

next page >

< previous page

page_111

next page >

Page 111 cepts and mastery of the native language are enhanced by the learning of scientific concepts and a new language (pp. 196–197). Metaconsciousness and Control Intellectualization of the higher psychological functions is not complete without development of metaconsciousness,4 that is, awareness and control of the intellectual plane itself (Vygotsky, 1986, p. 167). With the development of metaconsciousness, the child becomes aware of his/her own mental operations and starts thinking about them and exercising control over them rather than merely thinking through them, a process which is mediated by verbalized introspection and to a large extent promoted by development of inner speech and the acquisition of scientific concepts (Vygotsky, 1986, pp. 170–171). In fact, awareness and control of higher mental functions emerge only after these functions have been internalized: The general law of development says that awareness and deliberate control appear only during a very advanced stage in the development of a mental function, after it has been used and practiced unconsciously and spontaneously. In order to subject a function to intellectual and volitional control, we must first possess it. (Vygotsky, 1986, p. 168) In short, the intellectualization of the speech function in L1 acquisition is viewed from the Vygotskian perspective as a process characterized by revolutionary changes in the relationship between language and thought, starting in early childhood with the progressive internalization of social speech, continuing with the emergence of the inner speech function, and reaching its peak in adolescence with concept formation, a process that goes on throughout the lifespan, mainly through conceptual and word meaning development. Enabled by inner speech and the ability to think abstractly in concepts full realization of the speech function finally comes through metaconscious awareness and control. Sociocultural Theory and L2 Development Preintellectual Stage in L2 Development Although the literature on inner speech/verbal thought development from an L2 point of view is scant (see Guerrero, 2005, for a review of this literature), the limited evidence that exists would suggest that, as in L1 development, there is a preintellectual stage in L2 development. A preintellectual stage implies that the speech produced serves mainly a social or communicative purpose rather than a psychological one. While in native language

< previous page

page_111

next page >

< previous page

page_112

next page >

Page 112 development this phase would correspond to the period marked from the child’s earliest speech manifestations up until around the age of 3 (Vygotsky, 1986), in non-concurrent L2 learning (cases when the L2 is learned after an L1 has been acquired) a preintellectual phase would occur at any age at early stages of acquisition of the L2. In other words, with the exception of children who are learning an L2 simultaneously with the L1, in which case the preintellectual phase for the L1 would be similar to that of the L2, both children and adults starting to learn a language would undergo a preintellectual phase where the L2 would serve primarily a limited social, communicative purpose. A preintellectual stage in L2 development would involve the use of the L2 mainly for basic communicative purposes and would not entail merging of thought and the L2 at the level of inner speech. This stage is characteristic of learners who are beginning to learn a language. At this stage the learner’s language is mostly a mere reproduction of the speech in the environment, does not involve spontaneous or creative dialogic interaction, and relies mostly on formulaic expressions. The learner’s speech at this stage might resemble L2 social speech in sound, syntax, and lexicon, but has not really undergone the complex process of L2 internalization entailing syntactic, lexical, and, above all, semantic analysis and conceptual modification which would allow thinking to be mediated by the L2. At this stage there is no possibility of producing language creatively to communicate ideas or to engage in meaningful dialogic interaction in a sustained manner without recourse to direct translation from the L1. Thinking is still mediated by the L1 (John-Steiner, 1985, pp. 362–363; Kecskes & Papp, 2000, p. 64; Leontiev, 1981, p. 26). Reading and listening in the L2 are reduced to decoding written and oral text, laborious translation into the L1, and minimal comprehension. Significantly, there is no apparent use of the L2 to carry on everyday cognitive tasks. Stage of Private Speech As the process of internalization of the L2 gets under way, learners start giving signs of emergent intellectualization of the L2. A critical step is the use of audible private speech and subvocal (silent) self-directed use of the L2 (Lantolf & Thorne, 2006; Ohta, 2001). As learners continue experimenting with social uses of the L2, that is, utilizing L2 to communicate with others, learners also start privately manipulating the L2. This private speech is characterized by being usually imperceptible to outside ears (but vocalized enough to allow recording) and by being oriented to the self, not to others. There is enough evidence in the L2 literature to suggest that private speech among L2 learners has an important internalizing function, that is, a crucial role in turning inwards the L2 social speech in the environment and contributing to the formation of an internal mental (intellectual) plane mediated by the L2.5 As an ontogenetically interesting phenomenon, L2 private speech has been documented among children, adolescents, and adults.

< previous page

page_112

next page >

< previous page

page_113

next page >

Page 113 Case studies (see, for example, Leopold, 1949; Chen, 1988) conducted with preschool children raised in bilingual environments show use of private speech in both L1 and L2, indicating that the process of internalization—and thus intellectualization—of both languages can start before school age. Private speech in the L2 has also been documented among children who start learning the L2 after their L1 is well developed (take, for instance, Uguisu’s case in Hakuta, 1986). One of the most important studies of private speech among children learning an L2 in classroom contexts comes from Saville-Troike (1988). This study throws light on the role of private speech in the process of intellectualization of the L2 among child learners. In this study, six out of nine children of Chinese, Japanese, and Korean background, who were enrolled in nursery and elementary school classes (ages ranging from 3;3 to 8;3) in the US, were observed to produce private speech in English while ostensibly undergoing a “silent” period. Their L2 private speech took five forms: (a) selective repetition of others’ utterances, (b) recall and practice of words or phrases heard previously, (c) creation of new linguistic forms, (d) pattern practice of sentences involving substitution and expansion, and (e) rehearsal before speaking aloud to others. While some of the learners’ private speech production were “unanalyzed quotations” (Saville-Troike, 1988, p. 583) of someone else’s speech showing no comprehension of what was said, other utterances reflected more meaningful, intentional, and creative use of the L2. For example, the following learner’s sentences do not appear to be faithful repetitions of language in the environment and, as Saville-Troike comments, they were “clearly well ahead of his overt social speech” (p. 585): CHILD: Mine is let’s go (i.e. “I will leave”) CHILD: My it’s go house in (i.e. “I am going into the house”) (p. 585) Private speech in the L2 for internalization purposes was also found among older children in Broner and Tarone’s (2001) study of language play in a fifth-grade US Spanish immersion classroom. As seen, there is sufficient evidence to suggest that private speech is a common occurrence among L2 learners of preschool and elementary school age. At least one study (de Courcy, 1995) produced data showing L2 private speech use among adolescents. In de Courcy’s study, four Year 9 (Junior High) students at a late French immersion program in Australia were found to be participating in the lesson apparently silently but using self-directed speech in French. According to de Courcy, “when one student answered one of the teacher’s questions out loud in French, there were at least two others answering privately in French” (p. 15), as can be observed in the following episode:

< previous page

page_113

next page >

< previous page

page_114

next page >

Page 114 TEACHER: Avec un front froid et une pression basse, quel temps? ( With a cold front and low pressure, what sort of weather (do you have ) ?) STUDENT 1 [to self]: Il fait froid ( It’s cold) STUDENT 2 [to self]: Nuages … poussent ( Clouds … push ) (p. 15) Another use of private French among de Courcy’s (1995) learners was to practice privately before producing an oral answer: STUDENT [to self]: Cent quatre-vingts divisé par trois ( 180 divided by 3 ) [then she gave her answer out loud] (p. 15) Private speech has also been found in several studies among instructed adult L2 or FL learners (Centeno-Cortés, 2003; Lantolf, 2003; Lantolf & Yáñez, 2003; Ohta, 2001). A review of these studies reveals at least five functions of private speech as internalization of the L2 among adults: vicarious response (uttering to oneself a response to a question posed to others), repetition and imitation, manipulation of language forms, practice before speaking publicly, and metalinguistic awareness. Although most of these functions are also displayed by child L2 learners, two of them show differences between children and adults. First is imitation, which is viewed in SCT not as the simple mimicking of external models but as “an intentional, complex, and potentially transformative process” fundamental to L2 learning (Lantolf & Thorne, 2006, p. 176). According to Lantolf (2003), children differ from adults in their use of imitation through private speech in that children are more likely to engage in “persistent” imitation (creative reconstruction of models) (p. 366). Lantolf explains this difference by suggesting that children are less inhibited than adults in experimenting with language whereas adults, especially college learners, are more bound by the pressure to give “correct” answers. Another use of L2 private speech which seems to differ between children and adults is metalinguistic understanding of the L2 and overt comparisons between the L1 and the L2 (Lantolf, 2003). The following example, from Lantolf and Yáñez (2003), shows a college learner of Spanish vocalizing to herself the grammatical label “imperfecto ” in her search for the correct verb form: I need to change, I need to change fue … sería [20-second pause] Imperfecto [whispered] (p. 106) Private speech of this type was not observed among the young children in Saville-Troike’s (1988) study. In fact, metalinguistic comments in the

< previous page

page_114

next page >

< previous page

page_115

next page >

Page 115 private speech of instructed L2 or FL learners might be a consequence of the objectification of the language and metalinguistic awareness which come along with traditional language instruction. As Vygotsky (1986) pointed out, FL instruction results in increased awareness of linguistic functions, including those in the native language (p. 196). Metalinguistic awareness, however, may not reveal itself through private speech until the learner has made the L2/FL a conscious object of study, thus being limited to adults or older children who have already objectified language study. Despite the evidence suggesting that private speech has a significant role in learning the L2 among children, adolescents, and adults, overt private speech does not seem to be a necessary or even universal phase in the internalization of an L2. Some studies (Centeno-Cortés, 2003; Saville-Troike, 1988), in fact, show that not all learners engage in vocalized private speech, while others (Ohta, 2001) evidence only minimal private speech use for some learners.6 While private speech in the L2 may be widespread among preschool bilingual children or young L2 learners who have not yet distinguished between social and inner speech in their L1, older or adult language learners may refrain from using audible private speech because they fear it might be considered socially inappropriate, and may opt instead to engage in covert verbalizations, inaudible to observers. In this case, the internalizing function of private speech might be taken over by what Galperin (1967) calls “external speech to oneself,” that is, “ordinary speech without the volume” (p. 30), a transitional phase that temporarily serves as mediator of mental action while the more permanent stage of inner speech is being formed. External speech to oneself may be understood as an “incipient form of inner speech” (Guerrero, 2005) in which L2 learners engage alongside audible private speech or instead of it. Lantolf and Thorne (2006) believe that learners do not necessarily have to “produce audible private speech in order to internalize a language. Some may prefer, for whatever reasons, to keep things hidden either as subvocal private speech or as inner speech” (p. 184).7 Early Stages of Inner Speech Reports of silent or subvocal private speech show incipient inner speech processes, suggesting the progressive intellectualization of the L2. Incipient inner speech is not observable but has been reported through retrospective methods, such as diaries (Guerrero, 2004), questionnaires (Guerrero, 1994, 1999), and interviews (Guerrero, 1994, 2004). The diary study conducted by Guerrero (2004; see also Guerrero, 2005) among beginning English as a second language (ESL) college students revealed four types of silent self-directed speech in the L2: (a) concurrent processing of language being heard or read (such as analyzing words or associating them to images and events), (b) recall of language heard, read, or used previously (sometimes as uninten-

< previous page

page_115

next page >

< previous page

page_116

next page >

Page 116 tional playback or as deliberate recall), (c) preparation before writing or speaking (for example, planning a conversation before holding it), and (d) silent verbalization of thoughts for private purposes (that is, inwardly verbalizing feelings, ideas, and intentions for oneself).8 The following ESL student’s verbal report (translated from Spanish) is offered as illustration of (c) above: I had to call my brother who lives in the States, but because my sister-in-law might answer the phone and she doesn’t speak Spanish, I thought of the words I could use, like How you doing? [sic] and Where my brother? [sic]. But then I didn’t call her. During the transitional intellectualization phase comprising audible private speech and silent self-directed speech, it is possible that a gradual decline in mediation through the L1 takes place. The following verbal report, for example, evidences that the learner has not yet formed a functional internal plane mediated by the L2 and that he still relies on his L1 to conduct inner speech: [The learner is talking about his preparation for an oral activity.] Sometimes words came to my mind in English, but most of the time I said them in Spanish [the L1], and little by little, I tried to translate them into English. (Guerrero, 2005, p. 167) With the progressive internalization of the L2, however, comes increasing use of the L2 as a tool for thought and less reliance on the L1 as a translation process. The following comment from an adult ESL learner, cited in Guerrero (2005), suggests that considerable internalization of the L2 has taken place and that the learner is beginning to think spontaneously through the L2: Most of my oral practice in English [the L2] is with friends, but I do find myself talking alone in English and thinking about things in English. It has become a useful practice and very normal and easy. (Guerrero, 2005, pp. 166–167) Though most studies of incipient L2 inner speech processes focus on adults, a similar behavior among adolescents was reported by de Courcy (1995): “Many students reported that French just takes over their minds and keeps ‘popping into’ their thoughts when they least expect it. Some keep up an almost continuous conversation with themselves in the second language” (p. 16).

< previous page

page_116

next page >

< previous page

page_117

next page >

Page 117 Intellectual Stage Although there seems to be no L2 “end point” or ultimate attainment that would signal a learner’s capacity to think through the L2, some developments must and do occur for L2 learners to be able to carry on sustained intellectual activity in the L2: first, sufficient internalization of the L2 (in terms of grammar, lexicon, phonology, and semantics), and second, conceptual change involving grammaticized and lexicalized concepts (Pavlenko, 1999, 2005). Reports, mostly from adult learners, show that inner speech in the L2 can be attained, if not to mediate verbal thought most of the time, at least partially in some contexts and situations, and within certain cognitive domains related to a person’s public and private life. In Guerrero (1999), 98% of 46 advanced (superior proficiency) ESL college learners reported having experienced inner speech in English. A study conducted by John-Steiner (1985) indicates that fully developed bilinguals, such as interpreters, undergo conceptual change in such a way that a “largely unified meaning system” (p. 366) emerges supporting flexible and varied verbal thinking in multiple language codes.9 Autobiographical narratives analyzed by Pavlenko (1998) and Pavlenko and Lantolf (2000) also show extensive (and often traumatic) conceptual reconstruction among late bilinguals (people who started learning an L2 as adults and eventually attained superior competence in the L2). For some of these late competent bilinguals, full attainment of an L2 represented loss of their L1 inner speech and emergence of a new self mediated by the L2. Larsen, Schrauf, Fromholt, and Rubin (2002) specifically explored the use of inner speech in remembering past events among a group of adult Polish immigrants in Denmark. Results showed that the language in which the memories were retrieved depended on the language in which the memories were encoded. Participants tended to use L1 inner speech to remember events encoded before arrival in Denmark and L2 inner speech to recall those that were encoded after arrival. According to the researchers, these individuals had “encoded the world of the homeland in one language and the ‘new world’ in another language” (p. 53). Further evidence that changes in conceptual structures may affect the nature of an L2 speaker’s inner speech is provided by research on the relationship between bilingualism and thought and on the use of L2 gestures. Research on the cross-linguistic nature of bilinguals’ and multilinguals’ thinking processes and conceptual organization is too vast to be treated here with any fairness, yet recent studies (see review in Pavlenko, 2005) suggest that L2 learning may indeed result in conceptual restructuring and thereby in new ways of understanding and expressing the world.10 Studies on gestures also provide insights into the inner speech processes of L2 learners. Based on the assumption that thought, language, and gesture are intertwined at the level of inner speech, McCafferty and Ahmed (2000) explored the appropriation of typically American gestures by adult Japanese-L1 learners. The

< previous page

page_117

next page >

< previous page

page_118

next page >

Page 118 researchers found that gestures were indeed appropriated by the learners who had learned the L2 naturalistically but not by those who had been classroom-instructed. The results suggest that when learners are immersed in the L2 culture it is possible for them to change the gestural component of their thinking processes. As Vygotsky (1986) pointed out, concept formation and word-meaning development are ongoing lifelong processes. Inner speech, or the capacity to think verbally, is also likely to be in constant evolution, not just because an individual’s conceptual and semantic structures in the L1 keep changing throughout life but also because they may be significantly affected by the radical introduction of a new cultural, psychological tool: an L2. Research shows that it is indeed possible to develop the ability to “think words” in an L2 for a variety of cognitive, regulatory functions. The ability to think through an L2 has been observed particularly among individuals who have had intense exposure and participation in L2, bilingual, or multilingual communities. The level of intellectualization that a learner can reach in the L2, however, cannot be predicted—as Vygotsky aptly noted when discussing mental development—by isolated criteria, such as onset age of acquisition or degree of attainment. A multiplicity of factors conforming the “social situation of development” (Vygotsky, 1998, p. 198) will determine whether an L2 learner is capable or willing to use the L2 instrumentally. An L2 learner’s social situation of development will be affected at any given moment by the presence and interaction of myriad types of influences—personal choice, level of proficiency, extent of conceptual restructuring, degree of acculturation, learning history, context of learning, task difficulty, linguistic domain, and others. Furthermore, it is evident that most learners will not develop the capacity to think in an L2 uniformly across all cognitive domains or communicative situations. In most cases, there may be interim or ultimate use of L2 inner speech for certain cognitive functions and not for others. Importantly, for most L2 learners, bilinguals, or multilinguals, the L2 will always remain an additional and alternative tool for thought tightly interwoven with the L1. Research Areas in Need of Attention SCT has come a long way in elucidating major turning points in the ontogenesis of L2 development: private speech use, emergence of inner speech, conceptual formation and restructuring, metalinguistic awareness and volitional control, among the most salient. In terms of age, SCT has shown that whereas ontogenetic development of L2 learning may start at any age past birth, certain critical periods in the L2 may first appear later than in the L1. Egocentric speech, for example, usually makes its appearance among preschool children in the L1, whereas private speech for internalization can first emerge among adolescents and adults in the L2 if they are beginning to

< previous page

page_118

next page >

< previous page

page_119

next page >

Page 119 learn the language in adolescence or adulthood. It is clear, nevertheless, that though the intellectualization of L2 speech functions such as private speech, inner speech, concept formation, and metalinguistic awareness can occur simultaneously and integratedly with their corresponding L1 functions (as in cases of parallel bilingual learning), critical turning points in L2 development cannot occur earlier than in L1 acquisition (e.g. genuine concept formation in the L2 could not take place before adolescence). Although there are plenty of SCT studies that focus crosssectionally on children, adolescents, and adults, SCT research has not systematically looked at the intellectualization of an L2 over the lifespan among single individuals who start learning the language in infancy or early childhood and reach age maturity as L2 speakers. This type of longitudinal study would be useful in showing differences in the ontogenesis of L2 verbal thinking, ages when critical periods in the L2 manifest themselves, and comparisons with L1 intellectualization processes. It is also necessary to look in depth at the various critical periods, including the preintellectual L2 phase, and scan for differences across different age groups. As mentioned, some variations have already been noticed between children and adults in the specific functions of L2 private speech. A major gap exists in the study of the development of L2 inner speech (beyond the private speech phase) among children of school age and how this process compares with that of L1 learners of the same age as well as with older and adult L2 learners. It would be interesting, for example, to find out how children and adults differ in their conscious, deliberate deployment of private and inner speech in the L2. Similarly, it would be worthwhile to conduct studies of concept formation in the L2 among adolescents and compare the impact of conceptual change on inner thinking processes among this age group and older L2 learners.11 The study of inner speech processes in the L2 among children would also have important implications for an understanding of how L2 literacy develops in this age group. Researchers (Flavell, Green, Flavell, & Grossman, 1997; Liva, Fijalkow, & Fijalkow, 1994) have already suggested that silent verbal tasks, such as reading, writing, and mathematical problem-solving, contribute to the development and awareness of L1 inner speech. In turn, inner speech contributes to further growth in areas of literacy. Finally, there is need to conduct research from an SCT perspective on the verbal thinking processes of older L2 learners, bilinguals, or multilinguals. How does old age and age-related problems affect these individuals’ higher intellectual functioning in the L1 and other languages? Luria (1973, 1981), a Vygotskyan associate and follower who did pioneering work in neurolinguistics and aphasic disorders, including cases of multilinguals, studied the adverse effects of brain lesions or disease on inner speech activity and their consequences in speech production and comprehension as well as on regulative behavior. Luria’s work suggests that cerebral lesions and disorders that

< previous page

page_119

next page >

< previous page

page_120

next page >

Page 120 correlate with age may affect inner speech activity and the intellectual functions that depend on inner speech. KotikFriedgut (2001), a disciple of Luria, believes that the study of aphasic disorders in bilinguals should take a systemic– dynamic approach, one that is based on Vygotskyan–Lurian theory and that requires consideration of systems of factors—rather than a single factor—to account for specific syndromes, including not only neurological aspects but also a complex interplay of variables and characteristics of language acquisition and use. A systemic–dynamic approach, bolstered by modern neuroimaging techniques, may prove fruitful in providing answers to the myriad questions that remain on how older bilinguals or multilinguals handle higher intellectual functioning in one or more languages. Educational Implications Although SCT does not deny or underestimate the possibility of developing the capacity to think through a new language in naturalistic settings (an area in much need of research), a great emphasis is placed in Vygotskyan theory on the impact of organized instruction on the development of higher mental functions, among these inner speech and metalinguistic control, as well as on complex cognitive skills, such as reading and writing. Vygotsky (1986), for example, specifically cited the benefits of learning an FL on the development of the native one. From a pedagogical perspective, therefore, there are enormous consequences to the introduction of a new tool of thought in academic settings to learners of all ages. What should teachers expect and how should teachers prepare in terms of differential behaviors related to age in the intellectualization of an L2? Although comparative research is still needed, it is obvious that both children and adults undergo a preintellectual stage as they begin learning the L2, a stage characterized by mostly routine social speech and reliance on the L1 as a meaning-making tool. As learners begin to internalize the L2, teachers may encounter differences in the type of private speech produced by children and adults. Whereas children may be more open to audibly verbalize their self-directed speech in the L2 and to experiment with language form, adults may feel more inhibited to do so, turning instead to silent private speech. At any rate, teachers should not interpret learners’ silence at any age as necessarily lack of participatory engagement. Teachers may also want to promote awareness and activation of private and inner speech processes in the L2 among students who are old enough to exercise conscious control over their mental functions. Deliberate deployment and practice of the L2 through private speech and silent self-talk may contribute to the learners’ processes of internalization and production. Finally, an important insight from SCT as applied to the age question in L2 learning is that concept formation and restructuring are lifelong processes that can be strongly affected by schooling. L2 instruction may play a powerful role in determining the nature of an L2 learner’s intellectual activity by

< previous page

page_120

next page >

< previous page

page_121

next page >

Page 121 fostering the development of new concepts in the L2 or in modifying the L1 conceptual base. Notes 1. Vygotsky (1986) was quick to point out, however, that thinking and speaking continue to be independent processes, even in adulthood. Thought and speech remain separate lines of development despite their merging in inner speech. Thought can be non-verbal, as may occur during practical activity and unplanned behavior, and speech can be non-intellectual, as when repeating something we do not understand (pp. 88–89). 2. See Vygotsky (1998) for a thorough treatment of psycholinguistic development in adolescence, “the transitional age,” as he called it. 3. Vygotsky (1986) identified three phases in the process of concept formation: a first stage of syncretic thinking characterized by unorganized conglomerates of objects in the child’s perception; a more organized second stage of thinking in complexes where objects are grouped by actual bonds between them; and a final stage of forming genuine concepts in which common elements among objects are singled out, abstracted, and synthesized (pp. 110– 113). 4. The term “metaconsciousness” is used here to capture the notion of consciousness that Vygotsky had in mind: “We use consciousness to denote awareness of the activity of the mind—the consciousness of being conscious” (1986, p. 170). 5. In addition to the internalizing function, L2 private speech has been found to serve as a means of gaining or regaining self-regulation during challenging cognitive tasks (see, for example, Centeno-Cortés & Jiménez, 2004; Frawley & Lantolf, 1985, McCafferty, 1992). For reasons of space, the self-regulatory function of L2 private speech will not be discussed here, but an important treatment of studies is offered by Lantolf and Thorne (2006). 6. It is of course possible, however, that the learners who did not display any private speech during the taping sessions may have indeed engaged in L2 private speech when not being taped. 7. The absence of an overt private-speech-for-internalization phase among some learners is not surprising. As Lantolf (2007) points out, from a Vygotskyan perspective, cognitive development is a revolutionary process which may not always follow a predictable, stable trajectory. Variability in private speech occurrence among L2 learners could also be explained in terms of Vygotsky’s (1998) concept of “the social situation of development,” that is, the complex and unique relationship between the learner and the environment present at any given moment in the developmental process, a situation which is responsible for different dynamics of change and which may account for different developmental paths in the internalization of an L2. 8. Verbal reports in this study were 45% of type (a), concurrent processing of language, 38% of type (b), recall of language, 13% of type (c), preparation before writing or speaking, and 4% of type (d), silent verbalization of thoughts for private purposes. Not surprisingly, (d) was the least frequent type, indicating that these beginning students were still very far from making the L2 the medium of their inner speech. 9. According to John-Steiner, Meehan, and Mahn (1998), a Vygotskyan based “functional systems” approach provides powerful explanation for the integration of dynamic processes of conceptual development among bilinguals. 10. L2 learners may undergo conceptual change in various forms. Pavlenko (2005)

< previous page

page_121

next page >

< previous page

page_122

next page >

Page 122 lists seven different ways in which learning of an L2 may affect an individual’s conceptual store: (1) coexistence of L1 and L2 conceptual domains, (2) L1-based conceptual transfer, (3) internalization of new concepts, (4) shift from L1 to L2 conceptual domain, (5) convergence of L1 and L2 conceptual domains, (6) restructuring of a conceptual domain, and (7) attrition of previously learned concepts (p. 438). 11. Most studies on the conceptual bases of an L2 have been conducted among adults. A study by Collison (1974), however, applied Vygotsky’s theory of conceptual development to compare the concept formation processes of instructed Ghanaian children aged 12–14 in the L1 and the L2. Collison found that learners functioned at higher conceptual levels in the L1 than in the L2. References Broner, M., & Tarone, E. (2001). Is it fun? Language play in a fifth-grade Spanish immersion classroom. Modern Language Journal, 85, 363–379. Centeno-Cortés, B. (2003). Private speech in the second language classroom: Its role in internalization and its link to social production . Unpublished doctoral dissertation, Pennsylvania State University. Centeno-Cortés, B., & Jiménez, A. (2004). Problem-solving tasks in a foreign language: The importance of the L1 in private verbal thinking. International Journal of Applied Linguistics , 14, 7–35. Chen, R. (1988). The private speech of a Chinese–English bilingual child: A naturalistic longitudinal study (Doctoral dissertation, University of Illinois at Urbana-Champaign, 1997). Dissertation Abstracts International-A , 49(01), 60. Collison, G. O. (1974). Concept formation in a second language: A study of Ghanaian school children. Harvard Educational Review, 44, 441–457. de Courcy, M. (1995, March). Language learning experiences of Australian French immersion students . Paper presented at the Annual Meeting of the American Association for Applied Linguistics, Long Beach, CA (ERIC Document Reproduction Service No. ED 388 096). Flavell, J. H., Green, F. L., Flavell, E. R., & Grossman, J. B. (1997). The development of children’s knowledge about inner speech. Child Development , 68, 39–47. Frawley, W., & Lantolf, J. P. (1985). Second language discourse: A Vygotskyan perspective. Applied Linguistics , 6 , 19–44. Galperin, P. Y. (1967). On the notion of internalization. Soviet Psychology, 5 (3), 28–33. Guerrero, M. C. M. de. (1994). Form and functions of inner speech in adult second language learning. In J. P. Lantolf & G. Appel (Eds.), Vygotskian approaches to second language research (pp. 83–115). Norwood, NJ: Ablex. Guerrero, M. C. M. de. (1999). Inner speech as mental rehearsal: The case of advanced L2 learners. Issues in Applied Linguistics , 10, 27–55. Guerrero, M. C. M. de. (2004). Early stages of L2 inner speech development: What verbal reports suggest. International Journal of Applied Linguistics , 14, 90–112. Guerrero, M. C. M. de. (2005). Inner speech—L2. Thinking words in a second language . New York: Springer. Hakuta, K. (1986). Mirror of language. The debate on bilingualism . New York: Basic Books.

< previous page

page_122

next page >

< previous page

page_123

next page >

Page 123 John-Steiner, V. (1985). The road to competence in an alien land. In J. V. Wertsch (Ed.), Culture, communication, and cognition: Vygotskian perspectives (pp. 348–371). New York: Cambridge University Press. John-Steiner, V., Meehan, T. M., & Mahn, H. (1998). A functional systems approach to concept development. Mind, Culture, and Activity, 5 (2), 127–134. Karpov, Y. V. (2003). Development through the lifespan: A neo-Vygotskian approach. In A. Kozulin, B. Gindis, V. S. Ageyev, & S. M. Miller (Eds.), Vygotsky’s educational theory in cultural context (pp. 138–155). Cambridge, UK: Cambridge University Press. Kecskes, I., & Papp, T. (2000). Foreign language and mother tongue . Mahwah, NJ: Erlbaum. Kotik-Friedgut, B. (2001). A systemic–dynamic Lurian approach to aphasia in bilingual speakers. Communication Disorders Quarterly, 22(2), 100–109. Lantolf, J. P. (2003). Intrapersonal communication and internalization in the second language classroom. In A. Kozulin, B. Gindis, V. S. Ageyev, & S. M. Miller (Eds.), Vygotsky’s educational theory in cultural context (pp. 349– 370). Cambridge, UK: Cambridge University Press. Lantolf, J. P. (2007). Sociocultural source of thinking and its relevance for second language acquisition. Bilingualism: Language and Cognition, 10, 31–33. Lantolf, J. P., & Thorne, S. L. (2006). Sociocultural theory and the genesis of second language development. New York: Oxford University Press. Lantolf, J. P., & Yáñez, M. C. (2003). Talking yourself into Spanish: Private speech and second language learning. Hispania , 86, 97–109. Larsen, S. F., Schrauf, R. W., Fromholt, P., & Rubin, D. C. (2002). Inner speech and bilingual autobiographical memory: A Polish–Danish cross-cultural study. Memory , 10, 45–54. Leontiev, A. A. (1981). Psychology and the language learning process . New York: Pergamon. Leopold, W. F. (1949). Speech development of a bilingual child. A linguist’s record. Vol. 4, Diary from age 2 . New York: Northwestern University Press. Liva, A., Fijalkow, E., & Fijalkow, J. (1994). Learning to use inner speech for improving reading and writing of poor readers. European Journal of Psychology of Education , 9 , 321–330. Luria, A. R. (1973). The working brain. An introduction to neuropsychology (B. High, Trans.). New York: Basic Books. Luria, A. R. (1981). Language and cognition (J. V. Wertsch, Ed.). New York: John Wiley. McCafferty, S. G. (1992). The use of private speech by adult second language learners: A cross-cultural study. Modern Language Journal, 76, 179–189. McCafferty, S. G., & Ahmed, M. K. (2000). The appropriation of gestures of the abstract by L2 learners. In J. P. Lantolf (Ed.), Sociocultural theory and second language learning (pp. 199–218). New York: Oxford University Press. Mahn, H. (2003). Periods in child development: Vygotsky’s perspective. In A. Kozulin, B. Gindis, V. S. Ageyev, & S. M. Miller (Eds.), Vygotsky’s educational theory in cultural context (pp. 119–137). Cambridge, UK: Cambridge University Press. Ohta, A. S. (2001). Second language acquisition processes in the classroom. Learning Japanese . Mahwah, NJ: Erlbaum.

< previous page

page_123

next page >

< previous page

page_124

next page >

Page 124 Pavlenko, A. (1998). Second language learning by adults: Testimonies of bilingual writers. Issues in Applied Linguistics , 9 , 3–19. Pavlenko, A. (1999). New approaches to concepts in bilingual memory. Bilingualism: Language and Cognition, 2 , 209–230. Pavlenko, A. (2005). Bilingualism and thought. In J. F. Kroll & A. M. B. de Groot (Eds.), Handbook of bilingualism. Psycholinguistic approaches (pp. 433–453). New York: Oxford University Press. Pavlenko, A., & Lantolf, J. P. (2000). Second language learning as participation and the (re)construction of selves. In J. P. Lantolf (Ed.), Sociocultural theory and second language learning (pp. 155–177). New York: Oxford University Press. Saville-Troike, M. (1988). Private speech: Evidence for second language learning strategies during the “silent” period. Journal of Child Language, 15, 567–590. Tomasello, M. (2003). Constructing a language: A usage-based theory of language acquisition. Cambridge, MA: Harvard University Press. Vygotsky, L. S. (1978). Mind in society. The development of higher psychological processes. Cambridge, MA: Harvard University Press. Vygotsky, L. S. (1986). Thought and language. Cambridge, MA: MIT Press. Vygotsky, L. S. (1998). The collected works of L. S. Vygotsky. Vol. 5. Child psychology (R. W. Rieber, Ed.). New York: Plenum. Wertsch, J. (1985). Vygotsky and the social formation of mind . Cambridge, MA: Harvard University Press.

< previous page

page_124

next page >

< previous page

page_125

next page >

Page 125 5 A DYNAMIC VIEW OF SECOND LANGUAGE DEVELOPMENT ACROSS THE LIFESPAN Wander Lowie, Marjolijn Verspoor, and Kees de Bot1 While in previous chapters the focus was first language development (Behrens, van Geert) and first and second language development (Unsworth), this contribution will focus only on second language development (SLD) over the lifespan. In our view, second language acquisition is a process that is clearly related to the development of the first language. Following earlier suggestions by Cook (1995) and Herdina and Jessner (2002), we take the position that the language system is an integrated system consisting of languages, dialects, styles and registers that all interact. SLD is basically the development of the multilingual system over time. This view also implies that language acquisition and language attrition are manifestations of similar mechanisms of change, and the multilingual system can develop in many different ways, not only through the acquisition of new languages or dialects but also through the decline of language skills due to non-use or injury. In line with Behrens’s and van Geert’s contributions, our view on SLD is usage-based. To explain this further, changes in the system are the result of changes in language use, and the other way around: use is change and change is use (Larsen-Freeman, 1997). After briefly discussing a number of background theoretical assumptions, we will give an overview of variables that may affect SLD and then relate them to development across the lifespan. A Dynamic View of SLD Before presenting an overview of factors affecting SLD, we will clarify our theoretical points of departure. Dynamic systems theory is a theory of change. The most important characteristics of dynamic systems are that they are selforganizing sets of interacting variables which, depending on initial conditions and iterative growth, show patterns of unpredictable nonlinear

< previous page

page_125

next page >

< previous page

page_126

next page >

Page 126 and discontinuous development. Typical examples of dynamic systems are the weather, the behavior of a flock of birds, the shaping of sand dunes and the development of populations. All of these systems are characterized by numerous states that evolve over time. Even though each subsequent state may be predicted based on previous states, the eventual outcome is not predictable, as the development is chaotic rather than linear because of the dynamic interaction of factors affecting the system. In the mid-1990s several articles were published arguing that human cognition can be seen as dynamical systems, the most influential of which was Port and van Gelder (1995). Similar to other types of cognitive development, (first) language development can also be accounted for in terms of dynamic systems. This position was convincingly taken by, among others, van Geert (1995), who explains language development in terms of mathematical growth models. Various publications have recently appeared that argue that also second language learning can be seen as the dynamic development of a system and that findings from dynamic systems theory (DST) apply to second language development (de Bot, Lowie, & Verspoor, 2007; Herdina & Jessner, 2002; Larsen-Freeman, 1997; Larsen-Freeman & Cameron, 2008).2 These articles show that the development of language is not linear but chaotic, that the eventual attainment cannot be predicted, and that language development is dependent on the initial condition and shaped by a wide range of interconnected factors. The crucial point is that not only do these factors interact in shaping the system, but that they do so in a dynamic way. All factors change over time and affect the system differently at different moments in time. In DST terms, these factors are referred to as “resources” of the system. One of the great merits of this approach is that no separation for different types of resources is required. In this way it can be argued that the system is not only affected by external factors (e.g. the learner’s environment, the input, the feedback), but also by internal factors (e.g. the learner’s motivation, aptitude, previous language knowledge) and even by the development of the system itself. In this section, we will point out what these theoretical assumptions are for language development in general and SLD in particular. Language systems are complex sets of interacting variables at many different levels and sub-levels. Examples of levels are cultural, social, psychological and linguistic. Within each of these levels there are again many different sublevels. For instance, within the linguistic sub-systems there is the sound system, the lexicon, the grammar, and so on. These systems and their subsystems are interconnected. The linguistic sub-systems interact over time with other internal systems such as the individual’s cognitive and affective system and external systems, like the individual’s cultural, social and physical environment. When one variable changes in one sub-system, this change will affect not only the other variables in that subsystem, but also the ones in the related sub-systems and consequently the system as a whole. For example, we can

< previous page

page_126

next page >

< previous page

page_127

next page >

Page 127 consider an individual’s language system as one complex system, but as individuals interact with other speakers there is also a more general language system that is shared by all the individuals in a language community. When one individual starts to use a new word or expression, it may be taken over by the other language users. As a result, the more general system and the other individuals who constitute that system are also affected. Clearly, this is not to say that one minor change in one of the sub-systems will always have a huge effect on all other sub-systems. One speaker may introduce a new word or expression, but others may not start using it and the new word or expression disappears from the system. The point we are trying to make is that major changes in a whole system may have been initiated by one minor change in a single sub-system. These changes may take place at all different levels of language. The example we have used here concerns language communities, but the same developments can be found within the interacting sub-systems of an individual’s language system. For example, a fluent speaker may lose much of his or her language system because of one minor change in the neurological system. The effect of one factor on another in the same or in a different subsystem is very difficult to predict. The prediction of the outcome of development is particularly complicated because of the constant changes in the interaction of all variables. Yet, in spite of its chaotic nature, the development of language systems is not completely random. As they develop over time, sub-systems tend to move towards preferred states, so-called “attractor states.” Once a system has reached an attractor state, much additional energy is needed to achieve further development. N. Ellis (2008) argues that the “Basic Variety,” a variety of a second language that has been found to emerge in untutored SLD, is a good example of an attractor state. Larsen-Freeman (2005), however, questions the existence of “stable patterns of non-nativeness” (p. 194) and raises the question of whether it is justified to make a difference between fossilization and native-like patterns. From a DST perspective, no principled distinction is required; some attractor states can be characterized as “fossilization,” others as achieving target-like structure. When all sub-systems simultaneously end up in an attractor state this may result in a relatively stable system. However, this will hardly ever be the case. There is sufficient evidence that SLD goes in leaps and bounds with periods of stability and instability and identifiable stages in the non-linear developmental pattern. During the process of language acquisition, periods of rapid acquisition are followed by periods of delayed acquisition or even attrition. Language systems usually point to a high degree of variation right before the language system changes more drastically. In addition, even when they are in a stable period, language systems will always show some low-level degree of variation. Variation is functional because it allows the system to adapt to changing conditions. In this way, it is variation that enables evolution. Once language users have become adults, they are not likely to change their

< previous page

page_127

next page >

< previous page

page_128

next page >

Page 128 language use very drastically and will use a limited repertoire of expressions and constructions over and over again, even though there may be variation on a smaller timescale. Yet, the individual’s language system may change dramatically when there is a major event to change it, such as a move to another country or enrolment in a language course. Development is non-linear; in other words, developmental patterns at initial stages of a sub-system will be very different from patterns of development of that same sub-system at later stages. This also implies that the development of a language system may be highly dependent on its initial state, and minor differences at the beginning can have dramatic consequences in the long run. For example, if the first language (L1) of a learner is typologically very different from the second language (L2), it will be more difficult to acquire the L2 than when the L1 and L2 are very similar. Finally, a term that is commonly used in a DST framework is “resources.” Resources are all the available internal and external factors that enable the development of a dynamic system. Language development is dependent on external resources such as instruction, interaction with other speakers, and use of the language. Language development is also dependent on internal resources, like neurological and physiological factors. During development, the availability of resources, cognitive, affective, environmental, and sociocultural, will change too. Once we appreciate the dynamics of language development in general, we can also accept the fact that not all humans will show the same behavior even under seemingly similar circumstances. It is simply not likely that initial conditions, external resources (like cultural and social environment) and internal resources (like intelligence and aptitude) are identical in all respects for different individual language learners, especially since both the resources and the interaction of subsystems constantly change over time. Interacting Variables in Second Language Development In this section we will discuss the different variables that may play a role in SLD. We will move from the larger system at the social and cultural level in the community to the more deeply nested ones such as the individual’s psychological, physical and cognitive systems. In addition, we will consider the influence of other languages on the multilingual system. Subsequently, we will discuss the linguistic system and its sub-systems. Finally, we will examine how these variables may develop across the lifespan, which leads us to one of the most debated issues in SLD: age. Social and Cultural Variables An important distinction that is usually made in SLD research is whether the L2 is instructed or not. Of course, this is not a neat dichotomy but a con-

< previous page

page_128

next page >

< previous page

page_129

next page >

Page 129 tinuum. On the one hand, there are learners who only learn the language by following a language teaching program, in which everything is controlled and the learner only has to do what he or she is told. This was very much the norm in behaviorism-based pattern drills for which there was no room for creative additional language input on the part of the learner, who was basically seen as a mechanical learning machine. On the other hand, non-instructed learners acquire language through using it outside of a classroom. However, a non-instructed setting does not mean that there is no instruction at all. It is quite likely that some of the native speakers will adapt their language to a nonnative speaker and sometimes correct errors, but in such a setting instruction is not the systematic approach to language learning. Within instructed programs, the focus over the last 20 years or so has shifted to communicative language teaching (CLT), which is more similar to the non-instructed part of the continuum than the instructed one. There is also a trend towards autonomous learning in which the structuring of educational materials is minimal and the learner is supposed to find his or her own way using different sorts of resources and materials according to needs. The most recent insights put forward that a combination of communicative approaches and formal instruction seems to be the most fruitful. Therefore, most language learners seem to learn a foreign or second language in a way that is somewhere in the middle part of the continuum, and as the history of foreign language teaching has shown, language teaching has gone through various cycles going from more to less instructed and back over the ages (Wilhelm, 2005). In this context, the distinction between foreign and second language learning is also relevant. Again we are dealing with a continuum rather than a binary distinction. Foreign language learning takes place in a setting in which the language to be learnt is not the language of the larger community: for example, learning Spanish in Swedish secondary education. However, the foreign language may also be used as the language for instruction in school. For example, in many African countries, English or French is used in school rather than the children’s native language. In some cases the school language is phased in slowly; in other cases, the foreign school language is used from the first day onwards. In addition, the adequacy of such language instruction may vary. On the one hand, some schools may have huge classes, few facilities and materials and teachers who have not acquired the L2 sufficiently to talk about more complex school subjects. On the other hand, some other schools may have small classes, adequate facilities and good teachers who are not only fluent in the L2 but also experienced in helping the children acquire the new school language at the same time as they learn about the new subjects. The latter type of L2 teaching may occur in countries such as Luxembourg, where children speak Luxemburgish at home and learn to speak German and French at school without any problems (Davis, 1994).

< previous page

page_129

next page >

< previous page

page_130

next page >

Page 130 Second language learning occurs when the L2 learner learns the L2 in an L2 environment. For example, when entering school, Turkish children living in Germany who speak Turkish at home will be exposed to German, and Spanish-speaking children in the US who speak Spanish at home will be exposed to English. In addition, a large number of refugee children in many different countries have to deal with a new language at school. In all these cases, the home language is very different from the school language. This situation often gives the child a disadvantage compared to his or her monolingual classmates because the child has to deal not only with a totally new environment and culture at school but also with a new language system that he or she does not understand at all. Moreover, many studies have shown that these children are delayed in their school achievement and that they hardly ever catch up with their monolingual peers. A factor that plays a crucial role irrespective of the distinctions described above is the amount and type of contact with the L2-speaking community, that is, the amount of input and interaction the learner has with the L2. Superficially, it seems that input is more limited for foreign language learners than for second language learners. However, second language learning does not always guarantee a large amount of input. Some second language learners go to school or work with speakers of the L2, and thus receive a rich language input, which is likely to help develop language proficiency. For other learners the input is limited to “basic interpersonal communication” to use Cummins’s terms (Cummins 1977, 1993). Affective Variables Even though social and cultural variables play an important role, they do not on their own determine the course of SLD. They always interact with individual variables called social-psychological factors such as motivation, attitudes, aptitude, anxiety, beliefs about language learning (R. Ellis, 1994). There is some consensus in the literature that motivation is important for language learning. In the literature on motivation a distinction is made between “orientation,” which refers to the long-term goals for an activity, and “motivational intensity,” that is, the willingness of the learner to invest time and energy in learning the language (e.g. Dörnyei, 2001). A range of factors is likely to come into play in shaping and adapting motivation. Few learners will be motivated to learn a language because of its sheer beauty. They will typically have other motives, like reading literature in the original language, learning about the culture, studying abroad, and enhancing job opportunities. An important characteristic of motivation is that it is not a fixed factor, but an adaptable and interconnected one. For example, a strong motivation will lead to learning success, but learning success may also lead to a stronger motivation. As we will demonstrate in our discussion on lifespan development of SLD, motivation can

< previous page

page_130

next page >

< previous page

page_131

next page >

Page 131 hardly be seen as a constant, but will be dynamically variable3 at different timescales. Attitude is closely related to motivation. As Giles and Billings (2004, p. 193) indicate, language attitudes play an important role in social decision-making. Learners can have a positive or negative attitude towards a variety of things: languages, speakers of that language, the associated culture, but also towards teaching methods and materials. Compared to motivation, attitudes tend to be fairly stable. Even though a positive attitude usually helps in L2 acquisition, it is not a prerequisite. Students may have a negative attitude towards a language and still be motivated to learn it. Motivation is also related to learner beliefs. Learner beliefs refer to the beliefs of learners about language learning, such as their beliefs about their own language aptitude, about the difficulty of a language they want to learn, and about the best way to learn a language. Learners tend to be less motivated to learn when the teaching approach they are confronted with deviates from their beliefs about the best way to learn a language. For example, many learners believe explicit grammar learning is valuable, but in many countries there is a trend moving away from teaching grammar explicitly. Learner beliefs are often deeply rooted in culture: in societies in which teachers represent authority, beliefs will tend to support teacher-centered and conformity-oriented approaches, while in less authority-oriented cultures individual initiative and independent thinking are valued more. It is interesting to note that the relation between stated beliefs about language learning and L2 achievement is rather weak (R. Ellis, 1994, p. 544). Another affective variable is language anxiety, which refers to the anxiety learners experience when they have to use a language. Anxiety can be a personality trait, but it can also be related to the situation, for example when a learner has to respond in a foreign language in class (Horwitz, Horwitz, & Cope, 1986). Language anxiety is negatively related to achievement in the foreign language. According to MacIntyre and Gardner (1994), anxiety affects learning at different levels: at the input level when it blocks the perception of foreign sounds and units, at the storage level when language knowledge has to be stored in memory and at the level of retrieval when information on the language has to be retrieved from memory. As we will discuss below, there is also a relationship between anxiety and language aptitude. For a number of other social psychological factors “personality” is often used as an umbrella term and includes anxiety as a trait, risk-taking, empathy, self-esteem, and self-control. The trait that has been studied most extensively is the introvert/extravert distinction. As R. Ellis (1994, p. 541) points out, the correlations between personality traits and proficiency in the second language tend to be low; however, extraverts tend to have a higher fluency but a lower accuracy in their L2, while the opposite was found for introverts.

< previous page

page_131

next page >

< previous page

page_132

next page >

Page 132 Cognitive Variables Whereas affective variables are related to feelings and emotion, cognitive variables are related to language aptitude, intelligence, working memory, and processing speed. Language aptitude tests measure to what extent learners have a talent to learn a foreign language. Starting with the work by Pimsleur and Carroll in the 1960s (see for an overview Pimsleur, 1966; Pimsleur, 1968), a series of tests have been developed to measure language aptitude. Based on this research, a number of aspects have been found to show a high correlation with success in language learning: phonemic encoding ability, grammatical sensitivity, inductive learning ability, and rote learning ability. A recent language aptitude test developed at the University of Swansea (LLAMA language aptitude tests, Meara, 2004, consists of a test for vocabulary learning, a test for grammatical inferencing and a test of sound/symbol associations.4 Research on language aptitude by Sparks and his colleagues (Sparks, Ganschow, & Javorsky, 1992; Sparks, Ganschow, & Patton, 1995) has shown that there may be one particular factor that may explain a lack of learning success in many of the weak language learners. They have posited the Linguistic Deficit Coding Hypothesis,5 which is related to the first aptitude measure mentioned: phonemic encoding ability. According to the Linguistic Deficit Coding Hypothesis, some learners have problems coding language in general, which pertains not only to their second/foreign languages but also to their first language. This underlying deficit has a major impact on the learning process because perception of sounds and strings in the second language are the basis for (implicit) learning. It serves as a good example of how initial conditions can be crucial in SLD (see de Bot et al., 2007). In addition, aptitude has an impact on the social psychological and affective factors mentioned earlier: learners’ beliefs and their motivation will be affected heavily by the inability to even perceive relevant distinctions on the phonological and syntactic level. Language aptitude seems to relate to a component of general intelligence, but as far as there is research on this, the outcomes suggest that intelligence in its traditional definition is more related to the inductive learning ability part of aptitude than to other traits. Because memory is considered as one of the main traits of language aptitude, it comes as no surprise that working memory capacity is a relevant factor in SLD. The ability to keep an item longer in working memory considerably enhances the chance that it is being processed and stored in memory. Miyake and Friedman (1999) report high correlations between measures of working memory and language aptitude. They also show that larger working memory capacity is related to better syntactic comprehension. Although this tends to be interpreted as a onedirectional relationship, we suggest that it is not inconceivable that it is actually bidirectional, since better syntactic comprehension may allow for more efficient coding and storage.

< previous page

page_132

next page >

< previous page

page_133

next page >

Page 133 Another cognitive variable related to language is the time needed to process information in a language. There is quite some evidence that speed of language processing is lower in an L2 or L3 than in the L1 and that it is related to the level of proficiency in a language as measured in tasks like lexical decision and picture naming (N. Ellis, 2002; Mägiste, 1986; Woutersen, 1997). Moreover, Hulsen, de Bot, and Weltens (2002) show that in language attrition not only recognition scores but also reaction times go down. Even though there is a relation between speed of processing and SLD, it is not clear whether being slower affects acquisition negatively or being quick enhances acquisition. There is some evidence that, with old age, decline in processing speed may affect language production and perception (Pichora-Fuller, 2003; See & Ryan, 1995), but it is not clear what causes differences in processing speed. Is it simply a reflection of the fact that elements are used less frequently, or does adding another language to the language system in itself lead to a slowing down of the system? Mägiste’s (1986) work seems to support the latter view: she found that multilinguals when compared to monolinguals are also slower in their first language. This suggests that the total amount of resources may be limited and that having to process more than one language goes at the expense of speed for all languages. Age-related Variables In the previous section we discussed all kinds of variables that may affect SLD, but what we have not discussed yet is how these variables may interact with age. Considering the lifespan focus of this book, we will elaborate on this issue in more detail. There is quite some literature about early bilingualism, about early foreign language learning and immersion, about the development of language skills in the school years and in tertiary education, but after these phases we see a big gap in the descriptions. The problem is that we cannot really separate the variables of age and education or kinds of interaction that may take place; therefore, we are faced with a surprising lack of knowledge about what happens to our language knowledge, in particular our second language knowledge, after obvious moments of increased development. (See also Schmid’s contribution about attrition in this volume.) In research on language and aging, three sets of interacting variables appear to play a role in change over time: physical changes, psychological changes, and social changes (de Bot & Makoni, 2005). Physical changes refer to changes in metabolism, the functioning of the central nervous system, the locomotive system and all the other subsystems in the human body. Some of these changes may be largely irrelevant for SLD, but others may have a direct or indirect effect. Changes in vision and hearing are caused partly by changes at the neuronal level. Even though these lower-order changes are in themselves irrelevant for SLD, a change in hearing is

< previous page

page_133

next page >

< previous page

page_134

next page >

Page 134 \ relevant because it changes patterns of interaction and the possibilities to have input in the foreign language. Physical changes, psychological changes, and social changes are closely related and sometimes cannot really be distinguished. For instance, a decline in hearing capacity or speed of processing can be looked at as a purely physical change, but it has profound effects on cognitive functioning and may have important psychological and social consequences. Psychological factors will not change on their own, but are typically dependent on physical and social factors; moreover, psychological factors will have an impact on physical and social factors. Social changes like a new working environment, the death of a loved one, marriage or moving to another city can directly affect language development, but may also affect cognitive and psychological functioning. So although it may be useful for practical purposes to distinguish between physical, psychological and social factors, we should be aware of their interaction. The interaction between physical, psychological and social changes may change in the course of the lifespan. At a very young age and in old age physical changes will be more pronounced and influential than in adolescence and adulthood, whereas social factors play a more significant role in the latter two phases. In childhood, physical changes partly define first language learning: hearing and vision develop to allow for perception and production of sounds and longer units (see van Geert’s chapter on L1 in this volume). In old age changes in resources, such as memory and perception, lead to changes in language use and language skills (see Kemper in this volume). These examples illustrate that not only do physical, social and psychological factors interact, but that this interaction is dynamical and changes across the lifespan. In the literature about second language development, the age effect is referred to as the observation that young starters tend to be more successful in language learning than people who start learning an L2 relatively late. Traditionally, the age effect has been accounted for in terms of physical changes in the brain (for an overview, see Birdsong, 2006). Apart from attempts to explain the “age effect” from a physiological point of view, it has also been suggested that the causes must be sought in the changing sociocultural context or are due to affective factors. Motivation and contact with the language appear to be a major factor (Bongaerts, van Summeren, Planken, & Schils, 1997). Considering the dynamic interrelatedness of all factors, from social-cultural to maturational, it is not surprising that as yet not one factor has been singled out as the ultimate cause of the age effect. Here, too, a complex dynamic interaction of several factors prevents us from making valid generalizations about what in French is called le troisième âge. A hypothesis that is often mentioned in relation to the age effect for second language learning is the Critical Period Hypothesis. The Critical Period Hypothesis (CPH) states that it is not possible for learners to acquire a native-like level of proficiency of the L2 when learning the second lan-

< previous page

page_134

next page >

< previous page

page_135

next page >

Page 135 guage starts after a critical period, often associated with puberty. The discussion about the CPH has been going on since the 1960s and is still lively. On the one hand, it has been repeatedly shown that young starters ultimately tend to attain a higher level of proficiency, even though initially their rate of acquisition may be slower. On the other hand, several studies have shown that it is definitely not impossible for older starters to attain a native-like level of language proficiency. The evidence that younger learners are usually more successful L2 learners than older ones is abundant. For example, Johnson and Newport (1989) tested 46 native speakers of Chinese and Korean who had lived in the United States for at least 5 years on their ability to judge whether English sentences were grammatical or not. They found a very high correlation between score and age for young arrivals (age 6 and age 17), but not for the older ones. Agerelated effects were also found for French learners of English (Coppetiers, 1987) and for English and French learners of Italian (Sorace, 1993). These studies convincingly show that young starters perform better in L2 than late starters. However, this does not mean that adult learners cannot acquire an L2 effectively. For example, Bongaerts, Mennen, and Van Der Slik (2000) show that in some cases late starters can attain native-like proficiency. The point they make is that if there is a critical period, at least some people are not susceptible to it. Another argument against the CPH is that the end point of that period cannot be indicated. After a study of data on 2.3 million US immigrants, Hakuta, Bialystok, and Wiley (2003) conclude that while early starters do show higher second language attainment, there was not a sudden age at which learners no longer could acquire the second language: “the pattern of decline in second-language acquisition failed to produce the discontinuity that is an essential hallmark of a critical period” (p. 31). It is now generally accepted that the majority of post-puberty learners will not reach the native-like level (see Birdsong 2006 and Hopp 2007 for overviews); in addition, the focus of discussion is now more on what factors prevent them from reaching that level. It has been repeatedly suggested that the relative difficulty for older starters is due to physiological changes in the brain. Birdsong (2006) shows that there is a clear negative correlation between age of onset and the number of correct items on a grammaticality judgment task. This effect does not seem to stop shortly after the end of puberty, as argued by Johnson and Newport (1989), but is persistent throughout the age span. General patterns of cognitive aging have shown to appear at young adulthood (Birdsong, 2006, p. 28) and continues across the adult lifespan. One of the possible factors of decline, Birdsong reports, is the decrease of the brain volume, which already starts at age 20. Other factors include irreversible loss of plasticity of the brain or other neural modifications (for a recent overview, see Pallier, 2007). However, maturational changes of the brain that have been reported do not consistently coincide

< previous page

page_135

next page >

< previous page

page_136

next page >

Page 136 with what is usually regarded as the end of the critical period (i.e. the onset of puberty) and the evidence from this field is highly complex and far from conclusive. Even though there is much less research done on older learners, it is a common assumption that at one point people are too old to learn an L2. It has been suggested that older learners (after about 65) do not achieve high levels of L2 proficiency because of changes in brain structure, but now this idea has been modified in that it recognizes that brain structures result from experience and do not necessarily constrain development (Hagoort, 2006). Moreover, older L2 learners may do well in some sub-systems of the language and poorly in others. Older L2 learners often reach a very high level in vocabulary and grammar, but not in the L2 phonology, and keep speaking with a very distinct foreign accent. Therefore, not surprisingly, one of the explanations for age-related differences comes from the area of phonology. Flege (Flege, 1987; Flege, Yeni-Komshian, & Liu, 1999) attributes the general inadequacy of late starters’ L2 pronunciation to their perceptual capabilities. When children learn the sounds of their first language, they perceive the sounds in what Wode (1994) labels the “continuous mode”: all sounds are perceived and qualified equally. However, once children have established a linguistic sound system, they start categorizing the speech sounds they hear in terms of the sounds they already know (“categorical perception”). From that moment (around age 7), all L2 sounds that are similar to L1 sounds will be categorized as L1 sounds, so no new categories are created for “similar” sounds. A new phonological category will only be created for sounds that cannot be classified in terms of L1 sounds. This would account for the observation that L2 sounds that are phonetically similar to L1 sounds are the most difficult ones to attain. However, if the fixation of perceptual categories were indeed the reason for the existence of a critical period, again the question can be raised how it is possible that some “exceptionally good” learners are able to overcome this. From a DST perspective the explanation would be that even though adults will have more entrenched patterns (“attractor states”), with enough attention, motivation and practice these patterns can be changed. Cross-linguistic Variables The DST explanation for age effects in pronunciation suggests that depending on the stage in life when second language development starts, the system of the first language may be more or less established, which in turn would imply that especially for learners with a “mature” first language system, the development of a second language system is strongly dependent on the first one. The areas of difficulty in the second language could then be predicted based on a careful study of the differences between the two languages. This approach was popular in the 1960s, and is referred to as the Contrastive Analysis Hypothesis.

< previous page

page_136

next page >

< previous page

page_137

next page >

Page 137 Unfortunately, it did not turn out to be as simple as that. Given, it is true that languages that are typologically very different (for instance English and Chinese) tend to be more difficult to learn than languages that belong to the same language family (like English and Danish). For example, the US Foreign Service Institute (1985, discussed in Odlin, 1989, p. 39) has estimated how much time it takes to reach a given level of proficiency for a list of foreign languages, taking English as L1. With 30 hours a week in class, it takes the average learner 20 weeks to reach that level for Spanish, 24 weeks to do the same for Dutch and 44 weeks for a wide range of languages including Chinese, Greek, and Urdu. Typological distance from English seems to be the main factor for these differences. However, the “different = difficult” assumption cannot be generalized to all levels and settings of second language development. Several researchers have shown that areas where languages are similar but not identical may turn out to be more difficult than areas where the languages are completely different. For instance, Flege (1987) has shown that L2 sounds that are similar to L1 but not identical were more difficult to learn than sounds that are completely different, because learners think that the similar sounds are identical. A comparable effect is described by Kellerman (1986), who showed that the extent to which learners think that words in L2 are similar to the L1 determines their willingness to transfer these words. The amount of influence of one language on the other, or cross-linguistic influence , is dynamically dependent on a range of interacting factors. Apart from the distance between languages and the stage of development of the first language mentioned above, cross-linguistic influence depends on the amount of use of the languages, the learner’s level of proficiency, the type of language use (e.g. spoken or written), the language level considered (e.g. the sound system, the vocabulary, the syntactic structure, and the pragmatic conventions of language use), and so on. Obviously, all these factors interact with factors like the learner’s motivation, attitude, and language aptitude in a constantly changing system. The emerging picture is that the second language system is a language system in its own right rather than an “incomplete” target language system. Cross-linguistic influence is an important factor affecting that system. There is also a growing awareness now that cross-linguistic influence is not limited to L1 and L2, but that all sorts of influences between languages play a role, like the impact of the second on the third language or the impact of later learned language on the first language (Cook, 2002; Dussias & Sagarra, 2007; Kecskes & Papp, 2000). A Lifespan Perspective As a final issue in this chapter, we will relate the DST approach we have adopted here to developments in a person’s lifespan and discuss how these developments may affect the multilingual system.

< previous page

page_137

next page >

< previous page

page_138

next page >

Page 138 The dynamic systems perspective and the lifespan perspective go together well. Both assume that development continues over the lifespan and that initial conditions play an important role. The main difference is that in a lifespan perspective specific events, so called major life events, tend to more or less define an individual’s course of life. Examples are going to school, changing careers, but also divorce, disease and the loss of a beloved one. In the DST perspective such events are part of a more complex set of variables that interact over time. So, a specific life event is not seen as a monolithic factor exerting its role in a linear fashion but as a change of the total system. The death or departure of a family member may serve as an example. The effects of this event will not be limited to the absence of that individual but will impact on many other aspects of life, too. As a consequence, the roles in the family will be re-established, the living arrangements change, the financial situation changes, and the emotional links will change. Therefore, the removal of a part of the system leads to a complete overhaul and resettling of the system as a whole. This reorganization does not happen instantaneously, but gradually with the different aspects influencing each other over time. Major life events can explain why a system changes, while a DST approach may help to understand how the system changes, but the why and the how cannot always be clearly separated (de Bot, 2007). If we look at SLD from a lifespan perspective we can also look at the why and the how. The why can be understood by looking at those life events that are likely to have an impact on language use. In the earlier parts of this chapter, a number of factors have already been mentioned. Entering education is such an event. In many settings in the world where the language of the school is not the language of the home, a child’s entering the school system will have a major impact on the use of different languages, affecting the language use in the family and the community. The child finds out that there is more than one code to communicate and may develop positive or negative attitudes towards these languages, parents will be motivated to learn the other language in order to support the child at school and to be able to communicate with teachers and other parents, and eventually preferences for TV programs broadcast in specific languages may change. One possible result is that the home language becomes used less, which in itself may influence communication between generations. As research on language shift has shown, education in the dominant language is an important factor in shift from the minority to the dominant language (Kloss, 1966; Fishman, 1992). To summarize, a single event can start a chain of events the outcome of which is not always predictable. Other language-related life events are migration, study abroad, marriage with a partner speaking another language, having children in a multilingual setting, working in a multilingual setting, and retirement from such settings. Each of these events is likely to lead to a change in language use, with different languages dominating in different phases. These events are not always

< previous page

page_138

next page >

< previous page

page_139

next page >

Page 139 simple cause-and-effect chains. Belz (2007) reports on a German migrant to Sweden who completely refused to switch to the new language setting. So, a major change in the language setting will not necessarily lead to a shift in language use. Apart from major life events, smaller events can also be very relevant. The school may start teaching a foreign language from Grade 1, or a foreign language may become part of the curriculum in later years. Also, music or computer game preferences, having a penfriend abroad, or going on holidays in a foreign country may shift the balance between languages. Resources are also important to SLD, and changes in resources over the lifespan will have an impact on SLD. Most of the resources that are relevant for language are not stable and fixed, but rather variable. Examples of more stable traits are personality, language aptitude, and intelligence, while clearly more variable traits include contact and use of languages and motivation. Somewhere in between is working memory and speed of processing. Working memory increases in childhood and tends to be stable till old age (Behrens, Kemper, this volume). Speed of processing tends to be variable between and within individuals and declines with old age (Schrauf, this volume). All of these variables, also the more stable ones, interact over time. In the next paragraph we will take motivation as an example to show how one variable plays a role at different interacting timescales. There may be various reasons why an individual wants to learn a foreign language. On a timescale that spans years or decades, getting an interesting job some time in the future, wanting to integrate into a society after migration, or learning the language of a spouse can motivate someone at similar long-term timescales. Such global and long-term motivation will generate short-term and action-oriented types of motivation. Examples would be to opt for a foreign language at school, to enroll in a language course, or to start an e-mail exchange in the foreign language. In these cases, the timescale is in terms of weeks and months, which also involves motivation on a week-to-week or even day-to-day basis. Preparing homework, doing assignments, meeting for group work, all call for short-term motivation to get the work done. Finally, motivation is needed to actually do all the work on an hour-to-hour and minute-tominute basis. If the motivation to concentrate now and do this assignment is lacking, all the motivational resources at other levels will have little effect. The higher-level and more long-term motivational resources will feed the lowerlevel and more short-term activities, but also the other way around. A bad teacher, uninspiring materials, uninterested classmates and even a chilly classroom may drain motivational resources to the point that nothing gets done, and the motivation at the higher levels will go down as well. After all, what is the point of working through all those assignments that may or may not lead to that great job if doing this now is such a drag? Maybe taking classes in math or starting an MBA is much more effective in reaching that long-term goal. In sum, motivation as a factor will change over time on different timescales and changes

< previous page

page_139

next page >

< previous page

page_140

next page >

Page 140 on one scale will have an impact on other scales. Motivation is affected by and also affects other factors over time. Motivation is a clear example of how variation at different timescales works, but similar examples could be worked out for other factors, such as language proficiency or language contact. In addition, this example shows that changes are not completely dependent on life events that are external, but motivation can also be seen as an internal resource for SLD that is not caused by any specific life event. Conclusion In this chapter, we have looked at SLD over the lifespan from a DST perspective. We have argued that even though there are many separate systems, sub-systems, variables, processes, and factors that can be distinguished, none of these are ever static and none of these are independent of the others. On the contrary, all variables interact continuously with each other, and in doing so the variables themselves change, resulting in a continuously changing, somewhat unpredictable and chaotic “landscape,” reminiscent of the images produced by a lava lamp. It is obvious that in SLD many systems interact both within and between learners. Within a learner, the physical, social and cognitive systems interact and most second language learners will be in a setting with other individuals they interact with in social and economic communities. In agreement with what is found in other dynamical systems, these systems tend to be nested. The language system is nested within the cognitive system which is itself nested in the physical system of the body. At the same time, the individual shares cognition with other people and systems like computers. Individual learners are nested in the learning community (school, classroom, internet collaborative learning group) and the larger society. Languages are not learned from one moment to another: the learner processes a lot of input over time, taking every bit at the time and trying to make it fit in with the existing knowledge system and so iteratively causing the system to change to new input. In doing so, second language learners typically show variation in their development. There is not only variation in the acquisition process between individuals—some learn fast, others slowly—but also within individuals there is a great deal of variation. All developmental processes are dependent on resources. For SLD there are many such resources as described earlier: environmental factors like quality and quantity of input, instruction and motivation, but also learning materials, and cognitive resources such as aptitude, memory, and speed of processing. Even when L2 languagelearning situations are similar, there are some learners who will learn the L2 fluently and others who will not. This makes the SLD process to a certain degree unpredictable and non-linear in the sense that the developmental change may not always be proportionate to the input and resources invested.

< previous page

page_140

next page >

< previous page

page_141

next page >

Page 141 Dynamic systems are typically sensitive to initial conditions. Minor differences at early stages of development may have a significant impact in the course of time. One of the clear examples in SLD literature is the impact of relatively small perceptual processing deficits and their impact on language learning. Even the amount of education received prior to learning an L2 may have an impact on learning that L2. Many learners of a second language only develop their skills in that language to a limited degree. Through a lack of resources, like a lack of adequate input, attention and maybe instruction, they get stuck in what has been labeled fossilizations, that is, developmental stages that get frozen and seem to be impervious to input. These could be interpreted as attractor states of the language system. It is important to recognize that the variables that we have discussed may affect the same individual at different times differently or different individuals at the same time differently. We may acknowledge that there are general patterns to be recognized in language development, but that no individual will behave exactly the same and no exact predictions can be made about the multilingual development of individuals. This brings us to a final point in our discussion: how does the DST perspective affect our research agenda? If language development is not fully predictable and if research into isolated factors affecting SLD by definition implies a gross oversimplification of the true development of a dynamic system, how can we investigate SLD? Since language development is essentially an individual process, it will be obvious that we will need a shift from group studies to studies of individual development. One way of recording SLD is by carrying out microgenetic studies, in which the focus is on variability in the development (see, for instance, Verspoor, Lowie, & van Dijk, 2008). In microgenetic studies, a dense longitudinal analysis of individual learners is carried out to reveal the interactions between variables over time. Patterns of development, like an increased variation prior to a step in development, emphasize the process of language development rather than its products. If this approach is applied to multiple case studies, insight can be gained into both intraindividual variation and interindividual variation. Further progress in the field can be expected from attempts to apply dynamic growth models to language development (see, for instance, Bassano & van Geert, 2007, and van Geert, this volume, for applications to L1 development). Fitting mathematical models to actual developmental patterns can reveal the relative weight of the factors included in such a model. These outcomes can subsequently be empirically tested again. To investigate the dynamic nature of SLD, studies should be carried out at different timescales, to see if recursive patterns emerge. Having said that, we should realize that the DST approach to language development is still in its infancy and we are only beginning to grasp its real complexity. Statistical applications for DST models involve complex time series analyses that require huge datasets and

< previous page

page_141

next page >

< previous page

page_142

next page >

Page 142 that are much less obvious than relatively simple comparisons of group means. This implies that much time and effort will have to be invested in further developing this line of research and that it is going to take a while before the shift to proper DST research has become fully operational. Until that time, at least, the metaphorical interpretation of a dynamic view of second language development will help us appreciate its true complexity. Knowing the limitations of our knowledge will get us further than pretending that language follows a neat linear developmental pattern and has a limited set of clearly separable factors affecting it. Notes 1. We are grateful for the constructive feedback of an anonymous reviewer. We would also like the thank Joyce Haisma for her excellent proofreading and useful suggestions for improvement. Obviously, any remaining errors are our own responsibility. 2. And other contributions to the special issues of the journals Bilingualism: Language and Cognition (2007) and Modern Language Journal (2008). 3. The dynamic nature of motivation is explored in detail by Dörnyei (2001). 4. See www.swan.ac.uk/cals/calsres/lognostics.htm for the complete set and manual. 5. In later versions, the more politically correct term Language Coding Differences Hypothesis was used. References Bassano, D., & Van Geert, P. (2007). Modeling continuity and discontinuity in utterance length: A quantitative approach to changes, transitions and intra-individual variability in early grammatical development. Developmental Science , 10, 588–612. Belz, J. A. (2007). The differential loss and gain of identity in late bilinguals: The language learning memoirs of Werner Lansburgh. Internal paper, Department of German, Penn State University. Birdsong, D. (2006). Age and second language acquisition and processing: A selective overview. Language Learning , 56, 9–49. Bongaerts, T., Mennen, S., & Van Der Slik, F. (2000). Authenticity of pronunciation in naturalistic second language acquisition: The case of very advanced late learners of Dutch as second language. Studia Linguistica, 54, 298–308. Bongaerts, T., van Summeren, C., Planken, B., & Schils, E. (1997). Age and ultimate attainment in the pronunciation of a foreign language. Studies in Second Language Acquisition , 19, 447–465. Cook, V. (1995). Multi-competence and the learning of many languages. Language, Culture and Curriculum , 8 , 93– 98. Cook, V. (2002). Portrait of the language learner. Clevedon: Multilingual Matters. Coppetiers, R. (1987). Competence differences between natives and near-native speakers. Language, 63, 544–573. Cummins, J. (1977). Cognitive factors associated with the attainment of intermediate levels of bilingual skills. Modern Language Journal, 61, 3–12.

< previous page

page_142

next page >

< previous page

page_143

next page >

Page 143 Cummins, J. (1993). Bilingualism and second language learning. Annual Review of Applied Linguistics , 13, 51–70. Davis, C. (1994). Language planning in multilingual contexts. Policies, communities and schools in Luxembourg . Amsterdam/Philadelphia: John Benjamins. de Bot, K. (2007). Language teaching in a changing world. Modern Language Journal, 91, 274–276. de Bot, K., & Makoni, S. (2005). Language and aging in multilingual societies: A dynamic approach. Clevedon: Multilingual Matters. de Bot, K., Lowie, W., & Verspoor, M. H. (2007). A Dynamic Systems Theory approach to second language acquisition. Bilingualism: Language and Cognition, 10(1), 7–21. Dörnyei, Z. (2001). New themes and approaches in L2 motivation research. Annual Review of Applied Linguistics , 21, 43–59. Dussias, P. E., & Sagarra, N. (2007). The effect of exposure on syntactic parsing in Spanish–English bilinguals. Bilingualism: Language and Cognition, 10, 101–116. Ellis, N. (2002). Frequency effects in language processing. Studies in Second Language Acquisition , 24, 143–188. Ellis, N. (2008). The dynamics of second language emergence: Cycles of language use, language change, and language acquisition. Modern Language Journal, 92(2), 232–249. Ellis, R. (1994). The study of second language acquisition. Oxford: Oxford University Press. Fishman, J. (1992). Reversing language shift. Clevedon: Multilingual Matters. Flege, J. E. (1987). The production of “new” and “similar” in a foreign language: evidence for the effect of equivalence classification. Journal of Phonetics, 15, 47–65. Flege, J. E., Yeni-Komshian, G. H., & Liu, S. (1999). Age constraints on second-language acquisition. Journal of Memory and Language, 41, 78–104. Giles, H., & Billings, A. (2004). Assessing language attitudes: Speaker evaluation studies. In A. Davies & C. Elder (Eds.), The handbook of applied linguistics (pp. 187–209). Malden: Blackwell. Hagoort, P. (2006). What we cannot learn from neuroanatomy about language learning and language processing. A commentary on Uylings. In M. Gullberg & P. Indefrey (Eds.), The cognitive neuroscience of second language acquisition (pp. 91–97). Malden: Blackwell. Hakuta, K., Bialystok, E., & Wiley, E. (2003). Critical evidence: A test of the critical-period hypothesis for secondlanguage acquisition. Psychological Science , 14, 31–38. Herdina, P., & Jessner, U. (2002). A dynamic model of multilingualism. Perspective of change in psycholinguistics . Clevedon: Multilingual Matters. Hopp, H. (2007). Ultimate attainment at the interfaces in second language acquisition: Grammar and processing. Unpublished PhD, University of Groningen. Horwitz, E., Horwitz, M., & Cope, J. (1986). Foreign language classroom anxiety. Modern Language Journal, 70, 125– 132. Hulsen, M., de Bot, K., & Weltens, B. (2002). Between two worlds. Social networks, language shift and language processing in three generations of Dutch migrants in New Zealand. International Journal of the Sociology of Language, 153 , 27–52. Johnson, L., & Newport, E. (1989). Critical period effects in second language learning: The influence of maturational state on the acquisition of English as a second language. Cognitive Psychology, 21, 60–99.

< previous page

page_143

next page >

< previous page

page_144

next page >

Page 144 Kecskes, I., & Papp, T. (2000). Foreign language and mother tongue . Mahwah, NJ: Lawrence Erlbaum. Kellerman, E. (1986). An eye for an eye: Crosslinguistic constraints on the development of the L2 lexicon. In E. Kellerman & M. S. Smith (Eds.), Crosslinguistic influence in second language acquisition. London/New York: Pergamon Press. Kloss, H. (1966). German–American language maintenance efforts. In J. Fishman (Ed.), Language loyalty in the United States. The Hague: Mouton de Gruyter. Larsen-Freeman, D. (1997). Chaos/complexity science and second language acquisition. Applied Linguistics , 18, 141– 165. Larsen-Freeman, D. (2005). Second language acquisition and the issue of fossilization: There is no end and there is no state. In Z. Han & T. Odlin (Eds.), Studies of fossilization in second language acquisition. Clevedon: Multilingual Matters. Larsen-Freeman, D., & Cameron, L. (2008). Research methodology on language development from a complex theory perspective. Modern Language Journal, 92(2), 200–213. MacIntyre, P., & Gardner, R. (1994). The subtle effects of induced anxiety on cognitive processing in the second language. Language Learning , 44, 283–305. Mägiste, E. (1986). Selected issues in second and third language learning. In J. Vaid (Ed.), Language processing in bilinguals: Psycholinguistic and neurolinguistic perspectives (pp. 97–122). Hillsdale, NJ: Lawrence Erlbaum. Meara, P. (2004). Modeling vocabulary loss. Applied Linguistics , 25, 137–155. Miyake, A., & Friedman, N. (1999). Individual differences in second language proficiency: Working memory as language aptitude. In A. Healy & L. Bourne (Eds.), Foreign language learning: Psycholinguistic experiments on training and retention (pp. 339–362). Mahwah, NJ: Lawrence Erlbaum. Odlin, T. (1989). Language transfer. Cross-linguistic influence in language learning . Cambridge: Cambridge University Press. Pallier, C. (2007). Critical periods in language acquisition and language attrition. In B. Köpke, M. Schmid, M. Keijzer, & S. Dostert (Eds.), Language attrition: Theoretical perspectives. Amsterdam: John Benjamins. Pichora-Fuller, M. K. (2003). Cognitive aging and auditory information processing. International Journal of Audiology, 42, S26–S32. Pimsleur, P. (1966). Pimsleur language aptitude battery. New York: Harcourt Brace. Pimsleur, P. (1968). Language aptitude testing. In A. Davies (Ed.), Language testing symposium: A linguistic approach (pp. 98–106). New York: Oxford University Press. Port, R., & van Gelder, T. (1995). Mind as motion: Exploration in the dynamics of cognition. Cambridge, MA: Bradford. See, S., & Ryan, E. (1995). Cognitive mediation of adult age-differences in language performance. Psychology and Aging , 10, 458–468. Sorace, A. (1993). Incomplete vs. divergent representations of accusativity in non-native grammars of Italian. Second Language Research, 9 , 22–47. Sparks, R. L., Ganschow, L., & Javorsky, J. (1992). Diagnosing and accommodating college students with foreign language learning difficulties. Learning Disabilities: Research and Practice , 7 , 150–160. Sparks, R. L., Ganschow, L., & Patton, J. (1995). Prediction of performance in first-year foreign language courses: Connections between native and foreign language learning. Journal of Educational Psychology, 87, 638–655. van Geert, P. (1995). Growth dynamics in development. In R. Port & T. van Gelder

< previous page

page_144

next page >

< previous page

page_145

next page >

Page 145 (Eds.), Mind as motion: Explorations in the dynamics of cognition (pp. 313–337). Cambridge, MA: Bradford. Verspoor, M. H., Lowie, W. M., & van Dijk, M. (2008). Variability in L2 development from a dynamic systems perspective. Modern Language Journal, 92(2), 214–231. Wilhelm, F. (2005). English in the Netherlands. A history of foreign language teaching 1800–1920 . Unpublished PhD thesis, Katholieke Universiteit Nijmegen. Wode, H. (1994). Nature, nurture, and age in second language acquisition: The case of speech perception. Studies in Second Language Acquisition , 16, 325–345. Woutersen, M. (1997). Bilingual word perception . Unpublished PhD thesis, Katholieke Universiteit Nijmegen.

< previous page

page_145

next page >

< previous page

page_146

next page >

Page 146 6 BILINGUALS WITH PRIMARY LANGUAGE IMPAIRMENT Kathryn Kohnert By definition, individuals with language impairments have reduced ability in language—any language and all languages. Many language impairments are explained by obvious damage to sensory, cognitive or neurological systems, as is the case in hearing loss, mental retardation or dementia. For individuals with primary language impairments the most obvious area of difficulty is with language. Developmental and acquired primary language impairments are high-incidence disabilities among single-language speakers and most likely affect bilingual populations in similar numbers. For example, a subset of children learning one language from birth and a second language beginning in early childhood lag behind their sequential bilingual peers in communication skills for no readily apparent reason. This weakness in language persists through adolescence and into adulthood. In other cases, adults who once spoke two languages proficiently have chronic difficulty communicating in either language following a stroke, although other major systems seem well preserved. The focus of this chapter is on developmental and acquired primary language impairments in bilingual populations. The first section provides a general overview of communication at the intersection of bilingualism and primary language impairment. The second section focuses on primary developmental language impairment, first in monolingual children and then in children learning two languages. The third section is dedicated to primary acquired aphasia in bilingual adults. 1 Bilingualism and Primary Language Impairment: An Overview Individuals learn two languages under a wide variety of circumstances. Some have consistent experience with two languages beginning at birth. Others learn a single language from birth and a second language (L2) in childhood while attending educational programs that use a language different from the

< previous page

page_146

next page >

< previous page

page_147

next page >

Page 147 one spoken at home. For others, L2 learning begins later in life with visits, formal study or immigration to another country. In all instances, L2 learners vary in their maturational states, first language (L1) ability, world experience, and reasons for L2 acquisition. For present purposes, bilinguals are operationally defined from a functional or needsbased perspective. Individuals who have systematic experience with two or more languages or those who rely on two or more languages to meet communicative needs are considered bilingual (cf. Kohnert, 2008). By definition, individuals with language impairments have difficulty “doing language.” As such, proficiency-based definitions of bilingualism simply do not work in the present context as they exclude the key population of interest. 1.1 Language Proficiency, Language Impairment and MOM Language is a complex, multi-layered symbol system used to communicate with other individuals. The same interacting factors that contribute to success in one language are at play in the successful learning, use and maintenance of two languages. MOM—means, opportunities, and motive—affect the mastery and maintenance of any complex behavior. Means refers to learner-internal resources; those that affect language include the integrity of the cognitive, sensory, social-emotional, and neurobiological systems (e.g. Paul, 2001). Any developmental or acquired inefficiency in one or more of these systems that present challenges to one language will also undermine ability in two languages. Some endogenous deficiencies that affect language are readily apparent while others are subtle and difficult to explain. Opportunities refer to social factors, including the availability of rich language in the environment and diverse opportunities to develop or use a particular language for meaningful communicative interactions. Language opportunities come in spoken or written form and through private channels or public media. Potential language-use environments are home, school, work, sporting venues, or cyberspace. Each of a bilingual’s two languages may be used for different purposes and with different partners. Limited opportunities in one language will stifle development or ability in that language. Reduced outcomes or even declines in a single language because of limited opportunities are not language impairments but rather a natural consequence of evolving circumstances (e.g. Anderson, 2004; de Bot & Makoni, 2005; Jia & Aaronson, 2003; Montrul, 2005; Schmid, 2003, this volume). Opportunities may be viewed as a necessary but insufficient requisite for language acquisition or maintenance across the lifespan. The final “M” in “MOM” refers to motive. Motive reflects interactions between internal and external resources— between environmental needs and opportunities as well as personal preferences inextricably linked to social contexts. For bilingual individuals, motive may interact with both partners and purposes—one language may be preferred to communicate with a parent,

< previous page

page_147

next page >

< previous page

page_148

next page >

Page 148 child or spouse, and another to complete a business transaction. Another interacting aspect of motive is the social status or prestige associated with different languages in a given community. A status differential, in which one language is more valued by the broader community, contributes to shifts in personal preferences and proficiency levels (Lambert, 1977). This is particularly true for children who learn a minority L1 from birth and the majority community language as their L2 beginning with attendance in formal educational programs. The end result is that over time there is a shift to greater proficiency in the majority community language (e.g. Jia & Aaronson, 2003; Jia, Kohnert, Collado, & Aquino-Garcia, 2006; Kohnert & Bates, 2002). For both older and younger bilingual individuals, motive and opportunities to use language for rich, meaningful interactions go hand-in-hand. Thus, when MOM is sufficient, ability in each of a bilingual’s two languages will be developed and maintained. When one or more aspects of MOM are weak, either language—or both—may be affected. Low proficiency in one of a bilingual’s two languages because of reduced opportunities or motivation may be consistent with typical bilingual variation. In contrast, limited ability in both languages may reflect a weakness in the underlying language processing mechanism, not attributed to differences in learning opportunities or motivation. 1.2 Reduced Means: Primary Language Impairment The focus here is at the intersection of bilingualism and primary language impairments in children and adults. Primary language impairments are deficits or inefficiencies in the comprehension and/or production of spoken or written language alongside more preserved functioning in other areas. The underlying deficit that affects language processing is evident in each language of bilingual children and adults. Just as with typical bilinguals, however, the relative degree of cross-linguistic proficiency may vary in bilinguals with language impairments. That is, both languages may be equally affected, or one language may be relatively weaker than the other. Differences in cross-linguistic skill may be due to different opportunities to develop or use each language. For adults with acquired language impairment, differences in cross-linguistic skill may be related to premorbid levels of proficiency in each language, a direct result of the injury or consistent with the language in which systematic training opportunities were provided following injury. In healthy bilinguals, language decline may be a result of interactions between weakened means, changing social roles and vocational opportunities associated with advancing age (cf. de Bot & Makoni, 2005). For bilinguals with primary language impairments, communication difficulties exceed those found in chronological age peers, across the lifespan. Both languages must be systematically considered in assessment and intervention activities. The presence and severity of primary language disorders are determined

< previous page

page_148

next page >

< previous page

page_149

next page >

Page 149 by referencing typical or “normal” language performance. The parameters of normal or acceptable skills vary with age and language experiences, relative to the individual’s social environments. For example, to identify language impairment in an 8-year-old boy who has learned Turkish at home from birth and Dutch at school beginning at age four, we must consider his ability or proficiency, in both languages, as compared to typically developing children of the same age learning Turkish and Dutch under similar circumstances. In order to understand the severity of impairment in a bilingual Spanish–English speaking man with language deficits following a stroke in the left hemisphere, we must understand his skill in each language prior to the injury, as well as the functional need for each language in his daily life. Severity of the impairment is determined both objectively and subjectively. These combined perspectives are intended to determine the impact the endogenous difficulty in language has on the affected individual’s ability to participate in meaningful life activities (Centeno, 2005; Kohnert, 2008). The study of primary language impairment in bilingual populations is interesting and important for both theoretical and practical reasons. On the theoretical side, the understanding of language breakdown or inefficiencies in learning two languages can inform our basic understanding of interactions between the cognitive system that supports languages and the social environments in which they are acquired and used. This information is necessary in order to develop unified theories of language that encompass typical and atypical speakers of one or multiple languages across the lifespan (cf. Kohnert & Windsor, 2004). At the same time, the study of primary language impairments in children and adults is of enormous practical interest. Given that over half of the global population speaks two (or more languages) and that primary language impairments are high-incidence disorders, the ability to provide appropriate clinical and educational services to bilingual individuals with language impairments is a pressing need. In the following section we consider children with primary language impairment. 2 Developmental Primary Language Impairment (PLI) 2.1 Monolingual PLI Children who lag behind peers in language in the face of otherwise typical development are referred to by various names. These different names include late talkers, language acquisition disorder, learning disability, procedural language impairment, and specific language impairment or SLI. Differences in nomenclature emphasize changes in the most obvious characteristics of the affected population across developmental stages as well as divergent theoretical perspectives. The term primary language impairment, or PLI, is preferred here as it is most consistent with available evidence without presupposing a particular etiological cause onto the diagnostic category

< previous page

page_149

next page >

< previous page

page_150

next page >

Page 150 (Kohnert, Windsor, & Ebert, in press; Tomblin, Zhang, Buckwalter, & O’Brien, 2003; Windsor & Kohnert, in press). PLI is the most common developmental communication disorder and has been identified in a wide range of typologically different languages, including Spanish, Arabic, Cantonese, German, Icelandic, and English (Leonard, 1998). In the US, PLI affects an estimated 5–7% of children, boys somewhat more than girls (Paul, 2001; Tomblin et al., 1997). Some mild forms of delayed language identified in early childhood will resolve with time, while PLI will persist throughout the lifespan, changing in its manifestation as a function of development, social circumstances and personal style. Children with PLI perform within the normal range on standardized non-verbal intelligence and hearing measures, yet fail to make expected progress in language (American Psychiatric Association, 1994; Leonard, 1998). The social environments for children with PLI also do not differ from those of children who are developing language typically. Recent anatomical findings indicate a neurological component to PLI although there is no clear lesion site or frank neurological damage (see Ullman & Pierpont, 2005, for review). A positive family history of language or learning impairment is considered a risk factor for PLI (Plomin & Dale, 2000). 2.1.1 Language Characteristics Children with PLI show a protracted period of language development, are at high risk for reading and academic failure, and, by definition, perform below their age- and experience-matched peers on standardized measures of language. Within the language domain, the most salient features of PLI change across age, in part reflecting changing characteristics of unaffected age-peers on which performance standards are based. For example, about 15% of otherwise seemingly typical 2-year-olds do not meet early production milestones and are classified as “late talkers” (Klee, Pearce, & Carson, 2000). Unlike their peers, these “late talkers” do not yet have a minimal core vocabulary of 50 to 100 words and do not produce two- or three-word combinations (Girolametto, Wiigs, Smyth, Weitzman, & Pearce, 2001). About half of children who fall into the “late talkers” category will catch up to peers by age 3 without intervention and are thus referred to with the more benign term of “late bloomers” (Rescorla & Roberts, 2002). Late talkers at greatest risk for persisting delays appear to have reduced skills in understanding as well as producing language, a positive family history of language or learning disability, few gestures, limited symbolic play skills, and more frequent or lasting occurrences of otitis media (Kelly, 1998). In 4-to-7-year-old children with PLI, grammatical deficits are often considered the most distinguishing feature. However, it seems that both the types and prominence of grammatical errors in PLI differ as a function of the unique typological features of the language being acquired. English-speaking

< previous page

page_150

next page >

< previous page

page_151

next page >

Page 151 children with PLI tend to omit short, unstressed verb forms indicating tense or agreement such as third person singular –s, regular past tense –ed, and the verb “be” (e.g. Cleave & Rice, 1997; see Leonard, 1998, for review). Spanish-speaking children with PLI appear to have more difficulty with noun morphology, such as adjectiveagreement inflections, clitic pronouns (damelo/give me it), plural nouns, and articles (Bedore & Leonard, 2001; Restrepo & Gutiérrez-Clellen, 2001). Interestingly, evidence from French-speaking preschoolers with PLI indicates that morphosyntax is no more impaired than other areas of language (Thordardottir & Namazi, 2007). There is little information indicating the most salient symptoms of PLI in children who speak languages that do not rely on grammatical inflections to convey meaning, such as Vienamese or Hmong. Thus it seems the most observable characteristics of PLI reflect the internal language processing system interacting with the typological characteristics of the ambient language. During the school years, individuals with PLI may have difficulty with social discourse as well as academic language. They produce narratives that are shorter or less complex than those of their peers (Gutierrez-Clellen, 2004; Scott & Windsor, 2000). School-age children with PLI have significant difficulty in learning to read or write, as these are language-based activities (Kamhi & Catts, 1999). Children with PLI may also have marked difficulty in using language for social purposes which interferes with peer interactions (e.g. Fujiki, Brinton, Morgan, & Hart, 1999; Gertner, Rice, & Hadley, 1994). Children and adolescence with PLI process language more slowly or less efficiently than unaffected peers. Timed measures of language processing have revealed consistent differences at the group level between children with PLI and their non-language-impaired peers. For example, when asked to provide the name of familiar pictures or to match a string of word-like sounds to one of two pictures, children with PLI are slower than unimpaired age peers— even after controlling for differences in response accuracy (Kohnert, Windsor, & Miller, 2004; Windsor & Kohnert, 2004). Subtle differences in processing efficiency can have a big effect on language, considering the rapid pace at which natural language interactions take place. The reduced efficiency in processing information experienced by many children with PLI may have cascading effects, negatively affecting learning and classroom engagement (see Hsu & Karmiloff-Smith, in press, for related discussion). Although the majority of research on PLI has been conducted with children, the underlying weakness affecting language acquisition is chronic and lifelong. Adults with well-documented histories of PLI perform more poorly on a range of language-based tasks as compared to adults without histories of language impairment (Tomblin, Freese, & Records, 1992). PLI in adults is also associated with lower educational and vocational attainment, along with higher rates of conduct and psychiatric disorders (Brownlie et al., 2004;

< previous page

page_151

next page >

< previous page

page_152

next page >

Page 152 Johnson et al., 1999). Despite the long-term disadvantage in objective measures of long-term outcomes in PLI, selfreports are more positive. Records, Tomblin, and Freese (1992) found no significant differences between selfreported ratings of adults with and without histories of PLI on “quality of life” measures. That is, 29 young adults with histories of PLI in this study did not rate themselves as less happy or less well-off than the control group, despite lower educational and employment status (Records et al., 1992). Additional study of language performance and social outcomes in adults with histories of PLI throughout the lifespan will inform understanding of basic cognitive–linguistic mechanisms. For example, it may be that developmental vulnerabilities affecting language acquisition persist and exacerbate language declines observed in healthy aging (e.g. Connor, Spiro, Obler, & Martin, 2004; Dreary & Der, 2005; Kemper, 1987). 2.1.2 Beyond Language: General Processing Inefficiencies In PLI language is the most obvious area of functional weakness, perhaps because of the considerable demands of real-time language use. However, processing inefficiencies are not exclusive to the language domain. Subtle inefficiencies in the general cognitive processing system exist alongside the more obvious deficits in language and, from some perspectives, play a causal role in observed language deficits (Kohnert, 2004a; Kohnert & Windsor, 2004). General processing weaknesses in monolingual children with PLI have been documented on basic perceptual, motor, memory and attentional tasks that require little or no language ability (Miller et al., 2006; Windsor, Milbrath, Carney, & Rakowski, 2001; see also Leonard 1998 for review). For example, children with PLI are less skilled in detecting pure tones of brief duration, in replicating a series of colored lights, in rapidly tapping their fingers, in moving pegs along a board, in stringing beads, and in mentally rotating geometric shapes (see reviews in Bishop, 1992; Hill, 2001; Kohnert & Windsor, 2004; Leonard, 1998; and Ullman & Pierpont, 2005). Viding and colleagues (2003) found that PLI in one twin predicted poor non-verbal ability in a co-twin, with the strength of this predictive relationship greater for monozygotic than dizygotic twins (Viding et al., 2003). These results show that general genetic factors implicated in PLI extend beyond language to include challenges with non-linguistic information processing. It remains open whether non-linguistic and language deficits in PLI are correlated because of a common underlying neurological deficit (Ullman & Pierpont, 2005) or causally related, with subtle general motor, perceptual and cognitive weaknesses resulting in the obvious lags in language (see Leonard, 1998, and Ellis Weismer & Evans, 2002, for reviews). It also is unclear if or how associated cognitive weaknesses in PLI documented during development change across adulthood and with advancing age.

< previous page

page_152

next page >

< previous page

page_153

next page >

Page 153 2.2 Bilingualism and PLI Individuals with PLI learning two languages have the same backgrounds as their typical bilingual peers. Most children who learn two (or more) languages during childhood are typical; with appropriate time and experience they will become skilled in the languages consistently used for meaningful communicative interactions in their environments. Typically developing learners have intact language processing mechanisms—the internal means needed for language acquisition and use. However, just as with single-language learners, a subset of dual-language learners will have a language processing system that is somehow compromised. The underlying impairment will affect both languages. The admittedly limited evidence to date suggests that bilingual children with PLI exhibit impairment in each language similar to monolingual speakers of each language with PLI (J. Paradis, Crago, Genesee, & Rice, 2003). Adolescence and adults with histories of PLI will also be challenged in learning a second or foreign language, just as they experienced difficulties in L1. As with monolingual children with PLI, bilingual children with PLI need significantly more explicit opportunities to develop and use each language for the range of communicative purposes for which they are needed. As such, children with PLI are more vulnerable to the effects of limited language-specific experience than typical learners. They may exhibit early plateaus or rapid regression in a language that is not reinforced in academic, community or treatment settings (Håkansson, Salameh, & Nettelbladt, 2003; Kohnert, Yim, Nett, Kan, & Duran, 2005; Salameh, Håkansson, & Nettelbladt, 2004). The ability to accurately identify PLI among linguistically diverse learners is a first step in implementing adequate treatment programs. As described in the following section, this process is inherently complex. 2.2.1 Diagnostic Challenge: Separating Differences from Disorders PLI due to inefficiencies in children’s internal language systems is chronic. Timely, effective intervention is needed to improve long-term academic and social outcomes. A critical first step in this process is appropriate language assessment. Separating normal variation among dual-language learners from PLI is not an easy task. Identification of language impairment, that is the “ruling in” or “ruling out” of PLI, is predicated on a clear understanding of the parameters of typical performance in the face of diverse circumstances. Children learning two languages are an extraordinarily heterogeneous group, in terms of the combination of languages learned, the ages at which systematic experience with the different languages begins, and the variation in quantity and quality of learning opportunities. This heterogeneity significantly complicates the development of normative data that can be used to reference typical language skills in developing bilinguals, as distinct from that

< previous page

page_153

next page >

< previous page

page_154

next page >

Page 154 of children with PLI (Kohnert, 2008). As a result, children learning two languages are both under- and overidentified as PLI (e.g. Roseberry-McKibbin, Brice, & O’Hanlon, 2005; Silliman, Wilkinson, & Brea-Spahn, 2004). On traditional language measures, such as language samples or standardized vocabulary tests, there may be considerable overlap between children learning a second language and monolingual children with PLI (e.g. Håkansson & Nettelbladt, 1996; J. Paradis & Crago, 2000; Schiff-Myers, 1992; Windsor & Kohnert, 2004). The documented poorer performance on these traditional language measures by both children with PLI and typical sequential bilinguals, relative to their language-intact monolingual age peers, is explained in very different ways. Children with PLI are considered to have reduced means, a deficit internal to the child. In contrast, most children who are developing two languages sequentially have intact language processing systems but may perform poorly on traditional language tasks relative to monolingual age peers because of differences external to the child. Different language-specific opportunities often mean that school-age children who, for example, learned Urdu or Spanish as L1 and English as L2 beginning in early childhood may have had fewer opportunities to learn or use each language. Moreover, linguistic knowledge may be distributed across the two languages of bilinguals in ways that are not readily comparable to individuals who speak a single language (e.g. Kan & Kohnert, 2005; Peña, Bedore, & Zlatic-Giunta, 2002). Cultural mismatches between traditional language measures and the child’s previous experiences may also bias assessment results. Processing-dependent measures have been proposed as an alternative to traditional lexical, grammatical and narrative tasks. Processing-dependent measures of language are designed to assess the integrity of the underlying language learning system while simultaneously reducing the role of previous cultural or linguistic experiences. Nonword repetition is the processing-dependent task which has gained the most attention over the past decade. In nonword repetition, children are asked to repeat a series of phonemes that are word-like, but have no meaning, such as mifup or woginfer. The sounds and sound combinations in non-word repetition tasks are consistent with the conventions of the test language (Ebert, Kalanek, Cordero, & Kohnert, 2008). Performance on non-word repetition tasks has successfully separated English-speaking children with and without PLI (e.g. Campbell, Dollaghan, Needleman, & Janosky, 1997; Ellis Weismer et al., 2000). Similarly, performance on Spanish non-word repetition tasks has successfully separated Spanish-speaking children with PLI from their typical peers (Girbau & Schwartz, 2007). However, it seems that processing-dependent measures that use language stimuli, including non-word repetition, are not sensitive to PLI across all languages and do not overcome assessment bias when used with developing bilinguals. For example, Stokes and colleagues found that performance on a non-word repetition task did not separate Cantonese speaking children with PLI from their typically developing peers (Stokes,

< previous page

page_154

next page >

< previous page

page_155

next page >

Page 155 Wong, Fletcher, & Leonard, 2006). Kohnert and colleagues found that even sophisticated English L2 learners were at a relative disadvantage on a non-word repetition task as compared to monolingual English-speaking peers (Kohnert, Windsor, & Yim, 2006). In summary, to date there is no definitive method for separating differences from disorders among a linguistically diverse population. Rather, clinical researchers and speech-language pathologists must gather information in each of the child’s languages from multiple sources at different points in time, then interpret this gathered information with respect to the child’s language, cultural and educational opportunities (Kohnert, 2008). It requires considerable experience on the part of the diagnostician. In the following section we discuss potential contributions to this diagnostic challenge offered by investigations at the intersection of bilingualism and PLI that go beyond language to look at general information processing. 2.2.2 Typical Bilingualism vs. PLI: Contributions of Non-linguistic Task Performance Investigations at the intersection of bilingualism and PLI have generally been restricted to the language domain. This tradition is both consistent with the monolingual PLI literature and reasonable given that PLI is defined by deficits in language. However, it is informative to look beyond language, to investigate children’s performance on basic nonlinguistic information processing tasks to determine if there are fault lines that distinguish typical learners, independent of the number and types of languages learned, from peers with PLI. Recall that children with PLI may have subtle inefficiencies in the general cognitive processing mechanisms which challenge the acquisition and use of language codes present in their environments. Tasks which measure the speed or efficiency with which basic nonlinguistic information is processed are much less dependent on specific cultural or language experiences and therefore potentially less biased. Such an approach is inconsistent with modular perspectives of language or linguistic deficit accounts of PLI (which prefer the term specific language impairment). This approach is, however, very consistent with domain general interactive theories of language, including emergentism, dynamical systems theories and dynamic interactive accounts of PLI (e.g. Evans, 2001; Kohnert, 2004a, 2008). In a series of studies, Kohnert, Windsor and colleagues used a variety of non-linguistic as well as language-based processing measures to compare performance between sequential bilingual children and monolingual age peers, with and without PLI (Kohnert et al., 2004, 2006; Kohnert & Windsor, 2004; Windsor & Kohnert, 2008; Windsor, Kohnert, Loxtercamp, & Kan, 2008). Participants were typically developing 8- to 13-year-old Spanish–English bilinguals and English-only speaking children. Language-based processing tasks included non-word repetition, timed picture naming,

< previous page

page_155

next page >

< previous page

page_156

next page >

Page 156 picture–word verification and grammatical judgment tasks. Non-linguistic processing measures included choice auditory detection, choice visual detection, auditory serial memory and visual serial memory. Auditory detection required participants to respond to the presence of either a high or low tone by pushing one of two buttons: in visual detection children pushed a button associated with either a blue or red shape. The serial memory tasks required children to replicate either a series of high or low tones or a series of flashing colored lights by pushing associated buttons. In each of the memory tasks the sound or light sequences ranged from two to five items. No phonemes, words or other language stimuli were used on these measures. The response variables of interest were speed (auditory and visual choice detection tasks) or accuracy (auditory and visual serial memory tasks). Results from these studies revealed clear areas of overlap among typical monolingual and bilingual learners as well as fault lines that separated typical learners from children with PLI. The most robust areas of separation between children with PLI and their typical, albeit linguistically diverse, peers were on the non-linguistic tasks. Tasks that required language mediation generally placed even sophisticated L2 learners at a disadvantage. In summary, combined results from this first round of studies at the intersection of bilingualism, PLI and the general information processing system suggest that performance on selected non-linguistic information processing tasks may be useful in separating typically developing bilingual children from children with PLI (see Kohnert et al., in press, for review). Systematic exploration of task and child factors that underlie within- and across-group performance seems critical to help shape robust theory. Documenting more precisely where the non-linguistic weaknesses are in PLI as compared to diverse groups is an essential step as it paves the way to more refined methods for testing hypothesized causal factors (Ullman & Pierpont, 2005). From a practical perspective, findings from these combined studies point to a new direction for the development of less-biased assessment techniques. Specifically, it may be that performance on some set of non-linguistic processing tasks can help identify children with PLI in a linguistically diverse population. Importantly, the potential use of non-linguistic techniques would serve to complement, not replace, language assessment and treatment measures. In the following section we turn our attention to language and cognitive features of acquired primary language impairment in adults. 3 Acquired Primary Language Impairment in Adults: Aphasia Language is made possible by the anatomy and physiology of the human brain (see Hernandez, Hiscock, & Bates, this volume, for review and discussion). Damage to certain parts of the brain may result in aphasia, which

< previous page

page_156

next page >

< previous page

page_157

next page >

Page 157 appears as sudden and chronic difficulty in producing or understanding language, in spoken or written modalities. The parts of the brain affected in aphasia include those that have primary responsibility for processing language content and form. In their classic clinical text, Rosenbek, LaPointe, and Wertz (1989) define aphasia as: an impairment, due to acquired and recent damage of the central nervous system, of the ability to comprehend and formulate language. It is a multimodality disorder represented by a variety of impairments in auditory comprehension, reading, oral-expressive language, and writing. The disrupted language may be influenced by physiological inefficiency or impaired cognition, but it cannot be explained by dementia, sensory loss or motor dysfunction. (Rosenbek et al., 1989, p. 53) Chronic aphasia affects an estimated 1 of every 275 people (National Aphasia Association, 2008). The most common cause of aphasia is stroke, an interruption of normal blood flow to functional areas of the brain. Strokes are associated with heart disease, high blood pressure, and diabetes. Although strokes can occur at any age, stroke risk doubles for each decade after age 55 (National Stroke Association, 2003). Additional causes of aphasia are gunshot wounds, tumors, aneurysms or other focal brain trauma. Next we provide an overview of the aphasia profile, based on the monolingual literature. 3.1 The Aphasia Profile The nature and severity of language impairment in aphasia varies dramatically both within and across individuals. Within any individual, communication difficulties are greatest immediately following the sudden insult to the brain. This may be followed by marked improvement during the spontaneous recovery phase, which lasts 6 months or perhaps up to a year after symptom onset. Spontaneous recovery is a term use to describe the natural healing and functional process that takes place following sudden brain trauma. With time, functional recovery of language stabilizes and the individual moves into the chronic stage of aphasia. Some individuals with chronic aphasia seem to have little trouble with most language tasks, with the exception of difficulty in finding and producing the right words at the right time. Aphasia in others is so profound they are unable to understand or produce any words at all. The severity of language impairment for most individuals with aphasia falls somewhere in between these extremes. A wide variety of language-based treatments have been effective in improving communication skills in monolingual individuals with chronic aphasia (e.g. Meinzer, Djundja, Barthel, Elbert, & Rockstroh, 2005; Robey, 1998; Thompson & Shapiro, 2005).

< previous page

page_157

next page >

< previous page

page_158

next page >

Page 158 A number of different aphasia syndromes have been described based on the relative preservation or impairment of particular language functions. The focus of the clinical literature has often been on identifying and separating aphasia into subtypes. However, it seems that there is no exact correspondence between lesion site and discrete aphasia symptoms, and different language tests classify aphasia symptoms into different sub-types (e.g. Crary, Wertz, & Deal, 1992). For present purposes, aphasia is divided into three very general and readily recognizable types: non-fluent or expressive aphasia, fluent or receptive aphasia, and mixed receptive–expressive aphasia. Expressive or non-fluent aphasias are associated with focal damage to the anterior portion of the language-dominant cerebral hemisphere. The individual may produce slow, halting, non-fluent language consisting of short, grammatically simple utterances. Comprehension may also be affected but not to the same degree as expressive language. In contrast, receptive or fluent aphasias are more closely associated with damage to posterior portions of the language-dominant hemisphere. In receptive aphasias, the ability to understand linguistic symbols in all modalities is significantly impaired. The speech in receptive aphasia is produced with seemingly little effort, following the rhythmic flow of the native language, with few pauses or breaks. However, a close listen to the content of this speech reveals that it has little meaning. Because of this lack of content or meaning, the speech produced by individuals with receptive aphasia sounds like “word salad” or jargon. Most often, expressive as well as receptive language is affected to some degree. Mixed receptive–expressive aphasias result in mild, moderate, severe or profound impairment in the ability to produce and understand conventional linguistic symbols in meaningful ways. In all types and severity levels of aphasia, inefficiencies or weaknesses in the basic cognitive information processing system may accompany the more easily recognized difficulties in expressive and/or receptive language. Indeed, from general cognitive interactive theoretical perspectives, the linguistic manifestations in aphasia are epiphenomenal and secondary to a more fundamental information processing deficit (e.g. Chapey, 2001; Luria, 1966; McNeil, 1988; McNeil & Pratt, 2001). In some cases cognitive weaknesses may be obvious, with symptoms including impaired orientation or memory, significant fatigue, or difficulty in sustaining attention during simple activities. More often, impairment in the general cognitive processing system is subtle and overlooked in aphasia in light of the much more observable and characteristic language deficits. Findings from a number of recent empirical studies with monolingual adults reinforce the notion of aphasia as a primary, but not exclusive, deficit in language (Helm-Estabrooks, 2002; Van Mourik, Verschaeve, Boon, Paquin, & Van Harkskamp, 1992). Saygin and colleagues found robust correlations between performance on language and non-linguistic tasks in individuals with aphasia, suggesting that the two domains draw on common

< previous page

page_158

next page >

< previous page

page_159

next page >

Page 159 neural regions and or/cognitive processes (Saygin, Dick, Wilson, Dronkers, & Bates, 2003; Saygin, Wilson, Dronkers, & Bates, 2004). Recognition of cognitive weaknesses as part of the aphasia profile leaves open the possibility that, at least in some cases, treatment which strengthens underlying cognitive processing mechanisms will lead to gains in overall language ability (cf. Coelho, 2005; see Kohnert, 2008, for discussion). 3.2 Acquired Aphasia in Bilingual Adults For most neurologically intact monolinguals, the processing centers closely associated with the content and formal aspects of language are lateralized to the left hemisphere. This is also true for both languages in most bilinguals. The location of specific functional networks within this language-dominant hemisphere seems to be the result of interactions between the age of language acquisition, attained proficiency or skill in each language, the frequency with which each language is used in the environment and, critically, the particular language component measured (e.g. see Abutelalebi, Cappa, & Perani, 2005; for review). Given that both languages of bilingual adults are processed in the same hemisphere of the brain, in overlapping or adjacent areas, we can reasonably anticipate that both of these languages will be affected in similar ways in the vast majority of bilinguals with acquired aphasia (Fabbro, 2000; Obler, Centeno, & Eng, 1995; M. Paradis, 2004, for overviews). The general type of aphasia (expressive, receptive, mixed) holds for both languages. It is the case, however, that each of the bilingual’s languages could be affected to different degrees. The potential for differing levels of deficit in the languages of a bilingual adult following acquired brain damage has been a persistent theme in the bilingual aphasia literature. This issue has traditionally been referred to as “differential recovery.” 3.2.1 Patterns of Recovery or Impairment Observed in Bilingual Aphasia Historically, two questions have dominated much of the bilingual aphasia literature: Is one of the individual’s two languages disproportionately impaired or spared in aphasia? If so, which language and why? In response to this first question, several different patterns of cross-linguistic recovery or impairment have been observed in the literature. In some cases both languages are affected to a similar extent, in other cases either L1 or L2 is better preserved, and still in other cases the relative level of impairment has reportedly fluctuated between the two languages across time, as the individual’s neurological recovery progresses (e.g. Fabbro, 2000; Green, 1998; Obler et al., 1995; M. Paradis, 1997, 2004). For present purposes, patterns of preserved ability, or persisting impairment, in bilingual aphasia can be distilled into two general non-overlapping

< previous page

page_159

next page >

< previous page

page_160

next page >

Page 160 patterns of relative L1 and L2 proficiency. These patterns are parallel recovery and non-parallel recovery (cf. Centeno, 2005; Kohnert, 2008). In parallel recovery the relative level of L1 and L2 ability in aphasia is proportional to the individual’s proficiency in each language prior to the acquired brain injury. In non-parallel recovery, post-injury levels of ability in each language not only differ from each other (as is also common for most healthy, brain-intact bilinguals) but also differ from relative levels of proficiency in each language prior to aphasia onset. That is, nonparallel recovery occurs when there is a disproportionate impact on one language, as compared to previously attained proficiency levels. Despite findings across a number of studies that parallel recovery is most common, reports of non-parallel recovery have garnered more attention as researchers seek to explain why one language within the bilingual mind/brain may be more impacted than the other language (e.g. Fabbro, 2001). Reports of parallel and differential language recovery in bilingual aphasia must be considered in light of additional factors, including the amount of time that has elapsed since aphasia onset and the types of tasks used to measure skill in each language. Most studies investigating differential recovery or impairment have focused on language abilities in the first days or months following symptom onset: far fewer have investigated bilingual individuals with chronic aphasia. Most reports also consider only isolated linguistic behaviors rather than a broad spectrum of meaningful language tasks and few consider communicative performance with respect to the language needs in the post-injury environment. It is also important to consider true levels of pre-morbid proficiency in each language, rather than those that are assumed based on age or order of acquisition. Without understanding the full distribution of pre-morbid patterns of use and proficiency levels and without appropriate comparison groups for performance on selected measures, it is difficult to make precise claims about differential patterns of post-morbid recovery or persisting impairment. Another important issue in bilingual aphasia is the potential for cross-linguistic generalization. Evidence of improvement in communication functioning following treatment in monolinguals with chronic aphasia is clear (e.g. Robey, 1998). For bilinguals a question of interest is whether improvement in language functioning is confined to the language of treatment or whether a single input language will facilitate recovery in two languages. This issue is commonly referred to as cross-linguistic transfer or, in the clinical literature, cross-linguistic generalization. 3.2.2 Cross-linguistic Generalization Behavioral studies have consistently found functional connections between the two languages of proficient neurologically-intact bilingual adults (e.g. Finkbeiner, Forster, Nicol, & Nakamura, 2004; Kroll & Dijkstra, 2002). It is reasonable to ask if these connections are preserved following brain damage

< previous page

page_160

next page >

< previous page

page_161

next page >

Page 161 or if languages within the bilingual mind/brain dissociate as a consequence of the incurred damage. If crosslanguage links are preserved in aphasia, the generalization of treatment gains from a treated language, in which systematic input and training is provided, to an untreated language may be possible. Studies investigating the potential for cross-language generalization of treatment gains have produced mixed results. Some studies found that treatment in one language corresponded to improvement in another language. Junqué and colleagues investigated cross-language generalization following treatment in 30 Catalan (L1)–Spanish (L2) bilinguals with mild to severe aphasia (Junqué, Vendrell, Vendrell-Brucet, & Tobeña, 1989). Treatment was provided only in Catalan (L1). Results indicated that performance between pre- and post-treatment testing improved for both languages, although gains were greater for Catalan. These results were interpreted by investigators as evidence of cross-language generalization. However, the critical control for spontaneous recovery was not considered in this study as time post-stroke ranged from 1 to 6 months for all participants. Therefore, observed gains in the untreated language could also be attributed to a natural process of recovery rather than to direct links between the two languages. Goral, Levy, and Kastl (2007) reported limited analysis of word production and grammatical skills in a trilingual Hebrew (L1), English (L2), and French (L3) speaker with chronic aphasia. Treatment was provided only in English and focused on improving grammatical constructions and lexical access or “word finding.” The pre- and posttreatment comparison showed modest improvements in grammatical coding in English and Hebrew, with no change in French. There was no improvement in any language on the word generation probes. Authors suggested that these results indicate associations between L1 and L2 grammatical representations (Goral et al., 2007). However, no information was provided about the languages used in the man’s social environment or pre-morbid levels of functioning. Galvez and Hinckley (2003) found no evidence of cross-language generalization following naming treatment first in Spanish, then in English, with a bilingual 71-year-old man with primary acquired aphasia. Similarly, Meinzer and colleagues found no evidence of cross-language generalization in a 35-year-old German–French speaker with chronic aphasia. Treatment was administered in German; pre- and post-treatment gains in German and English were compared using a word retrieval task. Gains were restricted to the treated language (Meinzer, Obleser, Flaisch, Eulitz, & Rockstroh, 2007). Watamori and Sasanuma (1978) studied the effects of treatment in two English–Japanese bilinguals with aphasia in an alternating treatment design and found that gains were more restricted to the treated language. Edmonds and Kiran (2006) investigated the potential for cross-language generalization in three Spanish–English bilinguals with aphasia. Results indicated that cross-language generalization of treatment gains was linked to L1–L2 proficiency levels prior to aphasia onset. The participant with greater proficiency

< previous page

page_161

next page >

< previous page

page_162

next page >

Page 162 in Spanish and English prior to aphasia onset experienced greater cross-language generalization of treatment gains (Edmonds & Kiran, 2006). Kohnert (2004b) investigated the role of stimulus type on cross-language generalization in a Spanish–English bilingual with severe non-fluent aphasia. The study goal was to determine whether there would be generalization from trained to untrained items for cognates and non-cognates within and across languages. Cognates are translation equivalents that also share phonological or orthographic form (e.g. rosa–rose; elefante –elephant); noncognates are translation equivalents with little or no overlap in form ( mesa –table; caballo –horse). Following Spanish intervention, performance improved substantially for both word types. That is, there was generalization from trained to untrained test items for both types of words. However, the generalization of gains across languages, from Spanish to English, was apparent only for the cognate stimuli. Combined findings indicate that the potential for cross-language generalization in bilingual aphasia may depend on a number of factors. These factors include the similarities between the two languages spoken by the individual, the individual’s level of proficiency and patterns of use in each language prior to and following aphasia, as well as the modality and specific language component of interest. Future investigations are needed to more precisely define the parameters of generalization from one language to another, in terms of both possibilities and limitations. It is also possible that associations between the general cognitive processing system and language play a significant role in bilingual aphasia. 3.2.3 Cross-domain Associations in Bilingual Aphasia Many typical brain-intact proficient bilinguals engage in code-switching during communicative interactions. This comfortable alternation between two languages occurs at natural linguistic boundaries and is viewed as a positive and effective means of communicating among some bilinguals. Pragmatically appropriate code-switching has also been documented in bilinguals with aphasia (Muñoz, Marquardt, & Copeland, 1999). Unintentional or “pathological” language mixing is a sharp contrast to pragmatically appropriate code-switching. Unintentional language mixing appears as random, unwanted cross-language intrusions. The unintentional and unwanted mixing of two languages provides additional evidence of compromised cognitive abilities in bilinguals with aphasia (e.g. Green, 1998; Fabbro, Skrap, & Aglioti, 2000; Kohnert, 2004b). It is helpful to understand the frequency and conditions when unintentional language mixing is most likely to occur (e.g. when the individual is tired or excited, when communicating in one language or the other, on certain topics or fairly random). Of equal importance are affective factors or the individual’s response to unintentional switching. In receptive aphasias, it is possible that the individual is not aware of cross-

< previous page

page_162

next page >

< previous page

page_163

next page >

Page 163 linguistic intrusions; individuals with expressive aphasia may be very aware of the unintended language mixing and feel embarrassed or overwhelmed, which may further reduce their ability to engage in satisfying communicative interactions. Interventions may be directed at shaping and expanding the communicative intent of unintentional language switching, at shoring up the cognitive control system to reduce the occurrence of unintentional switching, as well as counseling with the individual and his or her family to reduce frustration associated with these unintentional language switches (Kohnert, 2008). With evidence of cognitive processing weaknesses in bilingual aphasia, it is reasonable to ask if treatment aimed at the general information processing system will improve one or both languages. Kohnert (2004b) investigated potential generalization across cognitive–linguistic domains with D. J., a 62-year-old bilingual man with severe nonfluent chronic aphasia (Study 1). Spanish was his L1 and English his L2. The stroke had occurred a year prior to the beginning of this cognitive treatment. The focus of treatment was on non-verbal skill performance. Specific training activities included card-sorting tasks to target perception and categorization skills; written single-digit math computations, visual number and letter searches to facilitate sustained and alternating attention, as well as several non-linguistic attention activities implemented with computer interface. No direct training was provided in Spanish or English. The critical question was whether improvement on these general non-linguistic tasks would be associated with improved performance on language tasks in Spanish and/or English. Pre- and post-treatment comparisons showed improvement on four of five language measures in Spanish. In English, there were gains on all five language measures following cognitive treatment (Kohnert, 2004b). These differences in performance for both Spanish and English following non-linguistic cognitive training are noteworthy given that the treatment reported here began a year after D. J.’s stroke. Consistent with cognitive views of aphasia, gains in Spanish and English may reflect the generalization of skill in information processing from the non-linguistic to linguistic domain. Of course, single case reports must be interpreted with great caution. Nonetheless, given the paucity of evidence directly addressing treatment outcomes in bilingual aphasia, such studies and theoretical framework should motivate future investigations. 4 Conclusions Primary language impairments are deficits or inefficiencies in the comprehension and/or production of spoken or written language. Developmental and acquired primary language impairments are chronic, high-incidence disabilities among both monolingual and bilingual populations. The presence of developmental or acquired primary language impairment may have a

< previous page

page_163

next page >

< previous page

page_164

next page >

Page 164 significant impact on the affected individual’s ability to engage in meaningful life activities. The two major types of language impairment considered here were developmental PLI and acquired aphasia. The underlying deficit that affects language processing is evident in each language of bilingual individuals. Just as with typical bilinguals, however, the relative degree of cross-linguistic proficiency may vary in those with language impairments for a variety of social and personal reasons. Individuals with developmental PLI or acquired aphasia experience a deficit in the basic “means’ for language, resulting in reduced absolute levels of language proficiency, beyond what is observed in typical bilinguals with similar backgrounds. The manifestation of language symptoms and their impact will change throughout the course of the affected individual’s life. Systematic investigations of language and cognitive functioning by diverse populations provide unique and necessary vantage points from which to consider language and lay the foundation for developing effective assessment and intervention methods. Investigations of language and non-linguistic processing in bilingual individuals with primary language impairment have the potential to inform our understanding of fundamental relationships between language, communicative opportunities and general cognitive processing mechanisms. Research which considers potential interactions between MOM, language means, opportunity and motives, across the lifespan will pave the way for more effective assessment and treatment procedures for bilinguals with language impairments. The study of developmental and acquired primary language impairments in bilingual populations is, in many ways, still in its infancy. The past decade has focused attention on both the basic and practical significance of research at the intersection of bilingualism and language impairments. Future investigations that systematically investigate crossdomain as well as cross-language relationships in bilinguals, with and without primary language impairments, will advance our clinical work and theoretical perspectives. References Abutelalebi, J., Cappa, S., & Perani, D. (2005). What can functional neuroimaging tell us about the bilingual brain? In J. F. Kroll & A. M. B. De Groot (Eds.), Handbook of bilingualism: Psycholinguistic approaches (pp. 497–515). New York: Oxford University Press. American Psychiatric Association, (1994). Diagnostic and statistical manual of mental disorders (DSM-IV) (4th ed.). Washington, DC: American Psychiatric Association. Anderson, R. T. (2004). First language loss in Spanish-speaking children: Patterns of loss and implications for clinical practice. In B. Goldstein (Ed.), Bilingual language development and disorders in Spanish–English speakers (pp. 187– 211). Baltimore: Brookes. Bedore, L., & Leonard, L. (2001). Grammatical morphology deficits in Spanish-

< previous page

page_164

next page >

< previous page

page_165

next page >

Page 165 speaking children with specific language impairment. Journal of Speech and Hearing Research , 44, 905–924. Bishop, D. V. M. (1992). The underlying nature of specific language impairment. Journal of Child Psychology and Psychiatry, 33, 3–66. Brownlie, E. B., Beitchman, J., Escobar, M., Young, A., Atkinson, L., Johnson, C., Wilson, B., & Douglas, L. (2004). Early language impairment and young adult delinquent and aggressive behavior. Journal of Abnormal Psychology, 32, 453–467. Campbell, T., Dollaghan, C., Needleman, H., & Janosky, J. (1997). Reducing bias in language assessment: Processing dependent measures. Journal of Speech, Language, and Hearing Research, 40, 519–525. Centeno, J. (2005). Working with bilingual individuals with aphasia: The case of a Spanish–English bilingual client. American Speech–Language–Hearing Association Division 14—Perspectives on Communication Disorders and Sciences in Culturally and Linguistically Diverse Populations , 12, 2–7. Chapey, R. (2001). Language intervention strategies in aphasia and related neurogenic communication disorders (4th ed.). Baltimore, MD: Williams & Wilkins. Cleave, P. L., & Rice, M. L. (1997). An examination of the morpheme BE in children with specific language impairment: The role of contractibility and grammatical form class. Journal of Speech, Language, and Hearing Research, 40, 480–492. Coelho, C. A. (2005). Direct attention training as a treatment for reading impairment in mild aphasia. Aphasiology , 19, 275–283. Connor, L. T., Spiro, A., Obler, L., & Martin, A. (2004). Change in object naming ability during adulthood. Journals of Gerontology: Series B: Psychological Sciences and Social Sciences , 59B(5), 203–209. Crary, M. A., Wertz, R. T., & Deal, J. L. (1992). Classifying aphasias: Cluster analysis of Western Aphasia Battery and Boston Diagnostic Aphasia Examination results. Aphasiology , 6 , 29–36. de Bot, K., & Makoni, S. (2005). Language and aging in multilingual contexts . Clevedon, UK: Multilingual Matters. Deary, I. J., & Der, G. (2005). Reaction time, age, and cognitive ability: Longitudinal findings from age 16 to 63 years in representative population samples. Aging, Neuropsychology, and Cognition, 12, 187–215. Ebert, K. D., Kalanek, J., Cordero, K. N., & Kohnert, K. (2008). Spanish nonword repetition: Stimuli development and preliminary results. Communication Disorders Quarterly, 29, 67–74. Edmonds, L. A., & Kiran, S. (2006). Effect of semantic naming treatment on crosslinguistic generalization in bilingual aphasia. Journal of Speech, Language, and Hearing Research, 49, 729–748. Ellis Weismer, S., & Evans, J. (2002). The role of processing limitations in early identification of specific language impairment. Topics in Language Disorders, 22, 15–29. Ellis Weismer, S., Tomblin, J. B., Zhang, X., Buchwalter, P., Chynoweth, J. G., & Jones, M. (2000). Nonword repetition performance in school-age children with and without language impairment. Journal of Speech, Language, and Hearing Research, 43, 865–878. Evans, J. (2001). An emergent account of language impairments in children with SLI: implications for assessment and intervention. Journal of Communication Disorders, 34, 39–54.

< previous page

page_165

next page >

< previous page

page_166

next page >

Page 166 Fabbro, F. (2000). The bilingual brain: Bilingual aphasia. Brain and Language, 79, 201–210. Fabbro, F. (2001). The bilingual brain: Cerebral representation of languages. Brain and Language, 79, 211–222. Fabbro, F., Skrap, M., & Aglioti, S. (2000). Pathological switching between languages after frontal lesions in a bilingual patient. Journal of Neurology, Neurosurgery, & Psychiatry, 68, 650–652. Finkbeiner, M., Forster, K., Nicol, J., & Nakamura, K. (2004). The role of polysemy in masked semantic and translation priming. Journal of Memory and Language, 51, 1–22. Fujiki, M., Brinton, B., Morgan, M., & Hart, C. H. (1999). Withdrawn and sociable behavior of children with language impairment. Language, Speech, and Hearing Services in Schools , 30, 183–195. Galvez, A., & Hinckley, J. J. (2003). Transfer patterns of naming treatment in a case of bilingual aphasia. Brain and Language, 87, 173–174. Gertner, B. L., Rice, M. L., & Hadley, P. A. (1994). Influence of communicative competence on peer preferences in a preschool classroom. Journal of Speech and Hearing Research , 37, 913–923. Girbau, D., & Schwartz, R. G. (2007). Nonword repetition in Spanish-speaking children with specific language impairment (SLI). International Journal of Language and Communication Disorders, 42, 59–75. Girolametto, L., Wiigs, M., Smyth, R, Weitzman, E., & Pearce, P. (2001). Children with a history of expressive vocabulary delay: Outcomes at 5 years of age. American Journal of Speech–Language Pathology , 10, 358–369. Goral, M., Levy, E., & Kastl, R. (2007). Cross-language treatment generalization: A case of trilingual aplasia. Brain and Language, 103 , 248–249. Green, D. W. (1998). Mental control of the bilingual lexico-semantic system. Bilingualism: Language and Cognition, 1 , 67–81. Gutiérrez-Clellen, V. (2004). Narrative development and disorders in bilingual children. In B. Goldstein (Ed.), Bilingual language development & disorders in Spanish–English speakers (pp. 235–256). Baltimore, MD: Brookes. Håkansson, G., & Nettelbladt, U. (1996). Similarities between SLI and L2 children: Evidence from the acquisition of Swedish word order. In C. E. Johnson & J. H. Gilbert (Eds.), Children’s language, Vol. 9 (pp. 135–151). Hillsdale, NJ: Lawrence Erlbaum. Håkansson, G., Salameh, E., & Nettelbladt, U. (2003). Measuring language development in bilingual children: Swedish–Arabic children with and without language impairment. Linguistics , 41, 255–288. Helm-Estabrooks, N. (2002). Cognition and aphasia: A discussion and a study. Journal of Communication Disorders, 35, 171–186. Hill, E. L. (2001). Non-specific nature of specific language impairment: A review of the literature with regard to concomitant motor impairments. International Journal of Language and Communication Disorders, 36, 149–171. Hsu, C. F., & Karmiloff-Smith, A. (in press). Language and Williams syndrome. Annual Review of Applied Linguistics . Jia, G., & Aaronson, D. (2003). A longitudinal study of Chinese children and adolescents learning English in the United States. Applied Psycholinguistics , 24, 131–161. Jia, G., Kohnert, K., Collado, J., & Aquino-Garcia, F. (2006). Action naming in

< previous page

page_166

next page >

< previous page

page_167

next page >

Page 167 Spanish and English by sequential bilingual children and adolescents. Journal of Speech, Language, and Hearing Research, 49, 588–602. Johnson, C. J., Beitchman, J., Young, A., Escobar, M., Atkinson, L., Wilson, B., Brownlie, E. B., Douglas, L., Taback, N., Lam, I., & Wang, M. (1999). Fourteen-year follow-up of children with and without speech/language impairments: Speech/language stability and outcomes. Journal of Speech, Language, and Hearing Research, 42, 744–760. Junqué, C., Vendrell, P., Vendrell-Brucet, J. M., & Tobeña, A. (1989). Differential recovery in naming in bilingual aphasics. Brain and Language, 36, 16–22. Kamhi, A., & Catts, H. (1999). Language and reading: Convergences and divergences. In H. Catts & A. Kamhi (Eds.), Language and reading disabilities (pp. 1–24). Boston: Allyn and Bacon. Kan, P. F., & Kohnert, K. (2005). Preschoolers learning Hmong and English: Lexical-semantic skills in L1 and L2. Journal of Speech, Language, and Hearing Research, 48, 1–12. Kelly, D. J. (1998). A clinical synthesis of the “late talker” literature: Implications for service delivery. Language, Speech, and Hearing Services in Schools , 29, 76–84. Kemper, S. (1987). Life span changes in syntactic complexity. Journal of Gerontology, 42, 323–328. Klee, T., Pearce, K., & Carson, D. (2000). Improving the positive value of screening for developmental language disorder. Journal of Speech, Language, and Hearing Research, 43, 821–833. Kohnert, K. (2004a). Children learning a second language: Processing skills in early sequential bilinguals. In B. Goldstein (Ed.), Bilingual language development and dis-orders in Spanish–English speakers (pp. 53–76). Baltimore: Brookes. Kohnert, K. (2004b). Cognitive and cognate-based treatments for bilingual aphasia: A case study. Brain and Language, 91, 294–302. Kohnert, K. (2008). Language disorders in bilingual children and adults . San Diego, CA: Plural. Kohnert, K., & Bates, E. (2002). Balancing bilinguals II: Lexical comprehension and cognitive processing in children learning Spanish and English. Journal of Speech, Language, and Hearing Research, 45, 347–359. Kohnert, K., & Windsor, J. (2004). In search of common ground—Part II: Non-linguistic performance by linguistically diverse learners. Journal of Speech, Language, and Hearing Research, 47, 891–903. Kohnert, K., Windsor, J., & Ebert, K. D. (in press). Primary or “specific” language impairment and children learning a second language. Brain and Language. Kohnert, K., Windsor, J., & Miller, R. (2004). Crossing borders: Recognition of Spanish words by English speaking children with and without language impairment. Journal of Applied Psycholinguistics , 25, 543–564. Kohnert, K., Windsor, J., & Yim, D. (2006). Do language-based processing tasks separate children with primary language impairment from typical bilinguals? Journal of Learning Disabilities Research and Practice, 21, 19–29. Kohnert, K., Yim, D., Nett, K., Kan, P. F., & Duran, L. (2005). Intervention with linguistically diverse preschool children: A focus on developing home language(s). Language, Speech, and Hearing Services in Schools , 36, 251– 263 . Kroll, J. F., & Dijkstra, A. (2002). The bilingual lexicon. In R. Kaplan (Ed.), Handbook of applied linguistics (pp. 301– 321). Oxford: Oxford University Press.

< previous page

page_167

next page >

< previous page

page_168

next page >

Page 168 Lambert, W. E. (1977). Effects of bilingualism on the individual: Cognitive and socio-cultural consequences. In P. A. Hornby (Ed.), Bilingualism: Psychological, social, and educational implications (pp. 15–28). New York: Academic Press. Leonard, L. (1998). Children with specific language impairment. Cambridge, MA: MIT Press. Luria, A. R. (1966). Higher cortical functions in man. New York: Basic Books. McNeil, M. R. (1988). Aphasia in adults. In N. J. Lass, L. V. McReynolds, J. L. Northern, & D. E. Yoder (Eds.), Handbook of speech–language pathology and audiology (pp. 738–786). Philadelphia, PA: W. B. Saunders. McNeil, M. R., & Pratt, S. R. (2001). Defining aphasia: Some theoretical and clinical implications of operating from a formal definition. Aphasiology , 15, 901–911. Meinzer, M., Djundja, D., Barthel, G., Elbert, T., & Rockstroh, B. (2005). Long-term stability of improved language functions in chronic aphasia after constraint-induced aphasia therapy. Stroke, 36, 1462–1466. Meinzer, M., Obleser, J., Flaisch, T., Eulitz, C., & Rockstroh, B. (2007). Recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fMRI. Neuropsychologia , 45, 1247–1256. Miller, C., Leonard, L., Kail, R., Zhang, X., Tomblin, B., & Francis, D. (2006). Response time in 14-year-olds with language impairment. Journal of Speech, Language, and Hearing Research, 49, 712–728. Montrul, S. (2005). Second language acquisition and first language loss in adult early bilinguals: Exploring some differences and similarities. Second Language Research, 21, 199–249. Muñoz, M. L., Marquardt, T. P., & Copeland, G. (1999). A comparison of codeswitching patterns of aphasic and neurologically normal bilingual speakers of English and Spanish. Brain and Language, 66, 249–274. National Aphasia Association (2008). Impact of aphasia on patients and families. Retrieved June 1, 2008 from www.aphasia.org/aphasia_community/impact_ of_aphasia_on_patients_and families.html. National Stroke Association (2003). What is a stroke/brain attack? Retrieved June 1, 2008 from www.stroke.org/site/DocServer/NSA_complete_guide.pdf?docID=341. Obler, L. K., Centeno, J., & Eng, N. (1995). Bilingual and polyglot aphasia. In L. Menn, M. O’Connor, L. K. Obler, & A. Holland (Eds.), Non-fluent aphasia in a multilingual world (pp. 132–143). Amsterdam: John Benjamins. Paradis, J., & Crago, M. (2000). Tense and temporality: A comparison between children learning a second language and children with SLI. Journal of Speech, Language, and Hearing Research, 43, 834–847. Paradis, J., Crago, M., Genesee, F., & Rice, M. (2003). Bilingual children with specific language impairment: How do they compare with their monolingual peers? Journal of Speech, Language, and Hearing Research, 46, 1–15. Paradis, M. (1997). The cognitive neuropsychology of bilingualism. In A. M. B. de Groot & J. F. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives (pp. 331–354). Mahwah, NJ: Lawrence Erlbaum. Paradis, M. (2004). A neurolinguistic theory of bilingualism . Amsterdam: John Benjamins. Paul, R. (2001). Language disorders from infancy through adolescence: Assessment & Intervention (2nd ed.). St Louis, MI: Mosby. Peña, E., Bedore, L., & Zlatic-Giunta, R. (2002). Category-generation performance

< previous page

page_168

next page >

< previous page

page_169

next page >

Page 169 of bilingual children: The influence of condition, category, and language. Journal of Speech, Language, and Hearing Research, 45, 938–947. Plomin, R., & Dale, P. S. (2000). Genetics and early language development: A UK study of twins. In D. V. M. Bishop & L. Leonard (Eds.), Speech and language impairments in children: Causes, characteristics, intervention and outcome (pp. 35–51). Hove, UK: Psychology Press. Records, N., Tomblin, J. B., & Freese, P. R. (1992). The quality of life of young adults with histories of specific language impairment. American Journal of Speech–Language Pathology , 1 , 44–53. Rescorla, L., & Roberts, J. (2002). Nominal versus verbal morpheme use in late talkers at ages 3 and 4. Journal of Speech, Language, Hearing Research, 45, 1219–1231. Restrepo, M. A., & Gutiérrez-Clellen, V. (2001). Article use in Spanish-speaking children with specific language impairment. Journal of Child Language, 28, 433–452. Robey, R. R. (1998). A meta-analysis of clinical outcomes in the treatment of aphasia. Journal of Speech, Language, and Hearing Research, 41, 172–187. Roseberry-McKibbin, C., Brice, A., & O’Hanlon, L. (2005). Serving English language learners in public school settings: A national survey. Language, Speech, and Hearing Services in Schools , 36, 48–61. Rosenbek, J. C., LaPointe, L. L., & Wertz, R. T. (1989). Aphasia: A clinical approach. Boston: Little, Brown. Salameh, E., Håkansson, G., & Nettelbladt, U. (2004). Developmental perspectives on bilingual Swedish–Arabic children with and without language impairment: A longitudinal study. International Journal of Language and Communication Disorders, 39, 65–91. Saygin, A. P., Dick, F., Wilson, S. W., Dronkers, N. F., & Bates, E., (2003). Shared neural resources for processing language and environmental sounds: Evidence from aphasia. Brain, 126, 928–945. Saygin, A. P., Wilson, S. M., Dronkers, N. F., & Bates, E. (2004). Action comprehension in aphasia: Linguistic and non-linguistic deficits and their lesion correlates. Neuropsychologia , 42, 1788–1804. Schiff-Myers, N. (1992). Considering arrested language development and language loss in the assessment of second language learners. Language, Speech, and Hearing Services in the Schools , 23, 28–33. Schmid, M. (2003). First language attrition: The methodology revised. International Journal of Bilingualism , 8 , 239– 256. Scott, C., & Windsor, J. (2000). General language performance measures in spoken and written narrative and expository discourse of school-age children with language learning disabilities. Journal of Speech, Language, and Hearing Research, 43, 324–339. Silliman, E. R., Wilkinson, L. C., & Brea-Spahn, M. R. (2004). Policy and practices imperatives for language and literacy learning. In C. A. Stone, E. R. Silliman, B. J. Ehren, & K. Apel (Eds.). Handbook of language and literacy (pp. 97–129). New York: Guilford. Stokes, S., Wong, A., Fletcher, P., & Leonard, L. (2006). Nonword repetition and sentence repetition as clinical markers of specific language impairment: The case of Cantonese. Journal of Speech, Language, and Hearing Research, 49, 219–236.

< previous page

page_169

next page >

< previous page

page_170

next page >

Page 170 Thompson, C. K., & Shapiro, L. P. (2005). Treating agrammatic aphasia within a linguistic framework: Treatment of underlying forms. Aphasiology , 19, 1021–1036. Thordardottir, E. T., & Namazi, M. (2007). Specific language impairment in French-speaking children: Beyond grammatical morphology, Journal of Speech, Language, and Hearing Research, 50, 698–715. Tomblin, J. B., Freese, P. R., & Records, N. (1992). Diagnosing specific language impairment in adults for the purpose of pedigree analysis. Journal of Speech and Hearing Research , 35, 832–843. Tomblin, J. B., Records, N. L., Buckwalter, P., Zhang, X., Smith, E., & O’Brien, M. (1997). Prevalence of specific language impairment in kindergarten children. Journal of Speech, Language, and Hearing Research, 40, 1245–1260. Tomblin, J. B., Zhang, X., Buckwalter, P., & O’Brien, M. (2003). The stability of primary language disorder: Four years after kindergarten diagnosis. Journal of Speech, Language, and Hearing Research, 46, 1283–1296. Ullman, M., & Pierpont, E. (2005). Specific language impairment is not specific to language: The procedural deficit hypothesis. Cortex , 41, 399–433. Van Mourik, M., Verschaeve, M., Boon, P., Paquin, P., & Van Harkskamp, F. (1992). Cognition in global aphasia: Indicators for therapy. Aphasiology , 6 , 491–499. Viding, E., Price, T. S., Spinath, F. M., Bishop, D. V. M., Dale, P. S., & Plomin, R. (2003). Genetic and environmental mediation of the relationship between language and nonverbal impairment in 4-year-old twins. Journal of Speech, Language, and Hearing Research, 46, 1271–1282. Watamori, T., & Sasanuma, R. J. (1978). The recovery processes of two English–Japanese bilingual aphasics. Brain and Language, 6 , 127–140. Windsor, J., & Kohnert, K. (2004). In search of common ground—Part I: Lexical performance by linguistically diverse learners. Journal of Speech, Language, and Hearing Research, 47, 877–890. Windsor, J., & Kohnert, K. (in press). Processing measures of cognitive-linguistic interactions for children at risk for language and reading disorders. In M. Mody & E. Silliman (Eds.), Language impairment and reading disability: Interactions among brain, behavior, and experience. New York: Guilford. Windsor, J., Kohnert, K., Loxtercamp, A., & Kan, P. F. (2008). Performance on non-linguistic visual tasks by children with language impairment. Applied Psycholinguistics, 29, 237–268. Windsor, J., Milbrath, R., Carney, E., & Rakowski, S. (2001). General slowing in language impairment: Methodological considerations in testing the hypothesis. Journal of Speech, Language, and Hearing Research, 44, 446–461.

< previous page

page_170

next page >

< previous page

page_171

next page >

Page 171 7 L1 ATTRITION ACROSS THE LIFESPAN Monika S. Schmid 1 Introduction Of all situations of bilingual development, first language attrition is arguably the one for which a lifespan perspective is most crucial, since it is conditioned by at least one major life event (most commonly emigration) and furthermore is a slow process where changes typically are witnessed only after years or decades. However, the very factors which make language attrition interesting from a developmental perspective and, theoretically, a prime candidate for longitudinal investigations also form the main methodological and practical obstacles. Investigations following a group of emigrants and potential attriters over a long timespan are faced with the twofold difficulty of keeping track of the same informants (sometimes over decades) and persisting in research which may only yield truly interesting findings after a considerable timespan (not to mention obtaining the necessary funding). For these reasons, there is at the present time only one truly longitudinal study of first language attrition, namely de Bot & Clyne’s investigation of German and Dutch immigrants in Australia, in which the same group of speakers was interviewed repeatedly after long gaps (de Bot & Clyne, 1989, 1994). Language attrition has been defined as “the non-pathological decrease in a language that had previously been acquired by an individual” (Köpke & Schmid, 2004, p. 5). In other words, attrition investigates the situation where a speaker (of an L1 or an L2) can no longer do something which s/he had previously been able to do, and this deterioration of a skill is not caused by insult to the brain due to illness or injury, but by a change in linguistic behavior and environment due to a severance of the contact with the community in which the language is spoken, in interaction with an increase in the use of a second language. Attrition is therefore a process in which cross-linguistic interference from the second language and lack of input from the first conspire to effect a change in language behavior or proficiency. While cross-linguistic influence is something that all bilinguals experience to some degree (e.g. Cook, 1992), the changes effected in the L1

< previous page

page_171

next page >

< previous page

page_172

next page >

Page 172 by lack of input are specific to the migrant condition, that is to life in an environment where the L1 is not generally spoken. Research on L1 attrition is therefore based on the assumption that without regular input from and use of the L1, that knowledge system may deteriorate and/or become more difficult to access. An intriguing question in this respect is whether this need for input in order to preserve linguistic knowledge varies throughout the lifespan. On the one hand, it may be possible that specific physical and neurological changes which typically occur at certain phases throughout a person’s life (such as maturational processes in adolescence or the decline of working memory in old age) bring with them a need for more input in order to maintain the language. On the other hand, there are typical events and occurrences in the life of a migrant which may be more or less favorable towards L1 attrition or maintenance, such as a steep learning curve in the L2 immediately upon emigration, or a change in linguistic environment due to retirement. At different stages in life, both the extent and the importance of input from the L1 may thus vary and have a varying effect on L1 attrition. The general bias in research of language development in general and the development of multilingualism in particular towards the very early and very late period of the lifespan is reflected in the existing research on L1 attrition. Investigations of migrants and their bilingual development typically focus on how they acquire the new language—how quickly, how successfully, and so on. It is far less common to find research on how the new situation impacts on proficiency, fluency and competence in the first language. Where attrition studies take a developmental perspective at all, they usually consider the age of the subjects at the moment of migration (“children” vs. “adults”) or the development towards the end of the lifespan, where emigrants reportedly show a tendency to revert to their L1. The “middle bit”—the longer-term development of L1 proficiency across the attritional period—and the impact that personal choices and environmental factors have upon this remains relatively uncharted territory. 2 The Front End: Age at Migration 2.1 Incomplete Acquisition and the Critical Period Hypothesis The age effect in the development of bilingualism is one of the most controversially researched topics in the area of language learning. It is often assumed that there is a “critical” or “sensitive” period in the development of human beings which favors language acquisition, although it has not been established to what degree the reasons for these favorable conditions are (neuro)biological, (neuro)cognitive or sociopsychological. In early work on the critical period it was assumed that maturation during puberty entails a biological change, referred to as “loss of plasticity” in the brain, which makes it increasingly difficult for language learners to achieve

< previous page

page_172

next page >

< previous page

page_173

next page >

Page 173 native-like performance in a second language if the onset of acquisition is after puberty (e.g. Lenneberg, 1967). Later work has questioned the premise that the effect of age of acquisition can be ascribed to irreversible neural changes caused by predominantly biological maturation processes (for an overview see Birdsong, 2006). In this context, it has been suggested that language acquisition itself may be a determining factor for the age effect: at the onset of puberty, normally developing speakers have learned, established and stored a large amount of linguistic knowledge, which then forms a tough competitor for the linguistic “newcomer” (e.g. Pallier, 2007). To what degree this critical period is of importance for language attrition is a subject that has received less (and less systematic) attention. While there is a host of studies comparing success in second language acquisition across virtually all age ranges and in a variety of different situations (see the literature overviews given by Birdsong, 2006; DeKeyser, 2003; and Hyltenstam & Abrahamsson, 2003), there is a striking lack of studies of first language attrition which compare pre- and post-puberty attriters. Where the attrition of a first language among young children is investigated at all, it is almost invariably through case studies. (Investigations of only one subject include de Bode, 1996; Isurin, 2000; Kaufman, 1992; Kaufman and Aronoff, 1991; Kravin, 1992; Ronowicz, 1999; Seliger, 1991; Turian & Altenberg, 1991; Uribe de Kellett, 2002; Vago, 1991; two subjects are investigated by R. T. Anderson, 2001; Brewer Bomar, 1982. Very few studies include more than two subjects; among them are Bolonyai, 1999, n = 6; Håkansson, 1995, n = 5; Schmitt, 2001, n = 5). The few quantitative investigations which count both pre- and post-puberty migrants among their informants do not, as a general rule, use age at migration as a predictor variable in their design (Armon-Lotem, 2000; El Aissati, 1997; Kaufman, 1998; Sharwood Smith, 1983). While it is therefore impossible to come to “hard” conclusions on whether language attrition will proceed differently if its onset is located before adolescence, the amount of attrition reported by this type of investigation almost invariably appears far more drastic than anything that studies investigating subjects who were older at the time of emigration have ever found (see Köpke & Schmid, 2004). It is regrettable that, due to the lack of quantitative studies, the rather inconsistent methodology used across the field, and the absence of studies which compare the attrition among children on the one hand and adolescents or adults on the other, this has to remain an impression, borne out quantitatively only through the findings of Ammerlaan (1996) and Pelc (2001), who did compare such groups and concluded that “age at migration” was one of the most important predictor variables. Preliminary findings from an ongoing investigation (Bylund, forthcoming) appear to corroborate this impression. With respect to language attrition the available evidence therefore suggests that the effect of the critical or sensitive period is reversed: if the onset

< previous page

page_173

next page >

< previous page

page_174

next page >

Page 174 of attrition takes place at a relatively early period in life, the deterioration of the linguistic system will be far more dramatic than if the change in linguistic environment occurs later. What precisely “relatively early” means in this context remains unclear and very hard to define. According to Köpke and Schmid (2004) the cumulative evidence suggests that a language system the deterioration of which begins before age 8 can vanish to the extent that later in life the speaker may be able to understand at most a few isolated words. First languages which have been fully learned until about age 12, on the other hand, appear stable—often surprisingly so. A further important factor with respect to attrition in general and attrition among children in particular is whether the language learning situation that the potential attriter finds him- or herself in is one of additive or of subtractive bilingualism (in the terminology of Gardner & Lambert, 1972). The studies cited above (with the exception of de Bode, 1996, and Isurin, 2000), investigate the linguistic development of children whose emigration took place within the family context. While in such cases a reversal of language dominance often occurs within a relatively short period of time, and the acquisition of the first language may be slowed down considerably, there is typically no total cessation of input. On the contrary, parents and caregivers often report making a conscious effort to transmit their heritage language, sometimes by insisting on the use of that language in the home and the family and sometimes by following the one parent, one language method. The linguistic situations investigated in this respect are therefore contexts of additive bilingualism : the speakers are exposed to the family or heritage language from birth, and then at some point in their linguistic development another linguistic system is added, either through instructed acquisition in school or through immersion. While this system can become the dominant language, it does not completely oust the earlier-learned one. Only in recent years has it been recognized that investigations of subtractive bilingualism , the situation where the new linguistic system completely replaces the old one, might provide additional insights. It is relatively rare, of course, for such a radical break in linguistic input to occur. One type of situation where exposure to the first language of a developing speaker ceases suddenly and totally, to be replaced by a different system, is that of children who are adopted into a different linguistic environment. This is the situation that de Bode (1996) and Isurin (2000) investigate in case studies of Russian children adopted by American families. Both studies report a rapid breakdown of first language proficiency, which after a relatively short period is followed by an unwillingness and refusal of the subject to interact in her first language at all with the investigator. A quantitative psycho- and neurolinguistic investigation of adult speakers of French who were adopted from a Korean-speaking background at ages 3–9 more than two decades ago (Pallier et al., 2003; Ventureyra, 2005; Ventureyra & Pallier, 2004) provides interesting insights. Eight adults were

< previous page

page_174

next page >

< previous page

page_175

next page >

Page 175 tested in a series of behavioral (sentence identification, word recognition and speech segment identification) and neurolinguistic (fMRI) experiments. The results revealed no differences between the Korean adoptees and the French controls with respect to any of the Korean stimuli1 (Pallier et al., 2003). These findings appear to suggest that the L1 attrition that takes place in cases where linguistic input ceases entirely before the onset of puberty can be far more radical than anything witnessed in other investigations of attritional situations—the first language appeared to have been erased. While exposure to an L2 before puberty may therefore facilitate its acquisition, the deterioration of an L1 may be quicker and more radical if the input is reduced or ceased during the same age span. How precisely this breakdown in linguistic competence happens, and what the relevant age span is, will hopefully become clearer as further studies are being conducted and more data becomes available. 2.2 Adolescence, Identification and Acculturation Studies of first language acquisition do not usually include speakers who have reached adolescence—the age from about 12 to 17 years old. It is assumed that the linguistic system, at least where its grammatical aspects are concerned, has stabilized before this age, and that few findings of interest to L1 acquisition can be found in investigations of this age group. On the other hand, it is generally acknowledged that the period of adolescence is characterized by the search for social identity, “that part of an individual’s self-concept which derives from his knowledge of his membership of a social group (or groups) together with the value and emotional significance attached to that membership” (Tajfel, 1981, p. 255). In this period, it is thus not the stabilization of a (linguistic) knowledge system but of the sociocognitive concept of identity and identification that impacts most heavily on linguistic development. Language is one of the predominant markers of identity: speakers adapt their language or variety to the variety spoken by the group of individuals to which they want to belong, they talk like the people they want to be like. This aspect of the development of language and identity is particularly important among adolescents, the age range where the search for group membership is prevalent (Eckert, 1988). While emigration is probably almost invariably an unsettling and to some degree difficult experience, it is arguably most difficult and traumatic for adolescents. On the one hand, these speakers have passed the age at which learning a new language through immersion may appear effortless, and they often have to work hard at achieving native-like proficiency in the new language. More importantly, however, adolescence is the age where a person ceases to define herself exclusively or even predominantly within the family context, and starts the search for integration and identification within the wider society (Eckert, 1988). At this stage, being “different” can often be felt

< previous page

page_175

next page >

< previous page

page_176

next page >

Page 176 with devastating sharpness. Eva Hoffman describes this experience in her autobiography Lost in Translation, and points out how the feeling of deviance is enhanced by reactions from the surrounding society towards what is perceived as a foreign accent (Hoffman, 1989). Schmid (2002) investigated a group of speakers for whom this situation was particularly traumatic: German Jews who during the Nazi regime were brought from Germany to the UK and the USA as adolescents (the lucky ones in the company of a sibling, most of them on their own) on the so-called Kindertransporte. All of their adult relatives were left behind, the majority of them to be deported and murdered. In this devastating situation, an added problem was the fact that they were initially recognizable as Germans through their non-native English. As one of those speakers later recalls, “the English children at the boarding-school were incredibly cruel, if you said anything at all that wasn’t totally perfect English, they’d immediately laugh at you.” After the outbreak of World War II their situation worsened in a terrible historical irony: having been deprived of their German citizenship and identity by the Nazis, they were now considered dangerous enemies by the English—because they were German. Many adolescent and adult males were confined to internment camps, others were exposed to resentment from the English population —paradoxically both because they were German (hence potential enemies or spies) and because their caregivers feared that, in the case of a German invasion, there might be repercussions for those who had given help or offered shelter to Jews. It is hardly surprising that these speakers invested every effort to shed themselves of their native language as quickly as possible (Schmid, 2002).2 The frequent wish of adolescent bilinguals to integrate linguistically into the host society is also borne out by reports from first generation emigrants who attempted to maintain their L1 as a home language and encourage their children to acquire and use it: almost invariably, they will report that this became difficult to impossible at the time that the second child reached school age, encountering more and diverse L2 contexts and L2 role models.3 At this stage, the siblings started to communicate with each other in the L2 and formed a “united L2 front” within the family. They often were reluctant to use the L1 any further or even refused this altogether. It remains unclear how this tendency to assimilate completely into the host society influences the development of L1 attrition among adolescents. Again, empirical studies which investigate the impact of age at onset of attrition are scarce for this age period, but two quantitative analyses which do make a distinction between “younger” and “older” emigrants during this age period do not find any difference: Köpke (1999) compared attrition in a group of speakers who were between 14 and 25 years at the time of emigration to that among speakers who were between 26 and 36 years, and Schmid (2002) set the groups at 11–16 and 17–28. In neither study is age of onset a significant predictor for language attrition, and all speakers investigated are

< previous page

page_176

next page >

< previous page

page_177

next page >

Page 177 able to communicate and perform a range of tasks in their L1, often after many decades of life in an L2 environment (and minimal exposure to the L1). Given the findings presented above for pre-puberty migrants, these findings again suggest a stabilization of the L1 system around the “critical period,” regardless of further language input, use or attitudes. The issues and findings presented above suggest that the age of onset of language attrition is a topic which can potentially provide converging evidence for the discussion on still controversial issues such as the critical period or social identity and identification among adolescents, particularly adolescent immigrants. The lack of quantitative studies and findings makes it difficult to come to hard conclusions. However, it is likely that a change of linguistic environment and reduction of L1 input before puberty will generally result in more drastic deterioration, while attriters for whom emigration takes place after puberty typically show surprising stability of their L1 knowledge, probably due to maturation processes. To what degree social identification processes during adolescence influence L1 maintenance and attrition will have to be assessed in more detail, preferably in studies with a cross-linguistic and cross-cultural approach to account for factors such as language prestige and demographic distribution of the heritage language community. 3 The Back End: Language Reversion and Old Age A phenomenon that appears to be often reported by older migrants themselves, their families or their caretakers, is a “reversion” in language dominance, whereby the second language which they have used in their daily lives for years or decades recedes and the first language becomes stronger again, allegedly sometimes to the point that communication with children who were not brought up to speak their parent’s first language becomes impossible. It is, however, very difficult to separate fact from fiction in such reports. References which are based on hearsay and common assumptions proliferate, and unsupported statements such as “it is common knowledge these days … that ageing is often accompanied by language reversion” (Haines, 1999) or “research and … experience clearly demonstrates that language reversion in later life is very common” (Leake, 2005, p. 4) are no exception. In other cases, the conclusion that language reversion has occurred is not based on actual observations of linguistic behavior (and comparisons to earlier behavior) but on self-reports (de Bot & Clyne, 1989). One problem with such anecdotal evidence is that it is not clear to what extent the problems reported here may not be linked to pathological changes caused by conditions such as early dementia. Research on multilingualism and pathological aging points out difficulties which patients may have with issues such as language choice, language separation and involuntary code-switching (e.g. de Bot & Makoni, 2005, pp. 66–70; Fabbro, 2001; Friedland & Miller, 1999).

< previous page

page_177

next page >

< previous page

page_178

next page >

Page 178 An interesting case in this context is the investigation of a large group of Dutch migrants in Australia reported by Clyne (e.g. 1981, 1982). In this investigation, based on data collection conducted in the early 1970s, he finds that phenomena such as switches to the L1 and loss of fluency in the L2 are more frequent among older speakers. A second data collection in 1987 (i.e., a good 15 years after the first), however, showed that proficiency in both L1 and L2 had remained surprisingly stable in the meantime. De Bot and Clyne conclude that for speakers who had reached a certain proficiency threshold in the L2, that proficiency had apparently remained stable into old age, and that in the light of these data, the reversion hypothesis could not be upheld (de Bot & Clyne, 1989, p. 167). In spite of these findings, self-reports from both the migrants and their families frequently maintained that the older speakers’ L2 had become less fluent, particularly in the case of speakers who used their L1 with the partner (de Bot & Clyne, 1989, p. 168). The premise that language reversion in healthy elderly bilinguals is a common occurrence thus stands on very shaky feet—it is based largely on anecdotal evidence and reports by older speakers and their younger interlocutors. Communication across age ranges—between “old” and “young” people—has been the focus of much attention in recent years. It has been shown that such interactional situations are often fraught with difficulty and frustrations. Younger people tend to experience their older communication partners as “underaccommodative,” “inattentive,” and “nonlistening,” and generally feel that interactions with older communication partners are less satisfying than those with same-age partners (Williams & Harwood, 2004, p. 121f.). Older people, on the other hand, often feel patronized, experiencing their younger interlocutors as “overaccommodating” in that they use overly simple language (p. 121f.). Some of these difficulties and misunderstandings are probably to some degree caused by the change in status people experience when they leave the job market, while others may indeed have to do with a change in communicative behavior by the elderly. In order to establish whether language reversion is actually a phenomenon that is as common as it is taken to be, it would therefore first have to be established to what degree problems reported by (otherwise healthy) elderly migrants and those in close contact with them are more frequent or more serious than those experienced in monolingual cross-generational interaction. Situations where communication with a close family member come to be perceived as problematic are threatening to both interlocutors, and it is very possible that at least some such problems are chalked down to the convenient and ubiquitous myth of language reversion, as this relieves the participants of responsibility: it is “just” a language problem. This impression may be furthered by a more general development common among the elderly. There is a large body of research demonstrating that one cognitive process that tends to become less efficient in old age is the inhibition of competing information in memory retrieval (Burke, 1997).

< previous page

page_178

next page >

< previous page

page_179

next page >

Page 179 While access to relevant information may not be lost or even impaired, it may get more difficult for older adults to suppress irrelevant information, and this difficulty is then interpreted as “becoming forgetful.” Inhibition is a process that bilinguals routinely depend on in their interactions: every time a bilingual accesses a word (or a language), the corresponding word in his or her other language (or the other linguistic system) has to be inhibited (Green, 1998; Paradis, 2004). If this mechanism does become impaired, it could lead to increased involuntary code-switching (as reported, for example, among the elderly bilinguals investigated by de Bot & Clyne, 1989). While in such cases the underlying linguistic systems and knowledge may well remain quite intact, and it is only selective access of one system that has become slightly impaired, the surface phenomenon of increased code-switching (and perhaps ensuing linguistic insecurity and searching for words) may well be interpreted by the speaker and/or his interlocutors as an indication of language reversion. Interestingly, there is some evidence to suggest that the cognitive aging effect on the effectiveness of inhibition is, to some degree, delayed or counterbalanced in active bilinguals (e.g. Bialystok, Craik, & Ryan, 2006; note, however, that Bialystok et al.’s findings were criticized and to some extent refuted by Colzato et al., 2008). This suggests that investigations into language reversions will have to choose their subjects and parameters with great care, since a number of confounding factors may conspire to bring about unexpected developments. All in all, while language attrition and language reversion in old age appear to be the situations that are surrounded by the most persisting myths, and which are potentially most confusing and disturbing to both the speaker and those closest to her, they are also the situations where there is least empirical research. It seems critical to investigate these claims in more detail in order to provide answers and advice to those who find themselves faced with such communicative difficulties. 4 The Middle Bit Identifying a time-frame for the attritional process is a complex matter. While de Bot & Clyne, based on their longitudinal findings, propose that attrition probably occurs largely within the first decade after emigration (de Bot & Clyne, 1994, p. 27), a more recent investigation of German migrants in the Netherlands and Canada with a migration span of between 15 and 58 years (mean 35.9 years) did reveal an impact of length of residence on language performance, implying that changes may take place after the initial decade (Schmid & Dusseldorp, forthcoming). There therefore clearly had been some change in the decades between emigration (which typically takes place when speakers are in their 20s) and old age. However, we are not even close to the ability to either chart the attritional process across middle age or

< previous page

page_179

next page >

< previous page

page_180

next page >

Page 180 to identify the factors and events that may contribute to either attrition or maintenance of an L1. The language choices that people make across this period are particularly interesting, as is the question of what will happen to the linguistic system that, until the moment of emigration, had enjoyed the uncontested state of being not only the first but also the dominant (if not the only) language. This system will now be vying for space with an L2 that may come to be used more frequently—at work, with friends, or at home. How will the choices that an immigrant makes in such a situation—the decision to integrate completely and adopt the new language even in the family, the decision to apply a strictly L1-only policy at home, or any road between these two extremes—affect the maintenance of and competence in the L1 after a long stretch of emigration? Does it help to work in a profession where use of the L1 is part of the job? Does it help to have a partner with whom one shares the same native language? Is it a bonus to be highly educated? Or is it all determined by the attitudes that people may have towards one or the other of their languages, and the identities that come with them? Any large-scale study of language attrition across a given population will reveal a substantial amount of interpersonal variation. There are well-documented cases of individuals who, after many decades of emigration, retain apparently perfect native speaker competence, whereas the speech of others contains a large amount of interferences on all linguistic levels and appears substantially reduced and/or simplified (e.g. Schmid, 2002). A number of explanatory factors have tentatively been proposed to account for this variation, ranging from the amount of exposure to and use of L1 through factors linked with attitude and motivation to the highly elusive concept of language attitude (see Schmid & Dusseldorp, forthcoming). 4.1 Language Use Of these factors, language use appears to be the most intuitively obvious one. It certainly is the one that is invoked as an important predictor by the majority of studies on first language attrition. Perhaps it is because this factor appears so intuitively obvious that it has so infrequently been used as an actual predictor in empirical investigations (for an overview see Schmid, 2007). In the case of Dutch migrants in New Zealand, a relationship between L1 use and L1 lexical access was demonstrated by Hulsen (2000), who found that the amount of L1 contacts speakers had, particularly in their primary network (those contacts who are more important in everyday life), correlated strongly with speed and accuracy on a naming task. Recent findings, however, call into question the simple “use it or lose it” prediction, as other investigations have failed to find substantial or even significant links between performance on a variety of linguistic tests and selfreported L1 use in a number of different everyday situations (Schmid, 2007;

< previous page

page_180

next page >

< previous page

page_181

next page >

Page 181 Schmid & Dusseldorp, forthcoming). The only variable linked to the use of the first language which appeared to have a substantial and consistent effect on L1 performance in these investigations is the use of the L1 for professional purposes. Regular L1 use in more informal settings (such as with the partner, children or friends) did not appear to be an important or reliable safeguard against language attrition, and passive exposure to the L1 (through TV, books or newsmedia) did not play any role whatsoever (for details of the investigation, see Schmid & Dusseldorp, forthcoming). It is very possible that language use is determined by the personal situation to such a highly diverse degree that quantitative research designs are unable to account for the interactions. It would therefore be desirable to conduct more qualitative investigations, taking into account reported language use and its change over time. Immigrants often report that in the first years after arrival in the new country they continued speaking the L1 with their partner, but that at some point they made the switch to the new language. It should be asked how, why and when this change in language use has occurred, and how language attrition is affected by events such as a divorce or a new relationship with a partner with a different first language. Such life events are, of course, common among immigrants, as is the situation that the decision to emigrate is itself taken for the sake of a partner who lives in the new country and speaks the language of that country. The Major Life Event approach (see Schrauf, this volume) may provide a promising framework for such investigations. A further important factor that should be investigated with regard to the family situation is whether transmission of the L1 to the next generation is an option for the speaker. There is not a single study on language attrition which considers whether or not an informant has children as a predictor. However, it is very possible that for some speakers the desire to pass their L1 on to the next generation might be an important reason for consciously attempting to maintain it. On the other hand, as was reported above, most immigrants who do have more than one child find it almost impossible to maintain their L1 as a family language after the oldest two children have reached school age. It may therefore also be a factor whether the attriter has one or several children. And what is the role of grandparents and grandchildren in this respect?4 All of these factors are linked to language use and the perceived importance of language maintenance in a number of intricate ways. It appears reasonable to assume that these interactions cannot be captured or easily quantified in empirical research. While it was argued with respect to the factors named above that large-scale quantitative studies should be conducted, it therefore seems necessary that the factors considered here should first be assessed from a qualitative perspective in a sociopsychological or sociocultural framework before any more quantitative work can be undertaken.

< previous page

page_181

next page >

< previous page

page_182

next page >

Page 182 4.2 Attitudes It has long been established that attitudes are one of the most important determining factors for ultimate attainment in second language acquisition—essentially it seems to be the case that if a speaker has a positive and integrative orientation towards the L2, SLA will be more successful (e.g. Gardner & MacIntyre, 1992; Krashen, 1981). Since motivation can have a strong impact on the development of one of a bilingual’s language systems, language attrition research has long been fascinated with the idea that the converse might also be true: if wanting to learn something (or wanting to be part of a community of speakers) can help with acquisition, can wanting to forget something (or wanting to no longer be part of a speech community) help forget it? The neuropsychological process of suppressing unwanted (or traumatic) memories and words has been established (e.g. M. C. Anderson et al., 2004), but can it also apply with respect to a linguistic system, that is, an abstract, interconnected knowledge system that has become very deeply entrenched during the process of L1 acquisition and pre-migration life? One of the first larger studies that attempted to answer this question was Waas’s investigation on the L1 attrition of German in Australia (Waas, 1996), which applied a two-way classification of identity along two criteria: residential status (“permanent resident” vs. “naturalized citizen”) and “ethnic affiliation,” measured by systematic exposure to German in daily life. This classification did not yield significant results when correlated with performance on a linguistic test. A more rigorous framework of ethnic identity is applied by Yagğ mur (1997) and Hulsen (2000), who apply the theory of Ethnolinguistic Vitality Theory, as developed among others by Giles, Bourhis, and Taylor (1977) in their respective investigations of Turkish in Australia and Dutch in New Zealand. Here, too, no significant impact of the predictor variables of Subjective Ethnolinguistic Vitality could be established. Recently, a large-scale investigation which applied a sociolinguistics questionnaire and an adapted version of Gardner’s (1985) Attitude and Motivational Test Battery (Schmid & Dusseldorp, forthcoming) has returned an unexpected and puzzling finding: in the data under investigation here (elicited from German L1 speakers in Anglophone Canada and the Netherlands with a variety of tests) a consistent (albeit weak) impact of reported attitudes towards the L1 was indeed found. However, strongly positive attitudes correlated with lower scores on the linguistic tasks under investigation; suggesting that for the sample under investigation here, a positive attitude actually had a negative impact on L1 maintenance. The findings from these three studies suggest that socio- and ethnolinguistic frameworks that rely on subjects’ selfreports about their attitude towards their L1 may be inadequate tools for the prediction of the linguistic aspects of language attrition. This may conceivably be down to the fact that people’s identity, affiliation, and self-concept does not remain stable across the life-

< previous page

page_182

next page >

< previous page

page_183

next page >

Page 183 span (Breakwell, 1986, p. 18). If a connection between attitudes or identities and attrition is to be found, it would have to be looked for at the moment of emigration, i.e. at the possible onset of attrition. Furthermore, all of the studies invoked above investigate migrants of only one ethnic background. While attitudes may vary individually across a community of immigrants, they will also to some degree be consistent across the group. More importantly, while groups consider themselves as stratified, they are often viewed as homogenous from the outside (Breakwell, 1983, p. 191). Thus, while an individual might have perceived herself as belonging to a social class with a certain social prestige while still in the country of her origin, after emigration she may find attitudes towards her influenced by stereotyped notions associated with “immigrants” or “immigrants of a certain ethnolinguistic background.” It is therefore important that studies of the impact of language attitude on attrition attempt to make a systematic comparison of groups of different backgrounds, or compare the same groups in different settings (as was done in the case of the German speakers in Canada and the Netherlands investigated by Schmid & Dusseldorp, forthcoming) in order to assess the impact of varying attitudes, prestige, and institutional support. These points were recognized in Schmid’s investigation of language attrition among German Jews who had emigrated from Germany during the Nazi regime (Schmid, 2002). Based on a historical classification of progressive radicalization of persecution during those years, Schmid investigated the impact of the period during which emigration had taken place, hypothesizing that exposure to increasingly traumatic events would lead to a greater wish among the refugees for disassociation from Germany and the German language. This classification proved to be highly relevant in explaining the outcomes of the linguistic investigation. These findings suggest that the attitude at the time of migration (which in this case was determined by the traumatic events which are on historical record) does indeed have a strong impact on language maintenance and attrition. 5 Conclusion The degree to which migrants maintain or lose their first language can vary considerably, both across and within migrant communities. One of the most important factors in this appears to be age at emigration. Interestingly, there does not seem to be a linear age gradient here, where “older” is “better,” but a kind of cut-off point around puberty. Speakers who emigrate before that age tend never to fully master their L1 and, later in life, not to use it in a manner that is fully native-like. Older speakers, on the other hand, tend to preserve astonishing levels of L1 proficiency, often after decades of little or no exposure to it. Moreover, after puberty, age at emigration does not seem to be a good predicting factor for L1 attrition or maintenance.

< previous page

page_183

next page >

< previous page

page_184

next page >

Page 184 The fact that quantitative studies of L1 attrition typically find a great deal of variance in the degree of maintenance of L1 skills among migrants who have an otherwise comparable background with respect to factors such as age at emigration, length of residence, education, and so on, strongly suggests that L1 attrition must be conditioned by some environmental factors. Possible candidates for these predictors are amount of use of and exposure to the L1 or the attitude which the migrant has towards the language and culture of her country of origin and/or her country of residence. So far the findings from quantitative investigations of L1 attrition and external factors appear somewhat unsatisfying, and often contradictory. This might be taken to suggest that the impact that environmental factors have for L1 maintenance can vary considerably among individuals, and can therefore not be reduced to a simple “language maintenance formula” on the basis of corpus-based, quantitative research. There can be no doubt that the process of migration in virtually all cases constitutes a major turning point or “life event,” and that its consequences are massive and unforeseeable. Where one speaker may experience it as traumatic, and spend decades feeling cut off from her roots and homesick, another may find it the most liberating and positive thing that could possibly have happened. More in-depth research will have to be conducted into how these and other migrant experiences interact and determine language maintenance vs. language attrition. Notes 1. The only difference between the Korean and the French group in these experiments was that with respect to stimuli presented in some languages (Japanese and Thai) the Koreans were better at indicating that these languages were not Korean. 2. It should be pointed out, however, that at the time of the investigation in the late 1990s, all speakers were still able to communicate and interact in German. While some had retained a more fluent and target-like command than others, this was suggested for the younger migrants investigated by Pallier et al. 3. The situation may be different in cases where there is a very strong L2 community, as has been reported, for example, with respect to the Turkish and Moroccan immigrant communities in the Netherlands. In such cases, demographic factors may help the heritage language to achieve sufficient prestige to make it appear desirable to maintain (e.g. Nortier, 2002). 4. It has been shown that within most immigrant situations, the use of the L1 gradually declines across generations, and that L1 maintenance is highest in interaction with grandparents (e.g. Clyne, 1981; Hulsen, de Bot, & Weltens, 1999; Schmid, 2002). References Ammerlaan, T. (1996). “You get a bit wobbly …” Exploring bilingual lexical retrieval processes in the context of First Language Attrition. Unpublished doctoral dissertation, Katholieke Universiteit Nijmegen.

< previous page

page_184

next page >

< previous page

page_185

next page >

Page 185 Anderson, M. C., Ochsner, K. N., Kuhl, B., Cooper, J., Robertson, E., Gabrieli, S. W., Glover, G. H., & Gabrieli, J. D. E. (2004). Neural systems underlying the suppression of unwanted memories. Science , 303 (5655), 232–235. Anderson, R. T. (2001). Lexical morphology and verb use in child first language loss: A preliminary case study investigation. International Journal of Bilingualism , 5 (4), 377–402. Armon Lotem, S. (2000). Language attrition: Why are resumptive pronouns so susceptible? In S. C. Howell, S. Fish, and K. Thea (Eds.), Proceedings of the 24th Annual Boston University Conference on Language Development (pp. 58–67). Somerville, MA: Cascadilla. Bialystok, E., Craik, F., & Ryan, J. (2006). Executive control in a modified antisaccade task: Effects of aging and bilingualism. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32(6), 1341–1354. Birdsong, D. (2006). Age and second language acquisition and processing: A selective overview. Language Learning , 56 (1), 9–49. Bolonyai, A. (1999). The hidden dimensions of language contact: The case of Hungarian–English bilingual children. Unpublished PhD dissertation, University of South Carolina. Breakwell, G. M. (1983). Identities and conflicts. In G. M. Breakwell (Ed.), Threatened identities (pp. 189–213). New York: John Wiley. Breakwell, G. M. (1986). Coping with threatened identities . London/New York: Routledge. Brewer Bomar, K.E. (1982). Second language lexical and syntactical interference on the first language of two four year old Spanish speakers. Dissertation Abstracts International (DAI), 42(12): 5105A. Burke, D. M. (1997). Language, aging, and inhibitory deficits: evaluation of a theory. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 52(6), 254–264. Bylund, Emanuel (forthcoming). Effects of age of L2 acquisition on L1 event conceptualization patterns. Bilingualism: Language and Cognition. Clyne, M. (1981). Second language attrition and first language reversion among elderly bilinguals in Australia. In W. Heid & K. Heller (Eds.), Sprachkontakt als Ursache von Veränderungen der Sprach- und Bewuβtseinsstruktur: eine Sammlung von Studien zur sprachlichen Interferenz (pp. 25–32). Innsbruck: Institut für Sprachwissenschaft. Clyne, M. (1982). Multilingual Australia . Melbourne: River Seine Publishers. Colzato, L. S., Bajo, M. T., van den Wildenberg, W., Paoliery, D., Nieuwenhuis, S., La Heij, W., & Hommel, B. (2008). How does bilingualism improve executive control? A comparison of active and reactive inhibition mechanisms. Journal of Experimental Psychology: Learning, Memory, and Cognition, 34(2), 302–312. Cook, V. J. (1992). Evidence for multi-competence. Language Learning , 42(4), 557–591. de Bode, S. (1996). First language attrition: Productive morphology disintegration and neurobiological support. A case study . Unpublished MA thesis, California State Polytechnic University, Ponoma, CA. de Bot, K., & Clyne, M. (1989). Language reversion revisited. Studies in Second Language Acquisition , 11, 167–177. de Bot, K., & Clyne, M. (1994). A 16 year longitudinal study of language attrition in

< previous page

page_185

next page >

< previous page

page_186

next page >

Page 186 Dutch immigrants in Australia. Journal of Multilingual and Multicultural Development , 15(1), 17–28. de Bot, K., & Makoni, S. (2005). Language and aging in multilingual contexts . Clevedon: Multilingual Matters. DeKeyser, R. (2003). The robustness of critical period effects in second language acquisition. Studies in Second Language Acquisition , 22, 499–533. Eckert, P. (1988). Adolescent social structure and the spread of linguistic change. Language in Society, 17, 183–207. El Aissati, A. (1997). Language loss among native speakers of Morrocan [sic] Arabic in the Netherlands . Tilburg: Tilburg University Press. Fabbro, F. (2001). The bilingual brain: Bilingual aphasia. Brain and Language, 79, 201–210. Friedland, D., & Miller, N. (1999). Language mixing in bilingual speakers with Alzheimer’s dementia: A conversation analysis approach. Aphasiology , 13(4–5), 427–444. Gardner, R. C. (1985). The attitude/motivation test battery: Technical report. Retrieved on 6 January, 2009 from http://publish.uwo.ca/~gardner/. Gardner, R. C., & Lambert, W. (1972). Attitudes and motivation in second-language learning . Rowley, MA: Newbury House. Gardner, R. C., & MacIntyre, P. D. (1992). On the measurement of affective variables in second language learning. Language Learning , 43(2), 157–194. Giles, H., Bourhis, R., & Taylor, D. (1977). Towards a theory of language in ethnic group relations. In Howard Giles (Ed.), Language, ethnicity and intergroup relations (pp. 307–348). London: Academic Press. Green, D. (1998). Mental control of the bilingual lexico-semantic system. Bilingualism: Language and Cognition, 1 , 67–81. Haines, Timothy. (1999). STTARS & Refugee resettlement in South Australia. Retrieved on May 13, 2008, from http://international.metropolis.net/events/washington/Haines.doc. Håkansson, G. (1995). Syntax and morphology in language attrition: A study of five bilingual expatriate Swedes. International Journal of Applied Linguistics , 5 (2), 153–171. Hulsen, M. (2000). Language loss and language processing. Three generations of Dutch migrants in New Zealand . Unpublished PhD dissertation, Katholike Universiteit Nijmegen. Hulsen, M., de Bot, K., & Weltens, B. (1999). Language shift, language loss and language processing: An investigation of three generations of Dutch immigrants in New Zealand. In E. Huls and B. Weltens (Eds.), Artikelen van de Derde sociolinguistische Conferentie (pp. 221–32). Delft: Eburon. Hyltenstam, K., & Abrahamsson, N. (2003). Who can become native-like in a second language? All, some or none? On the maturational constraint controversy in second language acquisition. Studia Linguistica, 54(2), 150–166. Isurin, L. (2000). Deserted island, or, a child’s first language forgetting. Bilingualism: Language and Cognition, 3 (2), 151–66. Kaufman, D. (1992). First language attrition as a creative interplay between two languages. Unpublished PhD dissertation, SUNY at Stony Brook. Kaufman, D. (1998). Children’s assimilatory patterns and L1 attrition. In A. Greenhill, M. Hughes, H. Littlefield, & H. Walsh (Eds.), Proceedings of the 22nd Annual

< previous page

page_186

next page >

< previous page

page_187

next page >

Page 187 Boston University Conference on Language Development, I II (pp. 409–420). Somerville, MA: Cascadilla. Kaufman, D., & Aronoff, M. (1991). Morphological disintegration and reconstruction in first language attrition. In H. W. Seliger and R. M. Vago (Eds.), First language attrition (pp. 175–188). Cambridge: Cambridge University Press. Köpke, B. (1999). L’attrition de la première language chez le bilingue tardif: Implications pour l’étude psycholinguistique du bilinguisme. Unpublished PhD dissertation, Université de Toulouse-Le Mirail. Köpke, B., & Schmid, M. S. (2004) First language attrition: The next phase. In M. S. Schmid, B. Köpke, M. Keijzer, & L. Weilemar (Eds.), First language attrition: Interdisciplinary perspectives on methodological issues (pp. 1–43). Amsterdam: John Benjamins. Krashen, S. D. 1981. Second language acquisition and second language learning . New York: Pergamon. Kravin, H. (1992). Erosion of a language in bilingual development. Journal of Multilingual and Multicultural Development , 13, 307–325. Leake, S. (2005). Kalimera—host home respite for Greek people with dementia. Retrieved on May 13, 2008, from www.alzheimers.org.au/upload/ KalimeraLeake.pdf. Lenneberg, E. H. (1967). Biological foundations of language. New York: John Wiley. Nortier, J. M. (2002). “Fawaka, what’s up?” Language use among adolescents in monoethnic and ethnically mixed groups. In J. M. Nortier & A. Hvenekilde (Eds.), Meetings at the crossroads. Studies of multilingualism and multiculturalism in Oslo and Utrecht (pp. 61–73). Oslo, Norway: Novus Forlag. Pallier, C. (2007). Critical periods in language acquisition and language attrition. In B. Köpke, M. S. Schmid, M. Keijzer, & S. Dostert (Eds.), Language attrition: Theoretical perspectives (pp. 155–168). Amsterdam/Philadelphia: John Benjamins. Pallier, C., Dehaene, S., Poline, J.-B., LeBihan, D., Argenti, A.-M., Dupoux, E., & Mehler, J. (2003). Brain imaging of language plasticity in adopted adults: Can a second language replace the first? Cerebral Cortex , 13, 155–161. Paradis, M. (2004). A neurolinguistic theory of bilingualism . Amsterdam: John Benjamins. Pelc, L. (2001). L1 lexical, morphological and morphosyntactic attrition in Greek–English bilinguals . Unpublished PhD dissertation, CUNY, New York. Ronowicz, E. (1999). Methodological issues of testing language attrition in children in a natural bilingual environment. Poznan Studies in Contemporary Linguistics (PSiCL), 35, 105–17. Schmid, M.S. (2002). First language attrition, use and maintenance: The case of German Jews in Anglophone countries . Amsterdam/Philadelphia: John Benjamins. Schmid, M. S. (2007). The role of L1 use for L1 attrition. In B. Köpke, M. S. Schmid, M. Keijzer, & S. Dostert (Eds.), Language attrition: Theoretical perspectives (pp. 135–153). Amsterdam/Philadelphia: John Benjamins. Schmid, M. S., & Dusseldorp, E. (forthcoming). Quantitative analyses in a multivariate study of language attrition: The impact of extralinguistic factors. Second Language Research. Schmitt, E. (2001). Beneath the surface: Signs of language attrition in immigrant children from Russia. Unpublished PhD dissertation, University of South Carolina. Seliger, H. W. (1991). Language attrition, reduced redundancy, and creativity. In

< previous page

page_187

next page >

< previous page

page_188

next page >

Page 188 H. W. Seliger & R. M. Vago (Eds.), First language attrition (pp. 227–240). Cambridge: Cambridge University Press. Sharwood Smith, M. (1983). On first language loss in the second language acquirer. In S. Gass and L. Selinker (Eds.), Language transfer in language learning (pp. 222–231). Rowley, MA: Newbury House. Tajfel, H. (1981). Human groups and social categories. Cambridge: Cambridge University Press. Turian, D., & Altenberg, E. P. (1991). Compensatory strategies of child first language attrition. In H. W. Seliger & R. M. Vago (Eds.), First language attrition (pp. 207–226). Cambridge: Cambridge University Press. Uribe de Kellett, A. (2002). The recovery of a first language: A case study of an English/Spanish bilingual child. International Journal of Bilingual Education and Bilingualism , 5 (3), 162–81. Vago, R. M. (1991). Paradigmatic regularity in first language attrition. In H. W. Seliger & R. M. Vago (Eds.), First language attrition (pp. 241–252). Cambridge: Cambridge University Press. Ventureyra, V. (2005). A la recherche de la langue perdue: Etude psycholinguistique de l’attrition de la première langue chez des coréens adoptés en France . Unpublished PhD dissertation, Ecole des Hautes Etudes en Sciences Sociales, Paris. Ventureyra, V., & Pallier, C. (2004). In search of the lost language: The case of adopted Koreans in France. In M. S. Schmid, B. Köpke, M. Keijzer, & L. Weilemar (Eds.), First language attrition: Interdisciplinary perspectives on methodological issues (pp. 207–221). Amsterdam: John Benjamins. Waas, M. (1996). Language attrition downunder: German speakers in Australia . Frankfurt/New York: Peter Lang. Williams, A., & Harwood, J. (2004). An intergroup perspective on intergenerational relationships in the family. In J. F. Nussbaum & J. Coupland (Eds.), Handbook of communication and aging research (pp. 115–138). Mahwah, NJ: Lawrence Erlbaum. Yagğmur, K. (1997). First language attrition among Turkish speakers in Sydney. Tilburg, Netherlands: Tilburg University Press.

< previous page

page_188

next page >

< previous page

page_189

next page >

Page 189 Part III NON-VERBAL ASPECTS OF LANGUAGE DEVELOPMENT

< previous page

page_189

next page >

< previous page

page_190

next page >

page_190

next page >

Page 190 This page intentionally left blank.

< previous page

< previous page

page_191

next page >

Page 191 8 THE DEVELOPMENT OF GESTURE Marion Tellier Human beings gesture every day while speaking: they move their hands, their heads, their arms; their whole body is involved in communication. But how does it work? How do we produce gestures and in what purpose? How are gestures connected to speech? When do we begin producing gestures and how do they evolve throughout the lifespan? These are questions gesture researchers have been trying to answer since the second half of the 20th century. This chapter will first define what a gesture is by describing the different kinds of gestures and by explaining some of the current theories about gesture production. Then the emergence of gesture along with language development will be exposed as well as its evolution during childhood. Finally, we will review studies about adults’ gestures and what we know about their change across adulthood. 1 What Is a Gesture? At first, gesture may seem easy to define: a movement of the hand or maybe of both hands produced by a human being. However, when one thinks more precisely about it, one may wonder if a gesture is only performed with hands or if it can involve other body parts such as head, face or arms. One can also wonder if there are different kinds of gestures: are nervous scratches, gestures accompanying speech and gestures used in deaf sign language the same kind of movements? Indeed, even if these are all called gestures, they differ. This first section will give a brief overview of the various types of gestures and of the main issues in gesture studies. 1.1 What Is a Communicative Gesture and What Is Not If we look at two persons involved in a face-to-face interaction, we will notice that they move their bodies continuously. One of the participants may be performing practical actions such as taking notes, smoking, driving, etc.: these are not considered as communicative gestures. Similarly non-

< previous page

page_191

next page >

< previous page

page_192

next page >

Page 192 verbal behavior such as postures, crossing the legs and nervous gestures like scratching, playing with an object, stroking the hair are not regarded as communicative gestures either. Nevertheless, these movements can have an impact on the interaction. For instance, if one nervously plays with a pen or taps on the table with their fingers, their addressee might end the conversation earlier than planned. Thus, as Kendon puts it (2004, p. 8) “usually ‘gesture’ is not used to refer to those visible bodily expressions of thoughts or feelings that are deemed inadvertent or are regarded as something a person cannot ‘help’ ”. A gesture is rather considered as intended to communicate something. Now that we have put aside what movements are not regarded as gestures, we are left with the idea that a gesture is an action related to ongoing talk and that has the features of manifest deliberate expressiveness (Kendon, 2004). This includes a whole range of movements such as a thumb up for OK, a finger pointing to a place or an object, and even a gesture of sign language. Researchers have proposed several classifications of these gestures in order to differentiate them. Classifications can rely on semiotic or functional distinctions, sometimes a mix of both (for an overview of the various classifications, see Kendon, 2004, Chapter 6). A very efficient and practical classification is called Kendon’s Continuum and will be detailed below for it is nowadays commonly used to explain the different kinds of gestures. 1.2 Kendon’s Continuum Based on Adam Kendon’s work, Kendon’s Continuum has been elaborated by David McNeill (1992 and 2000). McNeill placed four kinds of gestures on a continuum: gesticulation, pantomime, emblems, and sign language. Gesticulation refers to “idiosyncratic spontaneous movements of the hands and arms accompanying speech” (McNeill, 1992, p. 37); they are also called cospeech gestures and will be detailed below. Pantomime is used to define those gestures that mime an action or an object, a profession, etc., and that are mainly used when it is impossible to speak (because of the noise, distance, need to be discreet …) or in games of miming. Emblems are conventionalized gestures used in a specific community; they have a defined meaning. For instance, the thumb up meaning OK in some countries such as the USA, or the forefinger pulling down the skin under the eye, which means in the French culture: I don’t believe it (“Mon oeil ”). Emblems are most of the time associated with a fixed expression but can be used without speech. People belonging to the same cultural community understand these gestures because they have learned them along with their first language. Eventually, sign languages are “full-fledged linguistic systems with segmentation, compositionality, a lexicon, a syntax, distinctiveness, arbitrariness, standards of well-formedness, and a community of users” (McNeill, 1992, p. 38). Indeed, sign languages (no matter if they are languages used by the deaf or ritual and

< previous page

page_192

next page >

< previous page

page_193

next page >

Page 193 cultural languages used by the North American Plain Indians or by Central Australia Aborigines, for instance) are languages of their own and are mainly used without speech. Originally, McNeill (1992) organized these four kinds of gestures on a continuum according to their link to speech and to their degree of convention (see Table 8.1). Thus, on the left-hand side, gesticulation is made of gestures that require the presence of speech, whereas on the right-hand side, sign languages are used without speech. On the left-hand side, gesticulation is made of spontaneous idiosyncratic gestures, and on the right extremity, sign languages are strongly conventionalized and socially regulated signs. Table 8.1 McNeill’s Continuum Gesticulation  Pantomime  Emblems  Sign languages  Obligatory presence of speech absence of speech  ––––––––––––––––––– Not conventionalized conventionalized ––––––––––––––––––– In 2000, McNeill enriched this continuum by dividing it into four continua by using the original characteristics “relationship to speech” and “relationship to conventions” and by adding other characteristics such as “relationship to linguistic properties” and “character of the semiosis.” See Continua 1, 2, 3 and 4. Continuum 8.1 Relationship to Speech Gesticulation  Emblems Pantomime Sign Language Obligatory presence of Optional presence of Obligatory absence of speech Obligatory absence of speech speech speech Continuum 8.2 Relationship to Linguistic Properties Gesticulation  Pantomime  Emblems  Sign Language  Linguistic properties absent Linguistic properties Some linguistic properties Linguistic properties present absent present Continuum 8.3 Relationship to Conventions Gesticulation  Pantomime  Emblems  Sign Language  Not conventionalized Not conventionalized Partly conventionalized Fully conventionalized

< previous page

page_193

next page >

< previous page

page_194

next page >

Page 194 Continuum 8.4 Character of the Semiosis Gesticulation  Pantomime  Emblems  Sign Language  Global and synthetic Global and analytic Segmented and synthetic Segmented and analytic 1.3 Types of Co-speech Gestures Many gesture researchers have decided to focus on the study of co-speech gestures also called “gesticulation” in Kendon’s Continuum. They are movements of the hands and arms produced by people when they talk. They do not belong to a fixed repertoire as gestures of sign language, for instance; on the contrary, they are unique, personal and spontaneous. As mentioned before, there are several classifications of gestures; most of them descend from Efron’s (1972 [1941]), such as Ekman and Friesen’s (1969). Although these are relevant and fine classifications, they are extremely detailed and not always easy to use. That is why David McNeill and his team (1992) have worked on a simplified, easy-to-use scheme made of four categories: iconic, metaphoric, deictic and beats. Iconic gestures bear a close formal relationship to the semantic content of speech (McNeill, 1992). For instance, someone may say, “I was driving when I heard the news on the radio,” and mime holding a steering wheel while saying “drive,” or someone may say, “It was as big as that,” while showing a width with both hands open and facing. “Most of the time, iconics represent body movements, movements of objects or people in space, and shapes of objects or people. They do so concretely and relatively transparently” (Goldin-Meadow, 2003, p. 7). Metaphoric gestures are very similar to iconics except that they depict abstract concepts rather that concrete objects. If one cups their hands when saying the word “concept,” for instance, it is a metaphoric gesture because the cup acts as a symbolic image for the idea of a concept. Deictic gestures refer to things by pointing with the hand, the finger, the chin, etc. They can be either concrete, pointing to someone, something or somewhere, as when one says, “Your glasses are here on the table,” while pointing towards the table and the glasses. But they can also be abstract, pointing when referring to something/someone absent, or a place, or even a moment in time, as for instance when one points to the right to mean China or behind them to refer to the past. Abstract deictics can be shaped by cultural characteristics as geographical and time references differ between languages and cultures. Finally, beats are rhythmic movements that have no semantic connection to the speech they accompany. They rather stress important words or phrases. A typical beat would be a flick of the hand or of the finger. McNeill

< previous page

page_194

next page >

< previous page

page_195

next page >

Page 195 (1992) explains that the critical thing distinguishing beats from other gestures is that it has two movement phases— in/out, up/down, etc. 1.4 What We Know about Gesture Production 1.4.1 The Relationship between Gesture and Speech Gesture and speech are considered by most gesture researchers as being part of one single system (McNeill, 1992); this is why they should not be analyzed separately. There are two arguments to support the theory of the speech–gesture unified system. The first argument is that there is strong semantic coherence between the two modalities in an utterance. According to McNeill (1992), gesture and speech form a unified communication system; the coherence is possible because gesture and speech share a common cognitive representation: they are part of a single idea. When a speaker produces a message, most of the information s/he wants to share is conveyed in speech while part of the information may be channeled through gesture. However, gesture and speech convey information from different perspectives. In short, speech conforms to a codified, restricted and recognizable system of words and grammatical devices, whereas gesture is free from the standards of form language imposes and conveys meaning on a rather global and visual basis (Goldin-Meadow, 2003). With gestures, one can describe shape, motions or size far more easily than with words. Most of the time, information conveyed through gestures is visual imagery. Because they are so different, gesture and speech when both implied in the same message do not always bring the same information. Church and Goldin-Meadow (1986) talk about gesture–speech matches when gesture is elaborated on a topic already introduced in speech, and gesture–speech mismatches when gesture introduces new information not conveyed in speech. It is thus not rare in a message that gesture brings information that completes speech. For instance, a woman says, “She chases him out again” (talking about an old lady running after a cat) and moves her hand back and forth revealing that she uses her umbrella as a weapon (McNeill, 1992). In this example, the gesture provides us with information not conveyed in speech and shows us how much gestures can describe things speech cannot. Gesture is not restricted to a fixed form and can vary on several dimensions such as time, form, motion, trajectory, use of space, shape, rhythm, etc., which make it complex. The second evidence that gesture and speech form a unified system is that they are always synchronous. McNeill (1992) found that 90% of gestures are produced while the gesturer is speaking. It has also been found that gesture and speech are co-temporal in a single utterance: the stroke of the gesture lines up with the linguistic equivalent.

< previous page

page_195

next page >

< previous page

page_196

next page >

Page 196 1.4.2 Why Do We Produce Gestures? A first answer to this question could be: to help our listeners to understand what we say. Indeed, Alibali, Heath, and Myers (2001) have found that people gesture more when talking to a visible interlocutor, and that when they talk to someone hidden behind a screen they tend to use fewer illustrative gestures. Several other studies have come to similar findings (for a review, see Alibali et al., 2001, and Özyürek, 2002). Özyürek (2000, 2002) explored the communicative function of gesture by analyzing how speakers design their gestures according to the location of their addressees. She found that speakers oriented their gestures depending on where their interlocutors were sitting, so that gestures could be seen. In order to find out whether gestures were taken into account by the listeners, Kelly and colleagues (Kelly, Barr, Church, & Lynch, 1999) analyzed the role of deictic gestures on the understanding of indirect questions, like saying, “It’s hot in here,” while pointing to the window, inferring that the listener should go and open it. Results show that deictic gestures help listeners to understand better the hidden intention in the speaker’s message. Beattie and Shovelton (1999) showed that subjects listening to someone telling a story understand significantly more details when they see the speaker (and their gestures) than when they do not. Listeners also take into account information conveyed in gesture when it completes or contradicts speech (Cassell, McNeill, & McCullough, 1999). However, even if gesture helps listeners to better understand a conversation, it seems that this is not the main function. Indeed, in the study of Alibali et al. (2001) already mentioned, even if speakers produced fewer gestures when they did not see their interlocutors, they did still gesture. Moreover, Iverson and Goldin-Meadow (1998) have laid evidence that congenitally blind speakers spontaneously gesture even when they speak to blind listeners. Thus, we can assume that gesture does not solely convey information for the listener but also plays a role for the speaker. This can also explain why we gesture when we talk on the telephone, for instance. So if we produce gestures for ourselves, what is the function of gesture in speech production? There are several theories on this topic. The Lexical Retrieval Hypothesis (LRH) holds that gesture plays an active role in lexical access, particularly for words with spatial content (Rauscher, Krauss, & Chen, 1996). Thus gesture plays a role in generating the surface forms of utterances it infers directly in the process of speaking. Alternatively, the Information Packaging Hypothesis (IPH) (Alibali, Kita, & Young, 2000; Kita, 2000) is drawn from McNeill’s (1992) and McNeill and Duncan’s (2000) theory of gesture and speech as an integrated system (Growth Point). It argues that gesture and speech help to constitute thought and that gestures reflect the imagistic mental representation that is activated at the moment of speaking. In order to find out which theory (LRH or IPH)

< previous page

page_196

next page >

< previous page

page_197

next page >

Page 197 is likely to be true, Alibali et al. (2000) gave 5-year-olds two oral tasks: one was a description task (the children had to describe different objects) and one was an explanation task (a Piagetian conservation task like, for instance, judging whether two different receptacles contain the same amount of sand). Both tasks required similar lexical use (the same objects to talk about) but inferred different cognitive conceptualizations (one being description and the other explanation). According to the LRH, subjects should use the same gestures in both tasks since they need roughly the same lexical items. Conversely, according to the IPH, as conceptual planning is different in both tasks, gestures should be different. The hypothesis is that if children use different gestures in both tasks while using similar words, then gestures help not only to retrieve words but also to organize thought and conceptualize the message to be verbalized. Results show that, indeed, in the explanation task (more demanding cognitively), children used more gestures conveying perceptual dimensions of the objects and more gestures conveying information that differed from the accompanying speech. Thus, gesture helps cognitive activity. Alibali et al. (2000) conclude that “The action of gesturing helps speakers to organise spatial information for verbalisation, and in this way, gesture plays a role in conceptualising the message to be verbalised” (Alibali et al., 2000, p. 610). However, even if data tends to favor the IPH theory, the authors do not reject the LRH and admit that gesture helps both lexical retrieval and organization of spatial information for verbalization. One last noticeable element on gesture and production is that it has been found that preventing subjects from gesturing has an effect on speech: for instance, in a description task, gesture-restriction has an effect on the amount of time needed to describe an object (Cohen & Borsoi, 1996) and it also generally decreases speech rate (Morsella & Krauss, 2004). 2 Gesture Development in Childhood The first communicative gestures appear at a very early age. Many researchers have analyzed them and their occurrence with speech. It seems that gesture plays a crucial role in transitional knowledge. 2.1 What We Know about Gesture Development in Childhood From the age of 10 months, babies begin to produce some kind of gestures like pointing, giving, showing (Bates, Benigni, Bretherton, Camaioni, and Volterra, 1979; van der Straten, 1991). They repeat behaviors that they know will catch adults’ attention. Deictic gestures or pointing, which rapidly increase at the end of the first year of age, are considered by psycholinguists as prelinguistic gestures, for they constitute an important stage in the development of speech. Pointing, accompanied by eye contact with an adult, aims at seeking information or approval and acts as a precursor to spoken and sign

< previous page

page_197

next page >

< previous page

page_198

next page >

Page 198 naming. Indeed, the sequence of deictic gesture development reveals the gradual distancing of self from object that underlies symbolic development (Capone & McGregor, 2004). The child points to an object not to request it but to refer to it; this reveals that the child can isolate an object from the rest of his/her environment as s/he will soon do with words that will be isolated from the flow of speech the child is exposed to. The ability to decontextualize is crucial as it is related to the ability to use a word in the absence of the referent or to use it with other exemplars of the same referent. In the period between 9 and 13 months, ritualized requests appear, like open–closed grasping motions or pulling an open hand to obtain something (Bates et al., 1979). Representational gestures begin to emerge around the age of 12 months before the onset of the 25-word milestone. These are not instrumental gestures, for the infant does not manipulate objects but rather represents referents symbolically. For instance, the child represents the action of holding a glass and drinking or flaps his/her arms to represent a bird. Goodwyn and Acredolo (1993) consider that these representational gestures are real examples of language symbols and can be analyzed with the same criteria used to define spoken words. They argue that a gesture or a word is symbolic if it refers to multiple exemplars including pictures and absence of the referent, if it is produced spontaneously (without following the model of an adult), and if it is not part of a well-rehearsed routine (Goodwyn & Acredolo, 1993). Between 12 and 18 months, the child gestures in an isolated way, which means that s/he either gestures or speaks, but hardly both in the same time. The child thus chooses between the two systems s/he knows (McNeill, 1992). Iverson, Capirci, and Caselli (1994) found that 16-month-old children have a preference for either words or gestures, but by 20 months there is a significant increase in types and tokens of spoken words. As we have already stated, gesture and speech in adults seem to belong to a single system (McNeill, 1992). This hypothesis is supported by two characteristics: the integration of gesture and speech in a semantic coherent unit (the fact that gesture is combined with speech in a meaningful way) and the temporal synchrony between speech and gestures in a single utterance. But is that also true for young children? Is gesture–speech integration characteristic of the earliest communications of young children? Or does integration of the two modalities emerge at a consistent point in the young child’s linguistic development? To answer these questions, Butcher and GoldinMeadow (2000) have longitudinally observed three boys and three girls during the transition from one- to twoword speech. They started to videotape their subjects during play sessions when they were beginning their oneword period of language development (age range 12 to 21 months, mean 15.5 months) and until the stage of twoword combination (range from 18 to 26.5 months). During the one-word period, for five of the six children, 20%

< previous page

page_198

next page >

< previous page

page_199

next page >

Page 199 of the total number of their communications (speech and/or gesture) included a gesture (for the sixth child, it was approximately 40%). During the first session, data uncovered that most of the subjects (five out of six) produced the majority of their gestures without speech (to compare, McNeill, 1992, has found that only 10% of adults’ gestures are produced without speech). Then, during the following sessions, a general decline in the proportion of gestures produced without speech was observed. Thus, children began the one-word period producing gestures without speech, and by the end of this period they mainly used gesture–speech combinations. The two characteristics of adults’ productions of speech and gestures are the synchrony of both modalities and the semantic coherence. Consequently, Butcher and Goldin-Meadow (2000) observed whether children’s productions of speech and gestures bear these same characteristics. As far as synchrony is concerned, during the first session, five of the six children produced gesture–speech combinations that were not synchronous with speech (the sixth child produced synchronously timed combinations throughout the observation period). During the next sessions, combinations became more and more harmonious. The authors thus suggest that “gesture and speech do not form a completely integrated system from the start but may require some time to become aligned with one another” (Butcher & Goldin-Meadow, 2000, p. 246). As far as semantic content is concerned, McNeill (1992) discovered that gesture and speech “cover the same idea unit” (1992, p. 27) even if gesture and speech do not convey precisely the same information. When analyzing the gestures combined with meaningful words produced by their children, Butcher and Goldin-Meadow (2000) found that the number of gesture–speech combinations increased during the observation period. The children produced both occurrences of gesture conveying the same information as speech (point to the box and say “box”) and occurrences of gestures conveying different but related information (point to the box and say “open”). In this later case, the child can express two different elements in a single utterance (one in gesture and one in speech), something s/he is not yet able to do in speech only. “Thus the ability to combine gesture and meaningful speech in a single utterance greatly expands the child’s communicative range” (Butcher & GoldinMeadow, 2000, p. 248). By putting together all these findings, the authors highlighted the striking fact that the three events converge in time: gesture-alone communications began to decline and “synchronous gesture–speech combinations began to increase at just the moment when gesture was first combined in the same utterance with a meaningful word” (Butcher & Goldin-Meadow, 2000, p. 248). To sum up the observed developmental sequence: the child begins to produce communicative symbolic gestures mostly without speech; when gesture is combined with words, speech is meaningless and gesture is not synchronized with it. Then, gesture and speech become more fully integrated and the child begins to produce synchronized combinations of gestures and

< previous page

page_199

next page >

< previous page

page_200

next page >

Page 200 meaningful words. This is the beginning of gesture–speech integration as we find it in adult expression. Butcher and Goldin-Meadow (2000) explain that the emergence of combinations in which gesture and speech are semantically related but do not convey the same information represents a communicative, even conceptual, breakthrough for the child and announces the onset of two-word speech. Indeed, in the six children observed, the correlation between the onset of this type of gesture–speech combinations and the onset of two-word utterances is high and significant. During toddlerhood, children come to prefer verbal to gestural expression as they are learning more and more words. However, children still use gestures and there is a certain increase in the use of deictics, particularly accompanying expressions such as “this” and “that” (Iverson et al., 1994). In the second and third years of life, pointing becomes increasingly integrated with spoken language, particularly to supplement spoken messages (Iverson et al., 1994). From 16 to 20 months, there is a significant increase in pointing gestures co-occurring with representational words. As speech develops, gestures become more and more elaborated, especially in their relation to speech. Iconics tend to appear more and more with verbs and adjectives, rather with nouns, and the relationship between gesture and language extends to the domain of morphosyntax as the children advance in these areas (Capone & McGregor, 2004). Between the third and the fifth years of age, iconic gestures increase significantly. Iconics and speech become more and more synchronized. Nevertheless, children’s co-speech gestures do not yet refer to abstract contents; metaphorics are hardly found in young children’s gesture productions. From the age of 5, the rest of the gestural system develops and beats, metaphorics and abstract deictics become more and more numerous (McNeill, 1992). Colletta (2004) has conducted a vast quantitative study on the development of verbal and non-verbal activity of children from 6 up to 11. He confirms McNeill’s findings as far as the emergence of metaphorics and beats is concerned (after the age of 5/6). He also found that multimodal story-telling skills (linguistic, prosodic and gestural) develop together and simultaneously. Colletta also showed that the study of co-speech gestures enables researchers to gather clues and relevant information on the development of concept and mental imagery of children. As children grow older and develop, gestures develop too and appear in cognitive tasks very often, allowing researchers to better understand how the child acquires concepts. Studying the matches and mismatches in speech and gestures produced by children proves to be very relevant when one tries to understand their cognitive development (Goldin-Meadow, 2003). It appears that when some children explain something they have not yet understood (a math concept, for instance), they tend to convey the same incorrect information both in gesture and in speech, in a single procedure, so to speak. They then enter a discordant state in which they produce different procedures: one in speech and another in

< previous page

page_200

next page >

< previous page

page_201

next page >

Page 201 gesture. This means that the child is in a zone of proximal development. The information expressed in gesture is different from that expressed in speech. Most of the time, accurate information tends to be that conveyed in gesture. Then, when the concept is acquired by the child, gesture and speech again match in the child’s production (Alibali & Goldin-Meadow, 1993). This transitional state is thus characterized by the concurrent activation of more than one procedure, and provides further evidence that gesture can be a powerful source of insight into the processes involved in cognitive development. This phenomenon has been noticed for math and science concepts but is probably applicable to other general concepts. Therefore, gesture has a direct effect on the learning process and scaffolds the child’s cognitive development by structuring the various stages of the acquisition of a concept or a skill. As we have seen, the analyses of the gestures produced by a child can reveal stages of transitional knowledge. The first deictic gestures announce the emergence of the first words. Then the combinations of gesture and speech conveying different but related information precede the first two-word utterances. And finally, as the child develops, it seems that complex concepts emerge in gesture before they appear in speech (or in speech combined with gesture). Globally, gesture–speech mismatches occur in a wide variety of situations and at different ages, from childhood to adulthood (for a review, see Goldin-Meadow, 2003). The study of gesture–speech matches and mismatches offer a window to the mind of the developing child and of the teenager. Goldin-Meadow (2000, p. 237) suggests looking “beyond children’s words to the secrets that, until now, have been locked in their hands” to discover more about children’s learning. 2.2 For Further Research As far as development of the gestural system is concerned, most of the studies concern very young children who are acquiring their first language. Consequently, less attention has been devoted to older children and how they develop their way of gesturing while acquiring new discursive skills. Colletta’s study (2004) is thought worth mentioning, since it concerns gesture development between 6 and 11 years old. Gestures of children after 11 and of teenagers have not been much studied. This is perhaps due to the fact that most of the first language is acquired and that significant changes are very slow to occur. However, it seems relevant to study how a teenager develops his/her own style of gesturing at this particular period of self-constructing. The way somebody gestures depends on many factors (detailed in the next section), among them personality. Teenagers may also be influenced by fashion in their way of gesturing. Gestures used by rap signers, for instance, seem to influence young individuals, especially boys. Gender is a factor which seems worth studying as well. Whether boys and girls gesture the same way is something left to discover.

< previous page

page_201

next page >

< previous page

page_202

next page >

Page 202 Research on gestures and children is, as we have seen, relevant to understanding how a child acquires a first language and how gestures participate in cognitive development. These researches have implications in the field of education. For instance, it seems relevant to work on gestures children look at when they learn, and in the field of education this means teachers’ gestures. Singer and Goldin-Meadow (2005) have laid evidence that teachers’ gestures do not always convey the same information as their speech. This mismatch thus offers learners a second message (one conveyed by gesture and the other by speech). To determine whether learners take advantage of this information, Singer and Goldin-Meadow gave 160 children in the third and fourth grades instruction in mathematical equivalence (for example: “6 + 4 + 3 = _ + 3”). Children were taught either one or two problem-solving strategies in speech accompanied by no gesture, gesture conveying the same strategy, or gesture conveying a different strategy. The chosen strategies are commonly used by teachers when teaching mathematical equivalence: “(a) equalizer, a strategy highlighting the principle underlying the problem, and (b) add subtract, a strategy highlighting a procedure for solving the problem” (2005, p. 86). Results show that the children were likely to profit from instruction with gesture, but only when it conveyed a different strategy than speech did. Moreover, two strategies were effective in promoting learning only when the second strategy was taught in gesture, not speech. Singer and GoldinMeadow (2005) conclude that gesture has an active hand in learning. In the field of second language teaching to young children, it has also been found that teachers’ gestures help children to better understand the second language without translation. They also help the child to remember L2 lexical items better when s/he visualizes an illustrative gesture while listening to the matching word. Data has also shown that children who reproduce their teacher’s gestures while repeating new L2 words remembered significantly more items than those who just looked at them (Tellier, 2006). However, one may wonder whether a child always understands adults’ gestures, since gestures reflect mental imagery and adults’ and children’s mental imagery differs because of their different cognitive states and experiences of life. Adults tend to use a lot of metaphoric gestures that may not be understood by young children, since they do not use such gestures and do not represent the world in a abstract and symbolic way. Misunderstandings of adults’ gestures by 5-year-old children have been found (Tellier, 2006), but more data is definitely needed on this topic, extending to various age ranges to help teachers think about how they can improve their teaching gestures. 3 Gesture Development in Adulthood While many researchers work on the development of gestures during childhood, there seem to be very few studies on this development during adulthood. Studies on development focus essentially on acquisition and decline,

< previous page

page_202

next page >

< previous page

page_203

next page >

Page 203 for example loss of language and language-related gestures. However, we can also notice some temporary changes due to change of jobs or related to belonging to/integrating into a specific community. One explanation for this lack of studies may be that development in adulthood is much slower than in childhood, and therefore it would take longitudinal studies of several years to notice changes in the development of a single subject, while children change so quickly that studying a child during a period of a few months is enough to notice and analyze the changes in both his/her speech and gesture skills. This is probably why most studies concerning adults are comparative studies in which subjects of different age groups are given the same task so that results can be compared according to the age variable. Another explanation for the rarity of studies on adults is that interest is rather focused on childhood, when most of the development takes place. However, we cannot assume that gesture production does not change across the lifespan once an individual has reached the stage of adulthood. In this section, we will review the studies on adults and the evolution of their gestures across the lifespan, and we will expose what research needs to focus on in the years to come. 3.1 What We Know about Gesture Development in Adulthood 3.1.1 Different Adults, Different Ways of Gesturing Most branches of psychology examine how subjects behave in different settings or under various experimental conditions, assuming that they will behave/react the same way (Cooper, 2002). However, there is significant variation between people, and this is true at any age. Some children for instance develop more quickly than others, and they grow up to develop different personalities and mental abilities. Therefore, adults have to be considered as different human beings, and the way somebody gestures is very specific. We can try to figure out which parameters can influence the way somebody gestures. Every human being is brought up in and belongs to a certain community that will influence his/her development as a child but also the adult s/he will become. A variable such as cultural origin has a crucial effect on someone’s gestures. First, the emblems someone uses reveal his/her belonging to a certain cultural background, for they are, as mentioned before, culture-specific (Morris, Collet, Marsh, & O’Shaughnessy, 1979). However, we also know that more spontaneous forms of gestures can also bear cultural characteristics, though there are few studies on the subject. David Efron (1972 [1941]) studied the gestures of both Jewish (from Eastern Europe) and Italian immigrants freshly arrived in New York City, and was able to compare them to members of the same ethnic groups who were more assimilated to the American culture. He noticed significant differences between the “traditional” southern Italians and the “traditional” eastern Jews on various

< previous page

page_203

next page >

< previous page

page_204

next page >

Page 204 bases: the use of the gesture space (wide gestures or close to the trunk), the body parts implied in the gestures (the forearms or the whole arms, for instance), the gesture tempo, etc. For example, Efron found that the traditional Italians gestured in a wider radius than the traditional Jews. Also, the Jews seemed to hinge their gestures from the elbow, while the Italians’ axis of movement was rather from the shoulder. Efron also found out that Jews tended to gesture with one arm, or when they used both they tended to move them in an asymmetric way. On the contrary, the Italians were more likely to gesture with both arms and in a symmetric way. A last example taken from the many differences observed is the use of symbolic gesture, which was really more important in the traditional Italian community. These findings show that there is an effect of the cultural variable on the way people gesture. Efron also found far less differences between the “assimilated” Jewish and Italian communities, whose ways of gesturing resembled those of the Anglo-Saxon speakers. We may infer from this finding that our gesture style can be influenced by other cultures when we have a long contact with them. Connected to cultural origin, one’s first language is also known to have an impact on the production of co-speech gestures. Linguistic structures vary from one language to the next (Talmy, 1985) and so do gestures. For instance, data has shown that Dutch, French and Swedish native speakers give more importance to verbs and actions in a sentence than do Japanese, who rather stress the location and the settings for actions. Consequently, co-speech gestures produced by Dutch, French and Swedish speakers appear along with verbs, whereas the Japanese gestures provide information on the setting of the action (Gullberg, 2006; Yoshioka, 2005). Thus, speakers of various languages differ in the way they verbally describe motion events and space, and in their gestures. As McNeill and Duncan (2000, p. 154) put it, concerning English, Spanish and Chinese, Describing the same motion events, languages encourage different forms of thinking. English and Spanish … are predicative in their focus, but thinking differs in how motion-events semantics are focused. Chinese induces thinking in which a focus is a frame for other information. Observations thus show an effect of linguistic organization on thinking on two levels—predicative and discourse—and different patterns on both. (See also Kita & Özyürek, 2003.) Every human being also belongs to a certain social origin which probably has an effect on the gesture style (Calbris & Porcher, 1989), though it seems that there are hardly any studies on the subject. We nevertheless know that some gestures or ways of gesturing are considered rude for certain social

< previous page

page_204

next page >

< previous page

page_205

next page >

Page 205 classes, and others are considered posh or even ridiculous by others. There is clearly a need for data on this variable. We have just examined how the cultural, linguistic and social background of an individual can affect the gesture style and reveal the belonging of this person to a cultural/linguistic/social community. If we want to go further on differences, we have to remember that each community is composed of groups of individuals whose gestures are shaped by a certain combination of variables. A first variable is gender, though we do not have much information on how gender affects gestures, for there does not seem to be any systematic measurement of differences as far as co-speech gestures are concerned. However, studies in non-verbal communication have shown differences such as the way arms and legs are displayed, and the way people sit, stand or walk. Even the way they carry books can vary depending on gender. These differences seem to appear during childhood, then develop and increase with age (see Rekers, Sanders, & Strauss, 1981, for a review). However, there may be a difference between men and women as far as co-speech gestures are concerned, since it is sometimes said about some men that they are effeminate and this assumption is based on their gestures. Sexual orientation could be a factor influencing the way somebody gestures; research on this issue could give us more information on gender and gesture styles. There are other variables that may influence gestures like personality, for instance. How personality can affect nonverbal behavior has been studied (see Feyereisen & de Lannoy, 1985, for an overview), but there has been little work on hand gestures specifically. The effect of mental health and some specific psychological disturbances on gesture has also been analyzed, and it has been shown that depressive people tend to produce more self-touching gestures and that schizophrenics use more speech-related gestures. These gesture rates are likely to change along with the amelioration or deterioration of the mental health condition (Freedman & Hoffman, 1967; Freedman, 1972). Verbal skills and the level of proficiency in the language can also have an effect on the way someone gestures. This is true for both first and second language acquisition. In both cases, the lower the proficiency, the greater the number of gestures. As we have already mentioned, when children acquire their first language, their gestures are not replaced by speech but develop in parallel with it. As for second language acquisition, data has shown that learners produce more gestures when speaking in their L2 than in their L1 (Gullberg, 1998; Sherman & Nicoladis, 2004; Yoshioka, 2005). For adult L2 learners, gestures tend to be complementary from the beginning: “complementary strategic gestures serve both to elicit responses from listeners and to create redundancy” (Gullberg, 1998, p. 230). Also, in cases of disfluency and depending on the type of difficulty, L2 learners tend to use compensatory gestures. Indeed, as Gullberg (1998) has shown, they use

< previous page

page_205

next page >

< previous page

page_206

next page >

Page 206 gestures as communicative strategies to overcome expressive problems such as lexical shortcoming, grammatical difficulties and fluency-related problems. To find out how gestures can change in connection to language development, it is very relevant to work with adult subjects learning another language. As stated above, studying co-speech gestures can give us hints on the specific linguistic organization of each language. Two languages with different linguistic conceptualizations might lead to different gestures. Current research on gestures and L2 acquisition is analyzing how a learner acquires a target language by observing his/her gestures, with the assumption that if the learner has acquired the L2 conceptualization then his/her gestures should look L2-like (Gullberg, 2008). Finally, professional skills can have an effect on the gesture style, especially when a profession deals with rhetoric. Some studies on politicians (Calbris, 2003) and on second language teachers (Tellier, 2006; Cadet & Tellier, 2007) have highlighted the specificity of gestures produced intentionally for a large audience and in order to stress and illustrate major information. Empirical studies and observations aiming at showing how the development of professional skills can influence the way people gesture are definitely needed. 3.1.2 Gesture Development and Aging When looking at the scientific literature concerned with body movements and aging, one can notice that there is little work on spontaneous co-speech gestures and that there are a few studies on non-verbal behavior and how aging affects it. As already stated, this can be explained by the fact that longitudinal studies are difficult to set up, and maybe also by a lack of interest from researchers. It is also important to mention that there is a practical reason for studying young adult subjects rather than old. In many studies conducted in universities, local students often act as subjects, for it is, of course, very convenient to work this way. This, therefore, raises the question of the diversity of subjects of many studies who are, most of the time, under 25. Most of the available data concerning body movements and aging was collected in non-verbal communication studies, especially about the expression of emotions. This field of research has been particularly interested in how subjects perceive other people’s emotions on the basis of non-verbal cues. Studies exploring age-related differences in the perception of emotions from facial and vocal cues have found some evidence for age-group differences. Declines in the experience of emotion are more reliable for negative than for positive emotions (Montepare, Koff, Zaitchik, and Albert, 1999). It has also been found that older subjects seem better at identifying the facial expression of similar-aged peers than of younger adults, as uncovered by Malatesta, Fiore, and Messina (1987).

< previous page

page_206

next page >

< previous page

page_207

next page >

Page 207 Are there any age effects in the production of spontaneous hand gestures? Cohen and Borsoi (1996) have tried to answer this question by looking at what they call “descriptive speech gestures” (or representational gestures, linked to the verbal content) and at “non-descriptive speech gestures” (or non-representational gestures, which are connected to the flow of speech but which do not carry any part of the message) produced by different age groups in a descriptive communication task. They rely on previous comparisons of elderly adults with young adults on cognitive tasks which indicate that the elderly may suffer from a production deficiency. They also rely on the fact that gestures reflect difficulties in the verbal speech system and that gestures are produced in an attempt to compensate for language problems. They base this assumption on the findings of Marcos (1979) and Feyereisen (1983). Marcos found that bilinguals used more gestures when speaking their weaker language and Feyereisen highlighted the excessive use of gestures by aphasic subjects. Therefore elderly subjects are expected to compensate for relatively weak verbal communication skills by producing more gestures than young adults (Cohen & Borsoi, 1996). Similarly to Marcos’ (1979) findings, they also expect older adults to use more non-representational gestures than representational gestures. In a description task, Cohen and Borsoi (1996) asked 24 female undergraduates (age mean 19.92) and 24 females attending retirement courses at the university (mean age 69.42) to describe objects. They compared the subjects’ oral performance on the basis of the amount of time used for each description, the quality of the object description, and the rate of representational and non-representational gestures. They found no significant age effect for the description time or for the quality of object description. Thus, neither the compensation nor the production deficiency approach received support from their data. However, they found a significant age effect on the rate of representational gestures. Younger female subjects used significantly more representational gestures when describing objects than older ones. There was no difference in the use of non-representational gestures. Consequently, the expectation that difficulties in verbal description would be compensated by an increase of the use of gestures by older women was not fulfilled (Cohen & Borsoi, 1996, p. 53). In another similar experiment, Cohen and Borsoi (1996) added an extra within-subjects variable: gesture-restriction. Data uncovered an overall effect of the gesture suppression variable: time description was significantly longer in the suppressed condition (a finding later confirmed by Morsella and Krauss, 2004, who found that gesture-restriction decreased speech rate). However, this variable did not significantly interact with age. Moreover, neither age nor suppression variables affected the quality of verbal descriptions in a significant way. Once again, Cohen and Borsoi (1996) only noticed a significant age effect on the rate of representational gestures (and not on the rate of nonrepresentational gestures).

< previous page

page_207

next page >

< previous page

page_208

next page >

Page 208 In both experiments of Cohen and Borsoi (1996) the elderly women tended to take longer for their descriptions than the younger female subjects; however, this was not significant in the data because of large variation. No age effect on the quality of the descriptions was found. Data showed a significant age effect only on the types of gesture produced. Indeed, younger subjects produced significantly more representational gestures than older subjects. The authors suggest that this difference may be due to the fact that elderly people appear to be less involved with visual images than young adults (Fein et al., 1985) and that the production of representational gestures tends to be driven by a visual or visuomotor imagery system. Thus, young subjects may produce more representational gestures because they refer to more visual images. The authors conclude that further experiments are definitely needed on this topic. Studies could, for instance, involve male subjects to find out if there is a significant gender variable, and be based on another task to check if the observed age difference is task-related. Also, the hypothesis of the variation in the use of visual images should be further explored. Indeed, several studies have tried to find out whether the ability to generate visual images declined with age. A study by Dror and Kosslyn (1994), in which subjects had to imagine things and press keys accordingly to their visual representations, showed that older subjects were slower and less accurate in performing the task than younger subjects. Nevertheless, these could be explained by a general slowing and reduced efficiency in elderly people. This is supported by other experiments involving visual imagery and different age groups and showing no effect of the age variable (see Feyereisen & Havard, 1999, for a review). Feyereisen and Havard (1999) tried to find out whether different kinds of speech-related gestures depend on the same system or on different subsystems. This means an evaluation of McNeill’s (1992) theory that all kinds of gestures serve similar functions and that they belong to a single control system, with Hadar and Butterworth’s (1997) theory that iconics (i.e. representational gestures) are related to visual imagery whereas beats (i.e. nonrepresentational gestures) are connected to phonological encoding of sentences. According to this later hypothesis of two separate mechanisms underlying the production of gestures (representational vs. non-representational), iconics and beats should grow and decline at different rates. Feyereisen and Havard (1999) interviewed younger (M = 21, range 18–25) and older (M = 70, range 61–80) adults using various questions. Three questions were used to elicit visual imagery (for instance, “Could you describe a favorite painting or sculpture?”), three questions were used to elicit motor imagery (for instance, “Could you explain how to cover a book or to wrap a box in a paper for a present? ”), and three questions concerned abstract topics (for instance, “Do you think that more women should go into politics?”). Thus subjects were tested in three conditions: a visual imagery condition, a motor imagery condition, and an abstract condition.

< previous page

page_208

next page >

< previous page

page_209

next page >

Page 209 Results show that gesture production is affected by the content of the message. Questions eliciting motor imagery led to a larger amount of representational gestures than other questions, which can be explained by some specific influence of mental representation of movements. Representational gestures were also frequent in the visual imagery condition and less frequent in the abstract condition. Thus, representational gestures tended to occur with visual and motoric speech content, whereas non-representational gestures were associated with more abstract content. As far as age difference is concerned, it appeared that there was no global age difference. Rather, an age-related decrease was found in the proportion of representational gestures, especially in the visual imagery condition corresponding to an increase in the proportion of beat gestures. Therefore, to some extent, it matches the results obtained by Cohen and Borsoi (1996). However, Feyereisen and Havard (1999) hesitate to conclude that decreased rate of representational gestures indicate a reduced use of visual imagery. The fact that the subjects of Cohen and Borsoi (1996) could look at the objects while they were describing them implies that they did not need to activate visual imagery from memory. Also, the proportion of representational gestures produced by younger and older adults did not differ in the motor imagery condition, only in the visual condition, so we can suppose that iconic gestures are not exclusively controlled by visual imagery. Then, content analysis of the speech did not show age-related differences in the use of high-imagery words. Feyereisen and Havard (1999) recommend revision of the existing theories by suggesting that age-related variations in the iconic/beat ratio are due to several changes in speech characteristics. Indeed, younger and older adults have been found to use different speech styles (see Feyereisen & Havard, 1999, for a review). Because the size of the vocabulary continues to expand across the lifespan and because the culture and educational system have changed a lot across the 20th century, there are stylistic variations between younger and older subjects and they may be reflected in the gestural behavior. McNeill (1992) found that beat gestures often serve meta-narrative functions, and Feyereisen and Havard (1999) noticed such occurrences in their corpus. They admit that these phenomena might be more frequent in the conversational speech of older subjects. The authors also wonder whether beats could accompany more elaborate forms of language and whether these could be weaker forms of representational gestures. Indeed, iconic gestures were more frequent in the shorter responses of younger subjects. One could hypothesize that throughout the lifespan, iconic gestures are gradually being replaced by beat gestures, along with the development of speech, and that this process begins during childhood. This hypothesis would support McNeill’s theory of a single mechanism controlling the production of both representational and nonrepresentational gestures. Yet the fact that in the data of Feyereisen and Havard (1999) the production of beats by older subjects did not increase in the high-imagery condition does not support this

< previous page

page_209

next page >

< previous page

page_210

next page >

Page 210 hypothesis. It rather seems that different mechanisms control the production of representational and nonrepresentational gestures. Bassal and de Lannoy (2001) studied the production of spontaneous gestures of eight aging subjects (M = 86.7, range 77–92) in non-directive interviews. They found out that the subjects tend to have an essentially autobiographical discourse (about their personal life) and that the ratio of non-representational gestures (beats) is higher than the ratio of representational gestures, in particular in the autobiographical discourse. The authors consider that their results are consistent with Feyereisen and Havard’s (1999) idea that the changes in gesture production are due to changes in speech characteristics. Indeed, they hypothesize that older people are more likely to talk about their personal life (their family, their job, their childhood …) because (a) they suffer from isolation and tend to focus on themselves, (b) egocentrism is connected to end-of-life concerns, (c) the aging person feels the need to turn back to past experiences. 3.2 For Further Research There is a lot more to explore on the subject of gesture production and adulthood. First, the difference between individuals needs further experiments. Studies on co-speech gestures tend to shed light on similarities between subjects; however, we know that there is an important variability between individuals. It could be interesting to give a task (retell a story, describe a picture, give directions, explain something …) to subjects and analyze their gestures on the basis of their sex, language proficiency, social origins, professional activity, personality, education, cognitive style, mental health, etc. Parameters to look at would be gesture type, gesture rate, iconicity, rhythm and the use of gesture space, for instance. Second, gesture production and aging needs to be explored on various bases. First of all, age probably has an effect on gestures on a biological basis. Indeed, age-related decline in motor control is due to modifications in the central nervous system, specifically neural reduction of brain regions, and the loss of muscle mass that occurs with advanced age (Ketcham & Stelmach, 2001). This decline in motor control has an effect on everyday life as far as practical actions are concerned, but it might also have an effect on the production of co-speech gestures. Laver and MacKenzie Beck (2001) explain that the anatomical posture of every individual evolves throughout life. They call this evolution of the anatomical posture the individual’s baseline posture which is altered by “the interactive, cumulative effects on the skeletal and physiological framework generated by genetic inheritance, developmental growth, aging processes, gravity, habitual muscular action, diet and health record” (2001, p. 19). Thus, senescence has an impact on gesture production, especially on the use of gesture space.

< previous page

page_210

next page >

< previous page

page_211

next page >

Page 211 Each individual, over a lifetime of gestural performance, will have developed a characteristic three-dimensional space within which their gestures are habitually performed—their idiosyncratic gestural space. This space will represent some selection, as it were, from the total gestural space organically available to the individual. That organic gestural space will itself have changed through life, under the mechanical influences of the person’s stage in their life-cycle, and through any factors of trauma and pathology. (Laver & MacKenzie Beck, 2001, p. 20) For instance, changes in the spinal curvature, like a posterior curvature in the upper back area and in the neck, cumulated with muscular weakness may lead to shifting the gestural space forward (Laver & MacKenzie Beck, 2001, p. 22). Third, even if McNeill’s theory of gesture and speech being part of a unique integrated system seems to be true for the production of representational gestures (as we have seen both for children and adults), we can still wonder whether it is the same for non-representational gestures. As Feyereisen and Havard (1999, p. 169) put it, “as elderly speakers have acquired a great expertise in language use, it is worthwhile to further investigate the various ways in which beat gestures may serve their discourse.” Finally, it seems relevant to inquire how age-related disease (such as Alzheimer’s) can affect gesture production, on the one hand because studies involving subjects suffering from Alzheimer’s disease can help us to better understand how gestures are connected to speech, and on the other hand because knowing more about how patients suffering from Alzheimer’s disease produce and understand speech can help us to better communicate with them. Individuals with Alzheimer’s disease present complex and heterogeneous cognitive symptoms including memory, language and communication, perception, attention and executive functions. Communication problems encountered by these individuals are mainly word-finding and understanding the spoken language. As far as comprehension is concerned, Pashek and DiVenere (2006) have found that the use of pantomime gestures accompanying commands helped mild to moderate Alzheimer’s disease patients to comprehend spoken language. This is relevant information for caretakers who have to communicate with these patients. As far as production is concerned, Glosser, Wiley, and Barnoski (1998) found that co-speech gestures produced by Alzheimer’s patients reveal several parallels with their linguistic productions. The rate of gesturing of these patients does not differ from the one of healthy age-matched controls. However, when looking closer, one can notice similar disturbances in the specificity and clarity of the referential forms used in verbal and gestural channels by patients with Alzheimer’s disease (Glosser et al., 1998). Data show that their gestural communications

< previous page

page_211

next page >

< previous page

page_212

next page >

Page 212 revealed a proportional decline in the use of symbolically more complex gestures, such as metaphorics, as compared to gestures referring to concrete contents, such as deictics and iconics. Significant correlations between the severity of linguistic and conceptual impairments and the degree of impaired gestural clarity have been found. The authors suggest that this is consistent with the hypothesis that gestural and linguistic communications are closely related in terms of their semantic and conceptual characteristics (Glosser et al., 1998, p. 9). 4 Conclusions Gesture studies have rapidly developed over the past decades and many studies have been conducted in order to better understand how gesture is produced and what its functions are. In communication, co-speech gestures appear to be relevant not only to provide the listener with additional or redundant information but also to help the speaker to produce his/her message. Gesture’s function on lexical retrieval and on the conceptualization of verbal messages has been uncovered by several experiments. Therefore, gesture is widely considered as intimately connected to speech. It even seems that gesture and speech are part of a single integrated system (McNeill, 1992). Many studies on child development, in the acquisition both of language and of concepts, have highlighted the predominant role of gesture in these dynamic processes. Work on gestures and children is very important in the field of education since it enables us to discover more about the process of learning. Gesture and aging have not been studied much. We know very little about how age affects gestures. However, this is a relevant field of study for several reasons. First of all, the increasing numbers of elderly adults in occidental societies (through the aging of the baby boom generation, the improvement of medicine and life conditions that have extended life duration, etc.) are changing the composition of our social world. Second, analyzing how gesture evolves with age can improve our knowledge of gestures and of non-verbal behavior in general, as well as our knowledge of language. Third, different age groups need to communicate with each other, and it seems relevant to take a look at potential differences in gesturing (as well as in speaking) and find out whether or not this leads to communication problems. References Alibali, M. W., & Goldin-Meadow, S. (1993) Gesture–speech mismatch and mechanisms of learning: What the hands reveal about a child’s state of mind. Cognitive Psychology, 25(4), 468–523.

< previous page

page_212

next page >

< previous page

page_213

next page >

Page 213 Alibali, M. W., Heath, D. C., & Myers, H. J. (2001). Effects of visibility between speaker and listener on gesture production: Some gestures are meant to be seen. Journal of Memory and Language, 44, 169–188. Alibali, M. W., Kita, S., & Young, A. (2000). Gesture and the process of speech production: We think, therefore we gesture. Language and Cognitive Processes, 15, 593–613. Bassal, C., & de Lannoy, J.-D. (2001). La gestualité des personnes âgées: Une étude pilote. In C. Cave, I. Guaitella, & S. Santi (Eds.), Oralité et gestualité: Interactions et comportements multimodaux dans la communication (pp. 103–106). Actes du colloque Orage 2001, Laboratoire Parole et Langage—CNRS et Université de Provence. Paris: L’Harmattan. Bates, E., Benigni, L., Bretherton, I., Camaioni, L., & Volterra, V. (1979). The emergence of symbols: Cognition and communication in infancy. New York: Academic Press. Beattie, G., & Shovelton, H. (1999). Do iconic hand gestures really contribute anything to the semantic information conveyed by speech? An experimental investigation. Semiotica, 123 (1/2), 1–30. Butcher, C., & Goldin-Meadow, S. (2000). Gesture and the transition from one- to two-word speech: When hand and mouth come together. In D. McNeill (Ed.), Language and gesture (pp. 235–257). New York: Cambridge University Press. Cadet, L., & Tellier, M. (2007). Le geste pédagogique dans la formation des enseignants de LE: Réflexions à partir d’un corpus de journaux d’apprentissage. Les cahiers de Théodile , 7 , 67–80. Calbris, G. (2003). L’expression gestuelle de la pensée d’un homme politique . Paris: Editions CNRS. Calbris, G., & Porcher, L. (1989). Geste et communication. Paris: Crédif/DidierHatier. Capone, N. C., & McGregor, K. K. (2004). Gesture development: A review for clinical and research practices. Journal of Speech, Language, and Hearing Research, 47(1), 173–186. Cassell, J., McNeill, D., & McCullough, K. E. (1999). Speech-gesture mismatches: Evidence for one underlying representation of linguistic and non linguistic information. Pragmatics and Cognition, 7 (1), 1–33. Church, R. B., & Goldin-Meadow, S. (1986). The mismatch between gesture and speech as an index of transitional knowledge. Cognition, 23, 43–71 Cohen L. R., & Borsoi, D. (1996). The role of gestures in description-communication: A cross-sectional study of aging. Journal of Nonverbal Behavior, 20(1), 45–63. Colletta, J.-M. (2004). Le développement de la parole chez l’enfant âgé de 6 à 11 ans. Corps, langage et cognition . Sprimont: Pierre Mardaga Editeur. Cooper, C. (2002 [1998]). Individual differences (2nd ed.). London: Arnold. Dror, I. E., & Kosslyn, S. M. (1994). Mental imagery and aging. Psychology and Aging , 9(1) , 90–102. Efron, D. (1972 [1941]). Gesture and environment: A tentative study of some of the spatio-temporal and linguistic aspects of the gestural behavior of eastern Jews and southern Italians in New York City. The Hague/Paris: Mouton de Gruyter. Ekman P., & Friesen W. V. (1969). The repertoire of nonverbal behavior: Categories, origins, usage, and coding. Semiotica, 1 , 49–97.

< previous page

page_213

next page >

< previous page

page_214

next page >

Page 214 Fein, G., Feinberg, I., Insei, T. R., Antrobus, J. S., Price, L. J., Floyd, T. C., & Nelson, M. A. (1985). Sleep mentation in the elderly. Psychophysiology , 22, 218–255. Feyereisen, P. (1983). Manual activity during speaking in aphasic subjects. International Journal of Psychology, 18(6), 545–556. Feyereisen, P., & de Lannoy, J. D. (1985). Psychologie du geste. Bruxelles: Mardaga. Feyereisen, P., & Havard, I. (1999). Mental imagery and production of hand gestures while speaking in younger and older adults. Journal of Nonverbal Behavior, 23(2), 153–171. Freedman, N. (1972). The analysis of movement behavior during the clinical interview. In A. W. Siegman & B. Pope (Eds.), Studies in dyadic communication (pp. 153–175) . New York: Pergamon. Freedman, N., & Hoffman, S. (1967). Kinetic behavior in altered clinical states: Approach to objective analysis of motor behavior during clinical interviews. Percept Motor Skills, 24(2), 527–39. Glosser, G., Wiley, M. J., & Barnoski, E. J. (1998). Gestural communication in Alzheimer’s disease. Journal of Clinical and Experimental Neuropsychology , 20, 1–13. Goldin-Meadow, S. (2000). Beyond words: The importance of gesture to researchers and learners. Child Development (Special Issue: New Direction for Child Development in the Twenty-First Century), 71, 231–139. Goldin-Meadow, S. (2003). Hearing gesture: How our hands help us to think . Cambridge: Belknap Press of Harvard University Press. Goodwyn, S. W., & Acredolo, L. P. (1993). Symbolic gesture versus word: Is there a modality advantage for onset of symbol use? Child Development , 64, 688–701. Gullberg, M. (1998). Gesture as a communication strategy in second language discourse . A study of learners of French and Swedish. Lund: Lund University Press. Gullberg, M. (2006). Handling discourse: Gestures, reference tracking, and communication strategies in early L2. Language Learning , 56(1), 155–196. Gullberg, M. (2008). Gestures and second language acquisition. In N. C. Ellis & P. Robinson (Eds.), Handbook of cognitive linguistics and second language acquisition (pp. 276–305) . London: Routledge. Hadar, U., & Butterworth, B. (1997). Iconic gestures, imagery, and word retrieval in speech. Semiotica, 115 , 147– 172. Iverson, J. M., & Goldin-Meadow, S. (1998). Why people gesture when they speak. Nature, 396 , 228. Iverson, J. M., Capirci, O., & Caselli, M. C. (1994). From communication to language in two modalities. Cognitive Development , 9 , 23–43. Kelly, S. D., Barr, D. J., Church, R. B., & Lynch, K. (1999). Offering a hand to pragmatic understanding: The role of speech and gesture in comprehension and memory. Journal of Memory and Language, 40, 577–592. Kendon, A. (2004). Gesture: Visible Action as Utterance . Cambridge: Cambridge University Press. Ketcham, C. J., & Stelmach G. E. (2001 [1977]). Age-related declines in motor control. In James E. Birren & K. Warner Schaie (Eds.), Handbook of the psychology of aging (5th ed., pp. 313–348). San Diego, CA: Academic Press. Kita, S. (2000). How representational gestures help speaking. In D. McNeill (Ed.), Language and Gesture (pp. 162– 185). Cambridge: Cambridge University Press. Kita, S., & Özyürek, A. (2003). What does cross-linguistic variation in semantic

< previous page

page_214

next page >

< previous page

page_215

next page >

Page 215 coordination of speech and gesture reveal?: Evidence for an interface representation of spatial thinking and speaking. Journal of Memory and Language, 48, 16–32. Laver, J., & Mackenzie Beck, J. (2001). Unifying principles in the description of voice, posture and gesture. In C. Cave, I. Guaitella, & S. Santi (Eds.), Oralité et gestualité: Interactions et comportements multimodaux dans la communication (pp. 15–24). Actes du colloque Orage 2001, Laboratoire Parole et Langage—CNRS et Université de Provence. Paris: L’Harmattan. McNeill, D. (1992). Hand and mind: What gestures reveal about thought . Chicago: University of Chicago Press. McNeill, D. (Ed.). (2000). Language and gesture. Cambridge: Cambridge University Press. McNeill, D., & Duncan, S. D. (2000). Growth points in thinking-for-speaking. In D. McNeill (Ed.), Language and gesture (pp. 141–161). Cambridge: Cambridge University Press. Malatesta, C., Fiore, M., & Messina, J. (1987). Affect, personality and expressive characteristics of older people. Psychology and Aging , 2 , 193–203. Marcos, L. R. (1979). Non verbal behaviour and thought processing. Archives of General Psychiatry, 36(9), 940–943. Montepare, J., Koff, E., Zaitchik, D., & Albert, M. (1999). The use of body movements and gestures as cues to emotions in younger and older adults. Journal of Nonverbal Behavior, 23(2), 133–152. Morris, D., Collet, P., Marsh, P., & O’Shaughnessy, M. (1979) Gestures: Their Origins and Distribution. London: Jonathan Cape. Morsella, E., & Krauss, R. M. (2004). The role of gestures in spatial working memory and speech. American Journal of Psychology, 117 , 411–424. Özyürek, A. (2000). The influence of addressee location on spatial language and representational gestures of direction. In D. McNeill (Ed.), Language and gesture (pp. 64–83). Cambridge: Cambridge University Press. Özyürek, A. (2002). Do speakers design their cospeech gestures for their addressees? The effect of addressee location on representational gestures. Journal of Memory and Language, 46, 688–704. Pashek, G. V., & DiVenere, E. (2006). Auditory comprehension in Alzheimer’s disease: Influence of gesture and speech rate. Journal of Medical Speech–Language Pathology , 14(3), 143–156. Rauscher, F. H., Krauss, R. M., & Chen, Y. (1996). Gesture, speech and lexical access: the role of lexical movements in speech production. Psychological Science , 7 , 226–31. Rekers, G., Sanders, J. A., & Strauss, C. C. (1981). Developmental differentiation of adolescent body gestures. Journal of Genetic Psychology, 138 , 123–131. Sherman, J., & Nicoladis, E. (2004). Gestures by advanced Spanish–English second-language learners. Gesture, 4 (2), 143–156. Singer, M. A., & Goldin-Meadow, S. (2005). Children learn when their teacher’s gestures and speech differ. Psychological Science , 16(2), 85–89. Straten, A. van der (1991). Premiers gestes, premiers mots: Formes précoces de la communication. Paris: Collection Paidos, Centurion. Talmy, L. (1985). Lexicalization patterns: Semantic structure in lexical forms. In T. Shopen (Ed.), Language typology and syntactic description, Vol. 3 (pp. 57–149). Cambridge: Cambridge University Press.

< previous page

page_215

next page >

< previous page

page_216

next page >

Page 216 Tellier, M. (2006). L’impact du geste pédagogique sur l’enseignement-apprentissage des langues étrangères: Etude sur des enfants de 5 ans. Unpublished PhD dissertation, Université Paris 7—Denis Diderot, Paris. Yoshioka, K. (2005). Linguistic and gestural introduction and tracking referents in L1 and L2 discourse. Groningen: Groningen Dissertations in Linguistics.

< previous page

page_216

next page >

< previous page

page_217

next page >

Page 217 9 THE DEVELOPMENT OF SIGN LANGUAGE1 Deborah Chen Pichler The study of sign language development has become a broad field of inquiry over the last half century, benefiting from approaches across a variety of disciplines, including linguistics, psychology, sociology and neuroscience. An interdisciplinary approach helps us tease apart the myriad of factors that can influence sign language acquisition, such as whether the sign language is being learned as a first or second language, the age at which an individual is exposed to the sign language, the quality of sign language input, and many other variables. Although developmental studies have traditionally focused on the acquisition of American Sign Language (ASL) by deaf children from birth to 4 years, more recent work has begun to explore acquisition at later stages of life, and in other natural sign languages. Only by comparing development across a variety of populations and in multiple sign languages will we have the means to draw solid conclusions about the overall processes by which sign languages are acquired and used. Sign language acquisition research is still a fairly new discipline, which means that there have been preliminary studies on a great many topics, but comparatively few follow-up studies or attempts to replicate preliminary findings for other populations or sign languages. That said, there are several areas which have enjoyed a higher degree of research attention than others, or which have benefited from parallel investigations across multiple sign languages and/or multiple populations. This chapter will focus on four such areas of research: phonology, nonmanual markers, fingerspelling and spatial components of narratives. For readers new to sign language research, Section 1 will provide some background information on the structure and organization of sign languages. Section 2 will survey the literature on the development of sign language as a first language by children, by far the most frequently studied group in the sign language acquisition literature. Section 3 will discuss our four focus areas as they pertain to adult second language learners, before moving on to Section 4, where we will address language decline due to Parkinson’s disease. Finally, I will conclude in Section 5 by suggesting some directions for further research.

< previous page

page_217

next page >

< previous page

page_218

next page >

Page 218 1 Sign Languages: Some Background Information Despite the persistent public perception of sign languages as a form of pantomime or simplified English represented manually,2 sign languages are recognized in today’s scholarly literature as fully natural human languages with complex, rule-based organization. There is no universal sign language; deaf communities around the world currently use an estimated 121 national sign languages (Gordon, 2005) that developed and evolved indigenously over time. Although there are many striking similarities across sign languages, some due to historical contact and others due to fundamental iconic properties exploited by visual languages, each sign language possesses a grammatical structure that is distinct both from other sign languages and from local spoken languages. Thus, although ASL and British Sign Language (BSL) both exist in areas where spoken English dominates, they are grammatically distinct from each other and from spoken English. In order to discuss the development of sign languages, it is necessary to understand the basics of sign language structure. While a detailed exposition on sign language grammar is beyond the scope of this chapter, the following summary of sign phonology, nonmanual markers, fingerspelling and spatial aspects of narratives should serve to orient the reader to the basic focus areas central to this chapter. Readers interested in more detailed descriptions for a variety of sign languages are referred to Lucas, Valli and Mulrooney (2005) for ASL, Sandler and Lillo-Martin (2006) for ASL and Israeli Sign Language (ISL), Sutton-Spence and Woll (1999) for BSL, Meir and Sandler (2007) for ISL and Johnston and Schembri (2007) for Australian Sign Language (Auslan). 1.1 Phonology in Sign Languages Despite being visual–gestural languages, sign languages nonetheless display the same types of phonological organization found in spoken languages. Signs are composed of distinctive, sublexical parameters traditionally labeled handshape , movement, location , orientation , and, for some signs, nonmanual features (Stokoe, Casterline, & Croneberg, 1965). The ASL sign illustrated in Figure 9.1, for instance, involves the “5” handshape, with repeated contact (movement) between the thumb of the hand and the chin (location), the palm facing the signer’s left (orientation).3 Each of these sublexical elements, taken on its own, is meaningless, but together they form the ASL sign MOTHER. Individual sublexical categories such as handshape are themselves complex, made up of component features such as tension, selected fingers, joint specification, etc. Efforts continue in the area of sign language phonetics to identify similar componential features within the other sublexical categories such as movement and location (Johnson & Liddell, unpublished manuscript). These component features are likely candidates for errors across

< previous page

page_218

next page >

< previous page

page_219

next page >

Page 219

Figure 9.1 MOTHER in ASL. both first and second language learners of sign languages and, as such, represent an important area for further study. 1.2 Nonmanual Features in Sign Languages Nonmanual features refer to the constellation of facial expressions, eye gaze patterns, and movements of the mouth, head and upper torso accompanying manual signing. These features fall roughly into two categories: grammatical or affective. Grammatical nonmanuals are typically obligatory and linguistically constrained. For example, matrix whquestions in ASL are articulated with a brow furrow (notated in the following example by the label wh), while topics are articulated with a brow raise (top). Each of these nonmanual features must spread across a specific scope, indicated by the line above the ASL glosses, in order for the utterance to be well formed. ________________top ___________ wh (1) POSS-2 BROTHER, NAME WHAT “Your brother, what’s his name?” In contrast, affective nonmanuals, or nonmanuals that convey emotion or affect, are not obligatory in ASL and vary widely across sign production. In

< previous page

page_219

next page >

< previous page

page_220

next page >

Page 220 the following example, smiling functions as an affective nonmanual feature expressing happiness, but there are no specific timing requirements for its onset or offset, and its omission would not render the utterance ungrammatical. ________________smile (2) PRO-2 PREGNANT “You’re pregnant!” Many nonmanual markers in ASL appear almost iconic in that one can see similarities with relevant affective facial expressions (e.g. the ASL wh-marker involves furrowed brows, similar to the affective expression of puzzlement). Indeed, researchers have proposed diachronic paths of development for some nonmanual markers in ASL (Janzen, 1999; MacFarlane, 1998). With respect to linguistic inquiry of nonmanual signals, research has traditionally focused on the eyes, brows and head. Mouth movements have been investigated, but usually only those associated with specific adverbial or affective nonmanuals (sometimes referred to as mouth gestures, e.g. mouthing “cha” to mean “very large” in ASL) that have no discernible source in spoken words. These are widely accepted as genuine features of signed languages, whereas mouthing of spoken words while signing has often been dismissed as unsystematic artifacts of contact with the dominant spoken language. This idea has been challenged as researchers discover systematic patterns in mouthing distribution and function, indicating that at least some types of mouthing are bona fide elements of sign language phonology (see Brentari, 2001, and Boyes Braem & Sutton-Spence, 2001, for discussion). Still, linguistic mouthing continues to go unmentioned in many studies, particularly those focusing on ASL or language acquisition. 1.3 Fingerspelling in Sign Languages In ASL and other sign languages, words borrowed from spoken/written language can be represented by spelling them out using a manual alphabet. This practice is known as fingerspelling. The degree to which fingerspelling is considered an integral part of the language and the prevalence of finger-spelled forms used by skilled signers varies from sign language to sign language. In the French deaf community, for example, fingerspelling is regarded as a foreign component of the sign language, with strong (and somewhat negative) associations to written French; consequently, it is not often used by skilled signers (Michael Filhol, personal communication). In contrast, fingerspelling is an integral part of ASL (Padden, 2006), co-existing productively with signs, not only as a robust mechanism for representing proper names (e.g. T-O-Y-O-T-A) or other vocabulary borrowed from written languages (e.g. S-Y-L-L-A-B-U-S), but also as a source for new compounds (e.g. BLACK+M-A-I-L), abbreviations (e.g. F-L-A for Florida) and initialized

< previous page

page_220

next page >

< previous page

page_221

next page >

Page 221 signs (e.g. the sign MONDAY uses the M-handshape). Fingerspelling also allows signers to differentiate between potentially ambiguous noun–verb pairs. Padden (2006) cites as an example the ASL sign RENT, which only occurs as a verb. If a signer wants to refer to the noun form, the word is fingerspelled: R-E-N-T. In cases like these, one can perhaps think of finger-spelling as providing a neutral alternative to placing a sign with a strongly associated meaning into a context where that meaning would be inappropriate. A similar principle seems to be at work in compounds like BLACK+M-A-I-L. A lexical sign MAIL exists in ASL, but it is highly iconic and calls to mind the action of sticking a stamp onto the corner of an envelope. Given that blackmail has nothing to do with (postal) mail, the use of the sign MAIL in this compound would likely seem quite odd. 1.4 Space and its Use in Narratives As visual–gestural languages, sign languages have the option of exploiting physical space for linguistic purposes, and they do so in highly effective ways. Personal pronouns in sign languages consist of a pointing gesture directed towards its referent. If the referent is not physically present, the signer can still establish a locus for that referent in the signing space; subsequent pronouns or verbs directed towards that locus are then interpreted as applying to that referent. Spatial mechanisms are particularly effective for depicting complex spatial arrangements that might be difficult or impossible to convey using linear strings of standard lexical signs (i.e. the contours of the coast of Maine). For this reason, they are frequently found in sign narratives, such as the excerpt below, adapted from Reilly (2000, p. 424). (3)Direct quote of Baby Bear through referential shift. Note: +K indicates eye contact with addressee, –K indicates a gaze shift away from the addressee. _________ topic ___________________________ surprise __________+K ___________________________ –K BABY BEAR LOOK-AT LET’S-SEE MY BOWL “Baby bear looked at his soup, ‘Let’s see my bowl.’” ____________________________________________pouting ____________________________________________distress ____________________________________________–K HEY GONE SOUP Cl:1 SOMEONE FINISH EAT ALL “Hey, my soup’s gone! Someone ate it all up!” In this retelling of the story of Goldilocks and the Three Bears, the adult signer begins as narrator, then “takes on” the perspective of Baby Bear through a process known as referential shift. Referential shift is the mechanism by which

< previous page

page_221

next page >

< previous page

page_222

next page >

Page 222 direct quotes are produced in ASL and other sign languages (Emmorey & Reilly, 1998; Engberg-Pedersen, 1993, 1995; Morgan, 1999). Just prior to the shift, this signer (still as narrator) labels the character whose perspective he is about to take on; this label is sometimes accompanied by a point to the locus assigned to that character in signing space. Next, the signer breaks eye contact from his addressee, shifts his head and body to the side, assuming the surprised expression of Baby Bear, and signs that character’s utterance. The signer will maintain the character’s affective facial expression for the duration of the quoted utterance (in this case, it spans several clauses). While the signer is engaged in referential shift, pronoun reference is interpreted as being from the point of view of the quoted character. Thus the possessive pronoun MY refers to Baby Bear as possessor rather than the signer. Adults who signed the Goldilocks and the Three Bears story used numerous direct quotes such as the one in (3) and consistently produced head shift and eye gaze in tandem at the beginning of the referential shift function, suggesting these nonmanual changes are obligatory as signals of a shifted perspective (Emmorey & Reilly, 1998; Reilly, 2000). Affective facial expression of the quoted character varied somewhat in its onset point, beginning sometimes with the referential shift and sometimes slightly earlier, but consistently constrained by the length of the quoted material. Rossini, Reilly, Fabretti, and Volterra (1998) note that during dialogues between multiple characters that are already established in space, it is not necessary to label each character before delivering his/her direct quote; changes in nonmanual features are sufficient to signal referential shift from one character to another. 2 Child Signers: First Language Development Before moving on to our discussion of childhood development of sign language, a note on delayed exposure is in order. In the normal case of first language acquisition, children are exposed to their native language(s) from birth. This is, unfortunately, not commonly the case with deaf children, the great majority of whom are born to hearing families who do not know a natural sign language. Delayed exposure to one’s first language clearly has a significant impact on the course of development for both first and second languages. Research has demonstrated that deaf children with delayed exposure to a natural sign language display deficits in their grammatical competence and language processing abilities, as well as potential cognitive deficits (for an excellent summary, see Emmorey, 2002, Chapter 6). In a classic comparison of native signers (exposed to ASL from birth by deaf, signing parents), early learners (first exposed to ASL between 4 and 6 years) and late learners (first exposed to ASL after the age of 12) on a number of morphological and syntactic experimental tasks, Newport (1990) found significant performance gaps between native/early learners and late learners. All sub-

< previous page

page_222

next page >

< previous page

page_223

next page >

Page 223 jects had been using ASL as their primary language for 30 years, regardless of age of exposure, ruling out language experience as the source of this difference. Late-exposed signers also display deficits in phonological processing (Mayberry & Eichen, 1991; Mayberry & Fischer, 1989; Morford & Mayberry, 2000), requiring them to expend more effort to decode and recognize signs than their early-exposed counterparts. This additional demand on the cognitive system slows down access to lexical information, which can in turn impair comprehension. In particular, late learners may experience problems keeping track of spatial referents, as their working memory is already taxed by the task of decoding and recognizing phonological forms (Mayberry, 1995). In the summary of child sign language development in the next section, I will focus only on native and earlyexposed signers. This is in part due to space constraints, and in part due to the widespread opinion that native and early-exposed children provide the clearest picture of “normal” acquisition of sign language. Of course, the notion of “normal” can be turned on its head with respect to deaf children: although one may argue that acquisition from birth represents the ideal and normal pattern observed for all other languages, only a tiny minority of deaf children are born into an environment where such acquisition is possible. Until public policies are enacted to provide all deaf children with early, high-quality access to natural sign languages, the field of sign language acquisition has the charge of elucidating the course of sign development for the overwhelming majority of deaf children who experience delayed exposure to their first language, in addition to the course of “normal” development described below. Readers interested in more detail on the acquisition of sign language by children (both early- and late-exposed) are referred to the many excellent summaries that now exist on the topic, including Newport and Meier (1985), Emmorey (2002, Chapter 5), Bonvillian (1999) and Lillo-Martin (1999). In the following sub-sections, we will survey a sample of this growing literature, focusing on the four topics described in Section 1. 2.1 L1 Development of Phonology The development of sign language phonology is observable even before a deaf child begins to sign. Petitto and Marentette (1991) found that deaf babies between 10 to 14 months who had been exposed to sign language from birth produced more complex handshape and movement patterns than age-matched babies exposed only to speech. They called the manual behavior observed in sign-exposed babies manual babbling , and noted that frequently recurring patterns from the babbling behavior carried over into the babies’ first formal signs. For example, just as speech-exposed babies show a preference for certain classes of consonants (e.g. stop consonants) in both

< previous page

page_223

next page >

< previous page

page_224

next page >

Page 224 their babbling and their earliest words (Oller, Wieman, Dole, & Ross, 1976), sign-exposed babies favor the same set of handshapes in both their manual babbling and their first signs (cf. also Cheek, Cormier, Repp, & Meier, 2001). Petitto and Marentette (1991) pointed to such similarities between vocal babbling and manual babbling as an argument for abmodality of the human language faculty. Once a child has begun producing formal signs, we enter the realm of phonology proper. Although reports on early sign phonology exist for a variety of sign languages, most are based on very small data sets, due in part to logistical difficulties in locating sign-exposed children who are available for long-term study. Nonetheless, several patterns of development are observed repeatedly, across studies of different sign languages. Not surprisingly, sign-exposed children produce frequent errors of substitution (usually in the direction of less marked forms), most commonly with respect to handshape, and least commonly with respect to location. This general pattern of acquisition has been reported for a variety of sign languages (Conlin, Mirus, Mauk, & Meier, 2000, and Meier, 2006, for ASL, among others; Clibbens & Harris, 1993 for BSL; Karnopp, 2002, for Brazilian Sign Language; von Tetzchner, 1994 for Norwegian Sign Language; Takkinen, 2003 for Finnish Sign Language). This contrast between handshape and location accuracy has several plausible reasons. First, sign locations are fewer in number and more tolerant of variation than sign handshapes, for which a small variation could easily result in a different handshape. Thus, the statistical probability of a handshape error is higher than that of a location error. Second, accurate production of handshapes (e.g. distinguishing the U-handshape from the R-handshape) involves the fine muscle groups of the fingers, whereas accurate production of location (e.g. distinguishing between contact on the chest and contact on the forehead) involves much larger muscle groups. This same contrast also has potential effects on perception; children may initially have difficulty distinguishing similar handshapes in the input, leading them to substitute a single configuration for multiple targets in their production. The developing motor skills of young children influence not only handshape accuracy, but other aspects of early phonology as well. Richard Meier and his colleagues (Meier, 2006; Meier, Cheek, & Moreland, 2002; Cheek et al., 2001, among others) identified general properties of early motor development that potentially influence the development of sign phonology. Observing that all infants, regardless of language input, produce repeated movements, they predict that sign-exposed children will correctly produce target signs requiring repetition, and incorrectly add repeated movement to target signs that do not. Infants also have difficulty inhibiting the use of one hand while the other is active, a tendency that leads to sympathetic movement in signs where one hand should normally be held still. Many two-handed signs require the non-dominant “base hand” to remain static, serving as a surface on which the

< previous page

page_224

next page >

< previous page

page_225

next page >

Page 225

Figure 9.2 Cookie in ASL. dominant “active hand” acts. For example, the ASL sign for COOKIE, shown in Figure 9.2, is articulated with a twisting motion of the active bent-5 handshape (or lax C-handshape) on a static B-handshape base hand. The tendency to produce sympathetic movement might lead children to articulate COOKIE with identical handshapes (e.g. two bent-5 handshapes) and/or identical movements (e.g. both hands twisting). Cheek et al. (2001) reported just such articulation errors in their data, as well as apparent avoidance of two-handed base-hand signs by their infant subjects, a pattern that has also been observed in Brazilian Sign Language (LSB, Karnopp, 2002), BSL (Morgan, Barrett-Jones, & Stoneham, 2007), and Finnish Sign Language (Takkinen, 2003). The third and final property of motor development proposed by Meier and colleagues to influence sign phonology is proximalization, the tendency to substitute an articulator close to the torso for one that is further away. Meier (2006) describes errors in which children reliably used a more proximal active joint than was called for in the target sign. For instance, when signing HORSE, which in ASL involves a nodding motion of the hand originating at the first set of knuckles, children produced the nodding motion from the wrist, a more proximal joint. Proximalization errors have been noted in the early development of sign languages besides ASL (Takkinen, 2003, for Finnish Sign Language; Lavoie & Villeneuve, 1999, for Québec Sign Language, or Langue des Signes Québécoise, LSQ) and, as we will discuss shortly, has also been identified as a common error in L2 sign production by adult learners (Mirus, Rathmann, & Meier, 2001). 2.2 L1 Acquisition of Nonmanuals As discussed in Section 1.1, sign language grammars involve not only a manual component, but a nonmanual component as well. With respect to

< previous page

page_225

next page >

< previous page

page_226

next page >

Page 226 language development, one intriguing question is whether sign-exposed children are sensitive to the similarities that adults see between grammatical nonmanual markers and related affective facial expressions. If they are, they might potentially use affective facial expressions, which are acquired early, by the first year of life (Nelson, 1987), to “break into” the grammatical system of the target sign language. This question has been explored extensively by Judy Reilly and her colleagues over a number of studies (Anderson & Reilly, 1997, 1998; Reilly, 2000; Reilly, McIntire, & Bellugi, 1990a, 1990b, 1991), focusing on a range of ASL grammatical nonmanuals, including those for negation, wh- and yes–no questions, conditionals, topics, and adverbials. Results indicate that although children control affective facial expression early, their development of grammatical nonmanuals is more protracted and problematic, stretching into the fifth year and beyond (Reilly, McIntire, & Bellugi, 1990b; Reilly, 2006). Reilly and her colleagues argue that their findings point to several generalizations about the acquisition of sign. First, children obey the principle of unifunctionality (Slobin, 1973) in their development of grammatical nonmanuals, initially assuming a one-to-one mapping of grammatical form to function, and resisting marking multiple construction types with the same marker. This point is nicely illustrated by the case of brow raise in Reilly and colleagues’ data. Brow raise is a salient feature of several grammatical nonmanuals in ASL, including those for conditionals, yes–no questions, and topics. Prior to the age of 3 years, children have begun producing all three of these constructions, but only yes–no questions are correctly marked with brow raise. Both topics (or, more accurately, proposed objects that are plausible candidates for topics) and conditionals lack the obligatory nonmanual marking. However, the child is still able to mark these two structures, in the case of conditionals, thanks to the availability of lexical markers of conditionality (e.g. the signs SUPPOSE or #IF) that are optional in the adult system. At around 3;0, brow raise is observed marking topic structures, suggesting that children have relaxed their insistence on unifunctionality and are ready to consider using a single marker for more than one function. Second, children show a “hands before face” bias, treating the hands as the primary articulators of the language. They control the manual component of ASL grammar (e.g. producing signs with the proper form and in proper order) before the nonmanual component, and they choose manual over nonmanual marking in instances when both are available, despite the fact that omission of the nonmanual marker is ungrammatical (e.g. in the case of conditional constructions discussed above).4 Taken together, these two observations suggest that affective and grammatical nonmanual systems develop independently. Children’s early competence in affective facial expression does not lead to smooth and early development of grammatical nonmanuals. The latter begins between 1;6 and 2;0 in most cases, and

< previous page

page_226

next page >

< previous page

page_227

next page >

Page 227 stretches into the sixth year or later in some special cases (Lillo-Martin, 2000; Reilly, McIntire, & Bellugi, 1990b). 2.3 L1 Acquisition of Fingerspelling As mentioned in Section 1, fingerspelling is very common in ASL. For children born to deaf, signing parents, fingerspelling is present in their everyday environment. Padden (1991) reports that children as young as 2;0 spontaneously attempt to produce fingerspelled words for common objects such as I-C-E or T-V. These early attempts approximate the very salient movement contour of fingerspelled words, but lack a clear internal sequence of individual letter configurations (Akamatsu, 1982; Padden 2006; Padden & LeMaster, 1985). At this age, the children do not yet know how to read or write, but they have begun to develop what Padden calls the “first skill of fingerspelling,” learning the types of words represented by fingerspelling, and recognizing common fingerspelled words in the input. The “second skill of fingerspelling” is closely tied to development in English literacy, as it involves recognizing the internal structure of a fingerspelled word as a series of individual hand configurations corresponding to letters of the English alphabet. Full acquisition of fingerspelling requires mastery of both skills, a process that begins early but that stretches well into the school years. Investigations of reading development in deaf children point to a correlation between English reading comprehension skills and the ability to write down fingerspelled words (Padden & Hanson, 2000; Padden & Ramsey, 1998). Other studies have uncovered “indigenous” strategies intuitively employed by native signing parents and teachers that enhance children’s access to fingerspelled words (e.g. alternating fingerspelled words with signs related to that word). Such research holds great potential to inform and improve classroom practices in deaf education. 2.4 L1 Acquisition of Spatial Components of Narratives Before we discuss the narratives produced by child signers, we must briefly look at the development of spatial mechanisms in children’s emerging syntax. Much of the early literature on sign language development reports delays in the acquisition of pronouns and verb agreement with respect to spoken language development. Although children are able to direct pronouns and verbs towards present referents between 3;0 and 3;6 (Emmorey, 2002), and understand by 3;0 to 4;0 that non-present referents can be assigned locations in signing space (Bellugi, van Hoek, Lillo-Martin, & O’Grady, 1993; Lillo-Martin, Bellugi, Struxness, & O’Grady, 1985), they are not yet able to themselves direct pronouns and verbs towards non-present referents (Hoffmeister, 1978; Loew, 1984). Lillo-Martin (1999) suggests that the complexities of learning how to associate and maintain referents in space are the source of the observed delay, not a lack of morphological readiness.

< previous page

page_227

next page >

< previous page

page_228

next page >

Page 228 The difficulties of establishing and maintaining associations between non-present referents and their loci in space lead to interesting errors, readily apparent in children’s development of narratives. Loew (1984) and Bellugi et al. (1993) report that between 3;6 and 5;0, children often lose track of spatial relations over the course of a narrative, either “stacking” multiple referents on the same locus in signing space, or assigning a single referent to more than one loci. The resulting narrative lacks cohesion and leads to viewer uncertainty as to the antecedent of pronouns. This is strikingly similar to what has been reported for early spoken English narratives, in which children produce ambiguous pronouns (Karmiloff-Smith, 1986). In both cases, pronoun reference may be well-formed within an individual sentence, but children under 5;0 seem generally unable to relate coordinate reference across the multiple sentences of a narrative, resulting in a lack of cohesion. As for referential shift/direct quoting, Loew (1984) reported that the subject in her case study, Jane, was able to use changes in eye contact to mark referential shift by 3;6, although she failed to differentiate between multiple referents or change facial expression at the onset of a new shift. By 4;4 she consistently marked referential shift with eye gaze changes, but was still unable to coordinate multiple spatial referents over the course of a narrative.5 Jane was finally able to coordinate body shift and eye gaze to signal a shift, as well as to maintain multiple spatial referents across the narrative. However, once engaged in referential shift, she had difficulty assigning pronouns, that is, understanding that a first person pronoun no longer refers to the narrator when produced while quoting another character (Loew, 1984). A much larger and more controlled study of referential shift was conducted by Reilly and colleagues (Reilly, 2000) on 28 native deaf children ranging in age from 3;0 to 7;0. Their findings were generally in line with those reported by Loew (1984). In particular, eye gaze was used as the earliest indicator of referential shift, employed even by the 3year-old subjects. All age groups also used affective facial expression in their shifts. However, children younger than 6 years were inconsistent in their use of facial expressions, producing them with incorrect timing or extending the same affective non-manual across the direct quotes of more than one character. Children under 6;0 were also unreliable in their use of labels for quoted characters, compared to adult controls. A final divergence in the child data from adult controls was the use of the sign SAY to lexically indicate direct quotes. This strategy was used most frequently by the 5-year-old signers, in addition to eye gaze and affective facial expression. Reilly (2000) proposes that the appearance of SAY at 5 years is indication of a linguistic reorganization occurring. While children at 5;0 use affective facial expression in direct quotes, they still do not do so consistently or accurately, so SAY offers them an alternative way to mark these structures. The fact that deaf children apparently invent a new lexical marker for

< previous page

page_228

next page >

< previous page

page_229

next page >

Page 229 structures that they have difficulty marking nonmanually fits nicely with the findings reported by Reilly and her colleagues for the acquisition of conditionals, summarized earlier in this section. 3 Adult Signers: Second Language Acquisition of Sign Language In contrast to the rich and quickly expanding literature on childhood L1 development in a number of sign languages, there is only a very small and scattered literature on adult L2 signers. As we will see below, this literature focuses mainly on phonological aspects of non-native ASL, and on the ability of native signers to identify non-native signers on the basis of them. 3.1 Phonology in L2 Sign Language Second language research for spoken languages has identified phonology as one aspect of language for which a critical or sensitive period clearly applies. Studies as early as Scovel (1969) and Oyama (1976) reported that individuals beginning the L2 acquisition process before puberty generally spoke the target language with less “foreign accent” than those who began the process after puberty. Scovel (1981) further demonstrated that native speakers are able to reliably identify a foreign accent by the age of 10. The concept of “foreign accent” can be broadly defined as non-targetlike phonology, and is thus as applicable to sign languages as it is to spoken languages. Indeed, there has long been a popular perception that skilled signers can identify non-native signers on the basis of their “accent.” This claim appeared in an early paper by Cicourel and Boese (1972), but without empirical basis. Kantor (1978) set out to test whether identification of native vs. nonnative signers was possible on the basis of simply watching them sign, and whether any particular signer profile (e.g. native deaf vs. native hearing vs. non-native deaf vs. non-native hearing) was more successful than others in detecting non-native accent. Her results showed that native-deaf, native hearing and non-native hearing “judges” (signers who viewed the film clips and were asked to identify the profile of each signer) were all able to identify signers at the extremes of the ASL fluency continuum (i.e. native deaf and non-native hearing signers). Native deaf judges did so with the highest accuracy, followed closely by native hearing judges, then hearing non-native judges. The remaining category of judges, non-native deaf signers, was able to identify fellow non-native deaf signers, but not any other signer profile. Approaching the data from the perspective of which group was most accurately identifiable across judges, Kantor found that native deaf and non-native hearing signers were accurately identified the most often by all categories of judges (except the non-native L2 judges); both native hearing and non-native deaf signers were often mislabeled as native deaf. These results

< previous page

page_229

next page >

< previous page

page_230

next page >

Page 230 led Kantor to two main conclusions regarding “accent” in sign language: (a) the “accent” of native signers is consistent, regardless of hearing status, and (b) with enough practice, certain signers (i.e. non-native deaf) seem able to escape critical period effects and learn to sign with a native “accent.” Kantor also elicited feedback on specific aspects of signing that judges used to determine the hearing/deaf and native/non-native status of signers. The two features most frequently cited by all categories of judges were handshape and facial expression, followed by “rhythm” and choice of signs. Many judges also cited English mouthing as an indication of non-nativeness, and use of “mime” or depictive signing as an indication of nativeness. A similar study was conducted by Budding, Hoopes, Mueller, and Scarcello (1995), but with the focus on signers’ ability to identify foreign “accent” rather than non-native accent. In this study, deaf native ASL judges watched narratives signed in ASL by deaf individuals who had learned LSQ as a first language (although with significant delays in several cases, as late as 14 years) and ASL much later, as a second language. Half of the judges were informed at the onset that they were about to watch LSQ signers from Quebec. Predictably, these judges were the most likely to report that they noticed a foreign accent. Both sets of judges cited numerous features that contributed to a native accent in ASL, with particular emphasis on effective use of space, nonmanuals and classifier constructions. Drawbacks of the Kantor (1978) and Budding et al. (1995) studies are that they took on research questions that were very broad and involved a high degree of variation in the content of the video samples (in the case of Kantor, 1978). They also included signers who acquired their L1 later than usual, a population who have since been shown to process and produce signing differently from native signers. As a result, it is difficult to draw strong conclusions from these early studies regarding features that contribute to the perception of non-native accent in sign language. More recent investigations are more specific in scope and yield more generalizeable conclusions. Mirus et al. (2001) reported proximalization errors in isolated ASL signs produced by hearing L2 learners with no previous experience with sign language, parallel to those they found in infant L1 signing. Proximalization errors were also noted by Rosen (2004) in his examination of phonological errors produced in isolated lexical ASL signs by his L2 subjects. In contrast, native deaf German signers of German Sign Language ( Deutsche Gebärdensprache , DGS) who were asked to reproduce isolated ASL signs proximalized much less frequently than hearing L2 signers (Mirus et al., 2001). Mirus et al. conclude that proximalization is not simply the result of an immature motor system, but rather a universal reaction to learning to use the body in a new way. Mirus (personal communication) supports this proposal by pointing out that the awkward “too big” movements of adults skiing for their first time or writing with their nondominant hand are also due to proximalization.

< previous page

page_230

next page >

< previous page

page_231

next page >

Page 231 Rosen (2004) examines phonological errors produced in isolated lexical ASL signs by beginning students of ASL as a second language. After categorizing errors according to the phonological parameter most affected, Rosen attempts to determine whether the errors were caused by a lack of dexterity (i.e. substitutions, displacements, switches and incomplete formation) or incorrect perception by the student (i.e. mirrorizations, parallelizations, and addition or deletion of articulary segments). He concludes that early L2 handshape errors are caused solely by a lack of dexterity, while errors in location, movement and palm orientation are caused by both lack of dexterity and problems in perception. Chen Pichler (2006) explores the contributions of universal markedness patterns and transfer from L1 (i.e. conventionalized gestures common in American hearing culture) on handshape errors, two potential sources for nonnative accent that were not investigated by Rosen (2004). Preliminary results suggest that sign-naïve subjects did sometimes transfer familiar handshapes from their gestures to sign, but only for relatively unmarked handshapes. Marked handshapes such as R and Y were produced accurately in subjects’ gestures (“good luck/cross your fingers” and “call me,” respectively), but not in subsequent ASL signs requiring those handshapes. Chen Pichler speculates that naïve signers are generally able to visually extract handshape as an independent component of an unfamiliar sign (thereby facilitating transfer if the handshape is part of their gestural repertoire), but this ability is somehow blocked when the target handshape is highly marked. Thus markedness and transfer may interact to affect L2 sign phonology in ways previously overlooked. 3.2 Nonmanuals and Fingerspelling in L2 Sign Language There has been very little research on L2 sign development beyond the few studies of L2 sign phonology summarized above. Most are quite general in nature, collecting student and teacher opinions on various aspects of learning sign language as a second language. McKee and McKee (1992) report that adult students of ASL rate nonmanuals and fingerspelling as among the most difficult aspects of ASL to acquire. Difficulties with nonmanual markers are also observed by McIntire and Reilly (1988) in their study of American college students with two to three semesters’ worth of ASL classes. Subjects were asked to imitate signed sentences involving a variety of nonmanuals, some linguistic and with specific scope (i.e. grammatical nonmanual markers, adverbial nonmanual markers and nonmanuals associated with specific lexical items) and some paralinguistic with variable scope (i.e. affective non-manual expressions). Results indicated that subjects performed better with affective nonmanuals than linguistic nonmanuals. With respect to the latter, less-experienced learners initially employed a gestalt strategy, producing unanalyzed combinations of

< previous page

page_231

next page >

< previous page

page_232

next page >

Page 232 manual and nonmanual material. More experienced signers had reanalyzed these gestalt combinations and their manual and nonmanual components, but are not able to coordinate them with the correct timing and scope.6 In some cases, subjects incorrectly reanalyzed adverbial nonmanuals as optional, and omitted them from their imitations. L2 learners exhibited many of the same developmental patterns reported for L1 acquisition of ASL nonmanuals carried out by Reilly and her colleagues (described in Section 2.2). 4 Breakdown of Sign Language As we have seen, signing involves not only the hands, but the face and body as well. As the body ages, it is logical to expect that decreased physical strength and dexterity may influence the phonology of signed production. Aging may bring changes to the syntactic and pragmatic components of sign language as well, as signers experience problems with memory or processing. Unfortunately, the process of sign language attrition due to aging has not yet been empirically studied, to my knowledge. However, there is a healthy literature on the effects of disease or damage to the brain on sign language perception and production. These studies, in particular those focusing on sign formation by deaf Parkinson’s patients, allow us to begin to understand the process of sign language decline, an important but often neglected counterpart to developmental studies. 4.1 Language Decline: Parkinsonian Signers Parkinson’s disease affects the functioning of the basal ganglia, resulting in profound motor disturbances. Typical symptoms include tremor of limbs, slowed movements and a lack of facial expression. In a series of experiments, Poizner and his colleagues (Brentari & Poizner, 1994; Poizner & Kegl, 1993; Poizner, Brentari, Tyrone, & Kegl, 2000; Tyrone, Kegl, & Poizner, 1999, among others) have demonstrated that the formational features of Parkinsonian signing manifest profound and systematic disruptions. Poizner and his colleagues categorize these errors into two broad categories: reductions and timing disruptions (Brentari & Poizner, 1994). The most robust type of reduction in Parkinsonian signing is distalization, wherein movement is shifted from the target articulator to one that is further from the torso (this is the opposite effect from proximalization, observed for L2 and young L1 signers). Thus the sign for BETTER, normally signed with an upward diagonal movement of the hand that is articulated at the elbow, is articulated at the wrist and knuckles (Brentari & Poizner, 1994). A secondary effect of distalization is that the entire movement of the sign is contained in a smaller signing space than normal. The reduced signing space may also entail displacement of location. For example, Parkinson’s signers often drop signs that normally occur near the face to a more neutral level in

< previous page

page_232

next page >

< previous page

page_233

next page >

Page 233 front of the lower torso. Another type of reduction targets handshape, specifically, the selected (i.e. active) and unselected (i.e. less active) finger features. Normally, unselected fingers in a given handshape are specified as either fully “open” (extended) or fully “closed.” For example, in the “baby-O” handshape, the thumb and index describe a ring configuration, while the unselected fingers (i.e. the rest of the fingers) are fully closed. Brentari and Poizner (1994) noted that Parkinson’s signers produced “lax” or imprecise handshapes in which the unselected fingers are somewhere in between open and closed. Other reductions observed in Parkinsonian signing include shadowing (i.e. assimilation of movement from the dominant hand to the non-dominant hand), deletion of contact (i.e. the dominant hand stops short of its target contact location), and neutralization of orientation (i.e. altering palm orientation towards the mid-sagittal plane). Nonmanual expressions are also affected, appearing dampened or replaced altogether by the rigid, “masked” facial expression characteristic of Parkinson’s patients. Timing disruptions in Parkinsonian signing include excessive pause durations at the beginning of utterances, as well as mechanical acceleration or deceleration of subsequent pauses within the utterance. The normal rhythm of signing is lost, obscuring constituent borders that would normally be marked by brief pauses. Parkinsonian signers also take proportionally longer to effect changes in handshape while the hand is transitioning from one position to another. For instance, the sign for “flood” in ASL is a compound of WATER, articulated with a W handshape at the chin, and RISE articulated with a variant of the B handshape in front of the torso. In normal production of this compound, the hand changes from the W to the B handshape during the transitional movement from the chin to the area in front of the torso. Brentari and Poizner (1994) report that their control signer executed the W-to-B handshape change within the first 17% of the transitional movement, whereas their Parkinsonian signer took 30% of the transition movement to achieve the same task. Brentari, Poizner, and Kegl (1995) further observed that Parkinsonian signers produced about the same handshape change to movement ratio at word boundaries as they did within words (i.e. between two parts of a compound), giving their signing a monotonous quality. Together, reduction and timing disruptions have a particularly striking effect on handshapes, and correspondingly, on fingerspelling. Indeed, finger-spelling is particularly vulnerable to the effects of Parkinson’s disease, exhibiting not only the lax position of unselected fingers just described, but also blending and feature unraveling, two additional types of reduction. Blending occurs when the selected fingers feature for one letter handshape is combined with that of an adjacent letter handshape. For example, Tyrone et al. (1999) describe an instance of the fingerspelled word PI-L-L in which the deaf Parkinsonian signer produced P and I simultaneously, coalescing the

< previous page

page_233

next page >

< previous page

page_234

next page >

Page 234 two handshapes into one. In cases of feature unraveling, a single feature of one letter handshape is decoupled from the others during the transition between letters. For instance, while signing the last two letters of the sequence A-SL, one Parkinsonian signer waited until the thumb component of the L was almost fully extended before extending the index finger to complete the fingerspelled letter (Tyrone et al., 1999). This lack of coordination can seriously disrupt the flow of the fingerspelled word, since it appeared as though an extra handshape had been inserted between the S and the L. Poizner and colleagues argue that despite the extensive effects of Parkinson’s disease on sign production, the errors produced by this population are phonetic rather than phonological (Brentari & Poizner, 1994). Support for this proposal comes from the fact that the reductions and disruptions in timing discussed above all favor ease of articulation, consistent with documented difficulties that Parkinson’s patients have in controlling movement. This is quite a different pattern than is observed for deaf aphasics, for instance, who suffer breakdown at the phonological level of structure. Aphasics are often able to articulate signs clearly and sharply; their errors do not arise from the need to ease articulation, but rather from errors in selecting phonological targets (Corina et al., 1992; Poizner, Klime, & Bellugi, 1987). Aphasics also display serious problems in syntactic or morphological processing and production (paraphasias), not associated with Parkinson’s disease; deaf Parkinsonians employ the full range of sentence types, morphological and classifier forms typical of normal signers, and correctly use referential space (Brentari et al., 1995; Kegl & Poizner, 1997; Poizner & Kegl, 1993). Thus it appears that even though Parkinsonian signing is heavily compromised at the production level, their language capacity at the grammatical level is spared, unlike in the case of aphasics. 5 Conclusion and Directions for Future Research In this brief and limited survey, we have seen that sign language development, while certainly rapid and dynamic during the first few years of life, continues to be a worthy object of inquiry through later childhood and into adulthood and old age. We have noted a few developmental patterns that emerge for both child and adult signers, such as a tendency to proximalize movement, and difficulty coordinating manual and nonmanual signals. The fact that we have not noticed more similarities between child and adult language acquisition, or between errors in early development and errors due to aging and language decline, is surely due to a lack of research, rather than any lack of underlying similarities. The field of sign language acquisition is still relatively new, with many unexplored corners. Certainly, over the last half century we have made great strides in our understanding of sign language development. New technology and experimental methodologies are making it possible to perform studies on younger

< previous page

page_234

next page >

< previous page

page_235

next page >

Page 235 children than was previously possible, and to probe the regions of the brain where language is controlled. More cross-linguistic investigations are being conducted, making important comparisons between the traditionally wellstudied sign languages of Europe and the Americas with sign languages from Asia, Africa, Australia and elsewhere. Still, there are gaps in our research program, some of them quite big, that are still waiting to be filled in. With respect to childhood development of sign language, three areas in particular stand out as needing further study. The first, development of sign by late-exposed learners, I have already discussed in the beginning of Section 2. The second area is L1 sign language development in later childhood. Some possible candidates for study (using similar studies from the spoken acquisition literature as a guide) include advanced and low-frequency vocabulary, advanced forms of subordination, metaphors, humorous language and figurative speech. A third area that has traditionally been neglected is sign bilingualism. Hearing children are common in deaf families and many of them are bimodal bilinguals, acquiring both signed and spoken language. Yet with a few exceptions (i.e. Petitto & Holowka, 2002; van den Bogaerde, 2000), there have been no large-scale studies on how bimodal bilinguals develop their native languages, how these two languages interact, and how development in each language diverges from development in sign-only or speech-only monolinguals. There have been even fewer studies on bilinguals learning two sign languages at once. Turning now to adult signers, there are very few studies on the acquisition of sign languages as an L2, either by hearing learners or by deaf signers already proficient in a different sign language. Given the recent rise in popularity of sign languages in schools and universities in the US and elsewhere, it would seem in our best interest to understand the processes underlying second language development of sign languages. On a related note, I know of no studies examining the acquisition of multiple “varieties” or dialects of the same sign language. Sociolinguistic variables such as gender, age, region, socioeconomic status, etc., have been identified as factors contributing to phonological, lexical and syntactic variation in ASL (Lucas, Bayley, & Valli, 2001, 2003) and Australian Sign Language (Schembri, Johnston, & Goswell, 2006), but to my knowledge, there has not been any systematic study on the acquisition of these dialectal differences by native signers: for example, changes in one’s original dialect due to prolonged contact with another dialect. Finally, we know little about the process of sign language attrition, either as a result of the natural aging process, or (in the case of grown hearing children of deaf parents) departure from the signing environment of one’s childhood. L1 signers who move to an area where a different sign language is used may also experience attrition to their L1 sign language, as may L2 learners whose exposure to sign language is suspended for a long period of time (e.g. between academic terms at university).

< previous page

page_235

next page >

< previous page

page_236

next page >

Page 236 This list of suggested directions for further research is of course far from comprehensive, and it reflects my biases as a linguist studying language acquisition. Still, I hope it will convince students and other future sign language researchers of how much still remains unknown in this field, and how much they stand to contribute, should they choose to pursue it. Notes 1. I wish to thank the many colleagues who patiently answered my questions and offered helpful suggestions while I was writing this manuscript: Richard Meier, Ginger Pizer, David Quinto-Pozos, Martha Tyrone, Jenny Singleton, Paul Preston, Michele Bishop, Anne Baker, and Karen Emmorey. Thank you also to Julie Hochgesang for the photo illustrations. All errors are, of course, my own. 2. Signed versions of English such as Signed Exact English (SEE) are artificial systems invented by educators as a way to present English to deaf students. They do not qualify as natural sign languages and are not acquired in the same way as natural sign languages. 3. This is assuming the signer is right-handed, as the majority of signers are. In the case of a left-handed signer, this sign is articulated with the left hand, the palm accordingly facing the signer’s right. 4. There is even evidence that deaf parents share the “hands before face” bias, at least in child-directed signing (see Reilly & McIntire, 1991). 5. Note that the inability to maintain multiple spatial referents in one’s production does not preclude the ability to comprehend them in another’s signing (cf. the Bellugi et al. (1993) and Lillo-Martin et al. (1999) studies cited earlier). 6. Difficulties in coordinating grammatical nonmanuals with lexical materials have also been reported by Morgan, Smith, Tsimpli, and Woll (2002) in the BSL production of a polyglot savant, Christopher. References Akamatsu, C. (1982). The acquisition of fingerspelling in pre-school children. Unpublished doctoral dissertation, University of Rochester, NY. Anderson, D., & Reilly, J. (1997). The puzzle of negation: How children move from communicative to grammatical negation in ASL. Applied Psycholinguistics , 18, 411–429. Anderson, D., & Reilly, J. (1998). Pah! The acquisition of adverbials in ASL. Sign Language and Linguistics , 1 , 117– 142. Bellugi, U., van Hoek, K., Lillo-Martin, D., & O’Grady, M. L. (1993). The development of spatialized syntactic mechanisms in American Sign Language. In D. Bishop & K. Mogford (Eds.), Language development in exceptional circumstances (pp. 132–149). Edinburgh: Churchill Livingstone. Bonvillian, J. (1999). Sign language development. In M. Barrett (Ed.), The development of language. East Sussex, UK: Psychology Press. Boyes Braem, P., & Sutton-Spence, R. (Eds.). (2001). The hands are the head of the mouth: The mouth as articulator in sign languages. Hamburg: Signum-Verlag. Brentari, D. (Ed.) (2001). Foreign vocabulary in sign languages. Mahwah, NJ: Lawrence Erlbaum. Brentari, D., & Poizner, H. (1994). A phonological analysis of a deaf Parkinsonian signer. Language and Cognitive Processes, 9 (1), 69–100.

< previous page

page_236

next page >

< previous page

page_237

next page >

Page 237 Brentari, D., Poizner, H., & Kegl, J. (1995). Aphasic and Parkinsonian signing: Differences in phonological disruption. Brain and Language, 48, 69–105. Budding, C., Hoopes, R., Mueller, M., & Scarcello, K. (1995). Identification of foreign sign language accents by the deaf. In L. Byers & M. Rose (Eds.) Gallaudet University Communication Forum, Vol. 4 (pp. 1–16). Washington, DC: Gallaudet University Press. Cheek, A., Cormier, K., Repp, A., & Meier, R. (2001). Prelinguistic gesture predicts mastery and error in the production of first signs. Language, 77, 292–323. Chen Pichler, D. (2006). Handshape in L2 ASL: Effects of markedness and transfer . Presented at the 9th Congress on Theoretical Issues in Sign Language Research (TISLR 9), Florianópolis, Brazil. Cicourel, A., & Boese, R. (1972). Sign language acquisition and the teaching of deaf children. In C. Cazden, V. John, & D. Hymes (Eds.), Function of language in the classroom. New York: Teachers College. Clibbens, J., & Harris, M. (1993). Phonological processes and sign language development. In D. Messer & G. Turner (Eds.), Critical influences on child language acquisition and development . London/New York: Macmillan/St Martin’s Press. Conlin, K., Mirus, G., Mauk, C., & Meier, R. (2000). The acquisition of first signs: Place, handshape, and movement. In C. Chamberlain, J. Morford, & R. Mayberry (Eds.), Language acquisition by eye (pp. 51–69). Mahwah, NJ: Lawrence Erlbaum. Corina, D., Poizner, H., Bellugi, U., Feinberg, T., Dowd, D., & O’Grady-Batch, L. (1992). Dissociation between linguistic and nonlinguistic gestural systems: A case for compositionality. Brain and Language, 43, 414–447. Emmorey, K. (2002). Language, cognition and the brain: Insights from sign language research. Mahwah, NJ: Lawrence Erlbaum. Emmorey, K., & Reilly, J. (1998). The development of quotation and reported action: Conveying perspective in ASL. In E. Clark (Ed.), Proceedings of the Stanford Child Languages Forum (pp. 81–90). Stanford, CA: Center for the Study of Language and Information Publications. Engberg-Pedersen, E. (1993). Space in Danish Sign Language: The meaning and morphosyntax of the use of space in a visual language. Hamburg, Germany: SignumVerlag. Engberg-Pedersen, E. (1995). Point of view expressed through shifters. In K. Emmorey & J. Reilly (Eds.), Language, gesture, and space (pp. 133–154). Hillsdale, NJ: Lawrence Erlbaum. Gordon, R., Jr. (Ed.), (2005). Ethnologue: Languages of the world (15th ed.). Dallas, TX: SIL International. Retrieved December 3, 2007, from www.ethnologue.com/. Hoffmeister, R. (1978). The development of demonstrative pronouns, locative and personal pronouns in the acquisition of American Sign Language by deaf children of deaf parents. Unpublished doctoral dissertation, University of Minnesota. Janzen, T. (1999). The grammaticization of topics in American Sign Language. Studies in Language, 23(2), 271–306. Johnson, R., & Liddell, S. (2008). Sign language phonetics: Architecture and description . Unpublished manuscript. Johnston, T., & Schembri, A. (2007). Australian Sign Language (Auslan): An introduction to sign language linguistics . Cambridge, UK: Cambridge University Press. Kantor, R. (1978). Identifying native and second language signers. Communication and Cognition, 11, 39–55.

< previous page

page_237

next page >

< previous page

page_238

next page >

Page 238 Karmiloff-Smith, A. (1986). Some fundamental aspects of language development after age 5. In P. Fletcher & M. Garman (Eds.), Language acquisition (2nd ed.) (pp. 456–474). Cambridge, UK: Cambridge University Press. Karnopp, L. (2002). Phonology acquisition in Brazilian Sign Language. In G. Morgan & B. Woll (Eds.), Directions in sign language acquisition (pp. 29–53). Amsterdam: John Benjamins. Kegl, J., & Poizner, H. (1997). Crosslinguistic/crossmodal syntactic consequences of left-hemisphere damage: Evidence from an aphasic signer and his identical twin. Aphasiology , 11, 1–37. Lavoie, C., & Villeneuve, S. (1999). Acquisition du lieu d’articulation en Langue des Signes Québécoise: Etude de cas. In Variations: le langage en théorie et en pratique. Actes du colloque: Le colloque des étudiants en sciences du langage. Montreal: University of Quebec at Montreal. Lillo-Martin, D. (1999). Modality effects and modularity in language acquisition: The acquisition of American Sign Language. In T. Bhatia & W. Ritchie (Eds.), Handbook of child language acquisition (pp. 531–567). San Diego, CA: Academic Press. Lillo-Martin, D. (2000). Aspects of the syntax and acquisition of wh-questions in American Sign Language. In K. Emmorey & H. Lane (Eds.), The signs of language revisited: An anthology in honor of Ursula Bellugi and Edward Klima (pp. 401–414). Mahwah, NJ: Lawrence Erlbaum. Lillo-Martin, D., Bellugi, U., Struxness L., & O’Grady, M. (1985). The acquisition of spatially organized syntax. In E. Clark (Ed.), Proceedings of the Stanford Child Languages Forum 24 (pp. 70–78). Palo Alto, CA: Stanford University Press. Loew, R. (1984). Roles and reference in American Sign Language: A developmental perspective. Unpublished doctoral dissertation, University of Minnesota. Lucas, C., Bayley, R., & Valli, C. [in collaboration with Mary Rose, Alyssa Wulf, Paul Dudis, Laura Sanheim, & Susan Schatz]. (2001). Sociolinguistic variation in ASL ( Sociolinguistics in deaf communities, Vol. 7 ). Washington, DC: Gallaudet University Press. Lucas, C., Bayley, R., & Valli. C. (2003). What’s your sign for PIZZA? An introduction to variation in ASL . Washington, DC: Gallaudet University Press. Lucas, C., Valli, C., & Mulrooney, K. (2005). Linguistics of American Sign Language (4th ed.). Washington, DC: Gallaudet University Press. MacFarlane, J. (1998). From affect to grammar: Ritualization of facial affect in signed languages. Paper presented at the Theoretical Issues in Sign Language Research Conference at Gallaudet University. McIntire, M., & Reilly, J. (1988). Nonmanual behaviors in L1 & L2 learners of American Sign Language. Sign Language Studies, 61, 351–375. McKee, R., & McKee, D. (1992). What’s so hard about learning ASL? Students’ and teachers’ perceptions. Sign Language Studies, 75, 129–158. Mayberry, R. (1995). Mental phonology in language comprehension, or what does that sign mistake mean? In K. Emmorey & J. Reilly (Eds.), Language, gesture, and space (pp. 355–370). Hillsdale, NJ: Lawrence Erlbaum. Mayberry, R., & Eichen, E. (1991). The long-lasting advantage of learning sign language in childhood: Another look at the critical period for language acquisition. Journal of Memory and Language, 30, 486–512. Mayberry, R., & Fischer, S. (1989). Looking through phonological shape to lexical

< previous page

page_238

next page >

< previous page

page_239

next page >

Page 239 meaning: The bottleneck of nonnative sign language processing. Memory and Cognition, 17, 740–754. Meier, R. (2006). The form of early signs: Explaining signing children’s articulatory development. In M. Marschark, B. Schick, & P. Spencer (Eds.), Advances in sign language development by deaf children (pp. 202–230). New York: Oxford University Press. Meier, R., Cheek, A., & Moreland, C. (2002). Iconic versus motoric determinants of the form of children’s early signs. In B. Skarabela, S. Fish, & H.-J. Do (Eds.), BUCLD 26: Proceedings of the 26th Annual Boston University Conference on Lan guage and Development (pp. 393–405). Somerville, MA: Cascadilla Press. Meir, I., & Sandler, W. (2007). A language in space: The story of Israeli Sign Language. New York: Lawrence Erlbaum. Mirus, G., Rathmann, C., & Meier, R. (2001). Proximalization and distalization of sign movement in adult learners. In V. Dively, M. Metzger, S. Taub, & A. M. Baer (Eds.), Signed languages: Discoveries from international research (pp. 103–119). Washington, DC: Gallaudet University Press. Morford, J., & Mayberry, R. (2000). A reexamination of “Early Exposure” and its implications for language acquisition by eye. In C. Chamberlain, J. Morford, & R. Mayberry (Eds.), Language acquisition by eye (pp. 111–128). Mahwah, NJ: Lawrence Erlbaum. Morgan, G. (1999). Event packaging in BSL discourse. In E. Winston (Ed.), Story telling and conversation: Discourse in deaf communities (pp. 27–58). Washington, DC: Gallaudet University Press. Morgan, G., Barrett-Jones, S., & Stoneham, H. (2007). The first signs of language: Phonological development in British Sign Language. Applied Psycholinguistics , 28, 3–22. Morgan, G., Smith, N., Tsimpli, I., & Woll, B. (2002). The effects of modality on BSL development in an exceptional learner. In R. Meier, K. Cormier, & D. Quinto (Eds.), Modality and structure in signed and spoken language (pp. 422– 441). Cambridge, UK: Cambridge University Press. Nelson, C. (1987). The recognition of facial expressions in the first two years of life: Mechanisms of development. Child Development , 58, 889–909. Newport, E. (1990). Maturational constraints on language learning. Cognitive Science , 14, 11–28. Newport, E., & Meier, R. (1985). The acquisition of American Sign Language. In D. Slobin (Ed.), The crosslinguistic study of language acquisition (pp. 881–938). Mahwah, NJ: Lawrence Erlbaum. Oller, K., Wieman, L., Dole, W., & Ross, C. (1976). Infant babbling and speech. Journal of Child Language, 3 , 1–12. Oyama, S. (1976). A sensitive period in the acquisition of a non-native phonological system. Journal of Psycholinguistic Research , 5 , 261–285. Padden, C. (1991). The acquisition of fingerspelling in deaf children. In P. Siple & S. Fischer (Eds.), Theoretical issues in sign language research. Vol. 2, Psychology (pp. 191–210). Chicago: University of Chicago Press. Padden, C. (2006). Learning to fingerspell twice: Young signing children’s acquisition of fingerspelling. In M. Marschark, B. Schick, & P. Spencer (Eds.), Advances in sign language development by deaf children (pp. 189–201). New York: Oxford University Press.

< previous page

page_239

next page >

< previous page

page_240

next page >

Page 240 Padden, C., & Hanson, V. (2000). Search for the missing link: The development of skilled reading in deaf children. In K. Emmorey & H. Lane (Eds.), The signs of language revisited: An anthology in honor of Ursula Bellugi and Edward Klima (pp. 435–448). Mahwah, NJ: Lawrence Erlbaum. Padden, C., & LeMaster, B. (1985). An alphabet on hand: The acquisition of finger-spelling in deaf children. Sign Language Studies, 47, 161–172. Padden, C., & Ramsey, C. (1998). Reading ability in signing deaf children. Topics in Language Disorders, 18, 30–46. Petitto, L., & Holowka, S. (2002). Evaluating attributions of delay and confusion in young bilinguals: Special insights from infants acquiring a signed and spoken language. Sign Language Studies, 3 , 4–33. Petitto, L., & Marentette, P. (1991). Babbling in the manual mode: Evidence from the ontogeny of language. Science , 251 , 1493–1496. Poizner, H., & Kegl, J. (1993). Neural disorders of the linguistic use of space and movement. In P. Tallal, A. Falaburda, R. Llinas, & C. von Euler (Eds.), Annals of the New York Academy of Science, Temporal Information Processing in the Nervous System. Vol. 682 (pp. 192–213). New York: New York Academy of Science Press. Poizner, H., Brentari, D., Tyrone, M., & Kegl, J. (2000). The structure of language as motor behavior: Clues from signers with Parkinson’s disease. In K. Emmorey & H. Lane (Eds.), The signs of language revisited: An anthology in honor of Ursula Bellugi and Edward Klima (pp. 509–532). Mahwah, NJ: Lawrence Erlbaum. Poizner, H., Klima, E., & Bellugi, U. (1987). What the hands reveal about the brain . Cambridge, MA: MIT Press/Bradford Books. Reilly, J. (2000). Bringing affective expression into the service of language: Acquiring perspective marking in narratives. In K. Emmorey & H. Lane (Eds.), The signs of language revisited: An anthology in honor of Ursula Bellugi and Edward Klima (pp. 401–433). Mahwah, NJ: Lawrence Erlbaum. Reilly, J. (2006). Development of nonmanual morphology. In M. Marschark, B. Schick, & P. Spencer (Eds.), Advances in sign language development by deaf children (pp. 262–290). New York: Oxford University Press. Reilly, J., & Bellugi, U. (1996). Competition on the face: Affect and language in ASL motherese. Journal of Child Language, 23, 219–239. Reilly, J., & McIntire, M. (1991). WHERE SHOE: The acquisition of wh-questions in ASL. Papers and Reports in Child Language Development, 30, 104–111. Reilly, J., McIntire, M., & Bellugi, U. (1990a). FACES: The relationship between language and affect. In V. Volterra & C. Erting (Eds.), From gesture to language in deaf and hearing children (pp. 129–141). New York: Springer-Verlag. Reilly, J., McIntire, M., & Bellugi, U. (1990b). Conditionals in American Sign Language: Grammaticized facial expressions. Applied Psycholinguistics , 11(4), 369–392. Reilly, J., McIntire, M., & Bellugi, U. (1991). BABYFACE: A new perspective on universals of language acquisition. In P. Siple (Ed.), Theoretical issues in sign language research: Psycholinguistics (pp. 9–23). Chicago: University of Chicago Press. Rosen, R. (2004). Beginning L2 production errors in ASL lexical phonology. Sign Language Studies, 7 , 31–61. Rossini, P., Reilly, J., Fabretti, D., & Volterra, V. (1998). Non-manual behaviors in Italian Sign Language. Paper presented at the Italian Sign Language Conference, Genoa, Italy.

< previous page

page_240

next page >

< previous page

page_241

next page >

Page 241 Sandler, W., & Lillo-Martin, D. (2006). Sign language and linguistic universals . Cambridge, UK: Cambridge University Press. Schembri, A., Johnston, T., & Goswell, D. (2006). NAME dropping: Location variation in Australian Sign Language. In C. Lucas (Ed.), Multilingualism and sign languages: From the Great Plains to Australia (pp. 121–156). Washington, DC: Gallaudet University Press. Scovel, T. (1969). Foreign accents, language acquisition and cerebral dominance. Language Learning , 19, 245–254. Scovel, T. (1981). The recognition of foreign accents in English and its implications for psycholinguistic theories of language acquisition. In J.-G. Savard & L. Laforge (Eds.), Proceedings of the 5th Congress of AILA (pp. 389–401). Laval, Quebec: University of Laval Press. Slobin, D. (1973). Cognitive prerequisites for the development of grammar. In C. Ferguson & D. Slobin (Eds.), Studies of child language development . New York: Holt, Rinehart, and Winston. Stokoe, W., Casterline, D., & Croneberg, C. (1965). A dictionary of American Sign Language on linguistic principles. Silver Spring, MD: Linstok Press. Sutton-Spence, R., & Woll, B. (1999). The linguistics of British Sign Language: An introduction. Cambridge, UK: Cambridge University Press. Takkinen, R. (2003). Variations of handshape features in the acquisition process. In A. Baker, B. van den Bogaerde & O. Crassborn (Eds.), Cross-linguistic perspectives in sign language research: Selected papers from TISLR 2000 (pp. 81–91). Hamburg: Signum. Tyrone, M., Kegl, J., & Poizner, H. (1999). Interarticulator co-ordination in deaf signers with Parkinson’s disease. Neuropsychologia , 37, 1271–1283. van den Bogaerde, B. (2000). Input and interaction in deaf families. Published doctoral dissertation, University of Amsterdam. Utrecht: LOT. von Tetzchner, S. (1994). First signs acquired by a Norwegian deaf child with hearing parents. Sign Language Studies, 44, 225–257.

< previous page

page_241

next page >

< previous page

page_242

next page >

page_242

next page >

Page 242 This page intentionally left blank.

< previous page

< previous page

page_243

next page >

Page 243 Part IV METHODOLOGICAL AND NEUROLINGUISTIC ASPECTS OF DEVELOPMENT

< previous page

page_243

next page >

< previous page

page_244

next page >

page_244

next page >

Page 244 This page intentionally left blank.

< previous page

< previous page

page_245

next page >

Page 245 10 LONGITUDINAL DESIGNS IN STUDIES OF MULTILINGUALISM Robert W. Schrauf Cognitive aging or cognitive gerontology is the discipline that addresses the growth and decline of mental processing in the latter two-thirds of the human lifespan. Language production and comprehension, especially as addressed by psycholinguistics, has been a consistent interest in the field, as a brief review of recent handbooks reveals (e.g. Bialystock & Craik, 2006; Craik & Salthouse, 2000; Park & Schwarz, 2000). Nevertheless, while considerable research exists on language and lifespan development, the vast majority of this work is focused exclusively on monolinguals, with a few notable exceptions (for review, see Schrauf, 2008). On the other hand, studies of bilingualism/multilingualism are more often cross-sectional and rarely longitudinal. To be sure, there are longitudinal studies of second language acquisition, focused almost exclusively on children and young adults, in both educational (for a review, see Ortega & Iberri-Shea, 2005) and informal environments (see especially the European Science Foundation project, e.g. Dietrich, Klein, & Noyau, 1995; Klein & Perdue, 1992). However, the duration of such studies is relatively short, lasting from a few months to up to 2 years, largely because the focus is on the short-term acquisition of particular language forms. Further, in studies of language attrition, at the other end of the developmental continuum, there is one truly longitudinal study of language attrition. De Bot and Clyne (1994) studied first language attrition among Dutch and German immigrants to Australia over a 16-year period across three data periods of observation (1971–1972, 1987, 1991–1992). Clearly there is a great deal of work to be done on the specifically longitudinal, lifespan study of multilingualism. Fortunately, there are numerous lessons to be learned from lifespan psychology and cognitive gerontology to guide such research. The available longitudinal studies of monolingualism are eminently informative and lay a solid foundation for the study of multilingualism. In addition, the field of cognitive aging offers considerable wisdom concerning the design of longitudinal research and has witnessed exponential growth in the development of

< previous page

page_245

next page >

< previous page

page_246

next page >

Page 246 statistical methods for studying change over time. This chapter presents an overview of some of these resources. The chapter is organized as follows. First, a summary is provided of the common longitudinal designs in lifespan psychology and cognitive aging. This involves a careful comparison of cross-sectional vs. longitudinal designs and an evaluation of the relative contributions of both designs. Second, the literature on threats to the validity of longitudinal studies is reviewed. Of particular interest here are problems posed by cohort effects, test–retest or practice effects, drop-out or participant mortality, and selection effects. Of course, the study of multilingualism poses its own threats to validity—notably, the dynamic nature of language proficiency and shifting language environments. Third, summaries are provided of the currently available monolingual literature on longitudinal studies of language and aging. These are concerned with vocabulary recognition and recall, picture naming, and grammatical complexity/propositional content. All are restricted to English. Notions of Change and Research Designs Conceptual notions of change and measurement sensitivity to change are interdependent in lifespan psychology. Empirical research of change-overtime requires the conceptual and methodological disentangling of the intertwined effects of age (years since birth), time-of-measurement (date of testing or observation), and cohort (historical position due to year of birth). These are in fact necessarily confounded since knowing two automatically gives knowledge of the third. Two of the most influential scientists responsible for addressing these confounds early on in lifespan studies are K. Warner Schaie and Paul Baltes. In a 1965 article in the Psychological Bulletin, Schaie (1965) laid out a “general developmental model” for separating out these effects, and Baltes introduced the terms crosssectional and longitudinal sequences in studies of change over time (1968; Baltes, Reese, & Nesselroade, 1977). In the following paragraphs a brief summary of this work is offered, with notes concerning how the distinctions and designs might be relevant for studies of multilingualism over the lifespan. Other recent reviews of concepts and longitudinal methods in lifespan psychology and cognitive gerontology include Hofer and Sliwinski (2006) and Schaie and Caskie (2005). Cross-sectional Designs The basic cross-sectional design takes samples of individuals at different ages (multiple cohorts) and compares them on some variable of interest at one point in time (time-of-measurement). Figure 10.1 shows a cross-sectional study in which data are collected once in 2005 from individuals ranging in age from 25 through 65.

< previous page

page_246

next page >

< previous page

page_247

next page >

Page 247

Figure 10.1 Cross-sectional vs. longitudinal designs (source: adapted from Schaie, 1965, 2005). Note Rows represent the same participants tested at successively later ages. Columns represent different participants at successively later ages tested at the same time. Cross-sectional designs are maximally informative about age differences in cognitive or psychological functions. That is, such studies address interindividual differences as a between persons effect. Over 10 years ago, Hertzog (1996) noted that the modal design in cognitive aging studies was the extreme-age groups design in which a group of young adults (usually a university sample) is compared with a group of older adults (usually a community sample). A cursory look at the literature suggests that this is still the case, though it has become common to recruit multiple cross-sectional groups across the lifespan as well (30-, 40-, 50-, 60-year-olds, etc.). Because observations are made at only one time point, there are no effects of time-of-measurement in this design. Further, age differences are confounded with cohort differences, and hence inferences concerning within-person change over time are unwarranted. For instance, as we will review below in a study of monolinguals, differences in the educational experiences of different cohorts affect performance on cross-sectional tests of verbal ability, such that later cohorts who enjoyed more intensive educational exposure may outperform earlier cohorts. Obviously, from a study conducted in this way, it would be erroneous to draw the conclusion that verbal ability decreases with age. Cross-sectional studies of bilingualism and aging are reviewed in de Bot and Makoni (2005) and Schrauf (2008). Schrauf (2008) summarizes the available literature as follows. First, age-differences in decline and preservation of the two languages of bilinguals are similar to patterns and declines seen for the one language of monolinguals (Juncos-Rabadan, 1994; Obler,

< previous page

page_247

next page >

< previous page

page_248

next page >

Page 248 Albert, & Lozowick, 1986; Rosselli et al., 2000). Second, after controlling for equivalent proficiency in both languages, patterns of decline and preservation are similar in both first and second languages of bilinguals (de Bot & Lintsen, 1986; Juncos-Rabadan, 1994; Picciotto & Friedland, 2001; Xue, Hagstrom, & Hao, 2002). Third, the unique language-switching ability of bilinguals may play a special role in the preservation of cognitive abilities in aging (Bialystok, Craik, Klein, & Viswanathan, 2004; Bialystok, Craik, & Ryan, 2006; Hernandez & Kohnert, 1999). Given the critical difference between cross-sectional and longitudinal studies, it is important to note that these studies can speak only to age-differences between young and old on language performance and not to intraindividual trajectories or within-person changes in language performance. Hence, conclusions about patterns of acquisitions, retention, and attrition of the multilingual’s languages depend on more sensitive longitudinal designs. Longitudinal Designs In the basic longitudinal design, one sample of individuals (one cohort) is tested two or more times (at successively older ages) on the dependent variable of interest. Figure 10.1 shows a possible design for a longitudinal study in which data are collected on individuals aged 60 in the year 2005 (born in 1945) and then three more times in the years 2010, 2015, and 2020. Longitudinal designs have become the desideratum in cognitive aging because they make possible the investigation of age-related changes in function over time within persons. That is, such studies address intraindividual change as a within-persons effect. As Schaie (2000) points out, the shift from crosssectional studies to longitudinal studies made it possible in gerontology to address directly issues of antecedent and consequent conditions and events. As will be discussed below, many age-differences found in cross-sectional studies have not been replicated as age-related effects in longitudinal studies, and this has made the latter method crucial for testing theories about aging. In the simple case in which a single cohort is tracked over time, no information about cohort effects is available since only one cohort is studied, and age and time-of-measurement are confounded. In addition, there are other threats to the internal validity of longitudinal designs, particularly in the form of historical events that might occur between observations. For instance, the researcher who is interested in the effects of declining processing speed on production of the L2 must recruit a sample of individuals whose language environment undergoes no change for the duration of the study, since changes in social support for L2 would be confounded with changes in processing speed. Longitudinal studies are also affected by recruitment issues, practice effects, and drop-out, which will be examined below.

< previous page

page_248

next page >

< previous page

page_249

next page >

Page 249 Cross-sectional and Longitudinal Sequential Designs and Analytic Strategies Extensions of basic cross-sectional and longitudinal designs are known as sequential designs (Baltes, 1968; Schaie & Baltes, 1975). A design is sequential when data are collected over time, though not necessarily on the same participants. A cross-sectional sequential design involves making observations of a range of cohorts at one point in time and then at a later time point recruiting new samples from the same range of cohorts, and then comparing them to the original samples. A longitudinal sequential design involves making observations on a range of cohorts at one point in time and then making repeated observations on these same samples at later time points. This latter design makes it possible to assess intra individual change over time and to compare inter individual rates of change or trajectories. It is not hard to imagine, for instance, that some individuals experience more rapid declines than others, and a longitudinal sequential design opens a window onto this phenomenon. Designs are to be distinguished from analytic strategies. Thus, Schaie and colleagues have developed a taxonomy of analytic strategies for the data collected using sequential designs (Schaie, 2005; Schaie & Caskie, 2005). In particular, they distinguish three analytic strategies to interpret the data from sequential designs. In the cohortsequential strategy, observations are made on all cohorts at all ages. In Figure 10.2, two cohorts are examined: individuals born in 1945 and those born in 1950. The cohort born in 1945 is assessed twice: at age 60 in 2005 and again at age 65 in 2010. The cohort born in 1950 is also assessed twice: at age 60 in 2010 and again at age 65 in 2015. Thus, within-person change is tracked within cohorts (longitudinal assessment at ages 60 and 65), and between-person differences in change are assessed across cohorts (differences in the age 60–65 trajectories for the 1945 vs. the 1950 cohorts). Note that observations are made at three times (2005, 2010, and 2015). Assuming that time-of-measurement effects are minimal, this makes it possible to separate out cohort effects from age effects. Schaie and Caskie (2005) point out that this model is particularly appropriate for examining variables that are known to decline with age, such as loss of efficiency in peripheral sensory function (Baltes & Lindenberger, 1997) or processing speed (Salthouse, 1996). The cross-sequential strategy is common in studies of monolingualism and aging. Thus, in many studies, participants are recruited across a range of ages and then subjected to repeated testing over a specified time interval (e.g. three times over a 7-year period). The analytic strategy involves grouping them into cohort groups (usually by decade: 30s, 40s, 50s, etc.) and then analyzing within-cohort change over time (intraindividual trajectories) while comparing rates of change across groups (interindividual differences in

< previous page

page_249

next page >

< previous page

page_250

next page >

Page 250

Figure 10.2 Cohort-sequential and cross-sequential strategies (source: adapted from Schaie, 1965, 2005. Note Rows represent the same participants tested at successively later ages. Columns represent different participants at successively later ages tested at the same time. intraindividual change). Examples of this design in the third portion of this chapter include studies by Sliwinski and Buschke on vocabulary (Sliwinski & Buschke, 1999); Connor, Spiro, Obler, and Albert on naming (Connor et al., 2004), and Kemper, Greiner, Marquis, Prenovost, and Mitzner on grammatical complexity and propositional density (Kemper et al., 2001). In the cross-sequential analysis strategy, data are analyzed from at least two cohorts at two times of measurement (not three, as with the cohort-sequential strategy). Figure 10.2 shows a cross-sequential design in which data are analyzed from two cohorts: individuals born in 1955 and individuals born in 1980. Both cohorts are examined longitudinally: the 1955 cohort at ages 50 and 60, and the 1980 cohort at ages 25 and 35. This analysis focuses on comparison of cohorts and assumes that differences within-cohorts across time will not be great (i.e. age-effects will be trivial). Thus, for instance, a study of the use of the perfect conditional by Mexican American Spanish speakers living in a Los Angeles, California suburb might find 25-year-olds making little use of the form compared to 50-year-olds in 2005, with much the same difference 10 years later in 2015. (However, an interaction showing a more precipitous decline in usage for the younger vs. older cohort might lead to further questions about the effects of language environment.) In the time-sequential design data are collected on all ages at all times of measurement, but no repeated measures are taken. Table 10.1 shows a study that includes two data collection times separated by 20 years (2005 and 2025). In the year 2005, 20- and 70-year-olds (born in 1985 and 1935 respectively) are tested. Then in the year 2025, new samples of 20- and 70-year-olds (born in 2005 and 1955 respectively) are tested.

< previous page

page_250

next page >

< previous page

page_251

next page >

Page 251

Figure 10.3 Time-sequential analytic strategy (adapted from Schaie & Caskie, 2005). Note Cells containing years reflect different birth cohorts. Assuming that cohort effects are minimal, this approach makes it possible to separate out period effects from age effects. Such a design would probably be more useful in addressing questions about language attitudes and language ideology rather than psycholinguistic questions because these latter are arguably more period-dependent than are cognitive mechanisms. In sum, the different kinds and levels of research design in lifespan developmental psychology involve looking at (1) age differences (interindividual variability) between persons, (2) age-related changes (intraindividual variability) within persons, and (3) differential change over time (interindividual differences in intraindividual variability). In statistical terms, intraindividual patterns of change (slopes) are analyzed at the interindividual level (Hofer & Sliwinski, 2006). Many of the statistical models for analyzing longitudinal data fall under the general rubric of the general linear mixed model (Laird & Ware, 1982), of which multilevel modeling (Goldstein, 1995; Kreft & de Leeuw, 1998), hierarchical linear modeling (Bryk & Raudenbush, 1987, 1992), and random coefficient modeling (Laird & Ware, 1982) are examples. Such techniques model change for each individual in terms of fixed effects and random effects. Fixed effects include the coefficient for the intercept (initial level of performance) and a coefficient for the slope (rate of change in performance), plus possibly higher-order terms (e.g. non-linear functions). Random effects allow for individual variation among intercepts and slopes. Covariates may then be added to the models to assess the effects of additional independent variables on between-subject differences in intercepts and slopes. These models are particularly effective in dealing with problems of autocorrelation between successive observations, missing data, and differing intervals in times-of-measurement between persons. (For review of these and additional techniques for modeling change, see Hertzog & Nesselroade, 2003; Hofer & Sliwinski, 2006). Threats to the Validity of Longitudinal Studies There are several threats to validity posed by the nature of multilingualism itself. These include the dynamic nature of language proficiency, shifting language environments, frequency-of-use, and first and second language

< previous page

page_251

next page >

< previous page

page_252

next page >

Page 252 \ attrition (Schrauf, 2008). In addition, there are a number of threats posed by the nature of longitudinal design. These include cohort or historical effects, practice effects, drop-out or experimental mortality, and selection or sample bias (Campbell & Stanley, 1963; Schaie & Hofer, 2001). The following paragraphs describe these eight threats, beginning with the four posed by studying multilingualism itself. Threats to Validity in Studies of Multilingualism Changes in Language Proficiency In a cross-sectional study, individual differences in language proficiency are particularly problematic. For instance, comparing first language performance of young and old bilinguals requires knowing that the older subjects are (or were) possessed of the same original knowledge as the younger participants, or else differences in performance between young and old will be seen as evidence of change over time when in reality these differences existed at baseline. This is less problematic in a longitudinal study since measures are taken both at some baseline and then at subsequent ages. However, longitudinal studies can be problematic in their own right. Language proficiency in both first and second languages is a moving target. Initially, this is so because acquisition involves moving from a state of almost total ignorance to some level of ultimate attainment. Proficiency continues to develop or deteriorate, however, depending on circumstances and use (Lowie et al., this volume). Further, for even the most fluent bilingual, there are subtle differences in language proficiency in L1 vs. L2. Thus, depending on the dependent variable of interest, proficiency must be measured in both languages. For instance, if a potential research question concerns the effects of age-related, diminishing working memory on language performance in L1 vs. L2, then fairly precise measures of proficiency in L1 and L2, as well as working memory capacity, at earlier ages are required to assess a pattern of change. Language Environment Exposure to the L1 and L2 language environments provides both opportunities and motivation for use of the L1 and L2, but these environments are themselves subject to change. Although emigration or changes in language policy are obvious examples of extreme changes in language environment (Schmid, this volume), less dramatic occurrences can play a decisive role as well. Imagine the fluent bilingual who speaks her second language primarily in occupation-related activities and who retires to a home situation where the first language is primarily spoken. This subtle shift in environment portends a possible shift in language dominance as well, primarily because fre-

< previous page

page_252

next page >

< previous page

page_253

next page >

Page 253 quency of use of the L2 will probably decrease. If a longitudinal study addresses the influence of some cognitive variable (e.g. working memory) on language production or comprehension, such changes in language environment will be particularly unwelcomed since they serve to confound changes due to diminishing working memory with changes due to shifts in opportunity to use the language. Frequency-of-use Frequency-of-use is the mechanism through which language environment has its effects. Thus, it is related to proficiency and environment, though distinct from them. That is, on the one hand, the more one speaks a language, the more likely it is that his or her proficiency will be either maintained or increased or held constant. On the other hand, perhaps the best predictor for frequency-of-use is a social environment in which the specific languages are obligatorily spoken. This latter point is confirmed in the breach by immigrants who maintain residence and work in language-isolated neighborhoods and who acquire only low levels of proficiency in their L2 (Schrauf, under review). Thus, although such individuals may live in a largely monolingual society (e.g. the United States), they may live in a neighborhood that largely protects them from exposure to English. Language Attrition Finally, language attrition is a dynamic process of loss of proficiency (Goral, 2004; Schmid, this volume), due in part to changes in language environment and to well-established declines in cognitive mechanisms: slower processing speed, diminishing working memory, inhibitory deficits, and neural-sensory deficits (Schrauf, 2008). Again, however, the majority of studies of language attrition among bilinguals do not track intraindividual decline over time in one or the other language, but rather departures from “native norms” (Goral, 2004; Kopke, 2004). As noted in the introduction, the unique exception here is the longitudinal study conducted by de Bot and Clyne (1994) on German– English and Dutch–English immigrants to Australia. Interestingly, this study found little evidence of decline across three data collection points over a period of 16 years. Standard Threats to the Validity of Longitudinal Studies Cohort Effects Cohort effects reflect the influence of culture and history on participants. As both history and culture are dynamic in nature, participants from different birth cohorts are potentially subject to different social environments. Indeed,

< previous page

page_253

next page >

< previous page

page_254

next page >

Page 254 as pointed out above, studies of multilingualism are prey to changes in language environment, reflected for example in waves of immigration or changes in language policy at the level of society. In the literature on cognitive aging, changes in educational practices, opportunities, and expectations provide an excellent example of cohort effects. That is, more recent (younger) cohorts may obtain higher scores than more remote (older) cohorts as a result of higher levels of education and exposure to more affirmative environmental stimuli. In cross-sectional testing, such differences in education lead to more pronounced age-differences, favoring younger participants. Nevertheless, empirical work shows that cohort differences in cognitive abilities are not uniform across types of abilities. In the Seattle Longitudinal Study, Schaie analyzed birth cohorts from 1889 to 1966 and found linear positive cohort shifts over time for verbal meaning and inductive reasoning but negative cohort shifts for number skill and word fluency (Schaie, 1994, 2005). That is, more recent cohorts (younger individuals) scored higher than more remote cohorts (older individuals) on verbal meaning and inductive reasoning, whereas remote cohorts scored higher on number skill and verbal fluency than more recent cohorts. Schaie suggests that the former may be driven by increases in formal education for successively later cohorts and that the latter may be the result of shifts in educational strategies and exposure to relevant stimuli. In any case, these patterns point out the problems with reliance on cross-sectional studies alone to assess the effects of aging on cognitive abilities. As Schaie notes, crosssectional studies will therefore underestimate declines prior to age 60 in the case of abilities that show a positive gradient across cohorts and overestimate them for abilities that show negative gradients across cohorts (Schaie, 1994, pp. 308–309). Test–Retest Effects or Practice Effects Test–retest effects or practice effects describe increases in performance across time as a result of prior exposure to testing materials and/or the testing situation. These gains may occur for a number of reasons, but chief among them are memory for test-items resulting from prior exposure to them and participants’ growth in familiarity with the testing environment and task characteristics. The latter can occur because the kinds of tasks involved in cognitive testing are highly decontextualized, low in ecological validity, and associated with formal, academic learning. Individuals with low levels of formal education or individuals who have been away from classroom environments for many years (as is the case with most older adults) are disadvantaged at initial testing because they are unfamiliar with the tasks. Then the familiarity gained from the initial exposure to the tasks helps increase performance on the second and subsequent occasions. Attempts to reduce such effects include the construction and administration of alternate test forms for use at later time points.

< previous page

page_254

next page >

< previous page

page_255

next page >

Page 255 Both design and analytic strategies have been suggested as means of dealing with retest effects. At the design level, researchers have lengthened the intervals between testing occasions, presumably to induce forgetting during the intervening period. This raises the practical question of how long intervals should be. For instance, Rabbitt and colleagues found retest effects at intervals of 2 to 3 years on a measure of fluid intelligence (Rabbitt, Diggle, Smith, Holland, & McInnes, 2001). In a review of three longitudinal studies, including the Baltimore Longitudinal Study of Aging, the Iowa 65+ Rural Health Study, and the Victoria Longitudinal Study, Zelinski and colleagues found that 6 years was the “minimum retest interval to assure that longitudinal declines in memory are statistically reliable” (Zelinski & Burnight, 1997, p. 503). Analytic strategies for dealing with retest effects involve developing models that parameterize age and occasions of measurement separately in order to separate out the effects. Such modeling has implications for design as well, and therefore the issues are intimately connected. For instance, the fact that chronological age and periodic retesting advance in tandem produces a confound at the level of analysis. Ferrer and colleagues provide a succinct description of this problem. For example, imagine a study in which individuals of the same age are measured at intervals of exactly 1 year. The age-related changes in a particular variable will then be confounded with possible changes in performance that are due to retest. That is, increments in retest will correlate perfectly with increments in age. (Ferrer, Salthouse, Stewart, & Schwartz, 2004, p. 244) As Ferrer et al. note, one way of remedying this situation is varying the retest interval. This time-lag method is exemplified in McArdle and Woodcock (1997) in which a number of groups or individuals are tested at some initial time point, and then retested at different intervals for the second testing. McArdle and Woodcock used data from the norming sample of the Woodcock–Johnson Psychoeducational Battery, in which participants were retested once on particular subtests, to develop models that successfully separated out the effect of retest from either decline or stability over time, depending on the particular subtest and length of interval (McArdle & Woodcock, 1997; see McArdle, Ferrer-Caja, Fumiaki, & Woodcock, 2002, for a similar implementation). Retesting at more than one time point also helps in separating out practice effects because multiple retests provide more reliable estimates of change. In two studies, Rabbitt and colleagues (Rabbitt, Diggle, Holland, & McInnes, 2004; Rabbitt et al., 2001) analyzed longitudinal data over a 17-year period, from 5,895 participants, aged 49 to 93 at initial testing, and tested up to four times at two- to three-year intervals. By including separate

< previous page

page_255

next page >

< previous page

page_256

next page >

Page 256 parameters for chronological age and measurement occasion, they showed that, while scores generally declined with increasing age, performance increased over testing occasions. In fact, increases in scores due to testing were of the same order of magnitude as the declines due to age. Differences were also observed on type of task, with older participants showing more marked gains after initial testing on measures of fluid ability than did younger participants. Interestingly, the majority of this improvement took place from t1 to t2 but rates of improvement decreased after t2 . The authors suggest that older individuals may show greatest improvement early on as a result of overcoming their unfamiliarity with cognitive tasks from t1 to t2. If this is indeed the case, the authors point out that alternate or parallel forms are not likely to reduce such retest effects. Retest effects are also dependent on task type, as some of the studies already mentioned have noted. Ferrer et al. (2004) tested individuals ranging in age from 40 to 70 years four times over a 3-year period on three cognitive categories: verbal learning, spatial ability, and processing speed. Retest effects were modeled separately from age, and these effects differed depending on the variable. To wit, when retest effects were not included in the model, age had little effect on verbal learning, but when retest effects were included, the expected decline with age was evident, suggesting that retest effects masked the decline in the first model. However, on spatial ability and processing speed, there were no appreciable differences in models including retest effects vs. those that did not. As indicated above, these results suggest that retest effects can be task specific, and study designs would have to be planned accordingly. Of interest to psycholinguists is the fact that verbal learning in particular was subject to retest effects. The strategies investigated so far have involved individuals who varied in age at initial testing, and in part the models exploit this between-person difference in age to separate out age and retest effects. Thorvaldsson, Hofer, Berg, and Johansson (2006) have recently argued that this is problematic “because age-heterogenous differences do not converge to within-person changes because of other notable factors other than retest gains” (MacDonald, Hultsch, Strauss, & Dixon, 2003; Sliwinski & Buschke, 1999, p. 348). In addition, defining retest effects as the number of times an individual has been previously tested creates a confound with population mortality effects. Again, survivors in a longitudinal study comprise a progressively healthier subset of the original sample. One way of controlling for this is to recruit a sample of individuals at the same age from an original population for multiple retesting and then return to that same population at a later date to recruit a new sample matched with the prior sample. Thus, Thorvaldsson et al. (2006) used the H70 Study in which participants were recruited from the birth cohort 1901–1902 in Goteburg, Sweden, starting in 1971 and retested 10 additional times at ages 75, 79, 81, 85, 88, 90, 92, 95, 97, and 99. By sampling the same birth cohort

< previous page

page_256

next page >

< previous page

page_257

next page >

Page 257 again in 1986 and testing these new participants five times (at ages 85, 88, 90, 95, 99), investigators were able to compare a tested group to an untested group at age 85 and then track changes in both groups through repeated testing at later ages. Multiple slope linear mixed models were fit for five cognitive variables (synonyms, block design, figure identification, digits forward, and digits backward). Interestingly, the authors found small test–retest effects for the synonyms and the block design tests and no significant effects for the other three measures. They suggest that retest effects for synonyms and block designs may be due to participants’ remembering test material over subsequent occasions. This has interesting implications for language tests, which, like the synonyms task, may involve repeated items over testing occasions. Recruitment and Drop-out Recruitment and drop-out require particular attention in longitudinal studies of cognitive aging. In general, the older individuals who volunteer for such studies probably represent a healthier, more motivated, and more intelligent slice of the elderly population (Baltes, Schaie, & Nardi, 1971) with higher socioeconomic status (Powers & Bultena, 1972) than their age-mates who do not volunteer. In longitudinal studies, sample selectivity operates on the available survivors at later time-points via drop-out and population mortality. Drop-out is selective in that the oldest, frailest, and least able are more likely to leave studies (Lachman, Lachman, & Taylor, 1982), and population mortality functions to eliminate less healthy individuals (McArdle, Hamagami, Elias, & Robbins, 1991). Lindenberger, Singer, and Baltes (2002) refer to the former as “experimental selectivity” and to the latter as “mortality-associated selectivity,” and they propose different formulae for decomposing these effects. There are gender differences as well: men tend to drop out more often than women (Rabbitt et al., 1994). In sum, survivors who continue their participation in a study represent an increasingly gender-specific and elite subsection of the original sample, and this of course results in the underestimation of age-related change. Both design and statistical steps can be taken to ameliorate the effect of attrition in longitudinal studies. At the level of design, data collection at subsequent waves can involve the recruitment of new participants, matched on characteristics of the original sample, to substitute for drop-outs. However, insofar as these new participants differ in important (and unforeseen) ways from the original sample, this can create a “drop-in” effect that may affect the estimation of age-related change as well (Rabbitt, Diggle, Holland, & McInnes, 2004). At the level of analysis, it is common practice to assess the differences between responders and non-responders at later stages of data collection by comparing their data at the time of first assessment to see if significant differences are found there as well. Further, this

< previous page

page_257

next page >

< previous page

page_258

next page >

Page 258 missing data problem can be dealt with via maximum likelihood estimation methods or multiple imputation methods (Schafer, 1997). Multilevel or random effects models also allow for missing data as well as differing intervals of data collection (Bryk & Raudenbush, 1987; Goldstein, 1995; Hertzog & Nesselroade, 2003). Longitudinal Studies of Monolingualism As noted above, there is little in the literature concerning the longitudinal study of multilingualism in adulthood and old age, though there are a number of studies employing longitudinal methods among monolinguals. These latter studies fall into two categories. First, there are many studies on intellectual development and decline in midlife and old age that include language tasks as one among many different cognitive tasks. Research in this tradition addresses language, albeit as an indicator of something else: that is, accumulated knowledge. Usually, for instance, measurement of vocabulary knowledge (the lexicon) serves as an indicator of world-knowledge or knowledge accumulated from experience. Second, there are studies focused explicitly on the development and decline of language itself. There are far fewer studies of these latter than the former. In the correlational–psychometric tradition in cognitive aging, multiple measures are employed in longitudinal studies to assess gains and losses in cognitive abilities across the adult lifespan. A standard finding is that measures of online processing abilities show decreases across the lifespan while measures of accumulated knowledge typically show either preservation or gains into middle old age (e.g. Park et al., 1996). These two cognitive components are referred to as fluid vs. crystallized intelligence (Horn & Cattell, 1967) or the mechanics vs. pragmatics of intelligence (Baltes, 1987). Not surprisingly language is used extensively as both the measure and the object of measurement in assessments of accumulated knowledge of the world (crystallized intelligence, pragmatics of intelligence). In this sense, language figures prominently in lifespan developmental psychology. However, these measurements of “verbal ability” typically involve very low levels of language comprehension and production, such as vocabulary measures and picture naming. In the experimental tradition in cognitive aging, specific experimental interventions are designed and administered repeatedly over time. The work of Kemper and colleagues (Kemper, this volume) is representative of this strain of work. The following paragraphs provide summaries of representative studies of vocabulary, naming abilities, and grammatical complexity/propositional content and aging as these are found in the literature on aging and monolingualism (usually English).

< previous page

page_258

next page >

< previous page

page_259

next page >

Page 259 Vocabulary One of the most frequently cited cognitive findings in lifespan research on cognition is the distinct pattern of lifespan curves for crystallized intelligence (accumulated knowledge) vs. fluid intelligence (efficiency in cognitive processing). Fluid abilities are typically measured by tasks involving perceptual speed and mental manipulation, such as the Digit Symbol Substitution subtest from the WAIS (Wechsler, 1955, 1981) or the Visual Matching and Crossouts Subtests from the Woodcock–Johnson Psychoeducational Battery (Woodcock & Munoz-Sandoval, 1996). Crystallized abilities are typically measured by vocabulary tests, such as the Vocabulary subtest from the WAIS, the Shipley Institute of Living Scale (Shipley, 1946), the Mill Hill Vocabulary Scale (Raven, 1982), or one of the vocabulary tests in the ETS Kit of Factor Referenced Tests (Ekstrom, French, Marman, & Derman, 1976). The WAIS requires that subjects provide definitions of words and is a production test. The rest of the tests cited provide multiple-choice or other recognition formats and are comprehension tests. In short, the finding from both cross-sectional samples and longitudinal samples is that fluid abilities decline steadily across adulthood while crystallized abilities do not show evidence of decline until the seventh or eighth decade (Park, 2000; Schaie, 1996). By implication, then, lifespan vocabulary has been studied quite extensively in cognitive aging, though primarily via the very restricted tasks of word definition or recognition of word definitions. Nevertheless, there are some curious differences in the results from cross-sectional vs. longitudinal research. First, at a purely cross-sectional level, Verhaeghen (2003) conducted a meta-analysis of studies measuring vocabulary that were published in Psychology and Aging from 1986 to 2001. These compared young adults (between 18 and 30 years old) and old adults (60 years or older). Interestingly, the correlation between young and old scores on the WAIS-R Vocabulary subtest (a production measure) was .30, while the correlation on the Shipley Institute of Living scale (a comprehension measure) was .78. Vocabulary scores increase with age, but obviously the amount of increase is somewhat task-dependent. In general, Verhaeghen found that older adults scored .80 SD higher than young adults on vocabulary tests. This runs counter to the findings of longitudinal research, which shows stability and then late-life decline. Verhaeghen suggests several reasons for this: (a) older adults in this study were generally at or below 75 years of age, (b) cohort effects favor the earlier born (e.g. the Shipley scale was constructed in the 1940s), and (c) older adults in these studies often had more years of education than younger adults. Other researchers using cross-sectional designs have argued that the age-cohort confound masks an actual decline in vocabulary scores across the lifespan (Alwin & McCammon, 1999, 2001; Glenn, 1999). Thus, Alwin and

< previous page

page_259

next page >

< previous page

page_260

next page >

Page 260 McCammon (2001) used data from the General Social Survey (GSS; NORC, 1999) that included 14 repeated crosssectional surveys over 24 years of participants who were 24 years old or older. For the GSS vocabulary measure, participants saw 10 words, and for each word five other words, and were required to select the word closest in meaning to the target word. Analyses showed a growth in vocabulary over the lifespan through age 50 and then a gradual decline after age 60. However, after controlling for cohort effects in number of years of school completed, the effect was greatly reduced. That is, the authors found that later cohorts (notably after World War II in the US) benefited from more years of schooling than earlier cohorts, and this may explain both the increase in scores over time and effectively mask any actual declines. Finally, the very structure of the vocabulary scores has been called into question. Again using a cross-sectional analysis, Bowles, Grimm, and McArdle (2005) factor analyzed data from the GSS and found a two-factor solution that separated more difficult items (Advanced Vocabulary) from easier items (Basic Vocabulary). Interestingly, Basic Vocabulary scores seem to peak at around 35 years of age and show an age-related decline in later adulthood, while Advanced Vocabulary peaks around age 45 and remains constant in later ages. The authors suggest that vocabulary is at least a bidimensional and not a unitary construct, and that patterns of decline over the lifespan differ for either dimension. Longitudinal studies including measures of vocabulary and conducted on older adults as a group generally find declines in vocabulary over time. Hultsch and colleagues (Hultsch, Hertzog, Small, McDonald-Miszczak, & Dixon, 1992) investigated older adults aged 55–86 years old over a 3-year period and used three multiple-choice vocabulary subtests from the ETS Kit of Factor Referenced Cognitive Tests (Ekstrom et al., 1976). Results showed that scores declined over time. Sliwinski and Buschke (1999) tested 302 participants, aged 66–92, over five visits at 18-month intervals, on the Mill Hill Vocabulary Scale (Raven, Court, & Raven, 1986) and the WAIS-R Vocabulary subtest (Wechsler, 1981) and also found declines over time. Finally, Zelinski and Burnight (1997) tested 106 participants on a cognitive battery on two occasions separated by 17 years. They tested vocabulary via the Recognition Vocabulary Test from the Schaie–Thurstone Adult Mental Abilities Test (Schaie, 1985) and in which participants select an appropriate synonym to a target from among four synonyms. For groups of participants who were younger than 63 at the first testing, no declines were seen at the 17-year follow-up, but for groups aged 64–69 and 70–81 at baseline, declines were detected at follow-up by approximately one standard deviation. These various studies serve as representatives of many more longitudinal studies that consistently show declines, though often moderate, in vocabulary in the years from age 50 through late life. The results clearly diverge from findings in the cross-sectional literature and point to the absolute

< previous page

page_260

next page >

< previous page

page_261

next page >

Page 261 necessity of examining age-changes longitudinally in addition to looking at age-differences cross-sectionally. Picture Naming/Confrontation Naming Older adults commonly complain of problems with word-finding. However, laboratory tests involving confrontation naming or picture naming suggest a more complex pattern of maintenance and decline. Cross-sectional studies using the Boston Naming Test have showed that older adults named significantly fewer pictures than young adults (Borod, Goodglass, & Kaplan, 1980; Nicholas, Obler, & Albert, 1985). The pattern seems to show minimal declines until age 70, with marked declines afterwards (Albert, Heller, & Mulberg, 1988). Similarly, on the Action Naming Test, older adults named fewer pictures than younger adults (Obler & Albert, 1979). However, other studies suggested that controlling for education (a cohort effect) minimized the differences (LaBarge, Edwards, & Knesevich, 1986; Van Gorp, Satz, Kiersch, & Henry, 1986), and yet other studies showed consistent performance across age or indeed increases in performance (Farmer, 1990). Such differences may be explicable in part by differences in observation intervals: Borod et al. (1980) and Nicholas et al. (1985) used a 10-year interval, while LaBarge et al. (1986) and Van Gorp et al. (1986) used 5-year intervals. Shorter intervals may not be sensitive to change in naming. Indeed, in a comprehensive qualitative review, Goulet, Ska, and Kahn (1994) attributed the mixed results in the literature to differences in sample, methods, and health status of participants across studies. To clarify the issue, Feyereisen (1997) conducted a meta-analytic study of comparable cross-sectional findings and concluded that there is in fact an age-related decline in naming ability that is seen after age 70. Longitudinal studies show somewhat mixed results as well, though it seems clear that the total length of the observation period is the critical variable for detecting change. The first two studies reviewed here employ observation periods of 3 and 4 years respectively, and find no declines within age groups. Mitrushina and Satz (1995) tested 122 older participants annually for 3 years. Again, participants were divided into four age groups (57– 65, 66–70, 71–75, and 76–85). Mean performances were similar across age groups and correlation coefficients for test–retest effects across the three administrations were high, suggesting little within group decline. Cruice, Worrall, and Hickson (2000) examined scores on the BNT from 91 participants on two occasions, separated by 4 years. Longitudinal analyses showed no significant changes over time, whereas cross-sectional analyses showed a modest correlation between age and decreasing scores on the BNT. Longer periods of observation, however, do show declines over time for older adults, as the following four studies show. Au et al. (1995) tested 53 adults ranging in age from 30 to 79 three times over a period of 7 years.

< previous page

page_261

next page >

< previous page

page_262

next page >

Page 262 Participants were divided into four groups: 30-, 50-, 60-, and 70-year-olds. Within groups, declines over the 7-year period were seen for all groups, except the 30-year-old group. Declines were greatest within the 70-year-old group vs. the others, and equal declines were seen for the 50- and 60-year-olds (which were still higher than the initial performance of the 70-year-old group). Ramsay, Nicholas, Au, Obler, and Albert (1999) tested 66 participants using the Action Naming Test three times over a 7-year period. Participants were divided into four age groups (30s, 50s, 60s, and 70s). Analysis of within group scores showed declines for all age groups except the 30-year-olds. In what is probably the most advanced study to date, Connor et al. (2004) tested 160 participants in the age ranges 30–39, 50–59, 60–69, and 70–80, between once and five times over a 20-year period. By using random effects modeling (Laird & Ware, 1982; Verbeke & Molenberghs, 2000) the authors were able to take into account the fact that not all participants were tested the same number of times (drop-out) and that intervals between observations differed between participants. Longitudinal analyses of the data showed that naming ability declined approximately two percentage points per year. A comparison of cross-sectional vs. longitudinal analyses showed larger estimates of age-related change at cross-section. As the authors point out, this study differs from previous research in length of follow-up (20 years) as well as the age range of the sample (30–80 years), and these factors, as well as the random effects modeling used to analyze the data, probably provide the most sensitive test to date of the effects of aging on naming ability. In sum, longitudinal studies of naming ability that employ observation times lasting from 7 to 20 years show significant declines in naming ability. These findings may help to resolve the questions left by cross-sectional studies of naming that so far have provided ambiguous results. Grammatical Complexity and Propositional Density For many years, Susan Kemper and colleagues have conducted both cross-sectional and longitudinal research on the simplified speech register of older vs. younger adults by inspecting grammatical complexity and propositional density in both oral and written language samples. The simplified speech register of older vs. younger adults is operationally defined as lower levels of grammatical complexity and lower levels of propositional density. Grammatical complexity is a measure of the numbers and kinds of embeddings used to construct complex sentences. Examples are subordinate clauses, infinitives, gerunds, and right- and left-branching forms. Propositional density is defined as the mean number of propositions per 100 words, and it measures the amount of information in a sentence relative to the number of words (Kemper, Herman, & Lian, 2003). In cross-sectional work on sentence production, older adults typically show lower levels of grammatical complexity

< previous page

page_262

next page >

< previous page

page_263

next page >

Page 263 and propositional density than younger adults, and these results are related to diminishing levels of working memory among older adults (Kemper et al., 2003; Kemper, Herman, & Liu, 2004). The same results obtain longitudinally as well. Looking at oral production, Kemper, Thompson, and Marquis (2001) report results from cognitively healthy older adults (Study 1) and a sample of patients with dementia (Study 2). Study 1 included a sample of 30 older adults, 65–70 years of age at first testing, from whom oral language samples were drawn annually for up to 15 years. Data were collected from all 30 participants for 7 years, from 25 for 10 years, from 10 for 12 years, and from 5 for 15 years. Results showed that the decline in both grammatical complexity and propositional density both reflected a cubic function of age. That is, a period of relative stability was followed by decline in the mid-70s and then a period of more gradual decline. Initial levels of grammatical complexity were predicted by working memory capacity, and initial levels of propositional content were predicted by scores on a vocabulary test. Interestingly, individuals with initially higher levels of grammatical complexity and propositional density experienced more rapid declines on these variables with advancing age. The same patterns are seen longitudinally in written production. Kemper et al. (2001) analyzed written autobiographies and language samples spanning a 60-year time frame from the Nun Study (Snowden, 1997). Again, considering only cognitively healthy participants, both grammatical complexity and propositional content (idea density) showed a linear decline over the 60-year interval. That is, in initial samples taken when the participants were between 17 and 32 years of age between 1931 and 1943, participants used a greater range of embedded, subordinate clauses and related more ideas in fewer words than they did when annual cognitive testing was resumed some 60 years later in 1995–1996. Both of these studies are excellent examples of the benefits of random effects modeling (Laird & Ware, 1982) because the analyses were able to take into account participant attrition over time and varying intervals of data collection. Both involve differing intervals of data collection and experimental/participant mortality. In both studies, individual declines are modeled via separate intercepts and slopes and these are then used in subsequent analyses involving independent predictors (e.g. working memory capacity, years of education, socioeconomic status, etc.). Conclusion The current state of research on multilingualism reflects a paucity of studies employing longitudinal designs, though many cross-sectional studies have been done. However, as the extensive literature on cognitive aging shows, crosssectional studies cannot tell us much about within-person change over time. Given that multilingualism is inherently dynamic, in the sense that language

< previous page

page_263

next page >

< previous page

page_264

next page >

Page 264 proficiency itself is dynamic, it would seem that longitudinal studies are sorely needed to develop theory-based research in the field. The study of multilingualism across the lifespan is therefore in much the same position as gerontology some years ago, and “jump-starting” longitudinal research on multilingualism can profitably begin by attending to the history of cognitive gerontology. Of the longitudinal designs reviewed in this chapter, those termed “longitudinal sequential” designs are probably the most promising since they combine data gathering from multiple cohorts over multiple repeated occasions of measurement. Such designs facilitate the identification and characterization of interindividual change (within-person trajectories) and intraindividual comparison of these trajectories (between-person differences in rates of change; see also Schaie’s “most efficient” method; Schaie & Caskie, 2005). Perhaps the most frequent analytic strategy applied to such designs is Schaie’s cohort-sequential method in which at least two cohorts are studied across three occasions of measurement (Schaie, 1965, 2005). Statistically, such designs are especially facilitated by the use of random effects models (Laird & Ware, 1982; Singer & Willet, 2003) that have made it possible to model individual intercepts (initial levels of performance) and slopes (rates of change) and then compare them at higher levels of analysis. As with cross-sectional studies of multilingualism, longitudinal studies also require attention to particular threats to validity. Perhaps the most salient is the operationalization and assessment of language proficiency in the two (or more) languages of the participants. In this, longitudinal research has the distinct advantage of including a baseline measure against which changes in proficiency can be modeled over time. However, the investigator who is interested in how age-related changes in cognitive mechanisms affect language production may want to hold proficiency constant, and in this case changes in language environment become an additional threat to validity. Multiple measurement brings its own set of threats to validity. Practice or retest effects are particularly problematic when measuring language because linguistic stimuli may be recalled from one measurement occasion to the next. Similarly, selection effects exist when those returning for subsequent evaluation are healthier, more able, or more motivated than drop-outs. These and other threats to the validity of longitudinal research have received considerable attention in the cognitive aging literature, and many sensible solutions may be found there. Finally, the many existing longitudinal studies of language and aging provide an important point of departure for a longitudinal approach to multilingualism. Many of these are limited to very low levels of language production and comprehension (i.e. at the level of vocabulary or naming ability), and researchers of multilingualism may be interested in more molar forms of production and comprehension in second and later languages. Nevertheless, in contrast to many other (non-linguistic) cognitive abilities that show earlier declines, the notion that these particular abilities show little decline

< previous page

page_264

next page >

< previous page

page_265

next page >

Page 265 until later ages signals a robust retention of language abilities in middle age and early old age. This raises the question whether second languages experience the same pattern of retention. Though cross-sectional work suggests that they do (Schrauf, 2008), no longitudinal studies have made this point. Finally, the work of Kemper and colleagues (Kemper et al., 2003, 2004) showing declines in syntactic complexity and propositional density raises interesting questions concerning these abilities for second languages. If working memory is responsible for these declines, then one might expect similar patterns for second languages. Again, only a longitudinal study can definitively address the issue. References Albert, M. S., Heller, H. S., & Mulberg, W. (1988). Changes in naming ability with age. Psychology and Aging , 3 , 173–178. Alwin, D. F., & McCammon, R. J. (1999). Aging versus cohort interpretation of intercohort differences in GSS vocabulary scores. American Sociological Review, 64, 272–286. Alwin, D. F., & McCammon, R. J. (2001). Aging, cohorts, and verbal ability. Journal of Gerontology, Social Sciences, 56B(3), S151–S161. Au, R., Joung, P., Nicholas, M., Kass, R., Obler, L. K., & Albert, M. L. (1995). Naming ability across the lifespan. Aging and Cognition, 2 , 300–311. Baltes, P. B. (1968). Longitudinal and cross-sectional sequences in the study of age and generation effects. Human Development , 11, 145–171. Baltes, P. B. (1987). Theoretical propositions of life-span developmental psychology: On the dynamics between growth and decline. Developmental Psychology, 23, 611–626. Baltes, P. B., & Lindenberger, U. (1997). Emergence of powerful connection between sensory and cognitive functions across the adult lifespan: A new window to the study of cognition. Psychology and Aging , 12, 12–21. Baltes, P. B., Reese, H. W., & Nesselroade, J. R. (1977). Lifespan developmental psychology: An introduction to research methods . Monterey, CA: Brooks Cole (reprinted 1988, Hillsdale, NJ: Lawrence Erlbaum). Baltes, P. B., Schaie, K. W., & Nardi, A. H. (1971). Age and experimental mortality in a seven-year longitudinal study of cognitive behavior. Developmental Psychology, 5 , 18–26. Bialystock, E., & Craik, F. I. M. (Eds.). (2006). Lifespan cognition: Mechanisms of change . New York: Oxford University Press. Bialystok, E., Craik, F. I. M., Klein, R., & Viswanathan, M. (2004). Bilingualism, aging, and cognitive control: Evidence from the Simon Task. Psychology and Aging , 19(2), 290–303. Bialystok, E., Craik, F. I. M., & Ryan, J. (2006). Executive control in a modified antisaccade task, Effects of aging and bilingualism. Journal of Experimental Psychology, Learning, Memory, and Cognition, 32(6), 1341–1354. Borod, J. C., Goodglass, H., & Kaplan, E. (1980). Normative data on the Boston

< previous page

page_265

next page >

< previous page

page_266

next page >

Page 266 Diagnostic Aphasia Exam, Parietal Lobe Battery, and the Boston Naming Test. Journal of Clinical Neuropsychology , 2 , 209–215. Bowles, N. L., Grimm, K. J., & McArdle, J. J. (2005). A structural factor analysis of vocabulary knowledge and relation to age. Journal of Gerontology, Psychological Sciences, 60B, P234–P241. Bryk, A. S., & Raudenbush, S. W. (1987). Application of hierarchical linear models to assessing change. Psychological Bulletin, 10(12), 147–158. Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods . Newbury Park, CA, Sage. Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental designs for research. Washington, DC: AERA. Connor, L. T., Spiro III, A., Obler, L. K., & Albert, M. L. (2004). Change in object naming ability during adulthood. Journal of Gerontology, Psychological Sciences, 59B(5), P203–P209. Craik, F. I. M., & Salthouse, T. A. (Eds.). (2000). Handbook of aging and cognition . Mahwah, NJ: Lawrence Erlbaum. Cruice, M. N., Worrall, L. E., & Hickson, L. M. H. (2000). Boston Naming Tests results for healthy and older Australians: A longitudinal and cross-sectional study. Aphasiology , 14(2), 143–155. de Bot, K., & Clyne, M. (1994). A 16-year longitudinal study of language attrition in Dutch immigrants in Australia. Journal of Multilingual and Multicultural Development , 15(1), 17–28. de Bot, K., & Lintsen, T. (1986). Foreign-language production in the elderly. In B. Weltens, K. de Bot, & T. van Els (Eds.), Language attrition in progress (pp. 131–142). Dordrecht: Foris. de Bot, K., & Makoni, S. (2005). Language and aging in multilingual contexts . Clevedon, UK: Multilingual Matters. Dietrich, R., Klein, W., & Noyau, C. (1995). The acquisition of temporality in a second language. Philadelphia: John Benjamins. Ekstrom, R. B., French, J. W., Marman, H. H., & Derman, D. (1976). Manual for kit of factor related referenced cognitive tests. Princeton, NJ: Educational Testing Service. Farmer, A. (1990). Performance of adult males on the Boston Naming Test and the Word Test. Aphasiology , 4 , 293– 296. Ferrer, E., Salthouse, T. A., Stewart, W. F., & Schwartz, B. S. (2004). Modeling age and retest processes in longitudinal studies of cognitive abilities. Psychology and Aging , 19(2), 243–259. Feyereisen, P. (1997). A meta-analytic procedure shows an age-related decline in picture-naming: Comments on Goulet, Ska, and Kahn. Journal of Speech and Hearing Research , 40, 1328–1333. Glenn, N. D. (1999). Further discussion of the evidence for an intercohort decline in education-adjusted vocabulary. American Sociological Review, 64, 267–271. Goldstein, H. (1995). Multilevel statistical models (2nd. ed). New York: Halstead Press Goral, M. (2004). First-language decline in healthy aging: Implications for attrition in bilingualism. Journal of Neurolinguistics, 17, 31–52. Goulet, P., Ska, B., & Kahn, H. J. (1994). Is there a decline in picture naming with advancing age? Journal of Speech and Hearing Research, 37, 629–644.

< previous page

page_266

next page >

< previous page

page_267

next page >

Page 267 Hernandez, A. E., & Kohnert, K. J. (1999). Aging and language switching in bilinguals. Aging in Neuropsychology and Cognition, 6 (2), 69–83. Hertzog, C. (1996). Research design in studies of cognitive aging. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the psychology of aging (4th ed., pp. 24–37). New York: Academic Press. Hertzog, C., & Nesselroade, J. R. (2003). Assessing psychological change in adulthood: An overview of methodological issues. Psychology and Aging , 18(4), 639–657. Hofer, S. A., & Sliwinski, M. (2006). Design and analysis of longitudinal studies on aging. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the psychology of aging (6th ed., pp. 17–40). New York: Elsevier. Horn, J. L., & Cattell, R. B. (1967). Age differences in fluid and crystallized intelligence. Acta Psychologica , 26, 107– 129. Hultsch, D. F., Hertzog, C., Small, B. J., McDonald-Miszczak, L., & Dixon, R. A. (1992). Short-term longitudinal change in cognitive performance in later life. Psychology and Aging , 7 (4), 571–584. Juncos-Rabadan, O. (1994). The assessment of bilingualism in normal aging with the bilingual aphasia test. Journal of Neurolinguistics, 8 (1), 67–73. Kemper, S., Greiner, L. H., Marquis, J. G., Prenovost, K., & Mitzner, T. L. (2001). Language decline across the lifespan, Findings from the Nun Study. Psychology and Aging 16 (2), 227–239. Kemper, S., Herman, R. E., & Lian, C. (2003). Age differences in sentence production. Journal of Gerontology, Psychological Sciences, 58B, P260–P269. Kemper, S., Herman, R. E., & Liu, C.-J. (2004). Sentence production by young and older adults in controlled contexts. Journal of Gerontology, Psychological Sciences, 59B, P220–P234. Kemper, S., Thompson, M., & Marquis, J. G. (2001). Longitudinal change in language production: Effects of aging and dementia on grammatical complexity and propositional content. Psychology and Aging , 16, 600–614. Klein, W., and C. Perdue. (1992). Utterance structure. Developing grammars again . Amsterdam/Philadelphia: John Benjamins. Kopke, B. (2004). Neurolinguistic aspects of attrition. Journal of Neurolinguistics, 17, 3–30. Kreft, I., & de Leeuw, J. (1998). Introducing multilevel modeling. Thousand Oaks, CA: Sage. LaBarge, E., Edwards, D., & Knesevich, J. W. (1986). Performance of normal elderly on the Boston Naming Test. Brain and Language, 27, 380–384. Lachman, R., Lachman, J. L., & Taylor, D. W. (1982). Re-allocation of mental resources over the productive lifespan, Assumptions and techniques. In F. I. M. Craik & S. Trehub (Eds.), Aging and cognitive processes. New York: Plenum Press. Laird, N. M., & Ware, J. H. (1982). Random effects models for longitudinal data. Biometrics, 38, 963–974. Lindenberger, U., Singer, T., & Baltes, P. B. (2002). Longitudinal selectivity in aging populations: Separating mortality associated versus experimental components in the Berlin Aging Study (BASE). Journal of Gerontology, Psychological Sciences, 57B, P474–P482. McArdle, J. J., & Woodcock, R. J. (1997). Expanding test–retest designs to include developmental time-lag components. Psychological Methods, 2 (4), 403–435. McArdle, J. J., Ferrer-Caja, E., Fumiaki, H., & Woodcock, R. J. (2002). Comparative

< previous page

page_267

next page >

< previous page

page_268

next page >

Page 268 longitudinal structural analyses of the growth and decline of multiple intellectual abilities over the lifespan. Developmental Psychology, 38(1), 115–142. McArdle, J. J., Hamagami, F., Elias, M. F., & Robbins, M. A. (1991). Structural modeling of mixed longitudinal and cross-sectional data. Experimental Aging Research , 17, 29–51. MacDonald, S. W. S., Hultsch, D. F., Strauss, E., & Dixon, R. A. (2003). Age-related slowing of digit symbol substitution revisited: What do longitudinal age changes reflect? Journal of Gerontology, Psychological Sciences, 58B(3), P187–P194. Mitrushina, M., & Satz, P. (1995). Repeated testing of normal elderly with the Boston Naming Test. Aging , 7 , 123– 127. Nicholas, M., Obler, L. K., & Albert, M. L. (1985). Lexical retrieval in healthy aging. Cortex , 21, 595–606. NORC. (1999). General Social Surveys, 1972–1996: Cumulative codebook . Chicago: National Opinion Research Center. Obler, L. K., & Albert, M. L. (1979). The Action Naming Test, experimental edition . Boston: VA Medical Center. Obler, L., Albert, M., & Lozowick, S. (1986). The aging bilingual. In J. Vaid (Ed.), Language processing in bilinguals: Psycholinguistic and neuropsychological perspectives (pp. 221–231). Hillsdale, NJ: Lawrence Erlbaum. Ortega, L., & Iberri-Shea, G. (2005). Longitudinal research in second language acquisition: Recent trends and future directions. Annual Review of Applied Linguistics , 25, 26–45. Park, D. C. (2000). The basic mechanisms accounting for age-related decline in cognitive function. In D. C. Park & N. Schwarz (Eds.), Cognitive aging: A primer (pp. 1–21). Philadelphia, PA: Psychology Press. Park, D. C., & Schwarz, N. (Eds.). (2000). Cognitive aging: A primer. Philadelphia, PA: Taylor & Francis. Park, D. C., Smith, A. D., Lautenschlager, G., Earles, J., Frieske, D., Zwahr, M., and Gaines, C. (1996). Mediators of long-term memory performance across the lifespan. Psychology and Aging , 11, 621–637. Picciotto, J. de, & Friedland, D. (2001). Verbal fluency in elderly bilingual speakers: Normative data and preliminary application to Alzheimer’s disease. Folia Phoniatrica et Logopaedica, 53, 145–152. Powers, E. A., & Bultena, G. L. (1972). Characteristics of deceased dropouts in longitudinal research. Journal of Gerontology, 27, 350–353. Rabbitt, P., Diggle, P., Holland, F., & McInnes, L. (2004). Practice and drop-out effects during a 17-year longitudinal study of cognitive aging. Journal of Gerontology, Psychological Sciences, 59(2), P84–P97. Rabbitt, P., Diggle, P., Smith, D., Holland, F., & McInnes, L. (2001). Identifying and separating the effects of practice and cognitive ageing during a large longitudinal study of elderly community residents. Neuropsychologia , 39, 352– 543. Rabbitt, P., Watson, P., Donlan, C., Bent, L., & McInnes, L. (1994). Subject attrition in a longitudinal study of cognitive performance in community-based elderly people. In B. J. Vellas, J. L. Albarede, & P. J. Garry (Eds.), Facts and research in gerontology (pp. 203–207). Paris: Serdi. Ramsay, C. B., Nicholas, M., Au, R., Obler, L., & Albert, M. (1999). Verb naming in normal aging. Applied Neuropsychology , 6 (2), 57–67. Raven, J. (1982). Revised manual for Raven’s Progressive Matrices and Vocabulary Scale. Windsor, UK: NFER Nelson.

< previous page

page_268

next page >

< previous page

page_269

next page >

Page 269 Raven, J. C., Court, J. H., & Raven, J. (1986). Manual for Raven’s Progressive Matrices and vocabulary scales . Oxford: Psychologists’ Press. Rosselli, M., Ardila, A., Araujo, K., Weekes, V. A., Caracciolo, V., Padilla, M., and Ostrosky, F. (2000). Verbal fluency and repetition skills in healthy older Spanish–English bilinguals. Applied Neuropsychology , 7 (1), 17–24. Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review, 103 , 403–428. Schafer, J. L. (1997). Analysis of incomplete multivariate data. London: Chapman and Hall. Schaie, K. W. (1965). A general model for the study of developmental problems. Psychological Bulletin, 64(2), 92– 107. Schaie, K. W. (1985). Schaie–Thurstone Adult Mental Abilities Test. Palo Alto, CA: Consulting Psychologists’ Press. Schaie, K. W. (1994). The course of adult intellectual development. American Psychologist , 49(4), 304–313. Schaie, K. W. (1996). Intellectual development in adulthood. Cambridge: Cambridge University Press. Schaie, K. W. (2000). The impact of longitudinal studies on understanding development from young adulthood to old age. International Journal of Behavioral Development , 24(3), 257–266. Schaie, K. W. (2005). Developmental influences on adult intelligence: The Seattle Longitudinal Study. New York: Oxford University Press. Schaie, K. W., & Baltes, P. (1975). On sequential strategies in developmental research: Description or explanation. Human Development , 18, 384–390. Schaie, K. W., & Caskie, G. L. (2005). Methodological issues in aging research. In D. M. Teti (Ed.), Handbook of research methods in developmental psychology. Cambridge: Blackwell. Schaie, K. W., & Hofer, S. A. (2001). Longitudinal studies in aging research. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the psychology of aging (5th ed., pp. 53–77). New York: Elsevier. Schrauf, R. W. (2008). Bilingualism and Aging. In R. Heredia and J. Altariba (Eds.), An introduction to bilingualism: Principles and processes (pp. 105–127). Mahwah, NJ: Lawrence Erlbaum. Schrauf, R. W. (under review). English among older immigrants in linguistically isolated neighborhoods: Proficiency and patterns of use. Shipley, W. C. (1946). Institute of Living Scale. Los Angeles, CA: Western Psychological Services. Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York: Oxford. Sliwinski, M., & Buschke, H. (1999). Cross-sectional and longitudinal relationships among age, cognition, and processing speed. Psychology and Aging , 14(1), 18–33. Snowden, D. A. (1997). Aging and Alzheimer’s disease, Lessons from the Nuns Study. The Gerontologist , 37, 150– 156. Thorvaldsson, V., Hofer, S. A., Berg, S., & Johansson, B. (2006). Effects of repeated testing in a longitudinal agehomogenous study of cognitive aging. Journal of Gerontology, Psychological Sciences, 61B(6), P348–P354. Van Gorp, W. G., Satz, P., Kiersch, M. E., & Henry, R. (1986). Normative data on the Boston Naming Test for a group of normal older adults. Journal of Clinical and Experimental Neuropsychology , 8 , 702–705.

< previous page

page_269

next page >

< previous page

page_270

next page >

Page 270 Verbeke, G., & Molenberghs, G. (2000). Linear mixed models for longitudinal data. New York: Springer. Verhaeghen, P. (2003). Aging and vocabulary scores, A meta-analysis. Psychology and Aging , 18(2), 332–339. Wechsler, D. (1955). WAIS manual. New York: Psychological Corporation. Wechsler, D. (1981). WAIS-R manual. New York: Psychological Corporation. Woodcock, R. W., & Munoz-Sandoval, A. F. (1996). Bateria Woodcock–Munoz, Pruebas de habilidad cognitivaRevisada. Itasca, IL: Riverside. Xue, S. A., Hagstrom, F., & Hao, J. (2002). Speaking fundamental frequency characteristics of young and elderly bilingual Chinese–English speakers: A functional system approach. Asia Journal of Speech, Language, and Hearing , 7 , 55–62. Zelinski, E. M., & Burnight, K. P. (1997). Sixteen-year longitudinal and time lag changes in memory and cognition in older adults. Psychology and Aging , 12, 503–513.

< previous page

page_270

next page >

< previous page

page_271

next page >

Page 271 11 THE ROLE OF WORKING MEMORY IN LANGUAGE DEVELOPMENT OVER THE LIFESPAN Susan Kemper Working memory is generally assumed to contribute to language acquisition by children and to language processing by children and adults. This chapter begins with an overview of working memory, focusing on tests used to assess working memory capacity and executive function. It concludes by reviewing research on the development of working memory across the life span and the role of working memory in language acquisition and language processing. Concept of Working Memory Working memory is essential to many everyday tasks that involve the retention of information; working memory has two functions: the short-term retention of information and the manipulation of information. The prevailing model of working memory, as proposed by Baddeley and Hitch (1974), involves three components: two temporary storage mechanisms that buffer visual, for example, the visual scratchpad, and auditory information, for example, the phonological loop, and a central executive processor. A fourth component, an episodic buffer linked to long-term memory, was added by Baddeley (2000). Cowan (1995, 2001), McElree (2001), and Oberauer and Kliegl (2006), among others, have proposed mixture models linking attention and working memory. A chief characteristic of these multicomponent systems is that the system has limited capacity—to temporarily store information or to divide attention among processing tasks. Each component is also assumed to have unique characteristics: the auditory buffer is speech-based while the visual buffer is spatially defined. The executive component subserves different functions including attentional allocation and selection, inhibition, and information updating. Evidence for functional separation of these components of working memory comes from studies of healthy and impaired individuals responding

< previous page

page_271

next page >

< previous page

page_272

next page >

Page 272 to different task manipulations: (1) On tests of immediate serial recall, performance is worst for phonologically similar word lists, suggesting that verbal information is held in a phonologically-based short-term store; (2) Serial recall also varies with word length and with reading time, suggesting this phonologically based buffer has a limited capacity; (3) The continuous articulation of irrelevant speech, for example, repeating “the, the, the …” impairs recall, eliminates the phonological-similarity effect, and the word length effect, again suggesting that this buffer is speechbased; (4) Concurrent engagement in a spatial tracking task such as pointing to the source of a moving sound while blindfolded impairs performance on spatial memory tests, suggesting that the sketchpad is spatial in nature. One challenge to understanding the role of working memory in language acquisition and processing is the multiplicity of tests and assessments used to measure individual differences in working memory. Working memory is typically defined by tests of working memory span and by tests of executive function. Executive function itself is measured by neuropsychological tests such as the Wisconsin Card Sorting Test, the Tower of Hanoi problem, the Trail Making Test, or tests of verbal fluency, or specific tests of inhibition, time-sharing, updating, and switching. These tests are briefly described below; many variants of each test have been developed. Tests of working memory include a variety of different span tests, most derived from the Forward and Backward Digit Span tests (Wechsler, 1958). These span tests present participants with a series of sets of digits to be recalled in the correct serial order, either a forward order or backward order. The sets progressive increase in size, from two to nine digits per set. The measure of working memory capacity is the highest number of digits that can be correctly recalled in two of three sets at that length. Many variants of letter and word span tests have also been developed. In addition, Counting Span (Case, Kurland, & Goldberg, 1982) is widely used with children and clinical populations because of its simplicity. Participants count shapes, such as green dots, presented in a random visual array interspersed among other shapes, such as yellow dots, and remember the count total. A series of sets, increasing from one to five arrays per set, is presented. The dependent variable is the highest level at which the participant is able to get two of three sets correct. One of the most widely used tests is the Reading Span (Daneman & Carpenter, 1980) test. A related Listening Span test has also been developed. Participants read (or listen to) sentences and try to remember the final word of each sentence. Progressively larger sets of sentences (13 to 16 words in length) are presented, increasing from two to six sentences in a set with three sets being tested at each level. Reading span is defined as the level at which the participant is able to recall all of the items in two of the three sets tested at that level. Reading span has been shown to be highly predictive of reading skill (Daneman & Merikle, 1996; Daneman & Tardif, 1987).

< previous page

page_272

next page >

< previous page

page_273

next page >

Page 273 More recently, the OSPAN or operational span test has been developed by Turner and Engle (1989) as a working memory measure. OSPAN involves a presenting a sequence of trials, each consisting of a set of two, three, four, or five pairs of mathematical operations and to-be-remembered words. The mathematical operations, e.g. 9/3 − 2 = 1, are read aloud and then verified as correct or incorrect and the target word is read aloud and remembered. Operational span is defined as the level at which the participant is able to recall of the words in two of three sets. There is some debate regarding whether verbal working memory span is distinct from visual–spatial span; consequently, specialized tests of visual–spatial span have been developed. A frequently used measure of visual– spatial span is the Corsi blocks test (Milner, 1971). This test requires participants to remember a sequence of spatial locations, pointing to the correct sequence in either a forward or a backward order. Wooden blocks are unevenly distributed over a flat board; the experimenter taps a sequence of blocks and the subject is asked to tap out the same sequence or the reverse sequence. The sequences are random and the difficulty level is progressively raised by increasing the number of blocks tapped. There are three trials at each difficulty level. The subject’s spatial span is conventionally taken to be the longest sequence in which at least two out of the three sequences are correctly reproduced. Another measure of visual–spatial span is the Visual Pattern Test (Logie, Zucco, & Baddeley, 1990), designed to assess non-verbal visual short-term memory. The participant is presented with checker-board patterns, which have been designed to be difficult to code verbally. A visual pattern was created by filling in half the squares in a grid. The grids progress in size from the smallest, a 2 × 2 matrix (with two filled cells), to the largest, a 5 × 6 matrix (with 15 filled cells), complexity being steadily increased by adding two more cells to the previous grid. The patterns are displayed in a series of stimulus cards and the participant is asked to reproduce the pattern by marking squares in an empty grid of the same size as the one bearing the pattern just presented. The dependent measure is the number of filled cells in the most complex pattern recalled in the range 2–15 cells. Executive function is classically measured using the Wisconsin Card Sorting Test; a computerized version has been developed by Kimberg, D’Esposito, and Farah (2000). Participants sort cards based on color, number, or shape. Participants are not told the sorting criterion, but receive feedback concerning their sorts and continue sorting one card at time until eight correct sorts have been performed; at this point the sorting criterion changes and participants must again sort cards until eight correct sorts have been made. The test continues until 15 sorting categories or 288 sorts have been made. The dependent measure is the number of perseverative errors (when the sorting criterion changed but the participant failed to change sorting pattern). The Trail Making Test (Spreen & Strauss, 1991) is another commonly used neuropsychological test of executive function. It involves two tasks.

< previous page

page_273

next page >

< previous page

page_274

next page >

Page 274 Task A is a simple connect-the-dots task requiring the participant to connect sequentially numbered dots; Task B requires the participant to connect dots using an alternating sequence of letters and numbers. A difference score (BA) or proportional difference score (B-A/A) is usually computed. In addition, neuropsychologists commonly assess executive function using the FAS verbal fluency test (Benton & Hamsher, 1989). Participants generate as many words as possible beginning with a target letter, e.g., F, A, or S, in a one-minute period. The dependent measure is the total number of correct responses. There are many variants of the basic fluency test, including category and action fluency tests, and there is a current debate over how best to assess performance on these tests (see Kemper & McDowd, 2008). Executive function can also be assessed using the Tower of Hanoi test; a computerized version was developed by Hume, Welsch, Retzlaff, & Cookson (1997). Participants rearrange a set of disks varying in size placed on three pegs in order to match a target configuration. Moves are constrained such that only one disk can be moved at a time and a larger disk cannot be placed on a smaller disk. The dependent measure is the total number of moves required to complete the two problems. Often executive function is decomposed into subcomponents including inhibition, time-sharing, updating, and switching with specialized tests of each component. Inhibition is usually measured using the Stroop task. It involves two blocks of trials. On the first block of trials, strings of XXXXs are presented printed in different color inks and the participant must name the color of the ink for as many strings as possible during one minute; on the second block of trials, the words “red,” “green,” “blue,” and “yellow” are printed in ink of a contrasting color and the participant again must name the color of the ink for as many words as possible in one minute. A difference score (color wordsXXXs) or proportional difference score (color words-XXXs/color words) is computed. A more recently devised test of inhibition is the reading with distraction task (Connelly, Hasher & Zacks, 1991). It requires participants to read four short passages; two of the passages contain distracting words printed in a different type font. After reading the passages, the participants are tested on their comprehension. Two dependent variables are obtained: average difference in reading time for passages with and without distracters and average difference in comprehension accuracy for passages with and without distracters. The stop signal task (Logan, 1994) is also used to assess inhibition. It involves two blocks of trials. In the first block of 24 trials, participants categorize a sequence of words as animal/non-animal as rapidly as possible. In the second block of 192 trials, participants are instructed not to respond when they hear a tone (presented on 48 randomly selected trials—stop trials). The dependent measure is the proportion of stop trials on which the participant produces categorical responses.

< previous page

page_274

next page >

< previous page

page_275

next page >

Page 275 Time-sharing, another component of executive function, can be assessed using various combinations of tests in a dual-task paradigm. In one, the counting backwards plus connections task (Salthouse, Atkinson, & Berish, 2003) participants try to connect a series of numbered dots in order (the dots are scattered about the page) while counting backwards by three. The dependent measure is the average time per subtraction. Another dual-task measure of time-sharing is the tracking task plus paired associates task (Salthouse & Miles, 2002) which requires participants to use a track ball to keep a cursor on a randomly moving white circle while at the same time performing a pairedassociate learning task. The difficulty of the tracking task is individually adjusted to match a pre-specified level of accuracy. The learning task involves listening to a series of word pairs followed by the presentation of the first member of each pair and recall of the matching word. The dependent variable is the number of correct responses. Updating can also be considered a component of executive function and tests have been devised to assess updating. In the digit monitoring task (Salthouse et al., 2003) participants listen to a series of digits; they respond to every third odd digit by pressing the letter Z and all other items by pressing the letter M. The primary dependent variable is the percentage of correct responses. In the letter memory (N. Morris & Jones, 1990) lists of letters are presented. The task is to recall the last four letters presented in the list. The number of letters (5, 7, 9, or 11) varies randomly across trials. The dependent measure is the proportion of letters recalled correctly. A widely used test of updating is the n-back task (Mackworth, 1959). This task requires the participant to listen to a sequence of digits and to repeat the digit that occurred n items back; one-back, two-back, etc., tests can be used. The dependent variable is the number of errors. Switching is also considered to be a component of executive function that can be separately assessed. The plus– minus switching task (Miyake, Friedman, Emerson, Witzki, & Howerter, 2000) consists of three lists of two-digit numbers. On the first list, participants add three to each, on the second, they subtract three, and on the third, they alternate adding and subtracting three from each number. A difference score is computed between the time to complete the alternating list and the average time to complete the addition and subtraction lists; a proportional difference score can also be computed. The letter-letter switching task (Miyake et al., 2000) presents a number and letter pair in one of four quadrants of a computer screen. Participants are instructed to indicate whether the number was odd or even if the pair was presented in either of the two upper quadrants or whether the letter was a vowel or consonant when the pair was presented in either of the two lower quadrants. The number–letter pair is presented only in the two upper quadrants for a block of 32 trials, then only in the two lower quadrants for a block of 32 trials, and then alternating among all four quadrants for a block of 32 trials. The dependent measure is the difference in reaction time for the third block of

< previous page

page_275

next page >

< previous page

page_276

next page >

Page 276 trials minus the average reaction time for the first two blocks of trials; a proportional difference score can also be computed. The local–global switching task (Navon, 1977) requires a participant to focus on either a global figure (e.g. a triangle) or its component “local” figures (e.g. small squares) that compose the global figure. Participants are cued to say aloud the number of lines of the global figure (one for circle, two for X, three for triangle, four for square) or the number of lines in the local figures. The focus may shift across trials or not. A difference reaction time is computed for the response latencies between shift and no-shift trials; a proportional difference score can also be computed. There is considerable debate as to whether these tests assess separate but correlated abilities or a unitary construct, and their relationship to general intelligence. For example, Engle, Tuholski, Laughlin, and Conway (1999) suggest that span tasks may be differentiated into simple span tests assessing short-term memory and complex span tests also involving executive processes. A further issue is that working memory can be subdivided into verbal and nonverbal (or visual–spatial) domains. See also the “user’s guide” developed by Conway et al. (2005) for a discussion of many methodological and procedural problems in measuring working memory capacity using counting, operational span, and reading span tests. Executive function itself has been typically defined very broadly, as “those capacities that enable a person to engage successfully in independent, purposive, self-serving behavior” (Lezak, Howieson, Loring, Hannay, & Fischer, 2004, p. 35), or as “a multidimensional construct referring to a variety of loosely related higher-order cognitive processes including initiation, planning, hypothesis generation, cognitive flexibility, decision-making, regulation, judgment, feedback utilization, and self perception” (Spreen & Strauss, 1998, p. 171), and as bundle of “general purpose control mechanisms that modulate the operation of various cognitive subprocesses” (Miyake et al., 2000, p. 50). There is no single measure that serves as the “gold standard” for the assessment of executive function (EF). Salthouse et al. (2003), noting the complexity and breadth of notions of EF, undertook an examination of the construct validity of executive function in a sample of 261 adults ranging in age from 18 to 84 years. Their approach was to examine convergent and discriminant validity among a set of neuropsychological and cognitive tasks typically associated with executive function, and also a set of psychometric tasks including measures of verbal ability, fluid intelligence, episodic memory, and perceptual speed. A series of structural equation analyses were then conducted to look at the relations among these sets of variables. Their results indicated that the various neuropsychological measures were not very highly related to one another and were fairly highly related to other variables, particularly fluid intelligence. They concluded that individual differences in measures of working memory may in fact reflect differences in much broader abilities, such as fluid intelligence.

< previous page

page_276

next page >

< previous page

page_277

next page >

Page 277 Miyake and colleagues (Friedman & Miyake, 2004; Miyake et al., 2000) have addressed similar questions, but take a somewhat different approach, and reach different conclusions. Miyake et al. (2000) reported a study addressing “the unity and diversity of executive functions” (p. 49) using confirmatory factor analysis and structural equation modeling. They found that a three-factor solution fit the data better than any of the one- or two-factor solutions, indicating that there are separable dimensions of executive function. The authors conclude from this study that the three executive functions they measured (updating, shifting, inhibition) are “clearly distinguishable,” and that each plays a different role in more complex executive function measures such as the Wisconsin Card Sorting Test and the Tower of Hanoi. Working Memory across the Lifespan As measured by simple and complex span tests, working memory increases in childhood (Dempster, 1980; Gathercole, 1999; Pickering, 2001). Just as there is little agreement regarding the unity of working memory, there is little agreement regarding the differentiation or dedifferentiation of simple and complex spans in children, with some reports suggesting a dissociation between visual and verbal spans in children (Pickering, Gathercole & Peaker, 1998) and others reporting correlated span measures. What drives this increase in working memory capacity is also the subject of considerable debate. Case, Kurland and Goldberg (1982) proposed that differences in working memory span are determined by differences in processing speed. This conclusion was derived from the observation that counting span was linearly related to the speed of counting operations. An alternative view is that slower processing allows time for more forgetting to occur, or conversely, more time for the application of mnemonic strategies to boost retention; under this view, development differences in working memory span may be determined by the development of strategies for switching between processing and storage operations (Hitch, Towse, & Hutton, 2001). Thus, young children may have very limited working memories because of capacity limitations, limitations of processing speed, or limited executive function. An age-related decline in working memory is also apparent at the other end of the lifespan, and what drives this decline is under debate (Park et al., 1996). Some have suggested that the decline in visual–spatial measures of working memory, such as the Corsi Blocks tasks, evidence a faster rate of decline than verbal measures (Jenkins, Myerson, Joerding, & Hale, 2000) whereas others report similar rates of decline (Salthouse, Kausler, & Saults, 1990). There is a similar debate between those who hold that age differences to tasks that tap short-term storage are minimal and those who hold that storage and processing produce large age-differences (Baddeley, 1986; Dobbs

< previous page

page_277

next page >

< previous page

page_278

next page >

Page 278 & Rule, 1989). Executive function deficits have also been linked to Alzheimer’s dementia (Baddeley, Bressi, Della Sala, Logie, & Spinnler, 1991; R. G. Morris & Baddeley, 1988), leading to poor performance on tests requiring divided attention, memory retrieval, and short-term retention. A recent claim has been that working memory for emotional information is unimpaired with advancing age, consistent with socio-emotional selectivity theory (Mikels, Larkin, Reuter-Lorenz, & Carstensen, 2005; Carstensen, Mikels, & Mather, 2006) which holds that older adults seek to maintain a stable emotional balance. A variety of mechanisms have been proposed to account for the decline in working memory with advancing age. Salthouse (1994, 1996) has argued for processing speed as the fundamental mechanism; Lindenberger and Baltes (1994; Baltes & Lindenberger, 1997) have argued for neural integrity as measured by sensory acuity as the critical factor; and Hasher and Zacks (1979) have argued for a breakdown in inhibitory functions. Inhibition is critical for blocking irrelevant information from entering working memory, deleting irrelevant information from working memory, and restraining prepotent responses. Under this hypothesis, older adults with poor inhibitory mechanisms may not only be more susceptible to distraction, but they may also be less able to switch rapidly from one task to another and they may rely on well-learned “stereotypes, heuristics, and schemas” (Hasher & Zacks, 1979, p. 123; Yoon, May, & Hasher, 1998). A related issue is the hypothesis that cognitive abilities dedifferentiate with age, becoming more highly correlated (Cornelius, Willis, Nesselroade, & Baltes, 1983). Dedifferentiation is assumed to arise from the decline of a basic, fundamental mechanism, such as processing speed, whereas differentiation is assumed to arise from the development or breakdown of process-specific mechanisms. Rabbitt (1993; Rabbitt & Lowe, 2000) has suggested that aging leads to increasing individual differences, reflecting different rates and trajectories of change of underlying processes and/or neural structures. Despite these debates, there is widespread agreement that working memory is critical to a wide range of cognitive abilities. Performance on the reading and listening span tests of Daneman and Carpenter (1980) has been shown to be related to performance on reading and listening comprehension, learning to read, reading ability, arithmetic ability, and reasoning ability (Daneman & Blennerhassett, 1984; Daneman & Green, 1986; Daneman & Tardif, 1987; Hitch et al., 2001; Leather & Henry, 1994). Daneman and Merikle (1996) reviewed 77 studies involving 6,179 participants, confirming the link between reading/listening span measures and language comprehension, reporting correlations of 0.41 and 0.52 with global and specific tests of comprehension.

< previous page

page_278

next page >

< previous page

page_279

next page >

Page 279 Working Memory and Language Acquisition Gathercole and her colleagues (Gathercole & Baddeley, 1989; Gathercole, Willis, Emalie, & Baddeley, 1992) have linked vocabulary acquisition to phonological component of working memory. They assessed phonological memory in children using non-word repetition, reporting correlations between non-word repetition and receptive vocabulary of .52 to .56 for children between 4 and 6 years of age. They also reported that early non-word repetition scores were closely associated with later vocabulary scores, suggesting that phonological memory contributes to the rate at which children learn words. Further studies (Gathercole & Baddeley, 1990) also suggested that deficits in phonological memory may underlie language disorders in children as well as specific reading and sentence comprehension problems of poor readers (Crain & Shankweiler, 1988; Catts & Kamhi, 2005). Oakhill (1982; Oakhill, Yuille, & Parkin 1986) has suggested that children who have good oral reading skills but poor comprehension have a reduced working memory capacity, affecting the ability of the central executive to integrate information. Working Memory and Language Processing Just and his colleagues (Just & Carpenter, 1992; Just & Varma, 2002; King & Just, 1991; MacDonald, Just, & Carpenter, 1992) have claimed that working memory capacity constrains the interpretation of temporary syntactic ambiguities, limiting the ability of older or low span readers to make and sustain multiple interpretations of the ambiguous phrases. According to the Just and Carpenter (1992) capacity-constrained (CC) theory (see also the 3CAPS model of Just & Varma, 2002), older or low span readers should have difficulty processing the temporary syntactic ambiguities and should exhibit garden-path effects, initially misinterpreting reduced relative clause constructions as main verbs only to reinterpret the constructions once disambiguating information is encountered. Young or high span readers should be able to avoid garden-path effects by constructing multiple syntactic interpretations of the ambiguous phrases and retaining these interpretations until disambiguating information is encountered. This hypothesis has been carefully examined by Caplan and Waters (1999) who have considered a number of lines of evidence from studies of young and older adults as well as individuals with aphasia and dementia. They distinguish between immediate, interpretive syntactic processing and post-interpretative semantic and pragmatic processing. Caplan and Waters argue that there is little evidence to support the hypothesis that working memory limitations affect immediate syntactic processes; rather, they conclude that working memory limitations affect postinterpretative processes involved in retaining information in memory in order to recall it or use it, for

< previous page

page_279

next page >

< previous page

page_280

next page >

Page 280 example, to answer questions or match sentences against pictures. In a variety of studies comparing adults stratified into groups based on measures of working memory, Caplan and Waters (1999) note that effects of syntactic complexity do not differentially affect high versus low span readers or listeners. And they report that secondary tasks that impose additional processing demands on working memory do not differentially affect the processing of complex sentences. Caplan and Waters consider aphasic patients such as B. O. who had a digit span of only two or three digits but who was able to perform as well as normal healthy older adults on a wide range of tasks with complex sentences. They also note that patients with Alzheimer’s dementia, who also show severely limited working memory capacity, are able to make speeded acceptability judgments of complex sentences as accurately as non-demented controls. Waters and Caplan (1996a, 1996b, 1997, 2001) have directly examined the hypothesis that working memory limitations affect older adults’ ability to process complex sentences. These studies have used the auditory moving windows paradigm. This technique allows the listener to start and stop the presentation of a sentence and permits the analysis of phrase-by-phrase listening times, analogous to visual moving windows paradigms which permit the analysis of word-by-word or phrase-by-phrase reading times. The studies by Caplan and Waters typically examine the processing of subject- and object-relative clause constructions, such as those below: Object subject relative clause: The dancer found the musici,j that (tj) delighted the director. Subject object relative clause: The musici, j that the dancer found (ti) (tj) delighted the director. The object subject relative clause construction imposes few processing demands on the reader or the listener: the object of the main clause, (ti), is also the subject of the embedded relative clause, (tj). The subject object relative clause construction challenges the reader or listener to assign the correct syntactic relations, the subject of the main clause, (tj), must also be interpreted as the object of the embedded clause, (ti). Waters and Caplan (2001) compared how young and older readers allocate listening times to critical phrases of relative clause sentences. Despite differences in working memory, listening times were distributed similarly by young and older listeners. All paused longer when they heard the embedded verb in the complex object relative clause sentences than when they heard the corresponding verb in the simple subject relative clause version; this additional time is attributable to the extra processing required to recover its direct object. They found no evidence that differences in age or working memory lead to different processing strategies, supporting their theory.

< previous page

page_280

next page >

< previous page

page_281

next page >

Page 281 A recent study by Kemper, Crow, and Kemtes (2004) using eye-tracking methodology re-examined these issues. Eye-tracking is a more naturalistic task that imposes few restrictions on readers; they are free to skip words or phrases, read ahead and glance backwards, and re-read entire segments. Using this technology, Kemper et al. examined three aspects of reading: first fixations to key phrases, regressions to earlier phrases, and the total time key phrases were fixated. They examined reduced relative clause sentences such as those below: Reduced relative clause sentence: Several angry workers warned about the low wages decided to file complaints. Main clause sentence: Several angry workers warned about the low wages during the holiday season. Focused reduced relative clause sentence: Only angry workers warned about the low wages decided to file complaints. Kemper et al. (2004) found partial support for Waters and Caplan’s theory: young and older adults’ first pass fixations were alike and both groups showed a clear “garden-path” effect: a peak in fixation time at the second verb in reduced relative clause sentences but not at the verb in main clause sentences. This garden-path effect suggests that all readers initially interpret the first verb as the main verb and must reanalyze it when they encounter the second verb in the reduced relative clause sentence. However, Kemper et al. also observed an increase in regressions and in regression path fixations for older readers for the reduced relative clause sentences, suggesting that older adults were unable to correctly parse these sentences. Further, low span readers, identified by their scores on a battery of working memory tests, also produced more regressions and an increase in regression path fixations for reduced relative clause sentences, suggesting that they were unable to correctly parse the sentences. The results from the eye-tracking analysis of the focused reduced relative clauses sentences also posed problems for Caplan and Water’s theory: high span readers initially allocated additional processing time to the first noun phrase and then were able to avoid the “garden path” because the focus operator “only” led them to correctly interpret the first verb phrase as a reduced relative clause. Kemper and Liu (2007) also used eye-tracking to compare young and older adults’ processing of unambiguous object relative sentences and subject relative sentences, such as those below, which differed in the locus of embedding and the form of the embedded sentences. Young and older adults showed similar patterns of the first pass fixation times, regression path fixations, and leftward regressions to critical regions for both types of cleft sentences and for object subject relative clause sentences. However, older adults

< previous page

page_281

next page >

< previous page

page_282

next page >

Page 282 generally needed more time to process subject object relative clause sentences than young adults; they made more regressions back to both the main clause subject and the embedded clause subject than did young adults and, consequently, their regression path fixations for these critical regions were longer. These findings directly contradict Waters and Caplan’s hypothesis (2001) that working memory and sentence processing are unrelated. They also indicate that age-group differences, reflecting differences in working memory, arise for some, but not all, types of sentences. Whereas fixation patterns of young and older adults were similar for both cleft subject and cleft object sentences and object subject sentences, subject object sentences gave rise to marked age group differences in regressions and regression path fixations. Cleft subject and object subject sentences can be parsed as two sequential clauses: the main clause is followed by an embedded clause signaled by a “that” complementizer which is indexed to the preceding noun phrase. Cleft object sentences are somewhat more challenging to parse since the cleft object also serves as the object of the embedded clause and must be temporarily buffered while the embedded clause is processed. Subject object sentences impose yet greater demands for parsing since the subject of the main clause must also be assigned as the object of the embedded clause; further, the embedded clause interrupts the main clause, so that the main clause subject must be temporarily buffered if it is to be correctly linked with its verb. It may be that there is a threshold for parsing complexity such that differences due to age group—and, by inference, working memory span—are not apparent until this threshold is surpassed. What is apparent is that there are differences in the size of the temporary buffer required for syntactic analysis of subject object sentences, mirroring age differences in working memory as measured by traditional span measures. Compared to young adults, older adults, with smaller syntactic processing buffers, must make more regressions and allocate additional processing time to establish the main clause subject and relative clause subject of subject object sentences. Cleft subject: It was the tailor that altered the suit coat. Cleft object: It was the suit coat that the tailor altered. Object subject: The dancer found the music that delighted the director. Subject object: The music that the dancer found delighted the director. Conclusion Techniques to measure, manipulate, and fractionate working memory have proliferated in the past two decades, reflecting the central role of working memory in models of cognitive psychology. From a lifespan perspective, a

< previous page

page_282

next page >

< previous page

page_283

next page >

Page 283 number of key issues have been identified: Does working memory, and its component sub-systems including the phonological loop and central executive, differentiate or dedifferentiate across the lifespan? Does the growth of working memory capacity and executive functions in childhood foreshadow its decline in late adulthood? Do individual differences in working memory capacity and function map onto to individual differences in vocabulary acquisition, sentence processing, and skilled reading? These questions will continue to dominate discussions of working memory over the next decade. Acknowledgements Preparation of this chapter was supported in part by grants from the NIH to the University of Kansas through the Mental Retardation and Developmental Disabilities Research Center, grant number P30 HD-002528, and the Center for Biobehavioral Neurosciences in Communication Disorders, grant number P30 DC-005803 as well as by grants RO1 AG06319, K04 AG000443, P30 AG10182, RO1 AG09952, and RO1 AG025906 from the National Institute on Aging. Its contents are solely the responsibility of the author and do not necessarily represent the official views of the NIH. References Baddeley, A. (1986). Working memory. Oxford: Clarendon Press. Baddeley, A. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Science , 4 , 417–423. Baddeley, A., & Hitch, G. J. (1974). Working memory. In G. A. Bower (Ed.), The psychology of learning and motivation. Vol. 8 (pp. 47–89). New York: Academic Press. Baddeley, A., Bressi, S., Della Sala, S., Logie, R., & Spinnler, H. (1991). The decline of working memory in Alzheimer’s disease perseverative. Brain, 114 , 2521–2542. Baltes, P. B. (1994). Theoretical propositions of life-span developmental psychology on the dynamics between growth and decline (I. Ariyevich, Trans.). Psikhologicheskiy Zhurnal, 15, 60–80. Baltes, P. B., & Lindenberger, U. (1997). Emergence of a powerful connection between sensory and cognitive functions across the adult lifespan: A new window to the study of cognitive aging? Psychology and Aging , 12, 12–21. Benton, A. L., & Hamsher, K. D. (1989). Multilingual Aphasia Examination. Iowa City: AJA Associated. Caplan, D., & Waters, G. (1999). Verbal working memory and sentence comprehension. Behavioral and Brain Sciences, 22, 114–126. Carstensen, L. L., Mikels, J. A., & Mather, M. (2006). Aging and the intersection of cognition, motivation, and emotion. In J. Birren & K. W. Schaie, (Eds.), Handbook of the psychology of aging (6th ed., pp. 343–363). San Francisco, CA: Academic Press. Case, R., Kurland, D. M., & Goldberg, J. (1982). Operational efficiency and the growth of short term memory span. Journal of Experimental Child Psychology, 33, 386–404.

< previous page

page_283

next page >

< previous page

page_284

next page >

Page 284 Catts, H. W., & Kamhi, A. G. (Eds.). (2005). The connections between language and reading disabilities . Mahwah, NJ: Lawrence Erlbaum. Connelly, S. L., Hasher, L., & Zacks, R. T. (1991). Age and reading: The impact of distraction. Psychology and Aging , 6 , 533–541. Conway, A. R. A., Kane, M. J., Bunting, M. F., Hambrick, D. Z., Wilhelm, O., & Engle, R. W. (2005). Working memory span tasks: A methodological review and user’s guide. Psychological Bulletin and Review, 12, 769–786. Cornelius, S. W., Willis, S. L., Nesselroade, J. R., & Baltes, P. B. (1983). Convergence between attention variables and factors of psychometric intelligence in older adults. Intelligence, 7 , 253–269. Cowan, N. (1995). Attention and memory: An integrated framework. Oxford: Oxford University Press. Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24, 87–185. Crain, S., & Shankweiler, D. (1988). Syntactic complexity and reading acquisition. In A. Davison & G. Green (Eds.), Critical approaches to readability (pp. 167–192). Hillsdale, NJ: Lawrence Erlbaum. Daneman, M., & Blennerhassett, A. (1984). How to assess the listening comprehension skills of prereaders. Journal of Educational Psychology, 76, 1372–1381. Daneman, M., & Carpenter, P. A. (1980). Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Ability, 19, 450–466. Daneman, M., & Green, I. (1986). Individual differences in comprehending and producing words in context. Journal of Memory and Language, 25, 1–18. Daneman, M., & Tardif, T. (1987). Working memory and reading skill re-examined. In M. Coltheart (Ed.), Attention and performance XII: The psychology of reading (pp. 491–508). Hillsdale, NJ: Lawrence Erlbaum. Daneman, M., & Merikle, P. M. (1996). Working memory and language comprehension: A meta-analysis. Psychonomic Bulletin and Review, 3 , 422–433. Dempster, F. N. (1980). Memory span: Sources of individual and developmental differences. Psychological Bulletin, 19, 450–466. Dobbs, A. R., & Rule, B. G. (1989). Adult age differences in working memory. Psychology and Aging , 4 , 500–503. Engle, R. W., Tuholski, S. W., Laughlin, J. E., & Conway, A. R. A. (1999). Working memory, short-term memory, and general fluid intelligence: A latent-variable approach. Journal of Experimental Psychology: General , 128 , 309–331. Friedman, N. P., & Miyake, A. (2004). The reading span test and its predictive power for reading comprehension ability. Journal of Memory and Language, 51, 136–158. Gathercole, S. E. (1999). Cognitive approaches to the development of short-term memory. Trends in Cognitive Sciences, 3 , 310–419. Gathercole, S. E., & Baddeley, A. D. (1989). Evaluation of the role of phonological STM in the development of vocabulary in children: A longitudinal study. Journal of Memory and Language, 28, 200–213. Gathercole, S. E., & Baddeley, A. D. (1990). Phonological memory deficits in language disordered children: Is there a causal connection? Journal of Memory and Language, 29, 336–360. Gathercole, S. E., Willis, C., Emalie, H., & Baddeley, A. (1992). Phonological memory and vocabulary development during the early school years: A longitudinal study. Developmental Psychology, 28, 887–898.

< previous page

page_284

next page >

< previous page

page_285

next page >

Page 285 Hasher, L., & Zacks, R. T. (1979). Automatic and effortful processes in memory. Journal of Experimental Psychology: General, 108 , 356–388. Hitch, G. J., Towse, J. N., & Hutton, U. (2001). What limits children’s working memory span? Theoretical accounts and applications for scholastic development. Journal of Experimental Psychology: General , 130 , 184–198. Hume, G. E., Welsch, M. C., Retzlaff, P., & Cookson, N. (1997). Towers of Hanoi and London: Reliability and validity of two executive function tests. Assessment , 4 , 249–257. Jenkins, L., Myerson, J., Joerding, J. A., & Hale, S. (2000). Converging evidence that visuospatial cognition is more age-sensitive than verbal cognition. Psychology and Aging , 15, 157–175. Just, M. A., & Carpenter, P. A. (1992). A capacity theory of comprehension: Individual differences in working memory. Psychological Review, 99, 122–149. Just, M. A., & Varma, S. (2002). A hybrid architecture for working memory: A reply to MacDonald and Christiansen (2002). Psychology Review, 109 , 55–65. Kemper, S., & Liu, C. J. (2007). Eye movements of young and older adults during reading. Psychology and Aging , 22, 84–94 . Kemper, S., & McDowd, J. (2008). Dimensions of cognitive aging: Executive function and verbal fluency. In S. M. Hofer & D. F. Alwin (Eds.), Handbook of cognitive aging: Interdisciplinary perspectives (pp. 181–192). Thousand Oaks, CA: Sage. Kemper, S., Crow, A., & Kemtes, K. (2004). Eye fixation patterns of high and low span young and older adults: Down the garden path and back again. Psychology and Aging , 19, 157–170. Kimberg, D. Y., D’Esposito, M., & Farah, M. J. (2000). Frontal lobes II: Cognitive issues. In M. J. Farah & T. E. Feinberg (Eds.), Patient-based approaches to cognitive neuroscience (pp. 317–326). Cambridge, MA: MIT Press. King, J., & Just, M. A. (1991). Individual differences in syntactic processing: The role of working memory. Journal of Memory and Language, 30, 580–602. Leather, C. V., & Henry, L. A. (1994). Working memory span and phonological awareness tasks as predictors of early reading ability. Journal of Experimental Child Psychology, 58, 88–111. Lezak, M. D., Howieson, D. B., Loring, D. W., Hannay, H. J., & Fischer, J. S. (2004). Neuropsychological assessment (4th ed.). New York: Oxford University Press. Lindenberger, U., & Baltes, P. B. (1994). Sensory functioning and intelligence in old age: A strong connection. Psychology and Aging , 9 , 339–355. Logan, G. D. (1994). Spatial attention and the apprehension of spatial relations. Journal of Experimental Psychology: Human Perception and Performance, 20, 1015–1036. Logie, R., Zucco, G. M., & Baddeley, A. (1990). Interference with visual short term memory. Acta Psychologica , 75, 54–74. MacDonald, M., Just, M. A., & Carpenter, P. A. (1992). Working memory constraints on the processing of syntactic ambiguity. Cognitive Psychology, 24, 56–98. McElree, B. (2001). Working memory and focal attention. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27, 817–835. Mackworth, J. F. (1959). Paced memorizing in a continuous task. Journal of Experimental Psychology, 58, 206–211. Mikels, J. A., Larkin, G. R., Reuter-Lorenz, P. A., & Carstensen, L. L. (2005). Divergent trajectories in the aging mind: Changes in working memory for affective versus visual information with age. Psychology of Aging , 20, 542–553.

< previous page

page_285

next page >

< previous page

page_286

next page >

Page 286 Milner, B. (1971). Interhemispheric differences in the localization of psychological processes in man. British Medical Bulletin, 27, 272–277. Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., & Howerter, A. (2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. CognitivePsychology, 41, 49–100. Morris, N., & Jones, D. M. (1990). Memory updating in working memory: The role of the central executive. British Journal of Psychology, 81, 111–121. Morris, R. G., & Baddeley, A. D. (1988). Primary and working memory functioning in Alzheimer-type dementia. Journal of Clinical and Experimental Neuropsychology , 10, 279–296. Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive Psychology, 9 , 353–383. Oakhill, J. V. (1982). Constructive processes in skilled and less skilled comprehenders’ memory for sentences. British Journal of Psychology, 73, 13–20. Oakhill, J. V., Yuille, N., & Parkin, A. (1986). On the nature of the difference between skilled and less-skilled comprehenders. Journal of Research in Reading, 9 , 80–91. Oberauer, K., & Kliegl, R. (2006). A formal model of capacity limits in working memory. Journal of Memory and Language, 55, 601–626. Park, D. C., Smith, A. D., Lautenschlager, G., Earles, J. L., Frieske, D., Zwahr, M., and Gaines, C. (1996). Mediators of long-term memory performance across the life span. Psychology and Aging , 11, 621–637. Pickering, S. J. (2001). The development of visuo-spatial working memory. Memory , 9 , 423–432. Pickering, S. J., Gathercole, S. E., & Peaker, S. M. (1998). Verbal visuospatial short-term memory in children: Evidence for common and distinct mechanisms. Memory and Cognition. 26, 1117–1130. Rabbitt, P. (1993). Does it all go together when it goes? The Nineteenth Bartlett Memorial Lecture. Quarterly Journal of Experimental Psychology A: Human Experimental Psychology, 46A , 385–434. Rabbitt, P., & Lowe, C. (2000). Patterns of cognitive ageing. Psychological Research/Psychologische Forschung. 63, 308–316. Salthouse, T. A. (1994). How many causes are there of aging-related decrements in cognitive functioning? Developmental Review, 14, 413–437. Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review, 3 , 403–428. Salthouse, T. A., & Miles, J. D. (2002). Aging and time-sharing aspects of executive control. Memory and Cognition, 30, 572–582. Salthouse, T. A., Atkinson, T. M., & Berish, D. E. (2003). Executive functioning as a potential mediator of age-related cognitive decline in normal adults. Journal of Experimental Psychology: General , 132 , 566–594. Salthouse, T. A., Kausler, D. H., & Saults, J. S. (1990). Age, self-assessed health status, and cognition. Journals of Gerontology, 45(4), 156. Spreen, O., & Strauss, E. (1991). A compendium of neuropsychological tests: Administration, norms, and commentary . New York: Oxford University Press. Spreen, O., & Strauss, E. (1998). A compendium of neuropsychological tests (2nd ed.). New York: Oxford. Turner, M. L., & Engle, R. W. (1989). Is working memory capacity task dependent? Journal of Memory and Language, 28, 127–154.

< previous page

page_286

next page >

< previous page

page_287

next page >

Page 287 Waters, G. S., & Caplan, D. (1996a). The capacity theory of sentence comprehension: Critique of Just and Carpenter (1992). Psychological Review, 103 , 761–772. Waters, G. S., & Caplan, D. (1996b). Processing resource capacity and the comprehension of garden path sentences. Memory and Cognition, 24, 342–355. Waters, G. S., & Caplan, D. (1997). Working memory and on-line sentence comprehension in patients with Alzheimer’s disease. Journal of Psycholinguistic Research , 26, 337–400. Waters, G. S., & Caplan, D. (2001). Age, working memory, and on-line syntactic processing in sentence comprehension. Psychology and Aging , 16, 128–144. Wechsler, D. (1958). The measurement and appraisal of adult intelligence . Baltimore: Williams & Wilkins. Yoon, C., May, C. P., & Hasher, L. (1998). Aging, circadian arousal patterns, and cognition . Philadelphia, PA: Psychology Press.

< previous page

page_287

next page >

< previous page

page_288

next page >

Page 288 12 THE DEVELOPMENT OF NEURAL SUBSTRATES OF LANGUAGE OVER THE LIFESPAN Arturo Hernandez, Merrill Hiscock, and Elizabeth A. Bates For many years, researchers have been fascinated by cases of aphasia in which brain damage leads to the loss of one or many language abilities. Broca initially discussed the importance of the third frontal convolution in articulated speech. Wernicke followed by discussing the importance of the Superior Temporal Gyrus in processing receptive language. This model was further refined later as the Wernicke–Lichtheim model which has served as the basis of aphasia research from the early 20th century until the present (for a review see Cepanec & Judas, 2007; Graves, 1997; Head, 1926; Weems & Reggia, 2006). In brief, this model posited that language was in its essence a sensorimotor function. In this view, syndromes of aphasia could be described as consisting of damage to these sensory or motor centers or to the pathways that connected them. This model of language processing has come to form the default model for all research on the neural bases of language in adults and across development. In the current piece, we will begin by reviewing the current views on the neural substrates of language before proceeding to uncover the factors that modulate neural activity across the lifespan. Neural Substrates of Language in Neurologically Uncompromised Young Adults With the advent of current neuroimaging techniques researchers have begun to uncover the brain areas that are involved in language processing in normal uninjured adults (Bookheimer, 2002; Friederici, 2002; Friederici, Ruschemeyer, Hahne, & Fiebach, 2003; Hickok & Poeppel, 2004; Ni et al., 2000; Vigliocco, 2000). These studies for the most part have identified a large left frontotemporal network which is involved in language. While these results for the most part confirm the Wernicke–Lichtheim model of

< previous page

page_288

next page >

< previous page

page_289

next page >

Page 289

Figure 12.1 Neutral areas involved in language processing in neurologically intact adults. Note Brain areas commonly activated in neuroimaging studies of language processing: 1. BA 45 and 47 involved in semantic processing; 2. Auterior insula involved in anticulation; 3. BA 44 and 45 involved in grammar processing; 4. Superior portions of BA 44 involved in phonological retrieval; 5. Middle portion of BA 22 involved in phonetic processing; 6. Posterior portions of BA 22 which serve as an interface between acoustic and phonetic codes. The borders of these areas should not be taken as absolute. The anterior insula (2) is represented on the surface for display convenience. neurolinguistics, neuroimaging studies have identified the existence of areas beyond those based on aphasia syndromes (see Figure 12.1). In order to consistently identify brain areas, Brodmann established a classification system based on the physical appearance of neurons in a particular area. These areas are commonly referred to as Brodmann areas or BA for short. In the frontal lobe, researchers have uncovered a region which encompasses inferior portions of the inferior frontal gyrus (IFG), extending from inferior BA 45 into BA 47. This area has been found to be active in tasks which require semantic retrieval (Poldrack et al., 1999; Wagner, Desmond, Demb, Glover, & Gabrieli, 1997; Wagner, Maril, Bjork, & Schacter, 2001; Wagner, Pare-Blagoev, Clark, & Poldrack, 2001), processing of semantics at the sentence level (Dapretto, Bookheimer, & Mazziotta, 1999) and processing of metaphors (Rapp, Leube, Erb, Grodd, & Kircher, 2004). A second more superior portion of IFG which encompasses inferior and middle BA 44 and portions of BA 45 has been found to be involved in grammatical processing and the detection of syntactic anomalies. Researchers have found

< previous page

page_289

next page >

< previous page

page_290

next page >

Page 290 activity in this area when grammar is processed at the sentence level (Dapretto et al., 1999) and phrasal level (Kang, Constable, Gore, & Avrutin, 1999). Furthermore, researchers have found that the middle portion of BA 44 and 45 is active when people learn grammatical rules (Tettamanti et al., 2002). Finally, increased activity in the same portion of IFG has been found when comparing grammatical gender decisions to semantic processing of nouns (Miceli et al., 2002) and when comparing irregular to regular gender marked nouns (Hernandez et al., 2004). Taken together these results highlight the importance of the middle portion of BA 44/45 for grammatical processing. Finally, moving to the most superior point of BA 44 and extending posteriorly to BA 6, there is a posterior portion of IFG which shows increased activity for phonological retrieval tasks (Poldrack et al., 1999; Zatorre, Meyer, Gjedde, & Evans, 1996). In Broca’s original view, the inferior and middle portions of BA 44/45 were considered to be crucial for articulation (Broca, 1865, 1988 [1861]). In the past 15 years, however, new research has suggested that the anterior insula plays a more prominent role in articulation than was previously thought (Ackermann & Riecker, 2004; Bates et al., 2003; Dronkers, 1996; Kuriki, Mori, & Hirata, 1999; Wise, Greene, Buchel, & Scott, 1999). Evidence of the importance of the insula in articulation comes from studies which have used a lesion overlap method in which structural brain scans of different patients with articulatory speech problems are superimposed in order to determine areas of maximal lesion overlap (Bates et al., 2003; Dronkers, 1996). The findings from patients have been confirmed by studies of articulatory processing using Magnetoencephalography (Kuriki et al., 1999), Positron Emission Tomography (Wise et al., 1999) and functional Magnetic Resonance Imaging (Ackermann & Riecker, 2004; Hillis et al., 2004). Each method of neuroimaging captures slightly different aspects of the neural signal. Magnetoencephalography (MEG) captures the magnetic field that arises when a cluster of neurons fires in synchrony. Positron Emission Tomography (PET) and Magnetic Resonance Imaging (MRI) are methods which are sensitive to the hemodynamic properties of neuronal firing. Specifically, areas of the brain which show more activity draw in more blood. In the case of PET, participants are injected with a glucose solution which contains a mild radioactive tracer. Since brain areas that are more active require more glucose, neural activity can be tracked by the amount of decay in the tracer. In the case of MRI, blood oxygenation levels are related to blood flow. The change in deoxyhemoglobin can be detected by specialized MRI scanners. An important difference is the relative ability of these three methods to detect the localization and timing of neuronal activity. MEG is deemed to be better at detecting when brain activity occurs (up to 1 millisecond) but cannot necessarily detect where the brain activity is located. PET and MRI are better able to detect the location of brain activity (up to popu-

< previous page

page_290

next page >

< previous page

page_291

next page >

Page 291 lations of several hundred neurons) but are only able to resolve neural activity that differs by a few seconds. Much like in the classic Wernicke–Lichtheim model of the neural bases of language, more recent research has confirmed the importance of the temporal lobe in language processing. This includes regions in the middle portion of BA 22 (Superior Temporal Gyrus; STG) which have been posited to be involved in acoustic/phonetic processing and possibly syntactic processing (Friederici et al., 2003). A more posterior region in the STG (BA 22) has been posited to serve as an interface between acoustic and phonetic codes (Hickok & Poeppel, 2004; but see Stowe, Haverkort, & Zwarts, 2005 for a different view on the role of specific regions in language use). The main question of the current piece is how these neural substrates of language change across the lifespan. First, how does language processing in children lead to the configuration observed in the adult model presented in Figure 12.1? Second, how does this system change as adults mature past middle age and into older adulthood? To address these issues it will be important to consider research from both the neuroimaging and the neuropsychology literatures that inform a lifespan developmental cognitive neuroscience. Finally, it is important to note that language processing involves variation in activity on both an anterior–posterior and right–left axis. The latter has played a crucial role in thinking about early development. Hence, we will begin with discussion of the possible changes in right–left asymmetries in early development before proceeding to discuss other ways in which development affects the network of areas involved in language processing. Developmental Changes in Brain Anatomy Asymmetries in the Immature Brain Several anatomical asymmetries in the neonatal brain have been found (LeMay & Culebras, 1972; Teszner, Tzavaras, Gruner, & Hécaen, 1972; Wada, Clark, & Hamm, 1975; Witelson, 1983). These findings contradict the previously held belief that the two cerebral hemispheres in infancy are morphologically identical or nearly so (for an historical review see Dennis & Whitaker, 1977). Asymmetries in some cerebral regions can be attributed to more widely spaced cortical columns on the larger side (Buxhoeveden & Casanova, 2000), and it remains to be established that cortical volume is an index of any functional characteristic that differs between hemispheres. One can only conclude that some of the anatomical asymmetries that characterize the adult human brain are neither absent nor less marked in the infant brain. A similar conclusion applies to the anticlockwise torque that is manifested as a wider frontal lobe on the right side and a greater anterior extension of the right frontal lobe, coupled with a wider posterior left hemisphere and a

< previous page

page_291

next page >

< previous page

page_292

next page >

Page 292 greater posterior extension of the left hemisphere. Irrespective of its functional significance, if any, in the adult human brain, a similar torque is observed in the immature human brain as well as in the ape brain (LeMay & Culebras, 1972). Maturational Gradients Perhaps anatomical disparities between left and right hemispheres reflect differences in the rate of growth. Corballis and Morgan (1978) compiled evidence from diverse sources to support their assertion that the left cerebral hemisphere develops more rapidly than the right hemisphere. Although Corballis and Morgan’s arguments stimulated interest in the concept of a lateral gradient of maturation, which occurs during development and is very similar if not identical across individuals, the putative left-to-right direction of the gradient was challenged by subsequent evidence that the right cerebral cortex in fact develops more rapidly than the left (Dooling, Chi, & Gilles, 1983). Other accounts of brain maturation (e.g. Best, 1988) have been based on the premise that the right hemisphere matures more quickly than the left hemisphere. A difference between the maturation rates of the two hemispheres is but one aspect of brain development. Found within the cerebral hemispheres are regional patterns in the development of myelination, in neuronal, dendritic, and axonal density, and in the width of the different cortical layers (Campbell and Whitaker, 1986). In an attempt to consolidate all available data on the rate at which different brain regions mature, Best (1988) proposed a growth vector representing the resultant of four developmental gradients: right-to-left, anterior-to-posterior; primary-tosecondary-to-tertiary,1 and basal-to-cortical (from the middle of the brain out to the cortex). More recent studies using a variety of techniques have confirmed the importance of the three gradients or axes on which brain development occurs: right-to-left, anterior-to-posterior, primary-to-secondary-to-tertiary, and basal-tocortical. In young infants, sensory cortices develop earliest in life. The development of sensory cortex is followed by development of motor cortex and sensory bridgeways in the parietal and frontal lobe. The most anterior regions of the brain in the prefrontal cortex are known to develop the latest. Although less dramatic than the brain changes seen during infancy and early childhood, changes in brain structure have been found to continue from middle childhood, across adolescence and adulthood and through older adulthood. These findings have been confirmed by studies using direct observation of brain tissue as well as those using indirect methods of measuring the brain such as Magnetic Resonance Imaging (MRI) and functional Magnetic Resonance Imaging (fMRI). Early studies relied to a great extent on histological studies in which the brains of a few individuals were examined (Huttenlocher, 1990, 1994;

< previous page

page_292

next page >

< previous page

page_293

next page >

Page 293 Huttenlocher & Dabholkar, 1997; Huttenlocher & de Courten, 1987; Huttenlocher, de Courten, Garey, & Van der Loos, 1982a, 1982b). These studies have found that some areas of the brain develop earlier than others. This differential type of development is reflected in three distinct neurodevelopmental processes, including the production of interconnections between neurons, which are called synapses, as well as the reduction in these interconnections, which is called synaptic pruning. The third process involves myelination, the addition of a fatty sheet surrounding the axon of each nerve cell. Production of synapses is strongest in the occipital lobe (i.e. visual regions) of the cortex between 4 and 8 months of age (Huttenlocher & de Courten, 1987). In the frontal lobe, synapse production reaches its peak at 1 year 3 months of age (Huttenlocher & Dabholkar, 1997). The lag in the frontal lobes’ overproduction of synapses is also observed in the reduction of synapses via pruning (Huttenlocher, 1994). The advent of newer neuroimaging techniques has allowed researchers to look at brain development in vivo. Of particular use is the family of MRI, which includes both structural imaging of grey matter (called structural MRI), white matter (called Diffusion Tensor Imaging or DTI) as well as functional imaging (fMRI). Research using newer neuroimaging techniques has found evidence that developmental changes in brain areas can be seen as involving both a left–right, anterior–posterior gradient and primary-to-secondary-to-tertiary gradient. In a similar vein, recent neuroimaging studies of single-word processing with children have shown changes in the magnitude of neural activity in adults and children centered in areas of the frontal lobe (Schlaggar et al., 2002), the area that has been found to reach adult levels of synaptic connectivity the latest. These studies suggest that adults and children will show differences in brain activity in areas of the frontal lobe. In a groundbreaking study, Sowell et al. (2003), examined the grey matter density of high resolution structural MRI scans in a group of individuals who ranged between 7 and 87 years of age. Results revealed a number of changes in grey matter density, the part of the cortex composed of neural cell bodies which carry out the neural computations. Across age most areas showed a linear decrease in grey matter density. The authors speculate that changes in grey matter density may be due to increased myelination. For example, regions in the frontal lobe showed decreases in grey matter density from ages 7 and beyond. The authors attribute some of this to an increase in myelination up until age 40, with a decrease in neural density ensuing into older adulthood. One region which showed a different pattern of development was the left temporal regions (in particular the posterior portions of the middle and superior temporal gyrus) which revealed an increase in grey matter density until age 40. These studies converge with previous work which has found that left temporal cortices continue to show changes in grey matter density well into adulthood. This is most likely due to the involvement of

< previous page

page_293

next page >

< previous page

page_294

next page >

Page 294 this region in language processing. The late maturation of the temporal lobe and particularly in the superior and posterior parts is an example of primary-to-secondary-to-tertiary development. In this particular case, it would involve extensions of auditory processing into language processing. The results from studies with a number of methodologies including neuroimaging and histological studies are consistent with development which proceeds from anterior-to-posterior, left-to-right as well as from primary-tosecondary-to-tertiary across development. Future studies should seek to further explore changes across these dimensions as well as extend into exploring whether there are differences in the basal-to-cortical dimension. Clinical Evidence The literature on childhood aphasia, hemispherectomy, and recovery of function has been reviewed by a number of authors (Aram, 1988; Aram, Eisele, Rapin, & Segalowitz, 1992; Hiscock & Kinsbourne, 1994; Satz, Strauss, & Whitaker, 1990; Spreen, Risser, & Edgell, 1995; Woods & Teuber, 1978). Even though the hypothesis of progressive lateralization (Lenneberg, 1967) has drawn much of its support from cases of aphasia in children following right hemispheric lesions, the preponderance of evidence now suggests that childhood aphasia consequent to unilateral right hemisphere damage is as infrequent in children as in adults. Studies using quantitative measures of cognitive functioning following unilateral lesions are limited by a reliance on IQ and academic achievement tests, which constitute neither sensitive nor comprehensive measures of linguistic and visuo-perceptual functioning. Such measures are not optimal for differentiating between left- and right-sided lesions. Despite this handicap, several studies do show associations between left-sided damage and verbal impairments, and between right-sided damage and non-verbal impairments. Some studies suggest that the selectivity of unilateral lesion effects is less marked when the damage is prenatal or perinatal than when it occurs later in development (cf. Woods, 1980). Nonetheless, there is ample evidence that unilateral brain lesions of prenatal or perinatal origin can lead to differential impairment of verbal and non-verbal functions, especially when those functions are assessed using methods other than IQ tests (Aram et al., 1992). The implications of hemispherectomy performed at different ages are difficult to specify, mainly because of extraneous factors that confound comparisons between hemispherectomy in children and hemispherectomy in adults (St James-Roberts, 1981) but also because of the inferential limitations of small-sample and single-case studies (Bishop, 1983). If any conclusion is justified by the available hemispherectomy evidence, it is that the outcome of hemispherectomy performed at different ages—which traditionally has been used to support the concept of left and right hemisphere equipotentiality—

< previous page

page_294

next page >

< previous page

page_295

next page >

Page 295 now seems to suggest the opposite conclusion, viz., that specialization of the cerebral hemispheres is manifest early in life (Spreen et al., 1995). Evidence from Children with Focal Lesions More recent studies which have looked at focal lesion children, however, have qualified the specialization view to a certain extent (Stiles et al., 2002). These studies, conducted under the auspices of the San Diego Program in Cognitive and Neural Development, focused on infants with unilateral injuries sustained before 6 months of age. There has been a great deal of controversy about this population over the last century. Early studies were consistent with the equipotentiality hypothesis by revealing that early unilateral injuries have no effect at all on long-term language outcomes (Basser, 1962; Lenneberg, 1967). Subsequent studies conducted in the 1970s and 1980s uncovered subtle deficits on language measures in left hemisphere-damaged children (e.g. Aram, 1988; Dennis & Whitaker, 1977; Riva & Cazzaniga, 1986; Riva, Cazzaniga, Pantaleoni, Milani, & Fedrizzi, 1986). Based on these findings it was concluded that the left hemisphere is irreversibly specialized for language at birth. However, none of these later studies actually conducted a direct statistical comparison of children with left vs. right hemisphere injury (for a critical review see Bishop, 1994). Furthermore, these studies did not uncover anything resembling a childhood form of aphasia following early unilateral injury. More recent reviews of the literature suggest that aspects of equipotentiality and irreversible determinism are consistent with the findings (Bates, Vicari, & Trauner, 1999; Eisele and Aram, 1995; Stiles, Bates, Thal, Trauner, & Reilly, 1998; Vargha-Khadem, Isaacs, & Muter, 1994). First, it is now widely agreed that early unilateral injury does not lead to clinically significant language disorders in the vast majority of cases, if children with extraneous complications are excluded from the sample. That is, this population does not show the types of aphasia that are commonly observed in the adult literature. However, children with a history of brain injury tend to perform below neurologically intact age-matched controls on a host of linguistic and nonlinguistic measures. Their performance is usually in the normal to low-normal range, corresponding to an average drop in verbal and non-verbal IQ of about five–seven points. Second, when left and right hemisphere-damaged children are compared directly (with sample sizes large enough and sufficiently well matched to permit a statistical test), and measured after 5–6 years of age (when language acquisition is virtually complete), there is little evidence for a difference in long-term language outcomes as a function of lesion side (left vs. right), lesion site (e.g. anterior vs. posterior) or even lesion size (Stiles et al., 2002). In contrast with the results of retrospective studies, prospective studies of development prior to 4–5 years of age in this population demonstrate moderate to severe delays in all the early language milestones. These include delays in the onset

< previous page

page_295

next page >

< previous page

page_296

next page >

Page 296 of babbling and preverbal communication (Marchman, Miller, & Bates, 1991), and delays between 1 and 5 years of age in lexical development and grammar (Thal, Marchman, Stiles, & Aram, 1991; Vicari, Caselli, & Tonucci, 2000). Most importantly, there were more dissociations than would be expected by chance in this period of development, and there were significant correlations between these dissociations and specific lesion sites. Comprehension vs. Production Based on the adult aphasia literature, we might expect a profile of delayed production with normal comprehension to occur more often in children with left anterior involvement (by analogy to adult Broca’s aphasia). Conversely (by analogy to adult Wernicke’s aphasia), children with left posterior damage should display a profile in which comprehension vocabularies fall below the levels that are normally observed in children at the same level of production. This issue has been investigated by Thal et al. (1991) and more recently by Bates et al. (Bates, 1997). Results of both studies were quite surprising. The development of word comprehension is not selectively affected by lesions to left posterior cortex. Rather, the Wernicke-like profile was actually more common in children with right hemisphere damage, a finding which would not be predicted by findings in the adult aphasia literature. This finding is compatible with electrophysiological studies of normally developing children (Mills, Coffey-Corina, & Neville, 1993, 1997), which show that the difference in the brain’s response to familiar vs. unfamiliar words is bilateral (but somewhat larger on the right) prior to approximately 18 months of age. After that point (and strongly correlated with the “vocabulary burst”), there seems to be a reorganization in the brain’s response to familiar words, with a larger effect observed in the left hemisphere, primarily across frontal and temporal sites. To explain these findings, Bates et al. (1997) suggest that the right hemisphere plays a larger role in the first stages of word comprehension because that hemisphere appears to be particularly important for integration of information across multiple sources (Stiles et al., 1998). For adults who already know their language (and also for older infants), this kind of multi-modal integration may not be necessary in order to understand a familiar word. But for infants who are struggling to “crack the code,” right hemisphere resources may play a particularly important role. Analytic vs. Holistic Style Children who are acquiring English have a tendency to deal initially with pronouns by leaving them out entirely (analytic/referential style) or by producing them during early language acquisition, but only in a less productive but rote manner. When these definitions are applied to a focal lesion sample, Thal et al. (1991) reported a significantly higher incidence of

< previous page

page_296

next page >

< previous page

page_297

next page >

Page 297 holistic/pronominal types of children across a whole sample than would be expected if the sample were drawn randomly from the normal population (based on norms from Fenson et al., 1993). This finding is reminiscent of a report by Johnston and Kamhi (1984), showing that language-impaired children tend to increase the length of their utterances by extensive use of a handful of grammatical function words. There were, however, several cases of extreme referential style in the focal lesion data as well. These contrasting extremes provide us with an opportunity to examine two different hypotheses that have been offered to explain the analytic/holistic dissociations observed in normal children. The interhemispheric hypothesis is based on the claim that the left hemisphere is specialized for fine-grained analytic operations, while the right hemisphere is specialized for holistic/configurational operations (Bradshaw & Nettleson, 1981; Ivry & Robertson, 1998). By this argument, an analytic/segmenting approach to language learning should be less accessible if the left hemisphere is damaged, whereas a holistic/configurational approach should be less available if the right hemisphere is damaged. Although this prediction is reasonable, it is not borne out by studies of infants with focal brain injury. In fact, Thal et al. (1991) report a significantly higher incidence of pronominal/expressive style in children with right hemisphere damage, with proportionally more referential/telegraphic speech in children with left hemisphere damage. This finding suggests that left hemisphere processes may play an important role in the early production of pronouns and other function words—even when those words are used in rote or “formulaic” expressions. In short, although the right hemisphere account of holistic style in normal children seems reasonable, it is not supported by the data. Taken together the results from these studies suggest that language processing in children undergoes considerable reorganization during early development. This reorganization involves possible shifting of reliance on right and left hemisphere at different points in development. This pattern does not hold up for other cognitive functions such as spatial processing which seems to be more clearly related to specific neural substrates at a much earlier point in development. Finally, future studies should seek to test further subspecializations of the language process by looking at whether certain types of language processing (i.e. phonological processing) may be more directly tied to certain neural substrates during development. Changes Associated with Aging Normal Aging Physiological Changes Autopsy studies reveal no obvious predominance of neuropathology in either the left or the right hemisphere of the elderly (Ivry & Robertson, 1998).

< previous page

page_297

next page >

< previous page

page_298

next page >

Page 298 Studies based on positron emission tomography (PET scanning) report nearly identical levels of glucose metabolism in the left and right cerebral cortices and in the left and right basal ganglia of healthy elderly subjects (Duara et al., 1984). A study of 100 post-mortem brains revealed comparable age-related decreases in cerebral volume within the left and right hemispheres (Witelson, Beresh, & Kigar, 2006). Behavioral Changes Various psychometric data, typically cross-sectional, establish that certain kinds of tasks are more prone than others to show age-related declines in average performance. For instance, scores on the Performance subtests of the Wechsler Adult Intelligence Scale decline more rapidly with increasing age than do scores on the Verbal subtests. Much of this differential decline can be attributed to a cohort effect that affects non-verbal tests more strongly than verbal tests (Flynn, 2006). Nevertheless, a differential decrement of Performance and Verbal scores has been confirmed by longitudinal data, which may not be as susceptible to cohort effects (Mortensen & Kleven, 1993; but see Schrauf, this volume, on design problems). Similarly, tests of “fluid” intelligence—the ability to solve problems requiring novel information and strategies—are more likely to show age-related changes than are tests of “crystallized” intelligence—the ability to utilize previously acquired knowledge and strategies (Horn & Cattell, 1967). Age-related decline in average performance on some tests is quite marked. On Raven’s Standard Progressive Matrices, an untimed test that contains seemingly novel problems, a raw score at the 50th percentile for 18-yearolds falls at the 95th percentile for 65-year-olds (Lezak, 1995). Much of that difference, however, can be attributed to a cohort effect (Flynn, 1998). Psychometric data, including results from neuropsychological testing, sometimes have been construed as evidence that right hemisphere functions are more vulnerable than left hemisphere functions to deterioration with advancing age. This conclusion, however, is not supported with any consistency by findings from laterality tests. Although two dichotic listening studies (Clark & Knowles, 1973; Johnson et al., 1979) yielded larger right-ear advantages (REAs) in elderly people than in younger subjects—an outcome that is consistent with deterioration of the right hemisphere— the results are obfuscated by the use of multiple pairs of stimuli per trial. Given that subjects tend to report right-ear signals prior to left-ear signals, a larger REA may reflect nothing more than diminished overall performance, which in turn may reflect diminished working memory in the elderly. Two of the most salient cognitive deficits observed among the elderly are reduced speed of processing (Van Gorp, Satz, & Mitrushina, 1990) and reduced ability to perform complex or difficult tasks (Crossley & Hiscock, 1992). It is uncertain whether these deficits represent separate limitations—

< previous page

page_298

next page >

< previous page

page_299

next page >

Page 299 e.g. slowness of processing versus insufficiency of processing resources—or different manifestations of the same limitation (Salthouse, 1988). Nonetheless, since reaction time (RT) and dual-task paradigms are known to yield consistent age effects, these two paradigms are appropriate sources of evidence regarding lateralized impairment in the elderly. Should future use of these paradigms yield evidence of age-related changes in laterality, care must be taken to ascertain that the findings reflect differential deterioration of the left and right hemispheres rather than changes in the functioning of the corpus callosum (Hoptman, Davidson, Gudmundsson, Schreiber, & Ershler, 1996). Neuroimaging Evidence Recent work using in vivo structural MRI (Raz et al., 1997; Raz, Rodrigue, Head, Kennedy, & Acker, 2004) have shown that not all areas show reduced volumes with age at the same rate. Specifically, the volume of the lateral prefrontal cortex and hippocampus (to take two examples) have strong negative correlations with age whereas the inferior parietal cortex and primary visual cortex show a reduced negative correlation with age. Hence, there do appear to be differential changes in brain volume with age. This suggests that aging impacts some cognitive operations to a greater extent than others. Specifically, working and long-term memory (lateral prefrontal cortex and hippocampus) are affected to a greater extent, whereas sensory processing (visual cortex and parietal cortex) is affected to a lesser extent. This is consistent with findings that purely perceptual processing is slowed to a lesser extent in older adults than tasks which involve inhibitory processing. Although the differential cognitive decline and its concomitant neural substrates effects are not directly related to changes in language per se, they do carry implications for the neural substrates involved in language in older adults. The first implication is that older adults should show a relatively small decline in speed for tasks that require simpler cognitive and to some extent purely perceptual processing. Hence, there should be preservation of function in some of these simpler tasks. This to some extent might explain the relative preservation of function in older adults for some functions such as vocabulary tasks. Alternatively, it is possible that older adults show preservation of function but at a cognitive price. In two separate studies, Grossman and colleagues (Grossman et al., 2002a, 2002b) used fMRI to look at the differences in neural activity between young and older adult individuals processing sentences. Participants were shown sentences and asked to decide who was performing an action in a sentence. Results from these studies revealed overall group differences in performance between older and young adults. Interestingly, there was a great deal of variability in the older adult group which allowed Grossman et al. to identify a subgroup of high comprehenders and low comprehenders. Hence, Grossman

< previous page

page_299

next page >

< previous page

page_300

next page >

Page 300 et al. were able to compare young adults to older adults when both groups were matched on performance. The results revealed much stronger activity in the left posterior-temporoparietal region for young adults compared to high comprehending older adults. The latter revealed more activity in the posterior portion of the inferior frontal cortex and the right posterior superior temporal gyrus bordering on the inferior parietal lobule. These results can be interpreted using Figure 12.1. Specifically, it suggests that younger adults engage in more acoustic-to-phonetic processing whereas older adults show activity in areas which are involved in phonological retrieval (posterior IFG) and a homologous area of the right posterior superior temporal gyrus which is normally engaged by younger adults during complex phonological processing. Hence, older adults must recruit areas of the right hemisphere in order to accomplish the same goal. These results suggest that older adults engage additional brain areas when processing language. Finally, it is important to note that young adults do engage some of these same areas when processing more difficult sentences. This suggests that changes in the neural substrates for language involve a compensatory mode of processing rather than a “rewiring” of the circuit for language. Taken together, studies of aging in both the behavioral and neural domain suggest that older adults may be engaging additional cognitive or neural resources during processing of normal sentences. It remains to be determined how differences in general processing such as cognitive slowing may translate into the differential laterality or other compensatory processes such as higher use of working memory. By bridging work using both cognitive and neuroscience methods it may be possible to arrive at a more complete understanding of the changes that occur in normal aging. Clinical Evidence Aphasia The relative incidence of different aphasia types appears to shift as a function of age (Brown & Grober, 1983; Brown & Jaffe, 1975; Miceli et al., 1981; Obler, Albert, Goodglass, & Benson, 1978). Brown and Grober (1983) found that expressive disorders and mixed aphasias (i.e. non-fluent speech with moderately impaired comprehension) predominate throughout the life span, but the preponderance of these non-fluent aphasias is most striking during the first three decades of life. The incidence of global aphasia begins to rise in the fourth decade, and sensory (fluent) aphasias become relatively common during the seventh decade. Brown and Grober attributed these agerelated shifts to a gradually increasing degree of lateralization and focal brain organization: An initially bilateral and diffuse language substrate becomes unilateral and focal over time, with expressive functions lateralizing before receptive functions. It is not clear, however, that the changing inci-

< previous page

page_300

next page >

< previous page

page_301

next page >

Page 301 dence rates reflect shifts in lateralization rather than regional changes in vulnerability to cerebral disease, changes in compensatory ability (i.e. in plasticity), or other variables. Future studies are needed to further disentangle how these different potential factors may or may not be contributing to laterality effects. General Conclusions Regarding Aging Aging beyond early adulthood results in changes in the neural substrates involved in language. These changes do seem to involve both anterior–posterior and left–right axes. Specifically, older adults recruit more right hemisphere and frontal areas when conducting complex language processing (Grossman et al., 2002a, 2002b). Patient groups with language impairments also show differential impairment with age. Specifically, aphasia type varies as a function of age in a manner that is consistent with the view that language functions become even more focal between early and older adulthood. Interpretive Complexities General Problems in Studying Lifespan Development Developmental Patterns Must Be Constructed The ideal study of change across the lifespan would begin with the individual’s conception or birth, and then continue through successive stages of life. As a rule, the scientific examination of ontogenetic change in humans does not conform to this ideal. Instead, the inquiry typically begins with the mature organism. As interesting characteristics are observed and investigated in the adult, certain development questions invariably arise. At what stage of development does this attribute become manifest? What are the mechanisms responsible for the emergence the characteristic? To what extent is the attribute influenced by exogenous or endogenous environmental factors? What evolutionary advantage, if any, is bestowed by the characteristic? How does the attribute change as the organism approaches the end of its life trajectory? Studies of young or middle-aged adults are then extended downward to infants and children, and upward to the elderly. Lifespan Studies are Constrained by Methodological Limitations Unfortunately, the methods used with young or middle-aged adults often are not feasible for use with children or elderly people. Visual (tachistoscopic) experiments in which letters or words are presented to one visual half-field cannot be used with infants or young children because they are unable to read. Most elderly people can read, of course, but their ability to perform a

< previous page

page_301

next page >

< previous page

page_302

next page >

Page 302 visual task may be compromised by impairments of visual acuity. Dichotic listening is also problematic. Infant researchers have used habituation–dishabituation paradigms to study asymmetries of auditory perception in infants (e.g. Best, 1988; Entus, 1977), and some of the infant studies have shown a right-ear advantage for language sounds. Infant studies of this kind are difficult to do, however, and results will vary with the physiological and psychological state of the infant at the time of testing. It is especially difficult to test 1- and 2-year-olds, for whom infant methods are inappropriate and methods designed for older children may be unsatisfactory. Dichotic studies have been done successfully with 3-, 4-, and 5-year-olds (e.g. Hiscock & Kinsbourne, 1977), but even children in this age range do not respond well to the standard dichotic listening tests that are used with school-age children and adults. The elderly are difficult to assess because hearing impairment is common in older adults. Functional MRI is difficult to use with young children, but other functional imaging techniques (magnetic source imaging, in particular) seem to be more feasible. General Conclusions The neural substrates underlying language go through profound changes across the lifespan. These changes occur across a number of gradients including left-to-right, anterior-to-posterior and primary-to-secondary-to-tertiary cortices. Although there appears to be a left hemisphere bias toward language very early in life (perhaps even at birth), language cortices go through significant reorganization in early childhood and through early adulthood. Hence, the patterns of impairment for production and/or comprehension of language involve very different neural substrates across development. Our review also suggests that the frontotemporal network involved in language processing in adulthood does not go through the same reorganization in older adulthood. Taken together this review suggests that acquisition of a native language is a long drawn out process that involves multiple brain areas. Once language processing in this frontotemporal network takes root it, remains stable across most of adult life (but see Green, Crinion, & Price, 2006, for a different view). Note 1. An example of this can be seen in the visual cortex. Primary visual cortex represents visual space in a one-to-one fashion such that each point in the brain corresponds to a location in an individual’s visual space. Secondary cortices re-represent these areas using more complex information such as motion or color. An example of tertiary areas can be seen in the representation of single objects in space which do not correspond to that seen in a particular location in space. A similar phenomenon occurs in auditory cortex where primary auditory cortex is involved in more primitive forms of auditory processing, whereas secondary and tertiary auditory

< previous page

page_302

next page >

< previous page

page_303

next page >

Page 303 areas may re-represent other aspects of the auditory signal such as phonology, lexical, or syntactic aspects of the signal. References Ackermann, H., & Riecker, A. (2004). The contribution of the insula to motor aspects of speech production: Review and a hypothesis. Brain and Language, 89, 320–328. Aram, D. M. (1988). Language sequelae of unilateral brain lesions in children. In F. Plum (Ed.), Language, communication, and the brain . New York: Raven Press. Aram, D. M., Eisele, J. A., Rapin, I., & Segalowitz, S. J. (1992). Plasticity and recovery of higher cognitive functions following early brain injury. In I. Rapin & S. J. Segalowitz (Eds.), Handbook of neuropsychology, Vol. 6. (pp. 73–92). New York: Elsevier. Basser, L. (1962). Hemiplegia of early onset and the faculty of speech with special reference to the effects of hemispherectomy. Brain, 85, 427–460. Bates, E., Thal, D., Trauner, D., Fenson, J., Aram, D., Eisele, J., and Nass, R. (1997). From first words to grammar in children with focal brain injury. Developmental Neuropsychology , 13, 275–343. Bates, E., Vicari, S., & Trauner, D. (1999). Neural mediation of language development: Perspectives from lesion studies of infants and children. In H. Tager-Flusberg (Ed.), Neurodevelopmental disorders (pp. 533–581). Cambridge, MA: MIT Press. Bates, E., Wilson, S. M., Saygin, A. P., Dick, F., Sereno, M. I., Knight, R. T., and Dronkers, N. F. (2003). Voxel-based lesion-symptom mapping. Nature Neuroscience , 6 , 448–450. Best, C. T. (1988). The emergence of cerebral asymmetries in early human development: A literature review and a neuroembryological model. In D. L. Molfese & S. J. Segalowitz (Eds.), Brain lateralization in children: Developmental implications (pp. 5–34). New York: Guilford. Bishop, D. (1983). Linguistic impairment after left hemidecortication for infantile hemiplegia? A reappraisal. Quarterly Journal of Experimental Psychology A: Human Experimental Psychology, 35, 199–207. Bishop, D. (1994). Grammatical errors in specific language impairment: Competence or performance limitations? Applied Psycholinguistics , 15, 507–550. Bookheimer, S. (2002). Functional MRI of language: new approaches to understanding the cortical organization of semantic processing. Annual Review of Neuroscience , 25, 151–188. Bradshaw, J., & Nettleson, N. (1981). The nature of hemispheric specialization in man. Behavioral and Brain Sciences, 4 , 51–91. Broca, M. P. (1865). Sur la siege de la faculté langage articulé. Bulletins de la Société d’Anthropologie, 6 , 377–396. Broca, P. (1988). On the speech center. In L. T. Benjamin (Ed.), A history of psychology: Original sources and contemporary research. New York: McGraw-Hill. Brown, J. W., & Grober, E. (1983). Age, sex, and aphasia type. Evidence for a regional cerebral growth process underlying lateralization. Journal of Nervous and Mental Disease , 171 , 431–434. Brown, J. W., & Jaffe, J. (1975). Hypothesis on cerebral dominance. Neuropsychologia , 13, 107–110.

< previous page

page_303

next page >

< previous page

page_304

next page >

Page 304 Buxhoeveden, D., & Casanova, M. (2000). Comparative lateralisation patterns in the language area of human, chimpanzee, and rhesus monkey brains. Laterality , 5 , 315–330. Campbell, S., & Whitaker, H. A. (1986). Cortical maturation and developmental neurolinguistics. In J. E. Obrzut & G. W. Hynd (Eds.), Child neuropsychology: Vol. I. Theory and research (pp. 55–72). New York: Academic Press. Cepanec, M., & Judas, M. (2007). [The past and present status of “motor speech area”: from classical modular “center” to nodal and supramodal component of multiple neuronal networks]. Lijec Vjesn, 129 , 401–406. Clark, L. E., & Knowles, J. B. (1973). Age differences in dichotic listening performance. Journal of Gerontology, 28, 173–178. Corballis, M. C., & Morgan, M. J. (1978). On the biological basis of human laterality: I. Evidence for a maturational left–right gradient. Behavioral and Brain Sciences, 1 , 261–269. Crossley, M., & Hiscock, M. (1992). Age-related differences in concurrent-task performance of normal adults: Evidence for a decline in processing resources. Psychology and Aging , 7 , 499–506. Dapretto, M., Bookheimer, S., & Mazziotta, J. (1999). Form and content: Dissociating syntax and semantics in sentence comprehension. Neuron , 24, 427–432. Dennis, M., & Whitaker, H. A. (1977). Hemispheric equipotentiality and language acquisition. In S. J. Segalowitz & F. A. Gruber (Eds.), Language development and neurological theory (pp. 93–106). New York: Academic Press. Dooling, E. C., Chi, J. G., & Gilles, F. H. (1983). Telencephalic development: Changing gyral patterns. In F. H. Gilles & E. C. Dooling (Eds.), The developing human brain: Growth and epidemiologic neuropathy (pp. 94–104). Boston: John Wright. Dronkers, N. F. (1996). A new brain region for coordinating speech articulation. Nature, 384, 159–161. Duara, R., Grady, C., Haxby, J., Ingvar, D., Sokoloff, L., Margolin, R., Manning, R. G., Cutler, N. R., and Rapoport, S. I. (1984). Human brain glucose utilization and cognitive function in relation to age. Annals of Neurology , 16, 702– 713. Eisele, J., & Aram, D. (1995). Lexical and grammatical development in children with early hemisphere damage: A cross-sectional view from birth to adolescence. In P. Fletcher & B. MacWhinney (Eds.), The handbook of child language (pp. 664–689). Oxford: Blackwell. Entus, A. K. (1977). Hemispheric asymmetry in processing of dichotically presented speech and nonspeech stimuli by infants. In S. J. S. F. A. Gruber (Ed.), Language development and neurological theory (pp. 63–73). New York: Academic Press. Fenson, L., Dale, P., Reznick, J. S., Thal, D., Bates, E., Hartung, J., Pethick, S., & Reilly, J. S. (1993). The MacArthur Communicative Development Inventories: User’s guide and technical manual . San Diego: Singular Publishing Group. Flynn, J. R. (1998). Rising IQ scores: Implications for the elderly. Australasian Journal on Ageing , 17, 106–107. Flynn, J. R. (2006). Tethering the elephant: Capital cases, IQ, and the Flynn effect. Psychology, Public Policy, and Law, 12, 170–189. Friederici, A. D. (2002). Towards a neural basis of auditory sentence processing. Trends in Cognitive Sciences, 6 , 78– 84. Friederici, A. D., Ruschemeyer, S. A., Hahne, A., & Fiebach, C. J. (2003). The role

< previous page

page_304

next page >

< previous page

page_305

next page >

Page 305 of left inferior frontal and superior temporal cortex in sentence comprehension: localizing syntactic and semantic processes. Cerebral Cortex , 13, 170–177. Graves, R. E. (1997). The legacy of the Wernicke–Lichtheim model. Journal of the History of the Neurosciences, 6 , 3–20. Green, D., Crinion, J., & Price, C. (2006). Convergence, Degeneracy and Control. In P. Indefrey (Ed.), The cognitive neuroscience of second language acquisition (pp. 99–125). Malden: Blackwell. Grossman, M., Cooke, A., DeVita, C., Alsop, D., Detre, J., Chen, W., & Gee, J. (2002a). Age-related changes in working memory during sentence comprehension: an fMRI study. Neuroimage , 15, 302–317. Grossman, M., Cooke, A., DeVita, C., Chen, W., Moore, P., Detre, J., Alsop, D., & Gee, J. (2002b). Sentence processing strategies in healthy seniors with poor comprehension: an fMRI study. Brain and Language, 80, 296–313. Head, H. (1926). Aphasia and kindred disorders of speech . Cambridge: Cambridge University Press. Hernandez, A. E., Kotz, S. A., Hoffman, J., Valentin, V. V., Dapretto, M., & Bookheimer, S. Y. (2004). The neural correlates of grammatical gender decisions in Spanish. NeuroReport , 15, 863–866. Hickok, G., & Poeppel, D. (2004). Dorsal and ventral streams: A framework for understanding aspects of the functional anatomy of language. Cognition, 92, 67–99. Hillis, A. E., Work, M., Barker, P. B., Jacobs, M. A., Breese, E. L., & Maurer, K. (2004). Re-examining the brain regions crucial for orchestrating speech articulation. Brain, 127 , 1479–1487. Hiscock, M., & Kinsbourne, M. (1977). Selective listening asymmetry in preschool children. Developmental Psychology, 13, 217–224. Hiscock, M., & Kinsbourne, M. (1994). Phylogeny and ontogeny of cerebral lateralization. In R. J. Davidson & K. Hugdahl (Eds.), Brain asymmetry (pp. 535–578). Cambridge, MA: MIT Press. Hoptman, M. J., Davidson, R. J., Gudmundsson, A., Schreiber, R. T., & Ershler, W. B. (1996). Age differences in visual evoked potential estimates of interhemispheric transfer. Neuropsychology , 10, 263–271. Horn, J. L., & Cattell, R. B. (1967). Age differences in fluid and crystallized intelligence. Acta Psychologica , 26, 107– 129. Huttenlocher, P. R. (1990). Morphometric study of human cerebral cortex development. Neuropsychologia , 28, 517– 527. Huttenlocher, P. R. (1994). Synaptogenesis in the human cerebral cortex. In G. Dawson & K. W. Fischer (Eds.), Human behavior and the developing brain (pp. 137–152). New York: Guilford. Huttenlocher, P. R., & Dabholkar, A. S. (1997). Regional differences in synaptogenesis in human cerebral cortex. Journal of Comparative Neurology , 387 , 167–178. Huttenlocher, P. R., & de Courten, C. (1987). The development of synapses in striate cortex of man. Human Neurobiology , 6 , 1–9. Huttenlocher, P. R., de Courten, C., Garey, L. J., & Van der Loos, H. (1982a). Synaptic development in human cerebral cortex. International Journal of Neurology , 16–17 , 144–154. Huttenlocher, P. R., de Courten, C., Garey, L. J., & Van der Loos, H. (1982b). Synaptogenesis in human visual cortex —evidence for synapse elimination during normal development. Neuroscience Letters , 33, 247–252.

< previous page

page_305

next page >

< previous page

page_306

next page >

Page 306 Ivry, R. B., & Robertson, L. C. (1998). The two sides of perception . Cambridge, MA: MIT Press. Johnson, R. C., Cole, R. E., Bowers, J. K., Foiles, S. V., Nikaido, A. M., Patrick, J. W., & Woliver, R. E. (1979). Hemisphere efficiency in middle and later adulthood. Cortex, 15, 109–119. Johnston, J. R., & Kamhi, A. G. (1984). Syntactic and semantic aspects of the utterances of language impaired children: The same can be less. Merrill-Palmer Quarterly, 30, 65–85. Kang, A. M., Constable, R. T., Gore, J. C., & Avrutin, S. (1999). An event-related fMRI study of implicit phrase-level syntactic and semantic processing. Neuroimage , 10, 555–561. Kuriki, S., Mori, T., & Hirata, Y. (1999). Motor planning center for speech articulation in the normal human brain. NeuroReport , 10, 765–769. LeMay, M., & Culebras, A. (1972). Human brain—morphologic differences in the hemispheres demonstrable with carotid arteriography. New England Journal of Medicine , 287 , 168–170. Lenneberg, A. (1967). Biological foundations of language. New York: John Wiley. Lezak, M. D. (1995). Neuropsychological assessment (3rd ed.). New York: Oxford University Press. Marchman, V. A., Miller, R., & Bates, E. A. (1991). Babble and first words in children with focal brain injury. Applied Psycholinguistics , 12, 1–22. Miceli, G., Caltagirone, C., Gainotti, G., Masullo, C., Silveri, M. C., & Villa, G. (1981). Influence of age, sex, literacy and pathologic lesion on incidence, severity and type of aphasia. Acta Neurologica Scandinavia , 64, 370–382. Miceli, G., Turriziani, P., Caltagirone, C., Capasso, R., Tomaiuolo, F., & Caramazza, A. (2002). The neural correlates of grammatical gender: An fMRI investigation. Journal of Cognitive Neuroscience , 14, 618–628. Mills, D. L., Coffey-Corina, S. A., & Neville, H. J. (1993). Language acquisition and cerebral specialization in 20month-old infants. Journal of Cognitive Neuroscience , 5 , 317–334. Mills, D. L., Coffey-Corina, S., & Neville, H. J. (1997). Language comprehension and cerebral specialization from 13 to 20 months. Developmental Neuropsychology , 13, 397–445. Mortensen, E. L., & Kleven, M. (1993). A WAIS longitudinal study of cognitive development during the life span from ages 50 to 70. Developmental Neuropsychology , 9 , 115–130. Ni, W., Constable, R. T., Mencl, W. E., Pugh, K. R., Fulbright, R. K., Shaywitz, S. E., Shaywitz, B. A., Gore, J. C., & Shankweiler, D. (2000). An event-related neuroimaging study distinguishing form and content in sentence processing. Journal of Cognitive Neuroscience , 12, 120–133. Obler, L. K., Albert, M. L., Goodglass, H., & Benson, D. F. (1978). Aphasia type and aging. Brain and Language, 6 , 318–322. Poldrack, R. A., Wagner, A. D., Prull, M. W., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. (1999). Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex. Neuroimage , 10, 15–35. Rapp, A. M., Leube, D. T., Erb, M., Grodd, W., & Kircher, T. T. (2004). Neural correlates of metaphor processing. Cognitive Brain Research , 20, 395–402. Raz, N., Gunning, F. M., Head, D., Dupuis, J. H., McQuain, J., Briggs, S. D., Loken,

< previous page

page_306

next page >

< previous page

page_307

next page >

Page 307 W. J., Thornton, A. E., & Acker, J. D. (1997). Selective aging of the human cerebral cortex observed in vivo: Differential vulnerability of the prefrontal gray matter. Cerebral Cortex , 7 , 268–282. Raz, N., Rodrigue, K. M., Head, D., Kennedy, K. M., & Acker, J. D. (2004). Differential aging of the medial temporal lobe: A study of a five-year change. Neurology , 62, 433–438. Riva, D., & Cazzaniga, L. (1986). Late effects of unilateral brain lesions before and after the first year of life. Neuropsychologia , 24, 423–428. Riva, D., Cazzaniga, L., Pantaleoni, C., Milani, N., & Fedrizzi, E. (1986). Acute hemiplegia in childhood: The neuropsychological prognosis. Journal of Pediatric Neurosciences, 2 , 239–250. St James-Roberts, I. (1981). A reinterpretation of hemispherectomy data without functional plasticity of the brain: I. Intellectual function. Brain and Language, 13, 31–53. Salthouse, T. A. (1988). Resource-reduction interpretations of cognitive aging. Developmental Review, 8 , 238–272. Satz, P., Strauss, E., & Whitaker, H. (1990). The ontogeny of hemispheric specialization: Some old hypotheses revisited. Brain and Language, 38, 596–614. Schlaggar, B. L., Brown, T. T., Lugar, H. M., Visscher, K. M., Miezin, F. M., & Petersen, S. E. (2002). Functional neuroanatomical differences between adults and school-age children in the processing of single words. Science , 296 , 1476–1479. Sowell, E. R., Peterson, B. S., Thompson, P. M., Welcome, S. E., Henkenius, A. L., & Toga, A. W. (2003). Mapping cortical change across the human life span. Nature Neuroscience , 6 , 309–315. Spreen, O., Risser, A. H., & Edgell, D. (1995). Developmental neuropsychology . New York: Oxford University Press. Stiles, J., Bates, E., Thal, D., Trauner, D., & Reilly, J. (1998). Linguistic, cognitive and affective development in children with pre- and perinatal focal brain injury: A ten-year overview from the San Diego longitudinal project. In C. RoveeCollier, I. Lipsitt, & H. Hayne (Eds.), Advances in infancy research (pp. 131–163). Norwood, NJ: Ablex. Stiles, J., Bates, E. A., Thal, D., Trauner, D. A., Reilly, J., & Johnson, M. H. (2002). Linguistic and spatial cognitive development in children with pre- and perinatal focal brain injury: A ten-year overview from the San Diego longitudinal project. In M. H. Johnson, Y. Munakata, & R. O. Gilmore (Eds.), Brain development and cognition: A reader (2nd ed., pp. 272–291). Oxford: Blackwell. Stowe, L., Haverkort, M., & Zwarts, F. (2005). Rethinking the neurological basis of language. Lingua, 115 , 997–1042. Teszner, D., Tzavaras, A., Gruner, J., & Hécaen, H. (1972). L’asymétrie droite–gauche du planum temporale: A propos de l’étude anatomique de 100 cerveaux. Revue Neuerologique , 126 , 444–449. Tettamanti, M., Alkadhi, H., Moro, A., Perani, D., Kollias, S., & Weniger, D. (2002). Neural correlates for the acquisition of natural language syntax. Neuroim-age, 17, 700–709. Thal, D. J., Marchman, V. A., Stiles, J., & Aram, D. (1991). Early lexical development in children with focal brain injury. Brain and Language, 40, 491–527. Van Gorp, W. G., Satz, P., & Mitrushina, M. (1990). Neuropsychological processes associated with normal aging. Developmental Neuropsychology , 6 , 279–290.

< previous page

page_307

next page >

< previous page

page_308

next page >

Page 308 Vargha-Khadem, F., Isaacs, E., & Muter, V. (1994). A review of cognitive outcome after unilateral lesions sustained during childhood. Journal of Child Neurology , 9 , 2S67–62S73. Vicari, S., Caselli, M. C., & Tonucci, F. (2000). Asynchrony of lexical and morphosyntactic development in children with Down Syndrome. Neuropsychologia , 38, 634–644. Vigliocco, G. (2000). Language processing: The anatomy of meaning and syntax. Current Biology, 10, R78–80. Wada, J. A., Clark, R., & Hamm, A. (1975). Cerebral hemispheric asymmetry in humans. Archives of Neurology , 32, 239–246. Wagner, A. D., Desmond, J. E., Demb, J. B., Glover, G. H., & Gabrieli, J. D. E. (1997). Semantic repetition priming for verbal and pictorial knowledge: A functional MRI study of left inferior prefrontal cortex. Journal of Cognitive Neuroscience , 9 , 714–726. Wagner, A. D., Maril, A., Bjork, R. A., & Schacter, D. L. (2001). Prefrontal contributions to executive control: fMRI evidence for functional distinctions within lateral prefrontal cortex. Neuroimage , 14, 1337–1347. Wagner, A. D., Pare-Blagoev, E. J., Clark, J., & Poldrack, R. A. (2001). Recovering meaning: Left prefrontal cortex guides controlled semantic retrieval. Neuron , 31, 329–338. Weems, S. A., & Reggia, J. A. (2006). Simulating single word processing in the classic aphasia syndromes based on the Wernicke–Lichtheim–Geschwind theory. Brain and Language, 98, 291–309. Wise, R. J., Greene, J., Buchel, C., & Scott, S. K. (1999). Brain regions involved in articulation. Lancet , 353 , 1057– 1061. Witelson, S. F. (1983). Bumps on the brain: Right–left anatomic asymmetry as a key to functional lateralization. In S. J. Segalowitz (Ed.), Language functions and brain organization (pp. 117–144). New York: Academic Press. Witelson, S. F., Beresh, H., & Kigar, D. L. (2006). Intelligence and brain size in 100 postmortem brains: Sex, lateralization and age factors. Brain, 129 , 386–398. Woods, B. T. (1980). The restricted effects of right-hemisphere lesions after age one; Wechsler test data. Neuropsychologia , 18, 65–70. Woods, B. T., & Teuber, H. L. (1978). Changing patterns of childhood aphasia. Annals of Neurology , 3 , 273–280. Zatorre, R. J., Meyer, E., Gjedde, A., & Evans, A. C. (1996). PET studies of phonetic processing of speech: Review, replication, and reanalysis. Cerebral Cortex , 6 , 21–30.

< previous page

page_308

next page >

< previous page

page_309

Page 309 INDEX Italic refers to tables, bold refers to figures Action Naming Test 261 additive bilingualism 174 adolescence stage (12–17) 2, 3, 175–7 adult signers 229–32, 235 adulthood gestures 202–12 Advanced Vocabulary 260 affective variables 130–1 affordances 77–8, 79–80 age effect 134, 172 age-related variables 2, 133–6, 206–12, 246, 248, 259–62 aging changes 297–302; behavioral 298–9; physiological 297–8 Alzheimer’s disease 211, 280 American Sign Language (ASL) 217, 218, 219, 220, 221, 222, 223, 225, 230 analytic strategies 249–52 analytic versus holistic style 296 aphasia 156–64, 288–302 aptitude tests 132 ASL see American Sign Language attitude to language acquision 182–3 attractor states 84, 127 attrition 1, 11–12, 127, 171–84, 245, 253; sign language 232, 235 auditory detection 156 Australian Sign Language (Auslan) 218 BA see Brodmann’s areas Baltes, P. 246 Basic Variety (of second language) 127 Basic Vocabulary 260 Bassano, D. 86, 93 beat and rhythmic movement gesture 194–5, 209 Berend (case study) 93, 95 Berman, R. 2 Bernicot, J. 1 bilingual aphasia 159–63 bilingualism 11, 113, 117, 133, 146–64, 171–6, 245–65 Binding Theory 21 Boston Naming Test 261 brain: anatomy changes 291–5; placticity 135; see also hemisphere; neural areas British Sign Language (BSL) 218 Broca’s aphasia 296 Brodmann’s areas (BA) 289–91 capacity-constrained theory (CC) 279 case studies: Berend 93, 95; bilingual preschool children 113; Heleen 93–5, 94, 95; Jessica 93–5, 94, 95; Lisa 93, 95; Nun Study 263; Peter 87, 89; Ugiusu 113 catastrophe theory 66–7 categorical perception 43–4 central executive processor 271 child signers 222–9 childhood: change during 14–15; gesture 197–201; Piaget’s model 5 children 2 chimp experiments 81 Chomsky, N. 19, 20, 40, 81 CLT see communicative language teaching co-speech gestures 194–5 cognitive aging 245–6, 258 cognitive processes 51, 61, 110

next page >

cognitive variables 132–3 cohort effects 253–4 cohort-sequential strategy 249, 250 Colletta, J.M. 201

< previous page

page_309

next page >

< previous page

page_310

Page 310 combinatorial generator 8, 64, 65, 86 communicative gesture 191–2, 196 communicative language teaching 129 comparative fallacy 28 competence (Chomsky) 81 Complementerizer Phrase (CP) 22 complex sentences 52–4 comprehension versus production 296 computational–representational understanding of mind 60 confrontation naming 261–2 Contrastive Analysis Hypothesis 136–7 Corsi Blocks test 273, 277 cortical volume 291 Counting Span test 272 CP see Complementerizer Phrase Critical Period Hypothesis (CPH) 134 cross-language links 160–2 cross-linguistic variables 136–7 cross-sectional designs 246–8, 247, 249–52 cross-sequential strategy 250 cultural variables 128–30 de Courcy, M. 113–14 deafness 12 dedifferentiation 278 deictic gestures 194, 196, 197, 198, 201 dementia 263, 280 Determiner Phrase (DP) 22 development: chaotic 126; education and 120; environmental factors 5; gestures 12–13; and history (bioculture) 4–5; as lifelong process 4; linear 126; long-term dynamics 82–95; methodoloical aspects 13–15; neurolinguistic aspects 13–15; non-linear 128; non-verbal 12–13; and ontogeny 4–5, 107, 109; preintellectual 109, 111–12; prelinguistic 109; processes 5–6, 47–9; psychological 4, 70; see also first language; second language developmental agent model 68–71 DGS see German Sign Language diagnosis 153–5 dichotic listening 302 Diffusion Tensor Imaging (DTI) 293 digit monitoring test 275 Digit Symbol Substitution subtest 259 DP see Determiner Phrase drop-out rate in studies 257–8 DTI see Diffusion Tensor Imaging dynamic systems theory (DST): 9, 10, 96, 126, 155; competition-support systems 61–5; development and resource 61–5; developmental agent model 68–71; dynamic field theory 65–8, 69; dynamic growth model 66, 92; dynamic relationships 64; embedded action 74–81, 85; embodied action 74–81, 85; epigenetic model 82–5; examples of 126; grounded meaning 74–8; and lifespan perspective 138; long-term 73, 85; overview 60–71; robotic model 68–71; view of SLD 125–8 dyslexia 2

next page >

early adulthood stage (17–45) 2, 4–5 early childhood stage (0–3) 2, 19, 42 early linguistic development 43–52 EF see executive function egocentric speech 110 elderly 1, 2 Ellis, R. 131 embedded action 74–81, 85 emblems (gestures) 192–3 embodied action 60–1, 74–81, 85 emergentism 40, 155 emulation learning 42 epigenetic model (long-term dynamics) 70, 82–5, 96 episodic buffer 271 ethnic gesturing 203–4 ETS Kit of Factor Referenced Tests 259, 260 executive function (EF) 276, 278 experimental selectivity 257 expressive aphasia 158 eye-tracking 281 FAS verbal fluency test 274 feature strength 21 fingerspelling 217, 220–1, 227, 231–2 first language (L1) acquisition 1, 9, 22, 31, 40–57, 74–8, 111 first language (L1) development 6, 8, 12, 19–32, 126, 222–9 fluent aphasia 158 fMRI see functional Magnetic Resonance Imaging focal lesions 295–6 foreign accent in signing 229

< previous page

page_310

next page >

< previous page

page_311

Page 311 foreign language teaching 129 formulaic sequences 48 Forward and Backward Digit Span tests 272 frequency of use 253 Full Competence Hypothesis 22 Full Transfer (Full Access model 24 function words 87 functional Magnetic Resonance Imaging (fMRI) 292, 293, 299, 302 general developmental model 246 General Social Survey (GSS) 260 generalization 50–4 Generative Grammar 19–22, 40–1, 46 genetic analysis 108–9 German Sign Language (DGS) 230 gerontology 245 gesticulation (gestures) 192–3 gesture 12–13, 191–212 global aphasia 300 grammar (as complex ecological system) 85–93 grammatical complexity 262–3 growth models 87–93, 90 heads (in phrasal constructions) 21 Heleen (case study) 93–5, 94, 95 hemisphere: asymmetries in immature brain 291–2; damage 294, 297; disparities 292, 297–8; effects on comprehension 296; function deterioration 298; hemispherectomy 294–5; hemispheric lesions 294; injuries 295; interhemispheric hypothesis 297; left hemisphere 297; right hemisphere effects on comprehension 296; see also brain; neural areas Hoffman, E. 176 holophrastic generator 8, 64, 65, 86 hypothesized grammatical generators 65 iconic gestures 194, 209 identity and identification 175 imitative learning 42, 77 immature brain asymmetries 291–2 impairment 1, 10 inferior frontal gyrus (IFG) 289–90 Information Packaging Hypothesis (IPH) 196–7 information-processing theory 60 inhibition 278 innate faculty 19–20 inner speech 110, 115–16 instruction 129 intellectual stage 117–18 intellectualization of speech 109–10 interacting variables 128–37 interhemispheric hypothesis 297 internalization 109, 112, 113, 116 intraindividual variability 93, 248 IPH see Information Packaging Hypothesis Israeli Sign Language (ISL) 218 Jessica (case study) 93–5, 94, 95 joint attention 42 Kantor, R. 229–30 Kendon, A. 192 Kendon’s Continuum 192–4 Klein, W. 1 Kotic-Friedgut, B. 120 L1 see first language L2 see second language language: acquisition 182–3, 272, 279;

next page >

aptitude tests 132; attrition see attrition; decline 232–4; environment 252–3, 254; impairment 10, 11, 146–64; processing 133, 279–81; proficiency 147, 252; reversion in old age 177–9; use 180–1 Language Acquisition Device 6, 7, 19 Language development across childhood and adolescence 2 Lantolf, J.P. 114 late adulthood stage (65 plus) 3, 4, 5 late childhood stage (3–12) 2, 4, 19 late talkers 150 Levinson, D. 3 lexical knowledge 62–3, 87, 90, 91, 258; see also vocabulary Lexical Retrieval Hypothesis (LRH) 196–7, 212 Lifespan developmental psychology 1 Linguistic Deficit Coding Hypothesis 132 linguistic innovations 3 Lisa (case study) 93, 95 Listening Span test 272 literacy 2 LLAMA aptitude test 132

< previous page

page_311

next page >

< previous page

page_312

Page 312 logical problem of language acquisition 20 longitudinal designs 247, 248–52 longitudinal studies of monolingualism 258–64 Lost in Translation 176 Luria, A.R. 119–20 The lifespan development of individuals 1 McNeill, D. 192–5, 209, 211, 212 McNeill’s Continuum 192–4, 193 , 194 Magnetic Resonance Imaging (MRI) 290, 292, 293, 299 Magnetoencephalography (MEG) 290 manual babbling 223–4 maturational factors 3, 31, 108 maturational gradients 292–4 MEG see Magnetoencephalography memory 132, 139 mental operations 110 metaconsciousness 111, 121n metalinguistic awareness 115 metaphoric gestures 194 microgenetic studies in development 141 middle adulthood stage (45–65) 3 middle age 179–80 Mill Hill Vocabulary Scale 259, 260 Minimal Trees 24 mixed receptive–expressive aphasia 158 modularity (brain) 55 MOM (means, opportunity and motive) framework 10–11, 147–8, 164 monolingual PLI 149–52 monolingualism studies (longitudinal) 258–63 morality-associated selectivity 257 MRI see Magnetic Resonance Imaging multilingualism 12, 13, 245–65 myelination 293 n-back test 275 neural areas 289–91: 289; see also brain; hemisphere neural substrates of language 288–302 neurodevelopmental processes 293 neuroimaging 299–300 non-fluent aphasia 158 non-linguistic task performance 155–6 nonmanuals 217, 219–20, 225–7, 231–2 notions of change 246 noun phrase (NP) 21, 25 Nun Study 263 old age language reversion 177–9 old-old stage (85 plus) 3 one-word period 1 ontonogenesis 4–5, 107, 109, 118 operational span test (OSPAN) 273 Padden, C. 227 pantomime (gestures) 192–3 Parkinson’s disease 12, 217, 232–4 pattern recognition 43–4 PET see Positron Emission Tomography Peter (case study) 87, 89 phonemic encoding ability 132 phonology 217, 218–19, 223–5, 229–31 phrasal constructions 21 Piaget’s model of childhood 5, 68, 81 picture naming 261–2 plus–minus switching task 25 Positron Emission Tomography (PET) 290, 298 Poverty of Input 7 poverty of the stimulus 20, 56 practice effects 254–7 predicate raisat (Morton) 81 preintellectual stage 109, 111–12 prelinguistic development 1, 42–3, 109–10, 197 prepositions frequency 94, 95 primary language impairment (PLI) 10, 11, 146–64 private speech stage 112–14

next page >

productivity 50 progressive lateralization 294 property theory (of language acquisition) 32 propositional density 262–3 psychosocial factors 3 Raven’s Standard Progressive Matrices 298 REA see right ear advantages Reading Span test 272 Recognition Vocabulary Test 260 recruitment for studies 257–8 representational gestures 198, 209 representationalist stance 60 resources (in DST framework) 128, 140 right ear advantages (REA) 298 robotic model 68–71 San Diego Program in Cognitive and Neural Development 295

< previous page

page_312

next page >

< previous page

page_313

Page 313 Saville-Troike, M. 113, 114–15 Schaie, K. Warner 246, 248, 254 Schaie–Thurstone Adult Mental Abilities Test 260 SCT see sociocultural theory Seattle Longitudinal Study 254 second language (L2) acquisition 1, 7, 23, 26–31, 125 second language (L2) development 6, 9–12, 19–32, 107–21, 125–42, 229–31 semantics 63 serial recall 272 Shipley Institute of Living Scale 259 sign language (gestures) 192–3 sign language 12, 217–36, 221, 225, 236n Signed Exact English (SEE) 236n SLD see second language development social cognition 42–3 social origins and gesture 204–6 social-psychological factors 130 social referencing 42 social variables 128–30, 147, 203–12 sociocultural theory (Vygotskyan) 9, 107–21 space in sign language 221–2 span testing 272–6 spatial components of narratives 217, 227–9 speech and gesture 195, 199–200, 207–10 spontaneous gesture 210 starting small hypothesis 46–7 stochastic learning 44–5 storage mechanisms (memory) 271 structural MRI 293 structure mapping 45–6 subtractive bilingualism 174 Superior Temporal Gyrus 288 synapses 293 syntactic generators 8, 64, 65, 86 syntax 22, 61, 62, 63, 78 teachers’ gestures 202 Tense Phrase (TP) 22 test–retest effects 254–7 theory of mind 41 time-lag testing 255 time-sequential strategy 250–1, 251 Tower of Hanoi test 272, 274, 277 TP see Tense Phrase Trail Making Test 272, 273 transition theory (of language acquisition) 32 Truncation Hypothesis 31 UG see Universal Grammar Ugiusu (case study) 113 unifunctuality 226 Universal Grammar 6, 7, 8, 19–32, 41, 46, 47, 55, 56 unlearning (generalizations) 54–5 updating tests 275 usage-based theory of acquisition 7, 8, 40–1, 50–2, 55–7 V2 see Verb Second Valueless Features Hypothesis 24, 26, 29 variation 127–8 vector of motion 82 Verb Phrase (VP) 25 Verb Second (V2) 24, 25 verbal ability 258 verbal thought 110 Verhaeghen, P. 259 Verhulst equation 63 Very early Parameter Setting 31 Visual Matching and Crossouts Subtest 259 Visual Pattern Test 273 visual scalpel 271 vocabulary 259–61; see also lexical knowledge Vocabulary subtest (WAIS) 259 VP see Verb Phrase Vygotskian theory see sociocultural theory WAIS-R Vocabulary subtest 260

Wechsler Adult Intelligence Scale 298 Wernicke–Lichtheim model 288, 291 wh-questions 20–1, 47–9 Wisconsin Card Sorting Test 272, 273, 277 Woodcock–Johnson Psychoeducational Battery 255, 259 working memory 139, 271–83

< previous page

page_313