Acquiring Sociolinguistic Variation 9027265283, 9789027265289

Citation preview

Acquiring Sociolinguistic Variation

John Benjamins Publishing Company

Studies in Language Variation

Edited by Gunther De Vogelaer Matthias Katerbow

20

Acquiring Sociolinguistic Variation

Studies in Language Variation (SILV) issn 1872-9592 The series aims to include empirical studies of linguistic variation as well as its description, explanation and interpretation in structural, social and cognitive terms. The series will cover any relevant subdiscipline: sociolinguistics, contact linguistics, dialectology, historical linguistics, anthropology/anthropological linguistics. The emphasis will be on linguistic aspects and on the interaction between linguistic and extralinguistic aspects — not on extralinguistic aspects (including language ideology, policy etc.) as such. For an overview of all books published in this series, please see http://benjamins.com/catalog/silv

Editors Frans Hinskens

Paul Kerswill

Jannis K. Androutsopoulos

Peter Gilles

K. K. Luke

Arto Anttila

Barbara Horvath

Gaetano Berruto

Brian Joseph

Paul Boersma

Johannes Kabatek

Jenny Cheshire

Juhani Klemola

Gerard Docherty

Miklós Kontra

Peter Auer

Universität Freiburg

Meertens Instituut & Vrije Universiteit, Amsterdam

University of York

Editorial Board University of Hamburg Stanford University Università di Torino University of Amsterdam University of London Newcastle University

Penny Eckert

Stanford University

William Foley

University of Sydney

University of Luxembourg University of Sydney The Ohio State University Universität Zürich

University of Tampere Károli Gáspár University of the Reformed Church in Hungary

Bernard Laks

CNRS-Université Paris X Nanterre

Maria-Rosa Lloret

Universitat de Barcelona

Volume 20 Acquiring Sociolinguistic Variation Edited by Gunther De Vogelaer and Matthias Katerbow

Nanyang Technological University, Singapore

Rajend Mesthrie

University of Cape Town

Pieter Muysken

Radboud University Nijmegen

Marc van Oostendorp

Meertens Institute & Leiden University

Sali Tagliamonte

University of Toronto

Johan Taeldeman University of Gent

Øystein Vangsnes

University of Tromsø

Juan Villena Ponsoda Universidad de Málaga

Acquiring Sociolinguistic Variation Edited by

Gunther De Vogelaer Westfälische Wilhelms-Universität Münster

Matthias Katerbow Philipps-Universität Marburg

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

doi 10.1075/silv.20 Cataloging-in-Publication Data available from Library of Congress: lccn 2017018298 (print) / 2017037890 (e-book) isbn 978 90 272 3387 5 (Hb) isbn 978 90 272 6528 9 (e-book)

© 2017 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Company · https://benjamins.com

Table of contents

Chapter 1 Bridging the gap between language acquisition and sociolinguistics: Introduction to an interdisciplinary topic Gunther De Vogelaer, Jean-Pierre Chevrot, Matthias Katerbow and Aurélie Nardy Chapter 2 The effects of exposure on awareness and discrimination of regional accents by five- and six year old children Erica Beck Chapter 3 How do social networks influence children’s stylistic practices? Social mixing, macro/micro analysis and methodological questions Laurence Buson Chapter 4 Child acquisition of sociolinguistic variation: Adults, children and (regional) standard Dutch two-verb clusters in one community Leonie Cornips Chapter 5 Acquiring attitudes towards varieties of Dutch: A quantitative perspective Gunther De Vogelaer and Jolien Toye Chapter 6 What is the target variety? The diverse effects of standard–dialect variation in second language acquisition Andrea Ender

1

43

65

91

117

155

vi

Acquiring Sociolinguistic Variation

Chapter 7 The relationship between segregation and participation in ethnolectal variants: A longitudinal study Charlie Farrington, Jennifer Renn and Mary Kohn Chapter 8 Socializing language choices: When variation in the language environment supports acquisition Anna Ghimenton

185

213

Chapter 9 Language acquisition in bilectal environments: Competing motivations, metalinguistic awareness, and the Socio-Syntax of Development Hypothesis 235 Evelina Leivada and Kleanthes K. Grohmann Chapter 10 Acquisition of phonological variables of a Flemish dialect by children raised in Standard Dutch: Some considerations on the learning mechanisms 267 Kathy Rys, Emmanuel Keuleers, Walter Daelemans and Steven Gillis Chapter 11 Developmental sociolinguistics and the acquisition of T-glottalling by immigrant teenagers in London Erik Schleef Author index Subject index

305

343 345

Chapter 1

Bridging the gap between language acquisition and sociolinguistics Introduction to an interdisciplinary topic Gunther De Vogelaer, Jean-Pierre Chevrot, Matthias Katerbow and Aurélie Nardy

University of Münster / Laboratoire LIDILEM – Université Grenoble Alpes / Philipps-Universität Marburg / Laboratoire LIDILEM – Université Grenoble Alpes

Despite repeated calls for in-depth research, the acquisition of patterns of sociolinguistic variation has long been an underinvestigated topic both in sociolinguistics and in language acquisition research. With the exception of a few exploratory studies, most notably Labov (1964), it has long been rare for sociolinguistic research to focus on non-adults, whereas most research on language acquisition tended to take place in a sociolinguistic vacuum (see, e.g. Mills 1985: 142 and Labov 1989: 96 for statements to this effect). Over the last few years, however, the situation seems to be changing. Two reasons may be given for this: first, and quite trivially perhaps, technical advancements are making it possible to gather, store and explore data in cheap and efficient ways, providing researchers with the necessary data to conduct empirically sound research on the topic. And second, parallel to a paradigm shift from rule-based to usage-based conceptions of grammar, linguistic variation has moved into the centre of the attention of theoretical linguistics. As a result, the acquisition of variation can now be considered an ‘emergent topic’ in research on language variation in general. The aim of this book is to offer a state-of-the-art of current research on the topic, thereby focusing on two particular objectives: (1) the acquisition of sociolinguistic variation presents itself as an interesting research topic for sociolinguists and psycholinguists working on acquisition, but also for a broad range of other sub-disciplines of linguistics, including historical linguistics, dialectology, and for researchers working in different theoretical frameworks. This book aims at bridging the gap between these disciplines and frameworks and allowing an interdisciplinary perspective on the topic; and (2) in order to enable cross-­linguistic comparison, the book wants to bring together research carried out

doi 10.1075/silv.20.01dev © 2017 John Benjamins Publishing Company

2

Gunther De Vogelaer et al.

in different sociolinguistic constellations, as most obviously found in different language areas or different countries. Keywords: language acquisition, psycholinguistics, language variation, sociolinguistics, cognition

1. Introduction A few exceptions notwithstanding, it has long been rare both for sociolinguistic research to focus on non-adults and for research on language acquisition to pay attention to sociolinguistic variation. This is probably due to the way sociolinguistics and language acquisition research became prominent fields in linguistics, which, roughly speaking, dates back to the 1960s. The then emerging linguistic paradigms, i.e. the Chomskyan paradigm in language acquisition research and the Labovian one in sociolinguistics, became dominant and had an enormous impact for the decades to come, without really stirring the interest in the way patterns of sociolinguistic variation are acquired. This is in a way remarkable, as basic assumptions in both paradigms implicitly acknowledge that much is to be gained from including language variation in language acquisition research, or from an acquisitionist perspective on language variation. Linguistics’ interest in language acquisition was to a large extent stirred by the advent of the generative paradigm. Indeed one of the lasting contributions of the Chomskyan revolution is its focus on competence and ‘I-language’, which provided linguistics with a profoundly cognitive orientation, and linked aspects of linguistic structure to innate characteristics of the human mind. Among other things, acquisition research has been used to determine the structural status of linguistic elements in grammars. One common assumption is that the order in which parts of a grammar are acquired, reflects their position in the adult system: basic and/ or important traits are acquired before peripheral elements (cf. Mills 1986). In generativist thinking, variation has long been considered quasi unimportant for linguistic theory. This is obvious for instance in Chomsky’s homogeneity axiom. Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech-community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his knowledge of the language in actual performance. (Chomsky 1965: 3)

The homogeneity axiom should be read as a plea for descriptive abstraction rather than a principled exclusion of variation, however, and does not entail any claim to



Chapter 1. Introduction

the effect that linguistics should only be concerned with standard varieties. In fact, few generative linguists would dispute the idea that non-standard languages, such as dialects, have their own grammars and can be studied as proper I-languages, too (see De Vogelaer and Seiler 2012; De Vogelaer and Klom 2013 for discussion). Knowing how profoundly non-standard grammars can differ from standard languages and each other, the enormous potential of research into the acquisition of non-standard varieties in terms of broadening the empirical scope of linguistics must be acknowledged, which has been done in fields like dialect phonology and dialect syntax. Apart from insight into the structure of grammar, a variationist perspective also provides a better understanding of the nature of acquisition processes. For instance, studies on second dialect acquisition have been interpreted in support of the validity of the Critical Period Hypothesis for second language acquisition (e.g. Payne 1976; Chambers 1992). More recently, investigations on language learners who have been exposed to variation in the input are providing answers to the question of the extent to which regularities in language are internalized as ‘rules’, as assumed within the generative tradition, or rather derive from concrete linguistic experiences stored in memory, as assumed in exemplar-based or item-based grammars (e.g. Roberts 1994; Foulkes, Docherty, and Watt 2001; Díaz-Campos 2004; Nardy, Chevrot, and Barbu 2013). Some relevant case studies are discussed below. The rise of sociolinguistics is strongly associated with Labov’s work, which is deeply rooted in historical linguistics as well, as becomes evident in Weinreich, Labov, and Herzog (1968). The main methodological contribution of the Labovian paradigm is probably the apparent-time method, which allows to confidently detect processes of ongoing change within locations on the basis of quantitative data. The apparent-time method is only suitable to observe change-in-progress, however, because it crucially assumes a fixed adult grammar, which locates the pivotal moment for change in the intergenerational transmission of language, that means: in childhood and adolescence (cf. Labov 2007 on ‘transmission’). The apparent-time construct has been tested empirically and is proven to be, by and large, valid (see Sankoff and Blondeau 2007 for discussion). To reveal the underlying dynamics of language change, and to find empirically solid answers to questions such as the how and why of language change, therefore, it is important to conduct research on how patterns of sociolinguistic variation are acquired and differentially reproduced by non-adults, rather than restricting oneself to an investigation of diachronic sequences of synchronic stages (cf. Katerbow 2013). This is especially obvious in cases where the non-adults’ language is fundamentally different from the adult models, as in situations of extensive language contact, a case in point being the so-called New Towns, where children and adolescents appear to blend a new variety based on varieties spoken by older generations (e.g. Kerswill and Williams 2000).

3

4

Gunther De Vogelaer et al.

Sitting at the interface between language acquisition research (or the wider domain of psycholinguistics) on the one hand, and sociolinguistics on the other, research on the acquisition of sociolinguistic variation presents itself as a topic allowing or even necessitating an interdisciplinary perspective. In addition, such research will reveal relevant findings for, among others, historical linguists, dialectologists, contact linguists, and many other subdisciplines in the field. This book’s first goal is therefore to help bridge the gap between these disciplines. Second, in order to enable cross-linguistic comparison, the book wants to bring together research carried out in different sociolinguistic constellations, as most obviously found in different language areas or different countries. Section 2 of this introductory chapter provides an overview of the most important literature, including a number of exceptions to the general characterisation that there has been little interaction between sociolinguistics and language acquisition research. Section 3 then discusses recent tendencies that have stirred research interest in the acquisition of sociolinguistic variation, to the extent that, starting around the year 2000, it emerged as an important research topic. The chapter ends with an outlook on the remainder of the book, a characterisation of present-day research and some perspectives for future research in Section 4. 2. Research on the acquisition of variation before 2000 2.1

General characterisation: Three important lines of research

Among the oldest texts describing and evaluating language variation in children are descriptions by educators reporting on the linguistic situation in which they had to work, or instructions by legislators or governments. As a result of standardisation processes, formal education often required children to learn to write in a language variety quite different from their native variety, which, in many historical cases, was conceived of as a different language. As oral endoglossic standards developed, alternatives to regional dialects for everyday interaction became available, and were propagated via the education system. Even though the number of everyday speakers of standard languages in Europe is estimated to have been minor before 1900 (Auer 2005: 17), 19th century documents clearly illustrate that non-­standard varieties like regional dialects generally met with hostility in education. For instance, Hollingworth (1989: 294–295), in a discussion of the education system in 19th century Great-Britain, documents teachers’ attitudes towards dialects as likely to interfere with learning the standard and associated with ignorance and social inequality. In addition, governmental policies to use the education system as a means to establish linguistic unity in the area under their rule were widespread



Chapter 1. Introduction

across Europe. An infamous document in this respect is Abbé Gregoire’s (1794) Rapport sur la nécessité et les moyens d’anéantir les patois et d’universaliser la langue française (Report on the necessity and the means to eradicate dialects and universalise the French language), which refers to mandatory training in French as a means to oust the linguistic variation that was seen as a threat to France’s unity and its adherence to the ideas of the Enlightenment. 1 Similarly, 19th century Prussian education policies brought about a shift from Dutch to High German as the standard language in the parts of the Lower Rhine area under their rule, which is seen as the main cause for the subsequent loss of Low German dialects in everyday interaction (Cornelissen 1998). Interestingly, the 19th century also saw a number of pleas in favour of the use of dialects as a language of instruction, with standard varieties being considered too unfamiliar to pupils or even “the new Latin” (Rosenberg 1989: 62), and dialects standing nearer to the ‘essence’ of language (also in a historical sence) and indexing purity and moral simplicity (Hollingworth 1989: 296–299). Hollingworth argues that most 19th century scholars advocating tolerance towards dialects did so on linguistically doubtful, essentialist grounds, indeed in a similar vein as many propagators of the standard, and are best considered manifestations of the Romantic movement (see Ammon 1989: 113–114 and Geeraerts 2003 for historical descriptions and Blom and Gumperz 1972 for a more recent case study from Norway). The same era also saw an increased focus on dialects in linguistics, and for the first time systematic attention is demanded for regional variation in the acquisition of sound systems (Schultze 1880: 45). 2 We are not aware, however, of any substantial empirical work or theorising on language variation in children from that era. It seems to be Bernstein (1971–1975) who, almost a century later, has to be credited for raising the interest in sociolinguistic variation as found in children and adolescents. Bernstein hypothesized that academic underachievement of lower social classes is due to their exclusive use of a so-called ‘restricted code’, whereas middle and upper class also have access to an ‘elaborated code’. Although Bernstein resisted interpretations equating the restricted code with dialect, his work became 1. In his list of about 30 dialects Gregoire includes both Gallo-Romance varieties (Picardian, Normandian, Champenois, Bourguignon), Romance varieties more remotely related to French (like Occitan, Catalan), and non-Romance varieties like Breton, Alsatian, Flemish and Basque. 2. Schulze formulates it as follows: “es müssten die Beobachtungen nicht allein auf die Kinder verschiedenster Nationen ausgedehnt werden, sondern auch der Einfluß berücksichtigt werden, welchen auf die Entwicklung der Sprache eines Kindes der besondere Dialekt seiner Umgebung ausübt.” Translation: The observations should not merely be extended towards children from a variety of nations, but the influence should be taken into account that the particular dialect of a child’s surroundings carries out on the development of its language.

5

6

Gunther De Vogelaer et al.

associated with the ‘deficit hypothesis’, which locates the reasons behind the lower classes’ academic underachievement in the socialization of the pupils rather than in the educational system (see Jones 2013 for discussion), and which was discarded by sociolinguists in favour of the ‘difference hypothesis’, which places central the “doctrine of equality” (Cheshire et al. 1989: 7) of different language varieties. Bernstein’s work has inspired follow-up studies especially in educational research and also provoked intense discussion in sociolinguistics (Cheshire et al. 1989: 5–7), in which relevant hypotheses are formulated with respect to how sociolinguistic variation is acquired. 3 In addition, the rise of the generative paradigm raised the interest in how aspects of grammar were acquired, also of non-standard varieties. Both in the educational and in the linguistic strands, a number of recurrent themes can be observed. As most research deals with situations in which a clear native variety can be distinguished, a logical first issue is the attainment in the second dialect (§2.2), which may be another regional variety (a situation dubbed ‘naturalistic’ second dialect acquisition by Siegel 2010) or a standard (in ‘educational’ second dialect acquisition). Second, along with the transmission of the varieties themselves, attitudes towards varieties are passed-on between generations (§2.3). As language variation often results from or lies at the cause of language change, a third logical question addresses the role children and adolescents play in initiating or pushing forward processes of language change (§2.4). 2.2

Second dialect acquisition and other attainment issues

A substantial part of the research on the acquisition of sociolinguistic variation has focused on ‘attainment issues’, detecting whether or at which age children reach full proficiency in non-native varieties they are exposed to. Depending on whether such a non-native variety is learned in a school setting (the typical case being a standard variety) or in unmonitored conditions (typically in situations of speaker mobility), Siegel (2010) distinguishes between “educational” and “naturalistic” processes of second dialect acquisition. In many cases, however, the variety used in education plays a role in everyday interaction as well, so it is hard to maintain a clear-cut distinction between the two types. In line with Bernstein’s work in the 1960s and 1970s, it has been observed time and again that mismatches between children’s native variety on the one hand, and standard or prestige varieties that tend to be dominant in school contexts on the

3. Despite the fact that Bernstein is mentioned in most contemporary handbooks in the field, sociolinguistic follow-up studies engaging in proper testing of his theories, about which most sociolinguists were highly critical, appear to be rather scarce (see Kerswill 2007: 58–59 for discussion).



Chapter 1. Introduction

other, may negatively affect children’s chances on educational success. Differences between the varieties at play pervade the linguistic system, ranging from obvious domains like phonology and lexis that are most typically associated with non-standard language usage, over morphology and syntax, to pragmatics (Siegel 2010: 7–11). Example phenomena of the latter type include the organisation of narratives, topic development in conversation, or turn-taking mechanisms (see Hoyle and Adger 1998: 9–10 for discussion). Varieties may be in a situation of diglossia, as in the German speaking part of Switzerland (e.g. Stern 1988; Häcki Buhofer and Burger 1998; Ostermai 2000; Berthele 2002), or constitute a diaglossia, i.e. displaying a spectrum-like transition and involving intermediate varieties. Despite the fact that such contexts of diaglossia are believed to be the dominant linguistic repertoire in Europe and North-America (Auer 2005), and linguistic differences between varieties are often well-documented, the interest in how diaglossic patterns of variation are acquired seems to be quite recent (e.g. case studies like Rys 2007; Smith, Durham, and Fortune 2007; Katerbow 2013). Several approaches have been developed to deal with variation in the school context, some of which have focused on remedying linguistic disadvantages emanating from a non-standard linguistic background. Most educational problems associated with a non-standard background, however, appear not to result from any purely linguistic barrier. Of course linguistic differences between the home and the school variety can theoretically cause interference effects and misunderstandings (e.g. when non-standard speaking children have to acquire an orthography based on the standard variety; see Ammon 1989 for an illustration). 4 But most investigations on the influence of dialect backgrounds in school converge on the conclusion that educational underachievement rather results from a negative social evaluation of non-standard language usage (e.g. Goodman and Goodman 1981; Rosenberg 1989; Kroon and Vallen 2009). In this respect, instead of (merely) remedial approaches targeting non-standard speaking pupils, sociolinguists have developed Language Awareness programs geared towards furthering knowledge about and acceptance of linguistic diversity in both teachers and pupils. Even though most, if not all children are exposed to sociolinguistic variation from their infancy onwards, certain patterns of variation are acquired earlier than others due to developmental factors. In an attempt to take account of the importance of the age factor in the acquisition of sociolinguistic variation, Labov (1964)

4. The reverse, however, is also attested: Allerup et al. (1979) detect regions where children acquire better orthographical skills due to their dialect competence. In addition, bidialectism, i.e. proficiency in both a standard and a non-standard variety, is often found to correlate with eduational success (Rosenberg 1989).

7

8

Gunther De Vogelaer et al.

proposes a six-stage-model in which it takes children until late adolescence to acquire “the full range of spoken English”. Stage 1, the basic grammar (before year 5): mastery of the main body of grammatical rules and the lexicon of spoken English, normally achieved under the linguistic influence of the parents Stage 2, the vernacular (year 5–12): use of the local dialect in a form consistent with that of the peer group Stage 3, social perception (early adolescence): progressive awareness of the social significance of the dialect Stage 4, stylistic variation (late adolescence): learn to modify speech using standard forms in formal situations Stage 5, the consistent standard: ability to switch to a consistent style of speech and maintain that style, acquired primarily by the middle class groups Stage 6, the full range of spoken English: complete consistency in a range of styles appropriate, mostly achieved only by college educated persons with special interest in speech In Labov’s model, even in situations of diaglossia typical for most parts of the US, children essentially acquire different varieties successively: it takes until adolescence before children are able to accommodate their language to the situation. Chambers (2003: 174), in contrast, assumes that different variants (and, hence, also varieties) can be acquired simultaneously, and asserts that “there are no studies indicating a time gap between the acquisition of grammatical competence and the development of sociolinguistic competence.” Despite Chambers’ claim, there are empirical studies supporting either view. For instance, Rys (2007) documents a radically improving dialect proficiency in Flemish adolescents, even in those reporting that the dialect is their native variety. Ervin-Tripp (1973) and Roberts (1994), in contrast, provide illustrations of sophisticated sociolinguistic competences in young children. In a discussion of two variants from a variety of Scots, Smith, Durham, and Fortune (2007: 89–90) attest both scenarios: the use of third person -s is acquired as a variable rule right from the start, whereas for the hoose variable, the youngest children show a categorical usage of the standard variant (see also Cornips, this volume, for a related example). Neither linguistic level, complexity, or age appears to explain this contrast. While we may hypothesize that such differences have to do with the structure of the input to which the children and adolescents are exposed, most investigations do not analyse the input in enough detail to be able to test such a hypothesis. From studies analysing young children who are exposed to a diversity of influences, it is known that they model their language not only after that of their parents, but also, and perhaps predominantly, on the language they

Chapter 1. Introduction 9



observe in their peer group (see Nardy, Chevrot, and Barbu 2014 for results and extensive discussion). A different type of second dialect acquisition has several non-standard varieties as its target. This type often involves migration from one dialect area to another on the side of the speaker. Processes of this type have been surprisingly well-studied (cf. early studies by Payne 1976 and Chambers 1988; see Siegel 2010 for a comprehensive overview). Most research seems to be aiming, implicitly or explicitly, at a comparison with aspects of second language acquisition. For instance, Payne (1976) at length addresses the question to what extent there is a critical period within which native-like proficiency of the second dialect can be reached. Siegel (2010: 96–100), in his comprehensive discussion of the literature on second dialect acquisition, points out that, like in second language acquisition research, such a critical period has been stated to end at different ages between 7 and the mid-teens, depending on the type of data that have been studied and the context in which the investigation is carried out. The literature is equally ambiguous regarding the nature of such age-of-acquisition effects as the literature on second language acquisition, where explanations have been formulated in biological (e.g. loss of brain plasticity), social-psychological (e.g. a generally decreasing integrative orientation) and cognitive terms (e.g. a decreased ability to establish new grammatical categories). Apart from contributing to the evaluation of the Critical Period Hypothesis, research on second dialect acquisition has proven a rich source for hypotheses on factors influencing language learning processes. On the speakers’ side, relevant factors include length of residence, gender and class-­related factors, and attitudes and social identity (see Siegel 2010; Ender, this volume; and Schleef, this volume for discussion). Important linguistic factors include the linguistic distance between the varieties involved, rule complexity and (several operationalisations of the notion) salience. As for the very nature of the learning process, Chambers’ (1992) influential account assumes that rule-based learning is involved, whereas a number of more recent studies have proposed more gradual, lexical or exemplar-based learning (see also Rys et al., this volume). The fact that second dialect acquisition, or the acquisition of sociolinguistic variation in general, provides such an interesting test case for language learning models, lies at the root of the upsurge in relevant research starting in about the year 2000 (see below). 2.3

Transmitting attitudes from caregivers to children

Apart from linguistic building blocks (variants and varieties), the acquisition of sociolinguistic variation also encompasses the transmission of attitudes between generations of speakers. Attitudes are best considered theoretical constructs, and to

10

Gunther De Vogelaer et al.

describe them has proven to be a methodologically daunting task (see, e.g. Garrett 2007). Attitudes are believed to underlie actual behaviour, although the prognostic power of the results of attitudinal research is intensely debated (compare, e.g. Kristiansen and Normann Jørgensen 2011: 287 with Hinskens, Auer, and Kerswill 2005: 39). The transmission of attitudes is an essential part of socialisation, and since attitudes are considered to be context-specific, different socialization contexts need to be distinguished here. Regarding the family context, it is well-known from surveys that many parents are reluctant to speak dialect to their (young) children, and prefer a standard-­like variety, even if they would speak dialect towards each other. This tendency is considered one of the main mechanisms behind processes of dialect loss, and is documented for many regions, with Low German and to a lesser extent Dutch standing out as areas with a rich research tradition. See, e.g. Schulte Kemminghausen (1939) and Janssen (1943) for early studies on the transmission of Low German dialects and Mattheier (1980) for an overview of research on the underlying attitudes; for Dutch see, e.g. Weijnen (1967); Geerts, Nootens, and van den Broeck (1978: 36–37) and Daan et al. (1985). Yet it has also been pointed out that the children themselves may be the motor behind a language shift in families, for instance in the French Alsace region, where Cole (1975) reports that Alsatian adolescents’ increased proficiency in French causes French to become the dominant code in family interactions, despite the community’s general positive attitudes towards the Alsatian dialect. Similarly positive attitudes towards children using non-standard varieties are also reported in a more recent study in Germany (Katerbow, Eichele, and Kauschke 2011). An early study on how attitudes are reflected in actual language usage in families is Giesbers’ (1989) analysis of code-switching in the Dutch province of Limburg. This topic is also addressed in more recent studies like De Houwer (2003) and Smith, Durham, and Fortune (2007) (see also Ghimenton, this volume). A good deal of research has targeted the school context rather than families. An early study by Cremona and Bates (1977), for instance, ascribes the stigmatisation of Italian dialects during primary school to normative pressure from the educational system. In a similar vein, Day (1980) sees attitudes towards Hawaiian Creole English become more negative between kindergarten and first grade, to the benefit of attitudes towards Standard English. As noted above, negative attitudes towards non-standard language usage prevalent in schools have been found to play a key role in the reproduction of social inequality in many regions across the globe (see Cheshire et al. 1989 for a number of case studies from Europe, and also the articles by Buson and Farrington et al. in this volume). In general, some of the older research, including the studies by Cremona and Bates (1977) and Day (1980), has emphasized similarities between children and adults’ attitudes, in that even very young

Chapter 1. Introduction 11



children are sometimes assumed to have or develop adult-like attitudes. The agenda pursued is in line with the Imposed Norm Hypothesis, which states that standard varieties assume their prestige over other language varieties because of historical, cultural norms (Giles et al. 1974), which are transmitted between generations. Most recent research acknowledges that children and adolescents operate in different social environments than adults, and also points out differences between children and adults’ attitudes (Lafontaine 1986; Buson 2009). The role of the peer group is monitored especially in studies on adolescents, like Labov (1972a) or Eckert (1989). Still, research remains to be done for these differences to be properly understood, for instance because the role of the linguistic input is still largely unknown (Beck, this volume) and developmental factors are hardly investigated (Blum-Kulka 2004; see also De Vogelaer and Toye, this volume). 2.4

Children, adults, and language change

A lasting contribution from quantitative sociolinguistics to the toolkit of historical linguistics is the apparent time method, which operates on the assumption that adult speakers have stable grammars. Although cases have been described of grammar change in adults, critical investigations have proven this assumption to be by and large valid, at least for speakers not exposed to substantial geographical and social mobility (see Sankoff and Blondeau 2007 for discussion). The apparent time construct, then, implies that language change takes place as an outcome of the transmission process, which is made explicit by Labov (2007: 346–347) when he equates the altered replication during transmission with change from below or change within the system, whereas change from above (or ‘diffusion’) is a result of language contact, which mostly takes place between adults. Transmission is considered the “primary source of diversity”. Indeed “[t]o understand change, we have to understand how children identify the newer pattern in the community system that they are learning, adopt that pattern, and then move further in that direction” (Labov 2014: 34). In locating language change in the way languages are acquired, Labov’s approach resembles a long tradition of models of language change, ranging from Paul (1886) to more recent models stemming from generative or evolutionary approaches to grammar (see, e.g. Halle 1962 and Lightfoot 1999 for generative models or Croft 2000 for an evolutionary approach). The hypothesis that children are responsible for language change has stirred quite some research into the parallels between child language and historical change. Research that focuses on young children has not yielded enough similarities between acquisition and change to corroborate the hypothesis (Vihman 1980; Diessel 2012). A lot of the research, however, in addition to targeting standard languages as spoken by very

12

Gunther De Vogelaer et al.

young children, pays insufficient attention to quantification to qualify as a proper test of Labov’s framework, which emphasizes the (relative) structural faithfulness and stochastic nature of language change through transmission. In fact, a thorough study into three age groups quantifying the amount of regularised past tense marking in English (Bybee and Slobin 1982) reveals close parallels between the patterns found in third-graders and in adults, which leaves room for a model attributing language change to implementation by successive generations of speakers. Not all instances of language change amount to increasing diversity, however. In contact situations, which tend to be more common for adults than for children, Trudgill (1986) hypothesizes that change is often due to ‘long-term accommodation’, which takes place when speakers’ adaptations to their interlocutors’ speech get habitualized. The role of long-term accommodation in language change is controversial, though. For instance, Chambers (1992) prefers to study processes of longterm accommodation as second dialect acquisition. In addition, Auer and Hinskens (2005: 356) provide an overview of change-by-accommodation studies, and see as the driving force behind language change “not imitation of the language of one’s interlocutor but, rather, an attempt to assimilate one’s language to the possibly stereotyped characteristics of a group one wants to be part of, or resemble.” In essence, then, change through dialect contact presupposes a target variety to be acquired as a second dialect (cf. Siegel 2010: 71–73), and the distinction between change driven forward by adults and by children are more gradual than presented in Labov’s account, as are differences between first and second language acquisition. Kerswill (1996) assesses the role of different age groups in language change, and argues for a balanced approach, which also takes into account the linguistic level of innovations, their complexity, and their sociolinguistic patterning. For instance, children more readily acquire highly complex features and thus seem more capable of creating or adopting innovations involving structurally differential categorisations or sophisticated conditionings, but they are unlikely to serve as models for a speech community, and hence play a minor role in the diffusion of innovations (Kerswill 1996: 191). In the pre-adolescent stage, in contrast, Kerswill observes ‘focusing’ within peer groups, i.e. convergence in patterns of sociolinguistic variation, which creates opportunities for features to diffuse over the speech community. Adolescents, in turn, show a stronger desire for a distinct social identity, making them willing to adapt their speech (Kerswill 1996: 198). Clearly, then, they may move innovations even beyond levels observed in other members of a speech community. Concrete case studies dealing with acquisition and language change have mainly addressed either children or adolescents. One landmark study targeting children, is Kerswill and Williams’ (2000) investigation of koineisation in the English ‘New Town’ Milton Keynes. New Towns allow us to observe the development of a new local, mixed variety through dialect contact in real time. From the lack of linguistic



Chapter 1. Introduction 13

continuity between generations, Kerswill and Williams conclude that, at least in Milton Keynes, the generation of the children rather than the adults are creating the new local variety. As the koineisation process is influenced by demography and the structure of individuals’ social networks, however, it cannot be ruled out a priori that adults play a bigger role in other cases. Adolescents have been investigated most prominently by Eckert (1989, 1997, 2004), who considers them to be the “linguistic movers and shakers, at least in western industrialized societies” and “a prime source of information about linguistic change and the role of language in social practice” (Eckert 1997: 52). In particular, the social meaning attributed and exploited by adolescents reveals a vernacular variant’s potential, and may (but need not) foreshadow its future. In recent years, the language of youngsters in multicultural and often multilingual settings has received much attention (Bucholtz 1999; Clyne 2000; Kern and Selting 2011). Multiethnolects, as youth language in general, often are transitory phenomena, but some instances have shown to be more persistent, and there is the intriguing example of Sheng in Kenya, which has become nativized (Dorleijn, Mous, and Nortier 2015). On the variant level, innovations may be passed on between different generations of youth, or even spread to children or adults (see Cheshire, Nortier, and Adger 2015 for examples). In addition, multiethnic varieties often also contain phenomena that appear to be motivated language-internally, and as such can speed up changes already incipient in the relevant language (Wiese 2013). With language acquisition research taking a much more quantitative turn and incorporating variation as an inherent characteristic of all linguistic input, more thorough tests are becoming possible as to the extent to which linguistic change is pushed forward by children and adolescents. This is especially obvious in a number of contributions within the Usage-based framework, which are discussed below (Section 3.3). 3. Current trends in research on acquisition of variation 3.1

The acquisition of sociolinguistic variation as an emerging research field

Most older research on the acquisition of variation can be characterized as concerning questions relating to grammar, the social status of language varieties and their speakers, or the direction of language change. Since the 1990s, researchers have also proposed theoretical frameworks accounting for the actual mechanisms underlying the acquisition of sociolinguistic variation (Nardy, Chevrot, and Barbu 2013). The two main proposals are abstract variable rule formation and exemplar-based

14

Gunther De Vogelaer et al.

schema formation. These can be situated with respect to the distinction drawn by Ambridge, and Lieven (2011) between two major theoretical approaches to language acquisition. On one side, the nativist, generativist, Universal Grammar paradigm (UG) assumes that important aspects of linguistic knowledge are innate and that this knowledge consists of combinatorial operations applying to abstract linguistic categories (verb, noun, noun phrase, etc.). On the other side, the constructivist, emergentist, functionalist, socio-pragmatic, Usage-Based paradigm (UB) assumes that children have no innate knowledge of grammar but construct it from the input. Neither the adult nor the child’s linguistic knowledge is abstract and formal but it emerges from the memorization and the reorganization of concrete linguistic material due to general cognitive and socio-cognitive abilities (e.g. analogical processing or joint attention). The driving force for this process is the desire to use language and perform communicative and pragmatic functions in the social context. The UG and UB-paradigms are discussed in more detail in Section 3.2 and 3.3. Apart from theoretical reasons, however, two other developments seem to have stirred the interest in how patterns of variation are acquired. First, and perhaps quite trivially, technical developments have made it possible to record child speech in more convenient ways, and compile speech data in larger, enriched and fully browsable databases, which can be made available to the research community more efficiently than before. Thus, while child acquisition researchers until the 1990s typically only had recordings at their disposal from a few children speaking to a caregiver, current databases provide better opportunities to explore more children’s speech in a range of settings, and map it to the input received in various circumstances. And second, increasing contact between varieties in an era of linguistic ‘superdiversity’ (Blommaert and Rampton 2011) is giving rise to innovations which can only be understood properly if we gain insight into the varieties at play and the processes through which these influence one another. A case in point are contemporary urban vernaculars described as multiethnolects, which are typically formed in the interaction between youngsters with varying social and linguistic backgrounds. 3.2

From generative grammar to variable rules

Despite the fact that the UG-paradigm takes a highly idealized and allegedly homogeneous I-language as its object of investigation, it developed an interest in language variation decades ago, through the study of dialects. As described above, the most prominent research topic regarding acquisition within the UG-paradigm, was second dialect acquisition. Surprisingly little attention was (and remains to be paid)



Chapter 1. Introduction 15

to first language acquisition of non-standard systems. For instance, even though dialectal variation had been described in terms of parameters, it took considerable time for parametric explanations to be extended from cross-linguistic (e.g. Hyams 1986) towards language-internal variation in acquisition processes, with Henry’s (1995) study on Belfast English. Wilson and Henry (1998) locate the triggers for different parameter settings in the input. Hence they argue that “the mind/brain is endowed with a program sensitive to community grammar” (Wilson and Henry 1998: 16), which includes the probability matching capacities that play a central role in the quantitative Labovian paradigm. Yang (2010) proposes a computational model for how differences in the age at which a certain parameter is set in different languages, are explained by quantitative differences in the input, which could in principle be suitable to account for dialectal variation as well (cf. also Lacoste and Green 2016: 5–6 and Leivada and Grohmann, this volume, on UG-approaches geared at handling variation). The fact that dialectal differences between grammars carry over to differences in acquisition processes yields interesting perspectives for contrastive research, as evidenced by the flourishing field of cross-linguistic acquisition studies. The comparative study of dialect grammar acquisition, however, has been hampered by the fact that children typically acquire standard language competences in addition to dialectal ones, and the way in which dialects and standard varieties interact shows extensive variation in itself. It is well-accepted in the UG-paradigm that the heterogeneity resulting from a mixed input also affects the acquisition process, for instance because cues helping children to acquire a given pattern are less clear in more heterogeneous input. This becomes clear in Rodina, and Westergaard’s (2013: 60) study on gender and declension in the Tromsø dialect of Norwegian, where the alternation between dialectal and standard inflectional endings in the input is believed to slow down acquisition. Incidentally, the reverse situation in which a heterogeneous input leads to clearer cues and faster acquisition, also occurs (see Cornips and Hulk 2006 for a case study). In general, then, within the UG-approach proposals have been made to account for sociolinguistic variation and its role in acquisition. The most concrete such proposals assume a component in the language faculty steering grammar competition. While the grammar competition model is claimed to be compatible with findings from sociolinguistics, e.g. as couched within the quantitative Labovian paradigm (Yang 2000), it contrasts sharply with recent approaches in sociolinguistics, where competition is located at the variant rather than the variety-­level (e.g. Eckert’s 2012 third-wave sociolinguistics), which turns variation into an inherent property of grammar and even renders concepts like ‘variety’ or ‘system’ highly problematic, as would any notion referring to clearly delimited languages (Makoni and Pennycook 2007; Møller and Jørgensen 2009).

16

Gunther De Vogelaer et al.

The sharp distinction between linguistic and social knowledge is by no means an invention of the UG-paradigm, but presents a founding principle of 20th century linguistics. Based on the oppositions of Langue and Parole (de Saussure 1916) or Competence and Performance (Chomsky 1965), this distinction defines the structural linguistic properties as the legitimate object of the scientific study of language. Challenging this principle, sociolinguistic theories have attempted to account for the relationships between the social and the linguistic. But as Clark (2009) emphasizes, such a theory could not arise from adding an account of variation to a “purely asocial theory of grammar” because this construct allows no room for social knowledge within the theoretical framework itself. Elaborating a theory incorporating social meaning and structural linguistic properties in the same framework is thus an urgent task for sociolinguistics (Chevrot 2012). For this reason, this endeavour is central in our discussion of two main theoretical approaches to sociolinguistic acquisition (for other features opposing the models, see Nardy, Chevrot, and Barbu 2013). Weinreich, Labov and Herzog (1968) and Labov (1972b) forged the concept of variable rules forty years ago. Extending the idea of optional rules from generative grammar, their modelling of variation includes, within the structural description of the rule, the frequency of its operation and the weight of constraints upon its application. According to Labov (1989) and other authors who followed this theoretical framework (Patterson 1992; Roberts 1994, 1997; Smith, Durham, and Fortune 2007), at an early age the child constructs variable rules operating on abstract categories from the forms encountered in his environment. Early on, the general format of these abstract rules would correspond to those of adult speakers with the exception of some characteristics specific to children. Being exposed to input would then favour the adjustment of variable constraints in accordance with the speech used in the child environment. With regard to the distinction drawn by Ambridge and Lieven (2011), this modelling borrows features from both the UG paradigm and the UB paradigm. On the one hand, the proponents of variable rules assume that the child early generalizes general combinatorial devices that operate on abstract categories. However, they do not provide a more detailed description of the process that leads to the creation of these categories and rules. This assumption of the formal and abstract nature of linguistic knowledge is in line with the UG paradigm. On the other hand, the role of the environment is fundamental to the mechanism for learning variable constraints affecting rules, specifically the frequency of perceived variants in different linguistic contexts. This emphasis on input frequency is characteristic of the UB paradigm. On the question of the link between the social and the linguistic, Fasold (1991) noticed that it was sometimes the practice to include social cues (e.g. [+young], [−female]) along with linguistic cues (e.g. [+stress], [ tense]) into the formal apparatus of variables rules. But this

Chapter 1. Introduction 17



practice did not become mainstream in the variationist field and the majority of researchers see the social constraints as external factors that determine the degree of applicability of the rule. In this perspective, linguistic and social knowledge are held apart and the relationship between them is only established once the general format of the rule is in place. 3.3

Usage- and exemplar-based models

Contrary to variable rules, exemplar-based schema formation assumes an early link between linguistic and social information and explains how this link is constructed. This approach also highlights the impact of frequency of perceived forms. In this view, linguistic knowledge is constructed through memorization of episodic traces of an individual’s language experiences (Bybee 2006). Memorizing pieces of language experience amounts not only the linguistic information about words or sequence of words, but also social and contextual elements such as the type of speaker, the acoustic characteristics of his voice (Foulkes and Docherty 2006; Pierrehumbert 2001) or information about the communicative situation. In the memory, each unit or construction would be represented by a set of exemplars corresponding to heard realizations of the linguistic object (Pierrehumbert 2001, 2002, 2003). Frequent objects would thus be represented by more exemplars than less frequent ones even if the experience with the input is mediated by other factors than frequency, such as attention and saliency (Foulkes 2010; Pierrehumbert 2006). By connecting the memorized traces on the basis of their formal or functional similarities, children would then generalize low-level schemas allowing them to produce utterances that they have never heard before (Tomasello 2003). During the course of development, these low-level schemas are reorganized into networks forming more complex and abstract constructions. It should be emphasized that schema-formation does not erase the specific instances but implies concentrating on their similarities when the activity requires it. Schemas are thus nothing more than patterns of connections between memorized traces of language experiences. In contrast to rules, they do not exist independently of the stored instances from which they emerge (Kemmer and Barlow 2000). The exemplar-based schema formation model is clearly situated in the UB paradigm. In this framework, variation results from the competition between schemas or constructions (Bybee 2001a). Because constructions are interconnected on the basis of formal and functional similarities, competition arises between those that perform identical or similar functions. Viewed from this framework, the idea of variation encompasses all the alternatives licensed by the network of constructions. The mechanism for selecting the sociolinguistic variants would be identical to the

18

Gunther De Vogelaer et al.

general mechanism responsible for the choice between two lexical or syntactic alternatives (Chevrot 2012). Moreover, the acquisition of this mechanism is sustained by general learning abilities: “the ability to form networks among stored elements of knowledge and to register frequency of individual items and patterns” (Kemmer and Israel 1994). As for the encoding of social information, this framework assumes that it starts at an early stage, along with linguistic information, as soon as the child begins to memorize pieces of experience. As Foulkes, Docherty, and Watt (2001: 80) noted: “the details encoded at the holistic stage of representation will include features which have sociolinguistic relevance among the adult community”. Since the traces of language experience are not lost during generalization and continue to be stored after this process, the social information would be present, updated and reorganized at all stages of the acquisition process. Since social categories (e.g. ‘women’ or ‘local’) covary with certain linguistic features in the child environment, the structuration of each type of knowledge may reinforce the structuration of the other one. As Clark (2009) argues, this paradigm offers a “more unified approach to variation by incorporating social meaning into the theoretical framework” due to its fundamentally non-modular view of language (see also Docherty, and Foulkes 2014). Recent elements drawn from cognitive neuroscience reinforce the idea of strong relationships between social knowledge and linguistic knowledge. Indeed, the brain response to indexical incoherence is manifested via the same electro­ physiological activity as when constructing sentence meaning based on words and involves brain regions overlapping the processing of sentence meaning (Van Berkum et al. 2008; Tesink et al. 2009). The debate between the UG and the UB approaches is well established in the field of language acquisition (Ambridge and Lieven 2011; Gülzow and Gagarina 2011, inter alia). As far as acquisition of sociolinguistic variation is concerned, researchers have interpreted their results in the framework of the UG paradigm (Westergaard 2004), UB (Nardy, Chevrot, and Barbu 2013) or the variationist rulebased approach (Roberts 1997), but to our knowledge no research design has been developed to contrast the competing models. Yet some predictions emanate from the different models that are testable in future research. The first prediction addresses the issue of the relationship between linguistic and social knowledge. Put at its simplest, the advocates of the UG paradigm consider linguistic knowledge to be autonomous. During an early stage of development, children – guided by some innate mechanism – construct an abstract linguistic system which later interacts at a superficial level with their social knowledge. For the adherents of the UB paradigm, in contrast, linguistic knowledge is intimately linked to other knowledge systems. The child’s linguistic system gradually emerges from memory traces of language experiences in which the linguistic information is mixed with contextual information. Far from being noise to be eliminated, the socio-indexical information helps young



Chapter 1. Introduction 19

speakers to structure the linguistic variability of their environments (Docherty and Foulkes 2014). In contrast to the UG approach, the UB approach therefore predicts that indexical knowledge is an active component in the early emergence of language (Chevrot and Foulkes 2013). In the context of variable rules, social information is included in the weighting of external constraints that determine the degree of applicability of the rule. Therefore the relationship between linguistic and social information is only established once the general format of the rule is in place (Nardy, Chevrot, and Barbu 2013). The early awareness of socio-indexical information or its implicit activation during perception and production are thus specifically predicted by the UB approach. The second prediction concerns the role of lexicon in the generalization of sociolinguistic patterns (see Rys et al., this volume). During language acquisition, change toward native- or adult-like proficiency often comes in the shape of a generalization: a linguistic pattern experienced with known elements is applied by the child or the learner to new elements. While all theories recognize this process that is reflected by regularizations (e.g. goed instead of went), they do not account for this in the same way. On the one hand, the UG approach as well as the variationist rulebased approach assume that generalization is based on the appearance of general abstract mechanisms (parameter setting, rule) that process symbolic structures or units independently from the lexicon (Pinker 1999). On the other, the UB approach assumes that generalizations arise from the formation of schemas encoding patterns of regularity found in the lexicon, where words as well as frequent sequences of words are memorized (Bybee 2001b). As a consequence, the UB approach, but not the UG approach or the variable rule approach, predicts that the acquisition of sociolinguistic variables should be highly dependent of the lexical context in which the variants occur. For example, the early use and the age of acquisition of the standard and non-standard variants of words including a phonological variable should be strongly dependent on the ratio between the frequencies of the variants for each word. As far as sandhi phenomena and grammatical variables are concerned, the same prediction holds, on condition that one takes account of sequences of words rather than words. 3.4

Sociolinguistic cognition and cognitive sociolinguistics

3.4.1 When sociolinguistics meets psycholinguistics In the 1970s and 1980s, empirical studies observed child or learner sociolinguistic variation in light of the patterns established in native adults. During a second era, which started in the 1990s, the studies aimed to explore the acquisition process itself, proposing hypotheses concerning learning mechanisms. Thus, their objectives

20 Gunther De Vogelaer et al.

partially coincided with those of psycholinguistic research on language acquisition (Chevrot, Beaud, and Varga 2001). The key idea to which most researchers working in this field adhere is that initiated by Weinreich, Labov, and Herzog (1968): Languages are heterogeneous and evolving systems, due to their internal dynamics, the contacts between them, and their links with social organization, which is itself evolving, composite and multi-layered (Laks 2013). The internal heterogeneity of languages is described with reference to sociolinguistic variables, i.e. points of variation which enable speakers to say the same thing in different ways, with the variants being “identical in reference or truth value, but opposed in their social and/ or stylistic significance” (Labov 1972b). The same framework is applied to alternation between distinct languages, as bilinguals as well as monolinguals “continually engage in choices amongst alternatives which have the same referential meaning or function” (Poplack, Zentz, and Dion 2012). Although the view of sociolinguistic variants as having a pre-existing fixed social meaning was criticized in favour of a more flexible view of indexicality (Eckert 2012), the strength of this framework lies in its capacity to bring together linguistic, social and cognitive issues. The language environment the child or the learner is confronted with is variable but organized by social factors. Knowledge about social life and knowledge about language thus structure each other. The acquisition of sociolinguistic variation and the learning of indexical meaning are therefore not side effects of the general acquisition process, but an inherent and active part of it (Foulkes, Docherty, and Watt 2005; Chevrot and Foulkes 2013). The encounter of sociolinguistics and psycholinguistics is consistent with the current convergence between social and cognitive sciences, which manifests itself in the rise of new fields such as social neuroscience (Blakemore, Winston, and Frith 2004). In principle, the respective objects of social and cognitive sciences do not suggest that they might converge. Social sciences study collective human groups and their relationship to behaviours and representations, whereas cognitive sciences describe, explain and eventually simulate and amplify the abilities of the human mind and their basis in the brain (Andler 2006). However, the distinction is not as clear as it appears. Since the seventies, one strand of cognitive science has concerned itself with social aspects (Kaufmann and Clément 2011). Indeed, social cognition studies how people process, store and retrieve social information and how they implement it in social situations. The conception of the social dimension that prevails within this approach is different from the conception shared in social science. It is characterized by psychological reductionism and methodological individualism: social facts are limited to discontinuous and limited interactions between individuals, described in terms of mental mechanisms which are responsible for the processing of social information. Social aspects are therefore perceived



Chapter 1. Introduction 21

as a capability of the mind as it functions within the individual at the time of the interaction (Morgan and Schwalbe 1990). Since the work of Durkheim or Weber, social science has used concepts which refer to mental entities: ideologies, collective representations, subjective experience (Morgan and Schwalbe 1990). The type of mind at work here is very different from the mind within the individual. These mental entities are mental states shared by the members of a community. They find expression outside of the individual through the habits and regularities of social life (Kaufmann and Clément 2011). Within this perspective, a phenomenon such as ‘trusting’ cannot be reduced to a limited decision leading to a change of mental and cerebral state. Instead, it is a disposition which is progressively constructed and leads to entrust oneself to the good offices of another person (Quéré 2008). As seen by social science, mental phenomena have a long temporal dimension. They are anchored in a more extensive narrative, linked to a collective story (Molénat 2012). Thus, social science possesses a conception of mind and the science of mind possesses a conception of social dimensions. The two conceptions of the social differ in the time scales covered: a long-term time scale in social science and a short time scale linked to the moment of interaction in cognitive science. The two conceptions of mind differ in the place in which its expression is observed: ‘in the community’ in the case of social science and within the individual for cognitive science. These two approaches result in the construction of different observable phenomena. However, because both the social and the cognitive sciences possess conceptions of the social world and of the mind, it was inevitable that they would converge. The numerous points of convergence between the two scientific fields can be summarized as two tendencies (Kaufmann and Clément 2011). First, in social approaches to cognition, the starting point lies in the social world. The question is how situations of interaction and social life engender patterns of thought. Mental entities are thought of as being properties of a community and there would therefore be communities of thought in the same way as there are speech communities. Second, in cognitive approaches to the social, the starting point is the mind, clearly delimited within the individual. It is conceived of either in terms of automatic cognitive processes or in terms of rational decision-making processes. The question is how the implementation of these processes in interactions accounts for social phenomena. Sociolinguistics has contributed to these two types of convergence between cognitive science and social science through the emergence of two new fields: cognitive sociolinguistics (Kristiansen and Dirven 2008) and sociolinguistic cognition (Campbell-Kibler 2010; Loudermilk 2013). These fields strongly interact with research on the acquisition of sociolinguistic variation.

22

Gunther De Vogelaer et al.

Cognitive sociolinguistics, sociolinguistic cognition, acquisition of variation Kristiansen and Dirven (2008) consider that the most prototypical contributions to the field of cognitive sociolinguistics share three features. First, they explore language-internal or cross-linguistic variation linked to social dimensions. Second, they build on solid empirical methods. Third, they draw on the working hypotheses of cognitive linguistics: language is processed via the same cognitive mechanisms as other aspects of cognition, grammar is a conceptualization of the world and linguistic knowledge results from an emergent conventionalization of usage (Croft and Cruse 2004). Cognitive sociolinguistics accepts this framework and adds to it that languages are spaces of variation, linked to both cultural and social diversity. The fifteen contributions to Kristiansen and Dirven’s landmark book (2008) relate to diverse issues: Explaining sociolinguistic and sociosemantic patterns using cognitive concepts; exploring the variational dimension of large corpora in accordance with the methodological standards of sociolinguistics; establishing relations between cultural or political models and language policies or attitudes; observing how discourse conveys representations of social and cultural entities. Despite this diversity, all the contributions share characteristics which are typical of sociolinguistics. Linguistic knowledge and way of thinking are properties of communities. These communities are heterogeneous and situated at the sociological, cultural and ideological levels. Moreover, the preferred methodology consists of corpus studies. Cognitive sociolinguistics is therefore derived from the sociolinguistic tradition and social approaches to cognition. For its part, sociolinguistic cognition (Campbell-Kibler 2010; Loudermilk 2013) explores the cognitive and cerebral mechanisms underpinning the ability to encode sociolinguistic variation, to implement it during production and to process it during perception. Although this line of research is rooted in sociophonetics, recent contributions extend the work to morphosyntactic variation (Squires 2013). The research topics are the following: Cognitive representation of variation, involving the debate between abstract rules and exemplars; influence of indexical knowledge on the on-line processing of variation; retrieval of indexical information during the on-line processing of variation; study of the cerebral mechanisms that underpin the processing of indexical information. These topics and the methods (elicitation tasks, standardized tests, reaction time and eye-tracking experiments, neuroimaging) are typical of psycholinguistics. The mind is perceived in terms of the cognitive operations located within the individual. Sociolinguistic cognition is thus typical of a cognitive approach to social facts. Research into the acquisition of variation lies at the intersection of the two traditions (Chevrot and Foulkes 2013). The work is based on sociolinguistics and contributes to it. As a result, it focuses on the acquisition of linguistic knowledge 3.4.2



Chapter 1. Introduction 23

conceived of as properties shared by communities of speakers, both children and adults. At the same time, the work draws on the traditions of developmental psycholinguistics. This second affiliation leads to the exploration of the cognitive mechanisms involved in the acquisition of sociolinguistic variation. In brief, research on acquisition of variation mixes a social approach to cognition and a cognitive approach to the social, the two trends being present in this book. This dual affiliation should be seen as one of the most valuable assets of the developing field. 3.4.3 What sociolinguistics and psycholinguistics could bring to each other As a contribution to a series of papers on the place of anthropology in cognitive science (Beller, Bender, and Medin 2012), Levinson (2012) remarked that cognitive research was built on “the myth of ‘the human mind’ (…) idealized away from all the ‘noise’ of individual variation or systematic cultural diversity”. The idealization would be particularly at odds with the study of language, since “diversity is the most striking feature of human language – there is no other animal (…) which has such myriad variants of form and meaning at every level of its communication system”. The developing field of sociolinguistic acquisition is protected against temptation towards such an idealization due to its direct affiliation with sociolinguistics. The sociolinguistic perspective draws researchers’ attention to the inherently variable nature of language and the social diversity of situations in which contact occurs between languages or dialects. A research objective as obvious as establishing the age of acquisition of sociolinguistic patterns cannot be implemented without careful consideration of the indexical values and the social organization of the community under study. Within a given community, certain sociolinguistic variables are likely to be more strongly linked to identities and social roles. Their social meaning would thus be more accessible to children and more likely to influence the acquisition of the corresponding linguistic features. Furthermore, the caregivers control the use of these salient variables in child-­directed speech, according to the social role attributed to the child, and particularly the child’s gender (Foulkes, Docherty, and Watt 2005; Nardy, Chevrot, and Barbu 2013; Smith, Durham, and Richards 2013). More generally, the contribution of sociolinguistics to research into language acquisition is quite similar to the contribution of anthropology: renewing interest in variability and cultural diversity within and between communities and their influence on language and cognition. The contribution of psycholinguistics to the developing field of sociolinguistic acquisition is also obvious. Formulating hypotheses on mental states and learning mechanisms is a necessary step if we are to explain the changes that occur in the child’s or the learner’s linguistic behaviour. Moreover, studying the cognitive mechanisms involved in the acquisition of sociolinguistic variables and their social meaning provides fresh insight to old issues in sociolinguistics. One such issue is

24

Gunther De Vogelaer et al.

the degree of determinism in the individual choice of variants. Two opposing tendencies have divided sociolinguistic approaches to variation on this point (Chevrot, Beaud, and Varga 2001; Chevrot 1994). On the one hand, speakers are supposed to be free to manipulate variants in order to achieve communicative goals (e.g. manipulating interpersonal distance) or to claim or negotiate their identity. On the other, speakers’ linguistic behavior is seen as the result of social constraints (e.g. social status, gender, ethnicity, etc.) and cognitive limits (e.g. availability of the variants, amount of attention paid to speech). This debate could be renewed on the basis of three questions about the acquisition of sociolinguistic patterns in the first language. First, do the sociolinguistic differences appear early from the implicit learning of frequency patterns in the family environment? Or, alternatively, do these differences progressively emerge through the gradual learning of differentiated roles linked to social identities? Second, at what age does the choice of the variants depend on the knowledge of their social meaning? If children select variants on the basis of their meaning, they can make themselves more independent of earlier implicit learning. Third, is children’s usage of variants blocked by the early influence of the family environment or is it sufficiently flexible to adapt to the diversity of interactions? The available results on the acquisition of phonological variables in first language reviewed by Nardy, Barbu, and Chevrot (2013) suggest an initial response to the three questions. With regard to the first question, all the studies that have observed social differences in children (Reid 1978; Martino 1982; Romaine 1984; Chevrot, Beaud, and Varga 2001; Martin 2005; Díaz-Campos 2005; Chevrot, Nardy, and Barbu 2011) tend toward the same results: in informal speech, from the age of 3 to the age of 10, the higher the socioeconomic status (SES) of the parents, the more standard variants the child produces. Moreover, Smith, Durham, and Richards (2013) found a significant correlation between the use of non-standard Scottish variants in children aged 3 to 4 and their caretakers. Together, these results suggest the early transmission of sociolinguistic differences during family interactions. With regard to the second question, only one study we know of has been designed to observe the relationship between production of variants and knowledge of their social meaning. Barbu et al. (2013) examined how 185 children aged 2 to 6 years and raised in families from two contrasted social backgrounds produce and evaluate variable liaison in French (a well-documented sociolinguistic marker in adults). Production of liaison was investigated through a picture naming task and knowledge of its social value through a judgment task. Significant social differences appeared simultaneously in both children’s productions and judgments at the age of 5–6 years. However, one year earlier, the judgment of the higher-SES children, as well as the judgment of the lower-SES children, did not discriminate standard and non-standard variants. Thus no evidence was found that the awareness of social



Chapter 1. Introduction 25

meaning is a prerequisite of the ability of the higher-SES children to produce more standard variants at the age of 5–6 years. With regard to the third question on flexibility of the child usage, several studies have addressed the issue of stylistic ability in children (Reid 1978; Macaulay 1977; Chevrot, Beaud, and Varga 2001; Romaine 1984; Patterson 1992; Roberts 1994, 1997; Díaz-Campos 2005; Martin 2005; Smith, Durham, and Fortune 2007; Smith, Durham, and Richards 2013). All but one point in the same direction: children use the standard variants more frequently in formal contexts. This result applies to young children within parent-child interaction as well. Smith, Durham, and Richards (2013) observe that three-year olds and their caretakers use more non-standard variants during daily activities and games than during educational (What color is it?) or disciplinary (Behave! Get up) exchanges. Moreover, a longitudinal study of a group of pupils from the same class in kindergarten (Nardy 2008; Nardy, Chevrot, and Barbu 2014) showed that the children’s usages of phonological variables tended to converge during the observation period (4 to 5 years of age). Furthermore, the more two individuals interact, the more similar their use of standard and non-standard variants. Children’s sociolinguistic usage is thus not blocked by the early influence of the family. Rather, short-term stylistic changes are an inherent part of the early interactions within the family and midterm adjustments in the choice of variants may result from peer group influences. According to these results, it appears that the early usage of sociolinguistic variants is both strongly influenced by the vernacular forms used in the early child environment and highly flexible and likely to adapt to the context. Moreover, one should take seriously the hypothesis that the early contextual adaptations could occur through implicit alignment to the interlocutor’s speech (Garrod and Pickering 2004; Ghimenton, Chevrot, and Billiez 2013; Martin, Chevrot, and Barbu 2010), without awareness of the social meaning of the variants. As a result of the interest in diversity and variability in the field of sociolinguistic acquisition, these conclusions should be put in perspective. For example, the review of the available results on the appearance of the gender-related differences leads to a quite different conclusion than those on SES-related differences (Nardy, Chevrot, and Barbu 2013). Amongst the 11 reviewed studies, all the possible tendencies are found: no gender effect (7 studies, ages 2–9), more standard variants in girls (2 studies, ages 6–10), and more standard variants in boys (2 studies, ages 3–7). Everything suggests that the acquisition of gender-related patterns is highly dependent on the social meaning of the variable studied in a particular community. For example, variants with a strong association with gender in adults may be passed on to girls at an early stage because of the differences in speech addressed to girls and boys. Conversely, for other variants, gender-related differences would appear later in development through the influence of the peer group, the formation of gender identity, different educational practices and the accumulated effect of input.

26 Gunther De Vogelaer et al.

As we have just seen, the available results in the field of sociolinguistic acquisition provide sufficient grounds for formulating precise hypotheses to be tested or more general questions to be empirically addressed. 4. The need for cross-linguistic research: This book 4.1

Article summaries

Given the much more thorough integration of socio- and psycholinguistic perspectives taking place in current research on the acquisition of sociolinguistic variation, there is a need for publications that make accessible the insights that can be drawn from both fields. This book, therefore, aims at providing such a state-of-the-art at a moment in which the acquisition of variation is emerging as a proper field of investigation. The book is conceived as a selection of theoretically informed and empirically well-grounded case studies, providing a comprehensive discussion of research traditions and language situations involved. This should not only provide a clear overview of existing research on the acquisition of sociolinguistic variation, but also facilitate cross-linguistic comparison in future investigations. In order to illustrate the comprehensiveness of the volume, we not only provide a summary of the individual articles (4.1), but also an overview of answers given with respect to three main research questions that have been identified as crucial for understanding the very nature of the acquisition process (4.2; see also Chevrot and Foulkes 2013). Erica Beck investigates the extent to which five-year-olds from Philadelphia, Pennsylvania (US) are aware of regional variation in American English. Based on a number of word-based tasks, she observes that most children reliably discriminate between regional accents, and can identify a local regional accent. In contrast to expectations raised by earlier investigations (e.g. Kinzler and DeJesus 2013), the so-called outsiders in her sample, i.e. children who are exposed to a non-local accent at home, do not show an increased awareness in comparison to insiders. This is explained as a result of the children associating the non-local varieties with individuals rather than with regions, which is typical for the “library of speakers” stage (Munson 2010), in which children do not yet link language variation to regional information. Laurence Buson studies children’s ability to construct their identity with linguistic means. To investigate whether children already qualify as pluristylistic speakers, she carries out a case study on how 10-11-year-old children from Grenoble (France) use the discourse markers hein and oh. While peer-addressed speech is fairly similar in both schools in the investigation, adult-addressed speech



Chapter 1. Introduction 27

is more formal in the socially mixed school. Thus, children within socially heterogeneous networks show greater stylistic flexibility than those within socially homogeneous networks. In the first of a number of articles on Dutch, Leonie Cornips addresses variable word order in perfective two-verb clusters in Standard Dutch as spoken in Heerlen (the Netherlands). Experimental results show that older children (5;0–5;11) orient towards the adult pattern and display a preference for V2-V1 order, whereas the youngest children (between 2;8–3;10) prefer a rigid V1-V2 word order. This confirms Labov’s (1989) expectation that children acquire grammatical properties before stylistic and sociolinguistic constraints. The study by Gunther De Vogelaer and Jolien Toye targets Dutch too, and discusses attitudes towards a number of regional varieties used in the East Flemish town Kluisbergen (Belgium). While the 8-year-olds in their experiment do distinguish between varieties and orient towards Standard Dutch as a model for their own speech, only older children and adolescents assign any significant prestige to it. In addition, sensitivity to the ‘covert prestige’ of the local variety only emerges in adolescence, and peaks around the age of 16. The attitude measurements reveal significant parallels between sociolinguistic thinking and children and adolescents’ general psychosocial development. Andrea Ender documents how the coexistence of dialect and standard varieties of German in the Swiss context poses a challenge for adult learners of German as a second language. Data from three individuals using dialect and/or standard in free speech and in elicited language tasks illustrate variation in the extent to which dialect and standard variants are incorporated in the second language system. Language usage appears to be influenced by factors as diverse as second language learners’ sense of identity within the linguistic community and the linguistic complexity of certain variants in the target language. In the second paper on American English, Charlie Farrington, Jennifer Renn and Mary Kohn investigate the relationship between school demographics and the use of African American English (AAE) features. The results differ between the morphosyntactic and phonetic variables: whereas the use of morphosyntactic AAE variants not only correlates with the proportion of African Americans in the school population but also increases with age, phonetic variants remain constant with age. This pattern indicates that morphosyntactic variables are affected by stylistic choices related to development and social stance, leading to patterns of age-grading, and is consistent with earlier findings that the use of AAE features peaks in adolescence. Anna Ghimenton carries out an analysis of the production of a 25-month old boy growing up in the Veneto region in Italy, and his closest relatives: quantitative analyses show a preference for Italian in adult-child conversations, and a

28

Gunther De Vogelaer et al.

trend to use more Veneto dialect in inter-adult discourse. The child’s willingness to use dialectal elements appears to match the proportion of dialectal features in the interlocutor’s speech. In addition, qualitative analysis provides evidence that variation in the language environment helps in learning in which contexts one language is more appropriate than the other and hence provides guidance during the socialization process. In their article comparing clitic placement in Standard Modern Greek and Cypriot Greek, Evelina Leivada and Kleanthes Grohmann explore the connection between bidialectalism and metalinguistic awareness. Within the framework of their Socio-Syntax of Development Hypothesis, they argue that Cypriot Greek children face a pressure to adjust to the high variety and hence competing motivations are observed in their developing grammars, which causes Standard Modern Greek patterns to surface in their Cypriot Greek. Kathy Rys, Emmanuel Keuleers, Walter Daelemans and Steven Gillis discuss the acquisition of the Maldegem dialect by Belgian children with (a regional variety of) Standard Dutch as their first language. Their subjects’ data show evidence of word-by-word learning across the board and also show ‘neighbourhood effects’, i.e. the acquisition of a specific dialect feature in a specific word is influenced by the word’s lexical neighbours (i.e., words in the mental lexicon displaying the largest similarity). The data are used for a critical evaluation of Chambers’ (1992) rulebased account of second dialect acquisition and found to be more consistent with analogical, memory-based models of language acquisition. Erik Schleef’s investigation of immigrant teenagers in London addresses the acquisition of social and linguistic constraints on t-glottaling. Generally speaking, t-glottaling becomes available for stylistic work after two years of residence in Britain, even though complex constraints on the phenomenon are only acquired later. The article also explores different social meanings associated with use and non-use of t-glottaling, which include casualness or youth identity on the one hand, and Polishness or alignment with mainstream British culture on the other. 4.2

General characterisation and future perspectives

The articles in the volume combine insights from psycho- and sociolinguistics and are as such representative of recent investigations into the acquisition of sociolinguistic variation. All articles also provide an answer to at least one of three research questions that are deemed crucial for a proper understanding of how sociolinguistic variation is acquired (see also Chevrot and Foulkes 2013), viz.:



Chapter 1. Introduction 29

RQ1: Stages in the appearance of adult- or native-like sociolinguistic knowledge: at what age or proficiency level do sociolinguistic patterns appear? And which factors underlie these age-related changes (developmental or social ones, indexical knowledge) RQ2: The motor for acquisition and the influence of the environment (family, peers, teachers): what are the respective contributions of social factors (norms and identity) and learning of implicit patterns in the input? RQ3: The cognitive nature of the mechanisms responsible for the acquisition of variation: symbolic mechanisms (e.g. rules) or more concrete schemas linking together social and linguistic knowledge through exemplar-based learning? The answer to the first question, i.e. which stages can be distinguished in the acquisition of sociolinguistic variation, strongly depends on the type of acquisition process that is studied, and in particular on the age of onset of the process. From the three papers dealing with variation in language production by pre-schoolers, it becomes clear that very young children are indeed capable of reproducing sociolinguistic variation present in the input, both in circumstances in which two varieties are in a diglossic relationship (Leivada and Grohmann) and in cases tending towards diaglossia, in which extensive mixing between varieties occurs (Ghimenton). Cornips’ data, however, show that this also depends on their acquisition of the grammatical categories involved (in Cornips’ Heerlen Dutch case: perfective verb clusters). Farrington, Renn and Kohn, and Schleef provide evidence that adolescents may go further and extend patterns of variation beyond proportions observed in the input (e.g. peaks in African American English features) or exploit variation to reallocate social meanings (e.g. non-use of glottaling indexing Polishness in immigrant teenagers in London). The articles by Schleef and Ender underscore that sociolinguistic variation poses a particular challenge for adolescent and adult learners: Schleef ’s test persons only start exploiting the sociolinguistic potential of t-glottaling after two to three years of exposure, and of Ender’s three subjects only one reaches a native-like proficiency in using both Standard and Swiss German relatives. In both studies, usage patterns are linked to the subjects’ sense of identity within the respective speech communities, but the fact that learners start off from a stage with variable or reallocated patterns of variation and pass via intermediate stages with simplified patterns and/or overgeneralisations before they adopt native-­ like patterns (with/without variation), indicates serious difficulties in detecting both linguistic and sociolinguistic constraints on variation. Focusing on variation in the mother tongue, the attitudinal studies by Beck and De Vogelaer and Toye allow hypotheses to be developed about how the different behaviour of children and adolescents should be understood: Beck, in line with

30

Gunther De Vogelaer et al.

Munson’s (2010) ‘library of speakers’ account, suggests that children link variants primarily with individual interlocutors, and only later assign more abstract social meanings. De Vogelaer and Toye’s data indicate that this process of assigning abstract social meanings and aligning them with those of their peers may extend deep into childhood, even for varieties that are part of children’s everyday linguistic life. While faster formation of attitudes under more transparent sociolinguistic conditions than in Flanders cannot be excluded (and is indeed likely), the correlations between the structure of the children’s attitudes and aspects of general psychosocial thinking show that developmental factors play a role in the formation of attitudes, and children’s attitudes may differ from adults’ in fundamental ways. The scenario sketched in the attitudinal investigations to some extent also answers the second question, viz. what constitutes the motor for acquisition, and what is the role of the environment (family, peers, teachers) in this process? It appears that young children conceive of variation in relation to their different interlocutors (e.g. data by Beck, Ghimenton), and are able to extract frequency data from the input and reproduce them (e.g. the 5-year olds in Cornips’ data). The data by Buson and Farrington, Renn, and Kohn indicate that older children orient towards groups of people, too, in particular peer groups formed at school. Farrington, Renn, and Kohn’s morphosyntactic data furthermore illustrate how adolescents may exploit variation to construct an identity of their own, making the use of African American English features peak at age 16. Identity construction also plays a crucial role in choices made by the adolescent and adult learners described by Schleef and Ender, although they also report severe difficulties to reproduce variation from the input. Given the nature of the process, by which different ways of speaking are first associated with individuals and only later with groups of speakers (and, potentially, situations), varieties and styles are best considered emergent entities. The process by which they emerge may be a complex one, especially in situations of diaglossia, in which a lot of mixing between varieties is taking place. The third and final question that is addressed in this paragraph is which cognitive mechanisms are responsible for the acquisition of variation: are those symbolic mechanisms (e.g. rules) or rather more concrete schemas linking together social and linguistic knowledge through exemplar-based learning? Most articles in the volume subscribe to the common assumptions in variationist sociolinguistics, including the idea that usage frequencies form an integral part of linguistic competence. There are, however, different ways of modelling how frequency information relates to competence, which differ in the extent to which frequency information is kept apart from grammatical knowledge (see Section 3.2 above). While most papers do not take an explicit stance on this issue, both ends of the spectrum of possible answers to this question are represented in the book. Leivada and Grohmann maintain a clear distinction between grammatical knowledge and



Chapter 1. Introduction 31

frequency: they conceive of Cypriot bilectalism as involving a categorical use of proclisis in Standard Modern Greek vis-à-vis enclisis in Cypriot Greek. In order to explain the mixing of Cypriot and Standard Greek patterns in their data, they adopt a modular approach, in which children’s output is steered not only by grammatical but also by metalinguistic considerations, in this case a pressure to adopt variants from the prestige variety when performing an experiment on Cypriot Greek. The other end of the spectrum is represented by Rys et al. They argue that their results are best explained as a result of memory-based lexical learning, which is ‘lazy’ (i.e. no rule formation is assumed). Rys et al. also discuss the merits of Chambers (1992) rule-based account for similar data, and acknowledge that some of the differences between their results and Chambers’ may be a result of the fact that their subjects were exposed to the target dialect from a younger age than Chambers’ subjects. More concretely, it is suggested that “younger children are more attuned to lexical factors than older ones”. Other articles suggest further differences between learners of different ages: for instance, Cornips assumes that the rigid word order used by her youngest informants is explained as a result of a failure to distinct between perfective auxiliaries and modals, between participles and infinitives, or between different usages of participles (in copula vs. auxiliary constructions). Schleef, in contrast, observes that his adolescent learners of English are particularly good at acquiring functional categories, which he relates to findings from second language research showing that adult learners rely primarily on distributional and semantic cues in morphosyntactic segmentation, whereas L1 acquirers make more use of prosodic cues (Valian and Levitt 1996). Of course the articles in this book allow only preliminary answers to the three questions asked at the beginning of this paragraph, and more thorough answers require future research. Regarding the first question, whether any stages can be distinguished in the acquisition of sociolinguistic variation, future research would need to clarify the role of developmental factors and of relative clarity of variability patterns in the input. In addition, the process by which patterns of variation are aligned within children’s peer groups is still poorly understood, as is the extent to which this process influences their language usage towards their primary caregivers. With regard to later stages, Eckert’s (2004) conjecture that “adolescence is not a natural life stage” casts doubt over the applicability of the findings from psychosocial research used by De Vogelaer and Toye outside modern western societies. The second question asked above concerned the motor for the acquisition of sociolinguistic variation. As sketched above, it seems reasonable to assume that children start off by modelling their language on input observed in their interlocutors (typically their primary caregivers) and move on to form abstract concepts of varieties and styles. We know little about how this process takes place, however, and the extent to which it boils down to the mere internalization of patterns extracted from the input, or

32

Gunther De Vogelaer et al.

the extent to which this process is mediated by aspects of identity, or how far the process differs between age groups. With respect to the third question, i.e. which learning mechanisms are observed, the question of how to integrate frequency information into linguistic knowledge is currently intensely debated in several subfields of linguistics, and presents one of the most important research questions for future research on the acquisition of sociolinguistic variation, too (see Section 3.2 and 3.3). In addition, the extent to which age of onset determines which learning mechanisms will be used is still largely uncharted territory, as is the influence of prior linguistic knowledge in the case of second dialect learning.

References Allerup, Peter, Mogens Jansen, Henrik Löb, and Børge Prien. 1979. Retstavning: Nogle første analyser af folkeskolens afgangsprøve i 1978: Vol. III: Nogle statistiske analyser. København: Danmarks Pædagogiske Institut. Ambridge, Ben, and Lieven, Elena V.M. 2011. Child Language Acquisition: Contrasting Theoretical Approaches. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511975073 Ammon, Ulrich. 1989. “Aspects of Dialect and School in the Federal Republic of Germany.” In Dialect and Education: Some European Perspectives, ed. by Jenny Cheshire, Viv Edwards, Henk Münstermann, and Bert Weltens, 113–138. Clevedon, and Philadelphia: Multilingual matters. Andler, Daniel. 2006. “Sciences Cognitives. Encyclopædia Universalis.” Accessed October 11, 2016. http://www.universalis.fr/encyclopedie/sciences-cognitives/ Auer, Peter. 2005. “Europe’s Sociolinguistic Unity, or: A Typology of European Dialect/Standard Constellations.” In Perspectives on Variation: Sociolinguistic, Historical, Comparative, ed. by Nicole Delbecque, Johan van der Auwera, and Dirk Geeraerts, 7–42. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110909579.7 Auer, Peter, and Frans Hinskens. 2005. “The Role of Interpersonal Accommodation in a Theory of Language Change.” In Dialect Change: The Convergence and Divergence of Dialects in Contemporary Societies, ed. by Peter Auer, Frans Hinskens, and Paul Kerswill, 335–357. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511486623.015 Barbu, Stéphanie, Aurélie Nardy, Jean-Pierre Chevrot, and Jacques Juhel. 2013. “Language Evaluation and Use during Early Childhood: Adhesion to Social Norms or Integration of Environmental Regularities?” Linguistics 51 (2): 381–411.  doi: 10.1515/ling-2013-0015 Beller, Sieghard, Andrea Bender, and Douglas L. Medin. 2012. “Should Anthropology Be Part of Cognitive Science?” Topics in Cognitive Science 4 (3): 342–353.  doi: 10.1111/j.1756-8765.2012.01196.x Bernstein, Basil. 1971–1975. Class, Codes and Control: Volume 1–3. St. Albans: Paladin.  doi: 10.4324/9780203014035 Berthele, Raphael. 2002. “Learning a Second Dialect. A Model of Idiolectal Dissonance.” Multilingual 21 (4): 327–344.  doi: 10.1515/mult.2002.014 Blakemore, Sarah-Jayne, Joel Winston, and Uta Frith. 2004. “Social Cognitive Neuroscience: Where Are We Heading?” Trends in Cognitive Sciences 8 (5): 216–222.  doi: 10.1016/j.tics.2004.03.012



Chapter 1. Introduction 33

Blom, Jan-Petter, and John J. Gumperz. 1972. “Social Meaning in Linguistic Structures: Code Switching in Northern Norway.” In Directions in Sociolinguistics, ed. by John J. Gumperz, and Dell Hymes, 407–434. New York: Holt, Rinehart and Winston. Blommaert, Jan, and Ben Rampton. 2011. “Language and Superdiversity.” Diversities 13 (2): 1–21. Blum-Kulka, Shoshana. 2004. “The Role of Peer Interaction in Later Pragmatic Development: The Case of Speech Representation.” In Language Development across Childhood and Adolescence, ed. by Ruth A. Berman, 191–210. Amsterdam: John Benjamins.  doi: 10.1075/tilar.3.13blu Bucholtz, Mary. 1999. “‘Why Be Normal?’ Language and Identity Practices in a Community of Nerd Girls.” Language in Society 28 (2): 203–223.  doi: 10.1017/S0047404599002043 Buson, Laurence. 2009. Variation stylistique entre 5 et 11 ans et réseaux de socialisation scolaire: Usages, représentations, acquisition et prise en compte éducative. Ph.D. Dissertation University of Grenoble. Bybee, Joan L. 2001a. “Frequency Effects on French Liaison.” In Frequency and the Emergence of Linguistic Structure, ed. by Joan L. Bybee, and Paul J. Hopper, 337–359. Philadelphia: John Benjamins.  doi: 10.1075/tsl.45.17byb Bybee, Joan L. 2001b. Phonology and Language Use. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511612886 Bybee, Joan L. 2006. “From Usage to Grammar: The Mind’s Response to Repetition.” Language 82 (4): 711–733.  doi: 10.1353/lan.2006.0186 Bybee, Joan L., and Dan I. Slobin. 1982. “Rules and Schemas in the Development and Use of the English Past Tense.” Language 58 (2): 265–289.  doi: 10.1353/lan.1982.0021 Campbell-Kibler, Kathryn. 2010. “New Directions in sociolinguistic cognition.” In University of Pennsylvania Working Papers in Linguistics 15 (2): 31–39. Chambers, J.K. 1988. “Acquisition of Phonological Variants.” In Methods in Dialectology: Proceedings of the 6. Internat. Conference Held at the Univ. College of North Wales, 3rd–7th August 1987, ed. by Alan Thomas, and Martin Ball, 650–665. Clevedon. Chambers, J.K. 1992. “Dialect Acquisition.” Language 68 (4): 673–705.  doi: 10.1353/lan.1992.0060 Chambers, J.K. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Significance. 2nd ed. Oxford: Blackwell. Cheshire, Jenny, Viv Edwards, Henk Münstermann, and Bert Weltens. 1989. “Introduction.” In Dialect and Education: Some European Perspectives, ed. by Jenny Cheshire, Viv Edwards, Henk Münstermann, and Bert Weltens, 113–138. Clevedon, and Philadelphia: Multilingual matters. Cheshire, Jenny, Jacomine Nortier, and David Adger. 2015. Emerging Multiethnolects in Europe. London: University of London. Chevrot, Jean-Pierre. 1994. “La variation phonétique: Un point de vue cognitif.” Langage et société 70 (1): 5–33.  doi: 10.3406/lsoc.1994.2675 Chevrot, Jean-Pierre. 2012. “Que reste-t-il de la distinction compétence-performance dans les théories fondées sur l’usage?” Mémoires de la Société de Linguistique de Paris 22 (‘L’éventail des compétences linguistiques et la (dé)valorisation des performances’): 49–68. Chevrot, Jean-Pierre, Aurélie Nardy, and Stéphanie Barbu. 2011. “Developmental Dynamics of SES-Related Differences in Children’s Production of Obligatory and Variable Phonological Alternations.” Language Sciences 33 (1): 180–191.  doi: 10.1016/j.langsci.2010.08.007 Chevrot, Jean-Pierre, Laurence Beaud, and Renata Varga. 2001. “Developmental Data on a French Sociolinguistic Variable: The Word-Final Post-Consonantal /R/.” Language Variation and Change 12 (3): 295–319.  doi: 10.1017/S095439450012304X

34

Gunther De Vogelaer et al.

Chevrot, Jean-Pierre, and Paul Foulkes. 2013. “Introduction: Language Acquisition and Sociolinguistic Variation.” Linguistics 51 (2): 251–254.  doi: 10.1515/ling-2013-0010 Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge: MIT Press. Clark, Lynn. 2009. “Cognitive Sociolinguistics: A Viable Approach to Variation in Linguistic Theory.” Lacus Forum 33: 105–118. Clyne, Michael. 2000. “Lingua Franca and Ethnolects in Europe and Beyond.” Sociolinguistica 14 (1): 83–89.  doi: 10.1515/9783110245196.83 Cole, Roger L. 1975. “Divergent and Convergent Attitudes towards the Alsatian Dialect.” Anthropological Linguistics 17 (6): 293–304. Cornelissen, Georg. 1998. “Taal en onderwijs in Noord-Limburg in de Franse tijd (1794–1814).” Vereniging voor Limburgse Dialect- en Naamkunde 96: 3–30. Cornips, Leonie, and Aafke Hulk. 2006. “External and Internal Factors in Bilingual and Bidialectal Language Development: Grammatical Gender of the Dutch Definite Determiner.” In L2 Acquisition and Creole Genesis: Dialogues, ed. by Claire Lefebvre, Lydia White, and Christine Jourdan, 355–378. Amsterdam: John Benjamins.  doi: 10.1075/lald.42.21cor Cremona, Cristiana, and Elizabeth Bates. 1977. “The Development of Attitudes toward Dialect in Italian Children.” Journal of Psycholinguistic Research 6: 223–232. Croft, William. 2000. Explaining language change. An evolutionary approach. Harlow: Pearson Longman. Croft, William, and D. Alan Cruse. 2004. Cognitive Linguistics. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511803864 Daan, Jo, Kas Deprez, Roeland van Hout, and Jan Stroop. 1985. Onze veranderende taal. Utrecht: Spectrum. Day, Richard R. 1980. “The Development of Linguistic Attitudes and Preferences.” TESOL Quarterly 14 (1): 27–37.  doi: 10.2307/3586806 De Houwer, Annick. 2003. “Language Variation and Local Elements in Family Discourse.” Language Variation and Change 15 (3): 329–349.  doi: 10.1017/S0954394503153033 De Saussure, Ferdinand. 1916. Cours de linguistique générale, ed. by Charles Bally, and Albert Sechehaye. Paris: Payot. De Vogelaer, Gunther, and Guido Seiler. 2012. “The Dialect Laboratory: Introductory Remarks.” In The Dialect Lab: Using Dialects as a Testing Ground for Theories of Language Change, ed. by Gunther De Vogelaer, and Guido Seiler, 1–31. Amsterdam: John Benjamins.  doi: 10.1075/slcs.128.01vog De Vogelaer, Gunther, and Jan Klom. 2013. “Mikrovariation beim Erwerb des niederländischen Genussystems.” In Dialektologie in neuem Gewand: Zu Mikro-/Varietätenlinguistik, Sprachenvergleich und Universalgrammatik, ed. by Werner Abraham, and Elisabeth Leiss, 141–169. Hamburg: Helmut Buske Verlag. Díaz-Campos, Manuel. 2004. “Acquisition of Sociolinguistic Variables in Spanish: Do Children Acquire Individual Lexical Forms or Variable Rules.” In Laboratory Approaches to Spanish Phonology, ed. by Timothy L. Face, 221–236. Berlin, and New York: Mouton de Gruyter. Díaz-Campos, Manuel. 2005. “The Emergence of Adult-like Command of Sociolinguistic Variables: A Study of Consonant Weakening in Spanish-Speaking Children.” In Selected Proceedings of the 6th Conference on the Acquisition of Spanish and Portuguese as First and Second Languages, ed. by David Eddington, 56–65. Somerville: Cascadilla Proceedings Project.



Chapter 1. Introduction 35

Diessel, Holger. 2012. “New Perspectives, Theories and Methods: Diachronic Change and Language Acquisition.” In Historical Linguistics of English: An International Handbook, ed. by Alexander Bergs, and Laurel Brinton, 1599–1613. Berlin, and New York: Mouton de Gruyter. Docherty, Gerard J., and Paul Foulkes. 2014. “An Evaluation of Usage-Based Approaches to the Modelling of Sociophonetic Variability.” Lingua 142: 42–56.  doi: 10.1016/j.lingua.2013.01.011 Dorleijn, Margreet, Maarten Mous, and Jacomine Nortier. 2015. “Urban Youth Speech in Kenya and the Netherlands.” In Language, Youth and Identity in the 21st Century: Linguistic Practices across Urban Spaces, ed. by Jacomine Nortier, and Bente A. Svendsen, 271–289. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9781139061896.019 Eckert, Penelope. 1989. Jocks and Burnouts: Social Identity in the High School. New York: Teachers College Press. Eckert, Penelope. 1997. “Why Ethnography?” In Ungdomssprak i Norden, ed. by Ulla-Britt Kotsinas, Anna-Brita Stenstrom, and Anna-Malin Karlsson, 52–62. Stockholm: Stockholm University. Eckert, Penelope. 2004. “Adolescent Language.” In Language in the USA, ed. by Edward Finegan, and John R. Rickford, 361–374. New York: Cambridge University Press.  doi: 10.1017/CBO9780511809880.021 Eckert, Penelope. 2012. “Three Waves of Variation Study: The Emergence of Meaning in the Study of Sociolinguistic Variation.” Annual Review of Anthropology 41: 87–100.  doi: 10.1146/annurev-anthro-092611-145828 Ervin-Tripp, Susan M. 1973. “The Structure of Communicative Choice.” In Language Acquisition and communicative choice, ed. by Susan M. Ervin-Tripp, and Anwar S. Dil, 302–373. Stanford: Stanford University Press. Fasold, Ralph W. 1991. “The Quiet Demise of Variable Rules.” American Speech 66 (1): 3–21.  doi: 10.2307/455431 Foulkes, Paul. 2010. “Exploring Social-Indexical Knowledge: A Long Past but a Short History.” Laboratory Phonology 1 (1): 5–40.  doi: 10.1515/labphon.2010.003 Foulkes, Paul, and Gerard J. Docherty. 2006. “The Social Life of Phonetics and Phonology.” Journal of Phonetics 34: 409–438.  doi: 10.1016/j.wocn.2005.08.002 Foulkes, Paul, Gerard J. Docherty, and Dominic J.L. Watt. 2001. “The Emergence of Structured Variation.” University of Pennsylvania Working Papers in Linguistics 7.3 – A Selection of Papers from NWAV 29: 67–84. Foulkes, Paul, Gerard J. Docherty, and Dominic J.L. Watt. 2005. “Phonological Variation in Child Directed Speech.” Language 81 (1): 177–206.  doi: 10.1353/lan.2005.0018 Garrett, Peter. 2007. “Language attitudes.” In The Routledge Companion to Sociolinguistics, ed. by Carmen Llamas, Louise Mullany, and Peter Stockwell, 116–121. London, and New York: Routledge. Garrod, Simon, and Martin J. Pickering. 2004. “Why is conversation so easy?” Trends in Cognitive Sciences 8 (1): 8–11.  doi: 10.1016/j.tics.2003.10.016 Geeraerts, Dirk. 2003. “Cultural Models of Linguistic Standardization.” In Cognitive Models in Language and Thought: Ideology, Metaphors, and Meaning, ed. by René Dirven, Roselyn Frank, and Martin Pütz, 25–68. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110892901.25 Geerts, Guido, Johan Nootens, and Jef van den Broeck. 1978. “Flemish Attitudes towards Dialect and Standard Language: A Public Opinion Poll.” International Journal of the Sociology of Language 15 (1): 33–46.  doi: 10.1515/ijsl.1978.15.33

36

Gunther De Vogelaer et al.

Ghimenton, Anna, Jean-Pierre Chevrot, and Jacqueline Billiez. 2013. “Language Choice Adjustments in Child Production during Dyadic and Multiparty Interactions: A Quantitative Approach to Multilingual Interactions.” Linguistics 51 (2): 413–438.  doi: 10.1515/ling-2013-0016 Giesbers, Herman. 1989. Code-switching tussen dialect en standaardtaal. Amsterdam: P.J. Meertens-Instituut. Giles, Howard, Richard Bourhis, Alan Lewis, and Peter Trudgill. 1974. “The Imposed Norm Hypothesis: A Validation.” Quarterly Journal of Speech 60 (4): 405–410.  doi: 10.1080/00335637409383249 Goodman, Yetla, and Kenneth Goodman. 1981. “Twenty Questions about Teaching Language.” Educational Leadership 38: 437–443. Gregoire, Henri-Baptiste. 1794. Rapport sur la nécessité et les moyens d’anéantir les patois et d’universaliser la langue française. Paris. Gülzow, Insa, and Natalia Gagarina. 2011. Frequency Effects in Language Acquisition: Defining the Limits of Frequency as an Explanatory Concept. Berlin, and Boston: De Gruyter Mouton. Häcki Buhofer, Annelies, and Harald Burger. 1998. Wie Deutschschweizer Kinder Hochdeutsch lernen: Der ungesteuerte Erwerb des gesprochenen Hochdeutschen durch Deutschschweizer Kinder zwischen sechs und acht Jahren. Stuttgart: Franz Steiner Verlag. Halle, Morris. 1962. “Phonology in Generative Grammar.” Word 18: 54–72.  doi: 10.1080/00437956.1962.11659765 Henry, Alison. 1995. Belfast English and Standard English: Dialect Variation and Parameter Setting. Oxford, and New York: Oxford University Press. Hinskens, Frans, Peter Auer, and Paul Kerswill. 2005. “The Study of Dialect Convergence and Divergence: Conceptual and Methodological Considerations.” In The Convergence and Divergence of Dialects in Contemporary Societies, ed. by Peter Auer, Frans Hinskens, and Paul Kerswill, 1–48. Cambridge: Cambridge University Press. Hollingworth, Brian. 1989. “Education and the Vernacular.” In Dialect and Education: Some European Perspectives, ed. by Cheshire, Jenny, Viv Edwards, Henk Münstermann, and Bert Weltens, 293–302. Clevedon: Multilingual matters. Hoyle, Susan M., and Carolyn T. Adger. 1998. “Introduction.” In Kids Talk: Strategic Language Use in Later Childhood, ed. by Susan M. Hoyle, and Carolyn T. Adger, 3–22. Oxford: Oxford University Press. Hyams, Nina M. 1986. Language Acquisition and the Theory of Parameters. Dordrecht et al.: Reidel.  doi: 10.1007/978-94-009-4638-5 Janssen, Hans. 1943. Leben und Macht der Mundart in Niedersachsen. Oldenburg: Gerhard Stalling. Jones, Peter E. 2013. “Bernstein’s ‘Codes’ and the Linguistics of ‘Deficit’.” Language and Education 27 (2): 161–179.  doi: 10.1080/09500782.2012.760587 Katerbow, Matthias. 2013. Spracherwerb und Sprachvariation: Eine phonetisch-phonologische Analyse zum regionalen Erstspracherwerb im Moselfränkischen. Berlin, and New York: Mouton de Gruyter.  doi: 10.1515/9783110309768 Katerbow, Matthias, Lena Eichele, and Christina Kauschke. 2011. “Einstellungen zu kindlicher Sprachvariation: Eine empirische Studie zur Fremdbewertung regionalsprachlicher und/ oder störungsbedingter Variation von Kindern.” Linguistische Berichte 228: 467–484. Kaufmann, Laurence, and Fabrice Clément. 2011. “L’esprit des sociétés. Bilan et perspectives en sociologie cognitive.” In La sociologie cognitive, ed. by Laurence Kaufmann, and Fabrice Clément, 7–40. Paris: Ophrys.



Chapter 1. Introduction 37

Kemmer, Suzanne, and Michael Barlow. 2000. “Introduction: A Usage-Based Conception of Language.” In Usage-based Models of Language Use, ed. by Michael Barlow, and Suzanne Kemmer, VII–XXVIII. Stanford: CSLI Publications. Kemmer, Suzanne, and Michael Israel. 1994. “Variation and the Usage-Based Model.” In Papers from the 30th Regional Meeting of the Chicago Linguistics Society: Para-Session on Variation and Linguistic Theory: Vol. 2, ed. by Katherine Beals, Jeannette Denton, Robert Knippen, Lynette Melnar, Hisami Suzuki, and Erica Zeinfeld, 165–179. Chicago: Chicago Linguistics Society. Kern, Friederike, and Margret Selting. 2011. Ethnic Styles in European Metropolitan Areas. Amsterdam: John Benjamins.  doi: 10.1075/silv.8 Kerswill, Paul. 1996. “Children, Adolescents, and Language Change.” Language Variation and Change 8 (2): 177–202.  doi: 10.1017/S0954394500001137 Kerswill, Paul. 2007. “Social Class.” In The Routledge Companion to Sociolinguistics, ed. by Carmen Llamas, Louise Mullany and Peter Stockwell, 51–61. London, and New York: Routledge. Kerswill, Paul, and Ann Williams. 2000. “Creating a New Town Koine: Children and Language Change in Milton Keynes.” Language in Society 29 (1): 65–115.  doi: 10.1017/S0047404500001020 Kinzler, Katherine D., and Jasmine M. DeJesus. 2013. “Northern = Smart and Southern = Nice: The Development of Accent Attitudes in the United States.” The Quarterly Journal of Experi­ mental Psychology 66 (6): 1146–1158.  doi: 10.1080/17470218.2012.731695 Kristiansen, Gitte, and René Dirven. 2008. Cognitive Sociolinguistics: Language Variation, Cultural Models, Social Systems. Berlin: Mouton De Gruyter.  doi: 10.1515/9783110199154 Kristiansen, Tore, and Jens Normann Jørgensen. 2011. “Subjective Factors in Dialect Convergence and Divergence.” In Dialect Change: Convergence and Divergence in European Languages, ed. by Peter Auer, Frans Hinskens, and Paul Kerswill, 287–302. Cambridge: Cambridge University Press. Kroon, Sjaak, and Ton Vallen. 2009. “Discoursen over dialect en school.” In De klank van de stad: Stedelijke meertaligheid en interculturele communicatie, ed. by Jürgen Jaspers, 165–181. Leuven: Acco. Labov, William. 1964. “Stages in the Acquisition of Standard English.” In Social Dialects and Language Learning, ed. by Roger W. Shuy, Alva Davis, Robert Hogan, 77–103. Champaign: National Council of Teachers of English. Labov, William. 1972a. Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press. Labov, William. 1972b. Sociolinguistic Patterns. Oxford: Blackwell. Labov, William. 1989. “The Child as Linguistic Historian.” Language Variation and Change 1 (1): 85–97.  doi: 10.1017/S0954394500000120 Labov, William. 2007. “Transmission and Diffusion.” Language 83 (2): 344–387.  doi: 10.1353/lan.2007.0082 Labov, William. 2014. “What Is to Be Learned? The Community as the Focus of Social Cognition.” In Cognitive Sociolinguistics: Social and Cultural Variation in Cognition and Language Use, ed. by Martin Pütz, Justyna A. Robinson, and Monika Reif, 23–51. Amsterdam: John Benjamins.  doi: 10.1075/bct.59.02lab Lacoste, Véronique, and Lisa Green. 2016. “Child Language Variation: Sociolinguistic and Formal Approaches.” Linguistic Variation 16: 1–11. Lafontaine, Dominique. 1986. Le Parti Pris des Mots. Bruxelles: Mardaga.

38

Gunther De Vogelaer et al.

Laks, Bernard. 2013. “Why is There Variation Rather than Nothing?” Language Sciences 39: 31–53.  doi: 10.1016/j.langsci.2013.02.009 Levinson, Stephen C. 2012. “The Original Sin of Cognitive Science.” Topics in Cognitive Science 4 (3): 396–403.  doi: 10.1111/j.1756-8765.2012.01195.x Lightfoot, David. 1999. The Development of Language: Acquisition, Change, and Evolution. Oxford: Blackwell. Loudermilk, Brandon C. 2013. “Psycholinguistic Approaches.” In The Oxford Handbook of Sociolinguistics, ed. by Robert Bayley, Richard Cameron, and Ceil Lucas, 132–152. Oxford: Oxford University Press. Macaulay, Ronald K.S. 1977. Language, Social Class, and Education: A Glasgow Study. Edinburgh: Edinburgh University Press. Makoni, Sinfree, and Alastair Pennycook. 2007. “Disinventing and Reconstituting Languages.” In Disinventing and Reconstituting Languages, ed. by Sinfree Makoni, and Alastair Pennycook, 1–41. Clevedon: Multilingual Matters. Martin, Nathaël. 2005. Réseaux sociaux et variations phonétiques. Grenoble: Université Stendhal Master 1 dissertation. Martin, Nathaël, Jean-Pierre Chevrot, and Stéphanie Barbu. 2010. “Stylistic Variations in the Social Network of a 10-Year-Old Child: Pragmatic Adjustments or Automatic Alignment?” Journal of Sociolinguistics 14 (5): 678–692.  doi: 10.1111/j.1467-9841.2010.00459.x Martino, Josephine. 1982. “The Phoneme /θ/ and Its Alternative Realization as /f/: A Study of Variation in Australian English among Primary School Boys, According to Socio-Economic Background.” Working Papers in Linguistics 8: 39–42. Mattheier, Klaus J. 1980. Pragmatik und Soziologie der Dialekte: Einführung in die kommunikative Dialektologie des Deutschen. Heidelberg: Quelle & Meyer. Mills, Anne E. 1985. “The Acquisition of German.” In: The Crosslinguistic Study of Language Acquisition, ed. by Dan I. Slobin, 141–254. Hillsdals: Lawrence Erlbaum. Mills, Anne E. 1986. The Acquisition of Gender. A Study of English and German. Berlin: Springer.  doi: 10.1007/978-3-642-71362-0 Molénat, Xavier. 2012. “Les sociologues à l’assaut de l’esprit.” Sciences humaines 233 (1): 14–14. Møller, Janus Spindler, and Jens Normann Jørgensen. 2009. “From Language to Languaging: Changing Relations between Humans and Linguistic Features.” Acta Linguistica Hafniensia 41: 143–166.  doi: 10.1080/03740460903364185 Morgan, David L., and Michael L. Schwalbe. 1990. “Mind and Self in Society: Linking Social Structure and Social Cognition.” Social Psychology Quarterly 3 (2): 148–164.  doi: 10.2307/2786676 Munson, Benjamin. 2010. “Levels of Phonological Abstraction and Knowledge of Socially Motivated Speech-Sound Variation.” Laboratory Phonology 1 (1): 157–177.  doi: 10.1515/labphon.2010.008 Nardy, Aurélie. 2008. Acquisition des variables sociolinguistiques entre 2 et 6 ans: facteurs socio­ logiques et influences des interactions au sein du réseau social. Grenoble: Université Stendhal Ph.D. dissertation. Nardy, Aurélie, Jean-Pierre Chevrot, and Stéphanie Barbu. 2013. “The Acquisition of Socio­ linguistic Variation: Looking Back and thinking Ahead.” Linguistics 51 (2): 255–284.  doi: 10.1515/ling-2013-0011



Chapter 1. Introduction 39

Nardy, Aurélie, Jean-Pierre Chevrot, and Stéphanie Barbu. 2014. “Sociolinguistic Convergence and Social Interactions within a Group of Preschoolers: A Longitudinal Study.” Language Variation and Change 26 (3): 273–301.  doi: 10.1017/S0954394514000131 Ostermai, Guido. 2000. Sprachvariation im Grenzbereich: Eine Untersuchung zur Standard­sprache nordwestschweizerischer und südbadischer PrimarschülerInnen. Aarau: Sauerländer. Patterson, Janet Lee. 1992. The Development of Sociolinguistic Phonological Variation Patterns for (ing) in Young Children. Albuquerque: University of New Mexico Ph.D. dissertation. Paul, Hermann. 1886. Prinzipien der Sprachgeschichte. Halle: Niemeyer. Payne, Arvilla. 1976. The Acquisition of the Phonological System of a Second Dialect. Ph.D. Dissertation, University of Pennsylvania. Pierrehumbert, Janet. 2001. “Stochastic Phonology.” GLOT 5 (6): 1–13. Pierrehumbert, Janet. 2002. “Word-Specific Phonetics.” In Laboratory Phonology: Vol. VII, ed. by Carlos Gussenhoven, and Natasha Warner, 101–139. Berlin: Mouton de Gruyter.  doi: 10.1515/9783110197105.101 Pierrehumbert, Janet. 2003. “Phonetic Diversity, Statistical Learning, and Acquisition of Phonology.” Language and Speech 46 (2–3): 115–154.  doi: 10.1177/00238309030460020501 Pierrehumbert, Janet. 2006. “The Next Toolkit.” Journal of Phonetics 34 (6): 516–530.  doi: 10.1016/j.wocn.2006.06.003 Pinker, Steven. 1999. Words and Rules: The Ingredients of Language. New York: Harper Collins. Poplack, Shana, Lauren Zentz, and Nathalie Dion. 2012. “Phrase-Final Prepositions in Quebec French: An Empirical Study of Contact, Code-Switching and Resistance to Convergence.” Bilingualism: Language and Cognition 15 (2): 203–225.  doi: 10.1017/S1366728911000204 Quéré, Louis. 2008. “How Could Trust be Restored to Nature?” Intellectica 46–47: 105–122. Reid, Euan. 1978. “Social and Stylistic Variation in the Speech of Children: Some Evidence from Edinburgh.” In Sociolinguistic Patterns in British English, ed. by Peter Trudgill, 158–171. London: Edward Arnold. Roberts, Julie. 1994. Acquisition of Variable Rules: (-t, d) Deletion and (ing) Production in Preschool Children. Philadelphia: Ph.D. Dissertation, University of Pennsylvania. Roberts, Julie. 1997. “Acquisition of Variable Rules: A study of (-t, d) Deletion in Preschool Children.” Journal of Child Language 24 (2): 351–372.  doi: 10.1017/S0305000997003073 Rodina, Yulia, and Marit Westergaard. 2013. “The Acquisition of Gender and Declension Class in a Non-Transparent System: Monolinguals and Bilinguals.” Studia Linguistica 67 (1): 47–67.  doi: 10.1111/stul.12012 Romaine, Suzanne. 1984. The Language of Children and Adolescents: The Acquisition of Com­ municative Competence. Oxford: Blackwell. Rosenberg, Peter. 1989. “Dialect and Education in West Germany.” In Dialect and Education: Some European Perspectives, ed. by Jenny Cheshire, Viv Edwards, Henk Münstermann, and Bert Weltens, 62–93. Clevedon, and Philadelphia: Multilingual matters. Rys, Kathy. 2007. Dialect as a Second Language: Linguistic and Non-Linguistic Factors in Secondary Dialect Acquisition by Children and Adolescents. Ph.D. Dissertation, Ghent University. Sankoff, Gillian, and Hélène Blondeau. 2007. “Longitudinal Change across the Lifespan: R in Montreal French.” Language 83 (3): 560–588.  doi: 10.1353/lan.2007.0106 Schulte Kemminghausen, Karl. 1939. Mundart und Hochsprache in Norddeutschland. Neu­ münster: Karl Wachholtz Verlag.

40 Gunther De Vogelaer et al.

Schultze, Fritz. 1880. “Die Sprache des Kindes: Eine Anregung zur Erforschung des Gegenstandes.” Darwinistische Schriften 10: 1–46. Siegel, Jeff. 2010. Second Dialect Acquisition. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511777820 Smith, Jennifer, Mercedes Durham, and Liane Fortune. 2007. ““Mam, My Trousers is Fa’in Doon!”: Community, Caregiver, and Child in the Acquisition of Variation in a Scottish Dialect.” Language Variation and Change 19 (1): 63–99.  doi: 10.1017/S0954394507070044 Smith, Jennifer, Mercedes Durham, and Hazel M. Richards. 2013. “The social and linguistic in the acquisition of sociolinguistic variation.” Linguistics 51 (2): 258–324.  doi: 10.1515/ling-2013-0012 Squires, Lauren. 2013. “It Don’t Go Both Ways: Limited Bidirectionality in Sociolinguistic Perception.” Journal of Sociolinguistics 17 (2): 200–237.  doi: 10.1111/josl.12025 Stern, Otto. 1988. “Divergence and Convergence of Dialect and Standard from the Perspective of the Language Learner. Standard Language Acquisition by the Swiss-German Dialect Speaking Child.” In Variation and Convergence: Studies in Social Dialectology, ed. by Peter Auer, and Aldo Di Luzio, 134–156. Berlin: Mouton De Gruyter.  doi: 10.1515/9783110851601.134 Tesink, Cathelijne, Karl-Magnus Petersson, Jos Van Berkum, Danielle Van den Brink, Jan Buitelaar, and Peter Hagoort. 2009. “Unification of Speaker and Meaning in Language Comprehension: An fMRI Study.” Journal of Cognitive Neuroscience 21 (11): 2085–2099.  doi: 10.1162/jocn.2008.21161 Tomasello, Michael. 2003. Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, and London: Harvard University Press. Trudgill, Peter. 1986. Dialects in contact. Oxford, and New York: Blackwell. Valian Virginia, and Andrea Levitt. 1996. Prosody and Adults’ Learning of Syntactic Structure. Journal of Memory and Language 35: 497–516.  doi: 10.1006/jmla.1996.0027 Van Berkum, Jos, Danielle Van den Brink, Cathelijne Tesink, Miriam Kos, and Peter Hagoort. 2008. “The Neural Integration of Speaker and Message.” Journal of Cognitive Neuroscience 20 (4): 580–591.  doi: 10.1162/jocn.2008.20054 Vihman, Marilyn M. 1980. “Sound Change and Child Language.” In Papers from the 4th International Conference on Historical Linguistics, ed. by Elizabeth C. Traugott, Rebecca La Brum, and Susan Shepherd, 303–320. Amsterdam: John Benjamins. Weijnen, Antoon. 1967. “Sociodialectologische onderzoekingen in Limburg.” In ‘Taalsociologie’: Bijdragen en Mededelingen der Dialectencommissie van de Koninklijke Nederlandse Akademie van Wetenschappen, ed. by Jo Daan, and Antoon Weijnen, 16–27. Amsterdam: NoordHollandse Uitgeversmaatschappij. Weinreich, Uriel, William Labov, and Marvin Herzog. 1968. “Empirical Foundations for a Theory of Language Change.” Directions for Historical Linguistics, ed. by Winfried Lehmann, and Yakov Malkiel, 95–195. Austin: University of Texas Press. Westergaard, Marit R. 2004. “The Interaction of Input and UG in the Acquisition of Verb Movement in a Dialect of Norwegian.” Nordlyd 32: 110–134. Wiese, Heike. 2013. “What Can New Urban Dialects Tell Us about Internal Language Dynamics? The Power of Language Diversity.” In Dialektologie in neuem Gewand: Zu Mikro-/Varie­ tät­enlinguistik, Sprachenvergleich und Universalgrammatik, ed. by Werner Abraham, and Elisabeth Leiss, 207–245. Hamburg: Helmut Buske Verlag.



Chapter 1. Introduction 41

Wilson, John, and Alison Henry. 1998. “Parameter Setting within a Socially Realistic Linguistics.” Language in Society 27: 1–21.  doi: 10.1017/S0047404500019709 Yang, Charles. 2000. “Internal and External Forces in Language Change.” Language Variation and Change 12: 231–250.  doi: 10.1017/S0954394500123014 Yang, Charles. 2010. “Three Factors in Language Variation.” Lingua 120: 1160–1177.  doi: 10.1016/j.lingua.2008.09.015

Chapter 2

The effects of exposure on awareness and discrimination of regional accents by five- and six year old children Erica Beck

University of Michigan

This study compares bi-dialectal and mono-dialectal five-and six-year-old children’s sociolinguistic awareness and ability to discriminate regional accents in their native language. Children who regularly hear multiple regional varieties in their input are expected to have better awareness and discrimination accuracy. The children participated in two tasks: an awareness task, assessing their knowledge of regional variation, and a similarity judgment task, assessing their ability to discriminate between speakers based on accents. Results show that both groups reliably discriminate between regional accents, and can identify a local regional accent. However, no advantage is found for either group of participants on either task. The effects of exposure to regional phonological variation on perception and awareness are discussed in light of these findings. Keywords: regional accent discrimination, developmental speech perception, acquisition of socio-indexical knowledge, influence of socio-indexical information on speech perception

1. Introduction Adults have been shown to reliably use phonological regional variation to identify a speaker’s home region (Labov 1998; Clopper and Pisoni 2004; Clopper, Rohrbeck, and Wagner 2012), and long-term exposure to regional variation has been shown to improve identification of regional accents (Clopper and Pisoni 2004), as well as the processing of accented speech (Sumner and Samuel 2009). Comparatively less is known about children’s perception of regional variation, and the role exposure plays in developing the ability to perceive and interpret it (see Cristia et al. (2012) for a comprehensive review of extant studies on accent perception by children). Some studies have found children aged five and six are unable doi 10.1075/silv.20.02bec © 2017 John Benjamins Publishing Company

44 Erica Beck

to categorize speakers by regional accent in their native language, although they succeed at categorizing native and non-native accents (Girard, Floccia, and Goslin 2008; Floccia et al. 2009; Wagner, Clopper, and Pate 2014). In one study, 7-year-old children were shown to successfully distinguish between regional accents, with children from multi-dialectal families outperforming those from mono-dialectal families (Floccia et al. 2009), suggesting that even at an early age, exposure to regional variation may affect perception of regional accents. The present study asks whether exposure to regional variation has an effect on children’s awareness of regional accents and whether this awareness improves their ability to discriminate between them. The participants are two groups of children living in the same town, but having different histories of exposure to regional variation: children whose parents are natives of the same town as the children (Insiders), and those with parents who moved to that town from another regional dialect area (Outsiders). This chapter describes two tasks, one comparing these five-and-six-year-olds’ ability to discriminate between a local and non-local regional accent, and one comparing their meta-linguistic awareness of the two accents heard in the stimuli. A mixed effects logistical regression model is used to examine the effects of awareness on discrimination ability. 2. What five to seven year old children know about regional accent 2.1

Production of regional varieties

Children have been shown to produce linguistic variables specific to their region of residence from the age of three (Roberts and Labov 1995; Roberts 1997a; Roberts 1997b; Roberts 2005), but their acquisition of those variables depends heavily on the parents’ native dialect and use of regional variation in child-directed speech (Payne 1976; Foulkes, Docherty, and Watt 2005; Smith, Durham, and Fortune 2007; Smith, Durham, and Richards 2013). Roberts (1997b) found that for some of the regional variables studied, pre-school-aged children were most likely to adopt Philadelphia-specific sound changes when both parents were from Philadelphia. Studying older children who re-located to a new dialect region, Payne (1976) also found that acquisition of some regional phonological patterns depended on the native dialect of the parents. Smith and colleagues (Smith, Durham, and Fortune 2007; Smith, Durham, and Richards 2013) report that the frequency and patterning of particular regional variables in child-directed speech, along with contextual and grammatical constraints, influence the age and rate at which children acquire some markers of regional variation. These studies underscore the importance of



Chapter 2.  Children’s accent discrimination 45

the parents’ input in the acquisition of regional phonological features in early childhood. They also show that children are able to acquire the features and patterns of sociolinguistic variation at a young age, although the studies do not examine whether children are aware of that variation. The parents’ role in influencing production may be reduced once children reach school age. Kerswill and Williams (2000), in a study of children moving to a new dialect region, found that four-year-olds’ speech mirrored their parents’ regional dialects. By the age of eight, chidren were speaking the regional variety of their peers. Taken together, the results of these studies suggest not only sensitivity to the phonological patterns in the environment, but also to the social importance of those patterns in the children’s community at different stages of development. 2.2

Children’s discrimination of regional variation

Experimental work on children’s ability to discriminate regional variation suggests that the development of this skill does not progress in a linear fashion. Multiple studies find that between three and five months of age, children are able to distinguish the local accent from another accent in their language (Nazzi, Jusczyk, and Johnson 2000; Kitamura et al. 2006; Egerova 2010; Butler et al. 2011). However, between eight and eleven months of age, the ability to distinguish between local and non-local accents appears to decline (Kitamura et al. 2006; Phan and Houston 2008), and this inability persists through at least six years of age (Girard, Floccia, and Goslin 2008; Floccia et al. 2009; Wagner, Clopper, and Pate 2014). During the time period in which children are reportedly unable to discriminate between regional varieties, they maintain the ability to discriminate between foreign and native accents in their native language (Girard, Floccia, and Goslin 2008; Floccia et al. 2009; Wagner, Clopper, and Pate 2014). The three studies of five-year-olds’ ability to discriminate between regional accents cited above focus on three different language communities: native French speakers (Girard, Floccia, and Goslin 2008), native English speakers in the United Kingdom (Floccia et al. 2009) and native English speakers in the United States (Wagner, Clopper, and Pate 2014). These studies use a categorization paradigm, in which children are asked to sort sentence-length stimuli into two groups by accent (Girard, Floccia, and Goslin 2008; Floccia et al. 2009; Wagner, Clopper, and Pate 2014). In all three studies, five- and six-year-old children performed at chance, although in one of the studies, seven-year-old children were able to successfully categorize by accent (Floccia et al. 2009). In addition to the categorization paradigm, Girard, Floccia, and Goslin (2008) also test accent discrimination using an AX task, and an accent/voice matching task, in both of which 5-year-olds also failed to discriminate between accents.

46 Erica Beck

In the Floccia et al. (2009) study, seven-year-olds succeed in discriminating between accents in the categorization task, and seven-year-olds from bi-dialectal homes out-perform their mono-dialectal peers. Although not explicitly measured, it is surmised that bi-dialectal children have an advantage in awareness of regional variation, which improves their ability to identify and categorize the accents heard in that task (Floccia et al. 2009). 2.3

Association of regional varieties with social meaning

Awareness of regional variation appears to be still developing between the ages of five and seven, based on the findings of several studies of children’s social perception of accents. Cremona and Bates (1977) found Italian five-year-olds unable to distinguish their local regional dialect from Standard Italian. Despite this, the children in that study expressed negative evaluations of speakers of the regional dialect, which was also the children’s native dialect. Millar (2003) documented similar behavior in children evaluating speakers of Standard and regional Danish; some students as old as eight years of age incorrectly assigned stereotypes of regional dialect speakers to a Standard Danish speaker. This suggests that it is possible for children to be aware of the social status of regional varieties in their community, even if they cannot accurately identify them. The context of children’s exposure to regional variation may play a role in how it is interpreted for social meaning. Kinzler and DeJesus (2013) conducted a study of five- and six-year-old children in the Northern and Southern U.S., and found that they were not equally able to identify or assign social evaluations to those regional accents. Children from the North, where the more prestigious dialect is spoken, could reliably identify it and showed social preferences for speakers with a Northern accent. Children living in the Southern U.S., where a less prestigious regional variety is predominately spoken, were at chance when identifying the varieties, and showed no social preferences either accent. The crucial difference is that Southern children typically have exposure to both Northern and Southern accents whereas Northern children would mainly have exposure to the Northern accent. In this context, hearing multiple regional varieties did not give children an advantage in awareness of regional variation. Foulkes and Docherty (2006) suggest that regional variation may be hard for children to identify because in many contexts there are no visual correlates for regional variation, as there is for ethnic or gender variation in speech. They predict that children may have an easier time learning to associate linguistic variables with a speaker’s age, gender, or race, than with place of origin.



Chapter 2.  Children’s accent discrimination 47

Munson (2010) proposes a model explaining how children would begin to correctly identify and label social variation they hear in speech. He hypothesizes that children form initial representations based on individuals they interact with, creating an association between that person’s accent and the place the speaker is from. This is referred to as the “library of speakers” stage, in which representations of accents are associated with specific individuals, and not recognized as a category of linguistic variation. As children’s exposure to that regional variety broadens, their representation of the regional accent generalizes from the speech of a specific individual to all individuals from that geographical region. It follows from this model that children with more exposure to regional variation in childhood may be better at identifying and discriminating regional accents than those with less exposure, as they will have more established categories of regional variation. It should also be the case that because category formation is predicated on the knowledge that a speaker’s accent marks them as being from a particular region, children with exposure to multiple accents will also be aware of the association between regional accents and geographical locations. This, in theory, should give Outsiders an advantage in awareness of regional variation. However, if they have not abstracted their representations of regional accent beyond individual speakers, they may not be able to apply this knowledge to novel regional accents they encounter. Since no previous work had explicitly tested the effects of awareness and exposure on perception of regional accents in childhood, I designed a study to examine whether these factors play a role in five-year-olds’ ability to discriminate regional variation (Beck 2014). This chapter presents the results for two groups of participants from that study. The design and results of the larger study are reviewed in the following section to give context to the findings in the present chapter. 2.4

Description of the Beck (2014) perception study

In a series of three tasks, 66 five- and six-year-old children were tested on their ability to discriminate between an unfamiliar and familiar regional accent, as well as their ability to identify regional accents and interpret regional variation for social information about the speaker. The tasks were an accent awareness task, an ABX discrimination task, and a second discrimination task in which children had to select a speaker sounding most similar to their own speech. The same two accents were contrasted in the stimuli for all three tasks: the local regional variety from the children’s hometown and a U.S. Southern variety from Northern Louisiana in the United States. The subjects were all born and raised in the same town near

48 Erica Beck

Philadelphia, USA, and had been attending kindergarten for four months at the time of testing. The Awareness Task used a set of five questions to assess their ability to identify the specific regional accents heard in the study and their knowledge about regional variation more generally. The children were asked to identify the local accent and the non-local accent used in these stimuli. Seventy-two percent of the five- and six-year-olds were able to identify the local regional accent as being from their hometown. Only 10% could identify the non-local accent correctly, although approximately 50% could state the speakers with the unfamiliar accent were not from their hometown. Second, they were asked why the two sets of speakers heard in the task sounded different, to test whether they could identify regional variation. Forty percent correctly identified regional variation as the difference between the two sets of speakers heard in the task. Finally, they were asked to show the geographical location of their hometown relative to other places, so that the effect of geographical knowledge on understanding of regional variation could be assessed. Approximately 75% of the children could identify their hometown on a map, but only half knew where it was in relation to any other place in the U.S. Children also completed two discrimination tasks. The first discrimination task used an ABX paradigm 1 with single-word stimuli. The mono-dialectal participants discriminated between regional accents with a 79% success rate in the ABX task, although they were better at matching local speakers than the non-local speakers. In the second discrimination task, the Similarity Judgment Task, the participants were asked to choose the speaker that sounded most similar to them, given the choice between the local and non-local accent. An analysis was conducted using a mixed effects logistic regression model, to test whether their knowledge of regional accents, as demonstrated in the Awareness Task, affected their responses in the ABX and Similarity Judgment tasks. For the monolinguals, there was a strong effect of correctly identifying the local regional accent on correctly matching the local speakers in the ABX Task. There was also a strong effect of identifying the local speakers on choosing them as most similar sounding in the Similarity Judgment Task. This suggests that not only can children discriminate between regional accents, but that they can draw on the social meaning of those accents to guide them in discrimination. 1. In the ABX paradigm, participants hear two tokens of the same word, spoken in each of the two accents used in the study. These are the A and B tokens. They then hear a third token (the X token), which is spoken in one of the same two accents heard in the A and B tokens. Participants are asked to decide whether the X-token is more similar to the A or B token in that trial.

Chapter 2.  Children’s accent discrimination 49



3. Testing the effect of early exposure to accents on accent discrimination ability In the present work, I test whether children who receive exposure to multiple regional dialects are more aware of regional variation, and whether as a result, they are better at discriminating between regional accents than children who receive the majority of their input in a single regional accent. Because exposure to variation is thought to facilitate the construction of sociolinguistic categories (e.g. Munson 2010), children who regularly interact with speakers of different regional varieties should be more aware of regional variation than those who do not have that interaction. The hypotheses for this study are as follows: 1. Participants with more exposure to regional variation should be more aware of regional variation than children from mono-dialectal families, thus Outsiders should answer Awareness Task questions correctly at a higher rate than Insiders. 2. Outsiders (bi-dialectal children) will choose the local speakers more categorically in the Discrimination Task than Insiders (mono-dialectal children). 3.1

Experiment 1: Awareness Task

This task uses the same five questions described in the previous section to assess whether five- and six-year-olds are aware of the meaning of regional variation, and whether Outsiders are more aware of regional accents than their Insider counterparts. Results of two measures of awareness from the Parent Questionnaire are also reported in this section. Additionally, the travel history for these two groups of children, as reported by the parents, is detailed here. Including this factor will allow for examination of the effect of short-term exposure to regional variation on children’s responses in the Discrimination Task. 3.1.1

Methods

Participants Using the data on participants’ family and language background collected in the Parent Questionnaire, children were divided into two groups based on their parents’ residential history. The Outsiders (n = 10) were from families in which both parents had moved from another dialect region to the town in which the study was conducted. The participants themselves were all born and raised in that town. Since the regional dialect is limited to a relatively small geographical area, although the parents’ speech wasn’t analyzed, it can be said with a degree of confidence that

50

Erica Beck

adults from outside the child’s hometown did not speak with the regional accent of that town. The Insiders (n = 20) were children whose parents were native to the town where the study was conducted. The participants ranged in age from 61 to 77 months, average age 70.4 months, or 5;10. They had been enrolled in kindergarten in public school near Philadelphia, Pennsylvania, U.S.A. for four months when the study was conducted. None of the participants had any prior extended exposure to the non-local accent used in this study. Although they were Outsiders, 13 bilingual participants were excluded from the analysis to avoid confounds with English proficiency. The remaining participants were Caucasian, monolingual English speakers. None of the Outsiders were observed to speak with a non-local regional accent.

Materials Stimulus materials included five questions posed orally by the experimenter, a map of the United States and four audio clips of speakers representing two different regional accents (Philadelphia and General Southern). The accents used are typically associated with Caucasians in both Philadelphia and the Southern United States. For a detailed description of the stimuli, see the methods section in the description of Experiment 2. The five questions used in this task were as follows: 1. Can you show where we live (while looking at map of U.S.)? 2. Can you show me and name any other places you know (while looking at map of U.S.)? 3. (After hearing a short clip of a local speaker) Does this person sound like he lives here? (If answer is no, have the child say where person is from.) 4. (After hearing a short clip of a non- local speaker, pronouncing the same word as the local speaker above) Does this person sound like he lives here? (If answer is no, have the child say where person is from.) 5. Can you guess why these two people talk differently?

Procedure The study was conducted during school hours, in a quiet corner where children were often administered tests or individual instruction. The Awareness Task was always presented after the experimental tasks in the study. This was done in order to avoid biasing responses on the discrimination tasks, since no reference was made to accents or regional variation in the instructions in those tasks. Parents were paid $10 for filling out and returning a questionnaire to school prior to the experimental part of the study. The questionnaire asked about families’ residential history, whether the children interacted with speakers with non-local

Chapter 2.  Children’s accent discrimination 51



regional or foreign accents, exposure to accents on TV and in the media, travel to areas where other regional accents are spoken, and the child’s propensity to imitate or comment on others’ accents. 3.1.2 Results The five questions asked of the children were scored as 1 (correct) or 0 (incorrect). For the purposes of conducting statistical analyses, if no answer was provided, it was scored as incorrect. For Q4, any answer indicating the speaker was not from the child’s hometown was counted as correct. Table 1 below summarizes the responses for each group. Table 1.  Summary of responses from Outsiders and Insiders Sub-group Q1 Q2 Q3 Q4 Q5

Insiders (n = 20)

Outsiders (n = 10)

Fisher’s Exact Test, 2-sided p-value

  70% (14)   50% (10)   85% (17)   50% (10) 45% (9)

90% (9) 80% (8) 70% (7) 50% (5) 40% (4)

   0.372    0.235    0.372 1 1

The differences in proportions of correct answers for the Insiders and Outsiders were calculated using Fisher’s Exact Test for Count Data. As the p-values show, the differences in proportions of correct answers were non-significant for all five questions. These tests indicate that neither Outsider nor Insider children have a relative advantage in awareness of regional accents. In addition to the questions in the Awareness Task, the Parent Questionnaire included two questions on children’s awareness of accents. The first asked parents to report if the child spontaneously imitated accents. The second asked if the child ever commented on others’ accents. 45% of the Insiders (9/20) and 90% (9/10) Outsiders were reported by the parents to make comments on others’ accents. 45% (9/20) of Insiders and 70% (7) of Outsiders were reported to imitate accents spontaneously. Using the Fisher’s Exact Test, the proportions of Insiders and Outsiders who imitated other accents was not significantly different, while commenting was under the p = 0.05 significance threshold (p = 0.024 for commenting, p = 0.260 for imitation). The Questionnaire item asking about children’s travel is also included in the analysis, in order to account for effects of short-term exposure to different regional accents. Only travel to different dialect regions was included. Nine of 10 Outsiders were reported to travel regularly to different dialect regions, almost exclusively to the area where their parents were born. In contrast, only 6 of 20 Insiders reported regular travel, mainly to vacation areas such as beach towns in New Jersey and Florida.

52

Erica Beck

3.1.3 Conclusion This task provides a picture of what five and six year old children know about regional variation. It appears that many have begun to associate regional phonological variation with information about a speaker’s hometown. A majority of participants can identify the local accent of their hometown, and approximately half know that an unfamiliar accent represents a non-local speaker. Forty percent of all participants are aware that the difference between the speakers heard in the stimuli is regional variation. However, there are not significant differences between the knowledge of Insiders and Outsiders on any item in this task. Floccia et al. (2009) surmised that the superior performance in accent categorization by seven-year-olds from bi-­dialectal backgrounds may have been the result of increased awareness, as compared to their mono-dialectal peers. For the current group of five- and six-yearolds, it appears that early, intensive exposure to a different dialect at home does not confer increased awareness. However, this should be tested in future work with a larger sample size, in order to ensure the lack of difference is not a result of low statistical power. 3.2

Experiment 2: Discrimination Task

This task tests children’s ability to discriminate between two regional accents of their native language. The experiment was created to encourage children to consider the speaker’s social characteristics during discrimination, asking children to select the speaker who sounds most like them, given a choice between a local and non-local regional accent in each trial. None of the children in this task spoke like the speakers with the non-local accent, or had any significant prior exposure to that regional variety. The hypothesis for this task was that Outsiders were more likely to categorically select the local speakers as sounding similar to themselves, as they were thought to be more aware of regional variation generally, and thus better able to identify their own regional variety. 3.2.1

Methods

Participants The same set children from Experiment 1 participated in this task1. Materials A list of 30 stimuli words was created using six vowel-quality differences between the Philadelphia and General Southern accents (Labov, Ash, and Boberg 2006; Schneider 2008). The six vowel groups are (words in block letters following Wells



Chapter 2.  Children’s accent discrimination 53

(1982): FACE, PRICE, GOAT, GOOSE, peel and tail. In the peel and tail class, the there is a difference in the vowel quality before /l/ in the two accents. In Philadelphia “peel” is pronounced [pil] and in General Southern [pɪl], whereas for the word hill the pronunciation in Philadelphia is generally [hɪl] and in General Southern [hil]. For the tail-type words, [e] and [ɛ] are reversed, giving [tel] in Philadelphia but [tɛl] in General Southern, and the reverse vowel qualities for the word tell. In GOOSE class words, the /u/ is more fronted in Southern than in Philadelphia, and often preceded by the glide /j/. The PRICE vowel that is the diphthong [aɪ] in Philadelphia but is pronounced as the monophthong [a:] in Southern. Finally the GOAT class has a more fronted [oʊ] in Southern than in Philadelphia. The stimuli for this task were created from recordings of six Caucasian male speakers, three for each regional accent, 25 to 35 years of age, and all lifelong residents of their respective hometowns. The local speakers were all from the same town as the children. The non-local speakers were all from the same town in northern Louisiana, and speakers of General Southern American English. All words were recorded in the carrier phrase “say __ again” for uniformity of pitch. The stimuli were sampled at 44 kHz. The task was presented using Microsoft PowerPoint slides. For each sound clip presented in a trial, a small icon representing the sound clip was placed under a heading labeling it as either A or B, so children could indicate which speaker they chose. Each of the four pseudo-randomized orders of the experiment was presented to one-quarter of the participants, so that all orders were evenly distributed across participants, and post-hoc checks were done to ensure there was no effect of order or item.

Procedure The study was conducted during school hours and this task always directly preceded the Awareness Task. Children participated in a familiarization session prior to completing the Discrimination Task, in which they saw and were asked to name pictures of the stimuli words. Participants also were given three warm-up trials to ensure they understood the task. No feedback was given in the warm-up trials, as children were to choose similar sounding speakers on whatever measure they chose. Also, the children were never told that they were to use accent or the speaker’s hometown to choose similar sounding speakers. This was done to ensure they used whichever heuristic they found most appropriate for judging similarity. 3.2.2 Results The raw results are reported both as the average number of times the local speaker was chosen, as the children chose the local speakers in a majority of the trials. A mixed effects logistic regression model is used to test for effects of awareness on

54

Erica Beck

similarity judgments. The model used is a multi-level logistic regression model with binomial link function in the library lme4, written by Bates, Maechler, and Bolker (2011) in the R Statistical Package (R Core Development Team 2011). In this model, two kinds of effects on the responses are included in the analyses: random and fixed. The fixed effects are the independent variables taken from the Awareness Task and Parent Questionnaire. The random effects are those variables which are not measured, but may exert influence on the responses provided. In this case, the item (the stimulus word heard in that trial) and the individual participant are random effects. This accounts for individual variation in the responses, as well as any difficulty particular words present across the participant group, providing a more nuanced analysis of effects. The binomial link function allows the model to accommodate categorical, as opposed to normally distributed, data. All measures of awareness and exposure from the Questionnaire and Awareness Task were plotted against the Discrimination Task performance to look for potential effects. Those items that patterned with performance were included in the statistical model as fixed effects. Awareness Task questions Q3, Q4 and Q5 and three items from the Parent Questionnaire were included in the analysis. Q3 and Q4 asked participants to identify where the speakers heard in the task were from (local vs. non-local). Q5 asked participants to state why the two sets of speakers spoke differently, to see if they could recognize regional variation. The three factors taken from the Parent Questionnaire were travel to other dialect regions, whether children were reported to comment on others’ accents, and whether children were reported to spontaneously imitate accents. The average number of times the Insiders (n = 20) picked the local speaker as sounding most similar was 18.05/25 (72%). This was significantly higher than chance in a one-tailed t-test (t = 10.1187, p-value = 6.116e-10). The average number of trials in which Outsiders (n = 10) picked the local speaker was 17.2/25 (66%). In a one-tailed t-test, this was significantly different than chance (t = 5.6572, p-­ value = 0.0003). There was not a significant difference between the scores of the two groups, (t = 0.4262, df = 14.924, p = 0.6761) in a two-tailed t-test. Both Insiders and Outsiders selected the local accent as most similar sounding a majority of the time, and with equal frequency. Figure 1 below compares the average number of times each group judged the local accent similar to their own speech in this experiment, with error bars.

Chapter 2.  Children’s accent discrimination 55



25

20

15

10

5

0

Insiders

Outsiders

Chart 1.  Mean number of times local accent judged most similar

In Tables 2 and 3 below, the coefficients for the fixed effects used in the model are given along with their p-values for each of the two participant groups. Table 2.  Fixed effects for awareness and exposure items, Insiders (n = 20) Q3 Q4 Q5 Imitate Comment Travel

Estimate

Sth. error

Z value

Pr(>|z|)

 1.0073 −0.8173  0.8956  0.6120 −1.0498  0.2842

0.5571 0.5024 0.3896 0.3057 0.4202 0.3850

 1.808 −1.627  2.299  2.002 −2.499  0.738

0.0706 0.1038 0.0215 0.0453 0.0125 0.4604

Table 3.  Fixed effects for awareness and exposure items, Outsiders (n = 10) Q3 Q4 Q5 Imitate Comment/travel

Estimate

Sth. error

Z value

Pr(>|z|)

 0.49931 −1.95048  0.93704  0.05317 −3.26204

0.34813 0.41097 0.42152 0.39285 1.12331

 1.434 −4.746  2.223  0.135 −2.904

0.15149 2.07e-06 0.02622 0.89234 0.00368

In the mixed effects logistical regression model, no two fixed effects that have the same values across the subjects can be calculated, which is the case for commenting

56

Erica Beck

on accents and travel for the Outsider subjects. The model can be run twice, once with commenting as the fixed effect, and once with travel, and will result in the same coefficients for these two effects. They cannot, however, be included together, as they do not vary independently. For both Insiders and Outsiders, recognizing that regional variation was the main difference between the speakers (Q5) positively affects their choice of the local speakers as sounding most similar. Imitation was thought to be a indicator of awareness of accents, and Insiders show a positive correlation between parents’ reports of imitating accents and selecting the local speaker. Outsiders do not have the same positive effect of imitation on selecting the local speaker, but do have a negative effect of correctly identifying the non-local accent (Q4) on judging the local speakers as most similar. Both groups also show a negative effects of having been reported to comment on accents and choosing the local speaker. This was unexpected, as commenting was hypothesized to be an indicator of awareness of regional variation, thought to facilitate the choice of the local speaker as most similar. A post-hoc analysis of the effect of item does not find that any particular item is more likely to effect the selection of the local speaker in this task for either group. 3.2.3 Conclusion Contrary to the initial hypothesis, Outsiders showed no advantage over Insiders in discriminating between accents. Both groups discriminate between accents reliably, and choose the local speakers as most similar sounding a majority of the time. It appears that for this age group, exposure to multiple regional varieties in early childhood does not affect discrimination ability, either positively or negatively. The difference that exposure had on discrimination was found in the effects of awareness on responses to the Discrimination Task. Both groups show positive effects of correctly identifying regional variation on choosing the local speaker in this task. Surprisingly, for both Insiders and Outsiders, children who were reported to spontaneously comment on accents were less likely to choose the local speaker in this task. The main differences between the two groups was that for Outsiders, both travel and having correctly identified the non-local accent in the Awareness Task (Q4) were more likely to choose the non-local speaker as most similar. How each of these effects relates to children’s awareness of regional variation is discussed below.



Chapter 2.  Children’s accent discrimination 57

4. Discussion This study finds that both Insider and Outsider five- and six-year old children demonstrate the ability to discriminate regional accents, and have a nascent understanding of the association of accent with a speaker’s place of residence. No advantages of long-term exposure to regional variation are in evidence for children at this age. Previous work suggested that increased awareness of regional variation enhances discrimination ability (e.g. Floccia et al. 2009). This did not hold for the two groups examined in the present work, but measures of awareness are shown to influence children’s responses in the Discrimination Task, although not in the uniform manner expected, based on hypothesized differences in awareness of regional variation between the two groups. Below, the principal factors found to influence responses on the Discrimination Task are discussed with respect to how they reflect children’s history of exposure to regional variation. The main hypothesis for this study was that increased awareness would predict near-categorical choice of the local speaker in this task, as the participants would recognize the speaker as being from the same town as they. For both Insiders and Outsiders, there was a positive effect of correctly recognizing regional variation (Q5) on choosing the local speaker in the Discrimination Task. However, no effect of Q5 was found in the post-hoc item analysis. Thus it seems that some children are using this knowledge as a heuristic in making similarity judgments, but that its use depends on the individual participant and item. Insiders who were reported to imitate accents were more likely to choose the local speaker. This effect is in line with predictions made for the study that awareness, represented by imitating accents, increases the likelihood that the child chooses the local speakers as most similar. The Outsiders, however, showed several unexpected effects of awareness on discrimination. First was the negative effect of correctly identifying the Southern speakers on choosing the local speakers in the Discrimination Task. This knowledge was expected to support choosing the local speakers in the Discrimination Task, as children would realize the non-local speakers were not from their hometown. The negative effect of correctly identifying the non-local accent (Q4) on choosing local speakers may be an expression of some Outsiders’ realization that their families do not belong to the linguistic majority of their hometown. Children who realized the Southern speakers were non-local might express the fact they don’t perceive themselves or their families as locals in their town by affiliating themselves with the non-local speakers in some trials. Children were not asked to explain their

58

Erica Beck

choices in each trial, so for the current study, this explanation remains a point of conjecture. However, similar behavior was documented in the Beck (2014) study. Some participants in that study who attended remedial language classes consistently chose the non-local speakers in this task. One of these participants explained that she her responses reflected the fact she did not talk “normally,” and therefore sounded like the non-local speakers. A similar motivation may underlie Outsiders’ responses, in at least those trials in which they judged the non-local speakers tokens similar to their own speech. Travel to other dialect areas was included in the analysis in order to account for effects of short-term exposure to other regional accents. It had no effect for Insiders, but a negative effect on choosing the local speaker for Outsiders. The lack of effect for Insiders was surprising, given that they would benefit most from the contrast, and other sub-groups of subjects examined in Beck (2014) show positive effects of traveling to different regions on choosing the local accent in this task. For Outsiders, the negative effect of travel on choosing the local speakers may relate to the information they get about the distribution of regional accents through travel. On the surface, Outsiders’ input does not suggest that the distribution of regional accents relates to geography, given that they hear both the parents’ and the community’s regional accent in the same location. Most Outsiders were reported to travel to visit family members in the region where the parents were born. This means that through travel, many Outsiders get experience hearing the parents’ regional accent in two distinct locations, further obfuscating the link between accent and location. Only once they understand that the parents have moved to the child’s hometown from elsewhere would the distribution of accents become clear. Outsider children in this study were not tested on whether they knew their family was from another region, but further work will confirm whether this is a viable explanation for the negative effect of travel found here. Finally, the most unexpected result was the negative effect of commenting on choosing the local speakers in this task for both Insiders and Outsiders. Parental reports of children spontaneous commenting on accents was included as a measure of awareness, as it seemed likely that children would only comment on accents if they were salient to them. For Outsiders, commenting could possibly be a reflection of how the family perceives its regional accent with respect to the community’s. Families who comment more frequently on differences between the home and community accent may lead their children to believe that they too speak a non-majority variety. This possibility may be reflected in the Outsider’s responses on those trials in which they judged the non-local speaker tokens as similar to their own speech. They may have been marking the perceived difference between their own speech (or their home dialect) and the local speech by choosing the non-local speaker on those trials. This



Chapter 2.  Children’s accent discrimination 59

only occurred in a few trials for each Outsider, as they chose the local speakers a majority of the time. However, it shows that they were using different heuristics in judging similarity in each trial, and suggests they may have been trying to integrate social and acoustic information in making those judgments. An explanation for the negative effect of commentary for Insiders is less apparent. They should have no reason to believe that they speak differently from other local speakers, and awareness of accents should logically increase the likelihood that they choose the local speaker as most similar. Further data are needed to interpret the result for this group. It could be the kinds of commentary made by Insiders and Outsiders are fundamentally different, and represent different kinds of awareness. It also may be that commentary doesn’t represent children’s awareness, but rather the frequency with which parents comment on accents. Parental language practices have been shown to influence children’s own language use (e.g. Foulkes, Docherty, and Watt 2005; Smith, Durham, and Fortune 2007; Smith, Durham, and Richards 2013). It may also be the case that their meta-linguistic commentary and imitation of accents is likewise mirrored in their children’s practices. In some studies, five-year-olds are heard to express social evaluation of regional accents without being able to correctly identify them (e.g. Cremona and Bates 1977). More detailed observation of individual children and their families’ practices of imitating and commenting on accents is needed to better understand how these two factors are reflections of awareness for both Insiders and Outsiders, in particular how their beliefs about their own speech affect their perception of their own regional accent. 5. Conclusion The results show that all children in this study, regardless of their dialect background, perform at rates better than chance in discriminating between regional accents. Although this task was framed subjectively, asking children to pick the speaker sounding most similar to themselves, most children picked the local speaker as sounding most similar in a majority of trials. This suggests that most children at this age are aware of their own regional variety, or at least choose to associate their speech with a familiar variety from their environment. Previous research conducted with five- and six-year-olds indicated that they had difficulty with regional accent discrimination in categorization tasks (Girard, Floccia, and Goslin 2008; Floccia et al. 2009; Wagner, Clopper, and Pate 2014). The major difference between the previous and current works is that this study uses word-length stimuli rather than sentence-length utterances. Reducing the amount of information children need to hold in their short-term memory may allow them to focus on the phonetic properties of the accent across the tokens. It cannot be

60 Erica Beck

ruled out though that the difference lies in the context in which the study was conducted. This will be examined in future studies by conducting these tasks in different subject contexts. Contrary to expectation, Outsiders and Insiders did not show any significant differences in their abilities to discriminate between accents. Unlike the 7-year-olds in the Floccia et al. (2009), having exposure to multiple regional varieties at home and in the community does not seem to improve discrimination for five-year-olds. Both groups of participants in this study chose the local speakers with equal frequency in the Discrimination Task. Floccia et al. (2009) hypothesized that the difference between the bi-dialectal and mono-dialectal 7-year-olds in their study was that bi-dialectal children had an increased awareness of regional phonological variation, making them better able to discriminate between regional accents. They did not assess awareness in that study, but their finding motivated assessing awareness and its effects on discrimination in the current study. Again, contrary to expectation, no significant difference in awareness was found between Outsiders and Insiders for the measures assessed in the current study. However, the participants in Floccia et al. (2009) were, on average, a year older than the participants in this study. It may be that at age five and six, the benefit of receiving exposure to another dialect has not accrued to the point of positively affecting awareness or ability to identify accents. Supporting this interpretation is the study by Kinzler and DeJesus (2013), who examined accent identification by five-year-old Northern and Southern U.S. children. They found that the Northerners, who have regular exposure to only their native regional variety, outperform the Southerners, who have regular exposure to both Northern and Southern dialects. The important similarity between Kinzler and DeJesus’ finding and the current work is that long-term, robust, dialect exposure does not appear to confer an advantage in awareness or perception of regional variation before the age of seven. This finding corresponds to the stage in development of sociolinguistic categories which Munson (2010) refers to as the “library of speakers.” At this stage, children have regional accents associated with specific individuals and have not abstracted their experience to more general representations of regional variation. At this stage, the benefit of exposure to regional variation will not be evident in measures of awareness, as their knowledge of regional accents is still associated with specific individuals. In the present work, this is evidenced by the fact that most children correctly identify the local accent, with which they have the broadest range of experience. However, they have much less success in identifying the non-local accent, or identifying regional variation as the difference between speakers in the Awareness Task. For both Insiders and Outsiders, their exposure to other regional



Chapter 2.  Children’s accent discrimination 61

accents may be limited to a few individuals, and as a result, their representations of other regional accents are only associated with those individuals. More work is needed with Outsiders to understand how awareness of regional variation develops for this group. What is clear is that hearing multiple regional varieties at home doesn’t make five- and six-year-old children more aware or better able to identify accents than children without that experience. Instead, it appears that exposure to multiple dialects and accents at home changes how they interpret regional variation at this age. Future studies will look at larger numbers of Outsiders in different contexts, in order to understand how they assess differences in regional accents, and how their interpretation of regional variation develops between five and seven years of age.

References Bates, Douglas, Martin Maechler, and Ben Bolker. 2011. “lme4: Linear Mixed-Effects Models Using S4 Classes: R Package Version 0.999375-42.” Accessed October 8, 2016. https:// CRAN.R-project.org/package=lme4. Beck, Erica Lynn. 2014. The Role of Socio-Indexical Information in Regional Accent Perception by Five to Seven Year Old Children. Ph.D. Dissertation, University of Michigan. Butler, Joseph, Caroline Floccia, Jeremy Goslin, and Robin Panneton. 2011. “Infants’ Discri­ mination of Familiar and Unfamiliar Accents in Speech.” Infancy 16 (4): 392–417.  doi: 10.1111/j.1532-7078.2010.00050.x Clopper, Cynthia G., and David B. Pisoni. 2004. “Homebodies and Army Brats: Some Effects of Early Linguistic Experience and Residential History on Dialect Categorization.” Language Variation and Change 16 (1): 31–48.  doi: 10.1017/S0954394504161036 Clopper, Cynthia G., and David B. Pisoni. 2004. “Some Acoustic Cues for the Perceptual Cate­ gorization of American English Regional Dialects.” Journal of Phonetics 32 (1): 111–140.  doi: 10.1016/S0095-4470(03)00009-3 Clopper, Cynthia G., Kristin L. Rohrbeck, and Laura Wagner. 2012. “Perception of Dialect Variation by Young Adults with High-Functioning Autism.” Journal of Autism and Developmental Disorders 42 (5): 740–754.  doi: 10.1007/s10803-011-1305-y Cremona, Christiana, and Elizabeth Bates. 1977. “The Development of Attitudes towards Dialect in Italian Children.” Journal of Psycholinguistic Research 6 (3): 223–232. Cristia, Alejandrina, Amanda Seidl, Charlotte Vaughn, Rachel Schmale, Ann Bradlow, and Carolina Floccia. 2012. “Linguistic Processing of Accented Speech across the Lifespan.” Frontiers in Psychology 3: 479.  doi: 10.3389/fpsyg.2012.00479 Egerova, Natalia. 2010. Neural Correlates of Diaect Perception in 5-Month-Old Infants: A Near Infrared Spectroscopy Study. Master’s Degree Dissertation, University of Groningen. Floccia, Caroline, Joseph Butler, Frédérique Girard et Jeremy Goslin. 2009. “Categorization of Regional and Foreign Accent in 5- to 7-Year-Old British Children.” International Journal of Behavioral Development 33 (4): 366–375.  doi: 10.1177/0165025409103871 Foulkes, Paul, and Gerard J. Docherty. 2006. “The Social Life of Phonetics and Phonology.” Journal of Phonetics 34 (4): 409–438.  doi: 10.1016/j.wocn.2005.08.002

62

Erica Beck

Foulkes, Paul, and Gerard J. Docherty, and Dominic Watt. 2005. “Phonological Variation in Child Directed Speech.” Language 81 (1): 177–206.  doi: 10.1353/lan.2005.0018 Girard, Frédérique, Caroline Floccia, and Jeremy Goslin. 2008. “Perception and Awareness of Accents in Young Children.” British Journal of Developmental Psychology 26 (3): 409–433.  doi: 10.1348/026151007X251712 Kerswill, Paul, and Ann Williams. 2000. “Creating a New Town Koine: Children and Language Change in Milton Keynes.” Language in Society 29 (1): 65–115.  doi: 10.1017/S0047404500001020 Kinzler, Katherine D., and Jasmine M. DeJesus. 2013. “Northern = Smart and Southern = Nice: The Development of Accent Attitudes in the United States.” The Quarterly Journal of Experimental Psychology 66 (6): 1146–1158.  doi: 10.1080/17470218.2012.731695 Kitamura, Christine, Robin Panneton, Anna Notley, and Catherine Best. 2006. “Aussie-AussieAussie, Oi-Oi-Oi: Infants Love an Australian Accent.” The Journal of the Acoustical Society of America 120 (5): 3135.  doi: 10.1121/1.4787737 Labov, William. 1998. “The Three English Dialects.” In Handbok of dialects and language variation, ed. by Michael D. Linn, 39–81. San Diego: Academic Press. Labov, William, Sharon Ash, and Charles Boberg. 2006. Atlas of North American English. Berlin: Mouton de Gruyter. Millar, Sharon. 2003. “Children and Linguistic Normativity.” In Social Dialectology: In Honor of Peter Trudgill, ed. by David Britan, and Jenny Cheshire, 287–297. Philadelphia: John Benjamins.  doi: 10.1075/impact.16.20mil Munson, Benjamin. 2010. “Levels of Phonological Abstraction and Knowledge of Socially Motivated Speech-Sound Variation: A Review, a Proposal and a Commentary on the Papers by Clopper, Pierrehumbert, and Tamati; Drager, Foulkes; Mack; and Smith, Hall and Munson.” Laboratory Phonology 1 (1): 157–177.  doi: 10.1515/labphon.2010.008 Nazzi, Thierry, Peter W. Jusczyk, and Elizabeth K. Johnson. 2000. “Language Discrimination by English-Learning 5-Month-Olds: Effects of Rhythm and Familiarity.” Journal of Memory and Language 43 (1): 1–19.  doi: 10.1006/jmla.2000.2698 Payne, Arvilla. 1976. The Aquisition of the Phonological System of a Second Dialect. Ph.D. Dissertation, University of Pennsylvania. Phan, Jennifer, and Derek M. Houston. 2008. “Infant Dialect Discrimination.” Research on Spoken Language Processing 29: 316–329. R Core Development Team. 2011. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. Roberts, Julie. 1997a. “Acquisition of Variable Rules: A study of (-t, d) Deletion in Preschool Children.” Journal of Child Language 24 (2): 351–372.  doi: 10.1017/S0305000997003073 Roberts, Julie. 1997b. “Hitting a Moving Target: Acquisition of Sound Change in Progress by Philadelphia children.” Language Variation and Change 9 (2): 249–266.  doi: 10.1017/S0954394500001897 Roberts, Julie. 2005. “Acquisition of Sociolinguistic Variation.” Clinical Sociolinguistics, ed. by Martin J. Ball, 151–164. Malden: Blackwell.  doi: 10.1002/9780470754856.ch12 Roberts, Julie, and William Labov. 1995. “Learning to Talk Philadelphian: Acquisition of Short A by Preschool Children.” Language Variation and Change 7 (1): 101–112.  doi: 10.1017/S0954394500000910 Schneider, Edgar W. 2008. Varieties of English 2: The Americas and the Carribean. Berlin: Mouton de Gruyter.



Chapter 2.  Children’s accent discrimination 63

Smith, Jennifer, Mercedes Durham, and Hazel Richards. 2013. “The Social and Linguistic in the Acquisition of Sociolinguistic Norms: Caregivers, Children and Variation.” Linguistics 51 (2): 285–324.  doi: 10.1515/ling-2013-0012 Smith, Jennifer, Mercedes Durham, and Liane Fortune. 2007. ““Mam, My Trousers is Fa’in Doon!”: Community, Caregiver, and Child in the Acquisition of Variation in a Scottish Dialect.” Language Variation and Change 19 (1): 63–99.  doi: 10.1017/S0954394507070044 Sumner, Meghan, and Arthur G. Samuel. 2009. “The Effect of Experience on the Perception and Representation of Dialect Variants.” Journal of Memory and Language 60 (4): 487–501.  doi: 10.1016/j.jml.2009.01.001 Wagner, Laura, Cynthia G. Clopper, and John K. Pate. 2014. “Children’s Perception of Dialect Variation.” Journal of Child Language 41 (5): 1–23.  doi: 10.1017/S0305000913000330 Wells, John C. 1982. Accents of English. Cambridge: Cambridge University Press.  doi: 10.1017/CBO9780511611759

Chapter 3

How do social networks influence children’s stylistic practices? Social mixing, macro/micro analysis and methodological questions Laurence Buson

Université Grenoble Alpes

Studying children’s ability to construct themselves as pluristyle speakers goes hand-in-hand with studying their socialisation. Children’s conversational experiences are one of the key factors that allow them to integrate and participate in social life. This article will therefore present studies conducted in a school setting with the aim of better understanding the links between children’s stylistic usage and the social diversity of their social networks. The network approach has been widely used in adults but rarely with children and it allows the acquisition of variation to be considered in context, taking into account children’s horizontal socialisation beyond parental influences. The present analysis will begin by looking briefly at existing work drawing on the notion of social networks in adults and teenagers in order to outline a number of related key research questions in terms of first-language acquisition. Some results obtained from a study conducted in Grenoble (France) with 10–11-year-old children will then be described (Buson 2009a, b). Finally, micro-­ sociolinguistic analysis will be used to return to some of the methodological issues that remain to be resolved in order to reach a more detailed understanding of the links between the acquisition of variation and the influence of peer networks. Keywords: style, socialisation, social network, variation, methodology, acquisition of discourse markers

1. Social networks and linguistic practices in adults and adolescents: State of the art in sociolinguistics The concept of social network is borrowed from sociology (see, for example, Granovetter’s work, 1983) and considers social actors as interdependent rather than doi 10.1075/silv.20.03bus © 2017 John Benjamins Publishing Company

66 Laurence Buson

as isolated entities. This notion has been taken up in sociolinguistics by a number of authors. One of the first studies in linguistics to take networks into account was Labov’s study of Martha’s Vineyard (Labov 1963). In his more recent work, Labov has continued to insist upon the importance of networks: Social network factors do not replace the effect of age, social class, neighbourhood, or ethnicity. The social network effects are not the largest, but they add essential information to the description of the leaders of linguistic change. (Labov 2001: 344)

According to Gumperz, relationship networks between individuals can be open if several varieties coexist due to the presence of speakers from diverse backgrounds, or closed if only one variety or one language is allowed due to the presence of speakers with a strong level of complicity and a common culture. Situations of open networks are therefore “marked both by diversity of norms and attitudes, and by diversity of communicative conventions” (Gumperz 1982). Milroy and Milroy (2000 [1997]) gave greater precision to this category by introducing the notions of density and multiplexity and also referring to close-knit and loose-knit networks. A “dense” network is one in which a large number of people to whom an individual is linked also have relationships to each other, while the notion of “multiplex” relates to the nature of the link between individuals and refers to the fact that one individual is linked to another by different types of links, for example as both friend and colleague. Dense, multiplex networks correspond more often to groups from disadvantaged backgrounds, and strong close links within the local community encourage the maintaining of non standard forms. Close-knit networks are thus perceived as a locus for adhesion to the vernacular and for reinforcing local, and mainly non legitimate, norms. They slow down or inhibit language change and play a role in maintaining minority languages (Blom and Gumperz 1972; Gal 1979). Conversely, loose-knit or open networks allow individuals to be more open to change and less sensitive to the normative pressure of the group to which they belong, as indicated by Meyerhoff: Loose networks make individuals more open to change. The ties that the individual members have to other networks provide an opportunity for them to be exposed to and pick up innovations from outside their network. In addition, the normative pressure of any single network will be attenuated, because we assume speakers have to attend to the norms of the numerous different networks that their loose social ties give them membership to. (Meyerhoff 2006: 187)

Marshall (2004) updated Milroy’s (1987 [1980]) model: rather than considering social networks as systematic norm enhancers and individuals as belonging/not belonging to such networks, and having certain linguistic behaviour as a result, he



Chapter 3.  How do social networks influence children’s stylistic practices? 67

integrated the notion of personal orientation towards the network. In his framework, speakers harbouring a positive mental orientation towards a local network are free to choose whether or not they signal their integration in the network through their linguistic usage. Introducing this idea of choice linked to speakers’ individual characteristics allowed Marshall to explain the fact that the correlation between social networks and maintaining non standard forms is not systematic, as shown by his own study in a rural setting. This correlation also depends upon the desire to adhere to the group and to its local norms, and to the degree of solidarity felt by an individual towards the group. This amendment to the concept places it somewhere between the notion of social network and the idea of community of practice inherited from Eckert (2000). For Eckert, a community of practice supposes not only a certain frequency of interaction but also engaging in common activities and orientations. The social-network framework thus offers a level of social categorisation that can allow us to go beyond the usual macro-social categories and focus upon groups with shared activities, aims and values, as well as upon the (frequent) interactions between their members in their day-to-day relationships. It is a valuable tool in explaining and understanding variation: the latter cannot simply be explained in terms of access to education and/or a certain level of income, it also needs to be considered in terms of social positioning i.e. looking at relationships with others (domination, power, submission, complicity, loyalty, etc.) in specific contexts; in other words, focusing on the symbolic dimension of interactions. All these aspects take on meaning within groups of speakers, above and beyond their original social background, when we try and reach an understanding of how individuals interact on a local level. 2. Social networks and linguistic practices in children: The importance of horizontal socialisation Labov (1964) posits that social network has a role to play in the language of children as early as 5 years old and Ochs and Schieffelin (1995) have shown that integration in the social network in general – and the peer group in particular – is a determining factor in language acquisition. The influence of peer relationships on language begins from a very young age. Indeed, to quote Ensor and Hughes as cited by Hay, Caplan, and Nash: “Toddlers’ verbal ability is correlated with prosocial behavior with peers (Ensor and Hughes 2005)” (Hay, Caplan, and Nash 2009: 133). Some recent research even suggests that interactions with peers could be an even more important factor than mother-child interactions from early infanthood:

68 Laurence Buson

Recent studies of infants’ capacities for intersubjectivity concentrate attention on infant – adult interaction /…/, despite the fact that researchers have suggested that peer interaction may be a particularly suitable context for the development of intersubjectivity and shared meaning /…/. Furthermore, some writers have suggested that infants’ own capacities for social engagement and intersubjectivity may be overestimated in interactions with mothers, and are best demonstrated in interactions with peers. (Hay, Caplan, and Nash 2009: 136)

However, to our knowledge, few sociolinguistic studies have highlighted the possible link between linguistic usage and type of social network in children. Ito’s 1989 study showed similar patterns in usage of non standard forms, on a phonetic, lexical and syntactic level, within friendship networks in class. However, this study involved 13–14 year-old subjects, who were therefore already adolescents. Nardy (2008), taking up Ito’s approach, studied 4–5 year-old children in the same preschool class and was able to observe closer language usage between children who had frequent verbal interactions than between those who had no contact. Martin, Chevrot, and Barbu (2010) have also highlighted the pragmatic and identity-­related motivations behind stylistic adjustments in an extensive case study involving a 10-year-old child. According to the authors, the child modulates his production of a non standard regional variable (the Franco-Provençal [Y]) according to the supposed value of the variable for his interlocutor, and also depending on how long-standing his friendships are with interlocutors within the peer network: The use of the variants depends on the position of the addressee within the family (parents or siblings) or within the network of friends (recent vs. long-standing, native vs. non-native). (Martin, Chevrot, and Barbu 2010: 689)

However, research in which this question has been more or less central has sometimes yielded contradictory results. Marshall (2004: 201), in a study in a rural setting, observed that the 8–12 year-old pre-adolescents in his sample were less peer-­orientated than their elders and did not behave in the same way where usage of certain non standard variants was concerned. They appeared to be less sensitive to the pressure of their peers and to the city values – and associations with “fun” and youth culture – of certain non standard variants. Marshall considers that the effects of social network should be qualified by also taking into account speakers’ representations and conceptions, for example their degree of solidarity towards the group or the degree to which they have integrated urban values. However, he notes that the effects are less obvious in younger speakers, for whom the results were less consistent. The difficulty in analysing children’s social networks and their impact on language usage may stem from a lack of appropriate tools for working with young children. The relevance of the criteria of density and multiplexity when qualifying



Chapter 3.  How do social networks influence children’s stylistic practices? 69

children’s social networks can be questioned. It is possible that these criteria are associated with the values and models of socialisation in the adult world. This would be congruent with a somewhat “adult-centred” tendency in studies concerning children’s socialisation (see Kyratzis 2004). Definitions of “peer-group” can be more or less broad. Should the peer-group be considered as including all the children in a school, all the children in the same class or simply the small group of best friends? Relevant indicators are lacking to define these types of network in relation to the analysis of linguistic practices. As shown by Nardy’s (2008) study, the question of which interactional characteristics to take into account also arises: contact (verbal or non verbal), frequency of language interaction, the nature of these interactions or the hierarchical organisation and roles within these interactions, as influenced by factors such as the popularity of individuals? Moreover, according to Howes (2009), very little information is available regarding the developmental consequences – and thus the linguistic consequences – for children of frequenting peers who are unlike them (a much rarer tendency than frequenting peers who are like them). Our study therefore takes this criterion into account, looking at the degree of social diversity/social mixing of the network. The study is centred upon the classroom and the school context more generally. This is because this seems to be the principal locus for socialisation in children between 5 and 11 that can allow a certain degree of social diversity within peer contact (or not, depending on the school). How can social diversity/mixing be defined? Broadly speaking, it is an intermixing leading to interactions between socially diverse populations. The notion of interaction is central here, implying a dimension of meeting and of exchange. In France, studies in sociology such as those by Varro (2003) have shown that this social mixing, supposed to occur naturally, is declining in favour of a rapidly increasing degree of segregation in schools. Furthermore, the apparent social mixing in some of these establishments can actually mask a different sort of segregation, internal this time, between “good” and “bad” school classes as asserted by Payet (1999, 2000). Recent research focusing on the theme of ethnic segregation in groups of children and adolescents in North American schools has paved the way for social-­ psychological thought about the behavioural and developmental consequences of diversity in the school context. Killen, Rutland, and Jampol (2009), for example, give an account of perceptual studies carried out with first and fourth graders in primary schools of varying degrees of ethnic heterogeneity. 1 The authors show that 1. On the subject of the influence of multi-ethnic friendship networks on linguistic productions, see also Fox, Khan, and Torgersen (2011).

70 Laurence Buson

the diversity of the school context influences children’s representations: perception of friendship potential is modified according to the network’s social diversity (children in homogeneous schools were less inclined to imagine friendships with peers of different ethnic origin). 2 We can therefore question whether social diversity does not also influence children’s representations, and perhaps their language practice. 3. A study of the effect of social mixing on children’s stylistic practices in the French context: Research hypotheses and data collection methodology Based on a preliminary study (Buson 2008) in two “CM” classes (i.e. the last two years of primary school) in Grenoble and the surrounding area, two main working hypotheses were formulated: First, that children of around 10 years of age are pluristylistic and this includes children from disadvantaged backgrounds (this may seem self-evident, however it has sometimes been called into question, for example in Bernstein’s notion of “restricted code” (Bernstein 1975 [1971]), or more generally within public opinion). Second, that stylistic repertoires are larger within heterogeneous networks, due to greater familiarity with diverse linguistic usage. Given these hypotheses, the question arises at to what criteria should be adopted to define children’s social networks. In this study, in order to take into account two essential elements of horizontal socialisation between children of primary school age, two aspects were chosen for analysis: degree of social mixing in the school and in the friendship groups. These two focal elements allow the children’s network to be approached both from a relatively broad perspective (the social profile of the school space) and from a closer vantage point (the close friendships between children).

The sample Two schools were chosen due to their profile. Two criteria were retained for this selection: the distribution of social background in the school classes studied (were varied social backgrounds present in the school?) and the social mixity of the preferential friendship dyads (were children from different social backgrounds friends with one another?). As the children were divided into four social backgrounds, from the most disadvantaged to the most advantaged, the strongly mixed dyads 2. For a more detailed discussion of these notions of mixity, see Buson (2009b).



Chapter 3.  How do social networks influence children’s stylistic practices? 71

were those that combined children from distanced backgrounds, for example an advantaged background linked with a working-class background. These friendship links were derived from a standardised interview with each child, during which she was asked to state who her best friend was. The social classes were divided as follows: lower class (LC), lower-middle class (LMC), upper-middle class (UMC), and upper class (UC). They were established on the basis of the occupation of both parents of the children, as recorded on the information forms they provided to the schools. The first school (Portail Rouge) has a relatively balanced distribution of social backgrounds and the second school (Vercors) has a very low level of social mixing, as can be seen in Table 1: Table 1.  Social mixing in Portail Rouge and Vercors schools Portail Rouge school Distribution of the children across the four social backgrounds

LC LMC UMC UC 45.9% 16.2% 16.2% 21.6% (17) (6) (6) (8)

Vercors school LC LMC UMC 88.2% 5.8% 0% (15) (1) (0)

UC 5.8% (1)

In Portail Rouge school, a broad range of social backgrounds are represented and there is a fairly equal balance between these different backgrounds. Among the schools initially pre-selected, this school also had the highest percentage of mixed friendship dyads: Table 2.  Social mixity of the friendship dyads in the pre-selected schools Portail Rouge Strong Weak or null

  35% (13)   65% (24) 100% (37)

du Bois

Mille Chemins

Village

22% (9)   20.5% (8) 19% (9)   78% (32)   79.5% (31)   81% (38) 100% (41) 100% (39) 100% (47)

Renaudie & Vercors   9% (2)   91% (20) 100% (22)

 41 145 186

Conversely, the Vercors school is a disadvantaged school with very little social mixity, as can be seen in the distribution of social backgrounds (more than 88% of the children in the CM1/CM2 classes are from a disadvantaged background). Once the target schools had been selected, subjects were then selected within these schools. In order to limit the independent variables, a sample of only girls was chosen, as they were best represented in both schools. Tables 3 and 4 present some characteristics concerning each of the subjects chosen for this study.

72

Laurence Buson

Table 3.  The four girls from Portail Rouge school  Portail Rouge mixed school Marianne LauraG Morgane Lilia

Age   10.5 years   10.5 years 11 years 11 years

Academic Position among level 3 siblings A B B D

1/3 1/2 2/2 3/5

Father’s occupation

Mother’s occupation

House painter Cleaning operative Factory worker Unemployed Factory worker Child caregiver Waste collection Unemployed employee

Table 4.  The four girls from Vercors school Vercors nonmixed school

Age

Academic Position among level siblings

Dahlia Mélissa Abir

11 years   10.5 years 10 years

C B A

1/1 1/3 3/4

LauraS

  10.5 years

C

3/3

Father’s occupation

Mother’s occupation

Deceased Bus driver Advertising delivery man Absent

Cleaning operative Cleaning operative Factory worker Unemployed

The eight girls selected were therefore all of approximately the same age and social background, however they were part of noticeably different networks, the characteristics of which shall now be briefly described. Information was drawn from the preliminary individual interviews, during which the children had to describe their friendship ties with the other children in their class, in their school and in their neighbourhood. My presence in the school for several months also allowed me to reach an understanding of how the children worked, the types of relationships they had, their personalities, etc. This ethnographic approach was also supplemented by more quantitative analysis based on the children’s statements, such as a popularity index calculated as followed: the number of times the child was cited as a friend relative to the average for the children in the class, so as to neutralise any effects linked to the number of pupils. Rates above 1 therefore indicate children who were more popular than the average for children in their class.

The four girls from the mixed school The network of the four girls from Portail Rouge was dense and they had many and varied friendship ties. Lilia was a fairly popular child (1.3) who was often mentioned as being a friend, but never mentioned by the children she mentioned herself, which seemed to 3. Academic level is given on a scale from A to D, with A corresponding to a very good level. These assessments were provided by the teachers and are therefore only indicative.



Chapter 3.  How do social networks influence children’s stylistic practices? 73

indicate that she was very sociable and did not necessarily have any strong friendship ties with a particular child. She stated that she liked 3 children in her class, including LauraG, and that there were 2 children in her class that she did not like very much. She stated that she had 5 friends outside school. LauraG had 3 reciprocal friendships, which corresponded approximately to the average in her class. She stated that she did not like Lilia, who had mentioned her as a friend, which may indicate that there was sometimes a certain rivalry between them. She had 4 friends outside school, stated that she had 7 friends in school, and was mentioned by 6 other children, which made her a fairly popular (1.1) and social child. Marianne had 3 reciprocal friendships, with children from varied social backgrounds. She mentioned 4 friends outside school, 6 in school and was mentioned 8 times by other children (popularity index of 1.2). Her network profile was fairly similar to LauraG’s. Morgane had 4 reciprocal friendships, which was above the average for her class. These 4 girls therefore constituted a close-knit small group of friends. The other children she likes outside her class were in the CP class (first year in primary school), and she only mentioned one friend outside school. She was mentioned 5 times as a friend and had a popularity index below the average of the class (0.9). Morgane therefore had few interactions with children outside her small group, which seemed very close-knit. Conversely, Marianne, LauraG and Lilia were fairly popular children, who knew and were known by many others, even though Lilia seemed to be the biggest “social butterfly” of the three. In terms of the languages heard and/or spoken at home by the children at Portail Rouge school, Marianne and Morgane stated that they were monolingual in French. LauraG spoke and understood French and Italian (Italian was more the “language of complicity” between her parents, but she stated that she understood it and spoke it from time to time). Lilia said that she heard her mother speak in Armenian from time to time, but stated that she only understood a few words.

The four girls from the non-mixed school Mélissa seemed to be the most sociable and most popular of the 4 (and she was indeed the most popular in the class, with an index of 2.2). She had 2 reciprocal friendships, was mentioned 5 times and stated that she had 4 friends outside school. She seemed to have a certain enmity towards Laura and Dahlia’s group, whom she said she did not like. She did know, however, that Abir was a good friend. Abir was in a reciprocal friendship with Mélissa, whom she liked and knew she was liked by. She was only mentioned as a friend twice, which meant that she was not a very popular child (0.8), but she did not have any enmity towards anyone. She stated that she had 3 friends outside school.

74

Laurence Buson

Dahlia rejected adult authority and often entered into conflicts with everyone. She only gave one name as a friend in the class (LauraS, who also chose her) and was only mentioned by one other child aside from LauraS (popularity index of 0.8). She nonetheless stated that she quite liked 10 children in the school (children from the younger classes) and 3 children outside school. Finally, her partner, LauraS, designated 3 children in the school as being her friends, and was liked by 3 children, which made her a relatively popular child in this class (1.3). However, she stated that she did not have any friends outside school. The 4 girls from Vercors school were therefore generally divided into two subgroups: Mélissa-Abir and LauraS-Dahlia. The friendships in this class seemed to be quite watertight with few links between the mini-networks. In terms of the languages heard and/or spoken at home for the children of Vercors school, Dahlia and LauraS were, according to them, monolingual in French. Mélissa stated that she had Algerian origins and that she heard Arabic spoken at home. Abir said that she spoke Arabic and that she also knew how to write it a little as she had taken classes at school and had also learned a bit with her father.

The stylistic repertoires Having discussed how to define the children’s social networks, it is now necessary to address the issue of how to approach their stylistic repertoires. No previous studies exist allowing an overall view or description of speakers’ stylistic repertoires, particularly where children are concerned. No exhaustive list of features has been established and research is generally based either on a limited number of well-known linguistic features enabling quantification (see, for example, Martin, Chevrot, and Barbu 2010) or on qualitative analysis of interactions, used as significant case studies. At the current stage of knowledge about variation, no ideal methodology exists, as will be discussed in greater depth at the end of this article. The approach adopted here attempts to combine qualitative and quantitative analysis and to reach beyond the use of traditional, often phonetic, variants to explore the stylistic value of pragmatic variants. 4. Analysis of stylistic variation through two discourse markers: hein and oh In this regard, discourse markers (DM) were chosen for a number of reasons. First, they have rarely been taken into account in previous studies, despite being considered by several authors to be good indicators of children’s pragmatic development and being considered as salient traits in spoken French or in other languages



Chapter 3.  How do social networks influence children’s stylistic practices? 75

(Andersen 2000; Fernandez 1994; Kyratzis and Ervin-Tripp 1999; Andersen et al. 1999; Tagliamonte 2005). They are also markers of the social relationship between interlocutors. Andersen et al. (1999) observed, for example, that they are an important part of pragmatic competence in that their use reflects: the relative status of speaker and addressee, as well as their level of familiarity/ intimacy and the topic and setting of their discourse. (Andersen et al. 1999: 1340)

Despite the obvious interest of the DM in language acquisition, their usage in child everyday speech is still underexplored (several studies, however, use methods based on role-plays, e.g. Gillen and Hall 2001; Kyratzis and Ervin-Tripp 1999). For the purpose of this study, two specific DM “hein” and “oh” were selected. Given that no studies exist in French allowing strictly formal or informal markers to be listed or distinguished, this selection was based on a preliminary study. This study involved 6 children of 9/10 years (3 boys and 3 girls), from middle-class backgrounds, and a corpus of approximately 6,000 words, of which 3,000 in conversations between peers, and 3,000 in a formal context (oral speech, partly prepared, in front of an audience). The distribution of 5 DM – bah, ben, hein, oh and eh – was then analysed and the choice was made to retain the only two DM that were more systematically present in informal contexts: hein and oh (see Buson 2009b for the details of this prior study). It should be noted that the main study described hereafter supports the results of this initial preliminary survey: only occurrences of hein and oh were systematically more numerous in informal situations for the 8 children. For the purpose of the present study, the data were collected in an ecological setting by recording ordinary conversations in the school (during classes and break-time). Wearable recording systems have been worn by 8 girls (see Tables 3 & 4), during approximatively 10 hours each. The data were transcribed using orthographic transcription, for all utterances. 48,917 words were recorded during peer conversations and 10,244 words in adult-addressed utterances (see Table 5). A total of 977 “hein” and “oh” DM has been collected for the 8 girls (102 in formal situations and 875 in informal situations). The number of tokens for each subject ranges from 1 to 41 for the formal context, and from 20 to 261 for the informal context. 4

4. The number of tokens is quite low compared with studies dealing with other types of variables (e.g. phonological variables). However, even in the case of phonological variables, the number of tokens in the study of sociolinguistic variation in child may be low, due to the low density of interactions during play and school situations (for example, see Martin, Chevrot, and Barbu 2010: Table 1).

76

Laurence Buson

Table 5.  Number of DM tokens for each subject according to the situation Formal situations N hein+oh

N words

Portail Rouge (mixed school) LauraG   5  1234 Lilia   5  1038 Marianne   1   179 Morgane   8   507 Vercors (non-mixed school) Abir  41  2208 LauraS  19  1688 Mélissa  18  2577 Dahlia   5   813 102 10244

Informal situations

N hein&oh per 1000 words

N hein+oh

N words

N hein&oh per 1000 words

  4.1   4.8   5.6  15.8

126 135  85 101

 8537 10355  3919  4068

 14.8 13  21.7  24.8

 18.6  11.3 7  6.2

261  44  20 103 875

10593  3099  2214  6132 48917

 24.6  14.2 9 16.8

Concerning the interlocutors’ profiles, Figure 1 represents the number of tokens per 1000 words, the white bars corresponding to adult-addressed speech (in most cases, the teacher) and the grey bars to peer-addressed speech. The graph therefore emphasizes the relationship between the two types of situation for each child. The children from the socially mixed school are on the left of the vertical line in the middle of the graph. We can see that all the children mark the difference between adult- and peer- addressed speech, even if the proportions are different and even if some differences are not statistically significant. Figure 2 shows the values for the numbers of occurrences of hein and of oh produced by all children in each school. 5 By presenting this data grouping together the girls from each school, we can see that while peer-addressed speech is fairly similar, adult-addressed speech is more formal in the socially mixed school (less use of hein and oh). At Portail Rouge school, the frequency of DMs per 1000 words in adult-directed speech was significantly lower than the frequency of DMs in child-directed speech (p chi2 = 0.03), thus confirming the stylistic range of the children in this mixed school. In Vercors school, however, while there was a difference between the DMs in adult-directed speech and those in child-directed speech, it was not significant overall (p chi2 = 0.15).

5. The values in Figure 2 correspond to the average of the individual scores of number of tokens per 1000 words.

Chapter 3.  How do social networks influence children’s stylistic practices? 77



30

Formal Informal 24.8

25

24.6

21.7 20

15

18.6 16.8

15.8

14.8

14.2

13.0 11.3

10

9.0 7.0

5

0

4.1

LauraG

4.8

Lilia

5.6

Marianne Morgane

Abir

LauraS

Mėlissa

6.2

Dahlia

Figure 1.  Use of DM hein and oh in 8 girls according to the context of the exchange. Number of occurrences per 1,000 total words in a given context (formal for adult-directed speech and informal for peer-directed speech) 25

adult interlocutor child interlocutor 19.42

20 16.63 15 11.39 10 6.42 5

0

Portail Rouge

Vercors

Figure 2.  Distribution of DM according to interlocutor in the two schools

78

Laurence Buson

We can therefore see that children within socially heterogeneous networks show greater stylistic flexibility than those within socially homogeneous networks, as posited in the initial hypotheses, even if substantial inter-individual variations do exist as shown in Figure 1 above. These initial results about the possible effect of peer networks on children’s stylistic practices and the range of their repertoires are based on statistical analysis. A more qualitative analysis of the data can further complete these results, first by looking for meaning within interactions, and second, by asking a certain number of methodological questions. Global analysis can potentially flatten out some interpretations by neglecting micro-events. Micro-sociolinguistic analyses, on the other hand, can show that a number of phenomena must be taken into account so as to refine analyses of stylistic variation and perhaps also to change the explanatory variable for future quantitative studies. In particular, the focus here will be on how the ‘interlocutor’ variable is taken into account asking, first, why and in what ways this variable is complex, and second, how we should alter our ways of taking it into account so as to avoid over-simplification. 5. A few illustrations of the difficulty in analysing the corpora on a global level: Methodological issues linked to studying stylistic variation in children Closer analysis of what happens at the level of production is necessary. The intention is to qualify the results outlined previously and, above all, to highlight some methodological difficulties that remain unresolved as yet. At this stage, it should be specified that the discussion that follows does not, in my view, invalidate the previous results (which provide general trends without going into the details of specific situations and cases). Rather, this discussion aims to raise methodological questions with the potential to refine future research in this field where the choice of (dependent and independent) variants and variables remains largely under discussion. One of the first issues to be underscored regards the choice to separate adult-­ directed speech (considered to be a formal situation) and child-directed speech (considered to be an informal situation). The corpus in fact shows situations of spontaneous role-playing that call this dichotomy into question. We can find stylised parodies of formal styles within peer interactions, as in the following example: Example 3: Lilia: bonjou:r mes:dameS et messieurs nous allons vous présenter aujourd’hui Ophélie< – bonjour Ophélie< – oh les filles v:ous vous venez on joue à:: – à:: – style star académie /…/ allez venez – bonjour mesdameS et messieurs aujourd’hui nous allons vous présenter – mad(e)moiselle Ophélie



Chapter 3.  How do social networks influence children’s stylistic practices? 79

qui va nous interpréter je n(e) sais quoi – et: on va la laisser chanter /…/ ah ça va t’enregistrer::: – ça saoule venez on arrête de jouer ça saoule [Lilia: Hello: ladies and gentlemen today we are going to introduce Ophelie