51,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Geisteswissenschaft
- Sprache: Englisch
Drawing on major research developments in the field, Vihman has updated and extensively revised the 1996 edition of her classic text to provide a thorough and stimulating overview of current studies of child production and perception and early word learning.
- Offers a full survey of the thinking on how babies develop phonological knowledge
- Provides a much needed update on the field – one in which this book remains unique, and in which there have also been dramatic developments since the publication of the first edition
- Surveys what has been learned about phonological development and raises questions for further study
- The only book that includes balanced treatment of research in perception and production and attempts a synthesis of these fields, which have generally developed in isolation from one another
- Includes a new chapter providing an overview of communicative and attentional development, as well as perceptual and vocal development, in the first 18 months, with additional focus on both implicit and explicit learning mechanisms
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 957
Veröffentlichungsjahr: 2014
Ähnliche
Table of Contents
Praise Page
Title Page
Copyright
Dedication
Note on Second Edition
Acknowledgments
Chapter 1: Introduction
Biological Foundations of Language Development
Phonological Development: Goals and Challenges
Methodologies: Data Sources and Theoretical Perspectives
Overview
Chapter 2: Precursors to Language: The First 18 Months of Life
The Development of Linguistic Form and Function
1 Early Capacities: Birth to 2 Months
2 Early Capacities: 2 to 4 Months
3 Early Capacities: 4 to 6 Months
4 First Advances: 6 to 9 Months
5 Bringing the Strands Together: 9 to 12 Months
6 Transition to Language Use: 12 to 18 Months
Learning Mechanisms
Summary: Precursors and the Transition to Language Use
Note
Chapter 3: Development in Perception: Early Capacities, Rapid Change
Issues that Motivated the Study of Infant Speech Perception
Methods Used to Study Infant Speech Perception
Discrimination: Infant Capacities
Mechanisms Underlying Infant Perception
Developmental Change in Perception
‘Perceptual Narrowing’: Models of Developmental Change
Cross-Modal Perception
Summary: The Infant Listener—From Universal to Particular
Notes
Chapter 4: Infant Vocal Production
Early Vocal Production
The Social Context, I: Precanonical Period
The Emergence of Adult-Like Syllables
The Social Context, II: Canonical Period
Vowel Production in the First Year
Influence from the Ambient Language
Summary: Biological and Social Foundations
Notes
Chapter 5: Perceptual Advances in the First Year: Prosody, Segmentation and Distributional Learning
The Role of Prosody and Infant-Directed Speech (IDS)
Prosodic Bootstrapping
Advances in Knowledge of Accentual Patterns and Phonotactics
Experimental Studies of Segmentation
Distributional or Statistical Learning
Summary: Rhythms and Segmental Patterns
Notes
Chapter 6: The Transition to Language Use
Beginnings of Word Comprehension
Development of Intentional Communication
Referential Word Use
Phonetic and Phonological Development
Rhythm in Child Production
Emergence of Phonological Systematicity
Summary: Continuity and Change
Note
Chapter 7: Experimental Studies of Word-Form Learning
The Role of Phonology in Word Recognition and Word Learning
The Perceptual Basis of Word Learning: How Detailed Are Early Word-Form Representations?
Explorations of ‘Phonological Specificity’
Integrating the Findings
Neurophysiological Investigation of Word Learning: Event Related Potentials
Summary: Understanding the Development of Representation
Notes
Chapter 8: Phonological Development in the Bilingual Child
The Study of Child Bilingualism: Parental Strategy and Language Use
‘Non-selectivity’ in Adult Bilingual Processing: Implications for Children
Speech Perception and Processing
Bilingual Production
Separate Systems with Interaction
Summary: The Experience of the Bilingual Child
Notes
Chapter 9: Theories, I: Formalist and Perception Models
What is the Source of the Linguistic System?
The First Linguistic Models
Contemporary Formalist Models
Contemporary Formalist Models: Critique and Appreciation
Perception Models
Perception Models: Critique and Appreciation
Summary: Theory and Data in Developmental Models
Notes
Chapter 10: Theories, II: Functionalist or Emergentist Models
Biological Model
Self-Organizing Models
Usage-Based Models
Whole-Word Phonology
Functionalist or Emergentist Models: Critique and Appreciation
Summary: Emergent Phonology
Notes
Chapter 11: Conclusion
Linking Perception and Production
Effects of Lexical Learning and Language Use
Appendix 1: Data Sources
Appendix 2: Protoword Forms and Uses
Appendix 3: Template Analyses
Word Template Analysis: A Diary Study
Word Template Analysis: Research Studies
Generality of Template Use
Notes
References
Name Index
Subject Index
Praise for Phonological Development: The First Two Years
“This open-minded, comprehensive overview of the intersecting components of phonological development is a masterpiece that should shape new directions of research for many years to come. Vihman elucidates the many underlying assumptions, some in conflict with one another, that have guided research on phonological development, and lays out clearly the relevance of individual variability in very young children. Vihman's work will cause researchers in the disparate areas of perception, production, word learning, variation, and phonological universals to be informed by each other's results, potentially revolutionizing our understanding of phonological development.”
—Sharon Inkelas, University of California, Berkeley
“Phonological Development: The First Two Years is essential reading and a primary text for all specialists and students in language development as well as those interested in phonological development in particular. It provides in-depth and up-to-date coverage of all areas of research relevant to understanding phonological development, with comprehensive reviews of both empirical findings and theoretical frameworks. An emphasis is made on the need to relate the development of perception and production, and the study of phonological development to broader areas of language acquisition. Besides eleven chapters, it also contains valuable appendices on protowords and template analyses. To my knowledge it is the most thorough and important book on this topic to date.”
—David Ingram, Arizona State University
“Marilyn Vihman's work unfolds on the center court of child phonology research. This book gives a broad and insightful account of this complex topic—a treatment that is likely to serve, for a long time, as an indispensable reference on the early stages of learning to speak.”
—Björn Lindblom, Stockholm University
“This eagerly awaited second edition masterfully updates Vihman's review of research on earlier themes as well as on several new themes, much of which attests to the profound inspiration of the seminal first edition.”
—Mary Beckman, Ohio State University
This second edition first published 2014
© 2014 John Wiley and Sons, Inc.
Edition history: Blackwell Publishers Ltd. (1e, 1996)
Registered Office
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Offices
350 Main Street, Malden, MA 02148-5020, USA
9600 Garsington Road, Oxford, OX4 2DQ, UK
The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
For details of our global editorial offices, for customer services, and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell.
The right of Marilyn May Vihman to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Library of Congress Cataloging-in-Publication Data
Vihman, Marilyn May.
Phonological development : the first two years / Marilyn May Vihman.—Second Edition.
pages cm.
Includes bibliographical references and index.
ISBN 978-1-118-34279-4 (hardback)—ISBN 978-1-118-34280-0 (paper) 1. Language acquisition. 2. Grammar, Comparative and general–Phonology. I. Title.
P118.V54 2014
401′.93–dc23
2013039710
A catalogue record for this book is available from the British Library.
Cover image: Photo © Marilyn Vihman
Cover design by www.cyandesign.co.uk
Note on Second Edition
This book provides an extensive overview of research into child language production and perception. It focuses primarily on the first two years of life because, for the majority of children, that includes the whole of the single-word period, when phonological development and change are by far the most rapid, laying the foundation for further language learning. The new edition includes new chapters on development in the first 18 months, segmentation and distributional learning, word learning experiments, and bilingual phonological development; it retains and updates the original chapters on perception, vocal production and the transition into language. Although the new edition has no chapter on prosody, the sections of that chapter that pertain to final lengthening or to speech rhythm in general have been retained in Chapter 6; unfortunately, there is no longer space in a single volume to do justice to the field of prosodic development as a whole. The book also includes one chapter each on formalist and functionalist theoretical models and a new concluding chapter on the link between perception and production and the role of lexical growth in supporting further learning.
Acknowledgments
I would first like to thank the various colleagues who kindly read and commented on drafts of one chapter or another. These include Dunstan Brown, Paul Foulkes, Sam Hellmuth and Sven Mattys, all of whom I am lucky enough to have as colleagues at the University of York. In addition, three former colleagues from the University of Wales, Bangor (now Bangor University) also read and commented on certain chapters: Margaret Deuchar, Virginia C. M. Gathercole (now at Florida International University) and Guillaume Thierry. I also thank my colleagues Laura Bosch, University of Barcelona; Philip Carr, Université Montpellier III; John Esling, University of Victoria; Lorraine McCune, Rutgers University; Lise Menn, University of Colorado; Thierry Nazzi, Université Paris Descartes; Yvan Rose, Memorial University of Newfoundland; and Brenda Seal, Gallaudet University. And I am grateful to the four reviewers who took the time to read through the book as a whole in near-final draft form; their comments and critique saved me from some ‘bloopers’ and have undoubtedly made the book more ‘reader-friendly’.
I would also like to give special thanks to Nick Ellis, now at the University of Michigan, whose mentorship in implicit and explicit learning mechanisms has played a critical role in the thinking that underlies this new text.
I have reserved for special mention the two colleagues with whom I have worked the most closely since the last edition was written, Rory DePaolis (James Madison University), a speech scientist who established my lab when I moved to Britain in 1996 and who has been my mentor not only in acoustic phonetics and perceptual studies but also in pushing me to read and absorb the messages of dynamic systems theory, which has had such a critical impact on my thinking. Finally, it would be hard to overstate my appreciation and gratitude to Tamar Keren-Portnoy, who came to do postdoctoral study with me in 2003 and who stayed on to serve as co-investigator on several grants, moved with me from Wales to York, where we designed our MA programme in phonological development together, and joined the Department of Language and Linguistics as a lecturer in 2010. It has been my joy to work with both Rory and Tamar, and to develop ideas in lively, sometimes daily, dialogue as well as in the many research proposals and papers that we have written together.
I am also grateful to my commissioning editor, Danielle Descoteaux, and my project editor, Julia Kirk, both of whom, in supporting the shifts and changes that marked the evolution of this revised edition, have surely exhibited patience and understanding beyond what is usually called for.
Finally, I thank the ESRC, which has provided generous funding for my research ever since I came to Britain in 1997, and also the European Commission, whose Marie Curie initiative first enabled me to work with Tamar and later to renew my working relationship with Rory, who came to visit Bangor as an International Incoming Fellow.
Chapter 1
Introduction
Biological Foundations of Language Development
Phonological Development: Goals and Challenges
Methodologies: Data Sources and Theoretical Perspectives
Overview
The first two years of life constitute a period of dramatic change, not least because it is in that period that most children begin to make use of words or phrases of the adult language and to combine them into their first sentences. And indeed the first questions to be asked about phonological development, based on early diary studies, related to infant production of speech in the first two years. How universal is the order of learning of speech sounds, for example, in different language communities and different children, and how, if at all, is babbling related to speech (Jakobson, 1941/68)? And how do infants with bilingual exposure manage so successfully to produce two languages like native speakers (Leopold, 1939)?
Later, with advances in technology, it became possible to ask about speech perception: How do children learn to distinguish between the speech sounds that they hear, for example, and how do they begin to discover words in the rapidly changing speech signal, where words are not marked off by pauses (Jusczyk, 1997)? Perceptual discrimination is remarkably acute in the first months of life, as became clear from the first experimental studies in the 1970s, but these early capacities become increasingly attuned to the particular language or languages to which the child is exposed over the first year. We can then ask, how does this process of attunement support word learning (Werker & Curtin, 2005)? Equally basic is the issue of the relation of perception to production: How does the infant's early discriminatory skill translate into vocal practice and word formation (Kuhl et al., 2008)?
Additional questions have received attention and analysis only more recently: How important is the ‘music’ of speech, or speech rhythm or prosody, for phonological development (Mehler et al., 1988; Nazzi, Bertoncini & Mehler, 1998)? And to what extent does word learning itself support advances in knowledge of sounds and sound patterns (Ferguson & Farwell, 1975; Vihman & Keren-Portnoy, 2013)?
The essential mystery of language acquisition, the child's move from having no linguistic system to the beginnings of system, is deeply rooted in the first two years of development. However, to gain insight into that mystery we must look beyond studies of speech perception and vocal production per se to consider the findings of research into the developmental changes occurring in parallel in other domains. Before word use is observed, for example, the notion of intentional communication itself must emerge, followed by understanding of the possibility of communicating by vocal means. These pragmatic advances make up one of the strands that prepare the child for language use.
Changes in neuromotor control as well as in attentional mechanisms enable the child to participate more and more actively in social exchanges over the course of the first year. Early perceptual capacities come to be supplemented by increasingly adult-like vocal expression, leading to preparedness in terms of the phonetic prerequisites for word use. The third preparatory strand is increasing representational capacity. This can be understood as referring to advances in working memory, or the ability to maintain more than one item in mind while preparing a vocal or gestural action; these advances build on the child's emergent pragmatic and phonetic skills and social experiences to complete the set of essential precursors to language use.
We will largely restrict ourselves here to the foundational period of language development for several reasons. First, since change is so rapid, close attention is warranted to each of the successive phases of maturation and learning. Second, a wealth of research, deriving mainly from linguistics, psychology and speech science, has addressed these changes in the past several years; the literature available for review is now so considerable that a longer period could scarcely receive adequate coverage in a single volume.
Finally, the age of two is a sensible demarcation point, if only because this is the age at which ‘late talking toddlers’ are generally identified. These are 2-year-olds who have not yet reached the end of the single-word period—that is, who have fewer than 50 words in production and/or few if any word combinations. These children are considered to be at risk for specific language impairment (SLI), although at least half of them will reach the basic lexical and syntactic developmental milestones by age 2.5 and thus be reclassified as ‘typically developing’, or ‘[late-]bloomers’ (Fernald & Marchman, 2012). There is good reason to believe that difficulties with phonological and lexical development in the single-word period can account to some extent for later difficulties with overall language development (Vihman, Keren-Portnoy, Whitaker, Bidgood & McGillion, 2013). This is an important finding, underlining both the critical significance of the first two years and the relevance of research in phonological development for those interested in understanding language development more generally, and also for clinicians and other practitioners who deal with children.
Biological Foundations of Language Development
Some of the motivating questions of the field of phonological development pertain to language development more generally: How can children learn language so quickly, for example? What special skills or resources do they have? Alternative ways of responding to these questions, taking radically different approaches, fall together under the rubric of ‘biological foundations’. Here we consider the theoretical divide that is central to the field of linguistics and thus also language development, and take up the related question, what kinds of explanations are available? What sources of information can we draw on, to account for the timing and processes of phonological development?
Chomsky and the origins of the LAD and UG
A common non-specialist view holds that children learn language remarkably quickly and easily. This everyday view was enshrined in linguistic theory with the publication of Noam Chomsky's Aspects of the Theory of Syntax (1965), which soon became a key linguistic text, as structuralism was replaced by the theory that grew out of it, generative grammar (Harris, 1993). Chomsky was not centrally concerned with language development, which he has never claimed as a field of expertise. However, given the complexity of the adult linguistic system, for which he could provide ample evidence, he was presumably struck by the paradox of the rapidity and ease with which children—considered rather incompetent in most domains—appeared to ‘pick it up’, without the benefit of explicit instruction.
Chomsky's first foray into this territory was his review of Verbal Behavior, published in 1957 by one of the foremost behavioral psychologists of his day, B. F. Skinner. Based on his work on the effects of conditioning on animals, Skinner saw reinforcement as ‘a necessary condition for language learning’ (as cited in Chomsky, 1959, p. 36) and specified that acquisition of ‘verbal behavior’ occurs
when relatively unpatterned vocalizations, selectively reinforced, gradually assume forms which produce appropriate consequences in a given verbal community… Differential reinforcement shapes up all verbal forms… (Skinner, 1957, p. 31, emphasis added)
Chomsky's sharply critical review of Skinner's book has generally been seen as marking the end of behaviorism's acceptance as a potential account of language learning. In particular, Chomsky argued successfully that the idea that specific ‘reinforcement’ should be a cornerstone of language acquisition was untenable. Chomsky also attacked the idea that frequency (‘a very misleading measure of strength’: p. 34) plays any important role in learning. Instead, Chomsky argued in his later work that infants must be born with foreknowledge of linguistic principles, or, as he put it, with a ‘language-acquisition device’ or LAD (Chomsky, 1965). This clever acronym was soon replaced by Universal Grammar (UG: Chomsky, 1967, 1981b), although the basic idea remained the same.
Despite Chomsky's continued dominance or near dominance of linguistics, his rejection of frequency and reinforcement as playing any role in learning has begun to be quietly set aside, in light of current understanding of the role of implicit alongside explicit learning mechanisms (see Bybee & Hopper, 2001; Ellis, 2002a), of ‘statistical learning’ and its relevance for infants (Saffran, Aslin & Newport, 1996), and of the social context within which vocalizations gain value for the child, an indirect form of ‘reinforcement’ with no specific pedagogical intent (Bloom & Esposito, 1975; Hsu & Fogel, 2001; Goldstein & Schwade, 2008). We return to these issues in chapters 2–5.
Chomsky's radical claims galvanized researchers interested in child language. It is fair to say that the present highly dynamic field of psycholinguistics largely developed in response to Chomsky's ideas, which both inspired supporters and stimulated sceptics or potential critics. The ideas themselves gradually became linguistic orthodoxy in mainstream linguistics, especially in the United States, but they have now begun to be widely questioned by cognitive linguists and others who adopt an ‘emergentist’ or ‘usage-based’ stance, as we will be doing here (Barlow & Kemmer, 2000; Bybee, 2001, 2010; Pierrehumbert, 2003a, 2003b; cf. also the critique from typologists Evans & Levinson, 2010 and the commentaries that follow in Behavioral and Brain Sciences, 32).
Analysis of an argument
Before moving to issues more directly concerned with phonological development we will give Chomsky's position a little more attention, since it has been so very influential for such a long time. Chomsky clearly sets out his position in a single sentence:
A consideration of the character of the grammar that is acquired, the degenerate quality and narrowly limited extent of the available data, the striking uniformity of the resulting grammars, and their independence of intelligence, motivation, and emotional state, over wide ranges of variation, leave[s] little hope that much of the structure of the language can be learned by an organism initially uninformed as to its general character… (1965, p. 58)
This exceptionally complex statement can be more readily understood if divided into its component premises (1–4) and the conclusion that follows:
In his later work Chomsky has repeated some or all of these arguments in numerous books and papers, with no significant change in his thinking on this matter. For example, Elbers and Wijnen (1992) cite a similar passage from Chomsky (1981a, p. 356), in which two additional premises are included: ‘… a rich and complex system of rules and principles [premise 1, Complexity] is attained in a uniform way [premise 3, Uniformity], rapidly [premise 5, Speed of acquisition], effortlessly [premise 6, Ease of acquisition], on the basis of limited and rather degenerate evidence [premise 2, Limited sample]’ (Chomsky, 1981, p. 356).
Premises 1, 5 and 6: Complexity of the adult language system, speed and ease of acquisition. In a chapter illustrating the effort that children can be seen to put into the many years of active learning that are actually required to achieve adult-like command and fluency, Elbers and Wijnen (1992) comment as follows:
The only constituents of Chomsky's contention that do not seem to have invited much criticism are the claim that language knowledge consists of ‘a rich and complex system of rules’ and the claim that language acquisition is effortless… Yet, it is precisely in the conjunction of these two claims that a confusion of professional and nonprofessional reasoning is evident. The rich-and-complex-system claim [Premise 1] is a professional judgment, based on detailed and extensive linguistic investigation. The no effort claim [Premise 6], however, is a layman's contention, based on casual and superficial impression rather than on careful observation and research. But propositions of such a differing status should not be combined; it seems just as mistaken to hold that development is effortless just because it seems effortless as it would be to hold that language itself is simple just because it seems rather simple to the ordinary speaker who is not a professional linguist. (pp. 339–340)
Accepting Chomsky's first premise as beyond argument, then, we can go on to look briefly at each of the others.
Premise 2: Limited and poorly structured sample. First, does the speech that children hear actually provide only a ‘limited’ and ‘degenerate’ sample of the grammar? This premise, later elaborated as the ‘poverty of the stimulus’ argument in support of UG (see Pullum & Scholz, 2002), has led to decades of research into infant-directed speech (IDS) and its consequences (for reviews, see Soderstrom, 2007; Gathercole & Hoff, 2007; we discuss the prosody of IDS in ch. 5). The general finding is that talk to children is unlike talk between adults: It involves much shorter sentences along with a much higher rate of repetition of all or parts of utterances. Although it may feature sentence fragments (phrases rather than sentences), it includes few or no false starts or self-corrections, the kind of language use that Chomsky presumably intended by the term ‘degenerate’ but which is more typical of high-level academic discourse than of talk to small children.
Thus, the input, although certainly providing a limited sample, is in many ways tailored for the child, given its adherence to topics that a small child might understand (and, more importantly, its typical orientation toward the child's own actions and focus of attention) and the necessarily repetitive lexicon of words and phrases that accompany child-rearing routines. And yet, despite all of this, it has proven difficult to find evidence that ‘tailored speech’ is actually necessary for first-language learning.
Premise 3: Uniformity of adult grammars. Next we can ask, are adult grammars ‘uniform’—that is, the same for all adults in a speech community? This premise has been severely challenged by variationist sociolinguistics, which evolved in the 1960s and 1970s (Labov, 1963, 1980; Weinreich, Herzog & Labov, 1968). The general understanding today is that variation (within and across speakers) is a basic characteristic of language in any speech community (Docherty, Foulkes, Tillotson & Watt, 2008; cf. also the more specific demonstrations of individual differences in grammar among adult native speakers in Street & Dąbrowska, 2010; Dąbrowska, 2012).
What are the consequences of this potential criticism of Chomsky's premises for our understanding of language development, however? Although Labov, Dąbrowska and others have made it clear by now that we cannot assume that all speakers have ‘the same grammar’, it is pertinent here to ask whether the inter-individual differences have their origins in ‘errors of transmission’ (i.e., from parent to child), also known as ‘imperfect learning’. There has been a line of thinking within generative linguistics that this is the case (e.g., Kiparsky, 1965). The data so far brought to bear on the question have not supported the idea that children's errors are a factor in language change (yet the notion continues to be maintained and asserted as fact, despite the lack of supporting evidence: e.g., Lightfoot, 1999; Blevins, 2004; Kiparsky, 2008): If they were, then the variability that gives rise to change might be traced back to infant mislearning.
In reality, although children do make errors of many kinds, they come in due course to faithfully reproduce the language they hear around them, shifting from parental models to peer group models as they grow older. (Furthermore, despite the superficial similarity between some developmental and historical processes, the younger children whose language remains incompletely mastered are hardly influential members of a community whose speech patterns could be expected to lead to variability and change in the adult language: Foulkes & Vihman, 2013). Thus, adult grammars are not well characterized as ‘strikingly uniform’ in the first place. Second, the variability across different adults' speech in a given community is more likely to be related to differing social experiences than to have its origins in developmental differences, so that the extent of similarity of adult grammars would seem to provide no real clue, one way or another, to the nature of children's language acquisition.
Premise 4: Individual differences in children. Do children's differences in ‘intelligence, motivation, and emotional state’ result in differential success in language acquisition? Since the vast majority of children free of developmental disorder do learn to talk on roughly the same time scale, and since no non-human animals have demonstrated linguistic comprehension or expression in any way comparable to what children have typically learned by age 3 or 4, something in the human inheritance—some biological advantage—must support the learning process. Chomsky postulates that this advantage is linguistically specific knowledge rather than some more general cognitive capacity unique to our species; it is that foreknowledge of linguistic principles that he refers to by the term UG. Here, however, we will look for broader biological bases for language, drawing on the unique human production capacities and the powerful learning mechanisms that have just begun to be understood (for a similar view based on somewhat different supporting capacities, see Kuhl, 2004). We discuss these learning mechanisms briefly below (and see ch. 2).
It is important to add that not all children do manage to learn language successfully: ‘specific language impairment’ (SLI) affects around 7% of American and Canadian 5-year-olds (Tomblin et al., 1997; Johnson et al., 1999). If it is possible for children to be ‘impaired’ specifically—that is, exclusively—for language, this might support a nativist position that accounts for the difficulty in terms of a blocking of access to UG (Van der Lely & Marshall, 2011). However, the ‘specificity’ of SLI is debatable, since both low non-verbal ability and certain behavioral disorders and sensory impairments commonly co-occur with the linguistic difficulties (Snowling & Hayiou-Thomas, 2010). Furthermore, recourse to ‘non-access to UG’ merely moves the explanation a step further: What is the basis for such blocked access? An alternative view is that SLI (like dyslexia) is rooted in impairment to more basic aspects of brain function. Language, with its particularly complex demands in terms of both representational structure and processing, may simply be the behavior most affected (Bishop & Snowling, 2004).
Premise 5: Speed of acquisition. Finally, we return to the question, how rapid is language acquisition? On the one hand, the question cannot really be answered, since learning is never complete but continues over the lifetime. On the other hand, we can assume that Chomsky was referring to the apparent speed with which children move from not speaking—before age 12 months or so, say—to communicating in complete sentences, in a way that is intelligible even to non-family members—typically by around age 4 or 5 years. Whether we see a period of four to five years for language learning as ‘fast’ or ‘slow’ is a matter of taste—but it is important to note that a great deal of learning takes place in the first year, in the ‘prelinguistic’ period, and much of it is indeed remarkably rapid. We return to this issue below.
The course of language development
Chomsky (1959) rightly discredited the behaviorist approach. Imitation and selective reinforcement are wholly insufficient to explain the kind of creative construction, overgeneralization and idiosyncratic rule or pattern formation that is repeatedly seen in studies of language acquisition. In contrast with the insistence on the ‘meticulous training’ thought to be needed for children to learn word meanings and syntactic patterns in the 1940s and 1950s (Chomsky, 1959, p. 39, n. 17), Chomsky pictured language development as maturationally controlled behavior, with only minimal support from the environment (exposure to a ‘trigger’). The idea derives from ethology, or the study of animals in their natural habitat (Chomsky, 1959, pp. 41ff.). What was completely original with Chomsky was the suggestion that the kinds of automatically triggered behaviors seen in certain animals could be extended to language acquisition. But what are the criteria for identifying such behavior?
Upon reaching the critical stage, the individual should begin to show the behavior automatically; the behavior should not be seen before that stage.
An appropriate trigger in the environment is also required—so that even once the ‘stage is set’, the behavior may fail to appear in the absence of appropriate environmental stimulation.
For language, a critical question is how the trigger might be identified or recognized: How is UG meant to interact with information in the speech signal? Furthermore, if language is a ‘maturational program’ that unfolds when appropriately triggered by exposure—like the automatic ‘following’ of the mother duck by ducklings—then changes should occur instantaneously, ‘across the board’, when children identify the appropriate cues in the input; initially inaccurate forms (errors) should be followed by more accurate forms (this is known as ‘linear advance’).
Such ‘across-the-board’ changes are sometimes reported for phonological development but that is not what is typically found, as will be amply demonstrated in the chapters that follow. The single most consistent empirical finding in cross-linguistic longitudinal observational studies of child language production is that the earliest word forms and uses of inflectional morphemes (most notably, irregular morphology) are more accurate than later forms. A ‘U-shaped curve’ (or nonlinear advance) characterizes development in phonology and inflectional morphology, with the early accurate forms generally being of high input frequency. Generalization (‘rule’ or ‘pattern learning’) begins a bit later and results in a period of ‘regression’ or lesser accuracy, as a productive pattern is extended beyond its domain in the adult targets. We will illustrate this at some length in later chapters and also consider ways of accounting for it.
Alternative approaches to Chomskyan biological foundations
To propose that children learn to speak without the benefit of specifically linguistic foreknowledge of what all languages have in common—in other words, without access to anything resembling UG—is not to return to the ‘empiricism’ of the first half of the twentieth century. Instead, at the end of chapter 2 we introduce a contemporary model of learning that posits a dual memory system (Squire & Zola, 1996). In this model two quite distinct memory mechanisms combine in a uniquely human way:
In our view, only the combination of attention-based learning with ‘incidental’ or unconscious learning results in a sufficiently powerful mechanism to account for language development. Such a dual mechanism alone is capable of deriving from input speech, in context, both systematic relations and arbitrary form–meaning associations, and of both retaining specific items and deriving from them generalized patterns or categories. Postulation of a dual memory system that functions in children as well as adults responds to Chomsky's paradox in a way that relies on ‘biological foundations’ but not on specifically linguistic foreknowledge (Bates & Elman, 1996).
In addition, in its focus on phonological development this book will describe the early perceptual capacities of infants and how they are shaped, among other things, by more slowly maturing production capacities. Here we will draw on the developmental approach of dynamic systems theory, which sees perception and action as inextricably interwoven, with relatively simple skills interacting to create more complex ones (Thelen & Smith, 1994). Both the dynamic systems approach and the dual memory system conceptualization provide a strong biological basis for language, one which is increasingly supported by evidence from neuroscience.
One basic issue for the nativist line of thinking is the ‘explanatory cost’ of positing LAD or UG without such neuropsychological evidence. This has been glossed over by linguists for 40 years, but the time limit on what Lindblom (1992) termed a ‘loan on cognition’ may have expired. Mention is sometimes made of the ‘language areas’ of the brain—but the specialization by hemisphere and by area (auditory, visual etc.) comes with development and use, not as a pre-fixed template, as is now well understood (see Mareschal et al., 2007). To continue to subscribe to Chomsky's proposals, in the present state of our knowledge, we would require answers to two fundamental questions: (1) Where is UG located in the brain, or what neurological systems support it, and what phylogenetic or evolutionary processes can be supposed to have led to its presence in the newborn brain? (2) Given innate linguistic knowledge in the form of UG, by what mechanism does exposure to input speech trigger choices between differing possible structures? These difficult questions should be addressed if Chomsky's radical proposal is to continue to form a basis for acquisition studies.
Phonological Development: Goals and Challenges
In her chapter ‘Where's phonology?’ Macken (1992) raised another issue that divides formalist and functionalist approaches. Macken suggested that the study of phonological development in the 1980s was largely concerned with the ‘phonetics of acquisition’, which seemed to imply that no abstract phonology is learned. As her title suggests, Macken is contrasting the period of babbling and first words, which she takes to reflect variable phonetic production, with a later period of structurally informed patterns, or ‘phonology’. The distinction between these terms is defined somewhat differently by different linguists, not all of whom endorse the need to make a distinction at all (see Critique and appreciation, ch. 10).
Phonetics and phonology
Let us consider the basic constructs of phonology. The most essential of these is no doubt contrast between phonological categories, which refers to the occurrence, in a single phonological context, of phonetically similar speech sounds that signal different meanings: These are the phonemes of structural linguistics (that is, for English, /k/ : /ɡ/, as in coat : goat, /d/ : /n/, as in pad : pan, /ɪ/ : /i/, as in pick : peak). Beyond that, the natural classes of speech sounds or phonemes reflect the universal occurrence of what could be called paradigmatic patterning, that is, the more or less symmetrical distribution of different places of articulation (labial, coronal, dorsal …) across the different manner classes (stops, fricatives, nasals …). These natural classes can be formally expressed through the construct of distinctive features. The distinctive features, in turn, reflect, among other things, the repeated uses, within a single phonological system, of the same articulatory gestures (see Clements & Ridouane, 2011). At the same time, prosodic structure and phonotactics refer to syntagmatic regularities, or constraints on the possible sequencing of speech sounds, at different levels of linguistic organization (syllable, word, phrase …).
In contrast to these elements of phonological structure, phonetics is generally taken to refer to gradient (‘sub-categorical’) knowledge of speech sounds and sequences based on frequency and contextual effects. Phonetic forms are variable by speaker, speech rate, speaker's emotional state and conversational intentions, etc., whereas phonology encodes the essential differences between forms, and constraints on their possibilities of sequential combination, in a structured system that supports differences in meaning. The terms ‘phonetic’ and ‘phonological’ are by no means used consistently in the developmental literature, however.
The interaction of perception and production
The effort to trace links between perception and production in order to arrive at a more complete understanding of phonological development received little empirical or theoretical attention until the 1990s (e.g., Vihman, 1991, 1993a), with studies adopting this perspective few and far between for several years thereafter. In recent years, however, evidence of links between perception and production in the first year have begun to appear (DePaolis, Vihman & Keren-Portnoy, 2011; Lewkowicz & Hansen-Tift, 2012; Yeung & Werker, 2013; Majorano, Vihman & DePaolis, 2013; DePaolis, Vihman & Nakai, 2013).
Within the field of infant speech perception interest in production has been minimal (cf., e.g., Jusczyk, 1997; Kuhl, 2004; Werker & Curtin, 2005). As regards adults, proponents of the motor theory of speech perception have assumed that there is a critical link between the two domains but have not been interested in its development, taking a biologically based connection between perception and production to be axiomatic: The perceptuomotor link underlying speech is ‘not a learned association…but innately specified, requiring only epigenetic experience [or some exposure to speech] to bring it into play’ (Liberman & Mattingly, 1985, p. 3). Not all motor theorists take this position, however. Studdert-Kennedy (1993), for example, emphasized that ‘a central function of perception in the infant is surely to guide production: by learning to listen the child learns to speak’ (p. 150); he went on to endorse the view that ‘with the discovery of correspondences between the sounds it hears and the sounds it makes, the infant begins to focus attention on the phonetic (articulatory) properties of native sounds’ (p. 152).
Our view is that the interaction of perception and production is key to an understanding of the early stages of phonological development. Accordingly, we begin by surveying in alternating chapters infant speech perception (chs. 3, 5) and vocal development (ch. 4); in our review of the transition to language use (ch. 6) we bring the two areas together as we sketch a model of their mutual influences and growing linkage over the course of the first year. A production-oriented approach will be evident in our account of the word-learning studies reviewed in chapter 7, and perception–production interaction will be considered again in our concluding chapter (11).
Cross-linguistic perspectives
Although studies of phonological development in a variety of languages have long been available, much of the literature continues to make reference to specific characteristics of the acquisition of English as if they were universal properties of child language development. To take just one example, the prevalence of monosyllables among the early word productions of English-learning children is often cited as the characteristic starting point for phonological development. However, even within the period of production of the first 50 words, over half the words produced by children acquiring French, Japanese and Swedish are disyllabic or longer, in accordance with the adult models that they are attempting, while monosyllables dominate the production of children acquiring English (Boysson-Bardies et al., 1992; Vihman, 1993a; see also ch. 8).
Studies of the earliest period of development have revealed influence from the ambient language on both perception and production, alongside strong evidence of universal perceptual biases and production constraints; we review the extent of ambient language influence on both early perception and speech processing (chs. 3 and 5) and production (chs. 4 and 6). Furthermore, studies of children acquiring two languages from infancy provide additional insight into cross-linguistic similarities and differences in both perception and production (ch. 8).
The significance of individual differences
Each child must individually forge a path to language; this is clear from production studies, which focus on individuals, although it goes largely unremarked in perception experiments, which generally report only group results. Every careful production study of more than one child reveals a range of differences and individual strategies. Macken (1992) goes so far as to propose that
the central acquisition mechanism is a constrained hypothesis formation mechanism…where the linguistic constraints are not so restricted as to result in invariance but, rather, so closely replicate the formal constraints on languages in general as to render any set of ten or twenty learners (of even the same language) a virtual typological study of language parameters. (p. 250)
We would argue that both biological predispositions and salient aspects of the ambient language constrain the child's initial progress in language acquisition; nevertheless, multiple individual factors enter into the child's approach to language, as regards both timing and manner of acquisition. Attempts to reduce this individual variation to a single pair of contrasting learning styles (such as referential vs. expressive, analytic vs. holistic) have not, on the whole, yielded definitive or generally satisfying results (Nelson, 1981; Bates, Bretherton & Snyder, 1988; Lieven, Pine & Dresner Barnes, 1992; Bates et al., 1994). Instead, it seems that, within the given constraints, children vary widely in maturation rate and individual disposition along several parameters, both social (interest in communicative interaction) and linguistic (sensitivity to vocal patterns, motoric skill). Even more important perhaps is apparent variation in the child's deployment of the cognitive elements of acquisition: Attention and effort must be integrated through the internalized representations of adult forms and of the child's own vocal capacities; these must then be interrelated and molded into a viable set of production patterns which can gradually be brought into line—over a period of two to three years at the very least—with the target adult system.
Methodologies: Data Sources and Theoretical Perspectives
As increasingly rapid technological changes have made new methods available for the study of infants the field of phonological development has diversified dramatically. Today, such methods include audio and video recordings, with increasingly sophisticated ways of transcribing and coding while simultaneously listening to the audio and watching the video, often with further support from a view of the sound wave or full spectrogram, affording the option of carrying out acoustic analysis alongside segmental transcription and with due consideration of the situational context; audio-only or cross-modal experimental studies using the head turn and preferential looking paradigms, eye-tracking for more automatic analysis of infant responses to audio-visual correspondences, and event related potentials (ERPs), which permit investigation of the neurophysiological response to critical stimuli on a fine temporal scale (see ch. 7). Additional new techniques for gaining understanding of the infant brain include near-infrared spectroscopy (NIRS: see Gervain et al., 2011) and magnetoencephalography (MEG), which provides a way of combining temporal and spatial information. However, no studies based on these techniques, only recently adapted for use with infants, will be covered here.
A conference in 2001 brought together psychologists who use behavioral experimental techniques to study speech processing and word learning with developmental linguists whose investigations are largely based on formal phonological analysis of production data within a deductive theoretical framework (see ch. 9). Peperkamp (2003) summarizes the goals of the conference, expressing her disappointment at the lack of experimental tests of phonological hypotheses:
Since the founding work of Chomsky & Halle (1968), linguists have made detailed proposals concerning phonological representations and the derivations by which abstract underlying forms are mapped onto concrete surface forms. Most aspects of these proposals have not been looked at in experimental work on phonological acquisition. Furthermore, within the framework of generative grammar, it is assumed that there is a common core of phonological knowledge across speakers of all human languages. This common core is typically supposed to be innate (even though there is no a priori reason that universals could not emerge during the course of acquisition). Whether innate knowledge is used or not during phonological acquisition…remains to be investigated experimentally. (p. 88)
Peperkamp goes on to note that
the reverse is equally true: Linguistic approaches to phonological acquisition have been little concerned with experimental evidence regarding early phonological development…[Due to emphasis on production, these studies] sidestep the fact that before uttering their first words, infants acquire important parts of their native language's phonology. (p. 93)
These comments invoke all three of the main lines of current empirical research into phonological development:
1 Individual or small group production studies focus on one or more infants, whether family-member diary or outsider observational studies. These time-honored research methods have led to the formulation of many of the issues that continue to concern investigators. The results are generally based primarily on phonetically transcribed word lists, with more or less attention to variability across tokens. To counter the increasing tendency to disregard earlier work (cf. Menn's 2006 subtitle, ‘Making sure that old data survive new theories’) we list in Appendix 1 all of the readily available studies of this kind that we are aware of, categorized to indicate the language(s) to which the participant children were exposed, the number, age and lexical range of those children, the data source and whether or not a full list of child word forms is included.
Appendix 1 shows that although only three studies of phonological development were published between 1938 and 1967, in the following half-century 13 new studies of 1 to 10 children were published in every decade (Table 1.1; 24 of these studies include full word lists for one or more children). Of the 65 studies listed, however, 20 focus on 28 children learning English only. Germanic languages heavily dominate the picture, with 8 studies of the monolingual acquisition of Dutch or German accounting for an additional 19 children; Finnish is the only other language to include a substantial number of children. Altogether, phonological descriptions and/or analyses of the word patterns of over 100 children are available, covering the period of the first 50 to several hundred words in a total of 25 different languages (counting European and American variants separately). Of these studies, 16 focus on bilingual children (see ch. 8).
Table 1.1 Small group and case studies, 1938–2013
In addition, in the past few years Yvan Rose, in collaboration with Brian MacWhinney, has designed a software program, PHON, to support phonetic transcription and phonological analysis in conjunction with the well-established CLAN (Computerized Language ANalysis) programs that support analysis of data stored in the CHILDES database (MacWhinney & Snow, 1985). PHON supports multimedia data linkage, segmentation, multiple-blind transcription and systematic comparisons between target (model) and actual (produced) phonological forms (Rose et al. 2006; Rose & MacWhinney, 2013). The associated database, PhonBank, complements CHILDES; it currently includes data from 19 languages, six of them not represented in Appendix 1 (Berber, Cree, Greek, Indonesian, Swedish, Taiwanese). This new resource should make it far easier to conduct well-grounded cross-linguistic investigations of patterns of phonological development in the future.
2 Larger group studies of infant speech production, whether cross-sectional or longitudinal or a mix of the two, are typically designed to establish norms and/or to test phonetic and phonological theories of development. Recently, a number of PhD theses or large funded research projects have been dedicated to studies of phonological development in 10 or more children acquiring a range of different languages, including Finnish (Kunnari, 2000), French (Braud, 2003; Wauquier & Yamaguchi, 2013), Greek (Tzakosta, 2004) and Italian (Majorano, 2005). Whereas the case studies generally emphasize individual differences, the group studies are often intended to identify universal constraints on early development or to support or challenge theories of phonological development (Fikkert, 1994; Levelt, 1994; Kehoe, 1998; Kehoe & Stoel-Gammon, 1997a, 1997b, 2001; Wauquier-Gravelines, 2005). Where the study is longitudinal and provides phonological analyses of the word forms of individual children it has been included in Appendix 1.
3 Experimental studies of infant perceptual responses to speech address the nature of development in perceptual discrimination, segmentation, word recognition and comprehension over the course of the first two years. These group studies use a variety of techniques, including both head turn paradigms and, in more recent years, ERPs (ch. 7). Experimental studies of infant responses to speech have dominated the literature in phonological development for the last 15 years. The studies are generally based on groups of infants who may or may not receive training or conditioning with particular stimuli as part of the experimental procedure. In the first 15 years or so after the first such study of infant speech perception (Eimas, Siqueland, Jusczyk & Vigorito, 1971) infant capacities for discrimination of speech sounds were explored in some depth (ch. 3), with a primary focus on English. Later research turned to advances in knowledge of the ambient language and the role of infant speech perception in laying the groundwork for learning syntax, segmentation and statistical learning (ch. 5).
An important benefit of recent work in both perception and production has been the beginnings of an expansion from a heavily anglocentric field to the more common collection of data from other languages. Unfortunately, studies combining more than one methodological approach remain rare (although there have been some recent attempts to derive individual child measures from group perception studies: Tsao, Liu & Kuhl, 2004; DePaolis et al., 2011). Each methodology has also tended to have its own preferred vocabulary, framework of interpretation and conclusions.
Overview
The central goals of this book are to survey what has been learned about phonological development and to raise questions for further study. We will interpret these goals broadly, however, going well beyond the phonological rules or processes that were the primary concern of linguists interested in child phonology when it first emerged as a field in its own right in the 1970s. Placing our topic within the wider domain of inquiry into the possibility of a system arising (or ‘self-organizing’) out of no system, we will adopt a functionalist approach. We will be concerned with tracing the beginnings of phonology (and of language) in the infant's perceptual capacities, now well understood to be quite remarkable already at birth or soon thereafter, in the growing repertoire of vocal resources of the first year, and in the emergence of a link between perception and production. We will then consider the transition to language use, endeavoring to identify critical communicative and cognitive developments that permit the construction of a system of interconnected sound patterns along with a dawning understanding of the nature of naming and reference.
Before we embark on a roughly chronological account of phonological development in the first two years (focusing mainly on perception in chapters 3, 5 and 7 and production in chapters 4 and 6), however, chapter 2 provides a broad outline of infant development in the first 18 months, drawing on studies of the development of attention and joint attention, for example, to supplement a focus on more specifically language-related concerns.
Chapter 8, on bilingual phonological development, reviews another lively area of current research, language differentiation and category formation in bilingual infants, as well as the older and less dynamic area of bilingual child production, which has only recently begun to be enriched by studies of small groups of children in addition to the classic case study. Although bilinguals constitute the majority of the population, empirical study of infants raised bilingually is made more challenging by the many variables that affect bilingual language processing and use—such as extent of exposure to each language, in the home and in the community; differences in the phonological or rhythmic distance between the child's languages; changes in language dominance with changes in input conditions, and so on. In the light of the practical and theoretical difficulties involved, most researchers in the perception subfield, in particular, tended until recently to see establishment of the characteristics of monolingual development as an essential ‘baseline’ before the findings of bilingual studies could be fairly evaluated.
We consider the theoretical models that have been put forward to account for phonological development in chapters 9 and 10. These chapters are delayed to that point so that the reader initially uninformed about infant development could come to them prepared to use the information presented so far—that is, evidence regarding the nature and course of development and change—to better appreciate the strengths and limitations of the competing theoretical models. Finally, chapter 11 briefly reconsiders two of the main themes of the book, the critical interaction of perception and production and the effects of lexical advance, and of language use, on further development.
Chapter 2
Precursors to Language
The First 18 Months of Life
The Development of Linguistic Form and Function
1 Early Capacities: Birth to 2 Months
2 Early Capacities: 2 to 4 Months
3 Early Capacities: 4 to 6 Months
4 First Advances: 6 to 9 Months
5 Bringing the Strands Together: 9 to 12 Months
6 Transition to Language Use: 12 to 18 Months
Learning Mechanisms
Summary: Precursors and the Transition to Language Use
One goal for the field of language development is to arrive at an ecologically appropriate view of the infant and of the environment accessible to the developing infant brain—a brain that is in the process of maturing, learning, and integrating the various patterns or regularities to which it is exposed, including the experience of the infant's own actions and their perceptible consequences (Thelen & Smith, 1994; Campos et al., 2000). An account of infant language learning that aspires to be truly developmental should avoid predicating preformed adult-like linguistic or communicative goals for the infant. That is, there is no reason to believe that the infant is in any sense to discover the structure of language or how to most efficiently communicate information, as is sometimes implied. As Locke (1993) expressed it,
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
Lesen Sie weiter in der vollständigen Ausgabe!
