Understanding Language Contact 9780367766603, 9780367766597, 9781003167952

Table of contents :
Cover
Half Title
Series
Title
Copyright
Contents
List of abbreviations
Preface
Acknowledgments
Part I From milliseconds to minutes: what bilinguals do when they speak or sign
1 Interactive alignment and implicit priming
1.1 Unconscious alignment in interaction
1.2 Implicit priming
1.3 Case study of implicit cross-language priming and contact-induced language change
1.4 How to study bilingual phenomena experimentally
Exercises
Further reading
References
2 Conceptual transfer
2.1 Linguistic and non-linguistic conceptualizations
2.2 What happens when we use two languages that differ in what they code linguistically?
2.3 Case study of conceptual transfer for the concept “to be”
2.4 Case study of conceptual transfer for spatial representations
Exercises
Further reading
References
3 Cognitive costs and cognitive load
3.1 Cognitive costs
3.2 Cognitive load
Exercises
Further reading
References
Part II From minutes to years: what bilinguals do when they communicate with others
4 Code-switching, repertoires, and translanguaging
4.1 Acquiring and utilizing bilingual and multidialectal codes
4.2 Integrating the language varieties of bilingual repertoires
4.3 Embracing the multidimensionality of bilinguals’ codes
4.4 How to study code-switching
Exercises
Further reading
References
5 Social networks and accommodation
5.1 Accommodation and change
5.2 Case study: convergence in language choice in Montreal
5.3 Audience design and style
5.4 Language mode
5.5 Language mode and bilingual style
5.5.1 Style shifting in monolingual mode
5.5.2 Bilingual mode and style shifting
5.6 Social networks
5.6.1 Bilinguals’ social networks
Exercises
Further reading
References
6 Acquisition and attrition
6.1 Bilingual acquisition
6.2 Second language acquisition
6.3 Attrition of the first language
6.4 How to study bilingual child acquisition and attrition of the first language
Exercises
Further reading
References
7 Language ideologies and dispositions
7.1 Language attitudes and language ideologies
7.2 Attitudes, ideologies, and identity
7.3 Standard language and related ideologies
Exercises
Further reading
References
Part III From years to centuries: how languages change through contact
8 Contact-induced changes in grammar and borrowing
8.1 Different types of language change
8.2 Types of contact-induced changes
8.3 The consequences of borrowing: loanwords
8.3.1 Types of loanwords
8.3.2 The linguistic adaptation of loanwords
8.3.3 The impact of loanwords on the structure of a recipient language
8.4 Convergence
8.4.1 Quantitative convergence: convergence as a matter of degree
8.4.2 Qualitative convergence: a case study of Spanish in Texas
8.4.3 The effect of convergence on linguistic structure
8.5 How to study contact-induced changes in grammar and borrowing
Exercises
Further reading
References
9 Linguistic areas
9.1 What are linguistic areas?
9.2 The Balkans as a linguistic area
9.3 An overview of linguistic areas across the world
9.4 How to study linguistic areas
9.5 Case study of “have”-perfect in the Balkans
Exercises
Further reading
References
10 Creoles, pidgins, and mixed languages
10.1 Creoles
10.1.1 When did Creoles arise?
10.1.2 How did Creoles get formed?
10.1.3 What are the structural characteristics of Creoles?
10.2 Pidgins
10.3 Mixed languages
10.3.1 When do mixed languages arise?
10.3.2 What are the structural characteristics of mixed languages?
10.3.3 How do mixed languages get formed?
10.4 Urban youth languages
Exercises
Further reading
References
11 Minority languages, heritage languages, and immigrant linguistic practices
11.1 Different types of bilingual communities
11.2 Language maintenance and language shift
11.3 The outcomes of language contact on minority and heritage languages
11.4 Case study of Frenchville, Pennsylvania—U.S.A.
11.5 Case study of Romani
Exercises
Further reading
References
Conclusion
Answers to exercises
Glossary
Index

Citation preview

Understanding Language Contact Understanding Language Contact ofers an accessible and empirically grounded introduction to contact linguistics. Rather than taking a traditional focus on the outcomes of language contact, this book takes the novel approach of considering these outcomes as an endpoint of bilingualism and multilingualism. Covering speech production and comprehension, language difusion across diferent interactional networks and timeframes, and the historical outcomes of contact-induced language change, this book: • Discusses both how these areas relate to one another and how they correspond to diferent theoretical felds and methodologies; • Draws together concepts and methodological/theoretical advances from the related felds of bilingualism and sociolinguistics to show how these can shed new light on the traditional feld of contact linguistics; • Presents up-to-date research in a digestible form; • Includes examples from a wide range of contact languages, including Creoles and pidgins; Indigenous, minority, and heritage languages; mixed languages; and immigrants’ linguistic practices, to illustrate ideas and concepts; • Features exercises to test students’ understanding as well as suggestions for further reading to expand knowledge in specifc areas. Written by three experienced teachers and researchers in this area, Understanding Language Contact is key reading for advanced undergraduate and postgraduate students approaching bilingualism and language contact for the frst time. Evangelia Adamou is Senior Researcher at the CNRS, France. Barbara E. Bullock is Professor of French Linguistics at Te University of Texas at Austin, USA. Almeida Jacqueline Toribio is Professor of Linguistics at Te University of Texas at Austin, USA.

Understanding Language series Series Editors: Patience Epps, The University of Texas at Austin, USA Oliver Bond, University of Surrey, UK Consultant Editors: Bernard Comrie, University of California, Santa Barbara, USA & Greville Corbett, University of Surrey, UK The Understanding Language series provides approachable, yet authoritative, introductions to major topics in linguistics. Ideal for students with little or no prior knowledge of linguistics, each book carefully explains the basics, emphasising understanding of the essential notions rather than arguing for a particular theoretical position. Other titles in the series: Understanding Child Language Acquisition Caroline Rowland Understanding Semantics, Second Edition Sebastian Löbner Study Skills for Linguistics Jeanette Sakel Understanding Language Change Kate Burridge and Alexander Bergs Understanding Phonology, Fourth Edition Carlos Gussenhoven and Haike Jacobs Understanding Linguistic Fieldwork Felicity Meakins, Jennifer Green and Myfany Turpin Understanding Syntax, Fifth Edition Maggie Tallerman Understanding Corpus Linguistics Danielle Barth and Stefan Schnell Understanding Discourse Analysis Bernadette Vine Understanding Language Contact Evangelia Adamou, Barbara E. Bullock, and Almeida Jacqueline Toribio For more information on any of these titles, or to order, go to www.routledge.com/ Understanding-Language/book-series/ULAN

Understanding

Language Contact Evangelia Adamou, Barbara E. Bullock, and Almeida Jacqueline Toribio

Designed cover image: © Getty Images | cokacoka First published 2024 by Routledge 4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Tird Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 Evangelia Adamou, Barbara E. Bullock and Almeida Jacqueline Toribio Te right of Evangelia Adamou, Barbara E. Bullock and Almeida Jacqueline Toribio to be identifed as authors of this work has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafer invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identifcation and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Adamou, Evangelia, author. | Bullock, Barbara E., author. | Toribio, Almeida Jacqueline, 1963 – author. Title: Understanding language contact / Evangelia Adamou, Barbara E. Bullock and Almeida Jacqueline Toribio. Description: Abingdon, Oxon ; New York, NY : Routledge, 2023. | Series: Understanding language | Includes bibliographical references and index. Identifers: LCCN 2022061780 | ISBN 9780367766603 (hardback) | ISBN 9780367766597 (paperback) | ISBN 9781003167952 (ebook) Subjects: LCSH: Languages in contact. | Bilingualism. Classifcation: LCC P130.5 .A33 2023 | DDC 306.44/6—dc23/eng/20230313 LC record available at https://lccn.loc.gov/2022061780 ISBN: 978-0-367-76660-3 (hbk) ISBN: 978-0-367-76659-7 (pbk) ISBN: 978-1-003-16795-2 (ebk) DOI: 10.4324/9781003167952 Typeset in Minion by Apex CoVantage, LLC

Contents List of abbreviations Preface Acknowledgments

ix x xv

PART I

From milliseconds to minutes: what bilinguals do when they speak or sign 1 Interactive alignment and implicit priming 1.1 Unconscious alignment in interaction 1.2 Implicit priming 1.3 Case study of implicit cross-language priming and contactinduced language change 1.4 How to study bilingual phenomena experimentally Exercises Further reading References

2 Conceptual transfer 2.1 Linguistic and non-linguistic conceptualizations 2.2 What happens when we use two languages that difer in what they code linguistically? 2.3 Case study of conceptual transfer for the concept “to be” 2.4 Case study of conceptual transfer for spatial representations Exercises Further reading References

3 Cognitive costs and cognitive load 3.1 Cognitive costs 3.2 Cognitive load Exercises Further reading References

1 3 3 5 9 14 17 19 19

21 21 24 26 28 32 35 35

37 37 41 43 43 43

Contents

vi PART II

From minutes to years: what bilinguals do when they communicate with others 4 Code-switching, repertoires, and translanguaging 4.1 Acquiring and utilizing bilingual and multidialectal codes 4.2 Integrating the language varieties of bilingual repertoires 4.3 Embracing the multidimensionality of bilinguals’ codes 4.4 How to study code-switching Exercises Further reading References

5 Social networks and accommodation 5.1 5.2 5.3 5.4 5.5

Accommodation and change Case study: convergence in language choice in Montreal Audience design and style Language mode Language mode and bilingual style 5.5.1 Style shifing in monolingual mode 5.5.2 Bilingual mode and style shifing 5.6 Social networks 5.6.1 Bilinguals’ social networks Exercises Further reading References

6 Acquisition and attrition 6.1 6.2 6.3 6.4

Bilingual acquisition Second language acquisition Attrition of the frst language How to study bilingual child acquisition and attrition of the frst language Exercises Further reading References

7 Language ideologies and dispositions 7.1 Language attitudes and language ideologies 7.2 Attitudes, ideologies, and identity 7.3 Standard language and related ideologies Exercises Further reading References

47 49 49 52 59 60 62 62 62

65 65 66 67 69 71 71 72 73 75 77 78 78

81 81 89 91 92 94 94 94

97 97 101 107 112 113 113

Contents

vii

PART III

From years to centuries: how languages change through contact 8 Contact-induced changes in grammar and borrowing 8.1 Diferent types of language change 8.2 Types of contact-induced changes 8.3 Te consequences of borrowing: loanwords 8.3.1 Types of loanwords 8.3.2 Te linguistic adaptation of loanwords 8.3.3 Te impact of loanwords on the structure of a recipient language 8.4 Convergence 8.4.1 Quantitative convergence: convergence as a matter of degree 8.4.2 Qualitative convergence: a case study of Spanish in Texas 8.4.3 Te efect of convergence on linguistic structure 8.5 How to study contact-induced changes in grammar and borrowing Exercises Further reading References

9 Linguistic areas 9.1 What are linguistic areas? 9.2 Te Balkans as a linguistic area 9.3 An overview of linguistic areas across the world 9.4 How to study linguistic areas 9.5 Case study of “have”-perfect in the Balkans Exercises Further reading References

10 Creoles, pidgins, and mixed languages 10.1 Creoles 10.1.1 When did Creoles arise? 10.1.2 How did Creoles get formed? 10.1.3 What are the structural characteristics of Creoles? 10.2 Pidgins 10.3 Mixed languages 10.3.1 When do mixed languages arise? 10.3.2 What are the structural characteristics of mixed languages? 10.3.3 How do mixed languages get formed? 10.4 Urban youth languages Exercises Further reading References

11 Minority languages, heritage languages, and immigrant linguistic practices 11.1 Diferent types of bilingual communities

115 117 117 120 121 122 122 124 126 126 127 128 129 131 132 132

135 135 136 139 141 143 147 147 147

149 149 149 150 152 156 158 158 159 160 165 167 168 168

171 171

viii

Contents 11.2 Language maintenance and language shif 11.3 Te outcomes of language contact on minority and heritage languages 11.4 Case study of Frenchville, Pennsylvania—U.S.A. 11.5 Case study of Romani Exercises Further reading References

176 179 183 185 187 188 188

Conclusion

192

Answers to exercises Glossary Index

194 201 203

Abbreviations

1, 2, 3 ACC BENF COM COMP CONJ DAT DEF ERG FN FOC INAN L1 L2 M NEG NFUT NOM O OPT PAST PERF PL POSS PRET PROX PRS Q S SG SUB V VERD

frst, second, third person accusative benefactive comitative complementizer conjunction dative defnite article ergative fnite nominalizer focus marker inanimate frst language second language masculine negation non-future nominative object optative past perfect plural possessive preterit proximal present question marker subject singular subjunctive verb veridical mood

Preface WHO IS THIS BOOK FOR? Are you an undergraduate student studying language contact or societal multilingualism? Tis book is for you! You can use it as supplementary text if you wish to extend your understanding of language contact beyond your course syllabus. If you are a graduate student or a researcher and want to learn about the state of the art in contact linguistics, this book can ofer a good introduction. Are you an instructor? You can use this book as a primary text. Last but not least, if you are a language enthusiast and curious about language contact, you are most welcome!

WHY THIS BOOK? Te connection between language contact and the bilingual speaker goes back to foundational authors in the feld of contact linguistics such as Haugen (1950) and Weinreich (1953). Yet, in our exchanges with students, we realized that the nature of this connection is unclear for many and that, in practice, students have difculties deciding whether some data refect well-established language contact phenomena or whether they result from ongoing processes at the level of the bilingual language user. Tis book’s main goal is therefore to help the reader to disentangle the efects of bilingualism that bear the seeds of potential language change from language contact, that is, once an efect that started in the bilingual mind has spread and has become conventionalized in a language community. Understanding Language Contact therefore brings together concepts from the feld of bilingualism such as alignment, priming, conceptual transfer, cognitive cost and cognitive load, L2 acquisition and L1 attrition; the feld of sociolinguistics, such as repertoires, social networks, and language ideologies; and the traditional feld of contact linguistics, discussing historical outcomes such as contact-induced change in grammar and borrowing, linguistic areas, Creoles, pidgins, and mixed languages as well as how bilingualism afects users of Indigenous, heritage, and minority languages.

HOW TO READ THIS BOOK Understanding Language Contact is divided into three parts, starting with a timescale from milliseconds to minutes at the level of the bilingual language user, moving to the efects of bilingualism over years at the level of the society, and then to the long-term efects over centuries.

Preface

xi

Tis structure refects the idea that in order to understand the outcomes of language contact, we frst need to go into some detail about the mechanisms at work in the bilingual mind, in Part I (From Milliseconds to Minutes: What Bilinguals Do When Tey Speak or Sign), and how the efects at the level of the individual can spread in the language community, in Part II (From Minutes to Years: What Bilinguals Do When Tey Communicate with Others). It may seem like a long way to get to the actual language contact outcomes, in Part III (From Years to Centuries: How Languages Change through Contact), but then the reader will have acquired the basis to understand what has happened and how we got there.We have also included some examples of language contact outcomes already in Part I. Although these three levels are related, they correspond to diferent theoretical felds and methodologies. Te book will therefore combine the methodological and theoretical advances from the feld of bilingualism with the most recent fndings from the feld of contact linguistics. Understanding Language Contact will not only take you on a journey through time but also through space, with examples drawn from a wide range of languages from around the world and case studies from each author’s original research. Understanding Language Contact contains 11 chapters. Each chapter includes a chapter preview and a chapter summary and, in the main text, an overview of the literature and discussion of key issues. Tese are sometimes illustrated by case studies and “How to Study . . .” methodological sections. Chapters also include boxes with additional background information or highlighting an original aspect of the topic under discussion. Each chapter additionally contains a “Further Reading” section, with references covering key issues discussed in the chapter, exercises, and the list of bibliographical references.At the end of the book, you can fnd the glossary as well as a section of answers to the exercises. Our ambition is to ofer a textbook that students will enjoy reading while also learning about the most recent fndings in the scientifc literature.We therefore adopted as much as possible a conversational writing style that invites the reader to be active in the discovery process of what language contact is about. Questions are asked in the main text, and there are occasional “Stop and Tink” inserts. We followed the Linguistic Society of America (LSA) and the American Psychological Association (APA) guidelines for reducing bias: We use gender neutral language and the singular gender neutral third-person pronouns they/their. We also paid extra attention to avoid gender stereotypes in our linguistic examples. Languages, language groupings, and ethnic group names are capitalized. As such, we capitalize the names “Creoles,” “Indigenous,” “Native American,” and “Aboriginal” and describe specifc languages and groups when appropriate. We do not capitalize “pidgins,” as it is not a language grouping, but we capitalize a language named “Pidgin.” As much as we tried to use the preferred language and group names for each community, we acknowledge that terms continue to change over time. We capitalize “Deaf ” to promote a sense of community, even though not everyone who has hearing loss identifes as Deaf. When we are not referring specifcally to signers or speakers, we use the generic “language user” and “language community.”

xii

Preface

In agreement with Routledge guidelines, we provide detailed alt text to help people with visual impairments access the pictures and graphs in this book.

WHO ARE WE? Evangelia Adamou, Senior Researcher at the CNRS: I’m a Greek-born researcher working in France. I grew up speaking Modern Greek at home and at school. While visiting my grandparents on Sundays, I would sometimes listen to my mother and grandmother speak in a language that they never used to talk to me and that I didn’t understand (this was intended—both my grandmother and my mother grew up speaking Greek at home and learnt this other language from interactions in the village). I had to wait until I was in my 20s to study this mystery language: It is a Balkan Slavic variety, called Nashta “our language,” that was most likely spoken in the area since the 6th century but has no ofcial language status. I also learned several languages in the classroom: French, English, Latin and Ancient Greek, Italian, Chinese, and Spanish. At the University, I studied French in my hometown, Tessaloniki, and then moved to Paris to complete my graduate studies in General Linguistics. Four years afer receiving my PhD, I was fortunate enough to become a tenured researcher at the French National Center for Scientifc Research (CNRS). As a CNRS researcher, I collaborated with many people and studied their languages: Romani spoken in the Balkans and in Latin America, Balkan Slavic, and Ixcatec, an Otomanguean language of Mexico spoken by just four speakers who were 75 years and older. When I became a mother, I spoke both Greek and French to my daughter. Although I couldn’t teach Nashta to her (since I didn’t speak it fuently), she learned Breton by attending a Breton-French kindergarten and primary school (Breton is a Celtic language that used to be spoken in her father’s family). As an immigrant, using two languages is part of my daily life, in addition to the languages I study. I code-switch and I transfer structures from one language to another, but, thanks to all the work in bilingualism and language contact, I know that this is part of what it means to be multilingual. Barbara E. Bullock, Professor of Linguistics and French Linguistics My introduction to a language other than American English came very early, in my primary school, when we began learning French at age six, with a wonderful teacher from Guadalupe. My French learning continued during the summers of my childhood and adolescence when I always spent a month with my greatuncle, a WWI veteran, who had emigrated to the U.S. from the Channel Islands, and remained a proudly bilingual Guernsey French–English speaker throughout his life. Together, we spent many days code-switching between varieties of French and English with Franco-Canadian fsherman on the docks of the New England

Preface

xiii

coast. At age 19, I was fortunate to be able to study for a year in the region of my ancestors, the canton of Neuchâtel, Switzerland, where I added a new French variety to my repertoire. Afer graduation from college (university), I worked as a bilingual marketing coordinator for the international wing of a large corporation, exporting machinery and parts to all parts of the world. I became intrigued by the miscommunications that would occur when, for instance, our customers from Korea would use noun classifers (“one piece spectrometer”) rather than determiners (“a/ one spectrometer”). My curiosity propelled me to seek out a PhD in Linguistics. My dissertation tackled a well known phonological phenomenon in Italian so I competed for a fellowship that took me to Pisa, Italy, where I spent nearly a year, speaking only Italian with my peers. Much later in life, I decided to learn Spanish only by ear through our feldwork travels to the Dominican Republic. It took me a very, very long time to say anything of substance, but now I happily speak the “fronterizo” variety of Dominican. Spoken along the country’s northwest border with Haiti, the structures, vocabulary, and pronunciation of this variety ofen remind me of French. Today, I use structures from French when speaking Spanish and from English when speaking French. My languages mix when I speak and I am comfortable with that. Almeida Jacqueline Toribio, Professor of Linguistics and Spanish Linguistics: My scholarship is informed by my personal history as a “generation 1.5” transnational bilingual whose speech repertoire includes varieties of Spanish, English, and Spanglish. I was born in the Dominican Republic and lived there with my grandparents until the age of six, when my brother and I were reunited with our parents who had immigrated to New York to seek a better life for our family. Fortunately, my parents ensured that, in addition to learning the language or our host society, we maintained our heritage language and culture. Interestingly, while I have many vivid memories of the migration experience, I do not remember learning a new language—I recall that I didn’t know English and then suddenly I did! In high school I studied a third language (French) and was interested in math science, a likely combination for a budding linguist. I pursued studies in French Linguistics (and Psychology) as an undergraduate, and I took courses on the structure of Japanese, Korean, Icelandic, German, Hindi, and Sinhala as a graduate student in Linguistics. I also came to appreciate my own language as worthy of study: I investigated my dialect of Spanish, which departs in signifcant ways from other varieties, and the structured nature of the language mixing that my brother and I engage in. In fact, I would say that I am bi-dialectal in Spanish in addition to being bi-lingual in Spanish and English. As a university faculty member, my teaching and research have focused on language variation, bilingualism, and language contact. While my early scholarship largely centered on structural linguistics, my more recent work has been

xiv

Preface

grounded in sociolinguistics and the sociology of language, considering the ways in which variables such as ethnicity, race, gender, literacy, and national origin are encoded through linguistic features and language choices. And as a faculty mentor, I have endeavored to bring positive attention to the rich diversity of the linguistic varieties of my students.

Acknowledgments

Tis book would not have been possible without the support of our institutions, allowing us to fnd time to write it. We thank the participants in the various studies we have conducted over the years among bilingual communities. Tank you for your generosity and your trust! We thank Niki Costaouec for reading parts of the book and giving feedback about the style and clarity of the text from an undergraduate’s perspective. We also thank the editor of the series, Pattie Epps, and the reviewers for their insightful comments, as well as the editorial team at Routledge for their collaboration.

REFERENCES Haugen, E. (1950). Te analysis of linguistic borrowing. Language, 26(2), 210–231. Weinreich, U. (1953). Languages in contact. Mouton.

Part I From milliseconds to minutes What bilinguals do when they speak or sign

KEY QUESTIONS Why are we so good at interacting with others using language? How does our language behavior adjust to the language behavior of others? What happens when we use two languages that have diferent grammars? Are bilinguals confronted with extra cognitive costs and cognitive load as compared to monolinguals? What are the consequences of cognitive costs and cognitive load in the long run?

1 Interactive alignment and implicit priming

CHAPTER PREVIEW When we discuss with others, we unconsciously adjust our language use to the language use of our interlocutors. Tese adjustments take place at all levels, ranging from pronunciation to grammar. To achieve alignment with our interlocutor, we rely on a cognitive mechanism dubbed “automatic implicit priming.” Te efects of alignment and priming are not doomed to disappear once we have lef our interlocutor behind but can have persistent afer-efects in long-term memory and potentially lead to a change in our language habits. In that sense, mechanisms operating within milliseconds shape language change throughout the years.

Let’s start at the very beginning: we use language to communicate with one another, whether this is to gossip, have fun, argue, share ideas, or achieve things together like building a spaceship, preparing dinner, or electing our representatives. Tis means that language is grounded in our interactions with others. We produce sounds that form words and words or signs that form sentences to convey our message. Our interlocutor listens, watches, or reads the linguistic message and, more ofen than not, grasps the intended meaning thanks to prior knowledge of the world and of the specifc language in use. But why are we so good at communicating with others using language?

1.1

UNCONSCIOUS ALIGNMENT IN INTERACTION

A leading theory posits that when we communicate using language, we align our language use with the language use of our interlocutor during conversation. Tis is known as the interactive alignment account, elaborated by Pickering and Garrod (2004). According to this view, humans automatically and unconsciously adjust their language patterns to the language patterns of others. Tis ability helps children to learn to speak or sign but can also help adults to learn additional languages later in life. Garrod and Anderson frst reported interactive alignment in their 1987 article. In this study, participants were asked to play an interactive maze game. Two players were DOI: 10.4324/9781003167952-2

4

From milliseconds to minutes

seated in diferent rooms in front of two diferent mazes. Tey could communicate with one another but could not see each other’s maze. Tey had to cooperate, as one player could unlock gates for the other player to move freely. Te researchers noted that, in order to describe their position in a maze, the players built shared conceptual models by choosing similar words. For instance, one player would say I’m at C4. Ten, the other player would adopt the same type of description using letters and numbers. Tis shared model helped players navigate the maze more efciently and complete the game faster. Te kind of interactive alignment noted for lexical items and concepts in the original study is more widespread. Many studies have since found that speakers adjust their pronunciation of words (phonetics) and their organization of words (syntax) to those of their interlocutor. Regarding phonetics, in particular, these adaptive changes are found at all levels: speech rate (how fast we speak) and intensity and fundamental frequency of our speech (sounding more or less high pitched), as well as phonetic details such as the lapse of time that occurs between the burst in the pronunciation of consonants like p and b and the beginning of the following vowel sound. Next time you talk to someone who talks fast, observe whether you start talking slightly faster too. If you have many friends who are fast talkers, chances are you will become a fast talker too. More surprisingly, recent work shows that people adjust their speech merely based on how they expect others to speak! Wade (2022) fnds that participants in an experiment started pronouncing the vowel in words like ride and dine like rod and don afer listening to an English speaker with a U.S.southern pronunciation. Let’s take another example to illustrate speech adaptation in comprehension. Remember an instance where you frst listened to another student who had a diferent English pronunciation than yours, whether because they were an international student from another English-speaking country or English was a second language or simply because they came from a diferent region, city, or neighborhood and had a diferent pronunciation than the one most commonly used in your region, city, or neighborhood? Even if at frst, most of us have difculties understanding a pronunciation to which we have been exposed very rarely, studies show that we get accustomed to it in the long run (Cristia et al., 2012). In sum, studies show that these speech adaptation efects are not merely transient but can leave a long-lasting trace and transform the way we produce and perceive speech. Such long-term efects at the level of the individual mean that groups of people who interact with one another on a regular basis end up aligning as a language community (Garrod & Doherty, 1994). As research on the topic of interactive alignment intensifed, researchers realized that speech adaptation is not restricted to interaction settings, where one might be adopting a prosocial behavior. Rather, alignment seems to be taking place automatically even in non-interactional settings, where one doesn’t need to be collaborative, for example, when participants were asked to speak afer having listened to a recording. Some researchers therefore suggest that alignment is an automatic process that allows humans to tune into the language environment, whether this is interactional or not. We can therefore refer to it simply as “alignment.”

Interactive alignment and implicit priming

5

But what is the precise mechanism that allows us to adjust to the language behavior of others? Go to the next section to fnd out more about it.

1.2

IMPLICIT PRIMING

Alignment works mainly thanks to a cognitive mechanism that is called “automatic implicit priming.” Implicit priming refers to the fact that a language user will tend to repeat a previously used, heard, or read word without realizing it. In more technical terms,“priming” is the observation that processing one stimulus (the prime) unconsciously afects processing a subsequent stimulus (the target). Stop and think: Did you ever notice that when you write an essay and you reread what you wrote, you fnd out that you used the same word in two sentences that immediately follow one another? Tis is because the use of a word “primes” the use of the same word in the seconds that follow. Te brain made an efort to reach for this word so it might as well use it again (teachers and proof-readers tend to dislike the outcome and encourage us to fnd a synonym instead). Te same mechanism is at play for syntax, the order of words and their meaning and phonology. Interestingly, priming is cumulative across these various levels: for instance, priming of a syntactic structure is intensifed when some words are the same (Branigan et al., 2000). More concretely, this means that speakers are more likely to say: (i)

Te clown hands the balloon to the child afer hearing:

(ii)

Te seller hands the bag to the client than afer hearing:

(iii)

Te seller gives the bag to the client.

What is common in the frst two sentences when compared to the frst and third sentence? Not just the syntactic structure, X verb Y to Z, as is also the case for the frst and third sentences, but also the specifc verb “to hand” (i.e., hands). Tis is more important than the fact that sentence (ii) and (iii) share the same nouns (i.e., seller, client, bag). Te additional efect of priming achieved through the repetition of some important words in a sentence is dubbed “lexical boost.” A classic method to observe priming is by building an experiment in which participants have to repeat sentences and describe pictures (Bock, 1986). Tese studies overall fnd that when participants repeat a passive sentence like: A compromise is being suggested by the president, they are “primed” by this structure. In practice, this means that they are more likely to produce the same structure, in this case passive, when describing an unrelated upcoming picture that allows for this structure to be used, as opposed to when they have just repeated an active sentence like Te president is suggesting a compromise. Another method is to rely on corpus studies of written or spoken data to observe

6

From milliseconds to minutes

what people are doing in real life. For example, Gries (2005) analyzed large English freespeech corpora and found that syntactic priming is persistent across several utterances. Other phenomena that have been found to be sensitive to priming include a number of alternative structures available in a given language. Tis is the case for: • Noun phrase modifer type, such as adjective as in the red square vs. relative clause as in the square that’s red (Cleland & Pickering, 2003). • Te order of two adjectives, such as the large red chair vs. the red large chair (Goudbeek & Krahmer, 2012). • Ditransitive constructions (for example, with verbs “to give”) that can alternate in English between a prepositional object, such as the clown gives the balloon to the child, and double object constructions, such as the clown gives the child the balloon (Bock, 1986; Bock & Loebell, 1990). • Possessive constructions like the fork of the girl vs. the girl’s fork (Bernolet et al., 2012). • Transitive and intransitive clauses, such as the child eats the banana and the child eats (van Gompel et al., 2012). • Auxiliary verb-participle order; for example, in Dutch the participle can either be at the end of a sentence, as in De man belde de politie omdat zijn portemonee was gestolen, or the auxiliary can be at the end of the same sentence, as in De man belde de politie omdat zijn portemonee gestolen was. “Te man called the police because his wallet had been stolen” (Hartsuiker & Westenberg, 2000). • Relative clause attachment, as in “the secretary of the president [who is working on this fle]RC” where the relative clause introduced by “who” can, in some languages, refer to either the secretary or the president (high or low attachment) (Scheepers, 2003). More importantly for our focus in this book, researchers have found that implicit priming takes place not only within a single language, as in these examples from English, but also across languages for a bilingual language user. However, a meta-analytic study (see Glossary) that analyzes all published data together found that structural priming between the two languages of bilinguals is less robust than among monolinguals (Mahowald et al., 2016). Let’s see an example from a study that Torres Cacoullos and Travis conducted based on a conversational corpus from Spanish-English bilinguals residing in New Mexico, U.S. (2018). As you know, English is a language where personal pronouns like I and you are obligatory, but this is not the case in Spanish for yo and tu. It is not grammatical in English to say eat an apple if the intended meaning is I eat an apple, but it is perfectly fne in Spanish to drop the frst-person pronoun yo in the same sentence. Torres Cacoullos and Travis found that Spanish-English bilinguals in New Mexico tend to use this frst-person pronoun when they speak Spanish more than monolingual Spanish speakers do. More specifcally, bilinguals in New Mexico use the Spanish pronoun yo more ofen following a sentence with an English frst-person pronoun I. Tis is what we call “cross-language priming.” Yet the same speakers tend to use yo

Interactive alignment and implicit priming

7

even more following a sentence in Spanish with the Spanish pronoun yo. Tus, withinlanguage priming is stronger than cross-language priming. Tis diference in the strength of the priming efect may be due to the fact that cross-language priming benefts less from the lexical boost, discussed earlier. Tis hypothesis makes sense since priming in a single language relies on the exact same lexical items, but priming from one language to the other involves diferent lexical items that we refer to as “translation equivalents.” Tis does not mean that translation equivalents are not at all associated in the mind of a bilingual. For example, van Hell and de Groot (1998) found that translation equivalents are linked to some extent, although they are associated more strongly when the two words share similar phonological structure, for example, because the languages are closely related or because a word was borrowed from one language to another. Another factor that can modulate the strength of the priming efect among bilinguals is their profciency in the second language. Stop and think: What would be your prediction for the role of profciency when processing an L2? Would higher profciency levels in the second language provide an additional boost to the priming efect of the L1 or not? Read what researchers have found in the following. Studies report that priming is more robust among less profcient speakers of an L2 (a second language) than among profcient or L1 (frst language) speakers. Tis suggests that less profcient speakers of an L2 rely more strongly on lexical item-specifc representations in the L2 (e.g., clown, balloon, child), in other words, that they have not yet formed an abstract rule in the L2 for a structure of the type X verb Y to Z (Kim & McDonough, 2007; Bernolet et al., 2013). Finally, frequency of a given structure in usage in real life is important. For example, Torres Cacoullos and Travis (2018) found that the absence of Spanish pronouns favors the absence of Spanish pronouns, but the absence of English pronouns, which is rare in English, does not prime the absence of Spanish pronouns.Although frequent structures may have an important efect, less frequent structures, like passives in English as opposed to active sentences, can sometimes be more sensitive to priming precisely because they are so rare.You might ask why this would be the case. Listening to a rare structure is registered by the mind as an exceptional event and helps reset a language user’s expectations by bringing the rare structure to the foreground; this efect is known as the “inverse-preference efect” (Ferreira & Bock, 2006). But why is priming important for language contact? If we consider priming to provide insights into the linguistic representations in language users’ minds, then cross-language priming suggests that linguistic structures in the two languages share some aspects of their representation in the language user’s mind (either overlapping fully or being related in some looser ways). To better understand what shared conceptual representations are, let’s look closely at one example from a trilingual Dutch-English-German speaker.As you may have noted, all three are Germanic languages and are therefore related. Tese languages are represented in Figure 1.1. with three fags: a Flemish fag for Dutch (on the lef upper side), a

8

From milliseconds to minutes

Figure 1.1 Model for the representation of noun phrase structure in Dutch-English-German trilinguals

Conceptual level

shark

haai

Hai

Lemma level

N

AdjN

RC Verb-fin

RC Mod-fin

Source: Adapted from Bernolet et al. (2007, p. 945)

British fag for British English (in the middle), and fnally a German fag for German (on the right upper side). Te model represents the mind of the trilingual speaker as having a single concept, represented by the picture of a shark, associated with three “lemma” representations (see Glossary): Dutch haai, English shark, and German Hai. You can see again that each lemma is connected to a little fag for each language. At the bottom of the fgure, you can see that all three lemmas are connected to the same node “N” for “noun.” In addition, for all three languages, the three lemmas are connected to the node that indicates word order: “AdjN” signals that adjectives (Adj) precede nouns (N) in this language. Let’s turn to look at the remaining nodes. Do you notice a diference between the three languages? Take a minute to look at the lower part of the graph and see how “RC Verb-fn(al)” and “RC Mod(ifer)-fn(al)” are connected to the diferent lemmas. Since the word order difers in English in comparison to Dutch and German, authors suggest that priming between the English and Dutch or German relative clauses cannot successfully take place. In contrast, syntactic priming can take place between languages that share the same word order. As cross-language priming also takes place at the phonological level, researchers elaborated representations that consider the “word form level” which includes phonemes (see Glossary), such as /p/, /d/, /l/, in Figure 1.2. When words in the two languages share similar phonemes (as in /vork/ and /fork/, where /o/, /r/, and /k/ are shared), this strengthens the activation of their translation equivalent lemma (as vork is activated by fork) (Bernolet et al., 2012). Tis activation results in stronger priming between the syntactic structures of the two languages. Tis phonological boost is not observed when the two words do not share many similar phonemes (as in /pop/ and /dol/ that only share /o/), and the translation equivalent is not activated (in this case, doll does not activate pop).

Interactive alignment and implicit priming

9

Figure 1.2 Representation of shared syntax model including a phonological level

Conceptual level

doll

pop

vork

N

s-gen /pop/ /p/

/d/

of-gen

/dol/ /l/

/v/

/f/

Wordform level

/fork/

/vork/ /o/

Lemma level

fork

/r/

/k/

Source: Adapted from Bernolet et al. (2012, p. 511)

You can follow this in Figure 1.2, which represents a model illustrated for a bilingual Dutch-English speaker. At the top, the conceptual level is illustrated by a drawing of a doll and a fork. Te pre-linguistic concept of “doll” corresponds to two lemmas: pop in Dutch (see the Flemish fag) and doll in English (see the British fag). Te concept of “fork” corresponds to two-word forms that sound alike: vork in Dutch and fork in English. Ten it can be seen that both languages have two possible structures for expressing possession (also referred to as “genitive”): One structure is the “s-genitive,” as in the nurse’s fork (De zuster haar vork in Dutch), and another structure is the “ofgenitive,” as in the fork of the nurse (De vork van de zuster in Dutch). Tese models suggest that language users may share some aspects of their representations in their two languages depending on how closely related and typologically similar the languages are. More importantly, researchers hypothesize that since implicit priming has long-term efects, it could be a key mechanism for contactinduced language change (Kootstra & Şahin, 2018; Adamou et al., 2021). Is there any evidence to date demonstrating a link between cross-language priming (in the language user’s mind) and contact-induced change (across generations of language users)? Let us illustrate this possibility with a recent case study.

1.3 CASE STUDY OF IMPLICIT CROSS-LANGUAGE PRIMING AND CONTACT-INDUCED LANGUAGE CHANGE Our case study is word order in adjective phrases, that is, phrases that combine an adjective and a noun. Most languages exhibit a noun-adjective order, others an adjective-noun order, and others vary between the two.

10

From milliseconds to minutes

Discover: You can visit the World Atlas of Language Structures online to check the order of nouns and adjectives in various languages of the world (see https:// wals.info/feature/87A#2/17.9/144.3 Dryer, 2013). We focus on bilingual speakers of Romani, an Indic language, and Romanian, a Romance language, who live in Romania. Speakers of Romani live in many European countries as well as throughout the Americas. For more information about Romani, you can read the insert. Romani is a language in long-term and intense contact with a variety of languages in Europe, the U.S., and many Latin American countries. For a long time, it was not a standardized language that could ofer an ofcial “norm” to which speakers could turn, even though Romani is nowadays taught and standardized in Romanian schools and at the university.We can therefore observe how the Romani language developed throughout the centuries by comparing the contemporary varieties in various settings and whether they have become more similar with the languages that are spoken in those settings.

A few words about the Romani language Romani is spoken by the Roma people, most widely known as “Gypsies”; especially when used by outsiders, the latter is considered a slur by some groups, as it is related to historical anti-Gypsyism that led to the Holocaust and is still present today in many countries (read the Vogue article “Te Fight to Strike ‘Gypsy’ From the Fashion Lexicon” for an example of eforts by Roma people to reclaim their cultural identity: www.vogue.com/article/roma-activism-fashion). Romani is an Indo-European language like English but belongs to the branch of Indic languages, as Roma most likely lived in present-day India more than ten centuries ago. Te presence of Roma in Romania, which is the focus of the study reported in this chapter, goes back to the late fourteenth century.At present, Romania is home to the largest population of Romani speakers in Europe, with estimates ranging from 250,000 up to more than 2 million speakers.Virtually all Roma who live in Romania speak Romanian, which is the language of the Romanian State and the language spoken by most Romanians. Like Romani, Romanian is also an Indo-European language, but it belongs to the Romance branch, like Spanish.Although Romanian is the main language of instruction in Romania,Romani is nowadays recognized as a minority language and is taught at schools and at the university. Romani speakers can read the standard Romani spelling, taught at schools, but most can read Romani written informally using the same letters as the ones used in Romanian.

Let’s begin by examining what noun (N) adjective (ADJ) order is dominant in various Romani varieties in the world. We can do this rather easily by searching the Romani Morpho-syntax database, which is accessible online (see https://romani.

Interactive alignment and implicit priming

11

humanities.manchester.ac.uk/rms/). Tis database includes responses to a questionnaire by more than 100 speakers of Romani living in diferent countries. Te responses to the Romani Morpho-syntax questionnaire are not only transcribed, but the original recordings are also available online if we want to listen to the precise responses to be sure that the transcriptions are accurate or look for extra details in the pronunciation. Adamou et al. (2021) coded 3,000 occurrences of the adjective noun order in the questionnaire. Tey found that adjectives follow the noun in Romani spoken in some countries (e.g., dog brown), but they precede the nouns in others (e.g., brown dog). You can see in Figure 1.3 that Romani speakers use less than 20% of N-ADJ order (e.g., dog brown) when they live in Finland, Republic of North Macedonia, Bulgaria, Hungary, Serbia, Russia, Slovenia, Croatia, Slovakia, Poland, Ukraine, Lithuania, Estonia, Albania, Greece, and Latvia. Tese Romani speakers are currently in contact with languages with a dominant ADJ-N order (e.g., brown dog) like Slavic languages (e.g., Macedonian, Bulgarian, Serbian, Russian, Czech, Polish, Ukrainian), and the Finnic (Finnish and Estonian), Ugric (Hungarian), Baltic (Lithuanian and Latvian),

Figure 1.3 Percentage of all (DET)-N-ADJ responses in the RMS database by country. Te size of the dots corresponds to the number of speakers. Country codes: FIN (Finland), MK (North Macedonia), BG (Bulgaria), HU (Hungary), YU (former Yugoslavia), RUS (Russia), SLO (Slovenia), HR (Croatia), SK (Slovakia), PL (Poland), UKR (Ukraine), LT (Lithuania), EST (Estonia), AL (Albania), GR (Greece), LV (Latvia), CZ (Czech Republic), MD (Moldova), IT (Italy), RO (Romania), MX (Mexico)

Source: Adamou et al. (2021, p. 1639)

12

From milliseconds to minutes

and Germanic languages (German and Swedish), as well as Turkish and Greek. In contrast, Romani speakers who used the N-ADJ order more regularly, amounting to up to 70% of responses in total, are all in contact with Romance languages Italian, Spanish, Romanian, and the Moldovan variety of Romanian. Tese Romance languages have a preferred N-ADJ order (e.g., dog brown). In the samples from Romania, in particular, N-ADJ is the preferred order for all 19 speakers who responded to the Romani Morpho-syntax questionnaire, with ADJ-N order found in approximately 30% of the responses among individual speakers. In sum, this study shows that Romani-Romanian speakers use the N-ADJ order more frequently as compared to Romani speakers in contact with languages that do not have dominant N-ADJ order. Adamou et al. (2021) then examined whether the N-ADJ order is in use among RomaniRomanian bilinguals when they speak freely in Romani. Te analysis of a 10,000-word corpus of interviews with four older adult speakers reveals three diferent types of word order involving a noun, an adjective, and a determiner (DET). Te example in (1.1) illustrates the inherited DET-ADJ-N order with the color adjective “white.” Te example in (1.2) illustrates the innovative DET-N-ADJ order with the size adjective “big.” Te example in (1.3) reveals a third construction, DET-N-DET-ADJ, with the size adjective “big.” (1.1) DET-ADJ-N kärde kola parne ròkie had those white dresses “(women) had those white dresses . . .” (1.2) DET-N-ADJ äk nekàzo a grief “a big grief . . .”

baro big

(1.3) DET-N-DET-ADJ keldăs i vestea celebrated.3SG the news “s/he celebrated the great news . . .”

i the

bari big

Tis corpus study confrms that Romani-Romanian bilinguals alternate between ADJ-N and N-ADJ order in natural conversations in Romani. Adamou et al. (2021) then conducted a priming experiment with 90 Romani-Romanian bilinguals. In this experiment, a priming trial consists of a sentence trial and a picture trial. Participants were given a folded sheet of paper. On one side of the paper (the top in Figure 1.4), a sentence was written either in Romani or in Romanian. Tese were the “prime” sentences. Participants had to read them. Tese sentences would either be in Romanian and consistently follow the N-ADJ order (e.g., dog brown) or in Romani, where they followed either the ADJ-N order (e.g., brown dog) or the DET-N-DET-ADJ order (e.g., the dog the brown). Ten participants were asked to turn the sheet of paper, read the frst word in Romani, meaning “now,” and complete the sentences by looking at the pictures. Tey had to respond by repeating the sentence structure “Choose . . .” but use the object that was not named in the previously read sentence. For example, if

Interactive alignment and implicit priming

13

Figure 1.4 An example of a prime sentence (“Choose the brown dog,” here in Romani Alosar o maronio jukel) and a set of target pictures used in the priming experiment. Te sheet was folded in the middle (where the line appears) so that participants read the prime sentence without having seen the target picture. Tey then turned the sheet and only saw the two pictures and the Romani word Akana “now” without being able to see the prime sentence

Source: Adamou et al. (2021, p. 1644)

the sentence they had read was “Choose the brown dog,” they would look at a pink and brown dog and say “Now choose the pink dog.” If the cross-language priming efect is present, they should prefer a N-ADJ order in their responses in Romani afer listening to a Romanian sentence that has the same N-ADJ order. Te statistical analyses of the results reveal a signifcant efect of the Romanian prime N-ADJ (e.g., dog brown) on Romani N-ADJ responses as predicted. Overall, these results confrm the hypothesis that cross-language priming is driving the selection of adjective/noun order in Romani, favoring the use of N-ADJ order immediately following an N-ADJ prime from Romanian. Tis means that the priming mechanism may be the basis of the increased N-ADJ order that we fnd among Romani speakers living in Romania since they all speak Romanian, a Romance language with N-ADJ order. In contrast, Romani speakers who live in countries where the majority language speakers opt for an ADJ-N order seem to maintain an ADJ-N order in Romani too. If you are also interested in using experimental methods, read the next section!

14

1.4

From milliseconds to minutes

HOW TO STUDY BILINGUAL PHENOMENA EXPERIMENTALLY

When we think of experiments, we frst think of scientists in natural sciences. Indeed, scientists observe a phenomenon in real life or in the lab, but they also frequently conduct experiments in order to isolate, prepare, and manipulate the object of their investigation and its environment. Experiments, however, are not only reserved to physicists or chemists but are increasingly being used in linguistics. How do linguists run an experiment? First, keep in mind that experiments are quantitative methods, which means that you will need to conduct some kind of statistical analysis. If you don’t have any background in statistics, you can either follow some training at your university or attend classes online. Tis is a great skill-set that will serve you in the future. However, if you have always hated math, then you can just team up with a statistician at your university (remember to ofer co-authorship if you are to publish the study). When doing experimental research, you will invariably be confronted with the question of “statistical power.” In order for your results to be statistically signifcant, you will need many participants and many experimental trials. How many? It really depends on the topic, but a good number to start with is 24 participants. Of course, what is important is to know who your participants are. For example, if you just say that your participants are “bilingual,” it’s not enough: We now know that fne-grained diferences are important for understanding bilingual behavior, like age of acquisition of the two languages, profciency in the two languages, daily use, and frequency of some bilingual behavior like code-switching. You will then identify a “dependent variable,” something that is the focus of your study. Your goal is to test whether the dependent variable is afected by an “independent variable.” For example, the dependent variable can be the reaction times (in milliseconds) of participants in a response task and the independent variable whether the stimuli were in language A or B. Nowadays, complex statistical models allow for the examination of more variables and their interactions so that you can also consider each participant’s behavior, among other factors. When you are running an experiment, you will typically need to include a “control” group of participants that difers on the variable that is hypothesized to be the main cause for an observed behavior. For example, when studying bilingualism, you would like to compare a group of bilingual participants with a group of monolingual participants to make sure a linguistic phenomenon among the bilinguals is not also found among the monolinguals and is therefore potentially independent from bilingualism. Tis has been standard procedure in the past, but nowadays more and more researchers call for detailed comparisons at the level of bilingualism (you can also refer to this group as a “comparison” group). For example, you might want to test bilingual participants who learned language B as adults and bilingual participants who learned language B in childhood to see the diference in the age of acquisition, all other things being equal. Another characteristic of experiments is the use of “randomized trials.” Tis means that not all participants will respond to the same stimuli in the same order. Tis way scientists make sure that the order of appearance of the stimuli is not afecting the

Interactive alignment and implicit priming

15

results (we know participants get better as they perform a task, so that the early responses difer from the late responses). Don’t forget that you will also need to include stimuli that you are not interested in analyzing: the “fllers.” You may ask: “Why waste our time on something that we won’t study?”We do that so that participants cannot easily identify the research question and adjust their responses accordingly (you may be surprised to learn that participants sometimes just want to make the researchers happy and answer in specifc ways). However, you are also right: Researchers don’t like to waste responses. So they sometimes use the fllers as an opportunity to conduct a diferent study or just to make sure that participants responded to the task as expected. One important precaution before running an experiment is to “norm” your stimuli or use previously normed stimuli (e.g.,Wu & Hofman, 2021). Tis means that people who do not take part in the experiment have rated your stimuli to tell you whether they think these are well constructed and well understood (e.g.,“1” for “very bad” and “5” for “perfect”). Tere is a great variety of experimental designs to choose from: judgment tasks (e.g., participants judge the acceptability or naturalness of linguistic stimuli using a 5-point or 7-point scale, as in the norming study mentioned before), picture-matching tasks (e.g., participants see a pair of pictures and are asked to match one of the pictures with the content of a spoken word or sentence that they hear), and priming tasks, like the one we saw in this chapter. Tere are also many semi-experimental tasks that you can conduct like picture naming (e.g., participants name objects depicted in drawings or illustrated in pictures) and language production tasks (e.g., participants describe a video clip as they watch it or as they remember it). If you want to conduct an experiment, here are the main steps to follow: Step 1. Identify a research question When conducting an experiment, your main goal is to test and update a scientifc theory (see Glossary). Tis is why you should always start with a clear research question by reading the scientifc publications. Step 2. Clearly formulate your research question and predictions Make sure you clearly formulate your research questions and predictions before conducting the experiment.You can even pre-register your study using the Open Science Framework (OSF) (https://osf.io/). Step 3. Choose your experimental design Move on by carefully preparing your experimental design.You have to get the design right so as not to waste time and resources running an experiment that has huge faws. Don’t forget to ask for approval by the relevant Institutional Review Board (IRB)! Step 4. Prepare your stimuli and build your experiment

16

From milliseconds to minutes

It’s time to prepare your stimuli, whether they are written, audio, video, and picture stimuli or a combination of these. Tere are many tools available to build and run a psycholinguistic experiment. We recommend the open-source experiment builder Open Sesame, as it is free and easy to use: https://osdoc.cogsci.nl/. Step 5. Select your participants Depending on your research questions, you will have to make a call for participation by advertising the study, for example, on social media, or contact your participants, for example, through associations. Even if “convenience” samples are . . . convenient (and for this reason widespread in research), you should pay attention to the selection of your participants depending on the research questions you ask.Allow yourself to go beyond the typical student sample and open up to more diverse samples to include participants who do not have a high school diploma or equivalent, whose self-reported income is in the lower income brackets, and more generally may come from under-represented groups. Te typical student sample refects the “WEIRD” bias in research, where researchers mainly work with “Western, Educated, Industrialized, Rich, and Democratic” populations (Henrich et al., 2010). In any case, don’t forget to describe the sociolinguistic characteristics of your participants and of course assess their profciency and dominance in their languages. Before their participation in the experiment, participants will have to sign a consent form. Tis form briefy explains what participants are expected to do during the experiment; tells them that they can leave before the end of the experiment if they feel uncomfortable; briefy summarizes why the researchers are conducting this study; mentions some possible benefts for them or their language community; and informs them about whether they will receive compensation and who funds the research, where the collected data will be stored and published, how the researchers will ensure respect for the privacy and confdentiality of participants’ personal information, and fnally, whether they want to hear back from you about the results of the study. Step 6. Collect your data It is now time to conduct the experiment! Step 7. Analyze your data Well done! You now have your results. Statistical analyses are typically needed at this stage. Don’t hesitate to contact a statistician at your institution to do state-of-the-art statistics! Step 8. Interpret your results Based on the results of the data analysis, you can assess the predictions that you formulated in Step 2. Whether the predictions are borne out or not, it’s still interesting

Interactive alignment and implicit priming

17

to share your results. Unless you had a major faw in the design, all results are useful. One could argue that even major faws are important to report since they can help others avoid them.

Chapter summary When we communicate with others, we adjust our language to their language. Tis alignment to our language environment is automatic. It relies on the mechanism of implicit priming whereby we tend to reuse previously heard, spoken, or read sounds, words, or structures. As the efects of implicit priming can be long lasting and alter linguistic representations in the bilingual mind, implicit priming may be one mechanism that leads to contact-induced language change. It becomes apparent that a mechanism at the level of milliseconds is the starting point of linguistic change that can be observed afer many years.

EXERCISES 1.

Discover

Compare the prime and target structures in the following table and answer the following questions with “yes” and “no”: Is there a shared phrase structure? Is there a shared lexical item? Is there shared information structure as noted through word order? Is there shared semantics? Prime structure

Target structure

Te famous tennis player drove an old Mercedes to the restaurant

Te artist is handing a paintbrush to the child

Te worker dragged the bucket around the manager

Te zoo-keeper brought the food to the lion

Te musician needs to borrow a microphone from her friend tonight

While the poet traveled in France, she wrote many letters to her family

Shared phrase structure?

Shared lexical item?

Shared Shared information semantics? structure?  

From milliseconds to minutes

18

Ο πρόεδρος φύλαξε το χρυσό μετάλλιο μέσα στο συρτάρι “Te president kept the gold medal in the drawer”

Te hotel receptionist gave a key to the guest

Op de tafel ligt een bal “On the table lies a ball”

Naast het hok zit een hond “Beside the kennel sits a dog”

Te square that’s red

Te square that’s green

Te burglars broke the door down

Te high prices scared the customers of

. . . omdat de weg geblokkeerd was “because the road blocked was”

. . . omdat hij zijn been gebroken had “because he his leg broken had”

2.

Discover a.

Read the following sentences. What do these sentences have in common? Te senator was awed by the statue. Te swimmer was stung by the jellyfsh. Te escaping prisoner was illuminated by the guard tower. Te driver was confused by the blinking trafc light. Te Dalmatian was pursued by the fre truck. Te child was splashed by the drinking fountain.

b.

Do you predict the efect of implicit priming to be strong?

c.

Can you cite the factors that are expected to modulate priming in these sentences?

d.

How will these factors impact the size of the priming efect (will they make it stronger or not)?

Interactive alignment and implicit priming

3.

19

Discuss in class

Did you or your friends move from one place of your country to another to study? Did you notice any diferences in the speech of the people around you? Did you notice any changes in your own language behavior?

4. Test in class Record a conversation between two classmates. Replay the recording and notice in what ways, if any, your classmates aligned in their language. Think of all possible levels: from speech rate, to phonetic details, to shared structures and lexical items.

FURTHER READING Menenti, L., Pickering, M. J., & Garrod, S. C. (2012). Toward a neural basis of interactive alignment in conversation. Frontiers in human neuroscience, 6, 185. https:// doi.org/10.3389/fnhum.2012.00185 Tis freely accessed article is written by the two proponents of the interactive alignment account, Pickering and Garrod. In collaboration with neuroscientist Menenti, they aim to summarize fndings from neuroscience that help us to better understand how alignment in conversation takes place in the brain. Authors note that alignment is not only observed in production but also in comprehension and that these two aspects are closely related. Te authors further highlight specifc neural mechanisms that take place in communication.

REFERENCES Adamou, E., Feltgen, Q., & Padure, C. (2021).A unifed approach to the study of language contact: How cross-language priming drives change in noun-adjective order. International Journal of Bilingualism, 25(6), 1635–1654. https://doi.org/10.1177/13670069211033909 Bernolet, S., Hartsuiker, R. J., & Pickering, M. J. (2007). Shared syntactic representations in bilinguals: Evidence for the role of word-order repetition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33, 931–949. https://doi.org/10.1037/0278-7393.33.5.931 Bernolet, S., Hartsuiker, R. J., & Pickering, M. J. (2012). Efects of phonological feedback on the selection of syntax: Evidence from between-language syntactic priming. Bilingualism: Language and Cognition, 15, 503–516. https://doi.org/10.1017/s1366728911000162 Bernolet, S., Hartsuiker, R. J., & Pickering, M. J. (2013). From language-specifc to shared syntactic representations: Te infuence of second language profciency on syntactic sharing in bilinguals. Cognition, 127, 287–306. https://doi.org/10.1016/j.cognition.2013.02.005 Bock, K. (1986). Syntactic persistence in language production. Cognitive Psychology, 18, 355–387. https://doi.org/10.1016/0010-0285(86)90004-6 Bock, K., & Loebell, H. (1990). Framing sentences. Cognition, 35, 1–39. https://doi.org/10.1016/ 0010-0277(90)90035-I Branigan, H. P., Pickering, M. J., Stewart, A. J., & McLean, J. F. (2000). Syntactic priming in spoken production: Linguistic and temporal interference. Memory & Cognition, 28(8), 1297–1302. https:// doi.org/10.3758/BF03211830

20

From milliseconds to minutes

Cleland, A., & Pickering, M. (2003). Te use of lexical and syntactic information in language production: Evidence from the priming of noun-phrase structure. Journal of Memory and Language, 49, 214–230. https://doi.org/10.1016/s0749-596x(03)00060-3 Cristia, A., Seidl, A., Vaughn, C., Schmale, R., Bradlow, A., & Floccia, C. (2012). Linguistic processing of accented speech across the lifespan.Frontiers in Psychology 3. https://doi.org/10.3389/fpsyg.2012.00479 Dryer, M. S. (2013). Order of adjective and noun. In M. S. Dryer & M. Haspelmath (Eds.), Te world atlas of language structures online. Max Planck Institute for Evolutionary Anthropology. Retrieved August 16, 2020, from http://wals.info/chapter/87 Ferreira, V. S., & Bock, K. (2006). Te functions of structural priming. Language and Cognitive Processes, 21, 1011–1029. https://doi.org/10.1080/01690960600824609 Garrod, S., & Anderson, A. (1987). Saying what you mean in dialogue: A study in conceptual and semantic co-ordination. Cognition, 27, 181–218. https://doi.org/10.1016/0010-0277(87)90018-7 Garrod, S., & Doherty, G. (1994). Conversation, co-ordination and convention: An empirical investigation of how groups establish linguistic conventions. Cognition, 53(3), 181–215. https://doi. org/10.1016/0010-0277(94)90048-5 Goudbeek, M., & Krahmer, E. (2012). Alignment in interactive reference production: Content planning, modifer ordering, and referential overspecifcation. Topics in Cognitive Science, 4, 269–289. https://doi.org/10.1111/j.1756-8765.2012.01186.x Gries, S. T. (2005). Syntactic priming: A corpus-based approach. Journal of Psycholinguistic Research, 34, 365–399. https://doi.org/10.1007/s10936-005-6139-3 Hartsuiker, R. J., & Westenberg, C. (2000).Word order priming in written and spoken sentence production. Cognition, 75, B27–B39. https://doi.org/10.1016/s0010-0277(99)00080-3 Henrich, J., Heine, S., & Norenzayan,A. (2010). Te weirdest people in the world? Behavioral and Brain Sciences, 33(2–3), 61–83. https://doi.org/10.1017/S0140525X0999152X Kim,Y., & McDonough, K. (2007). Learners’ production of passives during syntactic priming activities. Applied Linguistics, 29, 149–154. https://doi.org/10.1093/applin/amn004 Kootstra, G. J., & Şahin, H. (2018). Crosslinguistic structural priming as a mechanism of contactinduced language change: Evidence from Papiamento-Dutch bilinguals in Aruba and the Netherlands. Language, 94, 902–930. https://doi.org/10.1353/lan.2018.0050 Mahowald, K., James, A., Futrell, R., & Gibson, E. (2016). A meta-analysis of syntactic priming in language production. Journal of Memory and Language, 91, 5–27. https://doi.org/10.1016/j. jml.2016.03.009 Pickering, M. J., & Garrod, S. (2004). Toward a mechanistic psychology of dialogue. Behavioral and Brain Sciences, 27, 169–226. https://doi.org/10.1017/s0140525x04000056 Scheepers, C. (2003). Syntactic priming of relative clause attachments: Persistence of structural configuration in sentence production. Cognition, 89, 179–205. https://doi.org/10.1016/ s0010-0277(03)00119-7 Torres Cacoullos, R., & Travis, C. (2018). Bilingualism in the community: Code-switching and grammars in contact. Cambridge University Press. van Gompel, R. P. G., Arai, M., & Pearson, J. (2012). Te representation of mono- and intransitive structures. Journal of Memory and Language, 66(2), 384–406. https://doi.org/10.1016/j. jml.2011.11.005 van Hell, J. G., & de Groot, A. M. B. (1998). Conceptual representation in bilingual memory: Efects of concreteness and cognate status in word association. Bilingualism: Language and Cognition, 1, 193–211. https://doi.org/10.1017/s1366728998000352 Wade, L. (2022). Experimental evidence for expectation-driven linguistic convergence. Language, 98(1), 63–97. www.muse.jhu.edu/article/849527 Wu,W., & Hofman, P. (2021, June 16). Validated measures of semantic knowledge and semantic control: Normative data from young and older adults for more than 300 semantic judgements. osf.io/9px7g

2 Conceptual transfer

CHAPTER PREVIEW Tis chapter is entitled “Conceptual Transfer.” You may already be asking yourself what “conceptual” means: What are concepts? Are concepts independent from language? Does this mean that animals can also form concepts? Or are concepts shaped by language? If yes, to what extent? Only when we use language or even when we don’t? And if concepts are shaped by language, what happens when we use two languages and these languages have very diferent concepts? Do we keep these concepts distinct, or do we favor one conceptualization over the other? Or do we mix concepts? Finally, how do these processes impact language in the long run?

2.1

LINGUISTIC AND NON-LINGUISTIC CONCEPTUALIZATIONS

Concepts are classifcations of real-world stimuli: Tey are not merely classifcations based on perception (these would be categorizations) but are formed while abstracting away from the specifc stimuli. We have long considered that only humans were capable of such sophisticated categorizations of the world. Recent research with animals challenges this view, as it increasingly shows that some species are capable of forming concepts. One example comes from bees. Despite their tiny brains, bees can learn spatial relationships such as “above” and “below” or “lef” and “right” and combine them with the concept of “diference” (Avarguès-Weber et al., 2012). Tis leads to a humbling conclusion: Human language is not necessary to form concepts. Studies also show that even though human babies do not productively use language, they seem to be having some conceptual representations of the real world. Tey can represent causal and spatial relations for objects that are neither too small nor too big, neither too close nor too far (wow, this ball over there just fell down!), they can represent number (there are quite a few balls here), and they can represent agents (this person just kicked a ball!). Tese pre-linguistic concepts are understood as being dynamic and shaped by the experience that the baby will have with the environment and with language. DOI: 10.4324/9781003167952-3

22

From milliseconds to minutes

Why brain size does not matter Are you wondering why brain size does not matter as much as we thought it mattered? Recent research reveals that the brain of some avian species contains numbers of neurons similar to those of mammals, even though mammals have much bigger brains. Indeed, mammals’ brains are heavy, and birds need to have the lightest possible brains in order to be able to fy. To solve this size/weight equation, avian brains just have much higher density than the brains of mammals (Olkowicz et al., 2016).

What happens once language is acquired? When we want to communicate a thought, we start by formulating a message in a conceptual system that is not linguistic; we refer to it as “pre-linguistic.” We then engage with language production and with the linguistic part of the message. However, researchers acknowledge that the conceptual pre-linguistic message is shaped to some extent by the specifc conceptual features that are available in a given language. Some researchers stress that when we use language, we necessarily flter our thoughts using language-specifc categories. Tis is known as thinking for speaking (Slobin, 1987). Other researchers go as far as to claim that even when we do not use language, we experience the real world through linguistic conceptualizations. Tis is known as the linguistic relativity hypothesis or Sapir-Whorf hypothesis (from the names of its proponents, Benjamin Lee Whorf and his professor Edward Sapir).Whorf formulates this idea as follows: Are our own concepts of “time,” “space,” and “matter” given in substantially the same form by experience to all men, or are they in part conditioned by the structure of particular languages? (1941/1956, p. 138) Te linguistic relativity hypothesis has been extremely controversial in the scientifc literature, and many researchers consider it frivolous (to say the least). Yet, thanks to the most recent techniques in neurosciences, it is possible to show that the idea is not that crazy and that it is actually to some extent true. Here is what the research shows. Let’s take the example of color.You probably already know that color is something that we perceive diferently. People who are color blind have difculties perceiving some colors like red, green, and blue (check the color blind simulator here www. color-blindness.com/coblis-color-blindness-simulator/). So color is not just something that exists out there and that we all access in the same way. When we look at color, we rely on our vision and analyze the information that we receive in the brain. Te lef hemisphere of our brain is generally where most language processing occurs. Our lef hemisphere is connected to our right visual feld because human brains are contralateralized, meaning that each hemisphere controls the opposite side of the

Conceptual transfer

23

body. Researchers found that language categories for color, based on color names, primarily afect our right visual feld! Regier and Kay (2009) ofer a great summary of these fndings in their review paper with the title “Whorf Was Half Right.” Researchers also found that young children start with color categorizations located in the right hemisphere of the brain, where language is not dominant.As they learn to speak, these right hemisphere categorizations are erased and replaced by the new lef-hemisphere categorizations largely based on color names! You may think,“So much for universalism, we literally see the world diferently depending on the language we use!” Not so fast! Science is always more complicated than you’d think. Te most viable approach is a weak version of linguistic relativity: Language does not determine everything we think or feel, but it can serve there to enhance discrimination and memorization of colors. Obviously, if a language does not have a name for a color, it doesn’t mean that someone is not capable of perceiving the color at all. Tis would be linguistic determinism, the idea that you are limited in how you think about the world by the way your language categorizes things.

Tell me how many color names your language has, and I’ll tell you how you live Did you know that some languages may have as little as two color names? What would these two color names be? Berlin and Kay (1969) hypothesized that these would always be black and white, in other words, dark and light. If a language had a third color name, this would be red. Why? Tink of blood and how important it is in the lives of humans. Te color names that would follow would be green and/or yellow, then blue, followed by brown, and fnally by color names for purple, pink, orange, and gray. Te discovery that some languages have few terms may lead us to think of these languages as “poor” or “archaic.” Yet, the range of color names depends on what you need them for. English has a wide range of color names to the extent that you may not even know them all. Indeed, how many color names you master depends on whether you need to work with colors in your daily life. If you are restoring Renaissance paintings, you are likely to have a wider range of color names than most of us do. Te same goes for entire languages. If you do not have access to human-made objects like colored textiles, paints, or dyes, what would be the main color names you may need to talk about? Not that many. In fact, color names are not needed as ofen as we’d think because colors of natural objects are largely predictable. If everyone around you has black hair, you may not need a variety of color names for hair color. Or, if you work with plants, you may just need to know the names of the plants rather than their colors. Te idea that we have the number of colors we need to communicate well with others is known as the communicative efciency hypothesis (see Gibson et al., 2017).

24

From milliseconds to minutes

Now, let’s consider space. Trough the special and general theories of relativity,Albert Einstein demonstrated that the way we experience and think of space and time greatly difers from the reality of spacetime.Whorf proposes another relativity principle: We are thus introduced to a new principle of relativity, which holds that all observers are not led by the same physical evidence to the same picture of the universe, unless their linguistic backgrounds are similar, or can in some way be calibrated. (1940/1956, p. 214) Whorf suggests that even though “the apprehension of space is given in substantially the same form by experience irrespective of language . . . the concept of space will vary somewhat with language” (1941/1956, p. 158). Since the original proposal, a number of cross-cultural and cross-linguistic investigations have ofered evidence to support the view that language categories play a key role in memorizing spatial arrangements (Levinson, 2003). In order to describe the location of two objects, including small ones like a glass of water and a plate, languages may draw on one or various types of descriptions: • an “egocentric” description where the speaker’s viewpoint locates one object to the lef or right of another object, like the glass is to the right of the plate; • an “intrinsic” description where an object is located with respect to another object’s properties (e.g., back, front, or side), like the glass is to the plate’s front; • a “geocentric” description where an object is located with respect to some coordinates such as cardinal points or environmental features such as the trajectory of the sun, tides, wind directions, or landmarks such as mountains and rivers. Tis would give something like the glass is to the north of the plate. Stephen Levinson argues that these linguistic preferences will also determine the preferred memorization strategy in a spatial array of small objects. To test this hypothesis, Levinson created a very simple memory task, where participants need to memorize how three small animal toys are positioned on a table, then reposition the animal toys a little further away afer a 180-degree rotation. If your language favors egocentric descriptions like “right” and “lef,” chances are you will memorize the positions of the three animal toys with respect to yourself. Te cow that was to your lef hand will still be to your lef even afer the rotation. But, if your language favors geocentric descriptions like “north” and “south,” chances are you will memorize the three objects in absolute terms. Te cow that was north of the pig will remain north of the pig even afer the rotation. Tink about it: It makes perfect sense, and it is even more accurate in objective terms!

2.2 WHAT HAPPENS WHEN WE USE TWO LANGUAGES THAT DIFFER IN WHAT THEY CODE LINGUISTICALLY? We have seen that conceptualizations difer depending on the language we use. So, what happens when someone uses two languages? Te answer, as always, is: It depends. When children learn two languages at the same time, they form concepts

Conceptual transfer

25

while learning new words and grammar.At present, we do not know exactly how they organize concepts in the two languages when these are very diferent, but fndings suggest they are capable of creating distinct conceptualizations. Yet it appears that concepts from one language infuence our way of perceiving the world even when we use another language. Let’s say that you use a language that has grammatical gender such as feminine and masculine; this would be the case for French, where “spoon” takes feminine gender (la cuillère) and“knife” takes masculine gender (le couteau) (keep in mind that grammatical gender is arbitrary, as opposed to conceptual gender, which follows closely on societal biases and general experience: for instance, hammer would be conceptually male and necklace conceptually female for many of us). Tese grammatical gender categories will then be recruited automatically in the speaker’s mind, as measured in several studies using the priming paradigm (see Chapter 1 for more details on priming). For example, speakers of French may be faster to associate a female face with an object with feminine grammatical gender independently of whether the object is female in conceptual gender. What is surprising is that this gender efect is also measured for French-English bilinguals who perform the task in English, a language that does not have grammatical gender so that no priming efect should be apparent! For more details, you can read the study conducted by Sato and Athanasopoulos (2018). Te participants in their study were French speakers with very good levels of profciency in English. Most had learned English right before adolescence. As the study on grammatical gender illustrates, when we learn a second language as adults or young adults, we do not need to form entirely new concepts, as children do, but can draw on concepts available in our frst language.When the two languages have similar conceptual representations, then it might even be helpful to draw on the conceptualizations of the frst language to speed up the learning process (MacWhinney, 2018). But when two languages have diferent conceptual representations and learners rely on their frst language’s conceptualizations, trouble begins: Tis is what we refer to as “conceptual transfer.” Clearly, conceptual transfer is taking place for those French-English bilinguals, but there is no way they can use grammatical gender in English, so we can’t really tell they are experiencing conceptual transfer from French when they speak English. In some cases, however, conceptual transfer will be associated with linguistic evidence that can help us spot the mechanism in place. For example, if you learn Spanish, you will quickly fnd out that you need to learn how to use the two Spanish “to be” verbs, ser and estar. Tis is difcult if you are a speaker of English, as for you, this is all the same: Tere is no “to be (ser) or to be (estar)” but plain “to be.” So, what most of us would do would be to choose one Spanish “to be” verb, let’s say ser, and use it all the time. Of course, this would be perceived as a mistake by an L1 speaker of Spanish, although it may be amusing for your interlocutors as your strategy might lead you to mistakes such as saying “I’m smart” (soy listo) instead of the intended “I’m ready” (estoy listo). Eventually, with time, practice, and some theory, you may be able to learn how to use these two verbs better, as your language teacher may correct you, or your friends may tease you (see “smart” versus “ready” previously). Indeed, as profciency increases, learners may start to “decouple,” that is, to access the second language directly without relying too much on the frst language.

26

From milliseconds to minutes

But what happens if everyone around you uses the same two languages, one having two “to be” verbs and one having a single “to be” verb? And there is no language teacher around to correct you, either? And you don’t really care because everyone around you talks the same? Tis is what we will fnd out in the next section, with an example from Romani (see Chapter 1 for a presentation of this language).

2.3 CASE STUDY OF CONCEPTUAL TRANSFER FOR THE CONCEPT “TO BE” Romani, as spoken in Europe, has a single “to be” verb, like English does: In third person singular, this is si “to be.” Interestingly, Mexican Roma, who speak both Romani and Spanish, use two “to be” constructions in Romani, like Spanish does: In third person singular, they use si “to be” and a series of l- (former-) pronouns that infect for gender and number as lo, la, and le (these are no longer in use as pronouns and can now be described as “non-verbal copulas”). In Adamou (2013), it was noted that the two Romani “to be” constructions seem to follow closely on the distinction of the Spanish “to be” verbs. See an example of these uses in (2.1a) and (2.1b): (2.1) Mexican Romani (Indic, Indo-European) a. le ʃave muᴚa bibiake si barbale def.pl children poss.1sg aunt.dat be.3pl rich “My auntʼs children are rich.” (Adamou, 2013, p. 1085) b. o raklo=lo felis def.m boy-3sg.m happy “Te boy is happy.” (Adamou, 2013, p. 1075)

Roma in the Americas Roma in the Americas are estimated at between 1.5 and 3.5 million people. Tey live in practically every country of the American continent: from Chile and Argentina in the south, to Mexico, the U.S., and Canada in the north. Te presence of Roma in the Americas is documented as early as the colonial period, from the ffeenth to the eighteenth centuries. Arrivals of newcomers resumed during the nineteenth century following the changes in the AustrianHungarian geopolitical space in Europe.World Wars I and II are other key historical moments that prompted new arrivals of Romani people in the Americas. Since the fall of the Berlin Wall and up to today, Romani mobility from Eastern European countries to the Americas has been continuous. Te itinerary of Romani people from Europe to the Americas varies greatly. Tey leave from harbors in England, France, or other European

Conceptual transfer

27

countries. Some families arrive in Latin American countries such as Argentina and Brazil, more rarely Peru and Chile; other families travel to the U.S. and Canada. Depending on historical circumstances, newcomers settle in specifc countries, where they obtain their ofcial documents, or move across the continent in search of better working opportunities. Roma in the Americas speak diferent Romani varieties, which are considered an intricate part of their identity. In fact, Frank Mitchell, a Rom with origins from Russia, declares in a sociolinguistic interview from 1964: “(if) they don’t speak the language, they’re not Gypsies” (California Language Archive; accessed at https://cla.berkeley.edu/list.php). Yet the professional activities and mobility of the Romani families push Roma to learn various languages depending on their communicative needs. A shif to these languages is followed by a shif in perceptions of who is a Rom to include those who no longer speak that language.

Analysis of responses to a copula choice task conducted among 60 RomaniSpanish bilinguals from the community of Veracruz in Mexico reveals a combination of factors in Romani that are similar to those described in Spanish (Adamou et al., 2019). It appears that when the speaker has an immediate experience with a referent, the use of the l- “to be” is preferred (as in Jesse is tall or Jesse has grown tall). In addition, the l- “to be” is preferred for objects whose attributes are compared to the attributes of other similar objects (as in the house is big compared to other houses) and for people whose attributes are compared to themselves at some earlier stage (as in Jesse has grown tall). Moreover, this study found that Mexican Roma generalize the l- “to be,” similar to the general trend of extending estar among Spanish Latin American speakers to contexts previously occupied by ser. In terms of chronology, the innovation did not take place at the level of the frst generation of migrants, who were late learners of Spanish and were likely to rely on their frst language-Romani representations. Tis is supported by the absence of the innovative uses of l- “to be” in Romani texts from Mexico published in the early 1960s from speakers who were mostly born in the beginning of the twentieth century. In the study in Veracruz, however, generational diferences in the use of l- “to be” were not apparent in the statistical analyses, with respondents ranging in age from 17 to 90. Tis suggests that the innovative uses of the l- pronouns took place for the generation of speakers born afer 1930. Tese were second- or third-generation Mexican Roma who had learnt both Spanish and Romani early in life. It is therefore possible to argue that convergence at the level of the conceptualizations of being took place among these early and highly immersed bilinguals, prompting the use of Romani linguistic material to express the subdomain of Spanish estar.

28

From milliseconds to minutes

2.4 CASE STUDY OF CONCEPTUAL TRANSFER FOR SPATIAL REPRESENTATIONS As noted in section 2.1 of this chapter, various studies show that language habits predict the preferred memorization of spatial relations. But what happens when speakers of a language with dominant geocentric conceptualizations use a language with dominant egocentric conceptualizations? Tis is what the paper “How to Speak ‘Geocentric’ in an ‘Egocentric’ Language” explores by focusing on a small Indigenous community of Mexico that is shifing from the ancestral language, Ngigua, to Spanish (Calderón et al., 2019). Eréndira Calderón conducted the study as part of her doctoral dissertation (Calderón, 2022). She visited San Pedro Buenavista, a small rural community with less than 200 inhabitants, of which 32 declared speaking Ngigua. We refer to them as the “Ngiguas.” In San Pedro Buenavista, people practice subsistence farming and home gardening, exploit forest resources, and rely on small-scale cattle breeding as an insurance policy in times of need. Te community has very strong social cohesion and is governed by a number of Indigenous customary law and legal practices, known as usos y costumbres “traditions and customs.” It is a relatively isolated community, as telephone landlines and mobile phone access are very limited in the village, restricting communication with outsiders. In addition, mobility to nearby cities is difcult, as few people in the community own a means of transport and need to hire a taxi, which is expensive.

A few words about the Ngigua language Ngigua is a language spoken in Mexico that belongs to the Popolocan branch of the Otomanguean stock. Ngigua is a self-denomination, meaning “our language.” In 2015, roughly half of the Ngiguas were speakers of either Ngigua or of the closely related variety Ngiba (more than 700 speakers for 1,500 people). According to these statistics, all the Ngiguas are bilingual with Spanish. Speakers are typically above 50 years of age. Half of the Ngiguas reside in the State of Oaxaca, the other half having moved to Mexico City, the city of Puebla, or other areas in the country. Information about Ngigua culture and history comes from an important corpus of written documents, relying on logographic-pictorial scripts (AD 950–1600). From the study of these documents, we learn that the Ngigua society comprised the royal couple and the nobility, who controlled the positions of power, the land, the natural resources, the production and distribution of goods and services, and the religious institutions, and received tribute levy by the people. Te other major groups were priests, administrators, artisans, tradesmen, and farmers. Te community was united by family ties, land, and

Conceptual transfer

29

religion. In the years before the arrival of the Spaniards, the city of Coixtlahuaca and the valley came under the control of the Aztecs. Te arrival of the Spanish conquistadores signifcantly diminished the size of the Ngigua population due to disease and slavery, as in most of the Americas. More specifcally, it is estimated that the population in the area where Ngiguas lived was drastically reduced from 700,000 people in 1520 to less than 25,000 a hundred years later. In the early sixteenth century, when the newly acquired territories became part of New Spain, Spanish became the language of relations with the new administration (in 1535). Ngigua/Ngiba was among the rare Indigenous languages to be extensively written from the sixteenth to the nineteenth century. Despite this status, the process of language shif from Ngigua/Ngiba to Spanish was accelerated through Spanish monolingual educational policies starting at the end of the nineteenth century and intensifying in the early twentieth century. Ngigua/Ngiba is one of the 364 ofcially recognized Indigenous languages of Mexico since 2003. Revitalization eforts include classes in Ngigua/Ngiba kindergartens and primary school and language classes organized by local associations such as the Regional Committee for the Revitalization of the Language Ngiba/Ngigua.

In order to investigate the full range of linguistic conceptualizations of the NgiguaSpanish bilinguals, Eréndira Calderón examined not only speech but also gestures that accompany speech (co-speech gestures). Co-speech gestures are actually very revealing about the way we represent the real world in our mind. Te task involved describing the location of two large entities with respect to one another, like two buildings. Tese buildings were either located in the community or in a nearby city. Participants were asked, for example: “Where is the primary school in relation to the clinic?” Seventeen Ngigua-Spanish bilinguals and 17 Spanish monolinguals from the community of San Pedro Buenavista participated in the study. Tey were flmed during the interview, and their responses were coded for language and gestures that were used to accompany their descriptions. Researchers predicted that participants could opt for a description of the location of the two entities from an imagined position and describe the position of the two entities from the observer’s viewpoint (“egocentric”), on orientations intrinsic to the layout (“intrinsic”), or the position of the entities with respect to some environmental feature or landmark (“geocentric”). In the case where buildings were located in the community, participants could also describe them from their body position (through pointing). If proficient bilingual speakers are able to maintain distinct linguistic conceptualizations in their two languages, then the last bilingual Ngiguas should

30

From milliseconds to minutes

use egocentric linguistic means (as left and right) and accompanying gestures in Spanish and geocentric linguistic means (as north and south) in Ngigua with a combination of deictic and geocentric gestures. If linguistic conceptualizations are changed under the influence of the dominant language, in this case Spanish, then egocentric linguistic means, associated with Spanish, and egocentric co-speech gestures should be preferred in both Spanish and Ngigua. If, however, linguistic conceptualizations are dependent on non-linguistic factors such as the interaction of people with the natural environment, then geocentric conceptualizations should be dominant in the two languages of the Ngiguas, as they reside in a small, rural community and engage daily with activities related to their natural environment. An analysis of the responses reveals that Ngigua-Spanish bilinguals speaking in Spanish and Spanish monolinguals from the community both produced a majority of geocentric co-speech gestures. Surprisingly, Spanish monolinguals also used cardinal terms, such as poniente “west,” in combination with geocentric gestures, something that is rare in most documented varieties of Spanish. In comparison, egocentric gestures were used by all the participants, although to a lesser extent: Egocentric gestures were more frequently used by the bilinguals when speaking Spanish and the monolingual Spanish speakers but also the bilinguals when speaking Ngigua. Egocentric gestures were used in combination with egocentric terms such as “right” and “left” more frequently by monolingual speakers, followed by bilinguals speaking Spanish and by bilinguals speaking Ngigua. Figure 2.1 shows what geocentric co-speech gestures look like. In this example, this Ngigua speaker says “Hmm, the big church, so it is farther above in the center, so the municipality is below.” She is facing north and frst places the church by gesturing behind her (to the south) with her palm open, followed by a gesture in the space in front of her to locate the municipality (to the north). Te gestures correspond to the exact locations in the neighboring city of Coixtlahuaca, so these are clearly not pointing gestures. In addition, geocentric gestures, using the palm, difer from direct pointing, which is more frequently expressed through fnger pointing. Figure 2.2 shows a monolingual Spanish speaker (facing east). He responds just by locating building B with respect to building A without mentioning them explicitly: “On the same street, more to the west,” while making a gesture toward the back of his body to the west in agreement with the Spanish term poniente “west.” He then adds “more downwards” using a topographical term, abajo “below/down,” to render the inclination of the land. Te Ngigua data document two-way conceptual transfer, as bilinguals combine egocentric conceptualizations that stem from Spanish with geocentric conceptualizations that stem from Ngigua. Surprisingly, the Spanish monolinguals relied

Conceptual transfer

31

Figure 2.1 Bilingual Ngigua-Spanish speaker facing north using geocentric gesture and topological expressions while using Ngigua

Source: See Calderón et al. (2019, p. 32) and videos online at 10.17632/bbzjg97smz.1

extensively on geocentric co-speech gestures and sometimes associated these gestures with the use of Spanish cardinal terms. Tis way, the younger Ngiguas found a way to strengthen the geocentric cognitive representations that were most likely present in the community and coded in Ngigua and that are still relevant to their everyday communicative needs, their mobility, and activities such as farming and herding.

32

From milliseconds to minutes

Figure 2.2 Monolingual Spanish speaker facing east using geocentric gestures and cardinal and topographical terms

Source: See Calderón et al. (2019, p. 36) and videos online at 10.17632/bbzjg97smz.1

Chapter summary In this chapter, we have seen that conceptualizations of the real world are to a large extent shaped by language. We have also seen that when an adult is learning a language that has diferent conceptualizations than the ones found in their frst language, at frst, they are likely to rely on these conceptualizations.With practice, they will create distinct conceptualizations and will draw on them without accessing the conceptualizations of their frst language. However, when many language users are bilingual and live in close-knit communities, they can combine the conceptualizations of their two languages more freely to accommodate their communicative needs. Tis may lead to contactinduced change.

EXERCISES 1.

Explore

Visit the World Color Survey to explore the data for color names and corresponding colors in more than a hundred diferent languages: www.icsi.berkeley.edu/wcs/.

Conceptual transfer

2.

33

Discover

Stephen Levinson and colleagues designed a task known as“Animals in a Row” where participants had to memorize the arrangement of three animal fgures and reposition them afer a rotation of 180 degrees. Following are the pictures from the responses that some pupils living in an Indigenous community of Mexico provided when performing the task. Tey no longer speak Ixcatec, the Otomanguean language of their community, and use Spanish in their daily lives. 2a.

Observe the various arrangements and decide whether the strategy used by the respondent is egocentric or geocentric. If it is difcult to decide, you can turn the book around to look at the second picture (afer the rotation).

Figure 2.3 Arrangement 1 before the rotation

Figure 2.4 Arrangement 1 afer the rotation

34

From milliseconds to minutes

Figure 2.5 Arrangement 2 before the rotation

Figure 2.6 Arrangement 2 afer the rotation

2b.

Discuss in class: If this were a task at school, do you think the teacher would consider these responses correct? Why? What could help the teachers understand how their students think of the world?

3. Test yourself Make a list of three pairs of buildings that everyone in the class is familiar with. Tese buildings can be in your city and in another city that everyone else is familiar with. Ten tell your classmates how building A is located with respect to building B. Take

Conceptual transfer

35

turns. If you can flm the descriptions, that’s great, as you can then observe what you did with your hands.

4.

3a.

What words did you use? What was the perspective you adopted?

3b.

Did you make any gestures? What kind of gestures?

Research essay

Over the years, the linguistic relativity hypothesis has been challenged, rejected, and updated. Do humans view the world diferently depending on the language they use?

FURTHER READING Jarvis, S., & Pavlenko, A. (2008). Cross-linguistic infuence in language and cognition. Routledge. If you want to learn more about conceptual transfer, you can read this excellent book that will walk you through the various ways that bilingualism shapes the mind.

REFERENCES Adamou, E. (2013). Replicating Spanish estar in Mexican Romani. Linguistics, 51, 1075–1105. https:// doi.org/10.1515/ling-2013-0045 Adamou, E., De Pascale, S., García-Márkina, Y., & Padure, C. (2019). Do bilinguals generalize estar more than monolinguals and what is the role of conceptual transfer? International Journal of Bilingualism, 23, 1549–1580. https://doi.org/10.1177/1367006918812175 Avarguès-Weber, A., Dyer, A. G., Combe, M., & Giurfa, M. (2012). Simultaneous mastering of two abstract concepts by the miniature brain of bees. Proceedings of the National Academy of Sciences of the United States of America, 109, 7481–7486. https://doi.org/10.1073/pnas.1202576109 Berlin, B., & Kay, P. (1969). Basic color terms: Teir universality and evolution. University of California Press. Calderón, E. (2022). Bilinguisme et cognition spatiale chez les Ngiguas (Oaxaca, Mexique) [PhD thesis, INALCO]. Calderón, E., De Pascale, S., & Adamou, E. (2019). How to speak “geocentric” in an “egocentric” language: A multimodal study among Ngigua-Spanish bilinguals and Spanish monolinguals in a rural community of Mexico. Language Sciences, 74, 24–46. https://doi.org/10.1016/j. langsci.2019.04.001 Gibson, E., Futrell, R., Jara-Ettinger, J., Mahowald, K., Bergen, L., Ratnasingam, S., Gibson, M., Piantadosi, S. T., & Conway, B. R. (2017). Color naming across languages refects color use. Proceedings of the National Academy of Sciences of the United States of America, 114(40), 10785–10790. https://doi. org/10.1073/pnas.1619666114 Levinson, S. C. (2003). Space in language and cognition. Cambridge University Press. https://doi. org/10.1017/cbo9780511613609 MacWhinney, B. (2018). A unifed model of frst and second language learning. In M. Hickmann, E. Veneziano, & H. Jisa (Eds.), Sources of variation in frst language acquisition: Languages, contexts, and learners (pp. 287–312). John Benjamins.

36

From milliseconds to minutes

Olkowicz, S., Kocourek, M., Lučan, R. K., Porteš, M., Tecumseh Fitch, W., Herculano-Houzel, S., & Němec, P. (2016). Birds have primate-like numbers of neurons in the forebrain. Proceedings of the National Academy of Sciences of the United States of America, 113, 7255–7260. https://doi. org/10.1073/pnas.1517131113 Regier, T., & Kay, P. (2009). Language, thought, and color: Whorf was half right. Trends in Cognitive Sciences, 13, 439–446. https://doi.org/10.1016/j.tics.2009.07.001 Sato, S., & Athanasopoulos, P. (2018). Grammatical gender afects gender perception: Evidence for the structural-feedback hypothesis. Cognition, 176, 220–231. https://doi.org/10.1016/j. cognition.2018.03.014 Slobin, D. I. (1987). Tinking for speaking. Annual Meeting of the Berkeley Linguistics Society, 13, 435–445. https://doi.org/10.3765/bls.v13i0.1826 Whorf, B. L. (1940). Science and linguistics. Technological Review, 42, 229–231, 247–248. Reprinted in J. B. Carroll (Ed.). (1956). Language, Tought, and Reality: Selected Writings of Benjamin Lee Whorf (pp. 207–219). MIT Press and John Wiley and Sons. Whorf, B. L. (1941). Te relation of habitual thought and behavior to language. In L. Spier (Ed.), Language, culture, and personality: Essays in memory of Edward Sapir (pp. 75–93). Sapir Memorial Publication Fund. Reprinted in J. B. Carroll (Ed.). (1956). Language, thought, and reality: Selected writings of Benjamin Lee Whorf (pp. 134–159). MIT Press and John Wiley and Sons.

3 Cognitive costs and cognitive load

CHAPTER PREVIEW In this chapter we introduce two distinct yet interrelated concepts in psycholinguistics: “cognitive costs” and “cognitive load.” We discuss, in particular, under which conditions bilinguals may be confronted with higher cognitive costs and cognitive load than monolinguals.We then explore some of the long-term consequences of these two mechanisms that can lead to language change.

Te idea that bilinguals have a higher “linguistic burden” has been suggested in linguistics for a long time, way before we had any of the sophisticated machines that we have today to measure this burden. Weinreich (1953, p. 8), for example, talks about the linguistic burden of bilinguals and suggests that one way of reducing it is through “interference,” that is, when some aspects of the two languages start becoming more similar in the bilingual’s mind. In the 1990s, Silva-Corvalán refers to the “cognitive load” of bilinguals and its consequences in the long run: Te general hypothesis investigated in this book is that in language-contact situations bilinguals develop strategies aimed at lightening the cognitive load of having to remember and use two diferent linguistic systems. (1994, p. 6) What more do we know today about the linguistic burden of bilinguals? In this chapter, we ofer an update on this issue based on the recent neurolinguistic and psycholinguistic literature on “cognitive costs” and “cognitive load.” If we admit that contact-induced change starts at the level of the bilingual speaker or signer, then it is important to understand what goes on in bilinguals’ minds.

3.1

COGNITIVE COSTS

When we rapidly switch tasks, a specifc part of our brain helps us break response habits formed during task A in order to successfully move on to task B and focus on the new task rules: Tis is the “executive control function” located in the front of DOI: 10.4324/9781003167952-4

38

From milliseconds to minutes

our brain (prefrontal cortex). Tis fexibility, however, comes with a small price that the brain pays and that is captured by the term “cognitive costs.” Tink of athletes who warm up running in one direction, and then their coach instructs them to start running in the opposite direction, and then back and forth, and so on. Tis constant change is likely to slow down running as opposed to keeping on running in a single direction. As far as cognitive costs are concerned, researchers consistently fnd that participants slow down when performing non-linguistic cognitive tasks such as the Simon, Stroop, and Flanker tasks that involve selection and inhibition of cues (see Box for more details).

The tasks that help scientists measure cognitive costs In the “Simon task,” participants have to press the blue button (on their left) when they see a blue circle on the screen and the red button (on their right) when they see a red circle on the screen. Sounds easy? Well, it isn’t that easy when you need to go fast and the red circle appears on the left side of the screen and the blue on the left side of the screen. Try this task and find out how often you will get tricked into pressing the button based on location and not on color! In the “Stroop task,” participants have to look at color words (blue, red, or green) but they have to name the color of the ink the words are printed in while ignoring the meaning of the word. Tat’s also trickier than it sounds! Another well-known task is the “Flanker task.” In this task, participants see fve letters on the screen but are asked to respond only based on the letter that is located in the center. Tat’s also a tricky task, as you have to focus on a single letter while ignoring all the others.

You may ask: “How does task switching relate to language and bilingualism?” Several researchers have drawn a parallel between cognitive costs associated with (non-linguistic) task switching and cognitive costs associated with language switching. For instance, neuroscientists found that the two mechanisms, task switching and language switching, are located in the same brain region (Abutalebi et al., 2012). Tis doesn’t mean that the two mechanisms are identical: Language switching relies on a specifc cognitive mechanism known as “language control.” Language control allows language users to choose specifc words and inhibit others according to the context; this can be the interactional context but can also depend on the rules set to perform an experiment in the lab. Is the process of lexical selection any diferent for bilinguals than for monolinguals? Some researchers argue that the search process for words in the mind is the same for bilinguals as it is for monolinguals. For example, you may need to select the word “voice” to say “the child’s voice” but instead you select “choice” under the infuence of the sound “ch” in the preceding

Cognitive costs and cognitive load

39

word “child.” In that case, you didn’t manage to select “voice” and inhibit the word “choice.” Te diference is that bilinguals have a potentially larger repertoire of words than monolinguals do, rendering the selection/inhibition task harder. Instead of having a repertoire of 10,000 words, a bilingual may have an extra 5,000 or 10,000 words in their mind. How does this diference in the size of the lexical repertoire afect the mental search process? Well, it’s a little bit like having to look for a book in a large library instead of a smaller one: the search may be slower for a specifc book when there are a lot of books among which to choose. Of course, you may not be slower in fnding your favorite books even in a big library; it’s like those bestsellers or librarian’s favorites that are showcased right at the entrance of a library or a bookstore.You fnd them in a minute. But what about those books that you don’t read very ofen and for which you need to go all the way to the corner shelf in the basement? You are likely to be slower in accessing those rare books in a larger library than in a smaller one. Similar to the search for less-accessed books in a physical library, bilinguals have fewer occasions to activate lexical items in both of their languages. According to the weaker links hypothesis, those words that we use more rarely also have weaker links between the form of the word (its sound shape) and its meaning (Gollan et al., 2008). As a result, researchers fnd that bilinguals are slower than monolinguals in picture-naming tasks and semantic verbal fuency tasks, that is, where they are asked to cite all the names related to, for example, body parts (see Bialystok et al., 2012 for an overview). Tis relative slowness in response times is sometimes referred to as the “bilingual disadvantage.” In terms of language behavior, bilingual speakers might then draw on the more frequent lexical items independently of whether they belong to one language: Tis language behavior is known as “code-switching” and “borrowing,” two notions addressed in Chapters 4 and 8, respectively. Language control is not just involved in lexical selection and inhibition but also allows switching from one set of grammatical rules to another. Language control allows you, for instance, to switch between Spanish and English in a conversation with your friends (see Chapter 4 for more details on “code-switching”). Tis bilingual behavior is not only observed in real life but has also been simulated using computational cognitive modeling without any code-switched input (Tsoukala et al., 2021). Tis means that code-switching is an expected outcome among bilinguals. But if code-switching is a natural way of speaking to your friends, is it still costly for your mind/brain? In the early studies, researchers found that switching from one language to another in the lab was associated with high cognitive costs whether participants were asked to perform a task in language comprehension or in language production (e.g., Costa & Santesteban, 2004). At the same time, some researchers noted that language switching costs can be reduced or even disappear for some speakers and some kinds of switching. For example, Yim and Bialystok (2012) found that language switching costs depend on code-switching frequency in the participant’s life: Frequent code-switchers tend to show smaller to no code-switch costs when compared to bilinguals who don’t codeswitch. Moreover, Adamou and Shen (2019) and Johns and colleagues (2019) found

40

From milliseconds to minutes

that code-switching costs closely align with the general code-switching habits of the members of a language community. Tis means that experience with code-switching is likely to reduce code-switching costs. But it’s not just about the act of codeswitching; it’s also about how predictable a specifc code-switch is for a given speaker. Beatty-Martínez and Dussias (2017), for example, found that frequent code-switchers have difculties processing rarely attested code-switches. Adamou and Shen (2019) also found that the size of these costs can even boil down to specifc lexical items, for example, whether they switch more frequently for the verb “to marry” than for the verb “to leave.” Hence, taking into consideration “ecological validity,” that is, what speakers do in their daily lives, is important to understand whether a speaker may exhibit cognitive costs in language switching. More evidence comes from sign languages, the full-fedged languages Deaf people use by relying on manual and non-manual (facial) signs. Tere is a variety of sign languages, ranging from American Sign Language, which is the sign language introduced in sign education across the world, to homesign languages, which are sign languages created by children who did not have access to any input. Researchers note that the hearing children of Deaf parents tend to use a sign language and a spoken language in parallel; these are referred to in the literature as “bimodal bilinguals.” Bimodal bilinguals do not explicitly aim at the use of the two languages. Rather, this co-articulation arises spontaneously and ofen becomes the default mode of communication. Tis indicates that both languages are active in the bilingual mind and an efort is needed in order to use only one of them. According to Emmorey et al. (2008), the fact that bimodal bilinguals prefer using both languages rather than suppressing one of them indicates that the inhibition mechanism is associated with high cognitive costs. Fun fact: researchers also found that bilinguals who switch frequently between two languages become overall better at switching from one non-linguistic cognitive task to another and are much better at inhibiting an inappropriate answer that pops up into the mind. Tis is a rather famous fnding in the literature on bilingualism and is known as the “bilingual advantage for executive control” (Bialystok et al., 2004). Recent research shows that it’s not just about being bilingual but more specifcally about having enough experience with code-switching. For example, Hofweber and colleagues (2016) compared the performance in non-verbal confict monitoring of two groups of German-English bilinguals: One group had the habit of engaging in frequent code-switching, and the other did not. Tey report that only the frequent codeswitchers exhibited an advantage in the non-linguistic task, not the bilinguals who were not accustomed to code-switching. Researchers nowadays agree that we need to move beyond a binary monolingual vs. bilingual distinction in order to understand how specifc aspects of bilingual experience can give rise to a bilingual advantage (Beatty-Martínez et al., 2020). To conclude on this topic, studies show that switching from one language to another language places an extra burden on the mind but that experience with code-switching ofers bilinguals good training that can translate into better efciency in some cognitive, non-linguistic tasks. Let’s now turn to discuss a closely related phenomenon, cognitive load.

Cognitive costs and cognitive load

3.2

41

COGNITIVE LOAD

When we accomplish a cognitively difcult task that draws on working memory, performance on another cognitive task is likely to decline. Tis suggests that we are experiencing an increase in “cognitive load.”

Stop and think: Maybe you had an experience in your math class where you had to learn something new and complicated while at the same time trying to fgure out the writing of the teacher on the board. Your efort to process the new information and read (let alone copy) what was written on the board means that there was most likely an increase in your cognitive load. In education, the solution is then to focus on one task and avoid any additional complications.

How is cognitive load different from cognitive costs? As we have seen in the previous section, cognitive costs are related to executive control, which is associated with abilities such as attentional control, response inhibition, and rule discovery, as well as working memory. In contrast, cognitive load is primarily associated with working memory, which is only one of the components of executive control. What is working memory? Working memory is a system that temporarily stores and processes information during cognitive tasks by managing the attention of our short-term memory: Its role is to update.Working memory is important in language processing, as it is responsible for language comprehension and production as well as for vocabulary acquisition and temporary storage of unfamiliar sound patterns in second language learning. In contrast, long-term memory is associated with lexical knowledge, phonology, and the formation of rules in syntax, as well as regular and irregular morphology. Tese diferent types of memory interact and compensate for one another. Luckily, there are at least two mechanisms that can help us modulate cognitive load: • good working memory capacity (which varies from one individual to another as well as for a single individual across the lifespan) (e.g., Huettig & Janse, 2016); • the transfer of a cognitive task to long-term memory afer some practice (e.g., Segalowitz & Hulstijn, 2009). Te cognitive load hypothesis in linguistics predicts that bilingualism should afect processing of information in working memory. Several studies that test cognitive load do not fnd any evidence to support this hypothesis (e.g., Ito et al., 2018; Contemori & Ivanova, 2021). Others fnd that working memory capacity modulates the results, as predicted by the theory (e.g., Chee et al., 2004; Hopp, 2014). Results from

42

From milliseconds to minutes

neuroimaging studies also reveal difculties in the processing of an L2 (for late L2 learners) depending on the degree of difculty of linguistic phenomena, once again in agreement with predictions. Hasegawa and colleagues (2002), for example, showed that processing negative L2 sentences (considered difcult: think of how sometimes you get mixed up with the negations in sentences) requires greater cognitive efort than in the L1. Te same efect was not found when processing afrmative clauses (considered easy). However, the previously mentioned studies all focus on late L2 learners. It can therefore be assumed that early bilinguals, who would have practiced the two languages long enough, would have access to long-term memory to a similar extent in both of their languages. In consequence, speakers should not experience any particular cognitive load efect, as processing should be more or less automatic thanks to long-term memory. Te cognitive load hypothesis is relevant to language contact, as it was linked to the “simplifcation hypothesis.” According to the simplifcation hypothesis, languages would tend to become simpler in language endangerment, among the last speakers of a language, and language shif, among speakers who abandon one language in favor of another language, as in migration. Silva-Corvalán (1994) makes specifc predictions about how simplifcation can change a language: • simplifcation of grammatical categories and lexical oppositions; • over-generalization of forms; • development of periphrastic constructions either to achieve paradigmatic regularity or to replace less semantically transparent bound morphemes; • transfer of forms from the dominant language. In a diferent theoretical frame, Sorace and Serratrice (2009) suggest that bilinguals may face difculties with access to knowledge, coordination of information, and allocation of resources. Tey propose that these difculties impact frst and foremost the coordination of syntactic phenomena with contextual discourse phenomena. For example, bilingual speakers would overgeneralize overt subject pronouns in Spanish if they also speak English (where pronouns are always used) and that they would do so in specifc contexts involving discourse factors (e.g., when speakers need to assess whether they are still talking about the same person). Simplifcation has also been extensively discussed for Creoles (learn more about Creoles in Chapter 10). However, there have been numerous critiques of this analysis. Among others, Mufwene (2001) stressed the importance of social and pragmatic factors to grasp Creole formation. Meakins and colleagues (2019) also tested the simplifcation hypothesis by examining the speech of Gurindji Kriol, a mixed language spoken in Australia (learn more about mixed languages in Chapter 10). Analysis of the Gurindji Kriol data reveals that simplifcation is not driving the selection of linguistic features; rather there is a bias toward the features found in the prevailing language (Kriol) as compared to those found in the traditional language (Gurindji).

Cognitive costs and cognitive load

43

Chapter summary In this chapter, we have seen that bilinguals have to deal with cognitive costs only when they have little experience with a specifc task and language. We have also seen that practicing code-switching improves performance in nonverbal task switching. At the same time, bilinguals have a much bigger lexical repertoire and therefore can be slower accessing words when compared to monolinguals. However, studies do not consistently fnd any cognitive load efects among L2 speakers. Indeed, such efects seem to be modulated by good working memory. Overall, all these cognitive pressures in a bilingual’s mind can have long-term efects and shape languages.

EXERCISES 1. Test yourself Try the Simon task, the Stroop task, and the Flanker task (and more) on www.psytoolkit.org/experiment-library/.

2.

Checkpoint What is the “bilingual advantage for executive control?” What is the diference between “cognitive costs” and “cognitive load?” Which mechanisms can help modulate cognitive load? What is the “simplifcation hypothesis?” What are the predictions it makes?

FURTHER READING Adamou, E. (2021). Te adaptive bilingual mind: Insights from endangered languages. Cambridge University Press. Tis book shows how bilinguals’ grammars and language mixing patterns adapt to cognitive constraints such as the need for simplifcation and how language processing concurrently adapts to their unique and complex bilingual experience. Te book is clearly written and achieves a double goal: It makes typologically rare data accessible to readers from the feld of bilingualism and cognitive sciences while making experimental methods accessible to readers from the feld of endangered languages and contact linguistics.

REFERENCES Abutalebi, J., Della Rosa, P. A., Green, D. W., Hernandez, M., Scifo, P., Keim, R., Cappa, S. F., & Costa, A. (2012). Bilingualism tunes the anterior cingulate cortex for confict monitoring. Cerebral Cortex, 22, 2076–2086. https://doi.org/10.1093/cercor/bhr287

44

From milliseconds to minutes

Adamou, E., & Shen, X. R. (2019). Tere are no language switching costs when code-switching is frequent. International Journal of Bilingualism, 23, 53–70. https://doi.org/10.1177/1367006917709094 Beatty-Martínez, A. L., & Dussias, P. E. (2017). Bilingual experience shapes language processing: Evidence from code-switching. Journal of Memory and Language, 95, 173–189. https://doi. org/10.1016/j.jml.2017.04.002 Beatty-Martínez, A. L., Navarro-Torres, C. A., Dussias, P. E., Bajo, M. T., Guzzardo Tamargo, R. E., & Kroll, J. F. (2020). Interactional context mediates the consequences of bilingualism for language and cognition. Journal of Experimental Psychology: Learning, Memory and Cognition, 46, 1022–1047. Bialystok, E., Craik, F. I. M., Klein, R., & Viswanathan, M. (2004). Bilingualism, aging, and cognitive control: Evidence from the Simon task. Psychology and Aging, 19, 290–303. https://doi. org/10.1037/0882-7974.19.2.290 Bialystok, E., Craik, F. I. M., & Luk, G. (2012). Bilingualism: Consequences for mind and brain. Trends in Cognitive Sciences, 16, 240–250. https://doi.org/10.1016/j.tics.2012.03.001 Chee, M.W. L., Soon, C. S., Lee, H. L., & Pallier, C. (2004). Lef insula activation: A marker for language attainment in bilinguals. Proceedings of the National Academy of Sciences, 101, 15265–15270. https://doi.org/10.1073/pnas.0403703101 Contemori, C., & Ivanova, I. (2021). Bilingual referential choice in cognitively demanding situations. Bilingualism: Language and Cognition, 24(1), 83–95. https://doi.org/10.1017/S1366728920000176 Costa, A., & Santesteban, M. (2004). Lexical access in bilingual speech production: Evidence from language switching in highly profcient bilinguals and L2 learners. Journal of Memory and Language, 50, 491–511. https://doi.org/10.1016/j.jml.2004.02.002 Emmorey, K., Borinstein, H., Tompson, R., & Gollan, T. H. (2008). Bimodal bilingualism. Bilingualism: Language and Cognition, 11(1), 43–61. Gollan, T. H., Montoya, R. I., Cera, C., & Sandoval, T. C. (2008). More use almost always means a smaller frequency efect: Aging, bilingualism, and the weaker links hypothesis. Journal of Memory and Language, 58, 787–814. https://doi.org/10.1016/j.jml.2007.07.001 Hasegawa, M., Carpenter, P. A., & Just, M. A. (2002). An fMRI study of bilingual sentence comprehension and workload. NeuroImage, 15, 647–660. https://doi.org/10.1006/nimg.2001.1001 Hofweber, J., Marinis, T., & Trefers-Daller, J. (2016). Efects of dense code-switching on executive control. Linguistic Approaches to Bilingualism, 6, 648–668. https://doi.org/10.1075/lab.15052.hof Hopp, H. (2014). Working memory efects in the L2 processing of ambiguous relative clauses. Language Acquisition, 21, 250–278. https://doi.org/10.1080/10489223.2014.892943 Huettig, F., & Janse, E. (2016). Individual diferences in working memory and processing speed predict anticipatory spoken language processing in the visual world. Language, Cognition and Neuroscience, 31, 80–93. https://doi.org/10.1080/23273798.2015.1047459 Ito,A., Corley, M., & Pickering, M. J. (2018).A cognitive load delays predictive eye movements similarly during L1 and L2 comprehension. Bilingualism: Language and Cognition, 21, 251–264. https://doi. org/10.1017/s1366728917000050 Johns, M. A., Valdés Krof, J. R., & Dussias, P. E. (2019). Mixing things up: How blocking and mixing afect the processing of codemixed sentences. International Journal of Bilingualism, 23, 584–611. https://doi.org/10.1177/1367006917752570 Meakins, F., Hua, X., Algy, C., & Bromham, L. (2019). Birth of a contact language did not favor simplifcation. Language, 95, 294–332. https://doi.org/10.1353/lan.2019.0032 Mufwene, S. S. (2001). Te ecology of language evolution. Cambridge University Press. Segalowitz, N., & Hulstijn, J. H. (2009).Automaticity in bilingualism and second language learning. In J. F. Kroll, & A. M. B. de Groot (Eds.), Handbook of bilingualism: Psycholinguistic approaches (pp. 371–388). Oxford University Press. Silva-Corvalán, C. (1994). Language contact and change: Spanish in Los Angeles. Clarendon Press.

Cognitive costs and cognitive load

45

Sorace,A., & Serratrice, L. (2009). Internal and external interfaces in bilingual language development: Beyond structural overlap. International Journal of Bilingualism, 13, 195–210. https://doi. org/10.1177/1367006909339810 Tsoukala, C., Broersma, M., van den Bosch, A., & Frank, S. L. (2021). Simulating code-switching using a neural network model of bilingual sentence production. Computational Brain and Behavior, 4, 87–100. Weinreich, U. (1953). Languages in contact. Mouton. Yim, O., & Bialystok, E. (2012). Degree of conversational code-switching enhances verbal task switching in Cantonese-English bilinguals. Bilingualism: Language and Cognition, 15, 873–883. https:// doi.org/10.1017/s1366728912000478

Part II From minutes to years What bilinguals do when they communicate with others

KEY QUESTIONS Why do bilinguals sometimes combine elements from their two languages when communicating with other bilinguals? Does our language behavior spread beyond our immediate interlocutor? Can we choose to adjust or not to the language behavior of others? Is it important if we learn two languages as an infant, as a child or as an adult? Is it possible to forget your frst language and what does this really mean? Does what we think about language matter?

4 Code-switching, repertoires, and translanguaging

CHAPTER PREVIEW “Code-switching” refers to the systematic alternation of languages within a conversation. In the past decade, however, many researchers in subdisciplines of linguistics and allied felds have criticized the notion of code-switching, highlighting the constant change and complexity of language practices. From such a perspective, language users are thought to possess a complex “repertoire” consisting of monolingual and bilingual word forms, constructions, and speech practices that become associated with specifc social activities that are also specifc to each individual language user.“Translanguaging” is a related concept that has emerged in the domains of applied linguistics and education in order to capture the dynamic nature of language practices.

4.1 ACQUIRING AND UTILIZING BILINGUAL AND MULTIDIALECTAL CODES In bilingual communities, language users may be observed to draw fully on their linguistic resources in interactions, variously moving between varieties in their rich language repertoires. But you may well wonder: What determines who speaks what to whom and why? More importantly, how is this knowledge acquired? A concise response to the latter question is that this knowledge is attained at a young age, as children interact with others in their language community. Tus, we begin this section by discussing language socialization, which focuses on how children are socialized through language and socialized to use language (Ochs & Schiefelin, 1984). We will illustrate language socialization by reference to the infuential ethnolinguistic work of Ana Celia Zentella, which documents life and language among working-class Puerto Rican families in the Spanish-speaking community of El Barrio in Harlem, New York. Zentella’s fndings, synthesized in Growing up Bilingual: Puerto Rican Children in New York (1997), register the multidialectal nature of bilingualism among community residents and foregrounds the children’s dexterity at deploying their linguistic codes separately and in tandem. DOI: 10.4324/9781003167952-6

50

From minutes to years

Puerto Ricans on the island and on the U.S. mainland Te island of Puerto Rico is a commonwealth territory of the United States; as such, Puerto Ricans are U.S. citizens, facilitating circular migration to and from the U.S. mainland. Spanish and English are recognized as co-ofcial in Puerto Rico, but Spanish is the primary language of business and education and the dominant language of the island. Following World War II, Puerto Ricans were attracted by work in the garment industry in New York City, where they established a vibrant presence. However, since the 1980s, Puerto Rican migrants to the U.S. have settled in Orlando, Florida. Notably, fndings disclosed in Lamboy (2012) point to signifcant contrasts between the New York and Florida communities in their sociodemographic make-up and also in their language ideologies.

Over a period of 18 months, Zentella conducted participant observations and recordings with 19 families and their 34 children on one street of tenement buildings known as el bloque “the block” in Spanish Harlem. Multiple patterns of language usage were evidenced among the families: In the vast majority, the parents spoke Spanish to each other, and at least one of them always spoke Spanish to their children. But in some families, the parents conversed mainly in English with their partners and with their children, and yet in others, the parents frequently mixed Spanish and English in addressing each other and their children. In addition, the social networks of the residents were dense and complex, with nine of the families related by blood or marriage, many adults jointly participating in work and leisure, and school-aged children enrolled in the same schools. In describing the everyday engagements among residents, Zentella remarks on the constant visiting and exchanging within the apartments, hallways, stairwells, and courtyards, such that all the children were exposed to the multiple languages and varieties of el bloque. Zentella’s careful analyses of language on el bloque uncovered a spectrum of linguistic codes that included standard, vernacular, and contact varieties of Spanish and English. Naturally, individual networks tended to interact in one variety more than in others. For example, the older adults primarily used Popular Puerto Rican Spanish, the younger males primarily used African American Vernacular English, and the children and teens primarily used Puerto Rican English. However, all of these codes were encountered by all members, and shifing among them was a commonplace practice. Te following examples (Zentella, 1997, pp. 33–34) attest to the availability of manifold language varieties within defned speech contexts: (4.1) Context: Doris and Blanca, both age 9, instruct another girl in the use of a microphone as they prepare to audio-record themselves. Noticeable in the exchange are phonological features that readily identify vernacular Puerto Rican Spanish and vernacular African American English. DORIS TO ISABEL:

Hol’ your head up.

Code-switching, repertoires, and translanguaging

51

BLANCA TO ISABEL: Cuando hablas tienes que hablar (‘when you speak you have to

DORIS TO ALL:

speak’), you know, regular. No vire[h] la cabeza pallá y eso (‘don’t turn your head that way and stuf ’). OK? Remember, don’ put your mou[f] in the—en el micrófono (‘on the microphone’). Blanca be actin’ big an’ baaad.

(4.2) Context: Nine-year-old Blanca is prompting the two-year old daughter of a recent immigrant to count in English; despite her threat in Spanish, she fails. BLANCA

Uno, dos, no (‘one, two, no’) —ONE, TWO THREE—no te doy dulce (‘I won’t give you any candy’). She knows how to say it but she don’t want to.

But how do language users select from among the linguistic codes available to them? Zentella’s study unveiled relevant variables: Code choice in this New York Puerto Rican enclave was correlated with the socio-cultural setting and the age, gender, status, and language profciency of interlocutors. For instance, children learned to adhere to the expectations surrounding the respective roles of Spanish and English, addressing older people in Spanish and teachers in English. When children defed the previous expectation, a reprimand could ensue: !Mira, no me hables así en inglés; soy tu madre y me debes respeto! “Look, don’t talk to me in English like that; I’m your mother and you owe me respect” (Urciuoli, 1991, p. 298). Children also learn to manipulate their codes for specifc communicative purposes. Zentella (1997) appeals to the foundational work of Gumperz (1982) in identifying several discursive functions that are served by the concurrence of codes in the interactions on el bloque. Several such functions are represented in these samples: shifing topics, marking quotations, reinforcing a message, changing footing, inserting parentheticals, and hedging or crutch-like switching (Zentella, 1997, pp. 94–97). (4.3) Sample functions of switching Topic shif: Quotation: Reinforcement: Footing: Apposition: Crutching:

Vamos a preguntarle (“let’s ask her”). It’s raining! El me dijo (“he told me”), “Call the police!” pero yo dije, “No voy a llamar la policia na” (“but I said, I’m NOT going to call the police”) Ella tiene (“She has”)—shut up! Lemme tell you! Mi nombre es Lourdes (“my name is Lourdes”). Now we’re going to my sister. She have a brother in the hospital, en el Bellevue (“in Bellevue”), and he was crazy. You shouldn’t take that out because you’re gonna stay mellá (“toothless”).

Nevertheless, Zentella and other scholars (e.g., Gardner-Chloros, 2009, with reference to European communities) caution that attempting to impute a precise intent to language switching in individual utterances is fraught with challenges, and a full

52

From minutes to years

accounting may not be possible, given the complexity of motivations that underlie communicative exchanges. Moreover, frequent switching may lead meaningful switches to lose their pragmatic force, and thus it may merely represent another way of speaking (see following discussion on language mixing). Tis is not to suggest, however, that switching is random or haphazard. In the next section, we review the regularities that can be discerned when bilinguals integrate the language varieties in their repertoires.

When the intermingling of codes is considered taboo Norms of code selection can vary widely across communities. Alexandra Aikhenvald observed the practices of the multilingual Tariana community in the Vaupés basin of Brazil. In this northwest Amazonian community, language serves as an emblem of ethnic identity, and the rules dictating language choice and mixing are quite rigid. As described by Aikhenvald (2003), mixing of Tariana with other community languages (Tucano, Baniwa, and Portuguese) is considered taboo. In fact, injunctions against switching are lifed in only two contexts—in direct quotations and in the expression of the speech of animals and evil spirits in narratives. In one story retelling captured by the author, the same character was quoted in Portuguese when behaving rudely (e.g.,“I won’t give you money”) but quoted in Tariana when acting more laudably (e.g.,“will pay you well”). Animals and evil spirits are even allowed to employ multiple languages within the same sentence. However, switching in other circumstances, albeit accidentally, is ridiculed. Even the blending of dialects of Tariana is also considered improper, as it indirectly signals a rejection of one’s language identity and community.

4.2 INTEGRATING THE LANGUAGE VARIETIES OF BILINGUAL REPERTOIRES As explained and exemplifed previously, the integration of language varieties in bilinguals’ repertoires may be facilitated (or hindered), contingent on assorted individual, interactional, and community considerations. But when codes are co-present within a single utterance, we can focus more narrowly on the specifc ways in which they are combined. Here, we examine lexical borrowing, the adoption of words from one language variety into another, as well as syntactic code-switching, typifed by the alternation and merging of constituents from multiple codes. In addition to detailing the structural patterns that characterize these language contact phenomena, we consider the external variables that account for variation across language users and communities. Of all bilingual and contact phenomena, lexical borrowing is most pervasive, and it is manifested in the speech of multilinguals and monolinguals alike. Lexical borrowings, also called “loanwords,” difer in form. Loanwords may be fully adapted into the recipient language; for instance, the Italian word fasco [ˈfa.sko] is borrowed into

Code-switching, repertoires, and translanguaging

53

English as [f. ˈas.koʊ], and its plural is [f. ˈas.koʊz] (cf., Italian plural faschi [ˈfa.ski]), obeying English-language phonotactic and infectional rules.Alternatively, borrowings may retain features of the source language, as with the French loans rendezvous and ennui, which are produced by English speakers with little knowledge of the lending language. In the speech of French-English bilinguals, of course, the status of singleword insertions such as rendezvous and ennui may be more difcult to ascertain—Are they borrowings into English or code-switches into French?

Incorporating verbs: borrowing or code-switching? Te preceding examples of lexical borrowings—fasco(s), rendezvous, and ennui—all belonged to the grammatical class of nouns. Tis is not accidental. In actual fact, nouns take pride of place in hierarchies of borrowability, followed by adjectives and verbs. Even verbs, when borrowed, ofen surface as nominals in the recipient language. Japanese, for example, borrows the English infnitive “to type” as taipu suru “to do typing,” and U.S. Portuguese borrows “to save” as fazer o save “to do a save.” Interestingly, U.S. Portuguese also allows verbs to be borrowed directly into the verbal category, with morphological assimilation, as with bordar “to live in a boarding house” and frizar “to freeze,” in which the English root is followed by the regular -ar morphological ending. Contact Spanish also makes available multiple strategies for incorporating English verbs—English root+-(e)ar (e.g., puchamos [pu.ˈʧa.mos] “we push” is found alongside an English bare verb (e.g., quiero push [ˈkjɛ.ɾo. ˈpʊʃ] “I want to push” and hacer “make/do” + English verb stem (e.g., hacíamos push [a.ˈsi.a.mos. ˈpʊʃ] “we pushed”). Tese forms difer in geographical dispersion: In Belize, on the eastern coast of Central America, there is a marked preference for the latter, and in the United States, the former prevail. Te forms difer in levels of adaptation: in puchamos, the English root push is phonologically and morphologically Spanish, but quiero push and hacíamos push could be argued to signal a code-switch to English. In a renowned quantitative study on lexical borrowing in French, Poplack et al. (1988) examined the frequency, dispersion, and adaptation of English-origin words in the Ottawa-Hull region of Canada. Te spoken-language corpus was collected via sociolinguistic interviews conducted with 120 speakers, balanced for age and gender, in francophone neighborhoods in Ottawa (in the French-majority region of Ontario) and in Hull (in the English-majority region of Quebec). Te researchers inspected nearly 20,000 tokens to determine their level of linguistic integration and social embedding. Most of the borrowed lexicon was identifed as “established loanwords” (e.g., fun). Tese English-origin words were widely used in the region and were phonologically and morphologically adapted into the local French. But the more profcient bilinguals in the study also produced idiosyncratic, momentary borrowings that are less broadly attested and were less aligned with the receiving-language phonology

54

From minutes to years

(though they could be morphologically adapted with respect to person and number); these were labeled “nonce words” (e.g., dépluger, which incorporates the English root “plug” and the French derivational prefx dé- and infectional ending (-er)). Interestingly, neighborhood was the principal determinant for the borrowing patterns in evidence: Ottawans employed a higher proportion of nonce borrowings and a lower proportion of established loans than did the residents of Hull. Age was also relevant: Te younger speakers who were profcient in English presented higher borrowing rates overall and used more nonce borrowings and fewer widespread loans than the other groups; in other words, they were the most innovative. Tis seminal work employs distributional and sociolinguistic analyses to achieve fndings that support the concept of nonce borrowing, a notion that Poplack has continued to confrm (cf., Poplack, 2012). Equally importantly, Poplack and colleagues maintain that nonce borrowing is distinct from code-switching. Specifcally, they note that in the borrowing dépluger, an English-origin verb is inserted into a French syntactic slot and adapted phonologically and morphologically, but in code-switching, there is no evidence of the French system once the speaker has switched—over dead bodies is decidedly English (Poplack et al., 1988, p. 52). (4.4) French-English samples from Ottawa-Hull corpus Borrowing Code-switching

Puis les parents ont jamais voulu qu’ils la dépluggent ([de.’plɔg]). “And the parents didn’t want them to unplug her.” II a dit que des fois là, quand il marchait là, il marchait over dead bodies. “He said that sometimes when he walked, he was walking over dead bodies.”

You might have noticed that we have discussed lexical borrowing. Estigarribia (2021) examines the phenomenon of Guarani grammatical borrowings in colloquial Paraguayan Spanish. In Guarani, questions are not marked by special prosody or constituent reordering, but by interrogative morphemes, the most common being -pa and -piko/pio, illustrated in (4.5a); similarly, evidentiality and emphasis are not marked by intonation but by attitudinal particles niko/nio/ngo, shown in (4.5b), where VERD marks veridical attitude mood (Estigarribia, 2021, pp. 670–671). (4.5) a. ¿Qué-pa es eso? what-Q is that “What is that?” b. No me hace-niko falta no me makes-VERD miss “I don’t miss him?”

It is notable that these morphemes are clitics or afxes, which are thought to be of relatively low borrowability. At the same time, these interrogative and emphatic particles are like discourse markers, which have high borrowability, because they are

Code-switching, repertoires, and translanguaging

55

commonly independent syntactically. What makes such Guarani borrowings even rarer still is that in the usual contact scenario, discourse markers from Spanish are borrowed into the Indigenous language (e.g., Quechua, Aymara, Nahuatl), the opposite of what is observed in Paraguay. Estigarribia argues that these occurrences can be explained by message conceptualization procedures (recall the discussion of conceptual transfer from Chapter 2).

Facts about Guarani and Jopara • Guarani is the only Indigenous language of the Americas that is widely spoken by non-Indigenous populations—over 7 million speakers. • Most Paraguayans communicate in a mixed code called Jopara (“mixed”), characterized as a mixed lect composed of a continuum of varieties that range from mostly Guarani discourse with Spanish borrowings/switches (“Guarañol”) to mostly Spanish discourse with Guarani borrowings/ switches (“Castení”). • Te Paraguayan flm Siete Cajas served as the main data source for Colloquial Paraguayan with Guarani mixing for the 2021 research discussed previously. Linguist Bruno Estigarribia (2021, p. 668) writes, “Te flm is of tremendous linguistic importance because most of it takes place in Jopara, exemplifying the complex relations that determine the choice of code between colloquial Paraguayan Spanish and Spoken Guarani.” Because it implicates the syntactic component of the linguistic system,code-switching is thought to be facilitated where the grammatical structures of the languages in contact coincide, an observation that is formulated as the Equivalence Constraint (Poplack, 1980). You might be asking yourself: Is code-switching possible between languages that are typologically distinct? Consider the case of switching between Japanese and English; Japanese objects precede their selecting verbs, whereas English objects follow. Yet, despite the lack of equivalence, switching is allowed. Te work of Chan (2009, 2015) elucidates the multiple strategies that bilinguals employ in reconciling the structural diferences.As shown in the following examples (Chan, 2009, pp. 89–92, citing Nishimura, 1997), the object (O) may either precede or follow the verb (V), as specifed by the language of the verb; the verb may be doubled, appearing on both sides of the object; or the confict may be resolved by dropping the object, if its referent is salient in the discourse. (4.6) Japanese-English code-switching OE-VJ Only small prizes moratta ne (“we got”) VE-OJ We never knew anna koto nanka (“such thing sarcasm”) VE-OEJ-VJ We bought about two pound gurai kattekika no (“about bought”) —V— She karita (“borrowed”)

56

From minutes to years

Languages may also difer in the ordering of elements in noun phrases. Such is the case with French and Arabic. In Arabic, demonstratives (and numbers) are followed by a determiner, for example, had l-medina “this the-city,” but French prohibits the combination, allowing only cette ville“this city”or la ville “the city.” Nevertheless,there are compromise strategies available to code-switching French-Arabic bilinguals: dakAr laFr chemiseFr “that the shirt;” cetteFr xubzaAr “this loaf.” Employing a web-based survey, Post (2010) investigated preferences for code-switching patterns among FrenchArabic bilingual speakers in Morocco and Tunisia. Specifcally, participants assessed the authenticity of such nominals inserted into sentences that mixed languages in both directions, such as L-bent lhih dart had les gateaux il y a deux jours “Te girl over there made these (the) cakes two days ago.” Te participants’ responses revealed signifcant diferences in acceptability, based on dialect and switch type. Tus, we observe again that the same language pairing can present diferent code-switching patterns in diferent communities. Lexical insertions and code-switching are also features of contact between two or more signed languages. Such is the case when signers of diferent signed language backgrounds interact; these might be diferent signed languages, as with Langue de Signes Québécoise and American Sign Language in Canada, or they might be diferent varieties of the same signed language (e.g., British Sign Language and Auslan, the majority sign language of the Australian Deaf community). Like spoken-language code-switching, the form and function of signed-language code-switching is predictable, and scholars have considered the similarity of code-switching in the two modalities. For instance, Nonaka (2014) studied unimodal gestural-visual interactions in the village of Ban Khor in Tailand, where signers switched between Ban Khor Sign Language and Tai Sign Language, for lexical need, as with the insertion of color terms, which difer across the two languages. And Quinto-Pozos (2009) documents the sequential use of synonyms from American Sign Language and Lengua de Señas Mexicana to mark emphasis or reiteration in exchanges between signers along the U.S.-Mexico border.

Sign languages and bilingualism Sign languages have recently been the focus of attention in studies of bilingualism.We can identify three main types of bilingualism involving a sign language (Zeshan & Panda, 2015): • “Sign bilingualism”: when bilinguals use in parallel a sign language and a spoken language in its written form. • “Bimodal bilingualism”: when hearing individuals simultaneously use a sign language and a spoken language. Tis is diferent from simultaneous communication as practiced by interpreters in formal settings, like meetings or lectures.

Code-switching, repertoires, and translanguaging

57

• “Sign multilingualism”: when bilinguals use more than one sign language at the same time. Studies on bimodal bilinguals allow us to observe the simultaneous articulation of two languages, a co-articulation which is impossible for spoken languages, as they both unfold in the same modality; the sequential use of a sign language and a spoken language, similar to codeswitching observed in spoken languages, is rare. Analysis of a variety of naturalistic corpora indicates that bimodal bilinguals do not continuously use the two languages jointly, that is, the spoken language and the sign language. Similar to the well-documented capacity of bilinguals to choose the appropriate code from an early age, studies of sign languages report that bimodal bilinguals are sensitive to the communicative context and to their interlocutors and can suppress one of the two languages when appropriate. Instead, constant parallel use of the sign and spoken language is attested in simultaneous communication among trained individuals. Since parallel use of a sign language and a spoken language is partial, the contexts in which such parallel use does occur are of great interest for understanding how the bilingual mind works. For example, İşsever et al. (2020, p. 184) show that the elements from the spoken language, Turkish, are not only partly mirrored in the sign language, Turkish Sign Language (TID), but that there is a common verb phrase to which the two languages contribute lexically. Tis can be seen at the level of the verb phrase, where “watch” is expressed in Turkish and “movie” in the Turkish Sign Language (TID), as illustrated in example (4.7). (4.7) Turkish: TID

Ben de gid-ip I too go-CONJ GO MOVIE NOT

izle-me-di-m watch-NEG-PAST_1SG

Zeshan and Panda (2018), however, report counter-evidence from Indian Sign Language and Hindi bilinguals (with minimal English) showing 48% syntactic and/or semantic mismatches. What they fnd instead is that the two propositions share the same communicative intent. To account for the diferences in these bimodal bilingual interactions, the researchers refer to their data as “sign-speaking.” To conclude, although there is semantic congruence between the sign and the spoken component in some cases, it is not required. Te previous discussion on lexical borrowing and code-switching further confrms that bilinguals have options in integrating their multiple codes through varied strategies available to them, and selection from among these options can lead to diferent patterns across communities and speakers (Muysken, 2000, 2014); these divergent outcomes are discussed next. Auer (1999) formalizes a dynamic typology of bilingual speech in which “codeswitching” is diferentiated from “language mixing” and “fused lects” (see Chapter 10 for more on fused lects).Auer restricts the term “code-switching” to cases in which bilingual speakers perceive their juxtaposition of codes as pragmatic and intentional, regulated by contextual cues, as we observed throughout this chapter. In “language mixing,” the

58

From minutes to years

use of two languages is frequent, the size and patterns of mixed constituents are variable (to include phrases and smaller units), and the alternation has lost its contextually meaningful function—the mixing is anticipated and unmarked. Tis phenomenon is illustrated in the following example, in which speakers interlace Swiss German and Italian such that the language of interaction cannot be identifed as solely Italian or German but both. In other words, the bilingual mixing ceases to signal any discourse-related function and may instead resemble a single code in itself (pp. 314–315). (4.8)

Italian immigrants in Switzerland: Swiss-German dialect and Italian (italicized) SPEAKER A perché meinsch che se tu ti mangi emmentaler o se tu ti mangi una fontina iscch au en unterschied, oder? schlussändlich è sempre dentro lì però il gusto isch andersch “because you mean, if you eat Emmental cheese of if you eat Fontina cheese, there is also there is also a diference, isn’t there? actually, it’s still there, but the taste is diferent” SPEAKER B è vero! “that’s true”

In cases such as these, the mixed code may be associated with ethnic identity, refected in the folk names coined to label them.Auer notes that the term italoschwyz emerged in Zurich in reference to the mixed variety that was generalized among second-generation Italian immigrants in Switzerland (Franceschini, 1998). Tese varieties may elicit highly pointed responses (largely negative) from monolingual and bilingual laypersons and scholars alike. In this regard, the nation of Belize presents an opportune site for observing language mixing practices in the absence of reproach. Tere is widespread multilingualism—Spanish is the language of the majority, English is the ofcial language, and Belizean Kriol (an English-lexifed Creole) is the lingua franca. Notably, the interweaving of varieties is commonplace and endorsed as legitimate; it is linked with Belizean identity at the societal level. Te acceptance of language mixing is evident in the following exchange, shared by Osmer Balam, the Belizean interviewer: (4.9)

INTERVIEWER STUDENT

INTERVIEWER STUDENT

¿Qué piensas de Sixth Form? Del transition de high school a Sixth Form, ¿Cuáles son algunos diferences que notaste? Un big diference es que los teachers, like, van a hace teach y si tú no haces pay attention, that’s your problem porque ahora en high school ellos hacen make sure que tu hagas pass, allá no, si no hiciste pass, like, tienes que agarra el course over. Ellos solo van a hace teach, hace su work y um, if you won’t pay attention, eres tú. ¿Qué tú le harías recommend a un student que está entrando en Sixth Form de high school? Em, que like, debes de hace tu work right away porque like se te hace pile up tu work y los teachers deh noh really care si no tienes time. Debes de usa tu . . . your time wisely. If they teach solo como un topic, further research on your own.

Code-switching, repertoires, and translanguaging

59

When what appears to be code-switching or language mixing is not Within Auer’s typology, frequent code-switching may lead to language mixing. In turn, language mixing may give way to a fused lect and then a full-fedged mixed language, with structurally regular mixing sedimented into a composite system. An example of such a stabilized mix is Media Lengua, spoken in Ecuador (see Chapter 10 for more). In this system, Quechua provides the word order and infectional morphology, while Spanish contributes the lexical stems. Tus, Unu faburta pidingabu binxuni “I come to ask for a favor”) presents roots (un- “a,” favor- “favor,” ped- “ask for,” ven- “come”) phonologically adapted Spanish accompanied by Quechua case markers and infections. Here, what appears to be code-switching or language mixing is not; in fact, speakers of these mixed languages may no longer have access to the full grammar of the contributing languages. In contrast, there are strong empirical grounds to sustain that code-switching and language mixing refect a high degree of profciency in multiple codes.

4.3 EMBRACING THE MULTIDIMENSIONALITY OF BILINGUALS’ CODES In recent years, professionals in the felds of applied linguistics and bilingual education have advocated for attention to the multidimensional language resources that learners bring from their homes and communities into the classroom, or “critical language awareness” (Alim, 2010; Blackledge & Creese, 2014). (Recall the discussion of language socialization of Puerto Rican children in New York.) Adopting this perspective, scholars have further advocated for recognition and development of “translanguaging” practices and pedagogies (García, 2009; García & Li Wei, 2014). Translanguaging validates home and community language norms in institutional spaces, bolstering students’ positive self-image and thereby supporting their learning. Rather than prioritizing discrete languages, proponents of pedagogical translanguaging seek to embrace and leverage the extensive fullness of language users’ repertoires. Te potential of a translanguaging orientation has been explored across instructional contexts and learner groups. Jones (2017) reports on the implementation of translanguaging in a history class in a secondary school in Wales, and Cenoz and Santos (2020) discuss the translanguaging practices observed in a science course in the Basque Autonomous Community in Spain. Canagarajah (2011) describes the translanguaging strategies in the narrative writing development of a Saudi Arabian university student, and Durán (2015) documents the benefts of translanguaging in an elementary classroom. Te latter classroom ethnography specifcally considered how the everyday language practices of bilingual frst-graders might serve as a resource for their literacy development. In that study, the teachers’ lessons were redesigned to

60

From minutes to years

include tasks that oriented students’ writing towards readers of diferent language preferences. For instance, the children exchanged letters with English-dominant pen pals who were learning Spanish, wrote notes to their Spanish-dominant caregivers, and prepared bilingual poems to share with their bilingual peers. Importantly, in the United States, the translanguaging approach has allowed for recognition of the rule-governed nature and communicative efectiveness of language practices that defy abstracted standards but that creatively merge other varieties, such as African American Vernacular English (Robinson, 2020).

When mixing is normalized A prominent aspect of the Belizean society previously discussed is the recognition and appreciation of linguistic diversity and innovation; use of the Kriol vernacular and of language mixing is normalized, and it appears to be the use of the “standard” varieties of English and Spanish that is marked. Oral production and attitudinal data from the northern region studied by Balam and de Prada Pérez (2017) reveal that there is little linguistic prescriptivism among adolescents and educators. In particular, teachers’ attitudes, ideologies, and linguistic behaviors are shown to align with the mixing norms that are prevalent in the community: Teachers accept mixing as a communicative and pedagogical strategy in the classroom by allowing students to engage in it, participating in it themselves, and employing it to facilitate language instruction. One teacher described language mixing as “an impressive skill that bilinguals have.” In fact, Balam and de Prada Pérez suggest that in the Belizean context, the challenge for teachers is to destigmatize the “standard” varieties. Regrettably, Belize is a special case; more widespread in multilingual communities and societies worldwide is a situation in which educators promote linguistic purism that insists on separate, bounded languages or “double monolingualism” (Heller, 2006; Lippi-Green, 2012).

4.4

HOW TO STUDY CODE-SWITCHING

Te studies discussed throughout this chapter represent a diversity of methodologies for studying code-switching. Let’s review a few of these: • Ethnographic observations were employed in understanding the bilingual practices of the Puerto Rican community in New York (Zentella, 1997), the Tariana community in the Amazon (Aikhenvald, 2003), the gestural-visual interactions of a Deaf community in Tailand (Nonaka, 2014), and the pedagogical interventions in bilingual classrooms (Durán, 2015).

Code-switching, repertoires, and translanguaging

61

• Corpora of spoken, signed, and written language were employed to examine lexical and grammatical borrowing: Poplack et  al. (1988) investigated English lexical borrowing in Canadian French, Zeshan and Panda (2015) examined the use of Indian Sign Language (ISL) and Burundi Sign Language (BuSL) in casual conversations, Adamou (2016) worked on a variety of small spoken corpora of minority languages, and Estigarribia (2021) examined grammatical transfer in Guarani-Spanish bilingual in Paraguay. • Participant interviews were the principal source of data for studies of KriolSpanish bilingualism in Belize (Balam & de Prada Pérez, 2017). • Grammaticality or acceptability judgments are employed by Post (2010) to assess preferred patterns of French-Arabic code-switching. • Another method for characterizing multilingual speech is elicited repetition or imitation (Gullberg et al., 2009). Tis task requires participants to listen to a sentence that contains code-switching and repeat the sentence following a brief delay. Te thinking is that the participant will process the sentence through their grammar in producing the output. In such studies, bilinguals may alter their repetitions to be more consistent with their grammars. Tis method is especially useful since it does not require participants to be biliterate. • In addition to naturalistic data, this bilingual behavior has also been simulated using computational cognitive modeling without any code-switched input, suggesting that codeswitching is an expected outcome among bilinguals (Tsoukala et al., 2021).

Chapter summary In this chapter we explored the diverse ways in which bilingual speakers combine elements from their linguistic repertoires in what is broadly termed switching in the linguistic and sociolinguistic literature.We noted that bimodal bilinguals efortlessly use a sign and a spoken language simultaneously (rather than sequentially), something that is possible thanks to the diferent modalities. We began by deliberating on how speakers acquire and utilize bilingual and multidialectal codes to communicate, make meaning, and construct identities. We also investigated the specifc ways in which the language varieties are integrated, at various levels of structure, for example, through borrowing of individual lexical items or at higher phrasal and clausal boundaries. Finally, we considered translanguaging, which refers to the continuity of bilinguals’ multidimensional repertoires and to the pedagogies that support these.

62

From minutes to years

EXERCISES 1.

Discover

Code-mixing is increasingly popular in social media postings and in musical performances across genres. Find several posts or musical selections that incorporate multiple languages that you are familiar with. Inspect the text, performances, or lyrics for the diferent bilingual phenomena presented in this chapter, such as borrowing and code-switching. For inspiration, visit your social media platforms (e.g., Facebook, Twitter, Instagram, Mastodon); listen to your favorite spoken word poetry performance or your favorite multilingual musical artist; you can also visit https:// en.wikipedia.org/wiki/List_of_multilingual_bands_and_artists.

2. Test yourself Studies have shown that language mixing can infuence language change. Sayahi (2015) examined the expression of attributive possession in Tunisia, where ArabicFrench switching is commonplace. In conveying “Salma’s book,” Tunisian Arabic allows two forms: the “synthetic” form (kita:bu Salma) and the “analytic” form (l-kte:b mte:c Salma), although the synthetic is used more frequently. (Consider similar alternations in English: the book’s title vs. the title of the book.) Sayahi conducted sociolinguistic interviews with 12 speakers; the sample returned 472 tokens of the attributive possession construction of interest. Analyses showed signifcant correlations. What results would you expect when the possessed noun is Arabic? Would you expect the analytic or synthetic form? When the possessed noun is French?

3.

Essay

Read Nonaka’s (2014) account of language shif in a sign community in Tailand and consider how the language patterns link to those described in Auer’s typology.

FURTHER READING Bullock, B. E., & Toribio, A. J. (2009). Temes in the study of code-switching. Cambridge University Press. Te book comprises work on a variety of language pairings by leading researchers in the study of bilingual code-switching, with prominent researchers addressing the linguistic, social, and cognitive implications of code-switching.

REFERENCES Adamou, E. (2016). A corpus-driven approach to language contact: Endangered languages in a comparative perspective. De Gruyter Mouton. Aikhenvald,A.Y. (2003). Multilingualism and ethnic stereotypes: Te Tariana of northwest Amazonia. Language in Society, 32(1), 1–21.

Code-switching, repertoires, and translanguaging

63

Alim, H. S. (2010). Critical language awareness. Sociolinguistics and Language Education, 18, 205–231. Auer, P. (1999). From codeswitching via language mixing to fused lects: Toward a dynamic typology of bilingual speech. International Journal of Bilingualism, 3(4), 309–332. Balam, O., & de Prada Pérez, A. (2017). Attitudes toward Spanish and code-switching in Belize: Stigmatization and innovation in the Spanish classroom. Journal of Language, Identity and Education, 16(1), 17–31. Blackledge, A., & Creese, A. (2014). Heteroglossia as practice and pedagogy. In A. Blackledge, & A. Creese (Eds.), Heteroglossia as practice and pedagogy (pp. 1–20). Springer. Canagarajah, S. (2011). Codemeshing in academic writing: Identifying teachable strategies of translanguaging. Te Modern Language Journal, 95(3), 401–417. Cenoz, J., & Santos, A. (2020). Implementing pedagogical translanguaging in trilingual schools. System, 92, 102273. Chan, B. H. S. (2009). Code-switching between typologically distinct languages. Cambridge University Press. Chan, B. H. S. (2015). A diachronic-functional approach to explaining grammatical patterns in codeswitching: Postmodifcation in Cantonese-English noun phrases. International Journal of Bilingualism, 19(1), 17–39. Durán, L. G. (2015). Audience and the writing development of young bilingual children [PhD thesis, University of Texas at Austin]. Estigarribia, B. (2021).A speech planning account of Guarani grammatical borrowings in Paraguayan Spanish. Journal of Language Contact, 12, 663–696. Franceschini, R. (1998). Code-switching and the notion of code in linguistics: Proposals for a dual focus model. In P. Auer (Ed.), Code-switching in conversation: Language, interaction and identity (pp. 51–72). Routledge. García, O. (2009). Education, multilingualism and translanguaging in the 21st century. In T. SkutnabbKangas, R. Phillipson, A. K. Mohanty, & M. Panda (Eds.), Social justice through multilingual education (pp. 140–158). Multilingual Matters. García, O., & Li, Wei. (2014). Language, bilingualism and education. In O. García, & Li Wei, Translanguaging: Language, bilingualism and education (pp. 46–62). Palgrave Pivot. Gardner-Chloros, P. (2009). Code-switching. Cambridge University Press. Gullberg, M., Indefrey, P., & Muysken, P. (2009). Research techniques for the study of code-switching. In B. E. Bullock & A. J. Toribio (Eds.), Te Cambridge handbook of linguistic code-switching (pp. 21–39). Cambridge University Press. Gumperz, J. J. (1982). Discourse strategies. Cambridge University Press. Heller, M. (2006). Linguistic minorities and modernity: A sociolinguistic ethnography. Longman. İşsever, S., Makaroğlu, B., Ergenç, I., & Dikyuva, H. (2020).A minimalist perspective on code blending in TİD-Turkish bimodal bilingualism. In U. Zeshan & J. Webster (Eds.), Sign multilingualism (pp. 171–200). De Gruyter Mouton. https://doi.org/10.1515/9781501503528-005 Jones, B. (2017). Translanguaging in bilingual schools in Wales. Journal of Language, Identity and Education, 16(4), 199–215. Lamboy, E. M. (2012). Language, Puerto Ricanness, and the new wave of Puerto Rican immigrants. In A. Cortazar, & R. Orozco (Eds.), Lenguaje, arte y revoluciones ayer y hoy: New approaches to Hispanic linguistics, literary, and cultural studies (pp. 140–161). Cambridge Scholars Publishing. Lippi-Green, R. (2012). English with an accent: Language, ideology and discrimination in the United States. Routledge. Muysken, P. (2000). Bilingual speech: A typology of code-mixing. Cambridge University Press. Muysken, P. (2014). Déjà voodoo or new trails ahead. In R. Torres Cacoullos, N. Dion, & A. Lapierre (Eds.), Linguistic variation: Confronting fact and theory (pp. 243–262). Francis & Taylor.

64

From minutes to years

Nishimura, M. (1997). Japanese/English code-switching: Syntax and pragmatics. Peter Lang Publications. Nonaka, A. M. (2014). (Almost) everyone here spoke Ban Khor Sign Language—Until they started using TSL: Language shif and endangerment of a Tai village sign language. Language and Communication, 38, 54–72. Ochs, E., & Schiefelin, B. (1984). Language acquisition and socialization. Culture Teory: Essays on Mind, Self, and Emotion, 276–320. Poplack, S. (1980). Sometimes I’ll start a sentence in Spanish Y TERMINO EN ESPAÑOL: Toward a typology of code-switching. Linguistics, 18(7/8), 581–618. Poplack, S. (2012). What does the nonce borrowing hypothesis hypothesize? Bilingualism: Language and Cognition, 15(3), 644–648. Poplack, S., Sankof, D., & Miller, C. (1988). Te social correlates and linguistic processes of lexical borrowing and assimilation. Linguistics, 26, 47–104. Post, R. (2010). Code-switching in the determiner phrase: A comparison of Tunisian Arabic-French and Moroccan Arabic-French switching [PhD thesis, Te University of Texas]. Quinto-Pozos, D. (2009). Code-switching between sign languages. In B. E. Bullock, & A. J. Toribio (Eds.), Te Cambridge handbook of linguistic code-switching (pp. 221–238). Cambridge University Press. Robinson, H. (2020). Translanguaging, performance, and the art of negotiation. In H. Robinson, J. Hall, & N. Navarro (Eds.), Translingual identities and transnational realities in the US college classroom (pp. 119–141). Routledge. Sayahi, L. (2015). Expression of attributive possession in Tunisian Arabic: Te role of language contact. In A. Butts (Ed.), Semitic languages in contact (pp. 333–347). Brill. Tsoukala, C., Broersma, M., van den Bosch, A., & Frank, S. (2021). Simulating code-switching using a neural network model of bilingual sentence production. Computational Brain and Behavior, 4, 87–100. Urciuoli, B. (1991). Te political topography of Spanish and English: Te view from a New York Puerto Rican neighborhood. American Ethnologist, 18(2), 295–310. Zentella, A. C. (1997). Growing up bilingual: Puerto Rican children in New York. Wiley-Blackwell. Zeshan, U., & Panda, S. (2015). Two languages at hand: Code-switching in bilingual deaf signers. Sign Language and Linguistics, 18(1), 90–131. Zeshan, U., & Panda, S. (2018). Sign-speaking: Te structure of simultaneous bimodal utterances. Applied Linguistics Review, 9(1), 1–34. https://doi.org/10.1515/applirev-2016-1031

5 Social networks and accommodation

CHAPTER PREVIEW In Chapter 1, we explored the unconscious cognitive processes of aligning to a conversation partner and of priming within and across languages. In this chapter, we explore the social and communicative aspects of how and why individuals tend to align with their interlocutors in what is known as a process of accommodation. Unlike alignment, which is an unconscious cognitive process, communication accommodation theory views accommodation in language during conversations as partly a conscious process, refecting the language user’s intention to express social solidarity or social distance. And since language users and their interlocutors connect with one another and form social networks, we consider the role that these networks might play in promoting language variation and change within a bilingual community.

5.1 ACCOMMODATION AND CHANGE Let’s begin this chapter with an unlikely example of accommodation. A group of researchers analyzed many decades worth of recorded broadcasts of Queen Elizabeth II to examine changes in her pronunciation over time; these are called longitudinal changes. Tey found that the Queen had adjusted her pronunciation over the decades so that, in her later recordings, some of her vowels sounded more like the vowels produced by younger, less afuent speakers from southern Britain (Harrington et al., 2000; Harrington, 2006). Tese results are surprising because we don’t think of members of royal families as regularly mixing with young folks from backgrounds that are diferent from their own. So, how, and why, could Queen Elizabeth II realign her speech in this manner? Did she deliberately change her speech habits or were these changes unintentional and, if these adjustments were unconscious ones, how did they come about in the absence of frequent face-to-face contact? Te case study of the Queen’s vowels provides us with an example of linguistic accommodation. Te social psychological motivations for accommodation are outlined in a framework known as communication accommodation theory (Coupland & Giles, 1988; Giles, 1984). Te direction of change toward other Britons in the Queen’s DOI: 10.4324/9781003167952-7

66

From minutes to years

vowels provides us with an example of linguistic convergence. In convergence, you de-accentuate certain features of your speech so that you speak more like those of people you interact with. Within the accommodation framework, the function of convergence is to reduce the social distance between yourself and others. Te opposite may occur as well. For instance, you might have found yourself in a situation in which someone you don’t really care for addresses you too intimately. If you have, you probably responded to them using a more formal language than they used with you. Using your language to accentuate social distance from others is known as linguistic divergence. Convergence and divergence are dynamic processes. We constantly shif properties of our language toward and away from our interlocutors, ofen in conscious ways that respond to the environment we fnd ourselves in. For instance, we recognize that children may have smaller vocabularies and less complex sentence structures than we do, so we tend to use diferent words with children than we would with adults and simpler sentences as well. In this way, we accommodate to the conversational style of children to enhance our communication with them.

5.2 CASE STUDY: CONVERGENCE IN LANGUAGE CHOICE IN MONTREAL Te drif of convergence and divergence can play out across languages among bilinguals. A person who uses more than one language can have a broad range of interlocutors; some of these individuals will share the same languages with that person and others may only share one. Gauging which language to choose when addressing a conversation partner requires an individual to decide how they would like to be perceived in an interaction. Even very young bilingual children have been shown to judge the language to use with an interlocutor they do not know based on their perception of the addressee (Genesee et al., 1996). In choosing a language, you may decide to be cooperative and converge to the language of your addressee, or you might decide to assert your diference, resisting the switch to their language. Te choices you make can have important social psychological consequences. Tis was shown in a series of studies of accommodation studies among anglophone (English-speaking) and francophone (French-speaking) Montrealers in Quebec, Canada (Giles et al., 1973; Bourhis, 1984; Bourhis et al., 2007; Moise & Bourhis, 1994). Beginning in the 1970s, the researchers sought to investigate whether the Loi 101 of 1977, a law that attempted to boost the status and use of French in Quebec, would increase the use of French among individuals in the public sphere. Te studies took place on the busy streets of Montreal during rush hour. A bilingual French-English young woman approached pedestrians and asked them for directions, sometimes using French and sometimes English. Te pedestrians did not know that she was a researcher until afer the interaction, and because she was only pretending to be a lost pedestrian, she is called a “confederate” in studies of this kind. Afer the interaction, the confederate wrote down how the pedestrian responded to her. If the confederate used a pedestrian’s frst language when initially addressing them, they always responded in that language. If the confederate frst

Social networks and accommodation

67

used the pedestrian’s second language and they responded to her fully or in part in that language, this was noted as a convergent response. If pedestrians did not switch languages to accommodate to the pedestrian in the language she used, this was recorded as a divergence. For the frst few decades afer the Loi 101 went into efect, the results of these studies showed the grip of English in Quebec. Te francophone pedestrians accommodated overwhelmingly to the confederate by switching to their second language, English, when she addressed them in that language, but the anglophone passers-by frequently resisted a switch to French. It wasn’t until the studies conducted in the 90s that the anglophone pedestrians began to accommodate to their second language, French, at rates equal to the francophone’s convergence when responding to the confederate. Te Montreal studies show us that there can be a socio-historical dimension to accommodation in communication. Te status of French and English among the populace in Canada is not equal, and this is refected, even today, in the patterns of communication accommodation that can be observed to take place between private citizens in the public sphere (Neupané, 2020). Sometimes accommodation in communication is afected by factors that are beyond an individual’s control, including regional attitudes, linguistic inequality, and national language ideologies.

5.3 AUDIENCE DESIGN AND STYLE From our excursion to the streets of Montreal, we learned that human communication is not neutral. Language use can refect the social inequalities and political realities of the world around us. And we know beyond a doubt that we accomplish much more with language than a mere exchange of information. We use language in the attempt to infuence others’ behaviors, to convey our attitudes toward someone or something, to express solidarity or opposition, to calm ourselves down, and to communicate who we are and how we want to be perceived in any given situation. To accomplish such diferent functions, we use our languages dynamically, shifing toward or away from our interlocutors as the social situation demands. Te sociolinguist Alan Bell (1984, 2001) formulated a sociolinguistic framework of style variation known as audience design that accounts for shifs in language as a response to someone we are speaking to, an addressee. Bell’s model includes the notion of referee design to describe situations in which we initiate style shifs according to audiences that are not actually present in our conversation, called referees. Initiative style shifing in language is indexical, which means that we use certain linguistic elements rather than others when we wish to accentuate our afliation to a social group or to project a particular persona (Eckert, 2008). Referee design was inspired by Bell’s observation of fve New Zealand public radio news reporters who were charged with reading the news on diferent radio stations. Te reporters from the pool would accommodate their speech style to their expected audience, adopting a British broadcasting accent called Received Pronunciation for presumed audiences of higher social status and shifing their speech to a local variety of English when presenting the same news stories for a New Zealand radio audience presumed to be

68

From minutes to years

of lower social standing. Teir behavior showed that accommodation is not just a responsive alignment to an interlocutor. Instead, accommodation can occur when we envision absent referees and we want to be perceived in a certain way by them. Tis may be a way to account for the changes observed in the vowels of Queen Elizabeth II. Referee design allows us to maintain or to project an identity through the way we use our language. Te efects of an absent referee on stylistic choices can be ofen observed when individuals slip in and out of using ethnolectal linguistic variants in their speech. Tese are features that index one’s ethnic orientation. Ethnolects of a language arise in communities formed by immigrants. Tese speech varieties are a product of language contact and multilingualism where speakers become used to using words, meanings, and structures from their minority language in the majority language of the community. To get a favor of what we mean by an ethnolect, a citation of “Yinglish” from Philip Roth’s novel, Portnoy’s Complaint, is provided in (5.1) (Prince, 1992, p. 6). (5.1) Tat night, afer dinner, I gave Aunt Gladys a kiss and told her she shouldn’t work so hard.“In less than a week it’s Rosh Hashana and he thinks I should take a vacation. Ten people I’m having. What do you think, a chicken cleans itself?” (Roth, 1963, p. 86) In the “Yiddish Movement” structure highlighted in 5.1, the focused object, “[T]en people” is displaced to the phrase-initial position, as it would be in the Yiddish translation of this phrase (Prince, 1981). Focus fronting in American English is a discourse strategy exclusive to Yiddish-English bilinguals and to their offspring, who may or may not actually speak Yiddish (Prince, 1992). Its use signals that the speaker wishes to be perceived as a member of the Jewish community. Referee design helps us to understand how individuals use linguistic features to express their solidarity with others, whether they are members of that group (ingroup) or not (outgroup). For instance, the former president of the U.S., Barack Obama, was also known to use more linguistic variants from African American Vernacular English when speaking on themes of race and civil rights than when he was speaking on other topics (Holliday, 2016).And a study of second-generation Punjabi speakers in England showed that those who self-identifed as “Asian” on a questionnaire pronounced “r” sounds (rhotics) afer vowels when speaking English signifcantly more frequently than did those who self-identifed as “British Asian” (Hirson & Sohail, 2007). Te British Asian participants generally dropped the “r” sound afer vowels, consistent with the non-rhotic pronunciation of the South East British varieties. Under referee design, the suggestion of an ingroup or outgroup audience can activate an aspect of our identities and trigger us to shif our linguistic style.

Social networks and accommodation

5.4

69

LANGUAGE MODE

To begin to think about how bilinguals shif styles to accommodate to addressees, present or absent, we can use the concept of language mode. Te psycholinguist François Grosjean proposed the concept of language mode to describe a continuum of diferent degrees of activation of each language in a bilingual speaker (Grosjean, 2001). We can interpret this as a continuum with monolingual poles at either end. A bilingual is in monolingual mode when activating only one of their languages and inhibiting the other. But the same speaker, under diferent circumstances, may also activate both simultaneously, and to diferent degrees, when in a bilingual mode. A representation of the continuum is found in Figure 5.1. where the end points represent the monolingual modes of a bilingual’s component languages. A language user’s behavior can fall anywhere along the continuum. Along the tails of the fgure, one language is more activated than the other and to varying degrees, depending on how far out the position falls along the tail. Only in the middle are both languages equally “on,” meaning that they are both highly activated. In laboratory experiments, language mode can be manipulated by the researcher. For instance, to replicate a monolingual mode, study participants can be asked to complete tasks that require them to name numbers or pictures in only one language. Ten, they can be asked to return to the laboratory in a diferent session to repeat the task, but this time, they must name the picture or number in more than one language. Usually, the participants are cued about the language to use in this task by a change in the background color of the number or picture (e.g., blue for one language and red for another) (Olson, 2012, 2016). In ordinary interactions outside a laboratory setting, language mode is guided by multiple factors. For instance, the topic of the message could skew the interaction more toward one language than another. Imagine, for instance, that you are a bilingual of foreign-born parents talking about a culturally rich topic with your family, like your favorite food from their country of origin. Tis topic could trigger a bilingual mode, as the discussion of food triggers words and phrases from Figure 5.1 Language mode: LA = language A, LB = language B

70

From minutes to years

the ancestral language. Te context of the interaction may also afect the language mode of a conversation. If you spend time in a community with few bilinguals, you may tend to operate more frequently in a monolingual mode than when you are in a community where bilingualism is prevalent. Or if you are with members of your family who use a minority language, you may switch into it more with them than in other circumstances. Intertwined with factors like the topic and context of the conversation is the fact that our language mode is dependent on who our audience is. When bilinguals are communicating with others who share only one of their languages, they operate mostly in a monolingual mode, using the words and structures that are common to all participants in the interaction. When they fnd themselves with addressees who share their languages, they can be observed to operate in a bilingual mode, using elements from both languages to diferent degrees and in varying ways. Tis can include using code-switches and borrowings but also using forms that show the infuence of sound patterns, words, meanings, and structures from the language that they are not currently using. Bilingual mode can have short- and long-term efects on language. In a phenomenon known as “phonetic or gestural drif,” speakers who are immersed in a second language environment have been shown to shif the phonetic boundaries of the sounds of their frst language toward those of the second language (Chang, 2012; de Leeuw, 2019; Flege, 1987; Major, 2008; Sancier & Fowler, 1997). Phonetic drif can happen over the very short term. Chang (2012) demonstrated that American students studying Korean in a six-week immersive program in Korea showed the efects of Korean on their pronunciation of English within two weeks. Drif can also be dynamic, with a speaker adjusting the pronunciation of one or both languages according to the surrounding speech environment. For instance, a study showed that Portuguese listeners could reliably judge from speech samples of a Portuguese-English bilingual whether the recording of the bilingual had been made when she was in Brazil versus when she was in the United States because the speaker adjusted the phonetic boundaries of her stop consonants in both languages toward those of the language she was immersed in at the time of each recording (Sancier & Fowler, 1997). Phonetic drif can have long-term consequences on the speech of a bilingual. For instance, a study of the German of the formerly #1-ranked tennis champion, Stef Graf, showed the efect of her long-term exposure to English (de Leeuw, 2019). Born in Germany, the athlete moved to the United States when she married the American tennis champion Andre Agassi. Several changes in Graf ’s German are attributable to English infuence, including the use of a velarized, or dark-l, [ɬ] as in the word fall [faɬ], a shif in the vowel [i], and alterations in her pitch. Tese studies of drif in pronunciation demonstrate that we can attune our languages to ft our environment. Te fne adjustments that the speakers made to their speech within these studies may not be noticeable to other listeners, but they are measurable, signifcant changes.An important consequence of the notion of phonetic drif is that our frst language is not “fxed” at a very early age, as is sometimes assumed.

Social networks and accommodation

71

Instead, we make changes to them throughout our lifetimes, particularly if they go from being the language that we use most ofen to being our non-dominant language. If we imagine an entire community of individuals like Stef Graf, who immerse themselves in a new language environment and adopt the new language as their everyday form of communication, we can begin to imagine how small adjustments made in language contact by individuals can lead to language changes over time.

5.5

LANGUAGE MODE AND BILINGUAL STYLE

Te cases of phonetic drif discussed previously add a bilingual dimension to the concept of accommodation, as it is likely that the traveling Brazilian student and the students studying Korean were aligning to their audiences in an unconscious, automatic way. We can view this type of convergence as responsive. Let’s look at some cases in which bilinguals perform shifs in their styles in a way that is indexical, that is, to project a particular identity. Some linguists have suggested that what monolinguals do with style shifing, bilinguals can do by code-switching between their languages (Romaine, 1995). However, this cannot fully describe linguistic style within a bilingual individual.

5.5.1

Style shifting in monolingual mode

In bilingual communities, people are ofen exposed to multiple varieties of the languages they use. For instance, they can be exposed to accented varieties of the dominant language as it is spoken by newly arrived members of the community who are only just learning the language or by older immigrants who have not been schooled in language. At the same time, they can be also exposed to “accented” varieties of their ancestral language as speakers from younger generations become dominant language users.We can say that members of multilingual communities have an expanded linguistic repertoire relative to members of dominant language communities where nearly everyone shares the same code. Research has shown that speakers in such communities are capable of acquiring more than one variety of their languages, an accented one and an unaccented one (Agnihotri, 1979; Heselwood & McChrystal, 2000; Khattab, 2009, 2013, 2003). Khattab (2013, p. 460), in particular, shows how three young Arabic-English bilingual children use diferent varieties of Arabic and English in acts of accommodation with their mothers. Te children are being raised in the United Kingdom, where English dominates, and they are sometimes reluctant to use Arabic. Khattab recorded them in separate English and Arabic sessions to control for monolingual mode so that she could get baseline vocabulary counts in each language and map out the phonological system for each child in their monolingual modes. She discovered that, even in a supposedly monolingual Arabic mode, the children calibrated their variety to an accented one. As an example, one 5-year-old child ofered an Arabic-accented English response to her mother’s question, even though she knew the actual word in Arabic and fnally uttered it when her mother persisted. Tis is shown in 5.2:

72

From minutes to years

(5.2) Conversation between 5-year-old bilingual child and her mother, adapted from Khattab (2013, p. 460) Mother: [ʃu haɪd-e?] what this-3F.SG? “What’s this?” Child (with a cheeky tone): [ʔʌmbrɛlla] “Umbrella” Mother (calling her bluf): [w-bɪ-l-ʕaɾabe]? and-LOC-DEF.ART-Arabic? “And in Arabic?” Child: [ʃamsijje] “Umbrella”

Te child’s response, with an initial glottal consonant [Ɂ] before the vowel, a trilled [r], and a phonetically long [l] show the transfer of Arabic phonetic features onto an English word. She also produced the word “umbrella” during a monolingual English session without such pronunciation. What is clear from Khattab’s study is that the children are capable of strategically shifing styles within a language in addition to code-switching between their languages in acts of accommodation to their mothers.

5.5.2 Bilingual mode and style shifting While we have already considered code-switching and language choice as a function of accommodation, we also need to consider that code-switching itself is subject to variation. As an illustration, we can refer to a study by Trefers-Daller (2011), who used an audience design model to investigate the efects of language mode on the speech of a Turkish-German bilingual in Turkey. Te speaker, Emre, was raised in Germany until early adolescence, when he moved to Turkey with his parents, who are Turkish.Although living in Turkey at the time of the study, Emre still had Turkish-German friends living in Germany. Trefers-Daller created the context for interviewing Emre in diferent bilingual modes by modifying his conversation partner. In one session, he spoke primarily in German with the researcher herself. Te researcher refers to this as monolingual (German) mode. In a separate session, Emre interacted on the phone with his TurkishGerman friend in Germany; this was called the bilingual (German) mode. In the third and last session, he interacted with a close Turkish-German–speaking friend who was also living in Turkey, using a bilingual (Turkish) mode. Te results of the study show that Emre calibrated the amount of each language used according to the diferent styles used by his interlocutors. With the researcher, he used predominantly, but not exclusively, German. At times, he would use a wellknown Turkish word to refer to concepts that were culturally specifc to Turkish people. In the conversations with his bilingual friends, he used both languages. However, he showed signifcantly diferent amounts of language mixing in these interactions. Te most mixing occurred with his close childhood friend living in Turkey. Te rates of code-switching that he used in his bilingual conversations closely tracked the rates

Social networks and accommodation

73

used by his conversational partners, a sign of accommodation. Te manner of mixing that he used in the bilingual mode environments also varied according to his addressee. In conversation with the German-dominant addressees, Emre used mostly tag switches and borrowings from Turkish that need not be integrated into the grammar of German. But with his Turkish-dominant friend, he used more variable codeswitching, including inter- and intra-sentential switches, requiring the speakers to integrate their grammars at a more complex level. Tus, with a bilingual mode, the same individual was observed to vary his linguistic behavior to accommodate to the quantity and quality of the language mixing of his audience. In other words, codeswitching is not a single style; rather it, too, can vary as a function of style. Overall, as Khattab (2013) has noted, we can say that accommodation is more variable in bilinguals than in monolinguals because bilinguals have a much larger array of audiences. Both monolinguals and bilinguals converge and diverge from their addressees or referees within a language in monolingual mode. But when in bilingual mode, bilinguals show stylistic variation within each language, aligning to audiences whose language behaviors and competencies can be situated anywhere along the bilingual mode continuum, producing “unaccented,” “accented,” and intermediate forms. In addition, bilinguals show variation in style while switching between their languages.

5.6

SOCIAL NETWORKS

Language users and their interlocutors have interpersonal ties with one another. Networks are active processes that imply change as people infuence one another. In a social network where connections between individuals are frequent and the ties between individuals are close or intimate, the linguistic behavior of an individual language user can afect an entire network. Te Chinese American novelist Amy Tan articulates the efect of her close family network when she provides a description of what we could call the “bilingual mode” English that she uses with her mother and her husband: (5.3) Just last week, I was walking down the street with my mother, and I again found myself conscious of the English I was using, the English I do use with her. We were talking about the price of new and used furniture and I heard myself saying this: “Not waste money that way.” My husband was with us as well, and he didn’t notice any switch in my English.And then I realized why. It’s because over the twenty years we’ve been together I’ve ofen used that same kind of English with him, and sometimes he even uses it with me. It has become our language of intimacy, a diferent sort of English that relates to family talk, the language I grew up with. (Tan, 1990) Te kind of “family talk” that the couple uses among themselves is not unusual in bilingual families.We get used to the language patterns of our friends and family, and

74

From minutes to years

Figure 5.2 Weak tie network

we adapt to the idiosyncrasies of the members of our network who have learned the language late in life. Stand-up comedians the world over can ofen get a laugh from their audience by imitating the way an older immigrant relative uses the dominant language because their experiences resonate with nearly anyone who has such speakers in their close social networks. Te sociologist Mark Granovetter pioneered a theory of how individuals form personal communication networks (Granovetter, 1973). Within social network theory, there are two main types of networks: dense networks, composed of strong, frequent, and intimate ties between individuals, and loose ones, in which the ties between individuals are infrequent. We can think of this as the diference between communication with close friends and family versus communication with acquaintances. We can illustrate the diference between a weak tie network in Figure 5.2 and a strong tie network in Figure 5.3. Tis theory has infuenced work on sociolinguistics because it predicts that networks composed of strong ties are resistant to linguistic variation that comes from the outside (J. Milroy & Milroy, 1985; L. Milroy & Llamas, 2013). In this sense, strong networks are more linguistically conservative and enforce local language norms. Networks with weak ties are the ones that are predicted to contribute to the spread of innovation because the individuals that compose them are ofen at the fringes of the network, so they have contact with a wider range of interlocutors and they are more open to external innovation.

Social networks and accommodation

75

Figure 5.3 Strong tie network

5.6.1

Bilinguals’ social networks

Studies on the social networks of bilinguals is ofen concerned with language choice. Tey are focused on an individual at the center of what is referred to as an ego network (Daming et al., 2009). In one such study, Dow (2016) interviewed the female members of multiple generations of 26 families of North African descent living in France to determine the amount of French,Arabic, and Kabyle (Berber) they use with diferent interlocutors, at diferent times, and at diferent periods of their lives. She asked a subset of these women to draw their own ego networks of interpersonal communication, indicating the languages they use with diferent interlocutors and the languages the interlocutors use with them. Tey were also asked to draw a timeline of how the relative use of one versus the other of these languages has changed over time. In addition, nine of these same women also participated in a mobile phone study where they would respond to a text message prompt, sent at randomized times over a 5-week period, to report the languages they used with their two most recent interactions. Tese diferent methodologies allowed the researcher to determine that some individuals almost always interact in dense social networks, communicating mainly or exclusively with family members in France or abroad, while others have more open networks. In both dense and loose networks, the heritage languages, Arabic and/or Kabyle, were used most ofen with intimate ties and rarely with new acquaintances. In this respect, the heritage languages had become a home language, rarely used with outsiders, a situation that is not unusual in the context of immigration. And, as many

76

From minutes to years

studies of bilingual social networks have shown, the members of the older generations, when interacting together, showed the most use of Arabic or Kabyle, and members of the younger generations spoke the heritage language more consistently with older family members than they did with member of their own generation, many of whom where French speakers that they knew from work or school. Te multiple social networks that bilinguals maintain may be more important than for monolinguals. If you do not maintain connections with others who use one of your languages, you are likely to cease using it.We commonly see a pattern of a threegeneration shif in immigrant communities where the frst generation to immigrate is dominant in the minority language, the second generation is bilingual in the minority language and the majority language of the community, and the third generation mostly shifs to the majority language. Only if you are part of a close, dense network of minority language users of your own age are you likely to maintain a heritage language. Certainly, bilingual social networks are more diverse than monolingual ones because of language mode and because of the mobility of the people who populate a network. Te individuals at the center of bilingual ego networks, like the Turkish-German and Arabic-Kabyle-French multilinguals discussed previously, have interactions that are monolingual in one or all their languages while also maintaining interpersonal ties that are conducted in bilingual mode. In addition, they may have moved from country to country throughout their lives, so they have dynamic networks, with ties of interpersonal contact that vary from one context to another. Tese factors complicate the study of bilinguals’ networks using the traditional sociolinguistic techniques of interviews or in-person observations. For this reason, the bilingual social networks studied by sociolinguists have been small sized and usually ego focused. Fortunately, the technology now exists to explore the efects of social networks on larger multilingual networks on micro-blogging sites, like Twitter (Eleta & Golbeck, 2014; Hale, 2014; Kim et al., 2014; Laitinen et al., 2020; Nguyen et al., 2015). Twitter data allow an analyst to map out the communication ties of many users. Doing so has revealed aspects of interpersonal communication that mirror those of smaller networks. For instance, Kim et al. (2014) found that the bilingual users they studied from Qatar, Quebec, and Switzerland tend to accommodate to the rates of code-switching of the other bilingual users in their network. In a study of 92 users who used English and one of 18 diferent languages in their ego networks, Eleta and Golbeck (2014, p. 431) showed that “multilingual Twitter users perceive the language composition of their network and interact accordingly.” In other words, they accommodate to the language choices of other users in their network. Much of the current research in multilingual social networks on Twitter focuses on the role of English among users of other global languages. A consistent fnding is that English users and multilingual users serve as a bridge between monolingual users of other languages (Eleta & Golbeck, 2014; Hale, 2014; Kim et al., 2014; Laitinen et al., 2020; Nguyen et al., 2015). Tis makes sense even if we think of smaller-scale ego networks in immigrant contexts, where multilingual children ofen serve as language brokers for their parents, translating and interpreting, when necessary, in interactions outside the home. Research on Twitter has also challenged the idea that networks with close ties resist innovation. In a study of a corpus of users from the Nordic countries of Denmark,

Social networks and accommodation

77

Norway, Iceland, Finland, and Sweden, Laitinen et al., 2020 showed that as the size of the network increases, the diference between strong and weak ties tends to collapse and innovations can spread through strong networks as well as weak ones (Laitinen et al., 2020). Tese results suggest that innovations may spread through dense networks, such as dense immigrant communities in urban centers such as New York City, London, or Dubai. And indeed, the evidence from sociolinguist studies also suggests that innovations can occur where large networks of minority language users may be found. A study of 142 Spanish speakers in New York City revealed elevated rates of subject pronoun usage among the New York City–born speakers relative to those born in their home countries (Otheguy et al., 2007). Tis suggests that there is accommodation among Spanish speakers in New York City, implying that some speakers serve as bridges between networks of families from diferent dialect regions and that features of accommodation spread through the networks of other Spanish speakers.

Chapter summary We began this chapter with an act of accommodation by one person, Queen Elizabeth II, and we’ve ended it with a consideration of how innovations can spread among users on vast networks like Twitter. It is through local acts of accommodation that we adapt new stylistic features, some of which take hold as innovations that we pass on to the interpersonal ties in our ego networks. And they, in turn, pass them on to others. We have seen that the strategies of accommodation among bilinguals are more complex than those of monolinguals because they can position themselves within a range of language modes, and they have an expanded repertoire of linguistic features to draw from, of which language switching is only one. Finally, bilinguals’ social networks are more complicated than monolinguals’, so there is still much more to learn about how innovations within and between languages circulate within bilingual social networks.

EXERCISES 1.

Discover

Consider the patterns of communication between users of minority languages or varieties with majority-language users where you live. What situations or context might motivate majority-language users to linguistically accommodate to minoritylanguage interlocutors?

2.

Discover

Ask a bilingual friend or family member to draw a timeline of their language use through their lifespan. If your friend or family member hasn’t noticed any change in their language use over their lifespan, ask them to consider how they use their languages

78

From minutes to years

during a typical day. What are the major events or contexts that cause them to activate one language more than another? Is it always the same language that is most activated?

3.

Create

Te network models illustrated in Figure 5.2 and 5.3 do not contain any information about the languages that people use to communicate with one another. How might we adapt a network to take into account the bilingual mode? Sketch out a network model that might link one node (representing a person) to another or others by language mode.

FURTHER READING Sachdev, I., & Giles, H. (2006). Bilingual Accommodation. In T. K. Bhatia & W. C. Ritchie (Eds.), Te handbook of bilingualism (pp. 353–378). Blackwell Publishing Ltd. Tis handbook chapter presents a social psychology view of bilingual communication, with a focus on how individuals use language to make community.

REFERENCES Agnihotri, R. K. (1979). Processes of assimilation: A sociolinguistic study of Sikh children in Leeds [PhD thesis, University of York]. Bell, A. (1984). Language style as audience design. Language in Society, 13(2), 145–204. Bell, A. (2001). Back in style: Reworking audience design. In P. Eckert & J. Rickford (Eds.), Style and sociolinguistic variation. Cambridge University Press. Bourhis, R. Y. (1984). Cross-cultural communication in Montreal: Two feld studies since Bill 101. International Journal of the Sociology of Language, 46, 33–47. Bourhis, R. Y., Montaruli, E., & Amiot, C. E. (2007). Language planning and French-English bilingual communication: Montreal feld studies from 1977 to 1997. International Journal of the Sociology of Language, 185, 187–224. Chang, C. B. (2012). Rapid and multifaceted efects of second-language learning on frst-language speech production. Journal of Phonetics, 40(2), 249–268. Coupland, N., & Giles, H. (Eds.). (1988). Communicative accommodation: Recent developments. Language and Communication, 8(3–4), 175–327. PsycINFO. Daming, X., Xiaomei, W., & Li Wei. (2009). Social network analysis. In Li Wei & M. G. Moyer (Eds.), Te Blackwell guide to research methods in bilingualism and multilingualism (pp. 263–274). Blackwell Publishing Ltd. https://doi.org/10.1002/9781444301120.ch15 de Leeuw, E. (2019). Native speech plasticity in the German-English late bilingual Stefanie Graf: A longitudinal study over four decades. Journal of Phonetics, 73, 24–39. https://doi.org/10.1016/j. wocn.2018.12.002 Dow, A. T. (2016). Mother daughter tongue: Te language use of North African women in France [PhD thesis, University of Texas]. Eckert, P. (2008). Variation and the indexical feld. Journal of Sociolinguistics, 12(4), 453–476. Eleta, I., & Golbeck, J. (2014). Multilingual use of Twitter: Social networks at the language frontier. Computers in Human Behavior, 41, 424–432. https://doi.org/10.1016/j.chb.2014.05.005 Flege, J. E. (1987). Te production of “new” and “similar” phones in a foreign language: Evidence for the efect of equivalence classifcation. Journal of Phonetics, 15(1), 47–65.

Social networks and accommodation

79

Genesee, F., Boivin, I., & Nicoladis, E. (1996). Talking with strangers: A study of bilingual children’s communicative competence. Applied Psycholinguistics, 17(4), 427–442. Giles, H. (1984). Te dynamics of speech accommodation. International Journal of the Sociology of Language, 46, 1–155. Giles, H., Taylor, D. M., & Work(s), R. B. R. (1973). Towards a theory of interpersonal accommodation through language: Some Canadian data. Language in Society, 2(2), 177–192. Granovetter, M. S. (1973). Te strength of weak ties. American Journal of Sociology, 78(6), 1360–1380. Grosjean, F. (2001). Te bilingual’s language modes. In J. L. Nicol (Ed.), One mind, two languages: Bilingual language processing (2005–03262–001, pp. 1–22). Blackwell. Hale, S. A. (2014). Global connectivity and multilinguals in the Twitter network. CHI’14: Proceedings of the SIGCHI conference on human factors in computing systems (pp. 833–842). Association for Computing Machinery. doi: https://10.1145/2556288.2557203 Harrington, J. (2006). An acoustic analysis of “happy-tensing” in the Queen’s Christmas broadcasts. Journal of Phonetics, 34(4), 439–457. Harrington, J., Palethorpe, S., & Watson, C. I. (2000). Does the Queen speak the Queen’s English? Nature, 408(6815), 927–928. http://doi.org/10.1038/35050160 Heselwood, B., & McChrystal, L. (2000). Gender, accent features and voicing in Panjabi-English bilingual children. Leeds Working Papers in Linguistics and Phonetics, 8, 45–70. CSA Linguistics and Language Behavior Abstracts. Hirson, A., & Sohail, N. (2007). Variability of rhotics in Punjabi-English bilinguals. Proceedings of the International Conference of Phonetic Sciences (ICPhS XVI) (pp. 1501–1504). Saarbrücken, Germany. http://www.icphs2007.de/conference/Papers/1623/1623.pdf Holliday, N. R. (2016). Intonational variation, linguistic style and the Black/Biracial experience [PhD thesis, New York University]. Khattab, G. (2003).Age, input, and language mode factors in the acquisition of VOT by English-Arabic bilingual children. Proceedings of the 15th international congress of phonetic sciences, Barcelona, Spain, 1–4. www.internationalphoneticassociation.org/icphs-proceedings/ICPhS2003/papers/ p15_3213.pdf Khattab, G. (2009). Phonetic accommodation in children’s code-switching. In B. E. Bullock, & A. J. Toribio (Eds.), Te Cambridge handbook of linguistic code-switching (pp. 142–160). Cambridge University Press. Khattab, G. (2013). Phonetic convergence and divergence strategies in English-Arabic bilingual children. Linguistics, 51(2), 439–472. https://doi.org/10.1515/ling-2013-0017 Kim, S.,Weber, I., Li Wei, & Oh, A. (2014). Sociolinguistic analysis of Twitter in multilingual societies. Proceedings of the 25th ACM conference on hypertext and social media (pp. 243–248). Association for Computing Machinery, New York. https://doi.org/10.1145/2631775/2631824. Laitinen, M., Fatemi, M., & Lundberg, J. (2020). Size matters: Digital social networks and language change. Frontiers in Artifcial Intelligence, 3, 46. https://doi.org/10.3389/frai.2020.00046 Li, Wei. (1994). Tree generations, two languages, one family: Language choice and language shif in a Chinese community in Britain. Multilingual Matters. Major, R. C. (2008). Transfer in second language phonology. Phonology and Second Language Acquisition, 36, 63–94. Milroy, J., & Milroy, L. (1985). Linguistic change, social network and speaker innovation. Journal of Linguistics, 21, 339–384. Milroy, L., & Llamas, C. (2013). Social networks. In J. K. Chambers & N. Schilling (Eds.), Te handbook of language variation and change (pp.  407–427). John Wiley and Sons, Inc. https://doi. org/10.1002/9781118335598.ch19 Moise, L. C., & Bourhis, R.Y. (1994). Langage et ethnicité: Communication interculturelle à Montréal, 1977–1991. Canadian Ethnic Studies, 26(1), 86.

80

From minutes to years

Neupané, R. (2020). A multidisciplinary approach to studying language attitudes and language use in the Ottawa-Gatineau region [PhD thesis, Te University of Texas at Austin]. Nguyen, D., Doğruöz, A. S., Rosé, C. P., & de Jong, F. (2015). Computational sociolinguistics: A survey. ArXiv Preprint ArXiv:1508.07544. http://arxiv.org/abs/1508.07544 Olson, D. J. (2012). Bilingual language contexts: Variable language switching costs and phonetic production [PhD thesis, University of Texas]. Olson, D. J. (2016). Te gradient efect of context on language switching and lexical access in bilingual production. Applied Psycholinguistics, 37(3), 725–756. https://doi.org/10.1017/S0142716415000223 Otheguy, R., Zentella, A. C., & Livert, D. (2007). Language and dialect contact in Spanish in New York: Toward the formation of a speech community. Language, 83(4), 770–802. Prince, E. F. (1981). Topicalization, focus-movement, and Yiddish-movement: A pragmatic diferentiation. Annual Meeting of the Berkeley Linguistics Society, 7, 249–264. Prince, E. F. (1992). On syntax in discourse, in language contact situations. In C. Kramsch, & S. McConnell-Ginet (Eds.), Text and context: Cross-disciplinary perspectives on language study (pp. 98–112). DC Heath & Company. Romaine, S. (1995). Bilingualism (2nd ed.). Blackwell. Roth, P. (1963). Portnoy’s complaint. Vintage. Sancier, M. L., & Fowler, C.A. (1997). Gestural drif in a bilingual speaker of Brazilian Portuguese and English. Journal of Phonetics, 25(4), 421–436. Tan, A. (1990). Mother tongue. Te Treepenny Review, 43, 7–8. Trefers-Daller, J. (2011). Variability in code-switching styles: Turkish-German code-switching patterns. In R. Jacobson (Ed.), Code-switching worldwide, Bd. I (pp. 177–198). De Gruyter Mouton.

6 Acquisition and attrition

CHAPTER PREVIEW In this chapter, we address language contact within the individual. Te chapter begins with a brief historical overview of the competing theories of “child language acquisition” in general before we consider how learning two languages from an early age might difer from acquiring only one. Because not every bilingual is born into a home or community where they are constantly exposed to two or more languages from birth, we look at diferent individual trajectories to see how they might difer, focusing in the second part of this chapter on what is known as “second language acquisition.” Finally, we introduce the notion of “language attrition” in bilinguals. We end this chapter with a consideration of how the outcomes of the difering individual processes of acquisition and attrition could lead to language change.

6.1

BILINGUAL ACQUISITION

All typically developing children learn how to speak or sign, whether they are exposed to only one language or to multiple languages. And they acquire language, which is full of structural complexity, at a very early age. In fact, before they begin formal schooling, most of the grammar of the language or languages that a child is exposed to is in place. Te function of primary education for children is not to teach them language, because they already have a language in place; rather it is to teach them literacy skills.

A note on bimodal acquisition Many children in primary schools are taught literacy skills in a language that they already know. For others, though, they may arrive at school using a diferent language from the one that they may be expected to use in school. Tis is ofen the case for students from minority-language or foreign-born families, who begin to acquire their second language only when they start to attend school. Tis is also very ofen the case for Deaf children. Deaf children who DOI: 10.4324/9781003167952-8

82

From minutes to years

arrive at school communicating in sign language must acquire skills, including literacy, in a second language. So it is important to remember that all Deaf students are bilingual, and they are also “bimodal,” meaning that they have acquired language competencies in two modalities, the visual, through the sign language, and the spoken, through training in lip reading and in reading in the local spoken language.

Children go through well-known milestones of language acquisition, beginning with cooing and babbling in the early stages. By 10–12 months, they start to pronounce their frst words in the“one-word stage” of acquisition.At 18–24 months, they will use strings of words in what is known as the “two-word stage” of acquisition. Tis is followed by the use of basic syntax from roughly the ages of 24–36 months.Around 36 months, children begin to use complex, multi-word sentences. Tere have been two major theoretical proposals about how children acquire languages so early: the “nativist” approach and the “usage-based,” or emergentist, approach. Tese approaches roughly run along the lines of whether human development is guided by nature, in which case behaviors are innate or instinctive, or by nurture, where a child’s environment plays a guiding role in development. Te nativist approach, most famously represented by Noam Chomsky, holds that language learning is to a large extent innate (Chomsky, 2006). Tis doesn’t mean that you are born with an instinct to acquire a particular language, like Igbo or Malay. Instead, Chomsky theorized that humans are born with a “language acquisition device” that allows children to use a set of innate rules for unlocking the structure of any language, whether they be exposed to Igbo, to Malay, or to any other language from birth. Tis hypothesis was built on the idea that the language that infants hear around them is too messy, and in amounts that are too limited, for them to develop the grammatical complexity that they do at such an early age. Nativists argued that babies must be “hard wired” to pay attention to certain properties of the languages they are exposed to. Once they fgure out certain things from the linguistic input around them, like the fact that objects tend to come afer verbs, or that facing your palm down, as opposed to up, is meaningful in a sign modality, they can use these observations to choose the correct grammar settings for their language. In contrast to the nativist view,usage-based approaches hold that language is acquired in social interaction using the types of general cognitive capacities that humans use for all kinds of behaviors (Tomasello, 2003). Tese behaviors include intention-reading, which means that children are able to fgure out the social goals that language users exploit to achieve certain ends, and pattern fnding, which allows learners to extract statistical regularities from the input they hear, for instance that the word, puppy is ofen preceded by the and ofen co-occurs with the word fufy (e.g., Te puppy is fufy. What a fufy puppy!). Under usage-based approaches, specialized language areas of the brain do not just grow on their own, they emerge as linguistic experience increases, and the

Acquisition and attrition

83

ability to create grammatical abstractions, like the fact that puppy belongs in the class of nouns, is formed through experience rather than acquired as an instinct. For this reason, usage-based models are sometimes also called “emergentist.” As its name suggests, proponents of usage-based models think that language structure comes from language use. Here, the exact nature of the “input” that children receive is key because it’s only through repeated use that children can abstract from chunks to isolated words. In the initial stages, children do not know where word boundaries are; instead, in the two-word stage, they use language in chunks that typically include a fxed word and an open slot (X): X gone, all X, no X. With language use, the individual words, and abstract categories of grammar, like noun, verb, and so on, emerge. You might remember having been taught to recite something as a child, like a prayer, the alphabet, or a pledge, without knowing exactly what words you were saying. For instance, English-speaking children learning to recite the alphabet commonly hear and pronounce the alphabetic sequence l-m-n-o as a chunk, like an adjective describing the letter p: “elemental p”! Since input is fundamental to usage-based models, researchers collect data that will aford them insights into the language that children receive from those around them. Typically, these researchers will record parent–child interactions to study the quality of language that parents use with the child. Tis will allow them to study the ways in which adults might modify their language to make it more accessible to children. Te words, phrases, and constructions collected from those who are communicating to a child can then be compared to the productions of the children to see the degree to which they match. Who addresses young children can be diferent across cultures. In some societies, most of the linguistic input directed to the child is from their adult caretaker, while in others, children receive their linguistic input from other children. Studies have investigated these diferences to see if they matter in the quality and quantity of language input that a child is exposed to. One study examined children of Sesotho-speaking parents who grow up in multi-generational, subsistence-farming households in South Africa (Loukatou et al., 2021). In this culture, most input to the younger language learners is provided by other children rather than by their parents. Te linguistic productions of the Sesotho-acquiring children and of their interlocutors were compared to those of the French-acquiring children living in urban France, who receive most of their language from their mothers. Te quality of the language directed to the children in South Africa and those in France turned out to be the same. So these cultural diferences do not appear to matter. Since bilingual children acquire two languages, rather than only one, there are three factors to consider in bilingual language acquisition because they can lead to diferent linguistic outcomes, and all of them relate to input in one fashion or another. Tese include: • the timing of acquisition of the second language, • the patterns of interaction in one language versus the other, and • the amount of input in one language relative to another.

84

From minutes to years

In monolingual language acquisition, we assume that children are exposed to only one language and that all their social interactions take place in this language. As a result, all the interactions they have with others are in the same language. Bilingual children, on the other hand, may acquire one language before another, so the timing of the input for one of their languages may be later than that for the other. Additionally, the input from one language to a bilingual child may come from a diferent set of interlocutors than the input in another; thus the patterns of interaction that a bilingual child experiences can vary in signifcant ways. Terefore, in bilingual acquisition, who addresses a child may indeed matter. And, fnally, bilingual children may experience less input in one language than the other, leading to one language becoming more dominant than another. Of these factors, the one that attracts the most attention from researchers concerns the timing of acquisition. Of concern is the “age of onset of bilingualism.” Tis specifes the point in time when an individual begins to receive continual exposure to a second language. Although linguistic experience is multifaceted and variable, we can distinguish three broad categories of bilingual profles: • “Early simultaneous bilinguals” are children who are exposed to two languages at the same time before the age of three, even if to diferent degrees. Such individuals can also be called “bilingual frst language acquirers.” • “Early sequential bilinguals” are children who are mainly exposed to one language at home and learn another language later, ofen when they enter preschool or elementary school. • “Late bilinguals” are those who acquire a second language afer the age of 12. We will discuss this type of bilingual, also called a “second language learner,” in the next section. Te timing of bilingual acquisition is important to cognitive psychologists, who claim that the mechanisms that allow children to acquire two complete linguistic systems quickly and efortlessly are due to special biological constraints. Tey propose that there is a “sensitive period,” a window of time in which young children must receive linguistic input in both languages to fully acquire the structures that monolingual speakers of those languages possess. Tis is generally thought to be before the age of 3. In this view, if individuals are exposed to a language afer the age of 3, they will not acquire all its structural subtleties, and they may demonstrate diferences from monolinguals and from simultaneous bilinguals. However, keep in mind that the linguistic diferences between bilingual frst language acquirers and early sequential bilinguals can be so small that they require careful study and analysis for detection. While there is a sensitive period proposed for early language acquisition, psychologists have also proposed that there is also an endpoint for the successful acquisition of a second language. Tis is called the “critical period,” which ends with puberty, a key moment in the development of the brain (see Johnson & Newport, 1989, among others). Te claim of a critical period is controversial, and other researchers consider that input and language experience are more relevant than the age of acquisition

Acquisition and attrition

85

(MacWhinney, 2018). Pliatsikas (2019), for instance, found no signifcant diferences in the stages of neuroplasticity among infants and children as compared to adults, so there is little basis to believe that your ability to acquire a language is determined primarily by your brain. Tis does not mean that there are no diferences in brain structure and connectivity depending on when you’ve learned your language. But those diferences can be accounted for by an individual’s linguistic experiences rather than by the timing in which they acquired their languages. Te second factor to consider in bilingual acquisition is how a child becomes bilingual. Here, we need to think about the pattern of interactions in diferent languages that an early bilingual might experience. Some children will become bilingual within their home when they have family members, or caregivers, who speak to them using diferent languages. One common interactional pattern is known as “one parent–one language (OPOL),” and it can be practiced anytime a child’s parents come from different language backgrounds. In the OPOL model, each of the child’s parents uses a diferent language to address the child, who is continually exposed to two languages from birth. In some cultures, OPOL is a required marriage pattern. Te most famous cases occur among communities of the northwest Amazon, where members of the Eastern Tukanoan and Arawakan language groups are required to marry outside their own linguistic clan. Tis is a cultural practice known as “linguistic exogamy.” In the northwest Amazon, it has resulted in widespread bilingualism in individuals and structural convergence between languages, a topic we will return to in Chapter 9 when we discuss “linguistic areas.” In another common scenario, one or more family members may communicate in a language within the home that difers from the language of the community around them. Tis is a frequent outcome when families relocate from their region of origin to a new area, whether for reasons of immigration, work, or traumatic social upheavals, like war or environmental devastation in their homelands. Tis can lead to a shif in language dominance in a child if the domain of use for the ancestral language becomes restricted to the home when the child begins to communicate in the majority language outside the home. As an individual’s exposure to and use of their home language diminishes over time, a partial loss, or “attrition,” of the minority language can set in, a topic we will return to in Section 6.3. Of the three factors that we have identifed as integral to the profle of a bilingual, the relative amount of input in the two languages is the most variable. Children differ in how, and how quickly, they acquire various grammatical aspects of language, and these diferences in their patterns of language acquisition appear to be related to the input they receive. Even some monolingual children use certain vocabulary words earlier than others do, while other children use more complex grammar but fewer words than other children (Frank et al., 2021). Hence, the exposure to two languages from an early age can afect a child’s pattern of language acquisition. Bilingual children need to process more variable linguistic input overall than monolinguals do because the input is split between their two languages, ofen to diferent degrees. Tis makes the results of the bilingual acquisition process fundamentally diferent from that of monolingual acquisition, and it has consequences on language development.

86

From minutes to years

As an example, bilingual frst language acquirers are exposed to two diferent linguistic sound systems even before birth. Abboub and colleagues (2016) demonstrate that prosody plays a key role for newborns from monolingual and from bilingual backgrounds based on their prenatal experience with their frst languages. From within their mothers’ wombs, prenatal babies cannot clearly hear the individual sounds of speech because they are mufed by the tissue of the womb, but they can perceive the rhythm and intonation of voices outside the womb. Te researchers showed that newborns who were exposed to French prosody in their prenatal environment tended to respond to rhythmic grouping based on the length of syllables rather than to changes to pitch, mirroring the way French rhythmic groups are formed. Babies who had prenatal exposure to French and to additional languages that use pitch, rather than length, diferences for rhythmic grouping were sensitive to both length and pitch when they were newborns. Tis means that simultaneous bilinguals, even from a very early age, are tuning in to diferent sets of perceptual cues, and they are acquiring them both. In addition, newborns hear variation in the sounds that are spoken around them. From this soup of sound, they need to pick out the distinctive phonemes of their language(s). Remember that these are the sounds of a language that make words distinct, like the [ɪ] in sit [sɪt] and the /i/ in seat [sit]. In many languages, these two vowels are not perceived to difer, so speakers from these languages will confuse these words when learning English.When hearing babies are born, they can detect phonetic diferences, like [ɪ] vs. [i], that could potentially occur in any language. However, they quickly lose this detection ability if it is not needed once they have zeroed in on the sounds that are distinct in their language(s). But it appears that infants who receive input in two or more languages from birth can retain this ability for several months longer than infants who receive input in a single language (Garcia-Sierra et al., 2011). Tus, the need to narrow down the meaningful sounds of two systems, rather than one, extends this discrimination timeline for bilinguals. A noticeable way in which bilingual acquisition difers from monolingual acquisition is that,in the beginning stages of early language acquisition,bilingual children mix their languages when they speak (Chapter 4). Tis sparked a theory, known as the “unitary system hypothesis,” that bilingual children start language acquisition with a single system. Volterra and Taeschner (1978) were among the frst to argue for this hypothesis. Tey conducted a longitudinal study of two German-Italian children, aged 1;05–3;06 and 1;02–2;06, who were brought up following the OPOL model. Teir mother always spoke German to them and their father Italian. Initially, the girls would use words from both languages with each parent, like in example 6.1, where Lisa uses the Italian là “there,” rather than the German da, in response to her mother’s question in German. (6.1) Lisa (1;10) to her mother (Volterra & Taeschner, 1978, p. 307) M: Wo ist miao? “Where is meow?” L: Là miao. “Tere meow.”

Acquisition and attrition

87

From evidence like this, the authors suggested that children start with a single vocabulary that includes words from both languages. At a later stage in their development, they develop separate vocabularies and, at a third stage, separate grammars for their two languages. It is problematic that the evidence supporting the unitary system hypothesis most ofen comes from children’s interactions with their parents in OPOL homes. Even if each parent only addresses their child in one language, the children are still exposed to a bilingual environment, where the use of both languages is acceptable and expected (Quick et al., 2021). In other words, at home they are constantly in a bilingual mode (Chapter 4). And we can’t rule out the possibility that their mixing refects an awareness that both their parents will understand them, no matter which language they use! It is also possible that because bilingual children are not always equally dominant in their two languages, they mix languages to exploit all the linguistic resources that are available to them (Genesee et al., 1995). While there is very little evidence to suggest that children confuse their languages or that they begin with a single linguistic system, there is increasing evidence that early simultaneous bilinguals have separate, diferentiated language systems from the start. Tis is known as the “dual system hypothesis” (Meisel, 2004). Tis view is supported by the fndings that bilingual frst language acquirers reach the same milestones of acquisition as monolingual children do and at roughly the same time (De Houwer, 1990; Meisel, 2004; Bosch & Sebastian-Gallés, 2001). And as soon as they reach the stage where they are combining more than one word in their speech into phrases, usually around age 2, they show evidence that they have acquired the word order and morphological infections of their respective languages. And, as we have seen in Chapter 4, young bilingual children are also sensitive to the context of language use and are able to calibrate their language use to match that of the language of their interlocutors (Lanza, 1997; Genesee et al., 1995). All these facts suggest that bilingual children do not have a single fused language system. Mixing aside, bilingual children still do not always use language like monolingual children. Troughout this book, we have seen, and will continue to see, that bilinguals can perform in various language modes according to their interactions. Tey can activate both languages at the same time in a bilingual language mode. So, bilingual children will show signs of cross-linguistic interaction in their language use. Examples of this interaction are shown in 6.2. In (a), a Spanish-English bilingual child was overheard using a lexical calque from English. Te equivalent Spanish expression for barefoot would be the adjective descalzo (masc) or descalza (fem) and, here, the child translates literally from English, where the expression to go with bare feet is common in some varieties. (6.2) Cross-linguistic transfer in children (a) English to Spanish: (Austin, 2020, p. 202) ¿Puedo salir con pies de oso? English: “Can (I) go out with bear (bare) feet?”

88

From minutes to years (b) Cantonese to English (age 3;04) English: Te snail why live in water? Cantonese: Zek3 wo1ngau4 dim2gaai2 zyu6 hai2 seoi2 dou6 ? CL snail why live in water there

Te example in (b) shows the syntactic transfer of a question formation strategy from Cantonese to English in a child aged 3;04 in 6.2b (Yip & Matthews, 2000, p. 198). In English, speakers use question words at the beginning of a sentence and invert the subject and auxiliary to form questions (why does-AUX the snail), while in Cantonese, speakers do not front the question word or use an auxiliary verb. Here, the child’s structure in English is a replica of the pattern of his dominant language, Cantonese. Tere are many possible reasons that children show these types of linguistic interactions in their language, including the quantity and quality of input they receive, the opportunity to participate in conversations in each of the languages, and the high frequency of some constructions in one language compared to the other. If they receive continued input and use in both their languages, the mixing that bilingual children show in early life will likely disappear. However, not all children who are exposed to language early in life have balanced input or usage of their two languages. Some hear and use one far more than the other. Tis becomes their dominant language. When one language is dominant in a bilingual, mixing and transfer can be persistent features in their less dominant language. As we will see in Chapter 11, this is a frequent pattern in the minority language within foreign-born and heritage language users, and such features can become anchored in minority language communities. We should point out that bilinguals should not be expected to have linguistic systems that are exactly like those of monolingual users of their languages. Even those who acquired both languages from birth and who have remained balanced through their lives show signs of behavior that is diferent from monolinguals of their languages. Tis is because the languages of bilinguals constantly interact (Kroll & Stewart, 1994). Interaction between languages of balanced bilinguals can occur in ways that might not be immediately noticeable but can be discovered through experiments. In one study, adult Belgian Dutch-French bilinguals, who had acquired each language in an OPOL home and maintained both languages in a balanced way all their lives, were asked to name pictures from two given sets of objects (Ameel et al., 2005, 2009). Te two sets of objects were household items, belonging to a “bottles set” or a “dishes set.” Te linguistic boundaries of words across languages do not always match. Te researcher wanted to fnd out if bilinguals would show linguistic boundaries for common objects that difered from those of monolinguals of their languages. Te study participants completed the naming task in Belgian Dutch and in French. Items from the bottle set were the Belgian Dutch and French equivalents of words like bottles, jug, jars, or containers, and objects from the dishes set included things that we might describe as dishes, plates, mugs, bowls, or ashtrays. What the researchers found is that the bilinguals provided a single mapping between an object and its name across their languages, showing fewer, merged linguistic boundaries for object names than monolinguals. So, language interaction in infancy can have long-term efects on a bilinguals’ language representations, regardless of when they acquired

Acquisition and attrition

89

their languages. Tis happens in more balanced bilinguals at all linguistic levels: phonology and phonetics (Simonet, 2010), morphology (Balaguer et al., 2005) syntax (van Gompel & Arai, 2018), and discourse (Matras, 2009/2020).

6.2

SECOND LANGUAGE ACQUISITION

It is ofen remarked that children who are exposed to two languages from an early age, whether simultaneously or sequentially, acquire them efortlessly. Acquiring a language later in life can be a very diferent matter. Early bilingual acquisition takes place in natural settings like the home or the playground. Here, interactions are varied, informal, and spontaneous. For some individuals, second language acquisition can be naturalistic, but it can also begin in a more formal setting, like a classroom where the language input is less varied, more formal, and sometimes written rather than spontaneous. Te outcomes of naturalistic versus instructed learning can be quite diferent. When someone begins to learn a second language as an adolescent or later, they have already been exposed to massive amounts of monolingual input. From this input, they have developed a fully formed vocabulary and grammar. Tey have also learned how to use their frst language for social and pragmatic functions like projecting their identity, showing their attitude about what they are communicating, or trying to achieve certain ends in a conversation. It is not surprising, then, that as they learn the second language (L2), the words, meanings, and structures that they use to express something in the L2 will be fltered through their dominant frst language (L1).“Transfer,” also called cross-linguistic infuence, describes the situation in which language users apply their knowledge of features of their L1 when producing, or perceiving, their L2 (Odlin, 2003). When you hear someone speaking your language as a second language, you can sometimes tell from their pronunciation that they have acquired it later in life. You can also guess what their frst language is likely to be. For instance, a Korean pronunciation would sound diferent from a Nahuatl pronunciation. So, one of the most obvious signs of L1 transfer in a second language learner is a diferent pronunciation. Te sound system of an L2 can be difcult to acquire later in life because transfer afects a learner’s perception, as well as their production. For instance, if your L1 is a non-tonal language and you are trying to acquire a tonal language, like Vietnamese, you would need to learn to perceive one pitch as meaningfully diferent from another so that you don’t confuse Vietnamese words with similar sounds but diferent tones, like huống “situation” and hướng “direction.” If your L1 does not use pitch in this way, these diferences can be difcult for you to perceive and to reproduce. Another complexity in perceiving the sounds system of an L2 is that the second language might have sounds that are phonologically similar, but not phonetically identical, to those of the L1. Tis means that the phonetic cues for certain sound differences can be diferent across languages. For example, some languages, like Spanish or Polish, have two series of obstruent consonants that are distinguished by the absence or presence of vocal fold vibrations. Tere is a voiceless series of stop consonants, /p,t,k/, made without vocal fold vibration and a voiced one, /b,d,g/, with vocal fold vibration. English has two series of consonants, too, but the diference between

90

From minutes to years

them is not one of voicing at the phonetic level. Instead, it is aspiration, the puf of air following the release of the closure of our articulators in the oral cavity when we pronounce a stop consonant. Tis is known as voice onset time, or VOT. VOT is measured as the delay between the release of the consonant and the onset of voicing in the vowel. In English, /p,t,k/ has a long-lag voice onset time, while /b,d,g/ has a short-lag VOT. So, an L1 English speaker is likely to perceive an equivalence between their aspirated stops /ph/, with a long lag, and the voiceless unaspirated stop in Polish /p/. As a result, they will show longer lag times in their speech than Polish speakers generally do. Tis will cause the L2 speaker to have a perceptible “accent,” that is, a pronunciation that is perceived by L1 speakers as diferent from their own. L1 transfer can be positive or negative, and it can afect all aspects of linguistic structure. In positive transfer, a structure of the L1 and the L2 overlaps fully or partially between the languages, and this can facilitate the acquisition of an L2 feature. For instance, there are many cognate words between English and other Indo-European languages, which facilitates word learning in languages like Portuguese or German, but not Arabic or Korean, for L1 English speakers. Negative transfer refers to cases in which the languages diverge in structure. For instance, in the example in 6.3, an L1 Chinese speaker leaves out the preposition with that is required to introduce an argument of the verb help in English because the translation equivalent of the verb in Chinese does not require a preposition (Li Wei & Liu, 2017): (6.3) Chinese L1–English L2 transfer Today he help dinner.

Similar examples of transfer abound in the grammars of second language acquirers who do not have access to an equivalent degree of input in the L2 as L1 users of the language do. Consequently, learners in the early stages of their acquisition process lack the linguistic resources of more profcient users. To compensate, an L2 learner might resort to the strategy of using syntactic calques, employing the words and structures of their L1 while using lexical items from their L2. Since words that share similar meanings between languages do not necessarily share the same grammatical requirements, this can result in sentences in the L2 that do not align with L1 users’ productions, as in 6.3. Just as pronunciation is an indication that someone has acquired their L2 later in life, the words and structures used by bilinguals in their L2 can provide clues to their L1s. Tis can be a matter of degree, where an L2 user tends to overuse or underuse a word or sequence relative to an L1 user. For instance, learners tend to always use the sequence I think in English rather than occasionally using an alternative like in my opinion. Tere can also be qualitative diferences in how L2 learners use words. For instance, they may use a mass noun, which normally cannot be counted, as if it were countable to produce an odd expression for L1 English speakers: I really like literatures. Te exact features that are likely to be transferred by individual learners of a given L1–L2 confguration is an open question in second language acquisition, but there is evidence that bilingual raters can use linguistic features to detect the L1 of an author

Acquisition and attrition

91

writing in their L2 at a level greater than chance. Tis is especially the case when the L1 of the rater and the writer match. In one study, Spanish raters were able to determine that the writer of an English document was an L1 Spanish speaker rather than a speaker of Czech, German, or Finnish with accuracy rates up to 92.5% (Jarvis et al., 2019). Machine models, called “native language identifcation” (NLI) systems, have also been shown to accurately classify the L1 of an L2 writer with levels greater than chance. Using data from the Chinese Learner Corpus, Malmasi and Dras (2014) created NLI models using structural features, like part-of-speech tags, function words, and simple grammatical parses of sentences, to identify the L1 of writers from 11 different languages, most from other Asian countries. Tat NLI models and human raters can accurately classify the L1 of L2 writers shows that transfer is pervasive in second language users, even among individual second language learners with advanced writing profciency. While certain transfer features may be subtle in the L2, there appear to be transfer properties or even clusters of properties that recur among individuals who use the L2 who share the same L1. Te outcomes of second language acquisition are dependent on a complex set of social, emotional, cognitive, developmental, contextual, and linguistic variables. It is undeniable that continued exposure to and use of the second language can lead to such advanced levels of profciency in some late bilinguals that they may be indistinguishable in their daily interactions from frst language acquirers of the same language. But what about their frst language? Is it impacted by the use of the second language? We discuss this in the next section.

6.3 ATTRITION OF THE FIRST LANGUAGE Some people will acquire a second language, either early or later in life, and then use it as much as, or more than, their L1. If the L2 becomes their dominant language, they may experience loss in some aspects of their L1. Tis is referred to as “language attrition.” We can think of language attrition as the reverse of language acquisition. In language acquisition, profciency increases with language exposure and use. In attrition, it decreases with lack of exposure and use. Language attrition can occur at the level of the bilingual individual who shifs away from their L1 to their L2. It can also occur at the societal level when the individuals in the younger generations of a community become dominant, or even monolingual, in their L2. We will address societal level shif and attrition in minority, heritage, and immigrant language communities in Chapter 11. L2 infuence on the L1 does not necessarily lead to language loss. Imagine this scenario: You visit a country for several months to immerse yourself in a language you have studied. By the time you leave, you have used your L2 much more frequently than your L1. And when you return home from your travels, some words from your L2 continue to pop up unbidden in your head when you are trying to say them in your L1. In this scenario, you have experienced a mild, and temporary, form of frst language attrition. Tis temporary forgetting process is helpful; the suppression of our L1 helps the mind deal with the learning task at hand, which is communicating in our L2. When we learn a second language, we need to exert cognitive efort to inhibit the

92

From minutes to years

words of our frst language because these will be the frst to come to mind (Gülsen & Schmid, 2019). Any changes to the L1 of bilingual users under the infuence of an L2 can sometimes be classifed by researchers as language attrition (Schmid, 2009; Schmid & Köpke, 2017). However, it is preferable to call this L2 to L1 transfer (Cook, 2003). L2 infuence on L1 categories can occur very quickly in the L2 acquisition process, even in novice adult learners (Chang, 2012). Transfer does not imply loss or forgetting. It refects the type of cross-linguistic infuence in the mind of a bilingual that we have seen throughout this book. And we must remember that not every change in an individual’s L1 over time is a sign of attrition. For instance, we wouldn’t say that Queen Elizabeth II was sufering attrition in English because she had shifed her vowels toward those of younger speakers. Instead, we view this as a form of communication accommodation. Like acquisition, attrition can correlate with age. Te time frame when a language shif from an L1 to an L2 takes place in an individual is critical, perhaps in relation to the biological constraints on language acquisition that we introduced in Section 6.1. It is well known that the language of adults is much more resistant to attrition than the language of children, so that adults who lose contact with their L1 still retain their profciency and can reactivate it when re-exposed to it. Children, though, can show more complete attrition.A study of Japanese children who were aged 3;10, 5;10, and 7;00 when they lef a Japanese language environment showed that the two younger children had a rapid and dramatic loss of Japanese syntax, while the older one did not (Yukawa, 1997). Tis suggests that a language that is more fully acquired is more resistant to loss. Additionally, those who continue to use their L1 frequently, no matter how young they are when they begin to use their L2, showed less attrition than those who do not (Montrul, 2008; Schmid et al., 2013). So, attrition is not the automatic consequence of a strong, or dominant, L2 in a bilingual even if increased L2 profciency is likely to bring about some restructuring of the L1. It is unclear whether bilingual people who have no further exposure to or use of their L1 afer a certain time can show complete and unrecoverable loss of their L1 or if traces of it remain. Research on Korean children who were adopted internationally shows mixed results, with some studies demonstrating that the children, now adults, have experienced complete loss (Pallier et al., 1997; Ventureyra et al., 2004) and others showing that traces of the L1 remain in the adoptees (Hylenstam et al., 2009). Te latter results suggest that knowledge of a language may not be entirely lost, merely suppressed, and a language user can reactivate it within a stimulating environment. Pierce (2015, p. ix) studied the morpho-syntactic and phonological systems of Chinese-born children adopted into French- and English-speaking homes before the age of 3 to conclude that “even brief experiences with a language early in life have a lasting impact on the brain.”

6.4 HOW TO STUDY BILINGUAL CHILD ACQUISITION AND ATTRITION OF THE FIRST LANGUAGE If you are interested in studying bilingual child acquisition, you can draw on the many methods available in the study of language acquisition in general. Ambridge and Rowland (2013) ofer an excellent review.

Acquisition and attrition

93

One good way to study language acquisition is by observing real-life behavior.You can study spontaneous language by recording it (nowadays there are light recorders that the child can wear during the day; see Pisani et al., 2021, for a review of available hardware). Less intrusive methods are those that elicit language production in semi-experimental tasks or ask children to repeat something.An old technique that is not intrusive yet allows accessing the child’s everyday language is to ask parents to write down the child’s new words and expressions in a diary.Also, even if it requires good memory, researchers ofen ask parents to fll in a report with the list of words and expressions that their child uses. You can always study comprehension by examining what children point to and what infants look at while listening to audio stimuli or where they turn their head to. Finally, you can use acceptability judgment tasks or truth value judgments. Having infants and children collaborate with you in a study is not an easy task: Tey may be shy, tired, or just bored with the task that you took so much time to prepare. So, you should look for ways to motivate your little participants and choose an age-appropriate task that makes communicative sense. And like in any experiment (see Section 1.4), don’t forget to use fllers, think of an appropriate number of trials and participants, have control conditions, and use an appropriate statistical test. To study L1 attrition, the best is to study language production (Schmid et al., 2013). You can either record conversations and narratives or conduct semi-structured interviews. You can elicit narratives through flm retelling (e.g., a sequence from Charlie Chaplin’s silent flm Modern Times). Your goal could be to analyze somewhere between 30 to 100 minutes per language user. To identify signs of L1 attrition, you can count words per minute, measure pauses, evaluate lexical diversity, and annotate disfuencies in the speakers’ speech, for instance, when there are repetitions and self-corrections. In L1 attrition studies, as in other studies of bilingualism, you can compare the performance of the bilingual participants with the performance of a monolingual control group (but don’t forget to match the two groups in age, gender, and education!) Alternatively, you can compare the productions of various bilingual users from the same or a diferent language community.

Chapter summary Bilingual experience is malleable: When we use our languages and with whom we use them are factors that can change across our lifetimes. Considerations of language experience, of how and when we acquire a new language in addition to our frst, and of how our abilities in our frst language might alter, or even fade, with the addition of other languages are central to the study of bilingualism. Te evidence considered in this chapter suggests, once again, that the languages of a bilingual continually interact, even when an individual no longer accesses one of their languages. Te language acquisition of an L2 and the language attrition of an L1 within an individual bilingual introduces variation into a linguistic system that can bring about language change in a contact setting.

94

From minutes to years

EXERCISES 1.

Research essay

Te “nature vs. nurture” debate is a classic debate about how we become who we are. Is it mainly due to our genes? Is it mainly due to our experiences and the way we grew up? Discuss how the “nature vs. nurture” debate is represented in child language acquisition studies.

2. Test yourself Test yourself and see how many words related to animals you can name in 30 seconds in two languages you know. You’ll fnd out that maybe you know more words in one language than in the other or that the words just don’t come to mind as easily in one language as they do in the other.

3.

Discover

Take pictures of three or more objects that you would call cups and three or more objects that you would call “bowls.” Vary the shapes, sizes, and material that they are made from.Ask individuals whose frst language is the same as yours how they would name each object. Allow them to name the objects with any label they want (“cup,” “bowl,” “mug,” “glass”). Now ask individuals whose frst language is diferent from yours how they would name each object in English. Do you notice any diferences? Compare your results with others.

FURTHER READING Genesee, F., & Nicoladis, E. (2007). Bilingual frst language acquisition. In E. Hof, & M. Shatz (Eds.), Blackwell handbook of language development (pp. 324–342). Blackwell Publishing. Tese authors discuss the psychological, social, and interactional aspects of the phenomenon of children acquiring two languages from birth. Odlin, T. (2003). Cross-linguistic infuence. In C. J. Doughty & M. H. Long (Eds.), Te handbook of second language acquisition (pp. 436–486). Blackwell. Tis classic chapter overviews various types of language transfer in language learning.

REFERENCES Abboub, N., Nazzi, T., & Gervain, J. (2016). Prosodic grouping at birth. Brain and Language, 162, 46–59. https://doi.org/10.1016/j.bandl.2016.08.002 Ambridge, B., & Rowland, C. F. (2013). Experimental methods in studying child language acquisition. Wiley Interdisciplinary Reviews. Cognitive Science, 4(2), 149–168. Ameel, E., Malt, B. C., Storms, G., & Van Assche, F. (2009). Semantic convergence in the bilingual lexicon. Journal of Memory and Language, 60(2), 270–290.

Acquisition and attrition

95

Ameel, E., Storms, G., Malt, B. C., & Sloman, S. A. (2005). How bilinguals solve the naming problem. Journal of Memory and Language, 53(1), 60–80. Austin, J. (2020). Cross-language contact in the developing grammars of bilingual children. In E. Adamou, & Y. Matras (Eds.), Te Routledge handbook of language contact (pp. 201–220). Routledge. Balaguer, R. D. D., Sebastián-Gallés, N., Díaz, B., & Rodríguez-Fornells, A. (2005). Morphological processing in early bilinguals: An ERP study of regular and irregular verb processing. Cognitive Brain Research, 25(1), 312–327. Bosch, L., & Sebastián-Gallés, N. (2001). Evidence of early language discrimination abilities in infants from bilingual environments. Infancy, 2(1), 29–49. Chang, C. B. (2012). Rapid and multifaceted efects of second-language learning on frst-language speech production. Journal of Phonetics, 40(2), 249–268. Chomsky, N. (2006). Language and mind. Cambridge University Press. Cook, V. (Ed.). (2003). Efects of the second language on the frst. Multilingual Matters. De Houwer,A. (1990). Te acquisition of two languages from birth: A case study. Cambridge University Press. Frank, M. C., Braginsky, M.,Yurovsky, D., & Marchman,V.A. (2021). Variability and consistency in early language learning: Te Wordbank project. MIT Press. Garcia-Sierra,A., Rivera-Gaxiola, M., Percaccio, C. R., Conboy, B. T., Romo, H., Klarman, L., Ortiz, S., & Kuhl, P. K. (2011). Bilingual language learning: An ERP study relating early brain responses to speech, language input, and later word production. Journal of Phonetics, 39, 546–557. https://doi. org/10.1016/j.wocn.2011.07.002 Genesee, F., Nicoladis, E., & Paradis, J. (1995). Language diferentiation in early bilingual development. Journal of Child Language, 22(3), 611–631. Gülsen, Y., & Schmid, M. (2019). First language attrition and contact linguistics. In J. Darquennes, J. C. Salmons, & W. Vandenbussche (Eds.), Language contact: An international handbook (pp.  198–209). Mouton de Gruyter. Hylenstam, K., Bylund, E., Abrahamsson, N., & Park, H. S. (2009). Dominant-language replacement: Te case of international adoptees. Bilingualism: Language and Cognition, 12(2), 121–140. Jarvis, S., Alonso Alonso, R., & Crossley, S. (2019). Native language identifcation by human judges. In M. Gutierrez-Mangado, M. Martínez-Adrián, & F. Gallardo-del-Puerto (Eds.), Cross-linguistic infuence: From empirical evidence to classroom practice (pp. 215–231). Springer. Johnson, J. S., & Newport, E. L. (1989). Critical period efects in second language learning: Te infuence of maturational state on the acquisition of English as a second language. Cognitive Psychology, 21(1), 60–99. Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33(2), 149–174. Lanza, E. (1997). Language contact in bilingual two-year-olds and code-switching: Language encounters of a diferent kind? International Journal of Bilingualism, 1(2), 135–162. Li Wei, & Liu, X. (2017). Te bilingual mental lexicon and language transfer in second language learning. International Journal of Language and Linguistics, 4(4), 11–23. Loukatou, G., Scaf, C., Demuth, K., Cristia, A., & Havron, N. (2021). Child-directed and overheard input from diferent speakers in two distinct cultures. Journal of Child Language, 1–20. https://doi. org/10.1017/S0305000921000623 MacWhinney, B. (2018). A unifed model of frst and second language learning. In M. Hickmann, E. Veneziano, & H. Jisa (Eds.), Sources of variation in frst language acquisition: Languages, contexts, and learners (pp. 287–312). John Benjamins.

96

From minutes to years

Malmasi, S., & Dras, M. (2014, April). Chinese native language identifcation. Proceedings of the 14th conference of the European chapter of the association for computational linguistics (2): Short papers (pp. 95–99). Association for Computational Linguistics. Matras, Y. (2009/2020). Language contact. Cambridge University Press. Meisel, J. M. (2004). Te bilingual child. In T. K. Bhatia, & W. C. Ritchie (Eds.), Te handbook of bilingualism (pp. 91–113). Blackwell Publishing. Montrul, S.A. (2008). Incomplete acquisition in bilingualism. John Benjamins. https://doi.org/10.1075/ sibil.39 Odlin, T. (2003). Cross‐linguistic infuence. In C. J. Doughty, & M. H. Long (Eds.), Te handbook of second language acquisition (pp. 436–486). Wiley Online Library. Pallier, C., Bosch, L., & Sebastián-Gallés, N. (1997). A limit on behavioral plasticity in speech perception. Cognition, 64(3), B9–B17. Pierce, L. J. (2015). Te consequences of early language experience: Neurocognitive evidence from international adoptees [PhD thesis, McGill University]. Pisani, S., Gautheron, L., & Cristia, A. (2021). Long-form recordings: From A to Z. https://bookdown. org/alecristia/exelang-book/ Pliatsikas, C. (2019). Understanding structural plasticity in the bilingual brain: Te dynamic restructuring model. Bilingualism: Language and Cognition, 6, 1–13. https://doi.org/10.1017/ S1366728919000130 Quick, A. E., Hartmann, S., Backus, A., & Lieven, E. (2021). Entrenchment and productivity: Te role of input in the code-mixing of a German-English bilingual child. Cognitive Linguistics, 32(2), 319–328. Schmid, M. S. (2009). Language attrition across the lifespan. In K. de Bot, S. Makoni, & R. Schrauf (Eds.), Language development over the life-span (pp. 171–188). Lawrence Erlbaum Associates. Schmid, M. S., & Köpke, B. (2017). Te relevance of frst language attrition to theories of bilingual development. Linguistic Approaches to Bilingualism, 7(6), 637–667. Schmid, M. S., Köpke, B., & de Bot, K. (2013). Language attrition as a complex, non-linear development.International Journal of Bilingualism,17, 675–682. https://doi.org/10.1177/1367006912454619 Simonet, M. (2010). Dark and clear laterals in Catalan and Spanish: Interaction of phonetic categories in early bilinguals. Journal of Phonetics, 38(4), 663–678. Tomasello, M. (2003). Construing a language: A usage-based theory of language acquisition. Harvard University Press. van Gompel, R. P., & Arai, M. (2018). Structural priming in bilinguals. Bilingualism: Language and Cognition, 21(3), 448–455. Ventureyra, V. A., Pallier, C., & Yoo, H. Y. (2004). Te loss of frst language phonetic perception in adopted Koreans. Journal of Neurolinguistics, 17(1), 79–91. Volterra, V., & Taeschner, T. (1978). Te acquisition and development of language by bilingual children. Journal of Child Language, 5(2), 311–326. Yip,V., & Matthews, S. (2000). Syntactic transfer in a Cantonese–English bilingual child. Bilingualism: Language and Cognition, 3(3), 193–208. Yukawa, E. (1997). L1 Japanese attrition and regaining: Tree case studies of two early bilingual children [PhD thesis, Stockholm University].

7 Language ideologies and dispositions

CHAPTER PREVIEW While in the classic sense, “language policy” refers to the set of activities that the state develops to promote and regulate the status of languages, it is more recently viewed as a combination of individual and group behaviors and beliefs about language that can also impinge on language use. From this perspective, not only the study of language structure and language use but also the study of language attitudes; and language ideologies become key to understanding the choices that language users make in their everyday lives. Te study of language attitudes and ideologies is critical in appreciating processes such as language shif and elucidating the role of language in understandings of unequal power relations. Te present chapter, then, is devoted to an exploration of the ways in which language users assess and invoke language in claiming or repudiating social identities and of the factors that imbue some languages with social capital and position certain varieties as standard models to be endorsed over others.

7.1

LANGUAGE ATTITUDES AND LANGUAGE IDEOLOGIES

As we learned in the preceding chapters, our observations of language can disclose abundant information about linguistic structure, the limits of linguistic variation, and the logics of language choice.We noted that language users can produce and perceive subtle diferences that typify regional, social, and ethnic varieties while at once recognizing that these are variants of the same language, distinguishable from varieties of other languages. For instance, British Received Pronunciation and Singapore English, though distinct, are recognized as varieties of English as opposed to being varieties of Welsh or Chinese. Likewise, we remarked that multilingual language users accommodate towards or away from their interlocutors along set parameters such as language selection, and they can defly mix the languages in their repertoires without defying grammatical principles when the communicative context allows. In brief, our observations of language are informative of the systematicity and rule-governed nature of language and language usage. We also learned that observations of our habitual language practices can ofer insights concerning the norms that regulate language usage by individuals within communities. Tus, as a member of the community of linguistic scholars, you would DOI: 10.4324/9781003167952-9

98

From minutes to years

likely be compelled to draw on highly specialized terminology, careful pronunciation, and formal grammar in delivering a lecture to your professional peers, and you would expect your audience to sit quietly until your language or bodily gestures signaled an ending to your lecture. But you would undoubtedly switch from the academic standards to a local vernacular that is slang-laden, unguarded, fast, and fragmented for a neighborhood game of pick-up football with your friends—like trading loafers for cleats. Tis is because in our exchanges, we follow communal conventions of appropriate behavior; doing otherwise would be censured by interactants. Te central point of interest for the present chapter is the fact that our behaviors are also mediated by our dispositions and beliefs about language(s)—whether language in general, our specifc language(s), or others’ language(s). In linguistics and allied social and behavioral sciences, the relevant conceptual tools are language attitudes and language ideologies. Te term “language attitudes” is used to refer to patterns of evaluative reactions towards languages or linguistic forms. Te aptly titled chapter“Italian Is Beautiful, German Is Ugly,” from the compendium Language Myths (Bauer & Trudgill, 1998) serves as a ftting example of commonly attested reactions among English speakers in the United States: Italian conjures connotations of elegance and sophistication, whereas German evokes connotations of harshness and austerity. Research on attitudes has also examined evaluations toward varieties of the same language: Parisian French is more positively evaluated than Canadian French, and non-standard varieties of English, such as London Cockney English and African American Vernacular English, are considered less attractive than those varieties that adhere to codifed standards with respect to grammar, pronunciation, and vocabulary. Te term “language ideologies,” in turn, is used to refer to the collective beliefs that shape these (dis)inclinations towards languages or varieties. By way of example, the ideology of “linguistic purism” ignores natural language change; it can lead to stigmatization of common hybrid language practices such as code-switching, and it depicts contact-induced changes as forms of linguistic or cultural defciency. As described by Dragojevic (2017), standard varieties are those that adhere to codifed norms defning correct usage, whereas non-standard varieties are those that depart from such norms in some manner (e.g., pronunciation). Standard and nonstandard varieties elicit diferent evaluative reactions, as we will see. Because standard varieties tend to be associated with dominant socioeconomic groups within a given society, their language users are typically attributed more status than non-standard language users. However, non-standard varieties are valued on dimensions of solidarity, loyalty, and in-group identity—we stand by our peer groups! Language ideologies are imbued with political, moral, and social values and, as such, they are shaped within socio-historical and cultural settings. Language ideologies are pervasive; they inform everything from scholarship (What varieties are suitable objects of study?), to policy (What variety is codifed and taught as standard language?), to linguistic landscape (What languages have political clout?), to language maintenance and second-language learning (What languages have cultural or instrumental value?). As should be clear, the concepts of language attitudes and language ideology implicate understandings of language—or linguistic subjectivity—as critically relevant in linguistic research. Linguistic subjectivity ofers insight into linguistic processes such as bilingual and second language acquisition and language shif. Tus, language ideologies and attitudes can regulate linguistic diversity.

Language ideologies and dispositions

99

Let’s delve further into scholarship on these core concepts. Language attitudes refect cognitive processes of social categorization and stereotyping (Garrett, 2010). As language users, we attune to linguistic cues (e.g., pronunciation, vocabulary) in inferring speakers’ social group memberships, and based on that categorization, we attribute to those speakers a set of stereotypic traits associated with those inferred groups. From the start,language attitude research focused on perceptions of languages as used by bilinguals, but this focus has expanded to the investigation of attitudes toward varieties of the same language.In studying language attitudes, researchers draw on direct and indirect methods. In the former, investigators administer questionnaires, surveys, polls, interviews, and rating scales that require participants to respond with their assessments of a particular language or language variety. For example, a participant might be asked:“How pleasant do you fnd Dutch?” An infuential project elicited evaluative data about dialects of Spanish from Cuban residents of Miami, Florida, arguably the most diverse Spanish-speaking metropolis in the world (Alfaraz, 2002, 2014).A questionnaire required the 148 respondents to rate on a 7-point scale their perceptions of the correctness and pleasantness of 18 Latin American Spanish varieties, Peninsular Spanish, and two varieties of Cuban Spanish—the Spanish spoken in Cuba before and afer the Cuban Revolution in 1959. Te overall correct and pleasant ratings correlated with the racial make-up and economic development of the countries in which the varieties are spoken. Te countries that are predominately White (e.g.,Spain) receive the highest ratings,followed by regions that are mostly mestizo (White and South or Central American Indigenous), and then the areas that are predominately indio (South or Central American Indigenous). Te region that is predominately Black, the Dominican Republic, is perceived as the country with the least correct or pleasant variety of Spanish. Te fndings further revealed that the variety of Spanish now spoken in Cuba (post-Revolution) is perceived as less correct and less pleasant than what was spoken before Castro assumed power, a perception that is founded in the political allegiances that shape Miami Cubans’ language ideologies. Individuals’ stereotypes about language can also be tapped indirectly via the matched guise or verbal guise technique. Tis method focuses the evaluators’ attention on the speakers of diferent language varieties rather than on the languages or varieties themselves. For instance, upon hearing a recorded sample of Cockney English speech, participants would be asked: How intelligent would you say this speaker is? In the seminal matched guise study carried out in Canada by psychologist Wallace Lambert and colleagues in 1960, participants were presented with recordings of English and French speech samples, afer which they were asked to rate the speakers on personal attributes, such as intelligence, patriotism, attractiveness, and friendliness. Te researchers found that there was a great disparity in the judges’ assessments of the speakers, which correlated with the prestige of the language variety being analyzed. Specifcally, the English guises were rated higher on measures that signal social status, while the French guises were rated higher on measures of solidarity or in-group loyalty. Notably, the listeners were unaware that they had rated the same bilingual individual in the English and French speech samples. Tat is, the same speaker evoked diferent reactions depending on the language being used. More recently, using the verbal guise technique, Kutlu (2020) found that when listeners saw the face of a South Asian woman, they would judge her speech as more “accented” whether they listened to an American English or Indian English sound fle. In contrast, when they saw the face of

100

From minutes to years

a White woman, they would judge her speech as less “accented” even if the sound fle was in Indian English. Tis exemplifes what has been called “accent hallucination.” Another study employing the matched guise technique investigated African Americans’ attitudes towards the use of African American Vernacular English, Standard English, and code-switching in a formal and casual setting; note that in this case, code-switching refers to African Americans’ ability to move between African American Vernacular and Standard English (Koch et al., 2001). A total of 102 university students were exposed to recordings of a man speaking in Black and Standard English, as well as code-switching “appropriately” (i.e., using Standard English in the setting of a job interview and African American Vernacular English in a casual discussion with a friend about the interview) or code-switching “inappropriately” (i.e., using African American Vernacular English in the formal setting and Standard English in the informal setting). Twelve polar adjectival pairings prompted participants to evaluate verbal guises along multiple dimensions of interpersonal attraction and aesthetic quality, such as “literate/illiterate,”“rich/poor,” and “pleasing/displeasing,” on a 7-point scale. Te results supported the hypothesis: Participants rated the Standard English-speaking and the socially appropriate code-switching models more positively than either the African American Vernacular English or the inappropriate code-switching guises on the attributes tested. Moreover, participants indicated that they would like to get to know and work alongside the speakers of Standard English and appropriate code-switching. Te results refute the idea that Standard English is preferred over African American Vernacular English in all circumstances; instead, it may be viewed negatively when used in situations where social norms call for standard language. We return to standard language ideologies in Section 7.3. While the study of language attitudes is grounded in social psychology, the study of language ideology fnds its origins in linguistic anthropology.Anthropologist Judith Irvine defnes a language ideology as “the cultural system of ideas about social and linguistic relationships, together with their loading of moral and political interests” (1989, p. 255). Scholarship on language ideologies is thus concerned with identifying and understanding the social and historical dimensions of these shared ideas about languages and the value they hold.As one example,the theocratic institutions of the Arizona Tewa urged resistance to the borrowing and shif of neighboring communities, and ritualized language practices elevated the conservative linguistic forms of ceremonies and political meetings to the community linguistic ideal.As discussed by Kroskrity (1993), in this Pueblo Native American community, language is embedded with ideas of indigenous purism and identity. As another example, Sandel’s sole and co-authored studies highlight the critical relevance of knowledge of historical policies to understanding shared notions of language among Mandarin-Taiwanese bilingual speakers in Taiwan. From 1945 until 1987, the Chinese Nationalist Party government enforced a strict Mandarin Chinese language policy; local languages and varieties were strictly sanctioned in the media and in schools. As a consequence, parents restricted Taiwanese language use in the home in order to more fully expose their children to Mandarin. Over successive generations, parents came to believe that school contexts could sufce for Mandarin exposure, and Taiwanese was used more ofen in the home. Sandel’s (2003) study shows that schoolbased policies had an impact on the family-based language practices. Te fndings also demonstrated the interplay between public and private sites in the development of linguistic ideologies and language as capital. Te government had strictly enforced the

Language ideologies and dispositions

101

ideology wherein Mandarin was the only language that carried linguistic capital, and home language practices mainly indexed ethnicity. Since 2001, schools have ofered mother-tongue instruction towards preserving Taiwan’s mother tongues. Nevertheless, Sandel et al. (2006) report signifcant language shif to Mandarin among parents and children, especially in urban environments. Parents commented on accommodation as a relevant factor, as they spoke Taiwanese to older people, mixed languages with peers, and spoke Mandarin to children. In addition, the parents viewed Taiwanese as the source of tradition and Mandarin as the language of public discourse. Finally, the urban parents did not link Taiwanese language with their identity, while the rural parents did. We turn to discussing language attitudes, ideologies, and the indexing of identity next. Te concepts of cultural and linguistic capital are critical to understanding the privileging of one language or variety over another. According to French sociologist and philosopher Pierre Bourdieu, cultural capital consists of resources that can be translated into social and economic advantage; one such form of cultural capital is linguistic capital, that is, the mastery of the language deemed legitimate in any context.

7.2 ATTITUDES, IDEOLOGIES, AND IDENTITY To recapitulate, whereas language attitudes are individual reactions to languages and their language users, language ideologies are communal, broadly circulated, shared beliefs regarding language use, ofen driven by unequal power relations among different groups. Let us consider the concepts of language attitudes and ideologies in the context of language contact and bilingualism. A pertinent example is found in the language situation of the Dominican Republic, a Spanish-speaking country in the Greater Antilles, and in the diasporic communities of Dominican immigrants and their families in the United States.

The Dominican Republic Te Dominican Republic is a nation on the island Hispaniola, the site of the frst European settlements in the Americas; the western third of the island is occupied by Haiti. Te histories of these present-day nations were shaped by Spanish and then French colonization, importation of enslaved persons from Africa, and several wars of independence—the Haitian Revolution secured independence from France, and the Dominican Republic won its independence from Haiti. In the present day, the two nations difer in signifcant respects. Te Dominican Republic is relatively more prosperous than Haiti, which stands as the poorest country in the Americas, and the majority race in the Dominican Republic is mixed (mulatto, mestizo, or multiracial), while that of Haiti is predominantly African. Important for our purposes, the countries difer with respect to language: Spanish is the ofcial language of the Dominican Republic, while French and Haitian Creole (or Kreyòl) are co-ofcial languages of Haiti, though the vast majority of the population (95%) speaks only Haitian Creole.

102

From minutes to years

Te Spanish spoken in this Caribbean nation is known to difer from other varieties of Spanish along all levels of linguistic analysis. Not surprisingly, scholars have attended to the structure and usage of Dominican Spanish, and discussions of the status and value of Dominican Spanish are commonplace in national educational fora and in mass and social media outlets. In sociolinguistic interviews conducted in the 2000s, Dominicans across social strata ascribe low status to their native language and express a profound sense of linguistic insecurity and shame over its poor state (Toribio, 2000, 2006). Dominicans report that they and their compatriots speak with “orthographical errors”; some reason that their speech is too far removed from its Peninsular Castilian origins, and others lament that their variety is tainted by the African substratum of Haiti, the neighboring nation with which the Dominican Republic shares the island of Hispaniola. Indeed, a verbal guise task administered to university students confrmed the disparaging attitudes towards the speech of residents of the border region: Te 60 participants identifed and negatively evaluated the guises of speakers from the border, whom they assumed to be “uneducated,” “poor,” and “African” (Bullock & Toribio, 2008). What accounts for these attitudes, refected in linguistic insecurity and linguistic discrimination and stereotyping? Te answer lies in the history of the Dominican Republic, especially the history of the border region that separates it from Haiti. In the Dominican Republic, language was central to nation-building under Rafael Trujillo, whose dictatorship promoted an ideology that positioned Dominicans as European and sought to suppress the African infuence of Haiti. As one example, government ofcials erected parochial boarding schools to inculcate Spanish language and Catholic religious practices. Te efects of such institutional interventions are patently manifest in the loss of the bilingualism that the Dominican-Haitian border region once boasted. Bullock and Toribio (2014) demonstrate the enduring legacy of the Eurocentric dogma. In interviews with 90 Dominican children and adolescents about language learning, there emerged negative attitudes towards Haitian Creole, which the youth described as feo “ugly” and disparate “gibberish,” sentiments that were overtly refected in their behavior—a refusal to learn the language. Remarkable in the children’s responses were the direct links they drew between language and identity, as in (7.1a). Also noteworthy is their impression that there was no incentive to invest in learning such a difcult language; instead, they viewed it as quite ordinary and simple for Haitians to learn Spanish, as in (7.1b). Of course, in interpreting Haitians’ accommodation to Dominicans’ speech as unexceptional, rather than driven by unequal power relations, the children and adolescents verbalize understandings that privileges Dominican Spanish and its speakers in the border regions of Hispaniola (Bullock & Toribio, 2014, p. 93). (7.1) a. [¿Aprender krèyol?] No! Porque yo no soy haitiano. “[Learn Kreyòl?] No! Because I am not Haitian.” b. Porque ese idioma nosotros no lo necesitamos saber; porque aquí ello hay abundancia de haitianos pero muy pocas veces uno los necesita y ellos también hablan español. “Because we don’t need to know that language; because there is an abundance of Haitians but rarely do you need them and they also speak Spanish.”

Language ideologies and dispositions

103

Facts about Kreyòl • Te Kreyòl lexicon is based on 18th-century French, but its distinctive grammar derives from West-African languages Fongbe and Igbo. • While 90% of Kreyòl vocabulary originates from French, it maintains its own distinct linguistic properties. • Kreyòl distinguishes itself among Creole languages because it is the most spoken in the world and has a standard writing system.

On the western side of Hispaniola, the language situation of Haiti presents an intriguing case: Kreyòl is the primary spoken language, but French is the primary language of instruction. And while virtually all Haitians speak Kreyòl and only 5% speak French, it is the biliterate minority occupies the administrative posts in the country, reinforcing French as the language associated with social mobility and achievement. As described by DeGraf (2005) and Hebblethwaite (2012), the Francophile elite subscribes to the language ideology of “Kreyòl exceptionalism”—the view that Kreyòl languages are defcient relative to their lexifer languages, unft for intellectual pursuits, and therefore a handicap for their speakers (for more about Creoles, go to Chapter 11). More generally, the diglossic relationship between French and Kreyòl spoken by the bilingual elite not only perpetuates the view of French as a more prestigious language than Creole, but it also excludes people who do not speak French from prestigious domains, thereby limiting their power. However, while it is stigmatized by the elite, Kreyòl simultaneously holds covert prestige as the in-group language of most Haitian communities of practice. Nevertheless, there persist linguistic inequalities rooted in colonial language ideologies. In a recent study, Ulysse and Burns (2022) explore the attitudes and ideologies of Haitians towards French and Kreyòl use, especially in educational contexts. Te researchers reason that educational institutions reproduce and disseminate language ideologies and that mother tongue instruction in Haiti could help foster more equal power relations through access to literacy and de-stigmatization of Kreyòl. In their study, the majority of the participants registered positive attitudes towards Kreyòl: Tey reported that they feel happiest when speaking Kreyòl, they thought that monolingual Haitians should be proud of speaking Kreyòl, they expressed support for making Kreyòl the sole ofcial language and educational policy, and they remarked that Kreyòl is the language most connected to their identity (7.2a). Still, there were some consultants who held more ambivalent attitudes towards Kreyòl.As shown in (7.2b), respondents are aware of language ideologies and practices that privilege French and marginalize Kreyòl among fellow Haitians who act as gatekeepers to human resources and services. (7.2) a. Fransè p ap ka idantite w. Ou pa blan Fransè. Se aprann ou aprann li. Lang ki pi konekte avè w la, ki nan san w, nan nanm ou se kreyol la. ‘French can’t be your identity. You are not a French foreigner. You learn it. Te language that is most connected to your identity, which is your blood, in your soul is Kreyòl.’ (Ulysse & Burns, 2022, p. 178)

104

From minutes to years b. Gen de ti sekretè nan biwo yon minis oubyen medsen. Si w adrese l an kreyòl li gen dwa pa reponn ou. Li gen dwa pa menm gade w. Alòske li pa ta dwe konsa. Ou vin pa alèz lè w pale l. ‘Tere are some places where you would go, even other fellow Haitians would humiliate you if you spoke Kreyòl. Tere are some receptionists in the ofces of ministers or doctors. If you speak in Kreyòl to them, they may not respond. Tey may not even look. at you. But it should not be like that. You become uncomfortable when you speak it.’ (Ulysse & Burns, 2022, p. 176)

Overall, the Haitian consultants displayed strong afective alignment with Kreyòl and an awareness of its covert prestige in their daily lives in Haiti. Tey demonstrated disparate orientations towards French—some viewed it as a resource and others as a problem—but they all recognized its overt prestige. Participants favored the promotion of Kreyòl language and culture in educational settings, but their attitudes reproduced ideologies of French as the language of social mobility and socioeconomic opportunity. You might wonder: Are language attitudes and ideologies altered when language users migrate to new contexts? Continuing with our Caribbean examples, we might ask: Does Dominicans’ linguistic insecurity persist in the United States? Does it hasten a shif to English, the socially dominant language? Do Haitians reproduce the persistent prestige associated with French to the ideological detriment of Creole in the U.S. diaspora? Do children of immigrants inherit their parents’ dispositions towards languages? In the U.S. diaspora, Dominican Spanish is stigmatized and held in low esteem. As documented in attitudinal studies conducted in Miami, participants ranked Dominican Spanish bottommost of all Spanish varieties on measures of correctness and pleasantness.And in New York, where Dominican immigrants and their families have largely settled, Dominicans themselves hold the view that they speak a “poor” Spanish that should not be taught in schools, and they express embarrassment on hearing their compatriots’ speech in public spaces. Tus, Dominicans’ language attitudes regarding Dominican Spanish remain unchanged—they evaluate their speech negatively on status. But do they abandon their Spanish and shif to English, the local language of power? No, they remain loyal to their language! Te heritage language is evaluated positively in terms of solidarity and thus enjoys some measure of covert prestige. Tese apparently contradictory feelings are not surprising: language users ofen accept the stigma attached to their way of speaking, indicative of their linguistic insecurity; nevertheless, despite the negative prestige associated with many language varieties, they persist, in large part because they serve important functions as markers of social identity. Interviews with Dominican Spanish-English bilingual adults and children in New York amply illustrate the critical importance of the heritage language as a resource in binding Dominicans to their culture and to their cohorts at home and abroad. Importantly, for Dominicans, language also serves in negotiating the Black/ White racial ideology of the United States. In particular, Dominicans who are racialized as African American by their appearance may contest this categorization through language display, such as by speaking Spanish (or by orienting towards a variety of English that avoids features attributed to African American Vernacular English) (Toribio,2003). In brief, language is a valuable resource in the display of the ethnic and racial self.

Language ideologies and dispositions

105

Research on the attitudes and ideologies of Haitian individuals and communities in the diaspora in the United States has focused on use of French and Kreyòl (and English). In an early study, Zéphir (1997) explored the social value of French expressed in interviews with 100 frst-generation immigrants, many of them FrenchKreyòl bilinguals. According to Zéphir, French-Kreyòl bilingual immigrants readily admit that their native language is Kreyòl, but they are keen to stress their command of French. What motivates this behavior? Haitian immigrants arrive in the United States with a defnite idea that knowledge of French connotes higher social class, a higher level of education, and a more refned and cultivated lifestyle, and they emphasize their French-speaking ability as a way of gaining status. Tus, French was commonly deployed for its symbolic power, especially in public spaces in the United States, as shared by a speaker in (7.3a). Te presence of French was also manifested in the home, where French-Kreyòl code-switching was the norm, as in the second excerpt (7.3b). Most notably, the bilingual parents reported using French with their young children in anticipation of social benefts that would accrue; they believed that the transfer of the linguistic capital of French through early exposure would enable the children to project a more positive image as they entered school and the larger society. (7.3) a. According to Haitian tradition, frst contact with gens de bien are made in French. (Zéphir, 1997, p. 400) b. La soupe est très bonne, fè yon ti goute, no. Te soup is very good, why don’t you taste it? (Zéphir, 1997, p. 401)

Zéphir’s study also revealed that bilingual Haitians in the U.S. diaspora use a “frenchifed” variety of Kreyòl (Kreyòl swa) that is markedly diferent from that of Kreyòl monolinguals (Kreyòl rèk). On the lexical level, the bilingual variety spoken by the bilinguals is characterized by a high proportion of borrowings from French. At the level of phonology, the bilingual Kreyòl has three additional front rounded vowels (e.g., paskeu, cf. monolingual Kreyòl: paske, French: parce que; peù, cf. monolingual Kreyòl: pè, French: “peur”). At the syntactic level, it demonstrates the use of the subordinator keu (e.g., m di nou keu m pa konprann anyen; cf. monolingual Kreyòl: m di nou m pa konprann anyen, French: je te dis que je ne comprends rien) and the syntactic relator deu in attributor constructions (e.g., yon seri deu pwoblèm; cf. monolingual Kreyòl yon seri pwoblèm, French: un série de problèmes). Tese linguistic innovations allow bilinguals to mark themselves of from monolingual Kreyòl speakers, highlighting the infuence of the French that they command. In a more recent study, Bonnefl (2018) undertakes a study of ten elite Haitian immigrants—French-Kreyòl-English multilinguals—in Washington, D.C. Qualitative and quantitative data gathered by the author show that the participants acknowledge the divisive efects of the diglossia in their homeland (7.4a), but they still hold French in high esteem (7.4b, c), buttressed by the same ideologies that circulated in Haiti. (7.4) a. Se lang nan ki separe, ki kreye fose a. C’est la langue qui crée le fossé. ‘Language is what separates, what creates the gap.’ (Bonnefl, 2018, p. 49)

106

From minutes to years b. In the minds of Haitian people living there in the day-to-day, French is still very much, like, the language that you need to show that you can speak a certain level of, to show that you’re a certain level of afuence or status. (Bonnefl, 2018, p. 53)

Like Zéphir, Bonnefl concludes that the colonial language ideologies that instill French with social privilege and power in Haiti endure in the U.S. diaspora.Advancing the argument, we might say that in the U.S. context, French is imbued with additional linguistic capital, as it sets Haitian immigrants apart from non-francophone Black Americans. A contrasting case with regard to the endurance of language ideologies of the family country of origin is presented by descendants of African immigrants born and raised in the United States. Unlike the Dominicans and Haitians discussed previously, the Ghanaian immigrants studied by Obeng (2008) underwent a rapid shif to English. In their New York enclaves, the Akan language and culture were shown to be circumscribed to the limited domains of ritual practices. Obeng notes that the shif to English was motivated by the immigrants’ marginalization in mainstream society and by limited opportunity for Akan use in the new context. Most importantly, the shif was compelled by the ideological framing of English as the language of progress in the ancestral communities. Tat is, English was already ideologically positioned as holding more instrumental value and social capital in the sending Ghanaian society. Makoni (2018) ofers additional insights on the attitudes and language learning behaviors of U.S.-born African immigrants enrolled in African language classes. For the 70 third-generation college students under study, African languages became important instruments of cultural identity and self-assertion. Te students felt that they were delegitimized, rejected as “native” speakers of English in U.S. society and rejected as Americans because of their African heritage, as expressed in (7.5). Feeling alienated, the participants chose to locate “home” elsewhere, and through language learning, they constructed alternative identities. (7.5) a. It’s strange, I know. I speak English right. I grew up in Queens . . . everywhere I go I’m asked the same question, it’s like where you from and when I say Queens then the next thing is, no I mean originally, where you originally from. When I was younger it was confusing but as I grew older, I . . . like I do not belong in America like not American maybe. I was with a white Slovene girl who could barely speak English . . . no one asked her that question. (Makoni, 2018, p. 80) b. Being profcient in English is not good enough . . . my name says it all . . . not English not American .  .  . take ESL classes. I cannot speak another language other than English. I must learn Akan .  .  . Akan has my roots written all over it. (Makoni, 2018, p. 82)

Te U.S.-born African students interviewed by Makoni had undergone a language shif to English. Notably, however, Makoni’s participants had no ties to the ideological stances regarding the relative prestige of English versus African languages—although their parents and grandparents did. While the learners held positive views about the family heritage language, they reported that their parents were indiferent to their language-learning eforts, and their grandparents openly disdained the family heritage language and discouraged its inheritance. Nevertheless, the fndings suggest the

Language ideologies and dispositions

107

possibility of language recovery following language shif, if not for communication, at least as a symbol of social identifcation.

7.3

STANDARD LANGUAGE AND RELATED IDEOLOGIES

Recall that language ideologies are conceptualizations about language that are shared and relate language and society. Because they are pervaded with political and moral interests, they promote the interests of the socially privileged groups at the expense of all less powerful groups. Tis is best demonstrated with reference to the well-studied “standard language ideology.” Te standard language ideology refers to the belief that a particular variety is aesthetically,morally, and intellectually superior to others. Te standard variety is usually the variety that has its roots in the language of the most powerful group in society, that is ofen based on the written language, that is highly homogeneous, and that is acquired through long years of formal education. Te ideology is taught in schools, promoted by the media and corporate sector, institutionalized in governmental services, and reinforced in public life (Lippi-Green, 2012). Standard language ideology creates and maintains linguistic and social hierarchies, elevating one variety as a crucial resource for social advancement and subordinating those groups who do not command it (Kroskrity, 2004). Te beliefs that constitute the standard language ideology become accepted to the extent that they are perceived as “common sense” and can make it seem “natural” and “normal” that speakers of that variety should occupy privileged positions in society, while non-speakers should be marginalized (Piller, 2015). Contrasted with standard language, a non-standard is any variety diferent from the dominant standard variety. Tis diferentiation points to another powerful language ideology, that of a language hierarchy, which allows for language varieties to be divided, labeled, and ranked (Weber & Horner, 2012). In some societies, such as the United States, the standard language ideology that privileges Standard American English is closely related to the one nation, one language ideology—the belief that the use of a single common language is important for national unity. Te political justifcations for an ofcial language are based on the principles described by the standard language ideology and the one nation, one language ideology. And the latter is linked to the ideology of monolingualism, the belief that individuals and societies are normatively monolingual, irrespective of actual linguistic diversity. In educational settings, standard language ideologies are found in the varieties that are promoted in the materials (e.g., dictionaries and grammars) that assure proper and correct usage. Te selection of the standard has important implications for language planning and policy. Consider the linguistic situation of Namibia. English is the sole ofcial language in Namibia, spoken as a second or third language by the majority in this highly multilingual country; 12 additional languages have been recognized as national languages, including 10 Indigenous languages spoken by 87% of the population.According to Da Costa et al. (2013) the post-colonial leadership wanted to form a nation defned by one single language, as the use of many languages in the country was (and still is) considered a problem rather than an asset. (Recall the one nation-one language ideology.) Te proclamation of languages as ofcial, national, and Indigenous imposes a power and status hierarchy among the languages and among the users of these languages.More critically, material resources for the development of languages depend on ofcial designations. In this context, English will be supported at the expense of local languages.

108

From minutes to years

In Namibia, both standard language and purist ideologies underlie teachers’ concern with language correctness. Teachers insist on teaching learners what they consider the “pure” form of English and look down upon the hybrid varieties spoken by the learners and their families. Te ideologies are widespread and extend beyond the classroom. Gauging ideological stances in call-in radio programs, Haingura (2020) found that while adults and youths practiced multilingualism, they maintained that the standard language should be used at all times rather than the hybrid forms of their everyday practices.

Rejecting abyssal thinking in the language and education of racialized bilinguals: A manifesto García and colleagues (2021) write a manifesto to draw attention to discrimination in education that treats racialized bilinguals harshly. Te authors challenge dominant assumptions about language, bilingualism, and education, arguing that these assumptions have their roots in colonialism. Tey propose a translanguaging perspective that rejects colonial boundaries of named languages. Tis decolonial approach considers that in order to attain justice in education, we must allow racialized bilinguals’ knowledge to be expressed fully. If you want to know more about translanguaging, go to Section 4.3. Here’s an example illustrating the problem: In the U.S. the assessment of fve-year old children entering kindergarten consists, for the most part, of having them orally describe pictures. One of us once observed the case of Margarita, a Mexican American girl born in the U.S. who was asked to describe a picture of a mother making cookies from dough. In this assessment, the word dough was assigned more points than the word cookie because it was considered academic language. Margarita was very familiar with her mother making tortillas with masa, but she had never made cookies with her mother or used the word dough. So even though she could describe the picture of the mother making cookies, she could not come up with the word dough.As a result of numerous examples of this kind, she was put in an English-as-a-second language program and kept out of the school’s dual language bilingual program, which was reserved for those students who tested as “gifed and talented,” and thus inappropriate for a child considered to have limited vocabulary. In the ESL program, Margarita was not challenged because the focus was simply on having her add more presumably academic vocabulary to her lexicon. Our question has always been: What would have happened to Margarita if she had been challenged by teachers and classmates who believed in what she could do with language? What if she had been engaged in funny, imaginative, and challenging work that built on her existing linguistic and cultural knowledge? (García et al., 2021, pp. 210–211)

Language ideologies and dispositions

109

In foreign-language educational curricula as well, teaching and learning are also organized around a chosen standard. As one example, standard Peninsular is the variety of Spanish that is broadly promoted in Spanish-language instruction, at the expense of non-Castilian varieties. Note that here, too, this standardization privileges the monolingual “hyperstandard,” erasing the linguistic diversity that is attested in the Spanish-speaking world. Since the standard is the variety that is ideologically associated with qualities that are conventionally valued, these linkages have important implications for the social world. Kircher and Fox (2021) investigate the standard language ideologies and the social stereotypes associated with speakers of Multicultural London English, an ethnolect spoken in parts of London (see Wiese, 2014 on the German ethnolect Kiezdeutsch in Berlin.) In Great Britain, the standard variety (Standard English) is codifed as the model of correctness that serves as an idealized reference for descriptions of variation (Milroy, 2000). Since the 1980s there has emerged an identifable ethnolect that refects the increasing ethnic and linguistic diversity and the extensive language and dialect contact that characterize some London neighborhoods. For this study, the researchers conducted a corpus analysis of responses submitted by 735 participants in order to explore the discursive representations of Multicultural London English and its speakers. Specifcally, they sought to understand the extent to which Londoners of diferent backgrounds had internalized the prevailing standard language ideology and corresponding stereotypes. Te non-ethnolect speakers adhered to the standard language ideology, using more negative than positive descriptors of the new variety (e.g., broken language, language decay,obstacle to success, fake variety).In addition,the discourse of this group displayed higher relative frequencies of stereotyped demographic descriptors of ethnolect speakers (e.g., ethnic minorities, teenagers, male) than did the ethnolect speakers themselves (cf., 7.6a vs. 7.6b, from Kircher & Fox, 2021, p. 802). Nonetheless, the responses of the ethnolect speakers also refect the infuence of the standard language ideology, as they referenced the variety as hindering employment opportunities. (7.6) a. First impression is that they are from East London, frst- or second-generation immigrants, ofen Bengali or Pakistani, or African. Tey are usually working class, parents may not speak English, ofen bilingual. b. As an MLE speaker who has a lot of contact with other MLE users, I think it doesn’t say a lot about its user. many diferent people I know (age, gender, class) speak in this manner.

As noted, standard language ideologies can have severe consequences in social life, such as in educational and employment advancement. Tis is also the case in the administration of justice. Rickford and King (2016) discuss the testimony of a key prosecution witness in a high-profle murder trial.Although the witness ofered valuable evidence, her testimony was dismissed as incomprehensible, unreliable, and non-credible, because it was delivered in vernacular African American English and Caribbean-infuenced English, rather than standard English. Her “non-standard” English was deemed unsuited for legal proceedings. According to the authors, the standard language ideology marginalized the witness in a domain

110

From minutes to years

Figure 7.1 National Fair Housing Alliance advertisement

Language ideologies and dispositions

111

where equality and justice are supposedly core pursuits. In this situation, language ideology manifested into associated attitudes (Albury, 2020). Similar discriminatory is attested in housing, prompting advertisements such as the following from the U.S. Department of Housing and Urban Development in Figure 7.1, also available online at www.hud.gov/sites/documents/NFHAHUDSOUNDSBLACK_EN.PDF. Finally, it should be said that standard language is ofen equated with the ofcial language, especially in postcolonial settings, to the detriment of users of local languages. Robertson and Evans (2019) report on Creolophone contexts, specifcally St. Lucia, which gained its independence from Britain in 1979. While Kwéyòl, not English, was and remains the dominant language of the majority of St. Lucians, it is not the ofcial language, and there are no language guarantees to protect Kwéyòl speakers. Tis has farreaching consequences in critical areas such as in the provision of healthcare and in law enforcement procedures.According to Robertson and Evans, even local magistrates and lawyers who are competent in Kwéyòl do not speak Kwéyòl in the courts. Rather, they display a marked uniformity with regard to their ideological stances towards Kwéyòl and English; competence in the latter is the principal identifer of education achievement and social standing, and the disregard for speakers of varieties other than the ofcial standard variety entails a distinct disadvantage. Tis is yet another example of the myriad of ways in which language articulates with systems of power.

Chapter summary In this chapter we have examined language attitudes, the evaluative reactions that we make about language varieties and their users, such as how pleasant or intelligent a variety or its speakers strike us as being. Studies on language attitudes are consistent in demonstrating that while users of minority varieties are more positively inclined towards those who speak like them, they assign greater status to socially dominant languages and language varieties. While language attitudes are individual, language ideologies are shared beliefs that are shaped by historical and political circumstances.We discussed scholarship on language ideology, which simultaneously attends to micro-details of linguistic practice and macro details of national and colonial histories, political movements, and/ or socioeconomic structures. We reviewed studies that pursue specifc explanations of the origin of language ideologies and their manifestation in how language is used and how it is viewed.We focused on standard language ideology, which reinforces social hierarchies, marginalizing groups with no access to the idealized standard.As we observed, standard language ideology can have deleterious efects; it can lead to linguistic discrimination and hinder language development in multilingual societies. It should not go without mentioning, however, that a standard language is just another variety; nevertheless, it is the variety that is privileged because, as we have seen, what people believe and feel about language refects structures of power.

From minutes to years

112

EXERCISES 1.

2.

Class discussion a.

We project and perform identities through linguistic means, and we interpret others’ language through our own ideological lenses. Can you think of a time when you deliberately oriented your own language use towards a variety of overt prestige or when you doubled down in insisting on using your local vernacular?

b.

Or can you recall an occasion when you made assumptions about a person based on their language use? What was the basis of your assumptions?

c.

Standard language ideologies can afect the ability of users of minority languages to succeed academically because of educators’ perceptions of what constitutes “good” language. Have you ever felt that a teacher disparaged your language? Describe the circumstances.

Paired discussion

One efect of standard language ideology is found in linguistic profling, that is, racial identifcation based on speech. In an important study addressing linguistic profling, Baugh and colleagues determined that housing authorities used racebased discrimination based on audio cues during telephone interviews (Baugh, 2000; Purnell et al., 1999); the listeners variously confrmed or denied appointments with callers who spoke standard or non-standard varieties of English. Can you think of other areas of everyday life in which linguistic profling and discrimination are manifested?

3.

Discover

Document multilingualism in the linguistic landscape of your community, for example, with images of supermarket signage, restaurant menus, school posters, advertisements on buses and trams, and so on (see Figure 11.1) Where are multilingual signs found in your community? Who is the intended audience? What varieties of the languages are incorporated? Which language takes precedence? What does the placement and organization of the languages say about the relative prestige of the languages and their users in the community?

4.

Explore

Our feelings about language may be a proxy for a myriad of sociopsychological reactions about people, some of which may be unconscious and not open to introspection. In social psychology this is called “implicit bias”—attitudes, stereotypes, and prejudices that afect our understanding, decisions, and actions in an unconscious manner (Greenwald & Banaji, 1995).You can test your implicit biases on various measures of social attitudes (e.g., race, religion, gender) at https://implicit.harvard.edu/implicit/.

Language ideologies and dispositions

113

FURTHER READING Alim, S., & Rickford, J. (2016). Raciolinguistics: How language shapes our ideas about race. Oxford University Press. Seminal compendium of research examining the interrelationship of race, ethnicity, and language in the United States context. Journal of Multilingual and Multicultural Development. 2021. Special Issue: In the Shadow of the Standard: Standard Language Ideology and Attitudes towards Non-Standard Varieties and Usages. Edited collection of eight papers focused on the impact of standard language ideologies on a broad range of “non-standard” varieties across political and cultural contexts.

REFERENCES Albury, N. (2020). Language attitudes and ideologies on linguistic diversity. In A. Schalley, & S. Eisenchlas (Eds.), Handbook of home language maintenance and development (pp. 357–376). De Gruyter Mouton. Alfaraz, G. G. (2002). Miami Cuban perceptions of varieties of Spanish. In D. Long, & D. R. Preston (Eds.), Handbook of perceptual dialectology (pp. 1–11). John Benjamins. Alfaraz, G. G. (2014). Dialect perceptions in real time: A restudy of Miami-Cuban perceptions. Journal of Linguistic Geography, 2(2), 74–86. Bauer, L., & Trudgill, P. (1998). Language myths. Penguin Books. Baugh, J. (2000). Racial identifcation by speech. American Speech, 75(4), 362–364. Bonnefl, C. E. (2018). Language attitudes of multilingual Haitians in the Washington, DC, area [PhD thesis, George Mason University]. Bullock, B. E., & Toribio, A. J. (2008). Kreyol incursions into Dominican Spanish: Te perception of Haitianized speech. In M. Niño-Murcia, & J. Rothman (Eds.), Bilingualism and identity: Spanish at the crossroads with other languages (pp. 175–198). John Benjamins. Bullock, B. E., & Toribio, A. J. (2014). From Trujillo to the terremoto: Te efect of language ideologies on the language attitudes and behaviors of the rural youth of the northern Dominican border. International Journal of the Sociology of Language, 227, 83–100. Da Costa, D., Dyers, C., & Mheta, G. (2013). Managing linguistic diversity through standardization. In Z. Bock, & G. Mheta (Eds.), Language, society & communication: An introduction (pp. 311–332).Van Schaik Publishers. DeGraf, M. (2005). Linguists most dangerous myth: Te fallacy of Creole exceptionalism. Language in Society, 34(4), 533–391. Dragojevic, M. (2017). Language attitudes. In Oxford research encyclopedia of communication. Oxford University Press. https://oxfordre.com/communication/view/10.1093/acrefore/9780190228613.001.0001/ acrefore-9780190228613-e-437. García, O., Flores, N., Seltzer, K., Li Wei, Otheguy, R., & Rosa, J. (2021). Rejecting abyssal thinking in the language and education of racialized bilinguals: A manifesto. Critical Inquiry in Language Studies, 18(3), 203–228, https://doi.org/10.1080/15427587.2021.1935957 Garrett, P. (2010). Attitudes to language. Cambridge University Press. Greenwald, A., & Banaji, M. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102(1), 4–27. Haingura, P. (2020). Te role of language ideologies in the development of indigenous languages in Namibia. Ministry of Education, 59. Hebblethwaite, B. (2012). French and underdevelopment, Haitian Creole and development: Educational language policy problems and solutions in Haiti. Journal of Pidgin and Creole Languages, 27(2), 255–302.

114

From minutes to years

Irvine, J. T. (1989). When talk isn’t cheap: Language and political economy. American Ethnologist, 16(2), 248–267. Kircher, R., & Fox, S. (2021). Multicultural London English and its speakers: A corpus-informed discourse study of standard language ideology and social stereotypes. Journal of Multilingual and Multicultural Development, 42(9), 792–810. Koch, L. M., Gross, A. M., & Kolts, R. (2001). Attitudes toward Black English and code switching. Journal of Black Psychology, 27(1), 29–42. Kroskrity, P. V. (1993). Language, history, and identity: Ethnolinguistic studies of the Arizona Tewa. University of Arizona Press. Kroskrity, P.V. (2004). Language ideologies. In A. Duranti (Ed.), A companion to linguistic anthropology (pp. 496–517). Wiley Online Library. Kutlu, E. (2020). Now you see me, now you mishear me: Raciolinguistic accounts of speech perception in different English varieties. Journal of Multilingual and Multicultural Development. doi: 10.1080/01434632.2020.1835929 Lambert, W. E., Hodgson, R. C., Gardner, R. C., & Fillenbaum, S. (1960). Evaluational reactions to spoken languages. Te Journal of Abnormal and Social Psychology, 60(1), 44–51. Lippi-Green, R. (2012). English with an accent: Language, ideology and discrimination in the United States. Routledge. Makoni, B. (2018). Beyond country of birth: Heritage language learning and the discursive construction of identities of resistance. Heritage Language Journal, 15(1), 71–94. Milroy, J. (2000). Historical description and the ideology of the standard language. In L. Wright (Ed.), Te development of standard English (pp. 11–28). Cambridge University Press. Piller, I. (2015). Language ideologies. Te International Encyclopedia of Language and Social Interactions. https://doi.org/10.1002/9781118611463.wbielsi140 Purnell, T., Idsardi,W., & Baugh, J. (1999). Perceptual and phonetic experiments on American English dialect identifcation. Journal of Language and Social Psychology, 18(1), 10–30. Rickford, J. R., & King, S. (2016). Language and linguistics on trial: Hearing Rachel Jeantel (and other vernacular speakers) in the courtroom and beyond. Language, 92(4), 948–988. Robertson, I. E., & Evans, R. S. (2019). Systemic linguistic discrimination and disenfranchisement in the Creolophone Caribbean: Te case of the St. Lucian Legal System. In I. E. Robertson, & R. S. Evans (Eds.), Te Routledge companion to the work of John R. Rickford (pp. 46–51). Routledge. Sandel, T. L. (2003). Linguistic capital in Taiwan: Te KMT’s Mandarin language policy and its perceived impact on language practices of bilingual Mandarin and Tai-gi speakers. Language in Society, 32(4), 523–551. Sandel, T. L., Chao,W.Y., & Liang, C. H. (2006). Language shif and language accommodation across family generations in Taiwan. Journal of Multilingual and Multicultural Development, 27(2), 126–147. Toribio, A. J. (2000). Language variation and the linguistic enactment of identity among Dominicans. Linguistics, 38(5), 1133–1159. Toribio, A. J. (2003). Te social signifcance of Spanish language loyalty among black and white Dominicans in New York. Bilingual Review/La Revista Bilingüe, 3–11. Toribio, A. J. (2006). Linguistic displays of identity among Dominicans in national and diasporic settings. In English and ethnicity (pp. 131–155). Palgrave Macmillan. Ulysse, G. M., & Burns, K. E. (2022). French and Kreyòl in multilingual Haiti: Insights on the relationship between language attitudes, language policy, and literacy from Haitian Gonâviens. Critical Inquiry in Language Studies, 19(2), 163–192. Weber, J. J., & Horner, K. (2012). Te trilingual Luxembourgish school system in historical perspective: Progress or regress? Language, Culture and Curriculum, 25(1), 3–15. Wiese, H. (2014).Voices of linguistic outrage: Standard language constructs and the discourse on new urban dialects. Working Papers in Urban Language and Literacies, 120, 1–25. Zéphir, F. (1997). Te social value of French for bilingual Haitian immigrants. French Review, 70(3), 395–406.

Part III From years to centuries How languages change through contact

KEY QUESTIONS How do languages change in contact with other languages? What happens when several languages are in long-term contact with one another? What are Creoles and pidgins? What are mixed languages? Can I observe language contact processes in my daily life?

8 Contact-induced changes in grammar and borrowing

CHAPTER PREVIEW In this chapter, we begin looking at the long-term efects of bilingualism and address the following overarching question: How do languages change in contact with other languages? Contact-induced change refers to cases where the functions and structures of two languages in contact come to overlap, or converge, not from accidental similarity but as the ultimate result of cognitive mechanisms that are constantly at play in the bilingual mind such as conceptual transfer, a phenomenon discussed in Chapter 3, and implicit priming, a mechanism discussed in Chapter 1. In this chapter,“borrowing” will be defned as the transfer of sound and form-meaning units, like words. Borrowings start at the level of the bilingual language user and then spread throughout a language community as language users accommodate to one another (Chapter 5), with the result that borrowings can become anchored in a recipient language as loanwords (see Chapter 4). Te same cognitive mechanisms can bring language users to fnd overlap between their languages outside their vocabularies. We discuss examples of the outcomes of language contact on the meaning and structures of various languages that result from convergence.

8.1

DIFFERENT TYPES OF LANGUAGE CHANGE

In this chapter we take a closer look at the infuences that long-term language contact can have on the structure of a language. Let’s consider how a language changes with time. Te key to language change is “language variation.” Language variation is the term used to describe the fact that there are competing ways of saying the same thing in the same language. Older adults use language diferently from adolescents, and people from one region may use language in ways that are noticeably diferent from residents of another. Language variation can impact every part of linguistic structure, from the intonation we use to ask a question to the way we use word order to form sentences. Language change occurs when the use of one linguistic form takes over the use of another.When change occurs, the displaced form might remain in the language, where it is only used in restricted contexts, like the word whom in English. DOI: 10.4324/9781003167952-11

118

From years to centuries

When, if ever, do you use that word in English? Over time, some older forms might be given up altogether. For example, the old plural form of cow was kine, but, over time, it was regularized to cows to match the other singular–plural pairs in English like cat, cats. Linguists are interested in studying the historical roots of a language because observations of similarities between languages have led to an understanding of how languages are related. And this can shed light on the movements of ancient populations and the spread of cultures. Because most languages do not have a long history of written records, and some have never been written down at all, linguists must use special methods to project back in time. One method compares structural properties, like the vocabularies of languages. Doing this, we’d fnd that the word for two is do in Farsi, Punjabi, and Hindustani; du in Kurdish; dos in Catalan and Spanish; ðʋo in Greek; dva in the Slavic languages; and zwei, twa, twee, or two in Germanic languages. Te overlap in this word found across diferent languages turns out not to be a coincidence. Instead, these languages, and many others found across Europe and Asia, are related to a single source language, known as Proto-Indo-European. While there’s no reason to believe that speakers of Proto-Indo-European didn’t have language variation, as languages generally do, the diferences between the sounds that make up the word with the meaning “two” in these languages today are the result of variation as speakers moved away from one another. Unrelated languages can also be similar because words or structures from one language are borrowed into another when cultures come in contact. Many modern languages contain vocabulary items that trace back to Arabic. A small sample of the words in modern languages that derive from the same Arabic roots are listed in (8.1): (8.1) Arabic words in modern European languages a) English: assassin, alcohol, magazine, apricot b) French: assassin, alcool, magasin “store,” abricot c) Spanish: asesino, alcohol, almacen “warehouse,” albaricoque

Compared to the words for the meaning “two,” which are inherited words in the Indo-European languages, these words are newcomers to the languages listed in (8.1). Unlike inherited words, loanwords are not inherited from a source language but are adopted into a language once there is contact between peoples using diferent languages.

Arabic loanwords in the languages of the Iberian Peninsula Te infuence of Arabic is very strong in the vocabularies of the Romance languages of Iberia: Catalan, Galician, Portuguese, and Spanish. Te Iberian Peninsula was the territory of Arabic-Muslim populations from 711 until the Reconquest of the Peninsula in 1492, which meant that the Romance languages that developed there were in constant contact with Arabic over eight centuries.

Contact-induced changes in grammar and borrowing

119

A property of these loanwords is that Arabic nouns entered the Romance languages with the Arabic defnite article attached to the noun, as in almojada “pillow,” which came from Arabic al-mikadda “the cushion.” Many Arabic loanwords are recognizable today because they begin with al- in the Iberian languages. However, the Arabic origin of some of the words used in today’s Iberian languages are not so easy to spot: azucar “sugar,” aceite “oil,” arroz “rice.” Tese begin with a but not with al-. Tis is because Arabic makes a distinction between sun letters and moon letters. Sun letters are what are called coronal consonants. To articulate them, the front of your tongue makes contact against your upper teeth or hard palate. Te sounds l, n, t, s, are all sun letters, while b, g, h are moon letters. With moon letters, al- does not change, but with sun letters, the /l/ of the morpheme al, which is a coronal consonant, melds with, or assimilates to, the following sun letter so the /l/ is no longer heard (unless the noun itself begins with /l/, as in al-lah,“the Deity, the God.” Notice that forms of these words, some beginning with al and some with a, means that the Romance-speaking borrowers were unaware that there was a determiner attached to these words. For them, the morpheme, al-, was perceived as just the frst syllable of an Arabic word since they used them with their own defnite articles, el, las, lo, etc.: el almojada (singular), las almojadas (plural). Tis implies that the Romance speakers of the Iberian Peninsula were probably not bilingual in Arabic when these words entered the language.

Lexical borrowing is an example of a contact-induced change, a type of external language change. Tis contrasts with internal change in which a language alters naturally through time in the absence of external infuences such as contact with another language or the infuence of correction from schooling. We can see evidence of an internal language change in the items listed in (8.1). Te meaning of the words for assassin, alcohol, and apricot are equivalent across these languages today. But the meaning of English magazine, which entered the language via French, has shifed away from referring to a physical space, like a store, and now refers to the storage of information in a specialized publication. Tis kind of language-internal change is called a “semantic shif.” Semantic shifs happen frequently in languages. Trough usage, words take on either more specifc or more general meanings from what they originally referred to. In this book, of course, we are focused on contact-induced changes rather than language internal changes. We are interested in the variations that enter a language through contact and how these variations lead to change over time. As we will see in §8.4.1, processes that occur due to language-internal change, like semantic shif, also take place because of language contact. Contact-induced language changes occur in a language when an innovation used originally by bilinguals is adopted by a community of language users. Let’s start our exploration of contact-induced change with an overview of the types of contact-induced variations that can afect a language.

120

From years to centuries

8.2 TYPES OF CONTACT-INDUCED CHANGES Tere are two main types of contact-induced changes. Te frst describes an overt form of language interaction, “borrowing” and “code-switching,” where a form and its meaning from a “source language” are used in a “recipient language” (Weinreich, 1953). In the second type of contact-induced variation, actual forms are not copied from the source language to the recipient language, only meanings or structures. Linguists refer to this second type of variation using a number of diferent terms, including “interference,” “copying,” “calquing,” “transfer,” “imposition,” “convergence,” or “pattern replication” (van Coetsem, 1988; Tomason & Kaufman, 1988; Jarvis & Pavlenko, 2007; Matras & Sakel, 2007; Trefers-Daller, 2009). Here, we will use the term “convergence”because it captures the idea that the languages become more alike. An example of convergence is shown in the scenario in (8.2). (8.2) A group enters a restaurant to ask for a table: Host: Welcome! How many are in your party? Guest: We are six.

Te reply given by the guest is in English and it is an acceptable sentence, but its phrasing is rare for that language, at least as it is spoken in North America. Usually, we would express this in English using the construction with an impersonal pronoun, there are (six of us). In comparison, speakers of Romance languages typically use a personal pronoun, as in French on est six “we are six.” When the guest in the scenario described in (8.2) says “we are six,” and we know that this person is bilingual in English and a Romance language, then it’s possible to conclude that they are translating the personal pronoun pattern of a Romance source language into English. If you have ever tried to learn another language, then you, too, will have literally translated words and expressions from your other language or languages when you are expressing something in a new language. Because this is such a common strategy of foreign language learners whose abilities in a new language are not yet well developed, convergence is sometimes thought to be limited to the language use of individuals who are stronger in one language than they are in the other. But all bilinguals show signs of convergence in their language use. Recall from what we learned in Part I of this book that interaction between the languages in the mind of bilinguals is expected, no matter how strong the bilinguals might be in their languages. In fact, the two English-dominant authors of this book, who are also Romance language speakers, frequently use the phrase in (8.2) when they speak English, even though they live and work in an English-dominant environment where other speakers are more likely to use the impersonal there are construction. Convergence also takes place among bimodal bilinguals (simultaneously using a sign language and a spoken language). For example, Lillo-Martin et al. (2016, p. 734) note that bimodal children produce structures in which the spoken component (e.g., English), follows the word order of the signed component (e.g., American Sign Language [ASL]) as in Mommy where? “Where is Mommy?” Tis happens when they use both languages simultaneously, but children who are not bimodal never produce such structures in English.

Contact-induced changes in grammar and borrowing

121

What borrowing and convergence have in common is that bilinguals are consciously or unconsciously combining elements from both their languages. Let’s take a closer look at these two avenues of language contact, beginning with a discussion of borrowing.

8.3 THE CONSEQUENCES OF BORROWING: LOANWORDS In Chapter 4, you were introduced to the concepts of “borrowing” and “code-switching.” As we’ve defned it in this book, “code-switching” involves the shif between two linguistic systems rather than the integration of the forms of one language within another. In some cases, the patterns of code-switching become so common within a community that a mixed language is formed that combines grammatical features of each of its component languages, and it becomes the frst language of children raised in the community when they no longer hear the source language used by itself. We will examine the topic of mixed languages along with other types of languages that are formed by language contact in specifc social conditions in Chapter 10. In this chapter, our focus will be on the efect of “loanwords,” or “lexical borrowings,” that are added to a language over time. Loanwords can become so integrated in a language that speakers may no longer recognize that they stem from another language. For example, in one study, speakers of Hindi and English were asked to identify words as either Hindi or English. Tis would seem to be a simple task, and when it came to Hindi, it was. Participants in this study agreed that a word was Hindi 98% of the time. However, if a word was originally English, they agreed only 80% of the time. Tis means that the bilinguals felt that one out of every six English words was of Hindi origin (Diab & Kamboj, 2011). Te act of borrowing can be motivated by the speakers’ need to refer to new terms for novel cultural concepts, things, or technologies. Loanwords can also bring more specifcity or precision to existing concepts than do their source language equivalents (Backus, 2001). For example, el bullying, an English loanword found in Spanish, refers only to acts of aggression against children, while the Spanish equivalent, el abuso, has the more general meaning of any kind of mistreatment (Serigos, 2017). Te feeling that the source language holds more prestige than the recipient language can be a factor in borrowing, as well. But there are many documented cases of borrowing that cannot easily be described as motivated by need, specifcity, or prestige. For instance, for what reason would French-English Acadian bilinguals have borrowed the words pickles and sink to replace cornichons and évier, respectively, in French (Mackey, 1970)? Studies of loanwords across languages have confrmed that content words (nouns, verbs, adjectives, and adverbs) are more ofen borrowed than function words (pronouns, prepositions, articles) that serve a grammatical function (Haspelmath & Tadmor, 2009). Of all word classes, nouns are the most borrowable across languages. Tis is probably because nouns are rich in cultural content. However, as we will see, there are cases when words that perform only grammatical functions are borrowed, and this can have an impact on the internal structure of the recipient language.

122

From years to centuries

Over time and with enough frequency, borrowing can substantially change a language, sometimes creating distinct layers in its vocabulary. Te most well-known example is Japanese. For centuries, this language has borrowed extensively from Chinese, to the extent that about 60% of its vocabulary has roots in Chinese. Te vocabulary of Japanese is divided into four diferent classes: inherited Japanese terms, known as yamatokotoba; onomatopoeia words that mimic sounds or actions; Chinese borrowings (kango); and foreign words (gairago). Te sound and word structures of Japanese are diferent across these classes. So even though Chinese borrowings are ancient in Japanese, they still do not sound like other Japanese words, nor are they written like them. For instance, the Japanese writing system represents these classes in diferent ways. Te kanji writing system is used for representing kango words, and it uses Chinese characters. For the inherited Japanese vocabulary, hiragana is used. Hiragana is a phonetic system where each symbol represents the pronunciation of a syllable. Katakana, another syllabic system, represents gairago and onomatopoeia words. So, while it might seem that we can integrate words from another language with little efect on the structure of the recipient language, extensive borrowings can have a real impact on the structural patterns in a language and on the cultural practices, like writing, that are associated with the language.

8.3.1 Types of loanwords Te direct copying of a form and its meaning is not the only form in which borrowings can be added to a language. Sometimes only part of a word from the source language is borrowed, and this creates what is called a “loan-blend.” An example is frage biro “information ofce” in Helsinki Yiddish, spoken in Finland. It is a compound word, which means it is composed of two content words, one formed from the Yiddish stem, frage “question,” and the other from the Helsinki Swedish stem, bira “bureau” (Muir, 2000). Loan-blends do not always consist of two content words; they can also occur between a stem and an afx. For instance, Arabic mōbaylāt “cell phones” is formed by adding the Arabic plural sufx, -āt to the English loanword mobile. Loan-blends can also involve part of a word that is not a prefx or sufx. American Portuguese, which was spoken in New England (U.S.), used a partial blending with the Portuguese syllable al- (see the box on Arabic loanwords) in place of the word initial o- in the English word. For example, overalls is pronounced as alverozes (Haugen, 1950).

8.3.2 The linguistic adaptation of loanwords Te American Portuguese examples of alverozes from overalls might make you ask if there are other ways that the source language words change when they are borrowed. Let’s consider some examples of loanwords from a feld that provides a lot of cultural borrowings, that of food. If the borrowers are also speakers of the source language, or if the incoming form is a close match to the structures in the recipient language, then the word that enters the recipient language might refect a very close imitation

Contact-induced changes in grammar and borrowing

123

of the item in the source language, as is the case in example (4a), where the only difference is a slight opening of the vowel from Brazilian Portuguese to French, but the pronunciation is otherwise similar. (8.3) Loanword adaptations a. Portuguese [ɐsɐ’i] > French [asa’i] “a kind of berry” b. English ice cream [ajskrim] > Japanese aisukariimu [ajsɯkɯrui:mɯ] c. Vietnamese phở [fəː˧˧] > English [fʌ], [fɜ], [fo] [pʰo] “a kind of meat broth, soup” d. Haitian French Jean Marie [ʒɑ̃ma’ʁi] > Dominican Spanish: yamagui [dʒama’ɤi] “a kind of mango”

If borrowers are not very familiar with the structure of the source language, or if there are elements of the source language that are difcult for them to perceive or produce, then they might adapt the item to various degrees so that it conforms more closely to the shapes of the words that already exist in their recipient language. To illustrate this case, we can see that each consonant of the [skr] cluster of the English loanword in (4b) is followed by a vowel as a loanword in Japanese. Tis pronunciation now fts the sound system of Japanese that has restrictions on the number and type of consonant that can end a syllable. In the examples in (4c), we see diferent possible renditions in English for the Vietnamese soup, phở. Te tone of the Vietnamese source word is always ignored when the loan is adapted into English because recipient language listeners are not used to perceiving a diference in pitch as something that is essential to word meaning. Te variants used in English range from some that sound like the consonants and vowels of the Vietnamese word, [fʌ], [fɜ], [fo], to one that refects no familiarity with the source language pronunciation but is instead a reading pronunciation: [pʰo]. It is not uncommon that a language show variation between a reading pronunciation and a heard one. In Spanish as spoken in North America, you can hear some speakers use a reading pronunciation of the city Bufalo [ˈbufalo], while others use [ˈbafalo], which sounds more like English [ˈbʌfəlo]. In example (4d), we see that the consonants of the Haitian loanword, proper noun Jean Marie, are adapted consistently with the phonology of Dominican Spanish, as in the word yamagui “a kind of mango.” Haitian [ʒ] is strengthened to [dʒ] and the uvular fricative, [ʁ], is replaced by a velar one, [ɤ], in Dominican. Notice, too, that the two morphemes of the Haitian source word, Jean Marie, are reanalyzed to be a single unit in the Spanish. Reanalysis of the boundaries of a loanword is not uncommon. A classic example concerns the word for “orange,” which began with in the language from which it was thought to originate, probably the Indic language Sanskrit. Some modern languages retain a form that begins with : Farsi na̅ rang, Arabic naranj, Spanish naranja, Greek nerantzi, Croatian narančast. In others, the loanword refects the misanalysis of the as part of a preceding article: German Orange, Italian arancia, Sicilian arancione, Czech oranžovy, and of course English orange.

124

From years to centuries

8.3.3 The impact of loanwords on the structure of a recipient language Loanwords can sometimes be integrated with little to no impact on the structure of the recipient language. For instance, the grammatical structure in (8.4a) with three borrowed words—chai (Hindi); barista (Italian); café (French)—is identical to that in (8.4b), a sentence that contains words of only English origin: (8.4) Loanword insertions with no change to grammatical structure a) Order me a chai from the barista at this café, please. b) Order me a hot dog from the waiter at this diner, please.

Borrowing can also introduce structures that are novel in the recipient language. For instance, afer the conquest of England by the Normans in the twelfh century, there was a massive importation of French loans into English. Today, about 30% of the vocabulary of English has French roots. Te infux of French words brought new words into the language with the sound /v/ at the beginning of a word, where it had not existed before. Tis had an impact on the English sound system, allowing new pairs of words that difer in meaning like vine and fne and very and ferry. More recently, borrowings from Yiddish added word-initial consonant clusters into English beginning with the sound [ ʃ ] that did not exist before this contact: schlep [ ʃlɛp], schtick [ ʃtɪk], schmooze [ ʃmuz]. Tere are examples like this in many languages, where sounds or combinations of sounds can only occur in borrowed words (Hall, 2013). Loanwords can alter the patterns of a language in other ways. For example, in languages that mark gender, borrowed nouns must be assigned into a gender class. Tis can create variability in usage, resulting in diferent gender assignment rules across varieties. For example, Sephardic Jews speaking a language called Judeo-Spanish or Ladino were in centuries-long contact with Turkish as part of the Ottoman Empire and borrowed many terms from both Hebrew (an important language for religious purposes) and Turkish (an important language for communication with outsiders). As in other varieties of Spanish, borrowed words for animate beings are classifed as masculine or feminine according to their sex. For inanimate nouns, words that end in -a are assigned to the feminine class, like in most varieties of Spanish. However, in Judeo-Spanish, borrowed words that end in stressed -á, as well as unstressed -a, are categorized as feminine, so Turkish çorba “soup” becomes la-FEM chobá, as can be seen here in the use of the feminine defnite article la. So, the presence of Turkish and Hebrew loanwords brought about new gender rules in the Judeo-Spanish variety of Spanish (Romero, 2009). Recall from Chapter 4 that when bilinguals borrow verbs, they need to fnd ways to conjugate them. Sometimes diferent speakers of varieties of the same language choose diferent strategies in a contact setting. For example, in 8.5a, the verb drive occurs with the regular verbal infnitive -er ending in Chiac, a variety of French in contact with English, spoken in Prince Edward Island, Canada (Comeau & King, 2011). In 8.5b, the same English verb is used in Cajun French, spoken in Louisiana, as a past participle without the regular French past participle ending -é (Picone, 1997).

Contact-induced changes in grammar and borrowing

125

As you can see, varieties of the same language can evolve in diferent ways due to language contact. (8.5) drive as a loanword in varieties of North American French (a) Tu es en train de driver? You are in process of driving “You are driving.” (b) On a drive en ville We PAST drive in city “We drove in town.”

Borrowing can also alter the syntactic structure of a language, as shown by the Moroccan Arabic of the region of Oujda, where Arabic is in contact with Berber (Lafioui, 2013). In Moroccan Arabic, the verb is surrounded by two negative markers: ma + verb + š. In Oujda, the second marker of negation is borrowed from the Berber variety, Taraft: ma + verb + bu, seen in example 8.6a. Te borrowing of the negation marker ma __ bu introduced a new grammatical pattern in the Oujda Arabic because any verb negated with ma __ bu must also be followed by a determined noun (‘sheep’ in the examples in 8.6). Tis pattern is a word order match of the Taraft example in 8.6a, and it difers from other varieties of Moroccan Arabic (Lafioui, 2013, pp. 80–81): (8.6) Oujda Arabic and Taraft Berber negation (a) Oujda Arabic ma sra-w bu l-.awli NEG PERF-3PL NEG DEF-sheep ‘Tey did not buy a sheep this year.’ (b) Taraft Berber u ssgi-n bu izmər NEG PERF-3M.PL NEG sheep ‘Tey did not buy a sheep this year.’

had PROX

l-ɛam. DEF-year

asəggɀas-a. year-PROX

Te use of loanwords can also tell us something about the way bilinguals narrate a story. Kildin Saami is a Uralic language of northwest Russia that has been in longterm contact with Russian. When people talk in Kildin Saami, they use Russian words to connect their ideas, so that the equivalents of adverbials like “then” or “in my opinion”; discourse connectors like “and,” “but,” “so”; and grammatical words that introduce subordinate clauses like “which” and “that” are ofen used in Russian even though the conversation is in Kildin Saami (Hakimov & Rießler, 2021). Tis means that the speakers are using Kildin Saami words but organizing their ideas in a Russian narrative frame. In this section, we have seen the ways that loanwords can impact more than the vocabulary of a language. Tey can afect any aspect of a language, from the types of sounds and the structures that speakers use to the way that speakers come to narrate a story in their language. In the next section, we’ll see that similar things happen when meaning and patterns, rather than words, converge in the minds of bilinguals.

126

From years to centuries

Borrowing in language genealogy studies “Glottochronology” (a term based on Greek “language” and “chronology”) is a method that was elaborated in the 1950s to date the points of separation of languages that were related. Tis method estimated that every 14 non-cognate pairs out of 100-word lists corresponded to a thousand-year separation between languages. Te most well-known word list is the one elaborated by Swadesh (known as the “Swadesh list”). Swadesh’s word list was elaborated based on the assumption that some concepts are more stable than others in that speakers would not borrow the word from another language. Such core words were thought to be those for body parts and generally body-related activities, close kin terms, pronouns, interrogatives, and basic concepts for nature and geography. Tis assumption was based on the idea that borrowings are flling gaps in the lexicon. Nowadays, there are alternative methods that no longer exclude borrowings when establishing language histories, since we now understand that language contact is ubiquitous (Holman et al., 2007). We are, in particular, in a position to show that diferent contact settings will impact diferently the types (and possibly rates) of change that languages may undergo. For example, in settings with “patrilocal exogamy” (where women move into their spouse’s group), we are more likely to fnd structural and phonological or phonetic features because women in these contexts are second-language speakers of their spouse’s in-group communication language. But in contexts in which we fnd persistent trade relations, we are more likely to fnd lexical borrowings for words introduced together with the traded artifacts.

8.4

CONVERGENCE

Contact-induced innovations occur in contact communities because bilingual speakers know that when they use words, structures, or meanings from a source language into the recipient language, they will be understood by other bilinguals of these languages. While code-switching and borrowing are obvious forms of contact-induced variation, convergence is sometimes more difcult to detect because people are communicating with all recipient language words, only they do so with a source language favor. Tis favor might be refected in their pronunciation and intonation, in their word choices, or in the way they use sentence structures.

8.4.1 Quantitative convergence: convergence as a matter of degree Convergence does not have to result in a completely new way of saying something. Instead, speakers tend to use certain structures more ofen in bilingual communities than speakers in monolingual ones do. Tis was shown in a study of French speakers who resided in communities in Ontario, Canada, that difered according to the percentage of the population within them that identifed as French speaking. Ontario

Contact-induced changes in grammar and borrowing

127

French shows variation between two structures that use diferent prepositional structures to express a similar meaning, as shown in the examples in (8.7). (8.7) Ontario French (Beniak et al., 1985) (a) J’ reste à I stay.1p.sg at “I say home” (b) J’ ai resté I aux stay.pp “I stayed home”

la maison (the) home chez at

moi mine

Te preposition chez, in example 8b, is used with a wide variety of meanings in French, so it can have several translations. For instance, un an chez les ados would translate as “a year among adolescents” rather than “a year at/in the home of adolescents.” Te preposition à in example 8a translates to English “to” or “at,” and it is used in the same syntactic context as English, à la maison = at (my) home. Both chez and à la maison are used in Ontario to express the idea of being at home, but the researchers of the study found a clear pattern of use: the more bilingual speakers there are in a community, the greater the chances that a francophone speaker will use à la maison. In other words, bilinguals prefer the structure that matches with English.

8.4.2

Qualitative convergence: a case study of Spanish in Texas

Sometimes, convergence results in new types of uses in the recipient language. So, rather than just being diferent by a matter of degree, as in the Ontario case, a contact variety becomes qualitatively diferent in some way from its source variety. For evidence, let’s look at another Romance language variety that has been in long-term contact with North American English, the Spanish of Texas (Bullock et al., 2021). Spanish speakers have been present in Texas since colonial times, well before English speakers reached the territory in 1821. Today, the shared border between Mexico and Texas is 2,000 kilometers long, and many counties of Texas house a population that is majority native Spanish speaking. Given generations of Spanish-English bilingualism in this border state, the varieties of Spanish spoken across the state show signs of contact-induced changes like borrowing and structural convergence. Te study looked at one construction in the Spanish of Texas that appeared to be a “loan translation,” also known as a “calque,” of English.A “loan translation” is a word or phrase that copies the meaning of an equivalent expression from a source language into a recipient language (Backus & Dorleijn, 2009). Te apparent loan translation used the Spanish verb agarrar, which means “to grab, to grasp” in ways like English “to get.” For example, speakers could say agarrar buenas notas “to get good grades.” Other varieties of Spanish would use a diferent verb, sacar“to take out, to pull out,” for this meaning: sacar buenas notas. We know that the meanings of words can change over time without any infuence from language contact. Tis was the case with the meaning of the word magazine, discussed in section 8.1. What the researchers sought to fnd out is whether the convergence of the meaning of agarrar with English “to get” was due to language contact. To

128

From years to centuries

examine how the structure is used, the researchers consulted a corpus of transcriptions of recordings from speakers around the state and counted every use of agarrar and any other verb that might be used in constructions where English speakers use “to get,” such as sacar (Bullock & Toribio, 2013; Bullock et al., 2018). Tey compared these counts to the counts they found of these words in Mexican, Spanish, and Argentine corpora. What they found is that both Mexico and Texas show much higher rates of the use of agarrar than the other corpora do. In fact, Spanish speakers from Spain show 80% fewer tokens of this verb than do Mexico and Texas Spanish speakers! Researchers also found that those speakers who were born in the United States and who rated themselves as having weaker reading and writing skills in Spanish than in English were more likely to use agarrar than others within the Texas corpus. Tis points to a contact efect because, like the Ottawa case, those who use more English were more likely to use the loan translation. But, also, the speakers from Texas used agarrar in a way that was qualitatively diferent from all other Spanish speakers, including their Mexican neighbors. Only the Texas speakers used the innovative verb with abstract nouns like agarrar trabajo “get a job,” agarrar crédito“get credit.” Only for these speakers, then, was agarrar semantically equivalent to English “to get.” So, just like the Spanish of Mexico and Texas have moved away from the varieties in Argentina and Spain, Spanish in Texas has moved away from Mexico due to the prolonged contact of Spanish speakers with English. Tis example of convergence, like the Judeo-Spanish example of borrowing, shows you how language contact can bring about diversity in language structure over time.

8.4.3 The effect of convergence on linguistic structure Extending the meaning of the verb agarrar in the Spanish of Texas might seem to only impact the lexicon of a language, like adding an additional meaning to a word already in the vocabulary. However, this contact-induced change also afects the grammar of Spanish, making it very diferent from other varieties. Tis is because the verb is no longer limited to being used with things that you can grasp, like animals, steering wheels, or ideas. In Texas, the verb is, in part, still a main verb with its core meaning of “grasping” like in other varieties of Spanish, but it has also become a support verb. As a support verb, it is semantically bleached, having lost its core meaning in certain phrases so that the noun carries the meaning in the innovative construction agarrar + noun. When content words become grammatical elements over time, the process is known as “grammaticalization” in historical linguistics. Te Spanish in Texas example shows us that grammaticalization can be contact induced. Grammaticalization of elements in a recipient language is frequent in language contact, where a word or a morpheme can be used to provide meanings from the source language that are not always easily expressed in the recipient language. Bao (2005) shows that the English word ever is used in Singapore English as equivalent to the aspectual meaning,“happened at least once in the past.” In the Chinese source language marker, this meaning is expressed by a morphological marker, guo.

Contact-induced changes in grammar and borrowing

129

(8.8) Singapore English (Bao, 2005) (a) I ever try this type of fruit before. “I have tried this type of fruit before.” (b) wǒ chı̅ guo Rìben fàn “I eat guo Japanese food (before).”

Since a distinct marker to express the meaning of guo is missing in English, bilingual speakers feel the need to create one, and they do so by choosing ever as a calque for Chinese guo. As Bao (2005, p. 237) describes it, “Singapore English is essentially the Chinese system fltered through English morphosyntax.” Such fltering can be detected at all levels of a contact language. In the domain of the sound system, we can point to many examples where the recipient language bears the traces of the patterns of the source language. As examples, the intonation contours of the Spanish spoken by bilinguals in the island of Mallorca, where Catalan is dominant, has converged toward the intonation patterns of Catalan (Simonet, 2011), just like the distinctive “Italianized” intonation of Argentine Spanish followed from the large-scale immigration of Italians in the nineteenth and twentieth centuries. At the level of sound segments, researchers have shown that phonetic convergence can be pervasive in long-term language contact settings. For example, the vowels of Welsh and the vowels of English, once distinct, have apparently converged in Wales due to long-term contact (Mayr et al., 2017), and, in some parts of the state of Wisconsin, the American English spoken there still bears phonetic traces of the German that was once widespread within immigrant enclaves in the area (Purnell et al., 2004). Perhaps there is no clearer example of the efect of convergence on the sound system of a contact language than the features that can be heard in English spoken in the countries that have a colonial connection to the United Kingdom and the United States of America. In these areas, the phonological system is fltered through the source languages, so we fnd, for example, that Indian English speakers merge [w] and [v] and retrofex the consonants [t] and [d], while speakers of Nigerian English use full vowels in syllables where we usually fnd reduced or deleted vowels, like the word police. Tese are all cases of convergence, where features of the source language are present in the recipient language.

8.5 HOW TO STUDY CONTACT-INDUCED CHANGES IN GRAMMAR AND BORROWING While early studies examined language contact phenomena in individual language pairs, since the 2000s, corpus-based approaches to language contact and the study of borrowing have become more central, as there are more and more digital corpora. You can work on written corpora, as they are typically large and allow for statistical analyses that increase the validity of your generalizations. You can also work with smaller spoken corpora that allow you to analyze the spontaneous productions of bilingual speakers. Such corpora ofer access to the phonetic details that may be crucial in determining the status of an item as a borrowing.

130

From years to centuries

To help you with the analysis, you can either apply automatic language tag models using n-gram methods that calculate the probabilities of character sequences to belong to a given language and look-up methods that check the languages’ dictionaries (Guzman et al., 2016). For less-studied languages, such tools for automatic annotation are not always available. In that case, you need to annotate and manually code your corpus. Depending on your research questions, you need to phonetically annotate the corpus and ofer an annotation for each word and morpheme.You can also tag the words depending on the etymologies of the languages involved: L1, L2, multiple or ambiguous, mixed, other language, or unknown (Adamou, 2016). At the level of the analysis, you can either look at tokens (e.g., how many instances of a specifc word you fnd in the corpus) or types (e.g., how many adjectives you fnd in the corpus). Here are some ideas about things you can study: Study 1. Calude and colleagues (2017) analyzed a million-word spoken corpus of New Zealand English. Their analysis shows that Maori borrowings are very frequent in New Zealand English, even though Maori is a minoritized language! To account for this finding, authors remind us that Maori has been revitalized since the 1970s, a process that was embraced by New Zealanders. Indeed, the authors show that higher use of Maori borrowings signaled support for the Maori identity. In addition, they show that Maori borrowings were more likely to be used when they had a cultural rather than a core meaning, when they did not have a clear English equivalent, and when they were content words rather than function words. Study 2. Zenner et al. (2015) analyzed a 35-hour spoken corpus from a reality TV show and examined the productions of 46 speakers of Dutch. Tey found that English insertions into Dutch were preferred by younger, more educated men. Tey also found that English swear words were very frequent in contexts charged with negative emotions. Study 3. Serigos (2017) analyzed English borrowings in a 24-million-word newspaper corpus of Argentine Spanish. She coded the “specifcity” of each borrowing by examining its linguistic context. Te hypothesis was that more specifc words would have less variation in their surrounding context. Te analysis confrmed the specifcity hypothesis: Serigos found that English borrowings are used in contexts that were similar and show less variation in comparison to their Spanish counterparts. Te author interpreted this as indirect evidence for the motivation of speakers to nuance the original concept. Study 4. Hakimov (2021) analyzed a Russian-German free-speech corpus of Russian immigrants living in Germany. Among many things, the author showed that usage frequency in the donor language predicts the insertion of infected words in bilingual sentences. For example, nouns are more likely to retain plural infection when they are more frequently used in plural (as opposed to singular form) in monolingual speech.

Contact-induced changes in grammar and borrowing

131

Chapter summary In this chapter, we reviewed the two types of contact-induced changes in language: borrowing and convergence.Each type of change is introduced at the level of an individual bilingual and then spreads through a community until it becomes anchored as part of the recipient contact language.We have examined the efects of lexical borrowing as they percolate through a linguistic system, afecting every level of a grammar, from the pronunciation and lexicon to the way that speakers organize their discourse. We have also seen that each type of contact-induced change can leave an imprint on every level of the recipient language.

EXERCISES 1.

Discover

Te words in the following table all came into English from diferent source languages.What do these words mean in English today? In what way do their meanings as loanwords difer from the meanings in the source language? Word in English

Source Language

Meaning in Source Language

chai

Hindi

tea

kebab

Arabic

grilled meat

safari

Swahili

travel

boondocks

Tagalog

mountains

tycoon

Japanese (kango)

great nobleman

2.

Discover

Sometimes researchers attempt to fgure out the frst language of individuals who are writing or speaking in their second language. How might English be fltered through another language that you know or study? Make a list of three features that might be afected by the second language you have in mind. Ten search social media to see if you can fnd these structures in use. Can you determine from the user profle the languages that the speaker uses?

3.

Discuss

Cognates refer to items that overlap in form and meaning across languages. False cognates are words that only overlap in form but not in meaning. For instance, carpeta in Spanish generally means “folder,” rather than “carpet,” and asistir means “to attend” rather than “to assist.” Sometimes bilinguals begin to use such words with the meaning of their false cognates; for example, carpeta can also be used to mean “carpet” in some Spanish-English bilingual communities. How do these examples ft into our

132

From years to centuries

two-way classifcation of contact-induced change? Is this an example of borrowing or an example of convergence? Or do we need another category?

FURTHER READING Poplack, S. (2017). Borrowing: Loanwords in the speech community and in the grammar. Oxford University Press. Te author of this book endeavors to show how borrowing difers from code-switching with data from many languages. Trefers-Daller, J. (2009). Code-switching and transfer. In B. Bullock & A. J. Toribio (Eds.), Te Cambridge handbook of linguistic code-switching. Cambridge University Press. Tis chapter provides an accessible introduction to convergence and its relationship to other forms of mixing.

REFERENCES Adamou, E. (2016). A corpus-driven approach to language contact: Endangered languages in a comparative perspective. De Gruyter Mouton. Backus, A. (2001). Te role of semantic specifcity in insertional code-switching: Evidence from Dutch-Turkish. In R. Jacobson (Ed.), Code-switching worldwide II (pp.  125–154). Mouton de Gruyter. Backus, A., & Dorleijn, M. (2009). Loan translations versus code-switching. In B. E. Bullock, & A. J. Toribio (Eds.), Te Cambridge handbook of linguistic code-switching (pp. 75–93). Cambridge University Press. Bao, Z. (2005). Te aspectual system of Singapore English and the systemic substratist explanation. Journal of Linguistics, 41(2), 237–267. https://doi.org/10.1017/S0022226705003269 Beniak, É., Mougeon, R., & Valois, D. (1985). Sociolinguistic evidence of a possible case of syntactic convergence in Ontario French. Linguistica Atlantica, 73–88. Bullock, B. E., Serigos, J., & Toribio, A. J. (2021). Exploring a loan translation and its consequences in an oral bilingual corpus. Journal of Language Contact, 13(3), 612–635. Bullock, B. E., Serigos, J., Toribio,A. J., & Wendorf,A. (2018). Te challenges and benefts of annotating oral bilingual corpora: Te Spanish in Texas corpus project. Linguistic Variation, 18(1), 100–119. Bullock, B. E., & Toribio, A. J. (2013). Te Spanish in Texas corpus project. Center for Open Education Resources and Language Learning (COERLL). University of Texas at Austin. https://corpus.spanishintexas.org/en/about-corpus Calude, A. S., Miller, S., & Pagel, M. (2017). Modelling loanword success—A sociolinguistic quantitative study of Maori loanwords in New Zealand English. Corpus Linguistics and Linguistic Teory, 16(1), 29–66. https://doi.org/10.1515/cllt2017-0010 Comeau, P., & King, R. (2011). Media representations of minority French: Valorization, identity, and the Acadieman phenomenon. Te Canadian Journal of Linguistics/La Revue Canadienne de Linguistique, 56(2), 179–202. Diab, M., & Kamboj, A. (2011). Feasibility of leveraging crowd sourcing for the creation of a large scale annotated resource for Hindi English code switched data: A pilot annotation. DTIC Document. http:// oai.dtic.mil/oai/oai?verb=getRecordandmetadataPrefx=htmlandidentifer=ADA562521 Guzman, G. A., Serigos, J., Bullock, B. E., & Toribio, A. J. (2016). Simple tools for exploring variation in code-switching for linguists. In M. Diab, P. Fung, M. Ghoneim, J. Hirschberg, & T. Solorio (Eds.),

Contact-induced changes in grammar and borrowing

133

Proceedings of the second workshop on computational approaches to code switching (pp. 12–20). ACL. Hakimov, N. (2021). Explaining Russian-German code-mixing: A usage-based approach. Language Science Press. https://langsci-press.org/catalog/book/289 Hakimov, N., & Rießler, M. (2021). Partial fusion in long-term bilingualism: Te case of vernacular Kildin Saami. International Journal of Bilingualism, 25(2), 401–424. https://doi. org/10.1177/1367006920924959 Hall, K. C. (2013).A typology of intermediate phonological relationships. Te Linguistic Review, 30(2), 215–275. https://doi.org/10.1515/tlr-2013-0008 Haspelmath, M., & Tadmor, U. (Eds.) (2009). Loanwords in the world’s languages: A comparative handbook. Mouton de Gruyter. https://doi.org/10.1515/9783110218442 Haugen, E. (1950). Te analysis of linguistic borrowing. Language, 26(2), 210–231. https://doi. org/10.2307/410058 Holman, E. W., Schulze, C., Staufer, D., & Wichmann, S. (2007). On the relation between structural diversity and geographical distance among languages: Observations and computer simulations. Linguistic Typology, 11, 393–421. Jarvis, S., & Pavlenko, A. (2007). Crosslinguistic infuence in language and cognition. Routledge. Lafioui, M. (2013). Reinventing negation patterns in Moroccan Arabic. In M. Lafioui (Ed.), African Arabic: Approaches to dialectology (pp. 51–94). De Gruyter. https://doi.org/10.1515/97831102 92343.51 Lillo-Martin, D., Müller de Quadros, R., & Chen Pichler, D. (2016). Te development of bimodal bilingualism: Implications for linguistic theory. Linguistic Approaches to Bilingualism, 6, 719–755. Mackey, W. F. (1970). Interference, integration and the synchronic fallacy. www.eric.ed.gov/ERICWebPortal/recordDetail?accno=ED042139 Matras,Y., & Sakel, J. (2007). Grammatical borrowing in cross-linguistic perspective.Walter de Gruyter. Mayr, R., Morris, J., Mennen, I., & Williams, D. (2017). Disentangling the efects of long-term language contact and individual bilingualism: Te case of monophthongs in Welsh and English. International Journal of Bilingualism, 21(3), 245–267. https://doi.org/10.1177/1367006915614921 Muir, S. (2000).Yiddish in Helsinki and its Swedish component. Nordisk Judaistik/Scandinavian Jewish Studies, 21(1–2), 139–148. https://doi.org/10.30752/nj.69573 Picone, M. D. (1997). Code-switching and lack of infection in Louisiana French. In C. Bernstein, T. Nunnally, & R. Sabino (Eds.), Language variety in the south revisited (pp. 152–162). University of Alabama Press. Purnell, T., Salmons, J., & Tepeli, D. (2004). German substrate efects in Wisconsin English: Evidence for fnal fortition. American Speech, 80(2), 135–164. Romero, R. (2009). Lexical borrowing and gender assignment in Judeo-Spanish. Ianua. Revista Philologica Romanica, 9, 25–35. Serigos, J. (2017). Using distributional semantics in loanword research: A concept-based approach to quantifying semantic specifcity of Anglicisms in Spanish. International Journal of Bilingualism, 21(5), 521–540. https://doi.org/10.1177/1367006916635836 Simonet, M. (2011). Intonational convergence in language contact: Utterance-fnal F0 contours in Catalan-Spanish early bilinguals. Journal of the International Phonetic Association, 41(2), 157–184. Tomason, S. G., & Kaufman, T. (1988). Language contact, creolization, and genetic linguistics. University of California Press. Trefers-Daller, J. (2009). Code-switching and transfer. In B. E. Bullock, & A. J. Toribio (Eds.), Te Cambridge handbook of linguistic code-switching. Cambridge University Press. van Coetsem, F. (1988). Loan phonology and the two transfer types in language contact. Foris Publications.

134

From years to centuries

Weinreich, U. (1953). Languages in contact. Mouton. Zenner, E., Speelman, D., & Geeraerts, D. (2015). A sociolinguistic analysis of borrowing in weak contact situations: English loanwords and phrases in expressive utterances in a Dutch reality TV show. International Journal of Bilingualism, 19, 333–346.

9 Linguistic areas

CHAPTER PREVIEW What happens when several languages are in long-term contact with one another? In this chapter, we see that languages can become more similar to one another through prolonged contact. Tis is captured by the term “linguistic area.” Te most well-known linguistic area is the Balkans, but more recently several authors have drawn attention to the existence of numerous linguistic areas around the world, as in North America, Amazonia, West Africa, Anatolia, South-East Asia, and Melanesia.

9.1 WHAT ARE LINGUISTIC AREAS? In the frst part of the book, we have seen processes that take place at the level of an individual who uses two languages. In the second part, we have seen what happens when language users communicate. We will now turn to see what happens beyond the level of individual language users. We will focus on efects that are observed in geographically large areas, encompassing many bilingual speakers and many language pairs. Te observation that some unique outcomes take place when bilinguals are in contact for a long time is not recent and emerged rather early in the history of modern linguistics. Nikolaï Trubetzkoy, a Russian linguist who taught at the University of Vienna, Austria, proposed in 1923 and then in 1928 at the First International Congress of Linguists that there might be two kinds of “language group(ing)s” (Sprachgruppen in German, based on the Russian term jazykovyje gruppy): “language family” (Sprachfamilie in German, translated from Russian jazykovaja sem’ja) and “language/linguistic union/league” (Sprachbund in German, from Russian jazykovoj sojuz). Language families were of course well-known to linguists, so the originality of Trubetzkoy’s idea is that a number of languages, while not necessarily belonging to the same language family, could still be viewed as belong to the same language group if they share some common linguistic features that had arisen not randomly but due to contact. Weinreich (1958) later coined the term “convergence area,” stressing the dynamic aspect of the process (through the term “convergence”) and introducing the DOI: 10.4324/9781003167952-12

136

From years to centuries

relevance of the geographic aspect (through the term “area”). At present, “linguistic area” is the most widespread term in English-speaking literature. Yet linguists struggle to adequately defne linguistic areas. Joseph ofers a good defnition: A Sprachbund can be defned as any group of languages that due to intense and sustained bilingual contact share linguistic features, largely structural in nature but possibly lexical as well, that are not a result of shared inheritance from a common ancestor nor a matter of independent innovation in each of the languages involved. (2010, p. 620) In this defnition, the geographical parameter is not explicit. Indeed, linguistic areas do not neatly correspond to geographic areas. Another issue is the number of languages needed to identify a linguistic area. Clearly more than one language is needed to form a group, but are two enough? Te compromise is to say that a group of languages is needed. Te number of language families involved is also problematic. Given that Trubetzkoy contrasts language family groups to language unions, it is understood that for a linguistic area to exist, we need to have languages from diferent linguistic families. Te issue can be settled if we focus on the linguistic features that have become more similar, as long as they don’t come from a common ancestor, as Joseph notes. But how many linguistic features are needed to consider that there is a linguistic area? Joseph rightly glosses over this issue, the idea being the more the better.And what kind of linguistic features are needed? In this defnition, structural features are a prerequisite, with lexicon being an option. In the face of these challenges, some authors take a diferent stance and suggest that “We should abandon the search for a defnitive defnition of ‘linguistic area’” (Campbell, 2006, p. 21). Ultimately, as Matras notes: Linguistic areas are therefore not real-life entities. Rather, they are constructions by linguists, who choose to grant their attention to situations in which, as a result of socio-historical coincidences, a series of conditions are met, and to label this kind of situation in a particular way. (2009, p. 274) Despite the difculty of convincingly defning linguistic areas in a convincing manner, the topic remains very popular in contact linguistics. Indeed, modern linguists prefer to pay attention to the processes of convergence and relate them, when possible, to socio-cultural contact. Let’s begin by introducing the most well-known linguistic area: the Balkans.

9.2 THE BALKANS AS A LINGUISTIC AREA Te Balkans are a geographic area in the southeastern part of Europe ofen delimited to the north by the Danube River. However, rather than a geographic region, the Balkans are best understood as a socio-political and cultural area where people

Linguistic areas

137

share a centuries-long common history, in particular during the Byzantine Empire (fourth to ffeenth centuries) and Ottoman Empire (ffeenth to nineteenth and early twentieth centuries). Within this socio-political space, multilingualism was widespread and constant over time. Te modern Balkan states have been reshaped multiple times, including recently following the Yugoslav wars in the early 1990s, and yet despite the current national borders, exchanges and contact in the Balkans are ubiquitous. Interestingly, the empirical observation that some Balkan languages shared some linguistic features preceded the theorization of the Sprachbund as proposed by Trubetzkoy (see Section 9.1). Indeed, the earliest works on the similarities between Balkan languages appeared in the late nineteenth and early twentieth centuries (see Miklosich, 1861 and Sandfeld, 1930 [1926]). Tese scholars listed several phonetic and phonological, morphological, syntactic, and lexical phenomena that were common in some of the languages of the Balkans. Most languages of the Balkans that are also considered “Balkan languages” belong to the Indo-European family, albeit to various branches: the South Slavic branch includes Macedonian, Bulgarian, and many nonstandard Balkan Slavic varieties; the Romance branch includes the Balkan Romance languages such as Aromanian and Meglenoromanian, as well as Romanian, but also Balkan Judeo-Spanish or Ladino, a Spanish variety spoken by Jews who arrived at the Ottoman Empire afer being chased from Spain; the Albanian language and to some extent the Greek language; and more signifcantly some Greek non-standard varieties in Northern Greece, as well as some of the Romani varieties belonging to the Indic branch of Indo-European and possibly also Western Armenian varieties. The non-Indo-European languages belong to the Turkic family (Sobolev, 2004; Friedman, 2021). Tere have been many discussions about the defnition of the Balkans as a linguistic area. On the one hand, Balkan linguistic features (also known as Balkanisms) are not necessarily restricted to the Balkans but can be found in other languages of Europe and the Middle East. So why draw a linguistic boundary at that specifc level? In a way, this question parallels the ambiguity in delimiting the Balkans as a geographic or cultural region since they are also part of a larger geographic and cultural area, namely Europe. Indeed, some Balkan states are still members of the European Union while forming smaller regional alliances, and speakers are in contact within the Balkans but also beyond. On the other hand, Balkan linguistic features are instantiated diferently in diferent language varieties, and multilingualism is distributed diferently depending on the regions in the Balkans. In that sense, it is more accurate to say that, rather than a homogeneous linguistic area, one could identify a number of small linguistic areas that compose the larger Balkan linguistic area. Wiemer (2004) nicely captures this superposition of areas by the metaphor of the Russian matryoshka dolls, where a big doll contains a smaller one that contains a smaller one and so on. Despite this complexity of superimposed areas and heterogeneity, researchers agree that there is good reason to consider the Balkans a linguistic area, as inhabitants were traditionally multilingual and had many ties across linguistic communities, as attested in many historical documents, and linguistic changes can ofen be traced

138

From years to centuries

in the written documents of the languages. In addition, comparison between the languages of the Balkans and languages of the same branch that are spoken outside of the Balkans further confrms the unique linguistic developments that took place in the Balkans. For example, linguists noted the rise of a modal future particle based on the volitive “want.” Tis feature is found in Greek, (Tosk) Albanian, Romanian, Macedonian, Bulgarian, Serbian and Croatian, and Romani (Joseph, 1992, p. 154). Written sources show that the rise of the “want”-future followed a similar grammaticalization path starting with a modal verb (infected for person and number), then becoming an auxiliary (free word order), then becoming phonologically reduced as a clitic (fxed word order) and fnally being used for modal future at some point between the fourteenth and seventeenth centuries (Assenova, 2002). Another well-known Balkan convergence phenomenon concerns sentences like “I want to know.” Speakers of many Balkan languages use an optative particle (e.g., Balkan Slavic da; Albanian të; Romani te; Balkan Romance să, si, s’; Greek na) and a fnite verb (i.e., a verbal complement that has tense and subject marking) following the loss or reduction of the infnitive (i.e., a verbal complement that is nonfnite, that is, has no tense or subject marking), a process that is frst described in Joseph (1983). Infnitive reduction is also noted in Balkan Turkish,Armenian, Judeo-Spanish/Ladino, and Judeo-Italian. Te use of an enclitic defnite article is another convincing result of convergence. Indeed, very few Slavic languages have grammaticalized defnite articles, and are those spoken in the Balkans. It is argued that the grammaticalization of postposed articles in South Slavic languages (as in Macedonian, Bulgarian, Torlak Serbian, and other Balkan Slavic varieties) results from both internal and contact-induced factors, as it coincides with a similar development in the Romance languages of the area (e.g., Romanian, Aromanian, and Meglenoromanian) and Albanian (Assenova, 2002). Once again, it is possible to trace the development of the defnite articles thanks to the written sources showing, for example, that the postposed demonstratives attested in Old Church Slavonic documents (between the ninth and eleventh centuries) were grammaticalized into clitic demonstratives and then into clitic articles (Mladenova, 2007). Regarding the lexicon, Friedman and Joseph (2023) summarize the trend as borrowings that are essentially rooted in conversation (ERIC). ERIC borrowings are generally lexical items that are either known to be rarely borrowed such as pronouns, numerals, kinship terms, and bound morphology or are more frequently borrowed lexical items such as discourse particles, interjections, and taboo expressions, among others. For example, in many languages, speakers use the structure “what+do.2SG/ other” to say “How are you?”: Greek ti kaneis?, Albanian ç’ka po bën?, Macedonian š[t]o pra[v]iš?, Bulgarian kakvo praviš?, Romani so kere[s]?, Romanian ce mai faci?, Aromanian tsi fats?, and Turkish n[e] [y]aparsın? Many languages share the same exhortative “c’mon!” from Turkish hay de (e.g., Albanian, Bulgarian, Romani, JudeoSpanish/Ladino, Greek . . .). Or the same word in agreement “just right” from Turkish tamam (e.g., Albanian, Aromanian, Romani, Romanian and Greek varieties . . .).

Linguistic areas

139

Tink of how widespread English “okay” and “bye” are today in our globalized world. Te same infuence was noted in the Balkans during Ottoman times. Finally, as Friedman (2021) notes, phonology ofers a diferent view, with several localized processes of convergence rather than a single process that would apply to all Balkan languages. In sum, the Balkans ofer a wonderful example of how languages that were not closely related (in that they were part of diferent branches) have structurally converged over the centuries. Tis convergence was driven by the multilingualism of the people of the Balkans, who in many cases lived in the same villages and cities and formed linguistically mixed households. Te historical written documents allow linguists to follow the development of a given phenomenon closely, demonstrating that the convergence process also involves internal processes specifc to each language.

9.3 AN OVERVIEW OF LINGUISTIC AREAS ACROSS THE WORLD While the Balkans are the most well-known linguistic area, many linguists have noted that there were similar convergent processes across the world. Tese diverse contact scenarios add to the discussion about linguistic areas. In what follows, we will focus on the two most linguistically diverse areas in the world: Amazonia and Melanesia. Although in both settings there is very little historical evidence, everything we know about the ways people live, their linguistic practices, and their languages’ contemporary grammars indicates that multiple language contact processes may have shaped the languages of these areas. Let’s start with “linguistic Melanesia,” which hosts some 1500 languages belonging to 20 to 40 language families, depending on classifcations. It is composed of speakers of Papuan languages (the term roughly refers to any language in the area that is not Austronesian or Australian and that may belong to a variety of language families) and Austronesian languages. Linguistic Melanesia has its center in New Guinea, probably the world’s most important hotspot, but goes beyond it to include several islands at the east and west. Language contact in this area is thought to have been ubiquitous without leading to language shif toward a dominant language. Schapper (2021) summarizes the linguistic features that characterize linguistic Melanesia. Te author notes that speakers of Papuan languages generally use a subject-object-verb order (SOV); literally “the child the banana ate” meaning “the child ate the banana.” What is interesting is to observe that SOV is only found among speakers of the Austronesian languages that are in contact with Papuan languages. In addition, the order of the noun and its possessor (e.g.,“the child’s father”) is an interesting feature, with Papuan languages using the possessor before the noun.Again, researchers note that Austronesian language speakers in contact with Papuan languages also use this order, in contrast to speakers of Austronesian languages beyond this area. Moreover, the order of the numeral and the noun (e.g., six bananas) is similar, with speakers of Austronesian languages within the area using the numeral afer the noun (as in “bananas six”), similar to speakers of Papuan languages and unlike speakers of

140

From years to centuries

Austronesian languages beyond this area. Other characteristics include the use of the “conjoined comparative,” where the standard of comparison and the comparee are expressed in two separate clauses, as in “Tis child is tall, that child is short” or “Tis child is tall, that child is not.” Most speakers of Papuan languages use such constructions, as do speakers of Austronesian languages from the area. When considering these and other linguistic features and knowing that the people in the area were multilingual, it makes perfect sense to assume that this structural linguistic homogeneity must have arisen following centuries of language contact. Moving now to South America, the focus is on Amazonia. Amazonia refers to the Amazon River watershed and surrounding areas, delimited by the Andes mountains, the Caribbean and the Atlantic oceans, and the Gran Chaco. Tis vast area is home to some 50 families and isolates, among them the Arawak, Carib, Tupí, and Macro-Jê stocks. One should add to this mix contact with European languages like Spanish and Portuguese following colonization and languages like Quechua and Nheengatú that played an important role in the post-Columbian setting. Before the arrival of the European colonizers, studies suggest that large groups of people in Amazonia entertained long-distance trade contacts. Similar to New Guinea, people ofen spoke several languages and didn’t shif toward a dominant language, as they did afer the arrival of the Europeans. Epps and Michael (2017) note that there is a trend favoring low levels of borrowings and code-switching among the various languages of the area, yet grammars at a regional level exhibit some striking parallels. In contrast, borrowing and codeswitching are characteristic of contact with the European Spanish and Portuguese and colonially mediated languages like Nheengatu and Quechua. One smaller area where researchers observe linguistic convergence is the Vaupés region, located at the northwestern part of the Amazon, and another one is the Caquetá-Putumayo area in the south of Colombia and the north of Peru. Epps and Michael report that in these areas, languages cluster geographically, in some cases cross-cutting genetic groupings. Tis is illustrated by comparing similarities and diferences among over 200 grammatical features. Results show that the Vaupés area languages are more similar to one another, including languages of the East Tukanoan family (like Koreguaje, Barasano, Makuna, Tanimuka, Cubeo, Desano, Wanano (Kotiria), and Tukano), the language of the Arawakan family called Tariana, the language of the family Naduhup called Hup, and the language of the Kakua-Nukak family called Kakua. In comparison, the Caquetá-Putumayo area languages are more similar to one another, including the two Arawakan family languages, Resígaro and Yukuna, that are similar to the Boran family language Bora; the Witotoan family language Minica; and the isolate language Andoke. Moreover, North Arawakan languages beyond these two regions cluster separately (Achagua, Bare, Yavitero, Warekena, and Piapoco). Several other such contact areas can be identifed, while others are yet to be described. In addition to these smaller contact areas, linguists have also argued for the existence of an east-west division, but more work is needed to determine the precise features that such an area would include.

Linguistic areas

141

The role of Shamanic incantations in contact-induced change in Amazonia Patience Epps is a specialist of language contact in Amazonia. She has recently drawn attention to the role of specialized discourse like shamanic incantations in the difusion of contact-induced language change. Epps notes that specialized discourse licenses creative linguistic processes that can reach networks beyond a specifc community and is therefore a vehicle for contact-induced changes. Tese changes are well accepted in the language communities because specialized discourse is prestigious, as are those who practice it (respected for their age, knowledge, and unique skill sets). Here we provide an excerpt from a Patawa-Spine incantation by Ponciano Salustiano, recorded in 2011 and transcribed by Epps and Ramos (2019, p. 210) with some additional explanations given to us by Epps: Yúp ãh bab’ ni dö’öp, hõp ków, hõp ków tëg b’ah, pu ág tëg b’ah, bíy’ tëg b’ah, dëh hiyö’ ni n’ɨh, bíy’. Dëh póh ùt tëg b’ah, ãh do’ nenep, ãh do’ neneh, të! Hɨ’ap ̃ dëh, dëh mi hiyö’, ãh do’ nenep, të! dëh k’et yòh ham k’ët yɨ’ɨh. Yúwan b’ɨyɨ’ do’ yö’, tɨh wɨ’ d’ö’ nɨh ten, nusö’, s’ùg sö’ ãh sopop b’ay. S’ùg sö’, sẽhẽ́ k, s’ùg húp ĩh, sẽhẽ́ k hup ĩh nɨh, tɨnɨh d’apbùy yup tɨh bɨ’ nip, sẽhẽ́ k b’ah tɨh bɨ’ nip, wahnáw b’ah, dëh sɨsɨw, sɨsɨw dö b’ah, sɨp säsäw, tɨh bɨ’ nip. Tese I list together, fsh chili, fsh chili tree splinter, genipap tree splinter, peach-palm (variety) splinter, those of the lower river, peach-palm. Te waterbubble-spine tree splinter, I list those (to come out), list those out, until! All those from the lower river, I call/list them out, until! I get to the headwaters (i.e., lists the types of trees all the way from the lower river up to the headwaters). Having listed all this, if it doesn’t listen (i.e., the illness continues), I go up into the forest. Te forest lord of the paricá (Anadenanthera peregrina), the paricálord’s, the weapon he made (i.e., the patawa spine; Oenocarpus bataua), the paricá splinter he made, the abiu splinter, the water-brazilwood splinter, the red brazil-wood, all brazil-woods, he made (i.e., trees with very hard wood) . . . Epps notes that this genre requires fne-grained fora-fauna descriptions and may be a key catalyst behind Hup’s extensive calquing of binomial fora-fauna terms from Tukanoan languages. How do these calques arise? In Amazonia, shamanic training involves apprenticeships with other ritual specialists, including those from diferent ethnic and linguistic groups. Tese incantations are then performed by speakers of many languages in the Upper Rio Negro region, allowing for any innovations to spread among the general population.

9.4

HOW TO STUDY LINGUISTIC AREAS

Interested in studying linguistic areas? Tere’s a lot to be done! In this section, you can fnd some methodological guidelines that can help organize your research.

142

From years to centuries

Step 1. Describe the linguistic feature in the various languages First, make a list of the languages spoken in a geographical area that all use a specifc linguistic feature. Try not to just look at ofcial, national languages, as they have been through a standardization process, but also examine as many non-standardized varieties as possible. Indeed, non-standardized varieties are the frst to show contact-induced changes, since there is no standardization to stop these changes from happening. Step 2. Search for traces of the linguistic feature in time Te second step is to look for traces of this feature in the past. Tis is ofen difcult to do without written sources. In practice, most varieties lack written documentation, and when this is the case, the flter of the literate authors does not allow us to access the spoken varieties of the time. So, if you are lucky enough to have access to written documents and studies about a language in diachrony, then don’t forget to take them into account. Stop and think: What can you do to reconstruct the history of a language if there are no written documents from the past? Read the following answer. If there are no written texts available, linguists reconstruct the development of a linguistic feature over time by observing its distribution in space among contemporary varieties. For example, the degree of grammaticalization is generally greater in varieties that have frst started the grammaticalization process. In general, comparing languages from the same language group allows proposing plausible scenarios about the presence and development of these features in the past. Which takes us to Step 3. Step 3. Search for the linguistic feature in other languages of the same family Another well-known step in language contact studies is to compare the changes in the language under study with the developments in other closely related languages. For example, a correlation between grammaticalization of a linguistic feature and language contact can be established by comparing all the Slavic languages with the Slavic languages of the Balkan area or by comparing diferent Balkan Slavic varieties that do not have the same grammaticalizations or the same kinds of contacts. Step 4. Assess the typological rarity of the linguistic feature A good idea is to assess the typological rarity of a candidate linguistic feature. Why? If a linguistic feature developed in a given linguistic area is a rare typological feature, it is all the more likely that it is due to language contact. But if a linguistic feature is typologically common, then it would be more likely to arise independently. Indeed, studies show that efciency in usage favors some features over others across languages and time (Gibson et al., 2019).

Linguistic areas

143

Step 5. Identify the source language(s) of the linguistic feature Try to identify the source language(s) (or model language). A source language is a language in which the feature under study predates language contact and may have served as a model.Again, in the absence of historical documents, you can consider its degree of grammaticalization (when applicable, this means that you observe how a feature became part of the grammar).You can also conclude that a language probably had this feature if you trace it in related varieties located outside this contact zone. Step 6. Trace actual contacts between speakers, region by region If you think that a linguistic feature arose in a given area because of language contact, think about when and where people speaking these languages were actually in contact. You can use historical, anthropological, or even genetic data to understand the history of an area. Of course, language contact, as materialized by contact between individuals and societies, varies depending on time and space. Tis fuidity raises many hurdles, since in order to decide on the nature of the linguistic phenomena, it is necessary to look at the linguistic and sociolinguistic confguration in each linguistic community. Indeed, all the varieties of a language spoken in a given area do not necessarily fall under the same explanatory framework. For example, a linguistic phenomenon common to several varieties of Language A can arise in one area because of contact with Language B and in the other by propagation via Language A varieties or even by internal development independent from language contact. Tus, for each linguistic variety examined, try to assess the type of language contact at the time of the emergence of the candidate linguistic feature. Step 7. Identify the linguistic process If the hypothesis of language contact seems plausible based on the previous criteria, try to identify the precise linguistic process: Is it a borrowing? Or do we observe convergence of the structures?

9.5

CASE STUDY OF “HAVE”-PERFECT IN THE BALKANS

Let us illustrate the various methodological steps with a concrete example: the “have”perfect in the languages of the Balkans (see Adamou, 2012 and references therein). Step 1. Describe the linguistic feature in the various languages Many languages spoken in the Balkans use a perfect with the auxiliary verb “to have.” Tis is the case of Modern Greek (Greek) (e.g.,“I have written” exo ˈγrapsi (have.1SG + non-fnite verb) or ˈexo γraˈmeno (have.1SG + verbal adjective)); Literary Macedonian; the varieties of the south (Slavic) (e.g., “he has died” jima umrjano (have.3SG + past participle)); Romanian,Aromanian,and Megleno-Romanian varieties (e.g.,“I have carried” am purtata (have.1SG + participle)); and Judeo-Spanish (Romance) and Albanian

144

From years to centuries

(Albanian) (e.g.,“I have worked” kam punuar (have.1SG + participle), as well as some varieties of Romani (Indic) (e.g.,“he has lef” ov therel našto (have.3SG + participle)). Step 2. Search for traces of the linguistic feature in time It is easy to look at some languages with a written tradition for traces of the development of a “have”-perfect. For example, Greek has such a long written tradition and is thoroughly studied in historical linguistics.We therefore know that Modern Greek uses a “have”-perfect, which is based on a former infnitive: [“have” + non-fnite verb], for example, ˈexo ˈγrapsi“I have written.” In addition, in some Greek varieties, another “have”-perfect has been in use since the thirteenth century: [“have” + verbal adjective], for example, ˈexo γraˈmeno “I have written.” Step 3. Search for the linguistic feature in other languages of the same family Let’s look at the “have”-perfect in Slavic, a language group that extends beyond the Balkans. We quickly fnd out that the “have”-perfect is not a typical feature of Slavic languages, which regularly use either a “be” auxiliary with a verb in -l (marking gender and number) or simply the verb in -l without the auxiliary. Terefore, the use of the “have”-perfect in Balkan Slavic languages rightly attracts attention as a rather exceptional feature that deserves our attention. However, in some other Slavic languages, a “have”-perfect is in the process of grammaticalization when these are in contact with languages with a “have”-perfect. For example, Germanic and Romance languages have driven the development of a “have”-perfect in Kashubian and a less grammaticalized “have”-resultative in Czech, Polish, and Slovene. Step 4. Assess the typological rarity of the linguistic feature Beyond the Balkan area, the grammaticalization of the “have”-perfect is a common feature in the languages of the world and across Western Europe. We know that possessive constructions (e.g., “I have a car”) are often grammaticalized for aspect (e.g.,“I have worked”) but also for conditional, deontic modality, time (in particular, the future), and existence and that they can sometimes function as copulas. More specifically, the “have”-perfect is composed by the auxiliary verb “to have” and the main verb. In the process of grammaticalization, the verb “to have” no longer functions as a possessive verb but becomes an auxiliary to the extent that it can combine with intransitive verbs like “to die” or even with the verb “to have.” Step 5. Identify the source language(s) of the linguistic feature In Balkan Slavic, one possible source for the grammaticalization of the “have”-perfect is through contact with Romance languages like Aromanian as well as with Albanian. Researchers further suggest Greek as a possible infuence.

Linguistic areas

145

Step 6. Trace actual contacts between speakers, region by region Te center of innovation of “have”-perfect for Balkan Slavic languages could be the region of the Prespa Lakes, on the contemporary borders between Albania, Greece, and North Macedonia. Tis is corroborated when looking at real-life contacts, since we still fnd Slavic-, Aromanian-, and Albanian-speaking people today who intermarry and live in linguistically mixed villages. Step 7. Identify the linguistic process Is the “have”-perfect in Balkan Slavic a case of borrowing or of convergence? Table 9.1. shows the various stages in the development of a “have-”perfect in Macedonian varieties in contact with Aromanian. At frst, the typical Slavic construction [“be” + verb-l] competes with the new construction with “have” that arises through convergence with Aromanian [“have” + participle] (Stage II). In Stage III, bilinguals use in Aromanian the Slavic “be” construction that they combine innovatively with the Aromanian participle. Finally, in Stage IV, bilinguals follow the Aromanian model [“be” + participle] in Macedonian and drop the typical Slavic l- forms. Table 9.1 Macedonian and Aromanian two-way convergence process Stage

Macedonian

Aromanian

I

sum večeral [“be” + verb-l ] (Slavic starting point) sum večeral [“be” + verb-l ] imam večerano [“have” + participle] (change: variation between the Slavic construction and a new “have” construction through convergence with Aromanian) imam večerano [“have” + participle] sum večeral [“be” + verb-l ]

amu cinata [“have” + participle] (Romance starting point) amu cinata [“have” + participle]

II

III

IV

imam večerano [“have” + participle] sum veceran [“be” + participle] (change: loss of the Slavic verb-l forms through convergence with Aromanian)

amu cinata [“have” + participle] esku cinatu [“be” + participle] (change: variation between the Romance construction and a new “be” construction through convergence with Macedonian) esku cinatu [“be” + participle] amu cinata [“have” + participle]

146

From years to centuries

Of course, many factors can be at the origin of a linguistic change. Tis is why we ofen discuss multifactorial scenarios. For example, language-internal motivations also play a role in the development of a “have”-perfect in Macedonian, namely because evidentiality (indicating the source of the information being shared) started being coded by the “be” + V-l forms. Te motivation was then to mark both evidentiality and perfect with distinct forms.

Chapter summary In sum, linguists have long noted that languages spoken in specifc areas of the world tend to become more similar. While these areas may come out clearly when we compare distant geographic areas with one another, the difculty that linguists are faced with is to delimit the borders of such areas. Tis is reminiscent of the Russian dolls where the biggest doll hides a smaller one that hides an even smaller doll (see Figure 9.1). Te same applies to linguistic areas, where a big linguistic area can hide a smaller one and an even smaller area. Tis doesn’t mean that the language contact outcomes are not there, and linguists are increasingly drawing attention to them. Figure 9.1 Linguistic areas are like Russian dolls, where the biggest area hides a smaller one that hides an even smaller area

Source: Photo by Julia Kadel on Unsplash

Linguistic areas

147

EXERCISES 1.

Explore

Visit the Atlas of the Balkan Linguistic Area Online: https://abla.cnrs.fr/ (Adamou & Sobolev, 2023). Explore the various linguistic features and the language datasets that share the same values. You can start by exploring the “Perfect: auxiliary verb in the active voice” and identifying the “have”-perfect that we have presented in this chapter.

2.

Checkpoint

___________ (1928) was the frst to express the idea that a number of languages, while not closely related, may be viewed as a group, called “____________,” for sharing some common linguistic features due to contact. Terms such as “____________” (Weinreich, 1958) and “_________” in English are generally used as equivalents despite the fact that they express slightly diferent theoretic approaches. Some of the areas that have been examined from this perspective include _____________, _____________________, and ____________________.

FURTHER READING Hickey, R. (Ed.). (2017). Cambridge handbook of areal linguistics (1st ed.). Cambridge University Press. Tis handbook brings together scholars covering a wide range of linguistic areas and discussing the theoretical and methodological issues related to the detection of linguistic areas. It is the perfect way to get a grasp of the dynamics of this sub-feld of contact linguistics.

REFERENCES Adamou, E. (2012). Le parfait avec “avoir” dans l’aire balkanique: approche multifactorielle et diversifée d’un balkanisme. In C. Chamoreau & L. Goury (Eds.), Conséquences linguistiques du contact de langues et changements linguistiques. Tendances dans le domaine de la prédication (pp. 101–119). CNRS éditions. Adamou, E., & Sobolev,A. (Eds.) (2023). Te atlas of the Balkan linguistic area online. https://abla.cnrs. fr/ Assenova, P. (2002) [1989]. Balkansko ezikoznanie [Balkan linguistics]. Faber. Campbell, L. (2006).Areal linguistics: A closer scrutiny. In Y. Matras,A. McMahon, & N.Vincent (Eds.), Linguistic areas (pp. 1–31). Palgrave. Epps, P., & Michael, L. (2017). Te areal linguistics of Amazonia. In R. Hickey (Ed.), Cambridge handbook of areal linguistics (pp. 934–963). Cambridge University Press. Epps, P., & Ramos, D. P. (2019). Hup bi’id ɨd: Shamanic incantation at the nexus of language and culture. Journal of Linguistic Anthropology, 29(2), 205–212. Friedman, V. A. (2021). Te Balkans. In E. Adamou & Y. Matras (Eds.), Te Routledge handbook of language contact (pp. 385–403). Routledge.

148

From years to centuries

Friedman, V. A., & Joseph, B. D. (2023). Te Balkan languages. Cambridge University Press. Gibson, E., Futrell, R., Piandadosi, S. T., Dautriche, I., Mahowald, K., Bergen, L., & Levy, R. (2019). How efciency shapes human language. Trends in Cognitive Sciences, 23(5), 389–407. https://doi. org/10.1016/j.tics.2019.02.003 Joseph, B. D. (1983). Te synchrony and diachrony of the Balkan infnitive. Cambridge University Press. Joseph, B. D. (1992). Te Balkan languages. In International encyclopedia of linguistics (pp. 153–155). Oxford University Press. Joseph, B. D. (2010). Language contact in the Balkans. In R. Hickey (Ed.), Te handbook of language contact (pp. 618–633). Wiley-Blackwell. Matras, Y. (2009). Language contact. Cambridge University Press. Miklosich, F. (1861). Die slavischen Elemente im Rumunischen. Denkschrifen der Kaiserlichen Akademie der Wissenschafen, Philosophisch-historische Klasse 12, pp. 1–70. Mladenova, O. (2007). Defniteness in Bulgarian. Mouton de Gruyter. Sandfeld, K. (1930 [1926]). Linguistique balkanique [Balkanflologien]. Kliensieck. Schapper, A. (2021). Linguistic Melanesia. In E. Adamou & Y. Matras (Eds.), Te Routledge handbook of language contact (pp. 480–502). Routledge. Sobolev, A. (2004). On the areal distribution of syntactical properties in the languages of the Balkans. In O. Miseska-Tomić (Ed.), Balkan syntax and semantics (pp. 59–100). John Benjamins. Trubetzkoy, N. (1928). Proposition 16. In Acts of the frst international congress of linguistics (pp. 17–18). Leiden. Weinreich, U. (1958). On the compatibility of genetic relationship and convergent development. Word, 14, 374–379. Wiemer, B. (2004). Population linguistics on a micro-scale. Lessons to be learnt from Baltic and Slavic dialects in contact. In B. Kortmann (Ed.), Dialectology meets typology (pp. 497–526). Mouton de Gruyter.

10 Creoles, pidgins, and mixed languages

CHAPTER PREVIEW Tis chapter is dedicated to outcomes of intense language contact that give rise to new languages such as Creoles, pidgins, and mixed languages. We will also take a look at African urban youth languages that are being used today in highly multilingual settings. What are these languages and how were they formed? Are they outcomes of language contact, unlike those we have investigated in this book, or are they formed by the same linguistic processes that we have studied so far? Trigger warning: Discussion of slavery.

10.1

CREOLES

10.1.1 When did Creoles arise? Creoles are relatively new languages that emerged through intense contact between multiple unrelated languages in adverse contexts involving slavery and racial segregation in colonial plantations (Aboh, 2015). In linguistic terms, Creoles result from the combination of Western European colonizer languages such as English, Spanish, French, Dutch, and Portuguese, with a variety of African languages. Te European languages provided large parts of the lexicon and are ofen referred to as “lexifer languages,” but these languages also provided some of the grammar. Te African languages of the enslaved people provided grammatical, semantic, and phonological features and are referred to as “substrate” languages. For example, Cape Verdean Creole and Guinea Bissau Creole result from the combination of the lexifer European Portuguese with various substrate NigerKordofanian languages like Wolof, Fula, Serer, Balanta, Manjaku, Mankan, Dyola, and Bola (West Atlantic languages) and Mandinka, Malinke, and Bambara (Mande languages) (Baptista, 2020). In comparison, Palenquero, a Creole spoken in Colombia, results from the combination of the lexifer Spanish, and the substrate Kikongo, a Bantu language. As with other outcomes of bilingualism and language contact, Creoles display a number of innovative linguistic features that are not found in either the lexifer or the substrate languages. DOI: 10.4324/9781003167952-13

150

From years to centuries

Although European languages are the most well-known lexifers of Creoles because of the scope and duration of the Atlantic slave trade, non-European languages, like Arabic, have also served as lexifers for Creoles with multiple local African languages as substrates. Tis is the case of Juba Arabic spoken in Sudan (Manfredi & Petrollino, 2013) and Kinubi spoken in Uganda and in Kenya (Lufn, 2013). And though African languages are the most well-known substrate languages, nonAfrican languages have also served as substrates in Creoles. For example, Kriol, spoken in Australia by several thousand Aboriginal people, draws from many Aboriginal languages such as Alawa, Marra, Ngalakgan, Wandarrang, Mangarrayi, Ngandi, and Nunggubuyu (Roper River); Jawoyn, Dalabon, and Rembarrnga (Barunga/Beswick); Jaminjung, Ngarinyman, and Wardaman (Westside); and Walmatjari, Jaru, Miriwoong, and Gija (Kimberley) (Schultze-Berndt & Angelo, 2013).

10.1.2 How did Creoles get formed? As discussed in Chapter 7, language ideologies around Creoles are still extremely stigmatizing. Tese ideologies, which cast Creole languages as “incomplete” or “simple” versions of their lexifers, have their roots in dehumanizing racial theories of colonialism and slavery, and they have also found their way into linguistic studies of Creoles (DeGraf, 2020). For instance, scholars of the nineteenth century argued that the enslaved Africans were incapable of learning the reputedly “advanced” European languages and that the Creole languages refected their simplifcation processes. Linguists of the twentieth century rejected the explicitly racial theories of the colonial era. But they continued to argue that Creoles were, in some sense, simpler than their European lexifers and sought explanations for why this was so. One longlasting idea held that Creoles were the “nativized” forms of pidgins, which are varieties that are spoken as second languages for special functions like bartering (see Section 10.2 for more on pidgins). According to this view, the enslaved Africans were brought together but didn’t share a common language, so ended up converging on the dominant colonial language in order to have a shared means of communication, in this case a pidgin. Creolized forms of the pidgin would have developed as children acquired the pidgin as their frst language. For years, linguists taught this “universalist hypothesis,” which held that Creoles ofer an illustration of the workings of a universal grammar in the minds of children who took a pidgin as input and imposed a grammatical structure upon it.Yet others believe that Creole languages show a process of the relexifcation of an African substrate language, which means that words and stems from the European language were inserted into the syntactic frame of an African substrate. However, all these origin scenarios view Creoles as unique among the world’s languages. Te idea that Creoles do not ft into the same category as the other languages of the world is known as the “creole exceptionalism hypothesis,” a position soundly criticized by DeGraf (2003). Today, many scholars, some of whom are L1 speakers of Creoles, propose an anti-exceptionalist explanation for the origin of these languages. According to linguists like Aboh (2015), Mufwene (2021), and DeGraf (2003, 2020), complex multilingualism is at the origins of the formation of Creoles. Tis means

Creoles, pidgins, and mixed languages

151

that the enslaved Africans and their European overseers used words, expressions, and structures from multiple languages spoken in plantations for their day-to-day communication. In other words, Creoles were formed through the same processes that operate in language contact in general (Mufwene, 2021). Te only diference is that they were formed quickly and in the restricted environment of a plantation setting. In support of the multilingual formation of Creoles, Aboh (2015) reports historical documents showing that enslaved people did not, in fact, lose their African languages. In many settings, they found themselves among speakers of their language or languages that were typologically similar, facilitating communication in African languages in the plantations. And we must not forget that the slave trade continued for centuries, so that contact with L1 speakers of African languages did not disappear in many colonies. In the case of Haiti and Suriname, speakers of Gbe and Kikongo languages could communicate with one another in their frst languages. In addition to speaking their African varieties, they would also learn the dominant colonial European languages of the homestead or the plantation. In other words, the enslaved Africans had multiple linguistic resources that they used among themselves and with their European overseers. A result of this multilingual contact is that there was competition among the available linguistic resources, and Aboh (2015) argues that the historical facts show that the speakers of the substrates and of the lexifer languages in such situations negotiated the language forms to be used.Accordingly, we can see Creoles as languages that developed in a multilingual colonial society by adult and early L2 speakers as well as by 2L1 and L1 speakers of the lexifer language. Aboh underlines the difering functions of the Creole in colonial society: From the perspective of the colonists, the creole therefore represents the language of efcient and orderly management of the enslaved population that guarantees a fourishing business: sugar and tobacco production or gold extraction. For (part of) the Africans, however, it represents the neutral language (just like ofcial languages in modern Africa) that guarantees successful daily interaction with everyone on the colony. (2015, p. 125) As long as the Creole and its function was still being negotiated, multilingualism would have persisted in the plantation. But as the Creole gradually took over more functions on the plantation, the African vernaculars lost ground. Although some scholars estimate that Creoles were formed as early as the seventeenth century, Mufwene suggests that they must have been formed in the eighteenth century when colonial plantations had become large enough that the majority of people within them were enslaved Africans who were segregated from the European population (Mufwene, 2021). On the linguistic level, Mufwene (2008) proposes a “competition-and-selection” framework regarding which linguistic features a Creole would retain. Baptista (2020) argues that features that are congruent (i.e., similar) between the languages in contact may have been favored. For example, in Cape Verdean Creole and Guinea

From years to centuries

152

Bissau Creole, the negation ka most likely results from the lexifer Portuguese nunca and substrate Mandinka buka, kana, kaka; Manjaku dika, kats(a), Balanta kè, and Manding kana, kuka, buka. Te conditional ál results from Portuguese há de and Wolof yall/yálla. Te passive -du can result from Portuguese -du and Wolof -(t)u. Te past-tense marker ba can result from Portuguese -va and Manjaku and Mankan -ba, Dyola ban, and Mandinka ka ban. Te complementizer kuma possibly results from Portuguese como and Mandinka kö. All these features share morphophonological, semantic, and syntactic properties and were therefore likely to be retained in the Creole as well. Diferences in the contribution of linguistic features by the lexifers and the substrate languages can also refect diferences in the socioeconomic settings (Yakpo, 2021). For instance, linguistic features from African languages are more present in the English-lexifer Creoles, moderately present in the French-lexifer Creoles, and the least present in the Spanish and Portuguese Creoles spoken in the Americas.

10.1.3 What are the structural characteristics of Creoles? Creoles are typologically similar to one another. Tis similarity was once considered evidence of their exceptional status, but, instead, Creoles are more likely to have similar structures because they are all formed from a combination of West European lexifers and West African substrate languages (Blasi et al., 2017). Let us take a closer look at some salient typological features of Creoles. Researchers note that Creoles are typically isolating languages, unlike some of their standard European lexifers that are fusional and some of the African substrate languages that are agglutinating or polysynthetic. Let’s examine an example of Haitian Creole (Kréyòl), spoken on the island of Hispaniola in the country of Haiti, compared with its lexifer, French. 10.1 Kreyòl Kreyòl

French

English Gloss

(a) Mwen manje I eat.pres

Je mange I eat.1p.sg.pres

I eat

(b) Mwen pral manje I fut eat

Je mangerai I eat.1p.sg.fut

I will eat

(c) Mwen te manje I past eat

J’ai mangé I aux.1p.sg.part

I ate

Te Kreyòl verb stem remains invariable; it is the same in its infected and infnitive forms: manje. When infected, for the future tense in example 10.1.b and the past in 10.1.c, the lexical stem is preceded by a stand-alone particle. Tis is what is meant by an isolating type of language. French shows a mix of morphological strategies. Tere is a synthetic infection in 10.1.b for the future, where the markers for person and for tense occur as sufxes attached to the verb and an analytical one for the past tense, where number is carried on the auxiliary and tense signaled by the participle.

Creoles, pidgins, and mixed languages

153

Isolating morphology in Creole languages was ofen considered an indication of the simplifcation processes that are more generally thought to arise in language contact settings (see Chapter 3 for the simplifcation hypothesis). Tis view is now challenged by many researchers who draw attention to the fact that we do not know with certainty the characteristics of the contributing varieties at the moment of contact.All too ofen, linguists compared the Creole to the standardized, written language of its European lexifer, ignoring the fact that the languages the European colonizers actually spoke were probably much more isolating than the standard, written language was. For example, in spoken French, 10.1.b is more likely to be expressed analytically, rather than synthetically, via an auxiliary: je vais manger “I go.1sg.pres eat.” Another argument against the simplifcation hypothesis is that some of the substrate African languages are isolating, like the Kwa languages that played a role in the formation of many Atlantic Creoles. Another feature that is often considered characteristic of Creoles is the absence of tones. This is also theorized as being part of a simplification process in language contact settings. Bordal Steien and Yakpo (2020), however, challenge this generalization and stress that while some Creoles exhibit stress, other Creoles exhibit tones (e.g., Creoles of the West African littoral zone and isolated Creoles in the Americas spoken by enslaved Africans who liberated themselves, such as the Maroon Creoles of Suriname) or a mix of both stress and tones (e.g., Iberianlexifier Creoles of the Caribbean like Papiamento, spoken in the Netherlands Antilles). These characteristics depend on the presence and conditions of life in the colonies for speakers of African (tone) languages and European languages (with stress). According to Yakpo (2021), European-lexifier Creoles of the Caribbean probably had tone systems that were lost at the end of the slave trade in the nineteenth century with the generalized dominance of European colonial languages. In contrast, in the Dutch Caribbean “ABC” islands (Aruba, Bonaire, and Curaçao), there were many speakers of African tone languages who lived in slightly less segregated contexts and had more economic opportunities. This meant that bilingualism in the Creole language of the ABC islands, Papiamento, and in Spanish was not uncommon. As a result, stress from Spanish and tone from the African languages were selected as features for the early Papiamento language. Such fine-grained analyses illustrate why researchers should pay more attention to the development of Creoles over time and the language ecologies that were in place at the time of their formation. Of course, Creoles are still dynamic languages and subject to general language contact effects. For example, speakers of Spanish and Palenquero, a Spanishlexified Creole spoken in Colombia, sometimes code-switch between Spanish and Palenquero; for example, they may use te digo “I’m telling you” when they speak Palenquero or use grammatical categories from Spanish that are otherwise not found in Palenquero, like Spanish verbs with person and number marking as opposed to invariant verbs used in Palenquero. Such uses either fall under the radar of Palenquero speakers and are evaluated as different from Palenquero, which is considered by its speakers a distinct language from Spanish (Lipski, 2019).

154

From years to centuries

Meet linguist Michel DeGraff Te term “Creole” started not as a linguistic term at all; it started as an ethnographic term. It was used to distinguish the Europeans that were born in the so-called New World—these were called Creoles, meaning Europeans who were born in the Caribbean and in the Americas, to distinguish them from the Europeans who were born in Europe. And the same term was also applied to the Africans, so the Africans that were brought from Africa were not Creole, but the Africans born in the Caribbean were called the Creole Africans. So, the term “Creole” was used to refer to those who were born and grew up in the Caribbean with ancestry from elsewhere, either from Europe or Africa. Te term was also applied to species and to food—there was Creole rice, Creole horses, Creole cows. Te speech of these populations was diferent from the speech of Europeans from Europe so it came to be known as Creole speech because it was spoken by Creole people, the people who were born and grew up in the Caribbean. Tat’s how the term got to be applied to varieties of speech that emerged from speech spoken in Europe but which had their own distinctive features. Tose languages came to be known as Creole: In Jamaica, they use Jamaican Creole, in Haiti, it’s Haitian Creole, in Barbados, Bajan Creole, in Trinidad, Trinidad Creole. Essentially, it’s more like an ethnographic term to refer to speech varieties that were specifc to the Caribbean but with ancestry from elsewhere, in the same way that the people were called Creole who were born in the Caribbean but with ancestry from elsewhere. So, for me, the term Creole doesn’t have any linguistic baggage at all. Specifc to Haitian Creole, we can step back to the 17th century when the French settlers frst arrived in Saint Domingue, which was the colonial name of Haiti. First there were relatively few enslaved Africans to work on the plantations but there were also indentured servants from France that they brought in to help them settle Saint Domingue.You have to remember that the speakers of French that came from Europe didn’t speak one standard French variety; they spoke diferently, what they called then “patois.” So those speakers of diferent French varieties came into contact with Africans who were also working as enslaved labor on the plantations. And there was a process of language acquisition on the part of the Africans and little by little there emerges this variety that had lots of features from the diferent 17th century French dialects plus, of course, the Africans’ speech.When they were learning French they brought in infuence from their own native African languages, mostly from the Niger-Congo family from the Ewe, Fon, Igbo, Congo, etcetera—a whole variety of Niger-Congo languages infuencing the learning of French. So basically, you had second-language varieties of French infuenced by those various dialects of French, plus the NigerCongo languages. And as new Africans were arriving into the colonial setting and appropriating these new varieties of French, they kept changing it until at some point there was a merger and the diferent idiolects from second-language varieties of French and Haitian Creole crystalized into what we know it as today.

Creoles, pidgins, and mixed languages

155

I like to compare the history of Haitian Creole coming from varieties of French and the diferent African languages as parallel to the history of French emerging from speakers of Latin coming into contact with speakers of Celtic and Germanic languages in northern Europe. In both cases you have population contacts and you have conquest. You have speakers of a prestige language—in Europe it was Latin—in contact with speakers of “barbaric” languages, meaning the languages of the oppressed; in this case it was the Celtic and Germanic. And out of this contact emerged Old French, which now has become what we know as French. In the context of Haiti, we have a parallel pattern but involving French as a prestige language in contact with languages spoken by the oppressed, in this case the Africans. So, in a way it’s a very similar process in terms of population contact and also in terms of power relations between different linguistic groups. I also like to say that a Creole language is just a language.When it comes to linguistic structure, there aren’t any special features that would keep Creole languages in a separate box from other languages. In fact, we linguists have brought in so many myths around this thing called “Creoles.” If we were to apply notions of mixedness based on this classic notion from the work of Derek Bickerton saying that in the history of Creole you can see what he called “macaronicity,” with lots of bits and pieces coming from diferent languages, we see that the vocabulary of English is way more macaronic than the vocabulary of Haitian Creole. Te vocabulary of Haitian Creole is 90% of French origin. From that perspective English has out-Creoled most Creoles, which shows how absurd linguists’ defnition of Creole languages is. I didn’t sit down one day and decide to study Creoles. My degree was in computer science and math. One of my frst jobs was at the Bell Labs in New Jersey and I had the luck to work for computer scientists who were very interested in natural language processing. My job was to write routines for text to speech programs, making sure that the sofware could pronounce French words like Catherine Deneuve correctly. It was my frst time thinking about language in a very computational symbolic way, and I would make comparisons between French and my native Haitian Creole. I realized that Creole as well has patterns that can be computed. But when I was back home in Haiti, I was always told that Creole is not a language, that Creole is broken French. Ten I started thinking that if I can write a computer program that can process Haitian Creole speech and syntax then it is a language, it’s not broken French. In fact, I could see that the rules of Creole were not at all broken, that it was systematic, like French. Ten it dawned on me that something was wrong with this picture. It’s an interesting interaction between power and knowledge where indeed, as Michel Foucault pointed out, in many cases knowledge is not neutral. Knowledge is produced in order to sustain particular regimes of power and vice versa. Regimes of power tend to favor particular types of knowledge that could, in turn, keep power in place. I had been told by

156

From years to centuries

very fancy teachers in the fancy French Catholic school that I shouldn’t speak Creole but here I was writing programs that proved to me that Creole is a language. I spent the frst 20 years of my life assuming that my native language was not a language. I became very interested in trying to understand what it means to be a language. And that’s when I decided I had to know more about this feld called linguistics. Shared by Michel DeGraf

10.2

PIDGINS

Pidgins are languages that emerged for communication in trade between people of various frst languages. Sebba (1997) notes that pidgins result from the need for communication among adults in multilingual settings and, generally, they serve as second languages for speakers. According to Sebba, pidgin formation can take one of many forms. Some, like Juba Arabic in Sudan, come into being as a result of military or police initiatives. Others arise from seafaring and trade and still others from labor, including mining, construction, and the forced labor on plantations we have discussed. Structurally, pidgins are more conventionalized than “jargons,” which are also used in trade. Tis means that pidgins have an agreed-upon lexicon and grammar among their users, even if there is variation in the way that individuals might use the language. Speakers of pidgins typically combine the languages spoken by foreign traders as lexifers with the languages spoken by the local populations as substrate languages (Mufwene, 2021). Unlike Creoles, pidgins are used as lingua francas in parallel with the other substrate languages and do not replace them. Indeed, pidgins are not the frst language of a language community. Pidgins show great variation, especially when spoken in large areas by numerous speakers. Unlike Creoles, they ofen have nonEuropean lexifers; for example, Chinook Jargon, once spoken in the Pacifc Northwest of North America, consists of contributions from Chinook and Nootkan, as well as from French and English. Hiri Motu in Papua New Guinea has an Austronesian language, Motu, as a lexifer. Fanagalo, spoken in South Africa, has Zulu as a lexifer with contributions from English and, to a lesser degree, Afrikaans. An example of a non-European lexifer pidgin is Bazaar Malay, which has been spoken for many centuries in a much-frequented trade center, Southeast Asia, and, in particular, in the area that extends from the Malaysian Peninsular through the Indonesian archipelago around Java, Sumatra, and Borneo (Ansaldo & Lim, 2021). Trade in this area was intensive between the Chinese, Arab, and Javanese well before European companies began to arrive in search of goods. Spices like cloves, pepper, and cinnamon were among the many traded products. Bazaar Malay is an important pidgin that has been used for centuries in trade and more generally in interethnic communication. Although Bazaar Malay is characterized by variation in time and space, it shows some consistent features. For instance, it has predominantly isolating morphology when compared to Literary Malay, personal pronouns are of Hokkien

Creoles, pidgins, and mixed languages

157

(Chinese) origin, possession is based on punya (“owner” > “to possess”), the Malay existential form ada is used as a progressive, and the causative constructions are based on Malay kasi “to give” and buat “to make” (Ansaldo & Lim, 2021). Pidgins may also have European lexifers, like Chinese Pidgin English or China Coast Pidgin, Chinese Pidgin Russian, Nigerian Pidgin English, and Ghanaian Pidgin English, to name but a few. Chinese Pidgin English emerged in Southern China, frst with the presence of Portuguese traders (sixteenth century), later with the arrival of the British (seventeenth century). Chinese Pidgin English has English as a main lexifer (e.g., pidgin < business, catchee < catch “fetch”) but also includes lexicon from Portuguese (e.g., sabbee < saber “to know”), Malay (e.g., chop < chapa from Hindi “to chop, to stamp”), and Cantonese (e.g., taipan “supercargo”) (Ansaldo & Lim, 2021). Chinese Pidgin English uses classifers, a typological feature of Sinitic languages. Classifers, or counters, accompany nouns and change according to the object being specifed.We use something similar in English with mass nouns or nouns that cannot be counted. For example, we would say a grain of sand, or a cup of tea, to specify the quantity of the material. In Chinese Pidgin English, the form of the classifers is based on English: You wantchee catchee one piecee lawyer < English “piece” (Ansaldo & Lim, 2021). Chinese Pidgin English has no copulas (like “is”), in line with Sinitic language typology, such as Englishman very good talkee (Ansaldo & Lim, 2021). Expressions like chop-chop, which we fnd in English with the meaning “quickly,” have their origins in Chinese Pidgin English. Another Chinese pidgin is Chinese Pidgin Russian, a pidgin with a Russian lexifer and Chinese, Mongolian, and Tungusic as substrate languages (Perekhvalskaya, 2013). Tis pidgin emerged at the end of the eighteenth century and was used until the twentieth century along the Russian-Chinese border for the purpose of trade and the colonization of Southern Siberia by Russia. Te frst evidence of a Russian-based pidgin is the Kyakhta language, a trade language that was standardized by the Chinese and taught to all the Chinese merchants who traded with Russia. Chinese Pidgin Russian became the lingua franca of Russian and Chinese communities in the large cities of the area. Chinese traders still travel to Russia for business and are confronted with the need to communicate in the absence of prior formal education in Russian. Frajzyngier et al. (2021) describe the Sino-Russian idiolects of some of these speakers who attempt to negotiate communication between these typologically distant languages. It does not appear that a new pidgin will arise in this context because the idiolectal Russian of the Chinese traders is severely restricted in use. Pidgins are not just a thing of the seafaring past or of the bustling trade bazaar. New pidgin languages continue to arise when economic or migratory pressures create new contexts for intense language contact, particularly between unrelated languages. For example, the widespread immigration of workers from South Asia has resulted in the emergence of a pidginized variety of Arabic, known colloquially as Urdubic or Gulf Pidgin, in the Persian Gulf. Urdubic, as its name implies, draws from a Hindi/ Urdu substrate and a Gulf Arabic lexifer. In conversation, it is also highly mixed with words and phrases from English, which has itself become a kind of lingua franca in multilingual cities like Dubai (Hussain et al., 2020).

158

From years to centuries

Some pidgins, like Tok Pisin in Papua New Guinea, become acquired as frst languages when children learn them from their parents. Tese are called “expanded pidgins,” and they can be as grammatically complex as Creoles. Tis is a reason to question the “nativist hypothesis” discussed previously by which Creoles were thought to be nativized pidgins. In the view of Mufwene (2015, p. 134), an expanded pidgin and a Creole can have the same structural outcome. Tey are considered diferent types of contact languages because they have very diferent histories, not because they necessarily difer in structure.

10.3

MIXED LANGUAGES

Mixed languages are yet another outcome of intense language contact. Tey can be broadly defned as “the result of the fusion of two identifable source languages, normally in situations of community bilingualism” (Meakins, 2013, p. 159).Although the existence of mixed languages was controversial for some time, mixed languages are now considered diferent from pidgins and Creoles both from a sociolinguistic and structural point of view.

10.3.1 When do mixed languages arise? From a sociolinguistic point of view, mixed languages are created in diferent contexts from Creoles and pidgins. Mixed languages are used in highly bilingual communities, ofen in a context of language shif. In comparison, Creoles are used when a speech community is not bilingual but speakers still need to communicate with others in a diferent language. Mixed languages also difer from cases of intense bilingualism, where speakers alternate from one language to the other dynamically in interaction (see Chapter 4 on code-switching) because the kind of mixing is more conventionalized. Unlike code-switching, mixed languages arise in a more limited set of social contexts, specifcally when speakers resist shifing to a dominant language and persist in using some elements of the community language. Tis tension can under some circumstances result in an entirely new language (Tomason & Kaufman, 1988; O’Shannessy, 2012; Meakins, 2013). Diverse social settings can result in mixed languages. In some cases, marriage practices between ethnically and linguistically diferent groups have brought about a blending of the languages. In others, mixed languages arise when ethnically and linguistically homogeneous groups migrate for work outside their language community. And mixed language varieties can also occur in the context of language shif and rampant code mixing. But then, why don’t we witness mixed languages more frequently, for example, in linguistically mixed households or in immigrant communities? We’ll briefy consider examples of each to understand the specifc circumstances under which mixed languages have emerged. Let’s begin with the emergence of a mixed language in linguistically heterogeneous households. Tis is illustrated by the case of Michif, a mixed French-Cree language, which is spoken in Canada and the United States by a few hundred Métis people (Bakker, 1994). Te Métis are the descendants of male European fur traders who spoke

Creoles, pidgins, and mixed languages

159

French and of the Native American women who became their wives and who spoke Cree to their children. At some point in its trajectory, Michif became the language of the wider community, replacing the two source languages. Tis is rare because, in most communities where exogamous marriage patterns are found, the respective source languages of each partner in a couple continue to be spoken, so there is no communicative reason to create a new language that would only be understood by members of a restricted community. Tus, the creation of the mixed language by the Métis should be understood as an act of identity. Another well-known example of a mixed language is that of Media Lengua from Ecuador. Media Lengua is a Quechua-Spanish mixed language that likely emerged when Quechua speakers moved to work in Spanish-speaking cities (Muysken, 1994). Tese speakers chose to replace many Quechua words with Spanish stems while keeping Quechua grammatical structure. Like the Métis, those who speak Media Lengua refect a new mixed cultural identity through their language. Lipski (2019) used a variety of experimental techniques with speakers of Quichua and Media Lengua to assess whether Media Lengua is a form of mixing that is independent from dynamic mixing (that is, code-switching) between Quichua and Media Lengua. Results confrm that Media Lengua is distinct from Quichua, as participants considered Quichua and Media Lengua mixed sentences less acceptable than sentences in either Quichua or Media Lengua. Another striking example of the formation of a mixed language is Gurindji Kriol. Gurindji Kriol emerged among the Aboriginal people of Kalkaringi and Daguragu in northern Australia. Before colonization by Anglos, people spoke Gurindji, a PamaNyungan language, as well as various local languages of neighboring groups. When Australia was colonized, cattle stations were established in the area by the colonizers. Tis led to the formation of the English-based Kriol that is nowadays spoken by more than 20,000 Indigenous people in Australia. While there was a process of shif from Gurindji to Kriol, this process was halted when language attitudes in the community shifed in favor of the maintenance of Gurindji (McConvell & Meakins, 2005). Tis is what gave way to the mixed Gurindji Kriol language that is nowadays spoken by the younger generations. Te older generation, who are still speakers of Gurindji, do not always evaluate the mixed language in positive terms, but the younger speakers consider their hybrid language, Gurindji Kriol, in increasingly positive terms as part of their identity.

10.3.2 What are the structural characteristics of mixed languages? From a structural point of view, mixed languages difer from Creoles and pidgins in that the elements of the two source languages are not substantially changed. Two major types of mixed languages are documented: 1)

Grammar–lexicon mixed languages

Grammar–lexicon mixed languages draw the grammar from one language and the lexicon from another.

160

From years to centuries

An example comes from Media Lengua spoken in Ecuador, combining Quechua grammar with Spanish lexicon (Muysken, 1994): (10.2) Media Lengua (Ecuador) < Quechua (in plain type) and Spanish (in bold) unu fabur-ta pidi-nga-bu bini-xu-ni one favor-acc ask-fn-benf come-prs-1 “I come to ask a favor.” (Muysken, 1994, p. 207)

In Media Lengua, Spanish words are integrated using Quechua phonology. 2)

Verb–noun mixed languages

Verb–noun mixed languages mix nouns from one language and verbs from another. An example of verb–noun mixing is Light Warlpiri, spoken in Australia by Warlpiri youth. In Light Warlpiri, most verbs come from English/Kriol, whereas verb structure comes from both Warlpiri and English/Kriol and nominal structure from Warlpiri (O’Shannessy, 2013). See example (10.3). (10.3) Light Warlpiri (Australia) < Warlpiri (in plain type) and English/Kriol (in bold) Junga mayi nyuntu yu-m go wati-kari-kirl mayi? true q 2sg 2sg.s-nfut go man-other-com q “Is it true that you went with another man?” (O’Shannessy, 2013, p. 330)

Michif is considered a grammar–lexicon and verb–noun mixed language. Te reason is that the mixed language uses predominantly Cree grammar with a French lexicon. However, at the same time, it shows a split between verbal and nominal morphology, with Cree providing most of the verbal morphology and French the nominal morphology. See an example in (10.4). (10.4) Michif (Canada) < Cree (in plain type) and French (in bold) maci-kîsikâ-w pas moyȇn si-misk-ahk bad-weather-3SG.INAN no way COMP-fnd-he.it son shack waisi-n his cabin be.lost-he “A storm came up, he got lost, he couldn’t fnd his way back to the cabin.” (Bakker, 1997, p. 6)

10.3.3 How do mixed languages get formed? How exactly mixed languages are created is still subject to discussion. Some researchers believe that mixed languages result from extraordinary language mixing processes. In one of the early studies on mixed languages, Bakker (1994) examined Michif and concluded that mixed languages result from “language intertwining,” that is, the use of lexemes from one language with the phonology, syntax, and morphology

Creoles, pidgins, and mixed languages

161

of another language. Tis process is roughly equivalent to “relexifcation,” proposed by Muysken (1994). Other researchers, however, consider that mixed languages emerge through ordinary processes of contact like borrowing and code-switching (Tomason & Kaufman, 1988). In a similar vein, Myers-Scotton (1998) proposes a code-switching account of mixed language formation based on the matrix language frame (MLF) model. She hypothesizes that mixed languages result from an arrested matrix language turnover, that is, when the switch from using the grammatical frame of one language to that of another is interrupted. Recall from Chapter 4 that Auer (1999, 2014) has proposed a continuum from code-switching to code-mixing to fused lects.When code-mixing becomes the norm in a community of practice, the result is a “fused lect” that eventually becomes its own language, a “mixed language” (see Figure 10.1). Auer defnes fused lects as follows: “the use of one “language” or the other for certain constituents is obligatory in FLs [fused lects]; it is part of their grammar, and speakers have no choice” (1999, p. 321). Auer (2014) recognizes two basic strategies for the fused lects. Te frst is what he terms “minimal insertion.” Tis occurs when speakers insert stems from one language into the grammatical frame of another, as is done in Media Lengua. In the other type,“maximal insertion,” speakers use elements within their own grammatical frames, ofen inserting entire phrases. For instance, a verb would be inserted with its infections. Maximal insertion can be seen in the example from Michif Cree, with its V-N split; the verb and their infections are from Cree, while the noun phrases and infections are from French. Strong evidence in support of the continuum hypothesis of the development of a mixed language from code-switching comes from Gurindji Kriol and from Light Warlpiri. McConvell and Meakins (2005) compare the productions of speakers of Gurindji from the decade of 1970 to recordings from the decade of 2000. In particular, they examine the speech of members of older and younger cohorts in each generation to fnd that code-switching between Gurindji and Kriol intensifes among the younger generations over time. Te conventionalization of the code-switching patterns from earlier generations of Gurindji and Kriol bilingual speakers resulted in an independent mixed language, Gurindji Kriol. O’Shannessy (2012) also fnds that code-switching in the input of children was the impetus for them to speak the mixed language Light Warlpiri. Te mixed language had its origin in the code-switched input between Kriol and Warlpiri in the talk directed to the children. Te children, of course, used this mixed input Figure 10.1 A continuum from code-switching to fused lects and mixed languages

Source: Adapted from Auer (1999)

162

From years to centuries

systematically. Just as with every case of language transmission to children, the structure of a mixed language continues to evolve (see Chapter 6). Tis means that the component parts of a mixed language system may not be identical to those of the source languages.). For example, the children who speak Light Warlpiri use a new auxiliary that draws on forms that are from Warlpiri, English, and Kriol (O’Shannessy, 2013). More evidence in support of the Auer’s continuum hypothesis comes from Adamou (2010) and Adamou and Granqvist (2015), who investigated two fused lects (that they also call “unevenly mixed languages”): Romani-Turkish as spoken in Greece and Romani-Finnish as spoken in Finland. In these varieties, all profcient Romani speakers insert morphologically intact verbs from the historically majority languages, Turkish and Finnish, respectively, into otherwise Romani-dominant speech. See example (10.5). (10.5) Romani-Turkish (Greece) < Romani (in plain type) and Turkish (in bold) ep always

me 1SG.NOM

ka will

me 1SG.NOM

da FOC

səndəm tired.PRET.1SG

dikh-av look-1SG

kale them

me da mang-av dineneəm 1SG.NOM FOC want-1SG rest.OPT.1SG “Am I always the one to look afer them? I’m tired of it! Me too, I want to rest.” (Adamou & Shen, 2019, p. 55)

You can see how the combination of Turkish verbs with Turkish verb morphology and Romani and Turkish nouns with Romani morphology looks like a V-N mixed language. In this example, you can see that the speaker uses the grammar of both. In Myers-Scotton’s terminology, the speaker is using two matrix languages: there is agreement between the subject (Romani frst person singular pronoun with Romani nominative case) and the fnite verb “tired” (in Turkish), and there is agreement between the subject (Romani frst person singular pronoun with Romani nominative case) and the Romani deontic fnite verb “want” (with present tense and indicative mood) and its complement clause verb, the Turkish fnite verb “rest” (with optative mood). In the Romani-Turkish fused lect, the use of agreement verb morphology from the two source languages, Romani and Turkish, is systematic. However, case comes from only one language, Romani. Te opposite pattern is not attested: Tere are no clauses with Turkish pronouns in combination with Romani fnite verbs. Tis means that there is a conventionalized way in which the two languages are mixed. Adamou and Arvaniti (2014, p. 228) also note that Romani-Turkish speakers use mixed strategies of phonological adaptation for the Turkish lexical material in a way that differs from the phonology of non-standard Turkish varieties. For example, Romani-Turkish speakers use metathesis: Turkish anlamayacak [anˈlamaˌdʒak] “he/she will not understand,” systematically becomes

Creoles, pidgins, and mixed languages

163

[alˈnamaˌdʒak]. Also, even though they use Turkish vowels [y ɯ œ] that are not part of Romani phonology, they do not use Turkish /h/ in either old borrowings like maala “neighborhood” from Turkish mahal or in recent borrowings like apo “pill” from Turkish hap. Despite its similarity to a V-N mixed language, Adamou and Granqvist (2015) do not argue that Romani-Turkish is a fully fedged mixed language. Tis is because there are few Turkish verbs in a conversation (i.e., approximately 12% of verbs come from Turkish). In comparison, full-fedged mixed languages such as Light Warlpiri and Gurindji Kriol are said to “derive their lexicon relatively evenly from their source languages” (O’Shannessy & Meakins, 2012, p. 384). Whether this is also apparent in conversation needs to be confrmed. In the meantime,Adamou and Granqvist (2015) suggest that the Romani data are most likely examples that illustrate the early stages of mixed languages, as has been documented for the Australian languages (McConvell & Meakins, 2005; O’Shannessy, 2012). Of course, unlike what happened in these Australian mixed languages, it appears that the mixing process has been interrupted in Romani-Turkish. How? Myers-Scotton hypothesizes that “the explanation would be that Romani speakers were in the process of shifing to Turkish, but that this was a Matrix Language Turnover that was arrested. For socio-psychological reasons, the shif stopped” (2013, p. 40). Adamou and Granqvist (2015) agree that an arrested matrix language turnover could be responsible for the creation of the Romani-Turkish variety. Tey hypothesize that a shif to Turkish may have started in the late nineteenth century. Indeed, similar Romani-Turkish mixing patterns are found in other modern-day Romani-speaking communities in the Balkans even though they no longer use Turkish for the past hundred years. Te shif to Turkish, however, must have been interrupted in the early twentieth century with the fall of the Ottoman Empire. Tis major change came with new borders afecting the mobility of the Roma populations who were working in trade and who traveled throughout the Balkans and relied on Turkish for their communication with clients. Te area of Trace became part of the Greek State in the 1920s, where Turkish became the minority language and Greek the language of the Greek state. Finally, in support of the hypothesis that fused lects are situated somewhere in between code-switching and mixed languages, Adamou and Shen (2019) examine the relative processing costs of mixed versus non-mixed sentences (see Section 3.1 for cognitive costs). Teir study confrms that Turkish verbs with Turkish verb morphology are processed similarly whether they occur in a Romani-Turkish mixed sentence or in an all-Turkish sentence. But the researchers also report differences that depend on the frequency of use of specifc lexical items; Turkish verbs that are frequently used when speaking Romani-Turkish are processed as fast in mixed Romani-Turkish sentences as in monolingual Turkish sentences. Tis means that a fused lect is composed of items that are highly frequent and conventionalized within a community but that it is also dynamic. Te RomaniTurkish data have now helped to revise the defnition of fused lects to specify that some elements of a fused lect may be obligatory, while others may still be variable (Auer & Hakimov, 2021).

164

From years to centuries

Roma from Drosero, Greek Thrace Te community of Drosero in Greek Trace is home to more than 7,000 Roma people. Many Roma from this community work in small trades or as seasonal workers, while men also work in ships. Te mixed Romani-Turkish variety is the home language and the language of in-group communication. Most children receive input in this Romani-Turkish variety, in the (extended) family and in the community of Drosero as well as in other Muslim Romani communities in Greek Trace, for example, in the Romani community settled in the outskirts of the town of Komotini, as intermarriages and exchanges between members of these communities are frequent. Roma from Drosero also use Romani with other Romani-speaking communities in Greek Trace (such as Christian Roma merchants, Dasikane Roma) and beyond (in other Greek cities or in the Balkans). Te Romani variety they use in these interactions is not the Romani-Turkish variety, Xoraxane Romane, in that they are careful to avoid the Turkish lexicon, in particular the Turkish verbs. In addition, most Roma children from Drosero also receive input by their families and community members in Turkish. At present, a wave of shif to Turkish is ongoing in the community and is completed among several Romani families. To understand such a language shif from the traditional minority language, Romani(-Turkish), to another minority language, Turkish, and not the ofcial language of the Greek state, Greek, one needs to keep in mind that Turkish has become in recent years an important language for the identity of Muslim Roma, as it is strongly related to the Muslim religion. Tis is supported by the fact that Turkish is also recognized as the language of the Muslim minority in Greek Trace (since 1923 by the Lausanne Treaty) and as such is the main language of education in several minority schools in the area. With the exception of the families who use Turkish in daily life, Turkish is generally used in the traditional market, the bazaar, and interactions with the religious authorities, as well as with other Turkish-speaking members of the Muslim minority. Moreover, Roma children from Drosero receive input in Modern Greek. Tis comes typically from outsiders, in communicational settings outside of Drosero, for example, when they accompany their families to work or to the city. Greek, which is the language of the Greek state, is also used in the administration, services outside of the community, and formal education. Unfortunately, access to schooling is strongly afected by the broader social exclusion and discrimination that Romani communities are confronted with both at the local and at the national level, despite the eforts that are made at the level of the European Union to ofer good schooling opportunities to all Roma children.

To conclude, mixed languages are forged in settings where individuals feel a need to retain and/or accentuate aspects of their blended culture.We can see these languages as the result of a resistance to societal pressures to shif to a dominant language, or

Creoles, pidgins, and mixed languages

165

they might be an act of identity construction in which speakers fashion themselves over time as culturally or ethnically distinct. Michif, Media Lengua, Young People’s Dyirbal, and other mixed languages are new languages, distinct from their source languages, and they have been formed, as fused lects, from conventionalized codeswitching practices. Tey are transmitted to children as the language of a community of speakers.

10.4

URBAN YOUTH LANGUAGES

In addition to mixed languages, there are well-known cases today of other hybrid linguistic language varieties that are created intentionally by their speakers. Tese, too, can become socially anchored as a frst language of a population, but their origin seems to be a distinctive multilingual speech style that is cultivated to index a certain toughness or coolness. Tese are usually called “urban youth languages” because they tend to be associated with young urban males, even though they are not restricted in use to such speakers, nor are such languages absent from rural settings (Makoni, 2017). Many of these linguistic creations are found in contemporary Africa where multilingualism abounds and the boundaries between languages are fuid (Nassenstein & Hollington, 2015; Makoni, 2017). Urban youth languages have been referred to as “anti-languages” because they are designed to be oppositional to standard language ideologies and because they are originally intended to disguise the message so that only those “in the know” can decipher it (Kiessling & Mous, 2004). Tough they are not usually structured very diferently from the dominant language that they use as a grammatical frame (matrix language), they are incomprehensible for monolingual listeners of the corresponding matrix language who are not familiar with the words and expressions used. In its original use, the term “anti-language” was associated with the cryptic language practices of criminal groups and street gangs. Today, the widespread distribution of urban youth languages across the continent and their use as frst languages for some African populations demonstrate that they have become mainstream. Te word formation processes of these youth languages are ofen playful, but structurally, many take the form of relexifcation, inserting borrowed and manipulated words and phrases from multiple source languages into the grammar of a matrix language, which can be colonial (French, English,Afrikaans) or African (Arabic, Zulu, Swahili, Lingala, Wolof, etc.) As an example, Zulu provides most of the grammatical frame for Iscamtho in South Africa, while a competing South African variety, Tsotsitaal, is based on the structure of Afrikaans. Similarly, in Nairobi (Kenya), Sheng uses a Swahili base mixed with insertions from Kikuyu, English, and other source languages, while Engsh, which is heavily used in social media and is identifed with “fancy” Kenyans, uses English as a base (Abdulaziz & Osinde, 1997). Sheng and Engsh are typical urban youth languages in that they are marked by lexical inventiveness, where the semantics of words and phrases take on new meanings. Te example of Sheng in 10.6 illustrates how the English elements, cold case and investigate, take on new meanings when embedded in a Swahili discourse.

166

From years to centuries

10.6 Sheng (adapted from Kioko, 2015, p. 126) ni-investigate cold case 1-sgS-eat:sub cold food “I had to eat cold food.”

African urban youth languages ofen confer a multilingual, youthful, swaggering persona on their users. Teir speakers are infuencers on social media and in real life, setting the style for how to speak for those who wish to cultivate a similar identity. To keep up with the latest styles, speakers need to learn new words and expressions. Nouchi (Ivory Coast), a mix of spoken Ivoirian French with words drawn from multiple linguistic sources, has a tradition of renewing its vocabulary in the media (Newell, 2009). To keep speakers up to date with the latest words and expressions, users share them in dictionary-like entries on social media and the internet. An entry from the Nouchi.com dictionary entry is shown in Figure 10.2. Te Nouchi dictionary is structured just like any other, registering the expression that means “heartbreak” in English. Te entry includes a synonym, goumin, also in Nouchi, along with a translation equivalent in French. Te example usage can be roughly translated as “the girl lef me heartbroken.” While the syntax is easily understood as French, the elements go“girl,” originally from English “girl” and the new term, tchindjouss, disguise the meaning for those who have not learned Nouchi expressions. Linguists and anthropologists alike see the hybrid linguistic forms found in African urban youth languages as manifestations of a continual process of identity formation and reformation that also involve modes of dress, music styles, and body language (Makoni, 2017). In other words, these hybrid linguistic forms are part of fashioning new ways of being in the contemporary world.

Figure.10.2 Screenshot of dictionary entry for “tchindjouss” from Nouchi.com

Source: Retrieved Sept 1, 2022 from www.nouchi.com/dico/liste-des-derniers-mots/item/tchasse.html

Creoles, pidgins, and mixed languages

167

Chapter summary We have seen three types of new languages that arise historically in language contact settings under particular social circumstances: creoles, pidgins, and mixed languages. We have also seen that new styles, urban youth languages, can also be cultivated in post-colonial settings and other multilingual/multiethnic contexts as acts of identity. Creoles are languages that arose in the settlement colonial plantations as part of the slave trade. Pidgins are those that are used in trade. Although many researchers in the past have argued that pidgins represent the initial stages of the development of a Creoles, this idea has been largely discredited. Now, we view Creoles as non-exceptional languages that developed in the context of multilingual colonial societies where power structures were not equal. Mixed languages are formed in bilingual communities with a strong cultural resistance to language shif.While both Creoles and mixed languages are spoken by children, pidgins typically are not. From a linguistic perspective, Creoles have a main lexifer language, ofen the language of the colonizers, and one or more substrate languages that provide most of the grammatical structure; these were the languages spoken by the enslaved people. Pidgins also have a lexifer language, usually the language of the foreign traders, and one or more substrate languages spoken by the local populations. Mixed languages can be either split along the lines of the grammar versus the lexicon, or the grammars and lexicons of both source languages can contribute to a mixed language in the form of a verb-noun split, where each contributing language is identifable. Mixed languages develop from the conventionalization of code-switching patterns.We underline the fact that Creoles and mixed languages include innovations that are not found in the contributing languages, so they are not merely the product of two languages added together but the outcome of a creative integration of linguistic features. Finally, the presence of urban youth languages in many locations of the world, but particularly in Africa where they are prevalent and well known, shows us that speakers can mix their languages intentionally to stand apart from and resist conservative attitudes and to project new ways of being.

EXERCISES 1.

Discover

Visit the website https://apics-online.info/ to discover the Atlas of Pidgin and Creole Language Structures, which provides information on 130 grammatical and lexical features of 76 Pidgin and Creole languages from around the world (Michaelis et al., 2013). Look at the map of where you fnd these languages. Where are they found and where are they absent? Tink about why they are not evenly distributed around the globe.

168

2.

From years to centuries

Discuss

Visit the MIT-Haiti initiative’s website https://haiti.mit.edu/about/. Teir goal? Promote the use of Kreyòl in science, technology, engineering, and math (STEM) disciplines to allow for Haitians to learn in the language most of them speak at home. Indeed, in Haiti, 95% of the population is fuent in Kreyòl only, while just 5% speaks French fuently. Despite that, education is conducted in French. For what reasons might a nation choose to educate its citizens in a language that the majority of its citizens do not speak? Can you think of parallel cases in other places of the world or at other points in history?

3.

Checkpoint

What are the main linguistic and sociolinguistic diferences between Creoles, pidgins, and mixed languages? Do these outcomes of language contact arise through exceptional linguistic processes, or can we understand them through general language contact processes?

4.

Discuss

In what ways do you think that African urban youth languages might difer from language styles that borrow from or code-switch freely with English, like Spanglish (Spanish-English), Chinglish (Chinese-English), and Hinglish (Hindi-English)?

FURTHER READING Auer, P. (2014). Language mixing and language fusion: When bilingual talk becomes monolingual. In J. Besters-Dilger, C. Dermarkar, S. Pfänder, & A. Rabus (Eds.), Congruence in contact-induced language change: Language families, typological resemblance, and perceived similarity (pp. 294–334). De Gruyter. Peter Auer describes how new language varieties, like mixed languages and antilanguages, are formed through the normal processes of language mixing and fusion. Mufwene, S. S. (2015). Pidgin and Creole languages. In J. D. Wright (Ed.), International encyclopedia of the social and behavioral sciences (pp. 133–145). Elsevier. One of the world’s most prominent linguists presents a non-technical overview of the controversies concerning Creoles and pidgins in an attempt to debunk myths about these language varieties.

REFERENCES Abdulaziz, M. H., & Osinde, K. (1997). Sheng and Engsh: Development of mixed codes among the urban youth in Kenya. International Journal of the Sociology of Language, 125(1), 43–64. https:// doi.org/10.1515/ijsl.1997.125.43 Aboh, E. O. (2015). Te emergence of hybrid grammars: Language contact and change. Cambridge University Press. http://doi.org/10.1017/cbo9781139024167 Adamou, E. (2010). Bilingual speech and language ecology in Greek Trace: Romani and Pomak in contact with Turkish.Language in Society, 39, 147–171. https://doi.org/10.1017/s0047404510000035

Creoles, pidgins, and mixed languages

169

Adamou, E., & Arvaniti, A. (2014). Greek Trace Xoraxane Romane (illustrations of the IPA). Journal of the International Phonetic Association, 44(2), 223–231. Adamou, E., & Granqvist, K. (2015). Unevenly mixed Romani languages. International Journal of Bilingualism, 19, 525–547. https://doi.org/10.1177/1367006914524645 Adamou, E., & Shen, X. R. (2019). Tere are no language switching costs when code-switching is frequent. International Journal of Bilingualism, 23, 53–70. https://doi.org/10.1177/1367006917709094 Ansaldo, U., & Lim, L. (2021). Language contact in the Asian region. In E. Adamou & Y. Matras (Eds.), Te Routledge handbook of language contact (pp. 434–461). Routledge. Auer, P. (1999). From code-switching via language mixing to fused lects: Toward a dynamic typology of bilingual speech. International Journal of Bilingualism, 3(4), 309–332. Auer, P. (2014). Language mixing and language fusion: When bilingual talk becomes monolingual. In J. Besters-Dilger, C. Dermarkar, S. Pfänder, & A. Rabus (Eds.), Congruence in contact-induced language change: Language families, typological resemblance, and perceived similarity (pp. 294–334). De Gruyter. Auer, P., & Hakimov, N. (2021). From language mixing to fused lects: Te process and its outcomes. International Journal of Bilingualism, 25(2), 361–368. Bakker, P. (1994). Michif, the Cree-French mixed language of the Métis bufalo hunters in Canada. In P. Bakker & M. Mous (Eds.), Mixed languages: 15 case studies in language intertwining (pp. 13–33). IFOTT. Bakker, P. (1997). A language of our own: Te genesis of Michif, the mixed Cree-French language of the Canadian Métis. Oxford University Press. Baptista, M. (2020). Competition, selection, and the role of congruence in Creole genesis and development. Language, 96(1), 160–199. https://doi.org/10.1353/lan.2020.0005 Blasi, D. E., Michaelis, S. M., & Haspelmath, M. (2017). Grammars are robustly transmitted even during the emergence of creole languages. Nature Human Behaviour, 1, 723–729. https://doi.org/10.1038/ s41562-017-0192-4 Bordal Steien, G., & Yakpo, K. (2020). Romancing with tone: On the outcomes of prosodic contact. Language, 96(1), 1–41. https://doi.org/10.1353/lan.2020.0000 DeGraf, M. (2003).Against Creole exceptionalism. Language, 79(2), 391–410. https://doi.org/10.1353/ lan.2003.0114 DeGraf, M. (2020). Toward racial justice in linguistics: Te case of Creole studies (Response to Charity Hudley et al.). Language, 96(4), e292–e306. Frajzyngier, Z., Gurian, N., & Karpenko, S. (2021). Language formation by adults: Te case of SinoRussian idiolects. Brill. http://doi.org/10.1163/9789004465848 Hussain, R.,Asif, M., & Din, M. (2020).View of Urdubic as a lingua franca in the Arab countries of the Persian Gulf. Research Journal of Social Sciences and Economics Review, 1(4), 225–232. Kiessling, R., & Mous, M. (2004). Urban youth languages in Africa. Anthropological Linguistics, 46(3), 303–341. Kioko, E. (2015). Regional varieties and “ethnic” registers of Sheng. In N. Nassenstein & A. Hollington (Eds.), Youth language practices in Africa and beyond (pp. 119–148). De Gruyter Mouton. Lipski, J. M. (2019). Field-testing code-switching constraints: A report on a strategic languages project. Languages, 4(1), 7. https://doi.org/10.3390/languages4010007 Lufn, X. (2013). Kinubi structure dataset. In S. M. Michaelis, P. Maurer, M. Haspelmath, & M. Huber (Eds.), Atlas of Pidgin and Creole language structures online. Max Planck Institute for Evolutionary Anthropology. Retrieved June 14, 2022, from http://apics-online.info/contributions/63 Makoni, B. (2017). Urban languages in African contexts: Toward a multimodal approach to urban languages. In O. García, N. Flores, & M. Spotti (Eds.), Te Oxford handbook of language and society (pp. 281–298). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780190212896.013.31 Manfredi, S., & Petrollino, S. (2013). Juba Arabic structure dataset. In S. M. Michaelis, P. Maurer, M. Haspelmath, & M. Huber (Eds.), Atlas of Pidgin and Creole language structures online. Leipzig: Max

170

From years to centuries

Planck Institute for Evolutionary Anthropology. Retrieved June 14, 2022, from http://apics-online. info/contributions/64 McConvell, P., & Meakins, F. (2005). Gurindji Kriol: A mixed language emerges from code-switching. Australian Journal of Linguistics, 25, 9–30. https://doi.org/10.1080/07268600500110456 Meakins, F. (2013). Mixed languages. In P. Bakker & Y. Matras (Eds.), Contact languages (pp. 159–228). Walter de Gruyter. Michaelis, S. M., Maurer, P., Haspelmath, M., & Huber, M. (Eds.) (2013). Atlas of Pidgin and Creole language structures online. https://apics-online.info/ Mufwene, S. S. (2008). Language evolution: Contact, competition and change. Continuum International Publishing Group. Mufwene, S. S. (2015). Pidgin and Creole languages. In J. D.Wright (Ed.), International encyclopedia of the social and behavioral sciences (pp. 133–145). Elsevier. Mufwene, S. (2021). Creoles and pidgins. In E.Adamou & Y. Matras (Eds.), Te Routledge handbook of language contact (pp. 300–324). Routledge. Muysken, P. (1994). Media Lengua. In P. Bakker, & M. Mous (Eds.), Mixed languages: 15 case studies in language intertwining (pp. 201–205). IFOTT. Myers-Scotton, C. (1998).A way to dusty death: Te matrix language turnover hypothesis. In L. Grenoble & L. J. Whaley (Eds.), Endangered languages: Language loss and community response (pp. 289– 316). Cambridge University Press. Myers-Scotton, C. (2013). Paying attention to morpheme types: making borrowability more precise. In C. de Feral (Ed.), In and out of Africa languages in question (pp. 31–42). Peeters. Nassenstein, N., & Hollington, A. (2015). Youth language practices in Africa and beyond. De Gruyter Mouton. Newell, S. (2009). Enregistering modernity, blufng criminality: How Nouchi speech reinvented (and fractured) the nation. Journal of Linguistic Anthropology, 19(2), 157–184. O’Shannessy, C. (2012). Te role of codeswitched input to children in the origin of a new mixed language. Linguistics, 50, 305–340. https://doi.org/10.1515/ling-2012-0011 O’Shannessy, C. (2013). Te role of multiple sources in the formation of an innovative auxiliary category in Light Warlpiri, a new Australian mixed language. Language, 89, 328–353. https://doi. org/10.1353/lan.2013.0025 O’Shannessy, C., & Meakins, F. (2012). Comprehension of competing argument marking systems in two Australian mixed languages. Bilingualism: Language and Cognition, 15, 378–396. https://doi. org/10.1017/s1366728911000307 Perekhvalskaya, E. (2013). Chinese Pidgin Russian. In S. M. Michaelis, P. Maurer, M. Haspelmath, & M. Huber (Eds.), Atlas of Pidgin and Creole language structures online. Max Planck Institute for Evolutionary Anthropology. Retrieved June 16, 2022, from http://apics-online.info/contributions/65 Schultze-Berndt, E., & Angelo, D. (2013). Kriol structure dataset. In S. M. Michaelis, P. Maurer, M. Haspelmath, & M. Huber (Eds.), Atlas of Pidgin and Creole language structures online. Max Planck Institute for Evolutionary Anthropology. Retrieved June 14, 2022, from http://apics-online.info/ contributions/25 Sebba, M. (1997). Contact languages: Pidgins and creoles. St. Martin’s Press. Tomason, S., & Kaufman, T. (1988). Language contact, creolization, and genetic linguistics. University of California Press. Yakpo, S. (2021). Social factors. In E. Adamou & Y. Matras (Eds.), Te Routledge handbook of language contact (pp. 129–146). Routledge.

11 Minority languages, heritage languages, and immigrant linguistic practices

CHAPTER PREVIEW In the fnal chapter, we consider contemporary language contact settings. We investigate the practices of individuals who are in the linguistic minority, those who use a language in their networks of family, friends, and neighbors that is diferent from the language spoken by the socially dominant population around them. We consider the social and psychological factors that lead to language maintenance or language shif in the types of bilingual populations under consideration. And, fnally, we examine the efects of language contact and restricted usage on the minority language. Tis chapter invites us to ask: What can we observe about language contact in minority language and immigrant communities?

11.1

DIFFERENT TYPES OF BILINGUAL COMMUNITIES

It is estimated that there are 7,000 languages in the world today. Tese are not evenly distributed throughout the world; some regions have many more languages than others. Asia, Africa, and the Pacifc combined contain 80% of the world’s Indigenous languages, while Europe and the Americas count for the remaining 20%. Te highest concentration of languages is found in the island nation of Papua New Guinea, where there are around 840 languages, most of them Indigenous and many of them endangered. Te languages of Papua New Guinea alone represent over 10% of the unique languages of the world (Kik et al., 2021). You do not have to live in areas that are rich in Indigenous languages like Papua New Guinea, Indonesia, or Tanzania to observe many diferent languages in use around you. In fact, over 700 languages are found among the current residents of New York City (NYC) (Kaufman & Perlin, 2018)! Te diversity of languages in NYC was brought about largely by people who have relocated from other countries, bringing with them “diaspora languages.” Diaspora languages are those that are spoken by a community of language users who have been displaced from their native lands. Even though many scholars focus on large, urban areas, like London or Dubai, as the location for a “superdiversity” of languages (Blommaert & Rampton, 2011), you DOI: 10.4324/9781003167952-14

172

From years to centuries

do not need to live in a large, urban setting like NYC to observe diverse languages in contact around you. Diaspora languages are found in small towns and rural settings, too, as migrants from a particular region tend to follow in the path of pioneers from their region who emigrated before. As a result, we can fnd diasporic language communities in rural areas as well as urban ones. For example, Somali and Hmong speakers relocated as refugees to small towns in Wisconsin, in the northern United States (Brown & Carpenter, 2018). And even though we think of Wales as the home for Welsh, there are a few thousand speakers of Welsh in rural Patagonia in southern Argentina. A large percentage of the population of the small town of Ōizumi, Japan was born abroad; this includes repatriated Japanese Brazilians who speak Portuguese and more recent diasporic groups from Nepal and other areas of Asia. And there are about 300,000 Chinese immigrants from Fujian Province who have relocated to small, rural towns across South Africa, adding to the multilingual tapestry of that country (Rumbaut & Ima, 1988). When a community of language users is dispersed to a new region where their language variety is in the minority, the result is language contact. From the diasporic movements we have mentioned, the following potential language contact scenarios arise: Somali and Hmong in contact with English in North America; Portuguese, Nepalese, and Vietnamese in contact with Japanese in Ōizumi; and Min, Mandarin, Hakka, and Taiwanese in contact with the languages of the Free State, South Africa (English, Sotho, Tswana, Afrikaans, Zulu, and Xhosa, among others). We can fnd evidence of diaspora languages close to home by using our ears as well as our eyes. Te “linguistic landscape,” which refers to the visual language displayed in public spaces, can provide clues to the languages spoken around you (Shohamy et al., 2010; Shohamy, 2012; Blommaert, 2013). Depending on where you live, you might notice that your linguistic landscape includes restaurants, salons, shops, or public health announcements and tweets that feature writing in Arabic, Hindi, Tai, Navaho, Spanish, Chinese,Yoruba, Turkish, Vietnamese, or Amharic.You might even see mixed language signs, like the one in Figure 11.1, which uses English and Spanish (desde “from”) to declare that the services on ofer cost (from) $7 and up. In many modern-day settings, we can identify two broad types of bilingual populations: relocated populations and Indigenous minority language populations. Relocated populations include “frst-generation immigrants” and their ofspring, called “heritage speakers.” “Indigenous minority language speakers” are native to the region in which they live, but their language is not the numerically dominant one in the region where they live. Some individuals who have little or no home exposure to a minority language choose to acquire it later in life for cultural or ideological reasons. Tese people are known as “new speakers” (O’Rourke et al., 2015). Moreover, Chen Pichler et al. (2018) propose that hearing children of Deaf adults (CODAs) and Deaf cochlear implant users from Deaf families are “heritage signers” because they use a minority language at home and a majority language beyond home. First-generation speakers are those who migrated in adulthood afer having acquired a language or languages of their homeland. Tey can be late bilinguals of the majority language in their new location, initiating their language learning when they are adolescents or adults. Tey ofen have a naturalistic learning experience in

Minority languages and heritage languages

173

Figure 11.1 Bilingual sign, Spanish-English, Austin, TX

the majority language rather than an instructed one, so they may not be literate in it. First-generation immigrants may be members of networks composed of speakers who share multilingual profles similar to their own, particularly immigrants who are part of a larger immigrant community like the New York and Los Angeles Spanish-speaking communities in the United States of America (Poplack, 1980; SilvaCorvalán, 1994) or Turkish-speaking communities in the Netherlands (Backus, 1992). Having communication networks in their language helps them to maintain its use in the diaspora. Immigrants who were born in their country of origin but arrive in their new destination prior to age 12 are referred to as “Generation 1.5” (Rumbaut & Ima, 1988). Unlike their parents, they are still of school age at the time of immigration, so they receive instruction in the majority language.As they build their social networks with classmates and peers, they tend to acquire the majority language quickly and profciently, to the point that it can become their dominant language. Because they were fully immersed in the ancestral language prior to immigration, many Gen 1.5 individuals remain regular users of both their languages throughout their lives. “Heritage speakers” are the children of migrant families who are born or arrive at an early age in the destination country. Tey, too, are frst language speakers of their family’s “heritage language.” A heritage language is spoken at home; it is not the

174

From years to centuries

dominant language of mainstream society (Montrul & Polinsky, 2019). For many heritage speakers, the interactions they have in the family language occur most intensely during early childhood and within a naturalistic setting rather than in a classroom or among peers. Heritage speakers do not always learn to read or write in their parents’ L1, and, sometimes, they become “receptive bilinguals,” meaning that they can understand the heritage language when it is spoken around them, but they do not speak it or like to speak it. As they age, heritage speakers grow to have more interaction with their peers, and more facility with reading and writing, in the dominant language. “Indigenous minority language speakers” use languages that are spoken by a minority of the population where they live. Examples include the Indigenous (or Aboriginal) speakers of Australian languages, like Warlpiri and Tiwi in Australia, Breton speakers in France, Bambara and Mandjak speakers in Senegal, Mapuche speakers in Chile, Khoisan speakers in South Africa, and Mi’kmaq or Dene in Canada. Indigenous minority language speakers are ofen bilinguals, typically speaking a socially dominant language as well as their ancestral language. Te socially dominant languages are used in most communication settings, including in the public sphere and powerrelated domains, such as education and the legal system. Te Indigenous and minority languages are used in a limited number of interactions, usually just in the family sphere or for transactions between members of the language group. Research shows that interactions with older adult members of the community can help preserve the Indigenous minority language; a study of the Gurindji language of the Pama-Nyungan family in Australia found that living with a member of the older generation who is fuent in Gurindji is the most important factor in predicting whether an individual community member uses the language (Bromham et al., 2020). Te label “new speakers” is used primarily in reference to adult learners of the minority languages of Europe, such as Breton, Basque, and Irish. New speakers are “individuals with little or no home or community exposure to a minority language but who instead acquire it through immersion or bilingual educational programs, revitalization projects or as adult language learners” (O’Rourke et al., 2015, p. 1). Te term is also applied to speakers who choose to switch their language dominance for ideological reasons. Tis is common in Spain within the autonomous regions of Galicia, Catalonia, and the Basque Country where minoritized languages are found and where people feel strong ties to their region and culture. New speakers can be crucial for the revitalization of an endangered language population; some examples include new speakers of Maori in New Zealand (Harlow et al., 2009) and of Irish in Ireland, where new speakers of Irish outnumber L1 speakers (McCloskey, 2008). New speakers are essential for the reclamation of a language that had no living L1 speakers, referred to for this reason as dormant languages. Trough the dedication of linguists, educators, parents, and children from the Wampanoag Nation in eastern Massachusetts, their language,Wôpanâak, has speakers again, 150 years afer it had disappeared as a spoken tongue (Baird et al., 2011). Ofen, the social economic situation of minority language speakers can be precarious. Many are from economically marginalized communities and bear the cultural baggage of this background. Migrants can be uprooted from their regions of origin and displaced for reasons that can be socially and psychologically traumatic,

Minority languages and heritage languages

175

including economic collapse, violence, famine, and environmental devastation. As migrants, refugees, or asylum seekers, they are not always universally welcomed by their receiving communities, and they may sufer from discrimination and harassment, particularly if they are perceived to be of a diferent ethnicity or religion from the dominant majority. Tey are expected to assimilate to the culture and habits of the receiving community, which means adopting the language of the majority in addition to, or in place of, the language of their family. Minority languages are spoken by all the members of a social group. Te term minority is generally taken to mean that language users of this group are in the numeric minority, but this does not have to be the case. It can also mean that the social group has less social, political, or economic power than surrounding groups. When the language has been marginalized through this power imbalance, we use the term “minoritized.” Tere are many areas of the world where languages might be spoken by millions of people, but they have less status than an ofcial language or a colonial language.An example is Taiwan, where Taiwanese is at risk of being displaced by Mandarin even though it is spoken by millions of people. In many regions, there is a recognition that language is an integral part of the culture and identity of a group and that protecting minority language communities respects the human rights of individuals. In others, though, minority languages are seen as a threat to national, or colonial, cohesion and identity. Despite the diferences in their socio-political setting and status, there are similarities in the language practices of speakers of immigrant and Indigenous minority languages that warrant their discussion together in this chapter. New speakers, too, should be included in the language ecology of contact communities because they participate in the circulation of linguistic variants of the minority languages they are learning. Nearly all minority language users are restricted in the contexts in which they can use them relative to a socially dominant language. Even if their language is viewed as a distinctive part of their heritage and identity, many speakers fnd it necessary to become bilingual in languages that have a greater reach outside their communities. So we ofen fnd that the language practices of members of older and young generations difer, with the older generation using the minority language more ofen than members of the younger generation. As we will see in the next section, generational diferences play a role in language choice.

Urban youth languages As young people are in the process of discovering their identity, they typically experiment with language. Urban multicultural and multilingual settings are particularly prone to this linguistic experimentation. Linguists note that specifc speech styles are adopted by these young speakers, drawing from the various languages that are spoken in the city. For example, young speakers in urban settings will deliberately use L2 phonetics, lexicon, and syntax in their speech. Tis is the case of the “Moroccan Flavored Dutch” that is used in the

176

From years to centuries

Netherlands, not just by Moroccan-origin groups but by various groups of young people (Dorleijn et  al., 2021). In fact, one important characteristic of urban youth speech is that it goes beyond the linguistic group whose speech characteristics are the source. A pioneering study making this apparent was about the infuence London Jamaican Creole spoken by young Jamaicans had on the speech of White working-class teens from South London (Hewitt, 1986). See Chapter 10 for more on urban youth languages.

11.2

LANGUAGE MAINTENANCE AND LANGUAGE SHIFT

Te linguist Joshua Fishman observed that unless a heritage or minority language serves specifc and necessary needs within society, it will lose ground to the majority language that people begin to use for a wider range of functions. “Language shif” occurs when “a community of users replaces one language by another” (Grenoble, 2021). Although there are exceptions that we will discuss later, language shif in immigrant communities in North America and in Europe ofen takes place within three generations in a family. Te original immigrants, who are ofen referred to as Generation 1, or Gen 1, will be fully profcient in the immigrant language and may acquire profciency in the majority language afer they relocate, or sometimes even before. Teir children, Gen 2, will be bilingual, using both the heritage language of the family and the majority language of the wider society. Ofen members of Gen 2, even as children, will serve as “language brokers,” translating between the family and majority language or interpreting for their parents. When their children, Gen 3, only use the majority language, language shif in the family is complete. Language shif to a socially dominant language is frequent in Indigenous minority language communities, as well. Sometimes, this occurs through “in-migration,” when individuals speaking a minority language relocate from their place of origin to a new locale, ofen moving from the countryside to a city. In Africa, minority languages are disappearing as children acquire major African languages like Urban Wolof, Hausa, Igbo,Yoruba, or Swahili in place of their Indigenous minority languages. For example, 80% of the population of Senegal now speaks Wolof, although only 43.7% are ethnically Wolof (Mc Laughlin, 1995). And in India, where multilingualism is the norm, speakers of Indigenous minority languages have nonetheless lost the social and economic beneft that their languages once held locally as they have been pushed out of village markets, where they used to barter in their own language (Mohanty, 2010). Some minority language speakers view their linguistic assimilation to the socially dominant culture as positive, or neutral, and do not fret if their children do not learn it. For instance, members of the Seereer group of Senegal, who speak multiple different minority languages, do not associate their languages with their identity. For them, the shif to Wolof when they relocate to the urban capital, Dakar, does not signal a threat to their culture or identity. But another minority language group, the HaalpulaarɁen, do resist Wolofzation because the varieties of Fula that they speak remain signals of their group identity (Mc Laughlin, 1995).

Minority languages and heritage languages

177

Figure 11.2 Tweeters expressing regret for their ancestral language; accessed June 2, 2022

Younger members of communities that are undergoing or have undergone language shif may come to regret not having learned, or learned more of, their ancestral language, as you can read in the tweets in Figure 11.2. Sometimes these individuals become new speakers, learning the language as second language speakers. But as they are ofen older when they begin to learn, or to relearn, a heritage language, their language use can be subject to L1 transfer. For example, researchers document the English-infuenced phonetic properties of new speakers in a Maori revitalization project in New Zealand (Harlow et al., 2009). While language shif within an immigrant family can happen relatively rapidly, shif at the community level might occur more slowly because new generations of speakers arrive, replenishing the presence of the heritage language in the community. Tis is the situation for Spanish-English bilingual communities in California or Texas and for millions of Turkish speakers abroad in Germany, Austria, France, and the Netherlands.Worldwide, increased globalization of travel and communication allows those who wish to retain ties to the language and culture of their sending communities to participate in transnational networks. But we do not yet know if transnational networks are efective in prolonging bilingualism over multiple generations within immigrant families. Even though there are economic and social pressures to shif to a socially dominant language, the long-term maintenance of immigrant and minority languages is possible. Tere are many possible factors that favor language maintenance over shif. One of the most well-known factors is religious or cultural isolation. Members of some

178

From years to centuries

immigrant religious groups live apart from the wider society and have maintained their language afer many generations of bilingualism. For instance, Old Order Amish and Mennonites continue to use Pennsylvania Dutch, the Hasidim still speak Yiddish, and adolescent Muslims from the Egyptian diaspora have been found to maintain Arabic in Athens while their Coptic Christian peers shif to Greek (Gogonas, 2012). But religious conservatism is only one possible factor leading to language maintenance. Tere are other cases of long-term bilingualism of a minority or immigrant language without complete language shif where religion is not a major factor. Molise Slavic, originally from the area of Herzegovina, has been spoken in the province of Campobasso, Italy, for 500 years (Breu, 2020). Veneto, a language from northern Italy, is still spoken in Chipilo, Mexico, following its implantation there a century ago (Mackay, 1992), and Hakka speakers from China preserved their language as they emigrated frst to East Timor in the nineteenth century and then to Australia (Hajek & Goglia, 2020).And populations of Indigenous minorities have managed to maintain their languages under ofen long-lasting, brutal colonial regimes and in the face of population loss in the Americas, South and Central Asia, Africa, and Australia. Language shif only occurs in situations of language contact but, as these examples show, language contact does not necessarily lead to language shif (Bromham et al., 2022). In fact, most Indigenous communities have been and continue to be in constant contact with other language communities in local or regional networks. Worldwide, minority language speakers tend to be bi- or multilingual, capable of communicating in the socially dominant language of their regions as well in their own language.A study of the languages of Dagestan, in the Central Caucasus, compared the size of the populations of Dagestani villages with reports of how many languages members of each village were able to speak (Dobrushina & Moroz, 2021). Te researchers found an inverse correlation between population size and the number of reported second languages; villages with a smaller population size reported a greater number of second languages among their residents. Tis suggests that exchanges of goods and services between smaller villages required communication between diferent groups, which fosters language learning and sustains bilingualism. People from larger villages do not need to establish social networks outside their local region and can remain monolingual. Many explanations for language maintenance and shif are based on the idea that minority languages are preserved when language users form dense social networks, as discussed in Chapter 5, and when the language is used across the domains of work, friendship, school, worship, and family life (Fishman, 2000). Scholars working on the history of shif among the Germanic-speaking heritage communities in the upper midwestern United States and the maintenance of Cherokee in North Carolina (Frey, 2013) have adopted a “verticalization model” for language shif (Frey, 2013; Brown, 2017; Salmons, 2022). Tey argue that shif occurs with a change in social structure from a horizontal model, where communities operate their own institutions, to a vertical model, where outside infuences—governments, national media, school boards, “big-box” stores—exert control. Tis idea is consistent with a fnding of Bromham et al. (2022) that the number of roads connecting previously remote endangered language communities to larger towns is one of the greatest predictors of language endangerment.

Minority languages and heritage languages

179

11.3 THE OUTCOMES OF LANGUAGE CONTACT ON MINORITY AND HERITAGE LANGUAGES In this section, we will consider what happens to minority and heritage languages in contact with socially dominant ones. Recall from Chapter 8 that linguists tend to identify two types of contact-induced changes, one in which items are borrowed from a source language and a second type in which only the meanings and structures from the source language are found in the recipient/replica language. Some studies in contact linguistics have made claims about what kinds of language elements should appear in diferent situations of language contact (Tomason & Kaufman, 1988). Tey propose that the two types of contact-induced changes result from diferent social contexts of contact, and each leads to a diferent type of efect on a language. Tomason and Kaufman (1988) defne borrowing as “the incorporation of foreign features into a group’s native language by speakers of that language: the native language is maintained but is changed by the addition of the incorporated features” (1988, p. 37). In the early stages, borrowing is limited to vocabulary. Tey call the other type of contact outcome “interference through shif” and propose that it occurs when a “group of speakers shifing to a target language fails to learn the target language perfectly” (9). Tey claim that this requires more extensive bilingualism on the part of the shifing speakers because they must know something about the structure, as well as the vocabulary, of the recipient language. So, the specifc type of change, borrowing versus structural interference, is linked to speakers of diferent profciencies. Te prediction Tomason and Kaufman make is that with more intensive language contact, there are more (source language) speakers who shif, and, as they do so, they use features from their source language in the recipient language, creating structural changes of all kinds. Tey propose a “borrowing scale” that links intensity of contact and degree of bilingualism with specifc linguistic efects (Tomason & Kaufman, 1988, pp. 74–75; Tomason, 2001). Tis is shown in Table 11.1. Winford (2013) makes similar predictions about a scale for minority languages undergoing language shif.According to Winford, in the earlier stages of contact, this infuence may be restricted to substantial lexical borrowing,” but“the greater intensity of contact during the phase of bilingualism and shif, as well as asymmetry in power and prestige of the languages involved, promote increasing structural infuence from the dominant language on the subordinate ancestral language” (2013, p. 382). While these proposals have been important in contact linguistics, they cannot explain some aspects of the linguistic outcomes of contact. Te problem is they assume that interference through shif, what we have called convergence in Chapter 8, is limited to non-dominant recipient language users who are using more of the socially dominant language than their ancestral one. It was another contact linguist of the era of Tomason and Kaufman who pointed out that bilingual speakers can be balanced, or nearly so, in their language dominance and that they can change their language dominance during their lifespan (van Coetsem, 1988). So what is called interference through shif occurs in the language of very fuent bilingual speakers, as well as in less fuent ones. Indeed, we have seen throughout this book that all kinds of bilinguals show signs of language interaction in the structures and lexicons of their languages and that

180

From years to centuries

Table 11.1 Borrowing scale adapted from Tomason (2001, pp. 70–71) Intensity of language contact Casual Slightly more intense contact More intense contact

Intense contact

Level of bilingualism

Linguistic elements

Borrowers need not be fuent speakers of SL More fuent individuals of source language among borrowers, but they are still the minority More bilinguals among borrowers, and societal attitudes favor borrowing

Lexicon: only content words Structure: none Lexicon: function and content words Structure: a few new phonemes in loanwords Lexicon: more function words Structure: more signifcant effects in all areas of linguistic structure Lexicon: heavy borrowing Structure: anything goes

Extensive bilingualism among borrowers

the interaction can be bidirectional (see examples for Ngigua-Spanish bilinguals in Chapter 2). Current adaptive models further predict that the more one uses the two languages, the more bidirectional transfer is likely to happen (Adamou, 2021). Another problem with these early proposals in contact linguistics is that they are difcult to test. If the linguistic outcome of language contact depends on intensity of the contact, as claimed, then researchers need to measure the degree of contact as well as the amount of infuence of the dominant language on the minority one. Te degree of contact is usually determined by considering the history and demography of a community. Some linguists have suggested that there are social and sociolinguistic factors that afect contact intensity, like population size (large vs. small), social network structures (tight vs. loose), and types of contact (long-term contact with child bilingualism vs. short-term contact with adult bilingualism) (Trudgill, 2009). But it is still not clear how we could distinguish a setting with “casual contact” from a setting with “slightly more intense contact” or from one with “intense contact” (see Table 11.1). To measure the linguistic efect of one language on another, linguists have turned to the analysis of transcribed interview data to count how frequently contact elements appear in the natural communication of bilinguals and how this frequency might vary depending on the settings in which they are spoken (Adamou, 2016; Guzmán et al., 2017). Adamou (2016), for example, compares 15 small bi- or multilingual spoken corpora. Te corpora cover a variety of contact settings, including Slavic and Romani in contact with western European languages, Ixcatec in contact with Spanish in Mexico, and data from several Afro-Asiatic languages (https://pangloss.cnrs.fr and http://corpafroas.huma-num.fr/ Archives/ListeFichiersELAN.php). She also analyzed contact features of Indigenous languages in contact with Spanish in the Americas from counts published in Gómez Rendón (2008) for Guarani in Paraguay, Quichua in Ecuador, and Otomí in Mexico.

Minority languages and heritage languages

181

Te analysis of these corpora focused on calculations of the frequency of borrowings from the dominant languages (L2) in the minority languages. Te results show that, across corpora, most contain less than 5% of words from the L2, so borrowing is infrequent. Tese low rates of borrowing are also confrmed in other language contact corpora analyzed by Poplack (2017). In rare cases, in the minority languages Adamou studied, 20%–35% of the word tokens come from the L2. Overall, the analysis of the corpora showed that there is no correlation between the intensity of contact and the amount of L2 word tokens found in a corpus. Instead, it appears that speakers of a minority language follow their community’s norms of language mixing. In other words, they accommodate to one another within their communities. For example, Molise Slavic speakers living in Italy are among the high borrowers (20%–35%), and their language shows changes at all levels of grammar from their long-term contact with Italian (Adamou et al., 2016). Speakers of Colloquial Upper Sorbian in Germany, also in long-term contact, use very few lexical borrowings from German (less than 5%). But German infuence is widespread in the structure of Upper Colloquial Sorbian in the forms of semantic extensions, lexical calques, and structural convergence. Each community shows intense contact with extensive bilingualism, yet the contact-induced outcomes detectable in the speech of the members of these communities proves to be very diferent. What appears to account for the diference between these contact communities is not length or intensity of contact but the presence of a literary tradition and prescriptive linguistic attitudes about the minoritized language. Molise Slavic speakers do not have a strong literary tradition,and it is only recently that any conservative attitudes toward the correctness of Molise Slavic have been introduced into the community, but they remain weak. Te low proportion of German words in Colloquial Upper Sorbian, on the other hand, seems to be connected to the strong linguistic tradition of Sorbian intellectuals who promote the use of Upper Sorbian. In this case, using German words while speaking Sorbian is frowned upon, but calquing and structural convergence are not. We might wonder why the language spoken by a minority and immigrant community continues to change as fewer and fewer people use it. In every language community, there are diferent ways of saying the same thing, and the choice of using one linguistic variant over another is determined, in part, by social factors that can carry positive or negative evaluations. For instance, in English, it’s considered more correct to use subject pronouns in their nominative forms,“he and I will do it” rather than in their accusative forms, “me and him,” even though many speakers prefer to use the less formal, vernacular form. In minority and heritage locales, the social pressures to use one form over another may be absent, as most people use only the informal vernacular. As a result, people can be observed to use a variety of forms without provoking negative reactions in their interlocutors (Schmidt, 1985; Dorian, 1994; R. Mougeon & Nadasdi, 1998).And as people are interacting with one another either directly or indirectly through intermediaries in the multiplex social networks that sustain a language in a minority language community, everyone within the community is exposed to, understands, and probably uses a variety of variants and contact-induced forms, like calques, semantic extensions, borrowings, and syntactically converged structures.

182

From years to centuries

Another reason that languages continue to change in minority and heritage settings is that sometimes the transmission of languages is horizontal (peer to peer) rather than vertical (older adults to young). In her work among the Aboriginal speakers of Dyirbal in North Queensland, Australia, Schmidt (1985) documented how the younger speakers, between 15–35 years of age, preferred to interact with one another in their own form of the language called Young People’s Dyirbal than to use the traditional form of the language. Tis allowed them to avoid the conservatism of the older generation, who ofen spoke critically of the tendency of young speakers to mix English with Dyirbal—much like older generations across the world may criticize the way teenagers speak. Te diference is that the community of Dyirbal speakers was undergoing a rapid shif from their ancestral language to a variety of English. In linguistic terms, Traditional Dyirbal, spoken by the older adults in the community, had a rich split ergative system where adjectives and third person pronouns were infected with ergative or absolutive case but pronouns for frst and second person were infected with nominative or accusative case. Speakers of Young People’s Dyirbal used one or two forms of the ergative case endings rather than the six forms used by the older adults. Tey also changed the basic word order of transitive sentences from OSV (object-subject-verb) to an English-like SVO, as in 11.1: 11.1 Changes to word order from Traditional Dyirbal (YD) (Schmidt, 1985, p. 102) TD: baun jugumbi-ru margi-gu DEM.ERG woman-ERG thin-ERG YD: baun margi jugumbil-du DEM.ERG thin woman-ERG “Te thin woman sliced the blackbeans.”

Dyirbal (TD) to Young People’s mirrany-Ø blackbean-ABS babin slice

babi-n slice-NFUT mirrany blackbean

Tis innovation to the case and word order properties of Young People’s Dyirbal highlights the kinds of linguistic changes that are found in minority and heritage communities undergoing language shif: convergence to the structures of the majority language may occur, sometimes leading to simplifcation and others to complexifcation of the minority and heritage language. Change in languages undergoing shif can also result from a change in cultural knowledge that has a connection to the way the language is used. For example, in Traditional Dyirbal, nouns were assigned to one of four classes according to their mythological connection with the class where, for instance, words associated with women fell into the same class with “birds” because they are mythologically associated with the souls of dead women in Dyirbal culture. As traditional mythology became less prevalent among the youth, speakers of Young People’s Dyirbal reorganized the complex noun classifcations system into three core noun classes: +animate +male, +animate +female, -animate. Te story of the horizontal networks of Young People’s Dyirbal underscores the personal and social dynamics of language communities that are undergoing shif. Similarly, when a language is located in a new setting, speakers sometimes fnd it difcult to maintain the connections between linguistic forms to social or cultural meanings that are no longer relevant in the new setting. For example, Korean heritage

Minority languages and heritage languages

183

speakers need to master a rich system of honorifcs that determines the linguistic forms to use according to the social status of the addressee and according to the status of the person being talked about. Using the wrong forms can be interpreted as too formal or overly rude. As a result, many Korean heritage speakers avoid speaking in Korean with individuals of a higher social status. Communities undergoing language shif illustrate what happens to a language when the prescriptive rules and constraints are relaxed. Ofen the forms that people use in these communities that appear to deviate from the traditional variety of the language are referred to as “attritional forms” (Dubois & Noetzel, 2005). Tis does not mean that individual language users are forgetting the language that they once knew (as with L1 attrition), rather that there is some change in the transmission of the language from one generation to another. A question that ofen comes up in contact linguistics is whether there are any limits on what can happen to a language in a contact setting undergoing shif. As we have seen in Table 11.1, Tomason and Kaufman (1988) have proposed that“anything goes” in situations of intense contact, when communities are shifing from the minority to the dominant language. However, many linguists have argued that diferent components of the grammar are more prone to change than others. It is widely accepted that the lexicon is directly afected by contact-induced change, but there is disagreement over the degree to which the syntax of a language can be “borrowed.” And even though there can be a great deal of change compressed into a short period of time, as shown by Schmidt (1985) for Dyirbal and by Dorian (1994) for East Sutherland Gaelic in Scotland, many of the changes that occur in contact are similar to those that occur in language change in monolingual environments over time, including the simplifcation and complexifcation of paradigms, the replacement of some forms by analogy with others, and the extending or narrowing of word meanings. Let’s end this chapter with the exploration of innovations to the linguistic structure of a particular heritage language that was in its last stages of language shif.

11.4

CASE STUDY OF FRENCHVILLE, PENNSYLVANIA—U.S.A.

In the 1830s, families from the Franche-Comté region of France began emigrating to the United States to settle in central Pennsylvania. Te village of Frenchville that they settled was linguistically and geographically isolated, permitting the residents to extend their heritage language maintenance past three generations. But, with the closing of the local village schools, the population began to shif toward the dominant language. Recordings were made at two diferent time points, 1988 and 2007, with two speakers at each point (F. Mougeon & Uritescu, 2006; Bullock, 2018). Te 1988 speakers, EV (born 1903) and her brother FP (b. 1909) were the elder cousins of the speakers from 2007, NB (b. 1930) and his brother, KB (b. 1933). Tey all remained in the small community all their lives, and they knew one another and used French together. Of these speakers, EV was the most French dominant, having married a French speaker and continuing to use the language daily all her life. Te others all married people from outside the community and switched dominance to English at home as young adults. All of them were bilingual in French and English but only literate in

184

From years to centuries

English. Te speakers of the younger generation characterized their French as rouillé “rusty” from disuse. Te corpora from both time periods showed very few overt borrowings from English (