Language Evolution - Contact, Competition and Change 9780826493705

Table of contents :
1. Introduction
I: Population Dynamics and Language Evolution
2. Language evolution: The population genetics way
3. Population movements, contacts, competition, selection, and language evolution
4. How population-wide patterns emerge in language evolution: A comparison with highway traffic
5. What do creoles and pidgins tell us about the evolution of language?
6. Race, racialism, and the study of language evolution in America
II: Competition, Selection, And the Development of Creoles
7. Competition and selection in language evolution
8. Transfer and the substrate hypothesis in creolistics
9. Grammaticization and the development of creoles
10. Multilingualism in linguistic history
III: Globalization And Language Vitality
11. Language birth and death
12. Globalization and the myth of killer languages: What's really going on?
13. Myths of globalization: What African demolinguistics reveals
14. A Case Study: The ecology of Gullah's survival
Conclusions
Bibliography

Index

Citation preview

Language Evolution

This page intentionally left blank

Language Evolution Contact, Competition and Change Salikoko S. Mufwene

continuum

Continuum International Publishing Group The Tower Building 80 Maiden Lane 11 York Road Suite 704 London SE1 7NX New York, NY 10038 © Salikoko S. Mufwene 2008 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photqcopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. British Library Cataloging-in-Publication Data A catalog record for this book is available from the British Library. ISBN: 978-08264-9370-5 (hardback) 978-08264-9369-9 (paperback) Library of Congress Cataloging-in-Publication Data Mufwene, Salikoko S. Language evolution : contact, competition and change / Salikoko S. Mufwene. p. cm. ISBN: 978-0-8264-9369 ISBN: 978-0-8264-9370-5 1. Linguistic change. 2. Creole dialects. 3. Language and languages— Variation. 4. Languages in contact. 5. Langugage obsolescence. I. Title P142.M84 2008 417'.7—dc22 2008005678 Typeset by Newgen Imaging Systems Pvt Ltd, Chennai, India Printed and bound in Great Britain by Cromwell Press, Wiltshire

For uniformitarianism in accounts of language evolution; for minority and disfranchised languages, whose evolutionary dynamics have been misinterpreted; and for their speakers, including the Bayansi, whose interactive practices deserve more adequate observation; they can inform us better about globalization and language vitality.

Praise for Language Evolution Professor Edgar W. Schneider Chair of English Linguistics University of Regensburg, Germany Inspired by evolutionary biology, Salikoko Mufwene's spectacularly comprehensive and thought-provoking new book goes for the big picture and illuminates fundamental principles of language evolution and language contact. Showcasing the peculiar (or not so peculiar, after all) evolutionary conditions of Creoles, Mufwene reaches novel and unorthodox insights which build upon concepts such as the importance of ecology, competition and selection, imperfect replication, and family resemblance. He questions and retunes some fundamental notions in linguistics like "system", "transmission" or "acquisition", thus coming considerably closer to an understanding of how language has evolved than earlier linguistic theory. Ingenious imagery like the highway traffic analogy show how patterns have emerged through "invisible hand" evolution, the convergence of communal behavior, and how imperfection, far from being imperfect, generates real-life structures. Principles like the ubiquity of contact and hybridism, the understanding of languages as species and complex adaptive systems, the relationship between mutual accommodation between individuals and emergent communal behavior, or the link between globalization and indigenization invoke a new down-to-earth linguistics in which the interactions of real-life individuals are at the core of far-reaching developments. A must-read for theorists of language change and language contact, and for anybody interested in how language really works.

Professor William S-Y Wang Chinese University of Hong Kong Hong Kong With great force and clarity, Mufwene shows how the structure and history of language may be illuminated by using the core concepts of evolutionary theory, such as social dynamics and competition and selection among variants. Rich in case studies from many parts of the world, this book is essential reading for anyone interested in what language is, including of course anthropologists, historians, psychologists and sociologists.

Contents Acknowledgments List of Maps

xi xv

1Prologue

1

Part One POPULATION DYNAMICS AND LANGUAGE EVOLUTION

2 Language Evolution: The Population Genetics Way 2.1 2.2 2.3 2.4 2.5

Introduction Why this return to a biological approach? Competition, selection, ecology, evolution, and other related notions Ecology and language evolution Conclusions

3 Population 3.1 3.2 3.3 3.4 3.5

Movements and Contacts in Language Evolution Introduction Some myths and facts about the development of Creoles The creole-like evolution of English and the Romance languages The linguistic consequences of the latest Indo-European expansion Conclusions

11 11 12 16 22 26 29 29 33 45 49 57

4 How Population-wide

Patterns Emerge in Language Evolution: A Comparison with Highway Traffic 4.1 Introduction 4.2 Highway traffic and language evolution 4.3 Conclusions

59 59 60 72

VIM

Contents

5 What Do Creoles and Pidgins Tell Us About the Evolution of Language? 74 5.1 5.2 5.3 5.4 5.5

Introduction Why Creoles have not developed from pidgins Why Creoles were not made by children What Creoles do not tell us about the evolution of language in mankind What Creoles tell us about the evolution of language

6 Race, Racialism, 6.1 6.2 6.3 6.4

and the Study of Language Evolution in America

Preliminaries Race and ethnicity in American history Race in linguistics Conclusions

74 75 78 83 88

93 93 96 99 110

Part Two COMPETITION, SELECTION, AND THE DEVELOPMENT OF CREOLES

7Competition and Selection 7.1 7.2 7.3 7.4 7.5 7.6

Preliminaries Hybridism in the normal and natural development of Creoles Variational evolution: development and change by selection More on hybridism in language Constraints on selections Conclusions

8Transfer 8.1 8.2 8.3 8.4

and the Substrate Hypothesis in Creolistics

Introduction The substrate hypothesis Transfer and the substrate hypothesis Conclusions

9Grammaticization in the 9.1 9.2 9.3 9.4 9.5

Development of Creoles

Introduction The Bruyn-Plag position: how justified is it? How "creolization" proceeded in Suriname Grammaticization in Creoles Conclusions

10 Multilingualism, 10.1 10.2 10.3 10.4

in Language Evolution

"Creolization," and Indigenization

Introduction Language "transmission" and "acquisition" The development of Creole and indigenized varieties Language shift, socioeconomic status, and language diversification

115 115 121 124 127 128 131

133 133 136 149 158

160 160 163 171 172 178

180 180 182 184 193

Contents

10.5 Language shift and loss in creole-speaking communities and elsewhere 10.6 Some clarifications 10.7 Conclusions

196 199 200

Part Three GLOBALIZATION AND LANGUAGE VITALITY

11 Language Birth and Death 11.1 11.2 11.3 11.4 11.5

12

Introduction The meanings of BIRTH and DEATH applied to languages Questioning some usual accounts of language birth and death A historical perspective into language birth and death By way of conclusion

205 205 208 209 213 221

Globalization and the Myth of Killer Languages: What's Really Going On?

225

12.1 12.2 12.3 12.4 12.5 12.6

225 226 234 241 245 250

Introduction Globalization and colonization Precursors of modern colonization and globalization Will there be an English-only Europe? The economics of language maintenance vs. shift Conclusions

13 Myths of Globalization: What African Demolinguistics Reveals 13.1 13.2 13.3 13.4 13.5

Introduction What GLOBALIZATION means and what it does not Language vitality in sub-Saharan Africa: a historical perspective Lessons from South Africa Conclusions

14 1 4 A Case Study: The Ecology of Gullah's Survival 14.1 14.2 14.3 14.4

Introduction Fears of Gullah's death: myths and facts Why hasn't Gullah died yet? Conclusions

253 253 254 258 265 269

273 273 273 275 283

Notes

285

Bibliography

313

Author index

337

Subject index

341

3

IX

This page intentionally left blank

Acknowledgments Books often have long, non-unilinear and non-rectilinear evolutionary trajectories, marked by multiple influences that instantiate polyploidy in biologists' language. This observation certainly applies to the present work. It is thanks to critical and constructive feedback from various colleagues and diverse audiences (of linguists and non-linguists) at my lectures "around the world" that I can now address more lucidly, I hope, some of the issues that I raised in The Ecology of Language Evolution (2001) and in Creoles, ecologie sociale, evolution linguistique (2005). I am most grateful to my colleague and friend WILLIAM WIMSATT, with whom I have team-taught, in alternate years, "Biological and Cultural Evolution" and "Language and Globalization" at the University of Chicago since 2000. Our interactions both in class and elsewhere have significantly influenced the way I have written and/or revised the essays included in this book. It is thanks to him that I have become more familiar with some of the issues in population genetics, where I have often sought inspiration in addressing issues in language evolution. He was the first scholar who advised me against cloning my analyses of linguistic phenomena on evolutionary theory in biology. I am also indebted to such great colleagues as MICHEL DEGRAFF and ROBERT CHAUDENSON, on whose regular feedback and support I could count while I was bouncing some of my "heresies" off of their bright minds, especially as we discussed ideas still in germination face to face or on the phone. I cannot overlook numerous other friends and colleagues who have encouraged me over the years, often through invitations to speak at their schools or to participate in important meetings that were germane to my interests. Although I know I may offend some by omission, I would like to thank especially the following, heartily, for giving me the privilege of teaching at their schools: ANNE PAKIR for setting things in motion for my appointment as a visiting professor at the National University of Singapore in autumn 2001; BERT VAUX for having me invited as a visiting professor at Harvard in winter 2002; CLAUDE HAGEGE for providing me the distinction of giving a series of lectures at the College de France in autumn 2003; DANIEL V£RONIQUE for inviting me to teach a course on "Creoles et revolution linguistique" at the Universite de Paris III in autumn 2004; HILDO DO COUTO and MARTA SCHERRE for having me invited to teach a crash course on the development of Creoles and language

XII

Acknowledgments evolution at the ABRALIN meeting in Brasilia in January 2005; JEAN BERNAB£ for inviting me to teach a seminar on "Dynamiques plurilingues et creolisation linguistique" at the Universite des Antilles et de la Guyane at Schoelcher, Martinique (spring 2005); and MICHEL DEGRAFF for successfully nominating me as a faculty member of the Summer Institute of the Linguistic Society of America, at MIT, in the summer of 2005. Besides the occasional courses on "The Ecology of Language Evolution" and "The Development of Creole Vernaculars and Cultures" which I have taught at the University of Chicago since 2000, the lectures I gave at all these places and the discussions I had with various colleagues and students after the lectures have been very edifying for me. They have helped me shape the manner in which I can advocate uniformitarianism about the emergence of Creoles and language evolution in general, especially about the central role of contact, competition, and selection not only in the way languages are restructured but also in matters of language endangerment and loss. I am also very much indebted to the following other colleagues and friends for various arrangements they made to provide me audiences to my positions: UMBERTO ANSALDO and WOLFRAM HINZEN, University of Amsterdam; TEJ BHATIA, University of Syracuse; EYAMBA BOKAMBA, University of Illinois at Urbana-Champaign; RUDOLF BOTHA, University of Stellenbosch; MORGAN CONTEH, Center for African and African-American Studies, Ohio State University; PATRICIA CUKOR-AVILA, University of North Texas; BRIDGET DRINKA, University of Texas at San Antonio; SYLVIE DUBOIS, Louisiana State University at Baton Rouge; JAN TERJE FAARLUND, University of Oslo; EDWARDO FAINGOLD, University of Tulsa; ELAINE J. FRANCIS, Purdue University (formerly at the University of Hong Kong); BERND HEINE, University of Cologne; VINESH HOOKOOMSING, University of Mauritius; TOMETRO HOPKINS, Florida International University; DICK JANNEY, University of Munich; BARBARA JOHNSTONE, Carnegie Mellon University; BRAJ and YAMUNA KACHRU, University of Illinois at Urbana-Champaign; ERNST KOTZE, Nelson Mandela University at Port Elizabeth, South Africa; BERNARD LAKS, Paris X (for the conference on the evolution of language, at the College de France in 2002); RICHARD LEWONTIN, Harvard University (for very useful feedback on The Ecology of Language Evolution); CRAIG MELCHERT, University of North Carolina at Chapel Hill (for invaluable feedback on the final draft of Creoles, ecologie sociale, evolution linguistique); RAJEND MESTHRIE, University of Cape Town; LIOBA MOSHI, University of Georgia; ISIDORE NDAYWEL, at the now defunct Agence Intergouvernementale de la Francophonie (for the inaugural lecture of the meeting "Les langues africaines et Creoles face a leur avenir," in Kinshasa); ROBERT NICOLAI, Universite de Nice (for my appointment as "professeur invite" at the Institut Universitaire de France in spring 2006); MARTIN NOVAK, Center for Advanced Studies, Princeton (for a conference on the evolution of language in 2000); CLYVE PERDUE, CNRS, Paris-St. Denis; MICHAEL PICONE and CATHERINE DAVIS, University of Alabama at Tuscaloosa (for the LAVIS III meeting); JAMES PRITCHETT, African Studies Center, Boston University; EDGAR SCHNEIDER and GUENTER HAUSKA, Regensburg; MONIKA SOKOL, University of Bayreuth; JACKSON SUN, Academia Sinica (for the joint meeting of the Annual Conference of the International Association of

Acknowledgments Chinese Linguistics and the International Symposium on Chinese Languages and Linguistics); JAMES TAI, Chang Kai Chek University; HILDEGARD L. C. TRISTRAM, University of Potsdam (for the Celtic Englishes IV meeting in 2004); ALBERT VALDMAN, Indiana University (for inviting me twice to write articles that would put research on L2 acquisition and on the development of Creoles in perspective, quite a challenge); and WILLIAM S-Y. WANG, Chinese University of Hong Kong, Hong Kong, whose support has become increasingly more significant to me. I very much appreciate the latest invitation he extended me to participate in the "Seminar on Language, Evolution, and the Brain" at the International Institute of Advanced Studies, in Kyoto in spring 2007. Much of what I have accomplished academically over the past few years has of course been enabled by the generous financial support I have received for my research from the Division of the Humanities at the University of Chicago. I am grateful to both Dean IANELLE MUELLER and Dean DANIELLE ALLEN for being particularly attentive to my research needs. I am likewise indebted to WALDO JOHNSON, Center for the Study of Race, Politics, and Culture, University of Chicago, for awarding me seed funding to do exploratory field research on language practice in South Africa in the summer of 2005. The completion of this book has especially benefited from the editorial assistance of ZUZANA TOMKOVA, who has kindly given me feedback on the accessibility of Chapters 2-8 and has integrated all the partial lists of references into one bibliography and adopted a uniform style sheet for the latter. I am even more grateful to her because she has equally shared some of the pressures I experienced in order to meet the publisher's ultimate deadline, after other obligations had forced me to ask for extensions twice before. I am likewise deeply indebted to my copy-editor ION BILLAM and proofreader JUDY NAPPER for meticulously attending to various editorial matters as the book was going into production. Last but certainly not least I am deeply indebted to my partner and colleague CECILE B. VIGOUROUX. I would have found it very difficult to complete this project without her love and intellectual and moral support. Love and support from my daughter TAZIE and the patient company of our cat, SABBY, always lying patiently behind the laptop while I write, have given me lots of strength over the past few years ... and they know why. Many hearty thanks to all of you who have had to cope with my more than occasional mental withdrawals into the universe of my work. May the following publishers and editors also accept my thanks for permission to publish in this book these revisions of my earlier publications or submissions: The Annual Review of Anthropology, for "Language birth and death," in ARA 33.201-22 (2004); Blackwell Publishing, originally published as "Multilingualism in linguistic history: Creolization and indigenization," in the Handbook of Bilingualism, ed. Tej Bhatia and William Richie, 460-88 (2004). Cambridge University Press, through Albert Vaaldman, for "Transfer and the substrate hypothesis in creolistics," in Studies in Second Language Acquisition 12.1-23 (1990).

XI11

XIV

Acknowledgments Duke University Press, for "The ecology of Gullah's survival" American Speech 72.69-83 (1997). Equinox, through Bernard Laks, for "What do Creoles and pidgins tell us about the evolution of language?" in The Origin and Evolution of Languages: Approaches, Models, Paradigms, ed. Bernard Laks, Serge Cleuziou, Jean-Paul Demoule, & Pierre Encreve. (Permission granted while the book was still being edited). Georg Olms Verlag, through Graham Huggan & Stephan Klasen, for "Globalization, and the myth of killer languages," in Perspectives on Endangerment, ed. Graham Huggan & Stephan Klasen, 19-48 (2005). Papia, for "Grammaticization in the development of Creoles," in Papia 16.5-31 (2006). Universitatsverlag Rengensburg, through Glinter Hauska, for "Language evolution: The population genetics way," in Gene, Sprachen, und ihre Evolution, ed. Giinter Hauska, 30-52 (2005). In the case of Chapters 6 and 13,1 sincerely thank the editors for allowing me to proceed with the revisions (as explained in the Prologue) before they had gone to press with the original contributions. Chapter 3 was originally published under the title "Population movements, contacts, competition, selection, and language evolution" in Journal of Language Contact 1.63-91 (2007) but was not copyrighted. The ancestor of Chapter 4 was circulated in Contemporary Linguistics (Working papers of the Department of Linguistics at the University of Chicago) 3, 39-57 (2002), edited by myself, under the title "Analogs anywhere: The flow of highway traffic and language evolution."

List of Maps Map 1 Complementary geographic distribution of Creoles and pidgins

35

Map 2 European colonial expansion since the fifteenth century

37

Map 3 Indo-European dispersal: The first colonial expansion

38

Map 4 A geographically adapted Indo-European cladogram

50

Map 5 The non-rectilinear expansion of the Indo-Europeans

51

This page intentionally left blank

Prologue This book is less about the evolution of language in the phylogenetic history of mankind than it is on language evolution, although it includes a Chapter on the former subject matter. I use the term language evolution in reference to all of the following topics: (1) structural changes that have traditionally interested historical linguists since the nineteenth century; (2) LANGUAGE SPECIATION, which I associate primarily with genetic and comparative linguistics; (3) LANGUAGE BIRTH, which has been central to research on the development of Creoles and pidgins, as well as that of "indigenized" varieties of European languages (chiefly English and French); and (4) LANGUAGE DEATH, an important preoccupation among linguists since the early 1990s. It is only the traditional concern of historical linguistics with form and rule change that does not receive extensive coverage in the following pages. Otherwise, I show how the other processes are all driven by competition and selection in specific ecologies of language practice and how what little has been learned to date about the evolution of Creoles can shed light on that of non-creole language varieties. The reader who has guessed already that I am arguing in the following Chapters for evolutionary uniformitarianism and against "creole exceptionalism" is not mistaken, although I maintain that Creoles are a legitimate group of vernaculars which fit together chiefly more because of the special colonial circumstances of their emergence than because of their structural peculiarities. They are structurally similar on the family resemblance model. Their features vary in accordance with the structural peculiarities of their "lexifiers" and of substrate languages these came in contact with, and depending on various ecological factors that determined the final outcomes of the contacts. As a matter of fact, they provide some of the best arguments for an ecology-based approach to language evolution. This book contains thoroughly revised editions of essays I have written on the evolution of language since the publication of its predecessor The Ecology of Language Evolution, which in many respects was exploratory and scraped only the tip of the iceberg on some issues. Chapters 1 and 2, the latest to have been published, are the only ones that are not substantially rewritten, because I intended them to provide the general background against which the other Chapters can be fully appreciated, as they have very broad orientations. All the others have been revised first to update the statements of my positions and to reduce redundancies in the ways I elaborate some of the topics and issues canvassed in the first two. Only Chapters 4, 6, and 13 are being published here for the first time, despite substantial reorganizations compared to the versions posted earlier at my website. Some of the previously published

D

2

Language Evolution Chapters, such as "Competition and selection in language evolution" (Chapter 7, originally written for biologists interested in language evolution), "Transfer and the substrate hypothesis in creolistics" (Chapter 8), and "How population-wide patterns emerge in language evolution" (Chapter 4), have been rewritten almost from scratch in order to update my takes on some issues, which have sometimes changed, and to integrate them better with the other Chapters. Those who have read some of these essays before will sometimes notice large chunks missing or some other parts added or elaborated, as well as positions that are worded differently. The Chapters have all benefited from my present positions on topics such as the following: (1) How far can the scholarship on naturalistic L2 "acquisition" inform us on the emergence of Creoles? (2) Are the ecologies in which Creoles emerged identical? (3) What can Creoles really tell us about the evolution of language in mankind? (4) Are the "global pressures" that bear on language endangerment and loss non-local? (5) How can we conceptualize the role of the INVISIBLE HAND in language evolution? The latter concept, borrowed from economics, has to do with how individuals pursuing their personal and occasionally conflicting interests wind up producing effects that converge at the population level. No speaker plans to change their idiolect or communal language through the innovations and deviations they produce during their communicative acts. Since they do not all do the same things at the same time, nor do they consult with each other about what to say (or how to express some new ideas), it is interesting to at least take note of the intriguing phenomena of how patterns emerge and spread into norms from these innovations and deviations. As direct evidence of PATTERN EMERGENCE is not easy to adduce from language behavior, given the protracted nature of language evolution (at the communal level), I have included an analogical study of the phenomenon, focusing on automobile traffic. It is not as far-fetched as the reader may fear. Chapter 4 is quite informative on the subject matter. This book has also inherited most of the working assumptions in The Ecology of Language Evolution and my other book Creoles, ecologie sodale, evolution linguistique (2005). I capitalize on the assumption that communal languages are linguistic species in their own right, with their evolutions determined not only by the ecologies in which they are practiced but also by some of their ontogenetic properties that make them different from biological species. Although they are in many respects comparable to viral species, we must remember that they have no geneses that are independent of their speakers, as the latter are not only their hosts but also their creators. Moreover, we must factor the peculiarities of their "transmission" and "acquisition," associated with "imperfect replication" as the default way of "copying" structural information, into our accounts of language evolution. The terms language transmission and language acquisition, always used between scare quotes in this book, are convenient misnomers for, respectively, a state in which current speakers produce utterances (identified as "primary linguistic data") from which learners work out their own ways of speaking/signing, inferring the relevant units and principles on their own. LANGUAGE ACQUISITION thus amounts to a gradual restructuring process, which can also be characterized as a

Prologue selective feature-recombination process, on the model of gene recombination in biology. Much of this is explained in Chapter 2. Imperfect replication, observable in how every idiolect is unique and none replicates any current speaker's, is a consequence of two fundamental facts: (1) Native speakers acquire their vernaculars naturalistically, through their interactions with current speakers and often with peers that are hardly better off than themselves. Everything proceeds piecemeal, in attempts to communicate; and there are no school-like institutions in which they are taught the vocabulary and drilled rules of grammar. This practice leaves a lot of room for inaccurate inferences and therefore deviations which can evolve into divergences. (2) Naturalistic language learners do not proceed like linguists, gathering sufficient bodies of representative data and producing analyses that yield significant generalizations. Pressure to communicate does not give them the luxury of waiting that long and working out a system. Attempts to communicate force them to rely on partial analogies, through well- and ill-formed utterances, improving their productions and interpretations of other speakers' utterances in the process. Rather than wide-ranging grammatical rules (which are good for linguistic analyses), analogies based on limited linguistic materials that are remembered during communication drive the emergence of idiolects, and consequently some of the changes that evolve concurrently. The patterns characterized by the linguist as RULES and/or CONSTRAINTS are EMERGENT PHENOMENA, not inherent systems that are independent of the scholar's theoretical framework of analysis. Like beauty, which lies in the eye of the beholder, they are little more than regularities inferred by the analyst both from the linguistic behavior of individual speakers/ signers and from cross-idiolectal averages or communal norms. I therefore also use "system" between scare quotes, to suggest that linguistic structures may not be as neatly integrated as traditionally claimed, as the conclusions of individual learners' partial analogical learning produce indeterminacy areas and are not identical from one speaker to the other. Also under pressure to communicate successfully, speakers often accommodate each other, through mutual copying of communicative strategies, which keep their idiolects and their languages in constant states of flux. Languages are thus constantly being (re)shaped by their speakers, their hosts and their makers, as the latter communicate, although not in such noticeable ways. Just as we can be concerned by how norms emerge, we can also be interested in the forces that cause communal languages to change. This is the gist of the ACTUATION QUESTION also discussed in Chapter 2. Although much of the reshaping goes unnoticed, amounting to producing variants that are already in currency within the relevant community of speakers or are rejected as abnormal, a few do contribute to structural changes. Much of this generally unpredictable behavior has to do with how the "invisible hand" works within a naturally heterogeneous population of speakers, producing new patterns of convergence among idiolects, driving some variants out, sometimes producing new ones, and generating new norms. As with biological species, the external ecology of a language works on variation within the latter to drive its evolution.

3

4

Language Evolution It is within the complex dynamics of coexistence among structural variants and how they are favored or disfavored by various ecological factors (largely defined by particular population structures) that we must seek the answer to the actuation question and understand how the "invisible hand" works. This book focuses on this state of affairs, arguing that COMPETITION and SELECTION operating on variants associated with the same communicative functions play the same role in structural change as in matters of language vitality. They determine whether or not the different structural variants within a linguistic "system" and/or language varieties in usage within a particular speech community can equally thrive, how some of them are (dis)favored, and under what particular ecological conditions. Like in biology, COMPETITION refers to the unequal ways the variants and coexistent language varieties are weighted by their speakers (the immediate external ecology through which all other ecological pressures are filtered), while SELECTION describes the resolution of the competition, determining which particular variant or language variety is favored over which particular competitors. It is thus essential to understand how competition and selection work in order to interpret language evolution adequately. While the state and process are invoked (jointly) in virtually every Chapter, two have specifically been written to explain some factors that drive selection: Chapters 7 and 8. Unfortunately what we know to date makes it impossible to be exhaustive. We must also bear in mind that while the factors are universal, every ecology determines which ones are significant and which values they assign to competing variants and language varieties. In the case of languages competing with each other, the literature on language endangerment has made of GLOBALIZATION an ecological factor whose significance has been calling for some explicit discussion since the early 1990s. Most of the literature has invoked or alluded to the association of the concept with the alleged Americanization or McDonaldization of the world, claiming that the world is becoming more and more uniform, thus at the expense of linguistic and cultural diversity. This book raises issues with the received doctrine, showing that international globalization - marked, among other things, by greater geographical mobility of people, by faster communication, and by world-wide diffusion of some industrial goods and cultural products - has actually left many parts of the world out of the circuit, creating inequities some of which bear variably on the vitality of languages. Speaker demographics are not informative without much necessary information about the relevant population structures and the local dynamics of language practice. English is far from functioning as a vernacular in many parts of the world and therefore does not threaten the vitality of some of the languages with very small numbers of speakers, although it endangers the imperial position of some other world languages. Language choices are driven by very local considerations in the hie et nunc of communicative acts. Here too one can examine how patterns emerge and which factors produce them that can be associated with the "invisible hand." Chapters 12 and 13 are particularly informative on the subject matter.

Prologue While the title of this book is very general, many of my discussions are inspired by my research on the development of Creoles, especially by my interest in finding out what evolutionary mechanisms have led many linguists to treat these linguistic phenomena as "exceptional," "unnatural," "unusual," "abnormal," i.e., different from other cases of language evolution. They are also inspired by my realization that the little that we now know about the emergence of Creoles invites us to reopen the books on virtually everything we knew about language evolution. So, I decided to raise issues with various aspects of the state of the art, asking particularly the following questions among others: Didn't koineization play a role in the emergence of Old English out of the different vernaculars the Germanic invaders had brought with them from continental Europe? What role did the "Roman legionaries" play in the development of the Romance languages? After all, many of them were not Romans, having been recruited from the "colonies," but they were the critical "transmitters" of Vulgar Latin to the Celts in southwestern Europe. Is their position as "transmitters" of a language nonnative to them not comparable to that of the Dutch colonists in Suriname who adopted English as the lingua franca of the colony they had bought from the English? Didn't the contact conditions of the gradual adoption of Vulgar Latin as the vernacular of Gaul and Iberia in particular and of its transformation into the Romance languages anticipate those under which colonial varieties of the same Romance languages would be transmitted to the enslaved Africans who restructured them into Creoles? Where do differences between these situations lie? Are there good reasons for stipulating that Creole vernaculars are not genetically related to their "lexifiers" and are therefore not Indo-European language varieties? Doesn't the refusal to classify Creoles genetically with their "lexifiers" have to do more with their speakers (typically populations many of whom have been denied the European side of their ancestries) than with any sound linguistic arguments? Generally, are genetic classifications of languages fundamentally not groupings of people rather than of languages? Chapter 3 canvases the above issues and more, preparing the reader especially for Chapters 6 and 8-10, notwithstanding Part 3. I have often turned the tables around in the Chapters of this book, assuming that Creoles raise all sorts of tough questions about their emergence because we actually have much more information about the ecological conditions of these evolutions than we do about the emergence of Old English or Old French, or about the speciation of Indo-European. I submit that similar questions can be asked about these older cases of language evolution. For instance, how homogeneous was proto-Indo-European? Didn't language contact as a concomitant of the well-accepted migrations of the Indo-European populations from their homeland play an important role, if not a critical one, in its speciation? What's the epistemological significance of the monoparental cladograms of genetic linguistics? Isn't the European colonial expansion since the fifteenth century a continuation of the earlier IndoEuropean dispersal? Don't the facts of the past half-millennium of European colonization

5

6

Language Evolution invite us to rethink our assumptions about what caused the changes that resulted in so many different Indo-European languages? Chapter 3 articulates the broad and complex spectrum in which such questions about language evolution and others can be addressed. It also prepares the reader for questions about the evolution of language and linguistic diversity in the phylogeny of mankind, which are discussed more explicitly in Chapter 5. In addition, Creoles remind us that their emergence as new vernaculars was a concomitant of the loss of their creators' heritage languages. If we take some time to put things in perspective, we realize that more or less the same was true of many Europeans in the colonies who now speak different vernaculars from what their immigrant ancestors did. The main difference may lie in the fact that these (descendants of) Europeans shifted rather late to a colonial variety that had already emerged, normalized, and autonomized from its metropolitan ancestor. Their influence on the extant colonial variety does not seem to have been more significant than that exerted by the latest immigrants to the same territories (especially North America and Australia). Overall, loss of pride in one's heritage does not appear to be the (main) reason why Native Americans are now losing their languages, not any more than the claim that forbidding the slaves to speak their ancestral languages is the primary cause of the loss of their languages in the New World and Indian Ocean. These issues are discussed especially in Chapter 11. One of the things linguists should have thought of is (adequate) comparisons of the conditions of language practice among the slaves on the plantations as contact settings with those of Black African cities, where the competition among ethnic vernaculars, ranked low ethnographically, is comparable to that among the languages the slaves had brought from Africa. While slavery was a factor in bringing heterogeneous populations into contact with each other, it is equally important to note that slavery was the product of a particular socioeconomic world order. The various cases reviewed in this book call for a more careful examination of the adaptive pressures that socioeconomic spaces exert on populations that inhabit them, as thoroughly explored in Chapter 13. Nevertheless, I would be remiss to mislead anybody into thinking that there was a general template of plantations as contact settings that can be applied universally. Chapter 10 reveals differences between the population structures of, on the one hand, the plantations of the New World and Indian Ocean, which used slave labor, and, on the other, the plantations of Hawaii, which were exploited after the slave trade had officially ended in most European settlement colonies and relied on contract laborers. The Hawaiian plantation industry operated within a different population structure, whose linguistic consequences were, from an evolutionary perspective, similar to those of the New World and Indian Ocean primarily in producing new language varieties. It impacted differently on language vitality. Even the ecology in which Hawaiian Creole and Pidgin developed is not so similar to those of the other plantation colonies. I show why it is too dangerous to base one's generalizations on superficial similarities.

Prologue We must remember that, in the case of the evolution of Creoles as of other language varieties, it is important to note not only similarities among them but also differences. In the final analysis, every ecology is unique and bears uniquely on the case of language evolution that is specific to it. This book invites the reader to capitalize on similarities across various cases of language evolution to ask, if not address, important questions that have been overlooked for too long. However, it does not at all argue that language evolution proceeded exactly the same way in different parts of the world and in different points in time. It is all a matter of family resemblance among Creoles as between them and other languages. We just don't have to assume that the emergence of Creoles was so different that it cannot teach us much about the presumably more "natural" or "normal" cases of language evolution. Unlike in The Ecology of Language Evolution, the Chapters of the present book are organized into three parts (see Table of Contents), only the first of which needs some explaining here. "Population Dynamics and Language Evolution" is about the big picture that I intended the book to paint about especially the following topics: (1) the working assumptions of my approach to language evolution and the benefits of a macro-ecological approach patterned on population genetics, (2) how we can think of the emergence of patterns in languages as unplanned phenomena, (3) what is the legacy of a racist nineteenth century in the state of the art (although this does not mean that linguists subscribing to this tradition are necessarily racist), and (4) how we can benefit from what we have learned to date on the development of Creoles by asking the right questions and avoiding misguided ones about the evolution of language in mankind. Most of all this is predicated on the fact that languages are species or populations of idiolects and the actuation of language evolution lies in interactions between speakers, hence contacts between idiolects and the ensuing competition among their peculiarities. It is in these interactional dynamics, within specific population structures largely determined by specific economic structures, that the "invisible hand" operates. Part 2, "Competition, Selection, and the Development of Creoles," focuses on particular mechanisms that bring about structural change in Creoles and apply to other languages. It ends with Chapter 10, which highlights both similarities and differences in the evolution of Creoles and between these and indigenized varieties of European languages. Because this Chapter also deals with language endangerment and death, it also smooths the transition to Part 3, "Globalization and Language Vitality," which is titled transparently. This ends with an elaborate discussion of the resilience of Gullah (Chapter 14), against expectations of "imminent death" by transformation or by shift since the late nineteenth century. It provides an opportunity to test some of the hypotheses presented in Chapters 10-13. I hope the skeptical reader will find in the following Chapters reasons to rethink many explanations that have been taken for granted. There is still a whole lot that we simply do not know about language evolution, let alone about the evolution of language in mankind and the emergence of linguistic diversity. Not every hypothesis is plausible and that is what this book intends to show about much of the state of the art.

7

This page intentionally left blank

Part One

Population Dynamics and Language Evolution

This page intentionally left blank

Language Evolution: The Population Genetic Way1 Chapter Outline 2.1 2.2 2.3 2.4 2.5

2.1

Introduction Why this return to a biological approach? Competition, selection, ecology, evolution, and other related notions Ecology and language evolution Conclusions

11 12 16 22 26

Introduction

I focus in this Chapter on only some aspects of language evolution, viz., language speciation and language death, although I also discuss some structural changes, more specifically phonological and syntactic ones. I argue, contrary to the dominant practice in the twentieth century, that a biological approach to evolution is applicable to languages, although I must clarify at the outset that languages should be analogized to species rather than to organisms. Moreover, the application works best when linguistic species are analogized with viral rather than with animal species. I argue eventually that the approach is analogical only to the extent that it is inspired by scholarship on biological evolution. Otherwise, linguistic species can be posited in their own right (Mufwene 200la). They share properties with other species, biological and otherwise, while they also differ from the latter in interesting ways that are specific to their ontogenetic, architectural peculiarities. Thus genetic linguistics can contribute to theories of evolution, adding hypotheses that are specific to, and/or inspired by, languages as species. The invocation of population genetics in the title of this Chapter underscores two working assumptions of mine since Mufwene (2001a), viz., (1) the agency of the evolution of a language lies in the individual communicative acts of its speakers, just like a biological population or species is cumulatively affected by the experiences or activities of the individuals it consists of;2 (2) the communicative activities that produce language evolution are largely determined by the socioeconomic ecologies in which speakers evolve, which is similar to saying that the ecology rolls the dice in evolution. However, I should now explain why

2

12

Language Evolution I adopt an approach that has been unpopular until recently and justify some of the terminology that I now prefer to use in my work, such as evolution, competition, and selection.

2.2 Why this return to a biological approach? By the time Charles Darwin published On the Origin of Species (1859), genetic linguists had already been addressing some of the questions that I address here. According to Maher (Koerner 1983: xxv), August Schleicher felt so vindicated by this book that he responded to it with an elaborate discussion of analogies that he perceived between linguistic and biological evolution (Schleicher 1863,1869). He covers notions such as "variation," "selection," "life of languages" (as new language varieties emerge and some of the current ones die), "gradual development" from the most "primitive" to more advanced and complex stages, "genus"/ "family," and "species" (not in the way I use it in the present discussion). He also remarks that confusion on whether a language variety should be identified as a language or as a dialect is parallel to what Darwin reports about identifying a class of biological organisms as a species or a subspecies. Overall, he concludes, in the spirit of his self-professed uniformitarianism: The rules now, which Darwin lays down with regard to the species of animals and [of] plants, are equally applicable to the organisms of languages, that is to say, as far as the main features are concerned, (cited from Koerner 1983: 30)

Schleicher saw in Darwin's principle of natural selection an explanation for the changes that produce the language speciation patterns represented by the Stammbaum "genetic tree" that he was developing, which apparently inspired the biologists' cladogram. He also assumed, like Darwin (Radick 2002), that different human populations had not reached the same stages of evolution - some being rather primitive and others particularly advanced - and that their languages reflected this putative variation. He thus posited an evolutionary trajectory of languages of the world from the isolating morphosyntactic type to the fusional morphology type. The former type, which he considered as the most primitive, includes, for example, Chinese, Thai, and the Kwa languages of West Africa, whereas the second type, which he claimed to have reached the most advanced level of evolution, includes Western European languages (Schleicher 1863,1869). This position, which was apparently widely accepted by the late nineteenth century even by Charles Darwin (Hull 2002: 13; Radick 2002: 7, 13) - led French philologists such as Adam (1882, 1883) and Vinson (1882, 1888) to treat Creoles and pidgins, which tend to have a predominantly isolating morphosyntax, as languages in their pristine, natural, or primitive state. For these philologists, these recent colonial phenomena were simply linguistic anomalies that reversed the evolutionary trajectory claimed by Schleicher. They were

Language Evolution: The Population Genetics Way 1 3 degenerations of the "refined" structures of the European languages they had evolved from, going back to some protolinguistic stages, so to speak, in human phylogeny. (See DeGraff 2003 for an elaborate discussion.) Although Creoles and pidgins have continued to be associated with non-ordinary, contact-based developments and as special cases that putatively do not fit naturally in the domain of genetic linguistics (however, cf. Mufwene 2003a), the Schleicherian evolutionary model has generally been either discredited or ignored over the past century. Except for Bickerton (1984a, 1990), who, through his language bioprogram hypothesis, has suggested that Creoles and pidgins give us an idea of the human protolanguage, no language typologist has ever suggested any particular evolutionary ranking of structural types.3 It is not clear why the whole biological approach was abandoned in the twentieth century, except that the comparison of languages with organisms was not particularly insightful, for reasons which I articulate below.4 Hoenigswald (1990:11) also notes that a "language is no corporeal something with physical boundaries, or molecular permanence, or genome." According to him, this ontological difference from a biological organism or species would make it difficult to account for language evolution on the biological model. He adds, "Try as we may, we can only think of it as the totality of the very traits that are subject to change" (11). Hull (2002:18) is right on the mark in observing that "most of [the] effort in [genetic] linguistics was expended to working out the histories of various languages. Much less attention was paid to the processes by which languages change." Indeed, the comparative method, on which genetic linguistics has critically relied, is used to determine the extent to which languages share linguistic materials (typically, words, sounds, and morphemes) and can be claimed to be genetically related. However, it cannot be used alone to determine whether the shared forms and structures have been inherited from a common ancestor or borrowed from a common, influential non-ancestral language they all came in contact with. Nor do the correspondences exclude the possibility of mutual influence among the relevant languages if they have been spoken in the same geographical area or in adjacent ones. In other words, used alone, the comparative method cannot distinguish between "genetic relatedness" (based on materials inherited from a common ancestor), "areal diffusion" (due to borrowings from the same language), and "convergence" or Sprachbund (due to mutual influence in a contact area). Areal linguistics (see, e.g., Heine & Kuteva 2005) underscores the importance of factoring the history of population contacts in genetic linguistics. Moreover, as shown in Mufwene (2003a), there has been very little cross-pollination between genetic linguistics and genetic creolistics (which focuses on the development of Creoles, under the conditions of population contacts in plantation settlement colonies). It is not clear why linguists, including some of the most influential creolists, have ignored the fact that genetic creolistics has dealt with ecological details of recent cases of language speciation. Yet, these bear on genetic linguistics, especially in regard to the seemingly gratuitous assumption of asexual parentage

14

Language Evolution (without mating/contact), i.e., the iterated splitting of languages into new varieties without external causation (see below). It is curious how closely genetic linguistics has stuck to Schleicher's Stammbaum, despite some important shortcomings associated with its misinterpretation as an account of language speciation rather than as a representation of the ultimate outcome of the process. For instance, it does not account for "substrate" influence, from the other languages that the relevant language came in contact with. Thus the role of substrate influence in the speciation of Latin into the Romance languages has received only lip service. The role of Celtic languages in the evolution of British English dialects other than Irish (either on the Germanic languages that the Jutes, Angles, and Saxons brought from continental Europe or since the development of Old English) is only now starting to receive some attention (e.g., Tristram 1999,2000; Vennemann 2001; Filppula etal. 2002). The monoparental, Stammbaum approach also makes it impossible to account for hybridization in language evolution as a natural process. I argue below and in Chapter 3 that the approach has been a major obstacle to learning from the development of Creoles lessons that should apply to all cases of language speciation such as the development of the Romance languages or, on a larger scale, the diversification of Indo-European and of Bantu languages. In Mufwene (200la), I submit the following reasons to explain why communal languages are more like species than like organisms, which I claim to be the counterparts of idiolects. Languages are indeed populations of idiolects and are as abstract as species, which are mere categories (Hey 2001) projected from structural or ontogenetic properties shared by classes of individuals or organisms. As the universe of our experience and knowledge is continuous, the boundaries of both species and languages are naturally fuzzy and operationally arbitrary, imposed by particular ideologies or other practical or theoretical considerations, especially within the same genetic family. It is difficult to draw the boundary between adjacent vernacular varieties of, say, Italian and French or Dutch and German, a problem aggravated by the arbitrariness of political boundaries. Unlike organisms, languages qua species are heterogeneous beyond the non-monolithic architecture of idiolects, a property that follows from the fact that idiolects vary among themselves and are alike by the family resemblance principle.5 They are thus like biological populations in which each organism preserves its genotypic and phenotypic individuality, despite the many features it shares with its conspecifics. Languages evolve in non-uniform ways, with some idiolects, sociolects (social varieties), or dialects (regional varieties) being more engaged in some changes than others. They are thus like biological species, whose members are differentially affected by, or react in different ways to, changes in their ecologies. As noted in Section 2.1, changes affecting languages are a cumulation of changes affecting individual speakers' idiolects, which raises the following interesting question: Under what particular conditions do ontogenetic features of individuals or organisms spread to become

Language Evolution: The Population Genetics Way those of (part of) a population? We should note, along with McMahon (1994: 248), that there are many idiosyncratic innovations that are short-lived and do not spread.6 Languages are like populations/species in the ways they die and are born, viz., in a protracted way (see Chapter 11). A language is dead when it has no more speakers left, just like a species is dead when no more specimen is left that can instantiate it.7 Cases of genocides precluded, languages do not typically suffer sudden deaths, because their speakers die or suffer linguistic dysfunction at different times - which entails gradualism. Moreover, they can die in one geographical or social setting but thrive in another, as is evidenced by Old World languages that have died in different places of the New World but continue to be spoken in their homelands or other parts of the New World. Like species, languages are not born in the same way as organisms. They have no moment of conception; nor do they have a gestation or incubation period. They are identified as new ones postfactum, after a particular variety (of idiolects or of organisms) has been identified that is considered (significantly) different from an earlier population. Other characteristics can be invoked, but the present list should suffice to support the position that as communal extrapolations, languages should not be thought of as organisms. This position will enable us to capitalize on the interaction of both variation and ecology to account for evolution. I assume that a communal language exists only as a social construct, suggested by the ability of speakers in a particular setting to communicate successfully when they use similar sequences of spoken or signed gestures. The bottom line is that every individual speaks or signs in a way that is internally systematic. Communication with other individuals triggers mutual accommodations, which make the individual systems converge and thus become similar both in the ways individual sounds or signs are produced and in how they combine together into more complex interpretable units. What is really required in such settings is ability on the part of each participant to understand and be understood by others. Nobody is required to speak or sign in exactly the same way, and sameness hardly happens among idiolects. However, doing things in similar ways must facilitate mutual intelligibility. Languages as communal systems are construed on the basis of similarities in the ways people speak or sign in order to facilitate mutual intelligibility. Perhaps it is no accident that folk definitions of languages do not include notions such as systems (consisting of units and principles) but rather boil down to "the way a particular group of people speaks." Thus German means the way the people called Germans speak and (Ki)Swahili means the way the people identified as (Wa)Swahili speak. The fact that world languages such as English and French are spoken in different parts of our planet now does not contradict my observation, because these languages can also be defined etymologically as, respectively, "(originally) language of the Angles or English people" and "(originally) language of the Franks or French people."

1b

16

Language Evolution All these considerations explain why it is necessary to focus on the agency of individuals and dynamics of their interactions. Accommodations made during individual communicative acts not only bear on the development of their communal properties but also drive evolution within their populations. Interesting questions include the following: Does a population interact collectively with its ecology? Or, is what is associated with a population rather a cumulation of the effects of individual interactions of its members with the relevant ecology? What kind of equation can be posited to account for such a cumulation of effects? How artificial is it to assume that the ecology acts on a population? If the ecology operates on individuals, what precisely is the ecology of the individual? If it is normal to assume an ecology for a population too, is the ecology of a population the same as the ecology of an individual? What's the significance of these epistemological questions? I must first explain some concepts that are fundamental in this intellectual exercise.

2.3 Competition, selection, ecology, evolution, and other related notions I should clarify again that I discuss "language evolution" in this Chapter (as in most of the others in this book), not "the evolution of language." As used in the latter phrase, language refers to the higher abstraction of what various languages share, with the ability of speakers/ signers to encode/store and communicate information through a spoken or signed "system," and with how this capacity evolved in mankind. Language evolution can apply to specific languages in a way similar to language change, to which it is closely related semantically. Since the latter term is more established in historical linguistics, the following question arises: Why don't I stick to it? As explained in Mufwene (200la, Chapter 1), the term evolution covers more than the traditional term change. In addition to traditional concerns with structural and pragmatic changes, it also covers language speciation and language birth and death, processes to which the term change has not applied in linguistics. Also, although language speciation has been a concern of genetic linguistics, a branch of historical linguistics, seldom has the topic been related to those of language birth and death, which are so germane to it. Thus, the development of Creoles and pidgins has typically been treated as anomalous or unusual (see, e.g., Hock & Joseph 1996), if not as an aberration, despite the fact that the topic is, at least from a phylogenesis perspective, closely related to historical dialectology. The traditional practice is made more bizarre by invocations of substrate influence in both genetic creolistics and genetic linguistics, especially in Romance linguistics, where the term substratum originated. It is as necessary to invoke substrate influence from the Celtic languages to account for the speciation of Latin into the Romance languages as it is to invoke African substrate influence to account for the evolution of French and Portuguese into

Language Evolution: The Population Genetics Way various Creoles, the role of other ecological factors discussed in Chaudenson (2001, 2003) and Mufwene (200la) notwithstanding. Both cases are clear instantiations of Pyrrhic victory - where the prevailing language is so clearly affected by the displaced ones - which makes the distinction between Creole and non-creole languages more sociological than linguistic, as made more obvious in Chapter 3. While the term change has been used in cases of substrate influence, it has not been applied to those of language speciation as consequences of divergent structural changes. The term evolution offers the advantage both of its applicability to this case and of aligning research on language evolution with that of species evolution, which covers a similar range of developments.8 At least in the way I approach the subject matter, looking for causation in the communicative activities of individual speakers (see below), the term evolution brings along a refreshing perspective, viz., the alternative of interpreting the relevant evolutionary processes as adaptive responses to changing communicative ecologies, both external (including other speakers) and internal. Contact plays an important role in my approach to language evolution, the most significant part being contact between individuals rather than between populations (Mufwene 200la, Chapter 2). As I also explain in Chapter 3, the coexistence of two populations in the same geographical area is not a sufficient condition for language contact. They must interact with each other. At the level of linguistic communication, this is made possible by interacting individuals, who can spread features from the other language among monolingual members of their respective languages. Moreover, as pointed out by Weinreich (1953), the real locus of language contact is the minds of individual speakers. Given the piecemeal way in which language acquisition proceeds (see below), speakers' minds are the arenas where selection resolves the competition that takes place among features received from various speakers and, in the case of multilingual speakers, also between features of the target "system" and linguistic "systems" that they have already worked out. The phenomenon known in contact linguistics as interference represents those cases when features of, say, a speaker's mother tongue are substituted for features of the target language - for instance when the English th in think is pronounced as s or/(thus sink of fink) by a nonnative speaker. From the point of view of the development of linguistic or communicative competence, the total amount of linguistic knowledge speakers have, including the variants that compete for the same structural or communicative functions, is comparable to a "gene pool" in biology, both at the individual and at the population levels.9 In the case of language, the term feature pool (Mufwene 200la) seems to be an apt analog. All of these observations make more sense once we re-examine briefly the concepts of "language acquisition" and "transmission," as I do below. Language transmission is a convenient misnomer in linguistics, because no speaker ever transmits a ready-made system to any learner, although caretakers simplify the learning task for the child by seriously limiting the number of topics for communication and the complexity of utterances. Both the range of topics and structural complexity (in terms of

1 7

1o

Language Evolution utterance length and embedding) increase as the child matures in age and in cognitive capacity, including processing skills. Otherwise, the burden of developing communicative competence rests on the learner, who, undoubtedly aided by Universal Grammar, must discover the units and principles of the targeted language or dialect and must (re)construct an approximation of its "system."10 "Language acquisition" is thus a (re)construction process comparable to gene recombination in biology (Mufwene 2001a, Chapter 2). It can be called a restructuring process, though the term has mistakenly been associated with structural divergences that make Creoles different from the European languages they have developed from. The main differences from gene recombination in biology lie in the fact that, in the best-known cases, gene recombination takes place at the conception of an organism, while its genotype is being formed, with no agency on the part of the gene carrier. In those species of organisms such as viruses whose genotypes can change during their lifetime, one can argue that gene recombination takes place more than once. Languages are species whose phenotypes - the linguistic features on which typological classifications are based - correspond to no genotypes.11 This peculiarity, which may appear strange to a biologist, is a consequence of the piecemeal way in which speakers develop competence in a language. From the point of view of population genetics, interesting questions arise regarding, for instance, why idiolects are not more different from each other than they are; why they wind up with similar, though not identical, combinations of features; and how some, but not all, of the individual changes they undergo can amount to communal changes. These questions, also discussed by Paul (1891), are similar to those regarding how, with genotypes that are different from each other's, organisms of a biological species do not wind up being more different internally than they actually are and why their phenotypes are so similar. Needless to say biology will not always inspire linguists who address such issues, nor will insights have to flow only one way in such comparisons between biological and language evolutions. Linguistics can contribute research avenues for theories of evolution in general, if ontogenetic properties of languages and their peculiar mode of "transmission" can provide alternative and independent ways of addressing some questions. One way of accounting for the piecemeal way in which a language is "acquired" is the fact that its features are copied (typically with modification) in ways that are closer to horizontal transmission in biology, as in epidemiology, than to vertical, generational transmission.12 This is apparently the consequence of the fact that languages have no genotypes and their so-called "transmitters" play no more important role than that of making available the primary linguistic data from which the learner can work out their idiolects. This is done incrementally as more and more communicative needs arise and the learner/speaker figures out how to express their ideas and feelings, thereby expanding and/or refining their structural strategies. Also, although interactions between caretakers and children suggest that language is "vertically transmitted," there is much more peer influence from the time the child interacts

Language Evolution: The Population Genetics Way 19 with other children, despite the fact that every child takes something from the older people they interact with. The importance of social interaction in language development should actually question the wisdom of capitalizing on age differences, rather than on experience, in the way language is learned. More remains to be thought out on this topic.13 Another important feature of language development is also polyploidy, a condition that makes it possible for the learner to incorporate in their emergent idiolects features originating in diverse idiolectal sources, including variants. These include alternative terms for the same concept (e.g., pail vs. bucket), alternative pronunciations for the same word ([fi'naens] vs. ['faynaens] finance), or alternative grammatical strategies (e.g., there's/there are/thereyre many things to do vs. it's many things to do).14 Identifying and articulating the principles which regulate learners' selections in the development of their own idiolects (i.e., determination of preferred ways of saying things) is a challenge that linguists must face. Individual learners make their selections in ways that do not seem so similar to the selections that take place during the formation of a genotype, for instance, what particular genes the recombination process makes dominant (determining the carrier's phenotypes) or recessive, and under what particular conditions. To begin with, no agency of the gene inheritor is involved in the biology case. If ecology plays a role at the level of the formation of genotype, it is not in the same obvious way it seems to do during the formation of idiolects. In other words, it is not clear what factors or what particular selection algorithm a biologist would invoke to account for why particular genes become dominant or recessive in the genotype of a particular organism or for why a particular individual winds up with a particular combination of phenotypes.15 Linguists can invoke factors such as the statistical frequency of a feature, semantic transparency, regularity, salience, and social status of the model speakers. For instance, a learner of English who is focused on regularities could easily and incorrectly substitute failed for fell as the past tense of the verb fall. However, although this kind of error often occurs, especially in child language, it usually does not crystallize into an idiolectal idiosyncrasy. The reason is simply that it cannot compete with the high statistical frequency of fell, which makes it less competitive. Interestingly, this is contrary to those cases where some speakers produce dove instead of dived as the past tense of dive, although the basic principle at work is the same. They do this by analogy to drove and rose as the past tenses of, respectively, drive and rise, because the latter also have high statistical frequencies. Likewise, because the comparative construction with more is more common, many speakers will say more fit instead of fitter, which should be expected to be more typical of monosyllabic adjectives such as short > shorter. On the other hand, one also hears funner instead of more fun, simply because it is more consistent with fitter and shorter, despite the fact that its status as an adjective is less clearcut. There are thus various factors that affect what particular linguistic habits (features) a speaker internalizes in his/her idiolect. This is how selection works, especially at the population level, often not resolving the competition

20

Language Evolution in exclusive terms, only in terms of dominance and differentially from one speaker to another. Before proceeding, it may not be redundant to clarify that, like in biology, the terms competition and selection entail no agency on the part of genes or features. The term competition refers to the condition of inequality that obtains among variants in a feature pool, with some factors of their internal or external ecologies (dis)favoring some of them for dominance. Contrary to what is suggested by Dawkins' (1976) memetics, it is speakers who make the selections. They are in fact the ones who impose the inequality condition, based on factors, sometimes non-structural, that guide their preferences. However, the way in which linguistic competence develops also suggests that in the linguistic species selection applies at the level of features (units or combinatoric principles). Although some ethnographic considerations suggest that selection also applies at the level of languages, when speakers target primarily features of a particular language over those of others, what we know about language mixing and the development of Creoles suggests otherwise. Languages are selected indirectly through the fact that their features (sounds, words, combinatorial rules, and particular ways of packaging meanings) wind up constituting the majority of those selected from the combined feature pool of the language varieties in contact. Although clearly favored, the indirectly selected language (variety) also bears the influence of (some of) the disfavored varieties and is therefore modified into a new variety. This is what I identified above as "Pyrrhic victory." It appears to explain how, under the influence of the Celtic languages over which it prevailed in especially southwestern Europe, Latin evolved into the Romance languages. The same evolution by selection, under the influence of substrate languages, is also true of Creoles in former European plantation settlement colonies of the New World and the Indian Ocean. As the relevant socioeconomic histories of the relevant plantation settlement colonies show, a European vernacular was typically selected over African and other European vernaculars, but the emergent colonial variety also bears the influence of these other languages. Thus, in partial support of Dawkins' (1976) hypothesis that it is genes rather than species which are involved in the selection process, it is indeed still features16 which are the units of selection, although being associated with a particular language becomes one of the factors that favor those features.17 In contact settings that produced Creoles, only one target language, often misnamed the superstratum or lexifier, lends most of its vocabulary and grammar to the emergent vernacular.18 Since the target has typically been structurally heterogeneous, availing competing variants, the selection of particular options, rather than others, into the emergent Creoles shows again that selection applies to smaller units and combinatoric principles that collectively make up a language; it applies only implicitly or indirectly to languages themselves. Selection is constrained by the specific external and internal ecologies of linguistic interaction. The external, socioeconomic ecology imposes a particular ranking of variants

Language Evolution: The Population Genetics Way 21 (dialects and structural features) to which the selections made by speakers are sensitive. It is, of course, not enough to know whether a feature is standard or nonstandard. Another important factor is whether a particular feature will promote or discourage the acceptance of a learner/speaker by the particular socioeconomic group that he/she wishes to fit in. The literature on covert prestige shows that in informal settings, where vernacular varieties are spoken, nonstandard features are normally favored over their standard counterparts. This behavior accounts very well for the resilience of nonstandard vernacular features, as stigmatized as they are by the elites of various populations. The personality of the speaker/learner is also another important factor, which partly determines which group he/she wants to be associated with, if given a choice. Similarity to epidemiology is not difficult to establish here, as social practices bear on how viruses spread in a population, although different species are subject to different constraints specific to them in the selection of their features. The internal ecology, which is no less important, consists of all the other variants that a particular form, structure, or rule coexists with. The factors that contrast them with each other determine, relative to the linguistic background of the learner (part of the external ecology), which of them is likely to become dominant. For instance, in the case of Creoles, the periphrastic comparative (with more in English) has been favored not only because the variant exists in the European language but also, and quite significantly, because the substrate languages typically have only a similar periphrastic option. This partial congruence of structural features favored the generalization of the periphrastic comparison in the new, plantation colony variety of the European vernacular. It is also important to remember in this case that the population of linguistic variants consists of those attested in the seventeenth- and eighteenth-century nonstandard varieties of European colonial languages, excluding some of the variants that we may be more familiar with today. These considerations prepare us not to be too easily impressed by the fact that structures of Creoles are quite different from those of the standard varieties of the same languages today. Thus, we should not claim uncritically and too hastily that the origins of their present nonstandard peculiarities must be African. Rather, they should prompt us to compare Creoles with nonstandard varieties of the same languages, which should give us a closer sense of the extent to which they have diverged. Then we must also recall that the target for those who made the Creoles consisted of several nonstandard varieties competing with each other, which should make it normal to notice that features of Creoles have diverse "super- and substrate" origins and therefore can differ maximally from non-creole varieties that did not develop under similar ecological conditions.19 Even the European target itself was structurally heterogeneous, including not only features of diverse dialects but also xenolectal features from especially the European indentured servants who did not speak it natively. In the case of non-creole colonial vernaculars (such as White American English varieties), whose beginnings have been associated with koineization20 (Montgomery 1995, Mufwene 2003b), the situation is thus comparable

22

Language Evolution to one where members of different subspecies can mate with each other, across the subspecies boundaries. In biological terms, it is like having a colony where different subspecies of dogs can interbreed across their normal social boundaries and produce new, hybrid canine subspecies. In the case of Creoles, the situation would amount to one where the dogs can also interbreed with members of related species such as foxes, jackals, and wolves, always bearing in mind that the modes of "transmission" in animal biology and in language are different, especially the fact that features of one particular language are likely to be heavily favored by the particular socioeconomic setting of the contact.21 However indirect the process maybe, community-wide targeting or selection of a particular language variety over (an)other alternative(s) can have particular ethnographic consequences. As will be shown in Part 3, languages that are less often selected for communication are endangered and may eventually die. There are various reasons why speakers in a multilingual or multilectal community would prefer a particular language variety or would often choose not to speak another one. The reasons are typically practical ones, associated with the hie et nunc conditions of interaction, and need not be discussed here yet. They can also vary from one speaker to another in the same population. In the vast majority of cases around the world, the decisions are individual and taken at different times (however repeatedly) and in different tokens or kinds of settings, though members of the population do influence each other. Similarities between language and biology arise here too in how competition and selection operating on features/genes and on organisms can lead to evolution at the species level. Future research should shed more light on details of similarities, how far they go and where they end. Note, for instance, that when such selections eventually lead cumulatively to situations where members of the population can no longer speak their language or dialect (fluently), loss of competence does not proceed uniformly within the overall population, confirming Mayr's (2001: 86) observation that "Since all changes take place in populations of genetically unique individuals evolution is by necessity a gradual and continuous process."22

2A

Ecology and language evolution

An important dimension that for too long has received little attention in historical linguistics is the motivation for change, i.e., the (chain of) factors that trigger the process. Weinreich etal. (1968: 102) formulate it under the name actuation problem: What factors can account for the actuation of change? Why do changes in a structural feature take place in a particular language at a given time, but not in other languages with the same feature, or in the same language at other times?

Language Evolution: The Population Genetics Way 2 3 McMahon (1994: 248) articulates it in the following fairly complementary way: the real actuation question is why some of these innovations [by individual speakers] die out and others catch on, spreading through the community, or why certain instances of variation become changes while others don't.23

Having focused on the subject matter for over 20 years, Labov (2001: 466) reformulates the "problem" as follows: Why here and now? The beginnings of change [applied to a population] are as mysterious as ever. Why not here and not now? Endings are equally difficult to understand. The obverse of the actuation problem is continuation. If change has already begun and is not conning to an end, it is continuing. What was the force that was missing a hundred years ago, that fuels the engine of the Northern Cities Shift24 today and keeps it moving?

He thus widens the scope of the question, showing in the rest of his book that a variety of social ecological factors account especially for the ways changes are driven in particular language communities, often affecting only segments of the relevant populations but not others. Recasting the actuation question the population genetics way, I submit that the interaction of the external ecology of a language with its internal ecology should shed light not only on causes of linguistic changes but also on how these phenomena spread. Here too, there are similarities between, more specifically, linguistic and viral species in the ways changes occur, as they are affected by the social practices of their hosts. Critical to this approach is recognizing that changes often consist of only modifications of patterns of variation within a language, not necessarily of the introduction of new variants (McMahon 1993: 248) or loss of some others. They can consist, at the population level, of shifts in the statistical frequencies of the variants, with some becoming dominant that used to be recessive. My basic assumption is that variation within a population is likely to remain stable unless something happens in its external ecology that disturbs the "balance of power" between the competing variants. In the case of linguistic and viral species, factors such as migration to a new habitat, contact with another population, and other changes in the composition of the population bear on actuation. Migration also implies split of the proto-population, which, if proceeding randomly, can affect variation both in the migrating, colonial sub-population and in the metropolitan one (the individuals left behind in the motherland). Depending on how it was (self-)selected, the colonial sub-population may carry less than the totality of variants that obtained in the proto-population. The "balance of power" between the variants can shift in the colony, so that (some of the) features that continue to be marginal in the metropolitan

24

Language Evolution population can become dominant in the colony. In the case of language, such a shift in the constituency of variants can eventually produce a new language variety, or it can simply cause the colonial and metropolitan sub-populations to evolve in divergent directions. Although this is not the full explanation, it contributes to accounting for why, for instance, English on the Falkland Islands is not the exact match of any dialect in England, despite the fact the Britons settled permanently in this archipelago in the late nineteenth century, much later than the English colonized North America, in the early seventeenth century. Other reasons for the linguistic divergence emerge from the discussion that follows. The above observations should remind us that even if the English had not come in contact with other European populations in North America and in Australia, their language would have speciated anyway. Shifts in the constituency of the variants or in their statistical frequencies would have changed the "balance of power" and (dis)favored different variants. Traditional accounts that have simply invoked geographical separation by rivers, mountains, and oceans to account for language speciation have barely addressed the question, especially if they assume a common ancestor for all the new varieties.25 Likewise, changes in the patterns and dynamics of social interaction - which migrations help produce - affect the vitality of a particular (strain of a) virus in two or more allopatric populations that have the same ancestor. Socioeconomic speciation in language and epidemiology is in some ways analogous to geographic diversification. Language contact can also disrupt the extant pattern of variation, provided the host and immigrant populations interact regularly with each other, at least through some of their members. We are quite aware of the most drastic effects of recent European migrations to settlement colonies of the Americas and Australia on indigenous languages. Owing to specific population structures26 - for instance who holds political and economic power, who works for whom, and who makes concessions to whom - significant proportions of indigenous languages have been driven to extinction. An important reason lies in the changing socioeconomic ecologies which have prompted the indigenous populations to assimilate the European populations' practices, including the adoption of the colonists' languages as vernaculars. In other cases where European vernaculars were appropriated by subordinated non-indigenous populations that have not been integrated by the Europeans, typically Africans in plantation settlement colonies of the New World and Indian Ocean, new language varieties have emerged that have been substantially influenced by African languages, which were also driven to extinction.27 These are the new, colonial vernaculars that have been disfranchised from the Indo-European language family by the name Creoles. One somewhat oversimplified but still plausible account of such developments is that the other languages that the European languages came in contact with affected the patterns of variation within the European languages, owing to various cases of (partial) structural congruence between them and the other features that the non-European languages sometimes introduced into the systems. (See Mufwene 2001a for a more accurate and nuanced account.)

Language Evolution: The Population Genetics Way 2 5 The same processes operated in less drastic ways to produce new colonial varieties of European languages in settings where only these came in contact with each other under varying population structure conditions. White American English varieties are generally byproducts of contacts among various metropolitan English dialects and other European languages.28 Mutatis mutandis, the same can be said of colonial varieties of other European languages. The whole evolution is similar to bringing populations infected by various strains of a particular virus in contact with each other, creating ecologies in which new strains can emerge, with each ecology producing its own dominant kind. In both cases, population structure, in terms of who interacts with whom, is an important ecological factor, to which I return below. Both contact and migration influence the ways competition and selection affect variants differentially in the colonial and the metropolitan (sub-)populations. Eventually they produce new and divergent language varieties. Population structure - interpreted now as who lives where and under what conditions he/she can interact with members of another group or population - has a great deal to do with how evolution starts and proceeds. Changes in the population structure also affect whether or not a particular change ends. In the way that Labov (2001) discusses actuation, changes occur even when the structure of a population does not seem to have changed. Assuming several changes to have started from below, in the way societies are economically or ethnically stratified, he considers social mobility and nonconformity as critical factors in the causation of changes. Social mobility is actually another way the structure of a population changes, when individuals leave one socioeconomic class for another and thus join new networks of interaction. "Non-conformity," which means refusal to abide by all the norms of the new socioeconomic class, suggests the expectation on the part of the current members that one's language remains the same all the time, thus that the current pattern of variation is there to stay.29 Other factors, of course, determine whether the non-conformists will have followers and which of their "innovations" will spread and be recognized communally as change. In population genetics, they are more or less like individuals who introduce new viruses in a population and who can spread them largely thanks to how the host population interacts with them. All the above observations make more sense once one factors in the following: (1) populations are constructed from individuals behaving singly most of the time, (2) they are not homogeneous, and (3) their members typically interact in dyads or triads within small, overlapping networks. At least from a geographical perspective, they are more like what ecologists call metapopulations, i.e., habitat patches interconnected by dispersing individuals (Hanski 1996). They are basically convenient groupings of individuals who interact with each other, within and across accepted social boundaries (racial, ethnic, gender, age, professional, economic class, etc.), as individuals, not as team members.30 An important question that will be discussed in Chapter 4 is: How do communal patterns emerge from individual speakers' behaviors?31 Changes typically apply in non-uniform

26

Language Evolution fashions, affecting idiolects differentially and applying only to some sub-groupings of the larger population. Thus, Labov (2001) reports that, overall, African Americans have not participated in the Northern Cities Vowel Shift in the United States. In this particular case, population structure provides a useful ecological explanation, bringing to bear the fact that, as observed by Labov (see also Wolfram 2000), Americans are sensitive to race distinctions. In fact, most Americans tend to socialize along race/ethnic lines and maintain a strong sense of ethnic identity. While most Whites would not like to be associated with their stereotypes of African American linguistic peculiarities, most African Americans, at least those associated with African American English, would not like the opposite evolution either. According to Wolfram (2000), young rural African Americans prefer to identify with urban African American Vernacular English (AAVE) (see also Cukor-Avila & Bailey 1996), considering the conservative variety of their parents too similar to White English.32 White and African American speakers would thus prefer to follow evolutionary trends within their respective ethnolects, ignoring or only following with interest those occurring in the other(s). The analogy with virology is clear, as viruses tend to be transmitted through interacting hosts. We are thus also in a position where we can justifiably speculate on divergence phenomena. Could a population's determination not to be assimilated by, or confused with, another population account for cases where two varieties such as European American and African American English varieties evolve in divergent directions, even in the Southern states where over two centuries of intimate cohabitation predated the last century of segregated life? Isn't this what the literature on the divergence of White and African American English varieties is ultimately telling us, thus indirectly underscoring the significance of population structure as an ecological factor in evolution?

2.5

Conclusions

It is evident that studies of language evolution stand to gain a great deal by adopting a population genetics approach, one that acknowledges that populations are groupings of individuals, that the agency of many diachronic processes lies in the activities of individual members, that such processes need not be uniform across populations, and that the activities are largely determined by the ecologies in which the individual agents evolve. I submit that languages or dialects are species of idiolects, based on family resemblance. They share with viral species several properties that are relevant to understanding evolution; therefore the biological evolution model that linguists should consider for inspiration appears to lie in virology and/or epidemiology but not in animal biology. Languages are essentially like viruses in that they are both parasitic species whose lives depend on the activities of their hosts, thus also on the latter's patterns of social interaction. However, it is equally important not to overlook some important differences between them, as these can account for their respective evolutionary peculiarities. Unlike viruses, idiolects

Language Evolution: The Population Genetics Way 27 as linguistic organisms do not have anything corresponding to a biological genotype, simply because, if one can at all claim that they are begotten through the communicative practices of their speakers, their structures develop piecemeal. They develop incrementally as individuals learn to communicate and grow to produce increasingly complex utterances.33 On the other hand, a virus starts its life with a fully structured genotype (by gene recombination at conception) and has an incubation stage. This peculiarity is worth bearing in mind, although, like an idiolect (even in its mature stage), a virus can change its ontogenetic structure several times over, in a Lamarckian style, and it can be modified when it spreads to other hosts. On the other hand, a virus may have started its life outside its current host, whereas an idiolect may not - only its features can. An idiolect's host is also its maker. Unlike the genes of a biological organism, the features of an idiolect cannot be transmitted - literally, in the form of a copy - from one host to another. They are typically copied with modification, under competition with similar inputs, by the learners, and they are recombined into new "systems" with some inter-idiolectal variation. Thus, no two speakers produce the same sound in physically and acoustically identical ways and no two speakers have exactly the same range of denotative and connotative meanings, as well as pragmatic constraints, associated with the same words. Cases of misunderstandings between speakers of the same dialect also suggest that no two idiolectal "systems" are identical, though speakers more familiar with each other have an easier time understanding each other. These observations follow from the fact that every speaker differs physiologically from other speakers and is thus equipped with a different configuration of speech organs.34 Moreover, from a social ecological perspective, no two speakers have had identical experiences of social interaction, hence of being exposed to identical primary linguistic data, which have influenced their language "acquisition" processes. Still, despite these differences, there are similarities striking enough to justify the approach advocated in this Chapter. Polyploidy suggests that gene selection and recombination apply to viruses in more or less the same ways feature selection and recombination apply to idiolects, gradually in the latter case, but subject to ecological constraints in both cases. Idiolects are more likely to be influenced by those of speakers that their hosts/makers have interacted the most frequently with. Speakers accommodate each other, minimizing chances of being misunderstood. Likewise, viruses in a population are more likely to share a lot of genetic materials when their hosts socialize with each other than when they do not. However, even under such conditions, it is still informative to understand under what more specific conditions a particular gene or feature becomes dominant or recessive, while the host has actually received, or been exposed to, all of them. How does selection work and what particular factors constrain it? As shown more specifically in Chapters 7 and 8, linguists can invoke markedness, or some optimality constraints, to explain why some variants are (dis)favored at the level of both idiolects and language varieties. While a wide range of

28

Language Evolution ecological factors can account for selection of particular genes for dominance within a biological population, it is not clear to me what factors can be invoked to account for a similar process at the level of (the conception of) organisms. Comparisons such as undertaken in this essay suggest that linguistics and biology can very well inspire each other in addressing evolutionary issues.35 Questions regarding the specific ways in which selection proceeds at different levels are, I hope, among those that future interdisciplinary research in biological and language evolution will explore.

Population Movements and Contacts in Language Evolution1 Chapter Outline 3.1 3.2 3.3 3.4 3.5

3.1

Introduction Some myths and facts about the development of Creoles The creole-like evolution of English and the Romance languages The linguistic consequences of the latest Indo-European expansion Conclusions

29 33 45 49 57

Introduction

Genetic linguistics, concerned especially with how language varieties are genetically related and how language speciation has occurred, is to date marked by a disjoint view of language diversification: speciation due to natural evolution versus speciation due apparently to the disruption of the normal process of language "transmission" by contact with other languages. As explained in Mufwene (1997b, 200la), Creoles, as outcomes of language contact and the epitome of "mixed languages," have been treated as children out of wedlock. Classic textbooks in historical linguistics such as Hock & Joseph (1996), which include a discussion of these vernaculars, caution the reader that they are genetically "exceptional," if not unnatural, because they have not emerged in the "usual" and/or "normal" way.2 A corollary of the above position is the distinction between internally and externally motivated language changes perpetuated especially by Thomason & Kaufman (1988) and Thomason (2001,2002). These linguists suggest that the two kinds of changes are different in nature and that those induced by language contact have not contributed to language speciation in the same way as those owing (principally) to language-internal mechanisms. Putatively, genetic classifications should be based only on correspondences based on internally motivated change. Thus, they conclude that Creoles cannot be classed genetically, because, in their case, contact was so extensive that the comparative method cannot be applied in informative ways. Heine & Kuteva (2005) are rather exceptional in arguing that Europe is a linguistic area, one in which language contact accounts for the wide diffusion of several grammatical features.

i

30

Language Evolution Thus, the correspondences that obtain among several Indo-European languages are not necessarily due to common inheritance. Some of them reflect borrowings by some from others, owing to layers of contacts among them, as I show below. It follows from Heine & Kuteva's work that one must know a substantial amount of the histories of particular populations before determining whether their shared linguistic materials are due to diffusion through contact or to shared ancestry. Thus, the comparative method can in principle not be used unilaterally, without taking into account the histories of the relevant populations, in order to determine whether or not two language varieties are genetically related (see, e.g., Aikhenwald & Dixon 2001; Lightfoot 2002). The method only reveals the extent to which languages share structural materials (sounds, morphemes, or syntactic structures) but not how these came to be shared. See Tremblay (2005) for a similar conclusion. Adducing arguments from external history, I argue below that the above problem stems from a misinterpretation of the genetic Stammbaum as an account of language diversification rather than as a representation of the outcome of the speciation process. The problem also comes from a basic assumption in genetic linguistics, since the design of the Stammbaum by August Schleicher in the nineteenth century, that normal language diversification proceeds monoparentally (a process identified by biologists as asexual transmission). Accordingly, language contact has been invoked especially to account for "irregularities" in the correspondences revealed by the comparative method but not as the trigger of the "regular" changes that lead to speciation. Thus, Creoles have typically been disfranchised by most linguists as separate languages that are genetically unrelated to the European languages from which they have evolved (see also Thomason 2002), despite the contrary sentiments of their speakers, as in the case of Gullah in the United States (Mufwene 1988b) or Tanaka English (Miihlhausler 1985). To my knowledge, however, the comparative method has never been applied in such cases (Mufwene 2003a). Faine (1937), Hall (1950), Goodman (1964), Posner (1985), and Trask (1996) should not be dismissed so casually when they claim that French Creoles and Papiamentu are new dialects of French, in the case of the former group, or new Romance languages.3 The whole distinction between internally versus externally motivated change has to do with another legacy from the nineteenth century, viz., the ideology of language purity, which is related to that of race purity (see Chapter 6). Accordingly, hybrids, products of race or language mixing, are less normal, if not simply abnormal phenomena. It is therefore not surprising that Creoles and pidgins, as putatively extreme cases of externally motivated change, have typically been suggested to be, if not downright isolated as, unnatural developments, even by creolists (e.g., McWhorter 2000,2001,2005). The putative distinction between the unusual emergence of Creoles and the normal evolution of non-creole varieties also appears to have to do with a myopic perception of colonization (at least by this name) as a recent phenomenon, correlated only with the dispersal of Europeans around the world since the Great Explorations of the fifteenth century.

Population Movements and Contacts Unfortunately, this position overlooks, or downplays, the important ecological fact that, for instance, the emergence of the Romance languages has to do with the Roman Empire, which is a past instance of colonization. The spread of Roman culture (including Rome's political, economic, and military systems, as well as its language) entailed population movements and language contacts. The latter resulted in language shifts, a concomitant of the gradual prevalence of the colonial language at the expense of the indigenous ones. This expansive and substitutive evolutionary process is also true of the birth and subsequent spread, with modification, of Old English as one of the consequences of the colonization of England by Germanic populations since the fifth century. The fact that this particular evolution also involved the concurrent demise of the continental European languages that the colonists had brought with them should not distract us from what the process shares ecologically with the emergence of Creoles, especially population movements and language contacts leading to language change in ways similar to the development of European koines in, for instance, the New World colonies since the sixteenth century. Increased and/or additional contacts with the indigenous Celtic languages and other continental European languages (e.g., Old Norse and Norman French) would likewise lead both to the demise of the latter and to more restructuring of English all the way to its modern forms. The experience of the insular Celts here is comparable not only to that of the continental Celts whose shift to Vulgar Latin produced the Romance languages but also to that of African slaves in the New World and Indian Ocean who developed Creoles while concurrently losing their heritage languages.4 Many other examples can be cited, including the colonization of most of sub-Saharan Africa by the Bantu populations (an expansion which started about 5,000 years ago), the gradual loss of Pygmy and Khoisan languages, and the concurrent diversification of the Bantu languages themselves. I submit that it is in fact possible to account for language diversification around the world as generally motivated by population movements and contacts, therefore by language contact even in the cases of the so-called "internally motivated changes." This ecological factor affects the markedness values of competing variants in a particular language (variety) setting it to evolve differently from before the contact. In kind, it does not matter whether the varieties in contact are separate languages or dialects of the same language, as is evident from Trudgill (1986, 2004). What matters is especially how typologically similar the varieties in contact are, aside from other ecological factors determining the way the target variety is being appropriated. This position entails acknowledging language and/or feature competition and selection as a central part of the engine which drives language evolution (Chapters 1 and 7). Once we explain, as in Mufwene (200la), that the fundamental and only kind of contact that triggers language evolution is inter-idiolectal, then the distinction between internally and externally motivated change becomes an artificial one, mostly sociological (see also Pargman 2002).5 All causes of change in any language are external to its structure, lying in

31

32

Language Evolution the communicative acts of speakers, such as the accommodations that speakers make to each other in order to be (better) understood and in the exaptations they make of old materials to convey new ideas. Contrary to the confusion in Trudgill (2004), mutual accommodations among speakers are part of the continual process of competition and selection which changes patterns of variation in a speech community, reducing them in some cases and increasing them in others.6 Kretzschmar & Tamasi (2003) actually show that languages have long-lasting memories and that variation is seldom obliterated. The dynamic fluctuations subsumed by competition and selection are where we should find the answer to the actuation question (Chapter 2). While it is necessary to situate the actuation of language change in individual speakers, it is only at the communal level that important cross-idiolectal patterns of evolution emerge. I focus on individual speakers here to highlight both their role as cumulative agents of changes and to situate their agency in their communicative acts. This is as important as situating them in the histories of their communities, from which they have both inherited a number of patterns while (re)shaping others or innovating new ones to meet their communicative needs (Mufwene 200la). This approach to language as practice (constantly in flux), rather than as a static "system," is consistent with Meillet's (1921,1929,1951) andHagege's (1993) observation that language "acquisition" is partly inheritance and partly recreation. It reflects an idea we also find in Lass's (1997) phrase "imperfect replication." While it explains why "internally motivated" changes occur in the first place, it also raises the following question: What distinguishes imperfect replication in LI acquisition from "imperfect second language acquisition" as invoked by Thomason (2001) and Thomason & Kaufman (1988) to account for the divergence evolution of Creoles? Aside from the influence of the language(s) previously spoken by the learner, which is an ecological factor, what differences are there in kind between imperfect LI and L2 acquisitions? As in Mufwene (2001a), I maintain that all language changes are externally motivated, in the sense that motivation for change is external to language structure, and contact (situated at the inter-idiolectal level) has always been an important factor causing changes in the "balance of power" among competing variants. I will deliberately ignore those other important changes that are due to innovations or exaptations, which are associated with simple inability to replicate the way other members of the community speak or with new concepts which cannot be literally expressed with traditional forms or constructions.7 To accomplish this demonstration, which will be kept at the macro-level (that of communal languages), we will review part of the history of Europe as a history of population movements and language contacts. Because of how colonization is implicated in this history, we will also discuss the differential ways in which European colonial languages have evolved outside Europe, including the kinds of issues that have recently interested linguists under the umbrella of "language endangerment and death." However, it will help to clarify some myths at the outset of this critical excursus.

Population Movements and Contacts

3.2 Some myths and facts about the development of Creoles 3.2.1 Among the facts are the following: Creoles of the Atlantic and Indian Ocean, our heuristic prototypes for this historic category of language varieties, have evolved from the contacts of non-European populations with European populations principally of low socioeconomic backgrounds, (native and nonnative) speakers of nonstandard European vernaculars in exogenous plantation settlement colonies (Chaudenson 1992,2001,2003). These territories, mostly insular and coastal, and typically between the tropics, were third-party places in which both Europeans and non-Europeans were newcomers. The colonists developed derivatives of European socioeconomic systems that depended on slave labor, which quickly became the overwhelming demographic majorities. These societally multilingual communities gradually evolved into monolingual ones in which the economically and/or politically dominant European language prevailed at the expense of all the others, an experience that is true also of non-plantation settlement colonies such as most of North America and Australia. Language shift among the free populations was experienced later than it was among the slaves and apparently also among the indentured servants. History suggests that because of extensive parochialism in settlement patterns, White Americans from continental Europe did not generally shift to English as a vernacular before the nineteenth century. New colonial varieties of European languages evolved concurrently, in the form of koines among the native speakers but in more divergent forms among speakers of different non-European ethnolinguistic backgrounds, especially after the European and non-European populations were segregated. That is, Creoles were diverging from other colonial varieties of European languages precisely while the latter were also diverging from their metropolitan ancestors. They thus evolved from varieties that were neither particularly metropolitan nor identical to their contemporary ex-colonial kin spoken primarily by descendants of Europeans. This correction must bear on how we discuss their "genesis," especially why this has been treated as exceptional. Even the extension of the label Creole to all the vernaculars that evolved primarily among non-Europeans in plantation settlement colonies associated with the cultivation of sugarcane, rice, or coffee begs the question, as does the position that they are not genetically related to their "lexifiers" and other colonial varieties with more or less the same lexical base that evolved concurrently with them.8 As much as linguists have denied it, the race of speakers seems to have been an important factor in this categorization of language varieties (Mufwene 2001a, DeGraff 2003), owing largely to the fact that the distinction between Creole and non-creole originates in the

33

34

Language Evolution nineteenth-century ideology of pure languages. Subsequent extrapolations of the term Creole to language varieties that evolved from colonial contacts involving no European languages (e.g., Lingala, Kikongo Kituba, and Sango in Central Africa) or to others that emerged from contacts of European languages before the fifteenth century (e.g., the Romance languages) only illustrate adequately the perception by some scholars of undeniable similarities between the contact ecologies of the emergence of Creoles and those of the evolution of other languages, as well as between the mechanisms involved in their restructuring processes. I revisit ecological aspects of these issues in Section 3.3 below.

3.2.2 Among the myths are the ones itemized below, which have been disputed principally by Chaudenson (1973-2003) and Mufwene (1996a-2005a). Space considerations make it unnecessary to repeat the arguments here; interested readers may find them in the above references. Suffice it here to expose and dismiss them very quickly, so that we can reinterpret European language history also from a colonization perspective. 3.2.2.7 Creoles are often claimed to have developed abruptly, formed by children from erstwhile pidgins made by their parents. History suggests, instead, a gradual development from the colonial koine ancestors spoken as vernaculars by the Creole populations of both European and non-European descent in the homestead communities that preceded the plantation communities. The intimate, though by no means equal, living conditions between slaves, indentured servants, and other early colonists made impossible the development of pidgins as reduced means of communication based on sporadic contacts. They precluded the possibility that Black and Mulatto Creoles would have spoken different varieties from White Creoles, especially if population structure rather than "biological race" determines what particular language variety a child "acquires" (see Chapter 6). Both Creole and Bozal children seem to have played an important role as transmitters of the colonial language and as an obstacle to a more rampant spread of substrate influence in the emergent varieties, although, as DeGraff (1999) observes, they must have contributed to the selection of some of their parents' xenolectal peculiarities into the new systems. As argued in Chapter 8, they must have identified many of their parents' features as xenolectal and avoided them. Moreover, as shown in Mufwene (2005a), there is an interesting geographical complementary distribution between the territories where Creoles developed and those where pidgins emerged, as illustrated in Map 1. The above considerations are backed by economic ones. Plantations hardly developed overnight. Most of the colonial founder populations among the free Europeans were too poor to begin a plantation. Although many were rewarded with (generous) land grants to

Population Movements and Contacts

Map 1 Complementary geographic distribution of Creoles and pidgins

develop the colony, they did not have the financial capital needed to get started. So they started with small farming and it took decades, and sometimes a whole century, before a successful family developed a sizeable plantation. As Dunn (1992) shows in the case of Jamaica, the most significant large plantations did not emerge until the eighteenth century, whereas England colonized the island in 1655. By the mid-eighteenth century, there were fewer than 15 large plantations on the island, suggesting that the evolution of English could not even have been uniform among the slaves, not any more than it was anywhere in the colonies among Europeans. (See Mufwene 2005a for more on the subject matter.)

3.2.2.2 It has been claimed that Creoles reflect imperfect L2 learning of the European colonial languages by the slaves (e.g., Thomason & Kaufman 1988, Thomason 2001). The emergence of other colonial varieties which are also divergent from the metropolitan ones, e.g., American and Australian Englishes, Quebecois French, or Brazilian Portuguese, suggests that imperfect learning applies to all groups. No operational yardstick has been proposed to measure the threshold past which one divergent variety qualifies as a Creole and below which it does not. No history-based characterization has been provided of the unique modes of language "transmission" and "acquisition" that must be associated with the emergence of Creoles. One must look for an explanation other than imperfect language learning to account for the undeniable structural differences between the Creole and non-creole colonial varieties of European languages we know today. They lie primarily in the specific ecologies of language appropriation, which have varied from one language contact setting to another, even among Europeans in the colonies.

35

36 Language Evolution 3.2.2.3 Since about Polome (1983a), it has also been claimed that structures of Creoles are so divergent from those of their non-creole kin because there was a break in the "transmission of their lexifier." As McWhorter (2001) puts it, a new language was invented from, or after, the "pulverization" of an earlier one. This "discontinuity hypothesis" does not explain why Creoles retain such large proportions of their vocabularies from their lexifiers (up to 95% in Gullah, according to Cassidy 1980, over 85% in Saramaccan, one of the most "radical" Creoles), nor why we can link so many of their structural features to their nonstandard European ancestors (see, e.g., Chaudenson 1992,2001,2003 and Corne 1999 for French Creoles, as well as Mufwene 200la for English Creoles). In fact, the position cannot explain why in all plantation settlement colonies the slaves developed their vernacular from a European, rather than an African, language. Even Berbice Dutch, which has retained more grammatical materials from a substrate language (Eastern Ijo, per Kouwenberg 1994) than any other New World Creole, has evolved primarily from Dutch. Although Surinamese Creoles have been heavily influenced by Dutch, most of their lexical and structural materials originate in English, the language of the founder colonists. The Dutch colonists simply decided to communicate with their slaves in the language that their English predecessors had used, although they themselves influenced the emergent varieties too, as evidenced by the materials of Dutch origin in Sranan. There is nothing in the structure of Creoles that supports the "discontinuity hypothesis." The normal process of "imperfect replication" under varying ecological conditions accounts not only for differences between Creole and non-creole colonial offspring of European languages but also for variation among and within the Creoles themselves.

3.2.2.4 It has also been claimed that the comparative method cannot apply to Creoles in order to determine whether they are genetically related to their "lexifiers" (Thomason & Kaufman 1988, Thomason 2002). As noted above, the method has never been tried (see also Chaudenson 1992, 2001; Mufwene 2003a). The myth is based on mistaken comparisons of Creoles' structures with those of the standard varieties of European languages. Creoles may actually reveal the artificiality of the comparative method itself based on "clean" data constituted by written records, which are not representative of the messy, internally variable, spoken vernaculars. Besides, as noted in Section 3.1, the comparative method only shows how much material is shared by a group of languages, regardless of whether the common stock was inherited from a common ancestor or spread among them by diffusion. This point is well made by most of the contributors to Aikhenwald & Dixon (2001) and more recently by Tremblay (2005). Meillet (1900) had already developed a similar argument, pointing out, in addition, that genetically related language varieties may share morphological structures or distinctions simply because they innovated (or borrowed) them under

Population Movements and Contacts similar ecological conditions, but not necessarily because they inherited them from a common ancestor.9 Moreover, we cannot ignore cases where a language diverges significantly from its genetic kin simply because it has been heavily influenced by other languages, as in the case of English, heavily influenced by Latin, French, and the insular Celtic languages (among others), compared to Dutch and German. In Part 3, I return to a position submitted in Mufwene (2005a), viz., that Creoles are, along with their other colonial kin, the latest linguistic outcomes of the Indo-European dispersal. The European colonization of the world since the fifteenth century, illustrated in Map 2, is just an extension of the original Indo-European dispersal which began about 5,0006,000 years ago (see, e.g., Forston 2004) and proceeded according to the cluster-bomb model illustrated in Map 3.

3.2.3 There are a few issues that must be addressed if linguists wish to stick to the assumption that Creoles are peculiar, with their unique kinds of ecological conditions (if not of restructuring processes) that set them apart from other cases of language speciation.

3.2.3.1 One of these issues arises from the association of the emergence of Creoles with large plantation colonies, especially those that thrived on sugarcane cultivation. If this association is true of most of the Caribbean, it is not true of Brazil. Although this polity was the first New World colony to have launched into sugarcane cultivation and the one with the largest

Map 2 European colonial expansion since the fifteenth century

37

38

Language Evolution

Map 3 Indo-European dispersal: The first colonial expansion

number of slaves engaged in this particular industry, it has produced no language varieties identified as Creoles. The closest thing to a Creole spoken in Brazil is Popular Brazilian Portuguese, which, based on Naro & Scherre (1993), developed basically from (ex-)colonial nonstandard Portuguese, whose features are clearly traceable to metropolitan Portuguese (see also Couto 2007). Attempts to classify it as a "semi-creole" (Holm 1989, 2004) underscore the metropolitan origins of most structural features of virtually any Creole, despite the undeniable role of substrate influence on their restructuring into novel "systems." See especially Chaudenson (1992,2001,2003) and Corne (1999) for French Creoles. The case of Brazil need not be confused with that of colonies such as Cuba and Santo Domingo (the present Dominican Republic). The latter territories had a protracted homestead development phase, during which their primary industry was animal husbandry, as they gave up sugarcane cultivation quite early in their colonial history. A large-size and relatively well-integrated Creole population had already emerged by the late eighteenth century when they resumed industrial sugarcane cultivation and imported more and more slaves till the late nineteenth century. Owing largely to the particular commitment of the Spanish colonial regime to teaching Castilian Spanish and literacy to their slaves, as much as to Christianizing them, the Creole slaves had been speaking the same kinds of colonial Spanish as other Creole populations. The integration of the Bozal slaves by the large

Population Movements and Contacts proportion of Creoles prevented the emergence of creole-like varieties, though at least one significantly divergent variety of Cuban Spanish has been reported, associated especially with nineteenth-century migrations from Haiti since its independence (Schwegler 2006). The extent of miscegenation in Brazil seems to reflect a different kind of colonial population structure which allowed relative integration of the labor force, regardless of whether they were African slaves, indentured servants, or Native Americans. The case of Cuba and the Dominican Republic shows that late implementation of segregation could not produce as much divergence as early adoption of the same practice, such as in the Anglophone and Francophone Caribbean. This linguistic aspect of their histories is in fact reminiscent of the continental USA (except Louisiana), where the late institutionalization of race segregation in the late nineteenth century has produced no significant structural differences between AAVE and White Southern English. The main differences lie in the statistical distribution of the variants. The reason for this state of affairs is that the separation of races came too late to induce a significant speciation of what until then had been the same vernacular of lowclass speakers. (See, for instance, Bailey & Thomas 1998, regarding their phonological features, or Wolfram & Thomas 2002 for AAVE in Hyde County, in coastal North Carolina.) On the other hand, segregation imposed by geographical isolation accounts for the development or maintenance of Creole vernaculars on islands such as St. Vincent (Prescod 2004) and Cariacou (Kephart 2000), on which majority populations have always been overwhelmingly of African descent.10

3.2.3.2 Some colonies such as the Netherlands Antilles and Cape Verde have produced varieties identified as Creoles (Papiamentu and Crioulo, respectively), although they hardly developed any noteworthy agricultural industry. For a long time they did not develop a large-size permanent slave population, as they served primarily as trade posts and slave depots where slaves were in transition to other destinations. Is it justified to associate the development of Creoles with large plantations and treat the emergence of Popular Brazilian Portuguese as an exception from the rule? The case of Brazil certainly highlights population structure as a more important factor than the high demographic disproportion between the native and nonnative speakers in favor of the nonnative group. However, it also appears that where segregation or isolation of the relevant population (that engaged in the process of language shift) is a factor,11 we must also take into account another factor discussed in Mufwene (1996a): rapid population replacement in a steadily growing overall population. This seems to be a critical factor for Cape Verde and the Netherlands Antilles. Brazil differs from these places in developing a less segregated society, although there is evidence of race discrimination against populations of African descent, as in most other colonies. Creoles must have evolved as significantly divergent and segregated vernaculars not only on the plantations but also in those colonies that

39

40

Language Evolution served as major slave depots. Thus, the development of Creoles need not be associated with one particular kind of economic activity.12

3.233 If (extensive) societal multilingualism is a condition for the development of Creoles, this factor does not seem to have mattered in the case of Berbice Dutch, since this variety developed out of the contact of primarily two major populations: the Dutch and Eastern Ijos. I raised this question in Mufwene (1993a), but I have not seen anybody address it yet. Also, if relexification were really involved in the development of Creoles, this is one particular variety that would prove it. However, even detailed studies such as Kouwenberg (1994) do not show exclusive or predominant origins of Berbice Dutch's structures from Eastern Ijo. It is possible to account for its emergence as a byproduct (not outcome, pace Thomason & Kaufman 1988) of language shift. The case of Berbice Dutch seems comparable to that of the emergence of the Romance languages as byproducts of the gradual shift of southwestern European populations from their Celtic languages to Vulgar Latin. An important difference lies in the fact that this ethnographic process was endogenous in the case of the Celtic populations and exogenous in the case of the Eastern Ijos in Guyana. This shows, incidentally, that the distinction between endogenous and exogenous evolutions is useful only when it can shed light on how much pressure a population must have felt to give up their heritage language and whether substrate influence must have been facilitated. On the other hand, the case of Berbice Dutch in Guyana is different from that of foreign workers' German and French in Western Europe, because even speakers of the "lexifier" were in an exogenous setting in the case of Berbice Dutch and other Creole vernaculars. This should direct attention to the contribution of the European populations themselves to the restructuring process, a role that was once recognized by proponents of the baby talk hypothesis.13 Moreover, no complete shift is involved in these more recent cases. If one considers the fact that children become gradually monolingual in French or German, then we may assume that the variety will die soon after the immigration of foreign workers ends (Chapter 8). These differences make the distinction between the development of Creoles and that of non-creole varieties non-structural, pointing to no particular language restructuring process that can be identified as creolization. Thus, Mufwene's (2000a) conclusion that creolization is a social process remains valid. Noteworthy in this connection is also the fact that, as pointed out by Chaudenson (2001), none of the studies of foreign workers' xenolectal varieties has produced evidence for relexification, as homogeneous as the populations that produce them have been ethnolinguistically, viz., Turkish speakers in Germany and Arabic speakers in France. (See, for instance, Klein & Perdue 1992 and Giacomi et al. 2000.) Their ethnolinguistic homogeneity has also been maintained thanks to residential and/or social segregation, which has left the adult migrants (not their children) to interact mostly among

Population Movements and Contacts themselves and to (learn to) speak the local vernacular occasionally, mostly at work and at the market with non-Turks and non-Arabs, respectively. If this comparison sheds any light on the development of Berbice Dutch, from an ethnographic perspective, it suggests that language shift among the Eastern Ijos in Guyana must have been more gradual than elsewhere, because they could communicate in Ijo among themselves. It also suggests that a key factor in cases of complete language shift is the "refusal" by children to speak their parents' languages, especially in situations where no more speakers of the ancestral languages are brought to the exogenous setting (Chapter 11; Mufwene 2005a).

3.2.3.4 Are the so-called "semi-creoles" evidence of a process of "creolization"? Or do they suggest a continuum of degrees of structural divergence from the "lexifier"? In the latter case, they may be interpreted to suggest that, subject to specific ecological conditions, the restructuring process did not always yield varieties that are equally divergent. At the communal language level, one can say that some varieties remain closer to the terminus a quo than some others.14 The best empirical answer to the question of whether "semi-creoles" are evidence of a "creolization" process comes from Schneider's (1990) study of English colonial varieties spoken by the descendants of Africans in the Caribbean and the United States. He presents a continuum of divergence from standard English, showing AAVE to be less divergent than basilectal Creole varieties of especially Suriname and Guyana. The study corroborates in quantitative terms what Alleyne (1980) had already observed. While recognizing the local social continua associated with the Creole communities of the Anglophone Caribbean, the latter also posited a broad geographical continuum of "Afro-American" extending from Saramaccan in Suriname to AAVE in the United States. He hypothesized this to be a concurrent of the differential development of Creoles, which he claims do not have pidgin ancestors (Alleyne 1971), rather than the outcome of "decreolization" as claimed by Schuchardt (1914), Bloomfield (1933), DeCamp (1971) and many later linguists who assume Hall's (1962) "life-cycle" theory. According to these linguists and contrary to the position of this Chapter, Creoles had their origins in the pidgins or jargons that they claim to have preceded them; coexistence with their "lexifiers" would cause them to shed away their "creole features" in favor of "acrolectal" ones, just like speakers of nonstandard varieties putatively get rid of their nonstandard features to develop vernaculars closer to their standard counterparts. As shown by Lalla & D'Costa (1990) and argued in Mufwene (199la, 1994), there is no convincing diachronic evidence (perhaps with the exception of Barbados, per Rickford & Handler 1994) for "decreolization" qua debasilectalization in creole societies, let alone for the evolution of AAVE from an erstwhile Gullah-like English that would have been spoken all over the USA in the eighteenth and/or nineteenth centuries (Mufwene 1994).

41

42

Language Evolution Research by Labov& Harris (1986), Bailey &Maynor (1989), Bailey & Thomas (1998),Poplack (2000), and Poplack & Tagliamonte (2001) suggests divergence evolution between AAVE and American White Southern English, just like Pollard (2000) suggests non-convergent evolution between Jamaican Creole and Jamaican English. The greatest problem with the semi-creolization hypothesis as the reversal of the decreolization hypothesis lies in difficulties with determining what "creole features" really are or what a "prototypical Creole" is supposed to be. McWhorter's (1998) attempt at identifying the prototypes by the combined absence of inflections, derivations, and tones was not particularly successful. With extensive counter-arguments, DeGraff (2001a, 2001b) demonstrated that the alleged lack of derivations was no more than a controversial parameter in the case of Haitian Creole. We may want to add that the "lexifiers" are not particularly tonal languages, which can naturally explain why Creoles need not be expected to be tonal, although most of their substrate languages are. The competition of features was naturally resolved in favor of toneless systems, except in Papiamentu. In addition, the nonstandard varieties from which Creoles have evolved are not as inflectionally rich and regular as their standard varieties (Chaudenson 1992, 2001). The rarity of inflections in Creoles is thus, as Chaudenson (ibid.) already observes, the ultimate conclusion of an inflection-erosion process that was already taking place in the colonial "lexifiers" themselves. Publications such as Brasseur (1997) and Chaudenson etal. (1993) make this observation more evident. Coming back to derivations, one can only thank DeGraff (200la, 200Ib) for showing that Creoles have not only selected many affixes from their "lexifiers" but also innovated other ones, despite the loss of some and the significance of multifunctionality in their morphosyntaxes. The innovations were of course facilitated by patterns emerging within Creole as it was gradually evolving into its modern form. Consistent with the fact that the Romance and Germanic "lexifiers" themselves represent impoverished morphologies compared to their older Latin and Germanic ancestors, respectively, we can safely assume, after Chaudenson (1979f), that the so-called "creolization" process is largely an extension of restructuring processes that had already started in the relevant "lexifiers." There cannot be a "creolization" process without a specific combination of "creole features" that would help us identify it. There is no exact measure that can be used to determine whether the restructuring has proceeded only half, or part of, the way. Rather, the reality in which every creole differs from every other creole that evolved from the "same" colonial European language shows family resemblance, in the same way that the Romance, West Germanic, or Bantu languages share, within their respective families, only parts of their structures with each other, while remaining different in a number of other respects. The "semi-creolization" hypothesis is thus an inadequate way of capturing the differential evolution of colonial European languages in plantation settlement colonies, with some of the new varieties being less divergent from their colonial non-creole kin than others. It reflects a misinterpretation of the ways in which differing ecological settings (e.g., proportions of

Population Movements and Contacts 43 speakers, changing strengths of the variants in the super- and substrate languages, patterns of population growth) have variably influenced the outcomes of the contacts of more or less the same languages. The remark applies to observations such as in Winford (1997), among others, which identify some Creoles as "intermediate" or "mesolectal" (e.g., Bajan, the Creole of Barbados), as if a uniform basilect was once spoken everywhere in the Caribbean (and North America) by all slaves. Alleyne's (1980) position on social and geographical continua remains the one most consistent with history. This "ecological setting" is also made more complex by Chaudenson's (2003) purler de seconde generation,15 which is also evident in Le Page & Tabouret-Keller (1985) and Winford (1997), in reference to varieties such as Belizean and Guyanese Creoles, whose evolution cannot be explained without factoring in the colonial language variety imported to the new colony from an earlier colony. Colonial history suggests that cross-colony influences must have been more common: for instance, in the case of Anglophone Caribbean, from St. Kitts to Barbados and other islands in the early seventeenth century (Baker & Bruyn 1998), from Barbados to Suriname and Jamaica in the mid-seventeenth century, from Barbados to South Carolina in the late seventeenth century, from South Carolina to Georgia in the early eighteenth century, and from Barbados and other islands to Trinidad and Tobago and to Guyana in the eighteenth century. As suggested by Le Page & Tabouret-Keller (1985), this complex web of colonial migrations defies a unilinear account of the evolution of Creoles.

3.2.3.5 The above discussions must prompt us to ask whether, in the first place, it is rewarding, from the point of view of language evolution, to assume a priori that Creoles have developed in their own unusual way. While it is informative to find out the respects in which they diverge structurally from (the standard varieties of) their "lexifiers" and their other colonial kin, how important is it to label them seemingly a priori as Creoles* For instance, from an evolutionary point of view, how much is to be gained from identifying the Cape Verdean vernacular as Creole when comparing it with Popular Brazilian Portuguese? The same question applies to Gullah, also identified as a creoley in comparison with its American hinterland kin AAVE identified by Holm (1989,2004) as "semi-creole." Even the identification of the latter as a semi-creole is question-begging in comparison with other colonial varieties such as Amish English, which has clearly arisen from the contact of English with some continental European languages (notably German or perhaps Pennsylvania Dutch and the like), under particular conditions of social isolation, but is not identified as a (semi-)Creole. Perhaps the extent-of-divergence question, arising from the term semi-creole, is revealing from a sociohistorical perspective. Insofar as the identification of particular colonial varieties as Creoles can generally be correlated with places where descendants of Africans have become overwhelming population majorities, the term presupposes some measure of the extent to which the non-Europeans were socially isolated and of whether miscegenation has

44 Language Evolution been extensive or limited. From this perspective, even varieties such as Cape Verdean are Creoles, although its structures are said to be (much) closer to nonstandard Portuguese than those of Sao Tomense, Angolar, and Principense (in the Bight of Biafra). The particular ecological conditions of their development, including patterns of population growth, the ensuing population structure, and the proportion of speakers of (approximations of) the colonial European language among non-Europeans can help us account for the differential outcomes of the structures of specific Creoles. However, what we are not learning is the more general fact that, regardless of whether or not the new varieties are identified as Creoles, the specific local ecological conditions of population contacts generally influenced the differential evolution of the European languages from one colony to another. For instance, the French variety of Quebec differs from that of Louisiana just as Martiniquais does from Haitien, because of specificities of the ecological settings of their emergence. We can explain in the same way differences between, for instance, American Southern English, New England English, Gullah, AAVE, Jamaican Creole, and Jamaican English. One important reason why the traditional approach to the development of Creoles should be given up is that it has generally failed to inform us, from the perspective of language restructuring, about the specific respects in which the emergence of Creoles generally differs from the presumably more continuous evolution of, for instance, the Germanic and Romance languages from their proto-varieties. As in Mufwene (2001a, 2005a), I argue below that external history suggests that there are actually no differences in regard to the restructuring processes involved. We could as well have spoken all along of the continuous, gradual evolution of Creoles from their "lexifiers." If anything, Creoles are prompting us loud and clear to question some of our assumptions about the putative differences between their development and those of the non-creole languages, especially regarding the critical role of language or dialect contact in answering the "actuation question" (Chapter 2). Variation in the specific language varieties involved in the contacts as well as in population structure, along with other ecological factors discussed in Mufwene (2001a, 2005a), go a long way toward explaining language speciation, highlighting similarities between the evolution of both Creoles and non-creole language varieties. Those like Thomason (2002, 2003) who are particularly wedded to the comparative method as proof of genetic kinship among languages should also remember that all shared items and structures among languages are not necessarily due to inheritance from a common ancestor; contact can also account for some of the formal and structural correspondences (Aikhenwald & Dixon 2001, Heine & Kuteva 2005) as can parallel innovations (Meillet 1900, Tremblay 2005). A contact-based approach to genetic linguistics can thus be informative not only about why some offspring of particular protolanguages have more divergent structures than others, but also about how the divergence can be correlated with patterns of other languages that a protolanguage came in contact with. The approach can thus inform us about shared forms and/or structures that are not attested in the protolanguage.

Population Movements and Contacts These are precisely the kinds of explanations that have been provided for why Creoles have, for instance, well-developed serial verb constructions, due to the influence of substrate languages that have such constructions. All things considered, the uniformitarianism advocated in this Chapter makes it also natural to invoke inheritance of features from the "lexifier" as an explanation for why, for instance, English Creoles have prenominal definite articles and attributive adjectives as well as stranded prepositions in questions and relative clauses, unlike their substrate languages. Below, I start an excursus, based on external language history, which supports this uniformitarian thesis.

3.3 The creole-like evolution of English and the Romance languages Similarities between the development of, on the one hand, Creoles and, on the other, English and the Romance languages have been pointed out before in creolistics. Bailey & Maroldt (1977) hypothesized that Middle English must have Creole origins, heavily influenced as it was by Norman French. They actually overlooked the fact that little, or no, language shift was involved in this particular case among the Germanic populations in England. The majority of English people continued to speak their vernaculars, little influenced by this continental European language. Only one sociolect, heavily influenced by the speech habits of the Norman colonists (aristocrats and administrators) and used then by their English colonial auxiliaries (a small minority), has evolved to show most of this French influence. It is standard English, used here as an abstraction of convenience subsuming several national standards. It is the sociolect that has undergone the Great Vowel Shift in its fullest form, representing an interesting case of differential evolution in a communal language. This was happening at a time when the more indigenous populations, the Celts (rather than the Germanics who had colonized them since the fifth century), were gradually shifting to English and seem to have been influencing the evolution of its vernacular varieties, as now made more obvious by students of Celtic Englishes (see, e.g., Filppula et al. 2002). The completion of the shift is marked more conspicuously by the emergence of vernaculars such as Irish English. In fact, it would have been more interesting to approach the genesis of English itself as a colonial, contact-based evolution that affected the Germanic language varieties brought to the British Isles, as noted in Section 3.1. Though it is not clear whether there was a single "lexifier" in this particular case, it is very interesting that one particular group, the Angles, bequeathed their name to both the land now known as England and the language, English, which can now hardly be defined as "the language of the English people" (Mufwene 2000b). While, from an ecological perspective, its genesis is apparently comparable to koineization in the emergence of colonial Englishes, its subsequent evolution in the British Isles is very

45

46

Language Evolution much influenced by important changes in the English population structure and by a continuous process of language shift among the colonized Celts, making the phenomenon no less interesting than that of later colonial English varieties, including indigenized Englishes and English Creoles (and also pidgins).16 We need not of course repeat Bailey & Maroldt's mistake in invoking a putative process of "creolization," whose meaning as a structural process remains illusive (Mufwene 2000a, 200la), any more than we must disfranchise Creoles a priori as having developed in their own unusual way. We should first abandon the nineteenth-century myth of language purity and approach the emergence of all of the new varieties from the point of view of language contact. The next step is then to let the facts inform us about more adequate distinctions that can be made. Schneider (2007) suggests a number of these, which are ecologically based, highlighting especially differences in population structure and the function of the language in the host territory. Like Bailey & Maroldt, Schlieben-Lange (1977) compared, in the same volume, the emergence of Romance languages with that of Creoles. If there really were "creolization" as a global language-restructuring process that had the peculiarity of transforming a non-creole language into a Creole variety, her comparison was perhaps even closer to being accurate than Bailey & Maroldt's. It was after the Romans had abandoned their southwestern European colonies that the masses of the continental Celtic populations gradually shifted to Latin, which had become the vernacular of the local aristocracy and a major trade language. Christian missions and urban centers played an important role in spreading it. The scenario is similar to that of the development of Creoles, after the European and non-European populations were segregated and mostly Black Creoles and seasoned slaves served as model speakers for the Bozal slaves (see below). On the other hand, as noted by De Landa (2000), this history of Latin also suggests that it did not spread in southwestern Europe in a rectilinear fashion. On the contrary, it spread from secondary dispersal points in the colonies: the Christian missions and the trade and/or administration points where it was already entrenching itself as a vernacular or as a dominant lingua franca. Also noteworthy is the fact that it is the nonstandard varieties, identified collectively as Vulgar Latin, which, as accepted among Romanists, spread within the masses of the indigenous populations and subsequently evolved into the Romance languages, under Celtic substrate influence. In this case too, we can note a bifurcated evolution. On the one hand, Vulgar Latin continued its natural, gradual evolution into the nonstandard and colloquial Romance vernaculars spoken today. On the other hand, the elite variety known as Classical Latin remained contained in the school system and among scholars until it was replaced by the standardized varieties of the new national vernaculars (especially French, Spanish, and Portuguese). From the point of view of language vitality, one can also tell that it is not always the privileged or the most prestigious varieties that thrive. Classical Latin is virtually dead, spoken today only

Population Movements and Contacts as a lingua franca at the Vatican, while Vulgar Latin may be claimed to still be alive, in mutated forms. The continental Celtic experience of language shift to Latin and its restructuring into the Romance languages is largely reminiscent of language shift and restructuring among the African slaves in the New World and Indian Ocean colonies of the seventeenth to nineteenth centuries. Just like the Celts learned Latin primarily from non-Romans (the legionaries and local administrators), most of the slaves had non-Europeans as their model speakers, especially during the plantation phase, when they became majorities in their respective colonies and segregation was then institutionalized to protect European minorities against possible uprisings. By the late eighteenth century, fewer and fewer Bozal slaves learned the colonial vernacular from those Creoles who were likely to speak closer approximations of European speech; they had seasoned slaves, L2 speakers of the emergent Creole, as their models. Like today's indigenized varieties of colonial European languages, the more speakers they gained, the more divergent they became, thanks to the mutual reinforcement of xenolectal features among nonnative speakers. The process was just as true during the gradual divergence of the Romance languages from Vulgar Latin. From the point of view of language shift, pressure to communicate in the new language came from people of one's own kind: urban and Christian missions' Celts in today's Romance countries, and the Creoles and seasoned slaves on the plantations. Ecology-specific considerations of adaptation costs and benefits to speakers drove the massive shift to monolingualism in the colonial language.17 Although the two cases of language shift differ in that the genesis of the Romance languages was endogenous and that of Creoles exogenous, both kinds of settings favored substrate influence, as acknowledged by Corne (1999), Chaudenson (2001, 2003), and Mufwene (200la, 2005a), i.e., primarily as factors influencing the selection of particular structural variants from among various competing alternatives in the "lexifier." In both cases, substrate influence was facilitated by the fact that the new speakers, those who had shifted, or were shifting, languages, were demographic majorities and spoke the new vernacular more among themselves than with its native or original speakers. We can indeed observe this phenomenon now in the so-called indigenized Englishes and lesfranfais africains, which display idiosyncrasies that can be related to features of their dominant substrate languages. Recall that the term substratum, so central in studies of the development of Creoles, came from Romance linguistics (Goodman 1993), in which its meaning was closer to its application in geology to lower layers of soil stratification. Having preceded Latin in the now Romance countries, the Celtic languages represented the "substratum" and Latin the "superstratum," in a way comparable to geological layers, which are also chronological. As pointed out by Chaudenson (1990) and Goodman (1993), Hall (1962, 1966) had extended the terms incorrectly, substituting a social stratification for the chronological order of arrivals in the colonies. In many cases, European colonies that developed Creoles or other divergent varieties associated with non-Europeans (e.g., Virginia, Barbados, Bermuda, and Reunion) had been

47

48

Language Evolution settled by Europeans before substantial numbers of African slaves were brought in. In others, Europeans and non-Europeans arrived literally at the same time (e.g., Jamaica, South Carolina, and Mauritius), as these "second-generation colonies" were being settled from earlier, "first-generation colonies" (in the language of Chaudenson 1979-2003). There are no exogenous settlement colonies that had been settled by Africans before the Europeans arrived. In fact, some of the islands, such as Cape Verde, Sao Tome, Principe, Reunion, and Mauritius, had been uninhabited before the European colonization.18 Nonetheless, creolists have learned to live with this notional inaccuracy, i.e., the incorrect way the term substratum was transferred from Romanistics by Hall (1962,1966). They have capitalized on the fact that African languages, ranked socially as low, influenced the gradual evolution of some colonial varieties of European languages into Creoles, just like the continental Celtic languages, also ranked as low varieties, played a role in the gradual transformation of Latin into the Romance languages. In both cases the influence was exerted through populations that were shifting from their ancestral vernaculars to a new one, the Celts in the case of Vulgar Latin and the African slaves in the case of colonial varieties of European languages. The historical facts in themselves, those having to do with population movements and language contacts, underscore the need to approach the evolution of Latin and European languages in the colonies from the same perspective of language competition and selection in varying ecological conditions, as explained in Mufwene (200la, 2003b). Chapter 8 elaborates on the meanings of competition and selection in this approach and how they operate in language evolution both regarding the vitality of languages and their structures. Suffice it here to note that in the development of both Creoles and the Romance languages, we must be careful not to confuse the question of the origins of features (lexical items, morphemes, grammatical rules, etc.), which lie predominantly in the so-called lexifier and sometimes in the "substrate" languages, with that of the influence that the latter have exerted on the selection of particular variants from a range of competing alternatives in the advantageous language.19 Many more ecological similarities emerge from this comparison of the development of Creoles and the Romance languages. They explain why Faine (1937), Hall (1950), Goodman (1964), Posner (1985), and Trask (1996) are probably not mistaken in claiming that Haitian Creole and Papiamentu, presumably like other Creoles lexified by French and Portuguese, are the latest Romance language varieties to have emerged. I return to this issue below. Recall, however, that the Latin variety appropriated and modified by the Celts was the nonstandard variety spoken by the (former) Roman legionaries and low-level colonial administrators, most of whom had been recruited locally. These are indeed the Latin speakers with whom the continental Celts interacted. It is therefore natural that the Latin which they modified gradually into Old Romance varieties (Old French and Old Iberian, in particular) developed from Vulgar Latin rather than from Classical Latin.

Population Movements and Contacts Also, contrary to what is often suggested in the literature referring to the older age, or longer history, of the Romance languages, it is actually Old Romance varieties (at a time when they were still identified as Latin) which developed directly from Vulgar Latin, not the modern Romance languages. The latter represent later, post-formative stages in their evolution. Arguments that have dodged the comparison with Creoles by claiming that it took the Romance languages over one thousand years to evolve (to their current structures) have missed the point about their initial formation. It did not take Old Romance much longer than it took Creoles to emerge, i.e., be disfranchised, as separate languages, viz., more or less two centuries from the adoption of the language by the masses of the Celtic populations. If anything, we have no idea what the structures of Creoles that survive will be like another thousand years from today. Nor can we surmise whether they will become so autonomous from their lexifiers as to develop their own separate standard varieties. In any case, there is no doubt that the same kinds of shift and restructuring processes have been involved in the evolution of the Romance Creoles as in that of the Romance languages and their recent non-creole offspring. Specific studies of the respective restructuring of Vulgar Latin and colonial European languages into new varieties can only help us articulate the specifics of this approach more concretely, consistent with the ecologies in which the varieties have emerged. I will now show why it would not be far-fetched to argue that Creoles are among the latest Indo-European vernaculars to have emerged.

3.4 The linguistic consequences of the latest Indo-European expansion The foregoing discussion can be interpreted in the following ways. The dispersal of the Indo-Europeans from their homeland (on which there seems to be no consensus) must have proceeded in more or less the same way as the settlement colonization of especially the New World, Australia, New Zealand, and Falkland Islands, not to mention the Atlantic and Indian Ocean islands around Africa, since the fifteenth century. The original populations need not have been politically and ethnolinguistically homogeneous, as well argued by Trubetzkoy (1939). They need not have departed from exactly the same geographical location either. Either Asia Minor or Caucasia, among the competing departure points, represents a broad geographical area. Nor need they have left the homeland at the same time, not any more than they could have taken the same dispersal routes. Nor did they reach their destinations at the same time. The histories of, for example, the Hellenic and the Roman Empires, as well as that of the Germanic expansion westward and southward from Scandinavia, all suggest also that the original colonization routes led them to new dispersal points from which they spread in all sorts of directions. This dispersal must have occurred in a way that is not matched at all by the traditional cladograms of genetic linguistics, with rectilinear and

49

50

Language Evolution

Map 4 A geographically adapted Indo-European cladogram

non-intersecting distribution branches, even if they were adapted as in Map 4 to match the present geographical distribution of Indo-European language families. Colonial expansions, often associated with mercantilism (Cowen 2001), were hardly ever planned. They were not at all like military expeditions, orchestrated by a central senior officer, although many of these were used to support them. For instance, no particular armylike general directed the Celts to be among the first to move out of the Indo-European homeland and to wait for the Hellenics, the Italics, and the Germanics to come and colonize them, at different times, in those territories where they (the Celts) had settled much earlier (Mufwene 2005a), as illustrated by Map 5. The history of the Indo-European expansion in Europe is one of population movements, therefore of language contacts, not only with the pre-Indo-European populations that had inhabited the territories they came to settle or control but also among themselves. In many cases where no military expeditions were involved, Renfrew (1987) seems correct in emphasizing that the Indo-European geographical expansion largely represents the cumulation of individuals' relocations singly or in small groups to other places. These are also the kinds of population movements that lead the newcomers to interact with the host populations; after all, it is through interacting individuals that contact can exert influence on language evolution (Mufwene 200la). Thus, as clearly shown by Heine & Kuteva (2005), one cannot account for the diversification of Indo-European languages and for their shared structures without factoring in

Population Movements and Contacts

Map 5 The non-rectilinear expansion of the Indo-Europeans (same as Map 3)

language contact. This Chapter shows that the rationale for approaching the subject matter as I just explained is the same as for accounting for the speciation of some Western European languages into Creoles and for how they variably share only some of their Indo-European structural features but not others, in ways that can be correlated with their contact histories. We could thus also conclude that the recent colonization of the world since the fifteenth century is but an extension of the same Indo-European expansion that started 5,0006,000 years ago. As in the case of the colonization of England by the Germanics, the recent colonization of the world by Europe started with small expeditions of pioneer colonists, who would be joined later by larger cohorts of immigrants who imposed their socioeconomic systems, although these were adapted to local physical ecologies. In settlement colonies, they became majority populations, with the exception of plantation settlement colonies, in which the non-European labor populations (slaves or contract laborers) became majorities. In all these cases, the indigenous populations were overwhelmed demographically and otherwise by the newcomers.20 The particular experience of settlement colonies is interesting because it entailed language shift in favor of (one of) the exogenous ruling populations. However, the latter's languages won only Pyrrhic victories, modified as they were by the languages that they displaced.

51

52

Language Evolution Just as Vulgar Latin was restructured into the Romance languages, most of the West Germanic languages taken to England died (like several European languages in North America), being displaced by the language of the Angles, which concurrently speciated into several British varieties and would continue to diversify into new ones outside the British Isles. Overall, in every extra-European settlement colony, the dominant European language has displaced not only indigenous languages but also other European languages with which it had competed. Scholars opposed to my perspective will want to capitalize on the fact that although they have evolved from the same colonial European koines, Creoles are structurally more divergent from their European metropolitan origins than their non-creole colonial kin (e.g., North American English varieties). In the first place, the reliability of such judgments depends on the kinds of varieties being compared. It is not so clear that structural differences are so strong between Cajun French, Louisiana French Creole, and Louisiana French, or between Gullah and Amish English, especially if one overlooks prosodic features. Be that as it may, the current literature has usually omitted considering population structure as an ecological factor that can account for structural differences between the colonial vernaculars spoken by descendants of Europeans and those spoken by descendants of African slaves, especially in the case of North America. Like the indentured servants of the colonial period (e.g., the period preceding the American Revolution), the slaves shifted early to their masters' languages. As soon as they became the plantations' dominant populations, segregation was instituted, quite early in the eighteenth century, forcing them to interact and socialize more among themselves than with populations of European descent. As the proportion of the Bozal slaves kept increasing, more and more Bozals learned the European colonial vernacular from seasoned slaves, whose command of the language was already divergent, than from the original Black Creole speakers, whose competence differed little, if at all, from that of their White Creole counterparts. During this continual language shift process, the colonial vernacular underwent more and more substrate influence and diverged increasingly to reflect this substrate influence, bearing in mind that, as explained in Chapter 8, the influences themselves were subject to competition and selection among the slaves. This situation is quite comparable to that of the gradual shift of the continental Celts to Latin as their vernacular. Being the majority populations increasingly communicating among themselves in the new, adopted vernacular, their Celtic linguistic habits apparently influenced the restructuring of Vulgar Latin into today's Romance languages, which are as different from Classical Latin as Creoles are from the standard varieties of their "lexifiers." There are very good ecological reasons why the colonial varieties spoken by descendants of Europeans diverge less from the metropolitan varieties than Creoles do, assuming correctly, like Chaudenson (1992, 2001, 2003), that the divergence is generally a later development. As also observed by Chaudenson, most of the nonstandard vernaculars of the European

Population Movements and Contacts languages that came in contact with each other are typologically more similar among themselves (as made more obvious by Heine & Kuteva's 2005 demonstration that Europe is an important linguistic area marked by convergence) than they are with the African languages. Equally important is also the fact that the European populations remained largely segregated by nationality until the early twentieth century. Until then, they kept their national vernaculars and used the economically (and politically) dominant language only as a lingua franca. Thus, the colonial English varieties now spoken by descendants of Europeans seem to have developed primarily among colonists from the British Isles, with only minimal "adstrate" influence21 from the other European languages that they displaced later. At the time of the shift in the late nineteenth and early twentieth centuries, the adults must have gradually died out with their accents, just like many adult immigrants do today, while their children, through regular interactions with the children of native speakers (especially in school), acquired the dominant vernacular natively (Chapter 11). This process, which can still be observed today among the non-Aboriginal populations of Australia (Clyne 2003), has kept adstrate influence to a minimum. In fact, it is still true of recent and current immigrant families in North America, like anywhere else in the world. The above explains why, in North America, it has become less and less common to speak of, for instance, German and Italian Englishes. The populations of English and non-English descents have increasingly blended to the point of reducing ethnic differences among them, at least among those who are not first-generation immigrants. Most of the linguistic peculiarities that justified those ethnic labels have disappeared, while a few of them may have spread in general American English. Only varieties spoken in communities that are not yet (fully) integrated in the dominant population still carry such ethnic labels (e.g., Amish English) in some way comparable to regional labels such as Ozark and Appalachian Englishes. Yiddish English, associated with Jewish first-generation immigrants from Yiddish-speaking territories, is considered moribund. Interesting from this perspective is also the fact that twentieth-century African and Caribbean immigrants to North America have exerted no noticeable influence on the structures of today's AAVE, largely because they have generally not settled in African American neighborhoods and have socialized with them in no more significant ways than they have with other Americans. When this has been the case, these immigrants' children have learned to speak AAVE natively. Caribbean immigrants who have settled in communities in which they either are dominant or constitute a critical mass have maintained their native Caribbean features, whereas their children have had to choose between their parents' and American English features.22 The same observation applies to any ethnic community, unless it has been formed recently and just reached its critical mass. Innovations generated by the internal ecology (i.e. the dynamics of both intra- and interidiolectal variation within a communal system) continue of course, such as in the Valley Girl

53

54 Language Evolution Talk in California, Hip Hop language among American teenagers, and Dread Talk in Jamaica. However, very little of this innovative speech has been motivated by contact with the languages of the recent immigrants. Inter-group variation within national varieties also persists, because it is determined by patterns of social interaction; schooling and the mass media have little bearing on it. Likewise, as explained in Chapter 2, Labov's (2001) discussion of the Northern Cities Vowel Shift reveals an interesting aspect of how changes spread. They are contained within populations whose members interact with each other. In this particular case, the shift has spread only among White American urbanites, who not only socialize primarily among themselves but also migrate from one city to another rather than to rural areas. This reflects the persistent de facto ethnic segregation in the American population structure. Also worth noting in this case is the fact that the Vowel Shift has spread not only among White speakers but also from city to city, hopping over rural areas, reflecting the fact that national populations are indeed ecological metapopulations in which smaller aggregates are interconnected by dispersing individuals driven by particular economic pressures. All these differential evolutions underscore the significance of POPULATION STRUCTURE as an ecological factor. It determines not only who interacts with whom but also who accommodates whom and under what ethnographic circumstances. This Chapter also makes more evident the fact that language shift, which has recently been associated with the demise of the ancestral languages of several indigenous and immigrant groups in European settlement colonies, has been a concomitant of the spread and diversification of Indo-European languages. From a longer perspective, this evolution has been in process since the first territorial expansion of Proto-Indo-Europeans in Asia and in Europe to the latest dispersal that has produced new colonial offspring of some of the modern Indo-European languages. Extinct languages have left their marks (substrate or adstrate influence) on those that displaced them. A more accurate account of language vitality in genetic linguistics should thus provide a balance sheet of losses and gains, instead of capitalizing exclusively on losses, bemoaning loss of diversity and ignoring the emergence of new patterns of diversity (Chapter 11). Interestingly, very few of the massive shifts from indigenous vernaculars to European languages have been observed in European exploitation colonies of Asia and Africa, where the same languages have been adopted as official languages and as elite lingua francas. Despite the fact that the latter have indigenized, as evidenced by the growing literature on indigenized Englishes (see, for instance, the references in Bolton 2003, Kachru 2005) and les fratifais africains (e.g., Lafage 1977, Manessy 1994, Manessy & Wald 1984), rare are the cases where Native Asian and African vernaculars are being driven out by the colonial European languages of their respective polities (Chapters 12-13). That Colored people in South Africa have appropriated Afrikaans as their vernacular is more a consequence of the dual colonization system in this polity, with the Afrikaners'

Population Movements and Contacts socioeconomic domination representing settlement colonization and the British the exploitation style. However, recent changing political tides are increasingly reverting the ethnographic status of Afrikaans to that of an ethnic language that is economically less advantageous, causing more and more Colored people to raise their children as native English speakers. Afrikaans may become one of the rare languages to be considered "at risk" while they have much more than a few hundred thousand speakers, millions in this particular case. Its feared demise reminds us that it is not so much the number of speakers in abstraction that determines the vitality of a language but how it negotiates speakers with the other languages it competes with and whether the proportion of children acquiring it as a vernacular increases or decreases. The explanation for this differential evolution of colonial languages is largely ethnographic. In Asia and Africa, the colonial languages have typically been appropriated as lingua francas, whereas they have been appropriated as vernaculars in the settlement colonies of the Americas, Australia, and New Zealand. The extent to which the indigenous languages have been endangered or lost in the settlement colonies is covariant with the extent to which the indigenous populations have been exterminated (as in the Caribbean) or the extent to which the indigenous socioeconomic system has been driven out and the Natives must function in the new socioeconomic world order in order to survive. Likewise, the extent to which those who have shifted to the new vernaculars have influenced its structures depends largely on how early during the colonial rule (which is still in process in the settlement colonies) the relevant population had to function in the new socioeconomic world order and use it for intra-group communication. Creoles are the products of situations in which language shift occurred early and segregation was institutionalized soon afterwards, favoring significant influence of the languages previously spoken by the African slaves on the structures of their new vernaculars. Creoles have emerged as distinct vernaculars with a high degree of vitality because the African slaves had to function quite early in the new socioeconomic world order of settlement colonies. The early imposition of socioeconomic segregation created an ecology in which the slaves socialized primarily among themselves and their colonial vernaculars could diverge structurally. As explained above, the particular ways in which their populations grew, more by importation than by birth in the eighteenth and early nineteenth centuries, with more and more Bozal slaves acquiring the vernaculars from seasoned slaves who had xenolectal features, paved the way for substrate elements to influence the divergence patterns of their new vernaculars (Chapter II). 23 At least in North America, the Natives were generally left on the margins of the new socioeconomic world orders. The Europeans traded with them in the indigenous languages, a practice that produced pidgin varieties such as Chinook Jargon, Mobilian, and Delaware Pidgin (Silverstein 1996). Although the later shift of Native Americans to English, in the nineteenth century, produced some sort of Native American English Pidgin (Mithun 1992),

55

56 Language Evolution the indigenous languages have exerted even less influence on the structures of North American Englishes than continental European languages have. Gradual absorption by the dominant populations, marked by the absence of Native American neighborhoods in American cities (unlike the tradition with, e.g., Irish, Italian, German, and Black neighborhoods at least up to the mid-twentieth century), led to the rapid dissolution of such transitional rural ethnic varieties, in the same way that Italian and German Englishes, for instance, have vanished. Overall, we learn that vernaculars compete with vernaculars and lingua francas with lingua francas. What has prevented indigenous languages in Asia and Africa from being displaced by European colonial languages is in part the fact that in exploitation colonies the latter have functioned primarily as lingua francas, and as vernaculars only within a negligible proportion of the elite. Interestingly, the European exploitation colonization also produced or promoted the expansion of (new) lingua francas "lexified" by indigenous languages, e.g., Fanakalo (in South Africa), Hausa (in Nigeria), Kikongo-Kituba, Lingala, and Sango (in the central African region), Swahili (in Eastern Africa), Town Bemba (in Zambia), and Wolof (in Senegal). Most of these function today as urban vernaculars. At least during the colonial period, there was no competition between the indigenous and the European lingua francas. Quite a few lessons can also be learned in this case: (1) The European languages have been transmitted through the school system and have been associated with the intellectual elite. Their practice has been contained in a particular social class, thus lending more theoretical significance to the notion of "population structure." This form of segregation has protected the indigenous lingua francas from their expansion. (2) The post-independence economic stagnation or demise experienced by most of the African countries has halted the expansion of European lingua francas, showing clearly that investment in a particular language can be assessed in terms of economic costs and benefits, although the degradation of the school systems takes part of the blame in these particular cases (Chapter 12).24 (3) To be sure, the degradation of the school systems can be blamed for the progressive indigenization of the European lingua francas. This is a process that probably would have been more rapid if the languages had vernacularized and their transmission had depended more on naturalistic acquisition, through socialization and face-to-face interactions, like the transmission of the indigenous lingua francas and urban vernaculars. These considerations prompt us to re-examine the past and think again over how Latin spread, at a time when only a few privileged children could attend school in Europe. They also tell us why Irish English was late in forming, although English was introduced to Ireland, as a trade language, as early as the ninth century. It is thanks to the potato plantations of the seventeenth century, under Oliver Cromwell, that Ireland would be colonized on the settlement model and English would spread informally, through naturalistic acquisition by the migrant workers. Indeed, the potato plantations gave to English an economic

Population Movements and Contacts value that the school system had not succeeded in doing, any more than it had done to Latin. This grass-roots valorization spread English faster among the Irish than the school system had in limiting access to the language only to a limited few. The spread by naturalistic "transmission" resulted, on the one hand, in the indigenization and vernacularization of English and, on the other, in the gradual extinction of Irish. More or less the same process can be assumed of Vulgar Latin as it (rather than Classical Latin of the Intelligentsia) was valorized as the language of the new, Romanized socioeconomic system, a language that would enable the subjects of the former Roman Empire to participate in the new world order. Since European colonial languages have been maintained as official languages of their former exploitation colonies and as lingua francas of the Intelligentsia, we can perhaps surmise the following: Whether or not they will vernacularize and spread through naturalistic "transmission" among the masses of the population depends largely on whether or not they are considered important for regular grass-roots jobs, which do not require high technical skills. (These are, incidentally, the only jobs accessible to the vast majorities of the Third World's populations.) So far, it is the indigenous urban vernaculars and regional lingua francas which are associated with such lucrative employment. With perhaps the exception of South Africa, the economies of sub-Saharan Africa are far from expanding, let alone involving the rural populations, which are still the majorities (Chapter 13). The European colonial languages have undoubtedly indigenized, but they have hardly vernacularized. If anything, it is the indigenous urban vernaculars and regional lingua francas that are threats to the ancestral ethnic languages, as much as they all bear essentially lexical influence from the European languages.

3.5

Conclusions

Little needs to be added here that has not been said already in Section 3.4. If Creoles have really developed in their own unusual or abnormal way, their different structures suggest that the ecologies of their emergence are far from being identical. On the other hand, if we assume a uniformitarian position and acknowledge that they have emerged by the same restructuring processes that have often resulted in language diversification, then it is normal that differing ecologies will produce new mutually divergent language varieties, even if, for all practical purposes, "exactly the same languages" are involved in the contacts. Local dynamics of competition and selection will favor different variants even from what may appear to be more or less the same feature pools. Thus, Creoles are not genetically unusual, nor abnormal, nor less natural than other language varieties. Instead, they offer linguists a precious opportunity to realize the extent to which language contact, subsuming also dialect contact but based really on idiolect contact, has been a critical catalyst of language change and speciation. This Chapter also shows that language diversification has typically proceeded hand in hand with language shift. Without invalidating the usefulness of the comparative method as

57

58 Language Evolution a tool for determining the extent to which languages share forms and structures, this Chapter shows that genetic creolistics can help us improve the practice of genetic linguistics and broaden usefully the scope of its concerns. Insights learned from research on the evolution of Creoles can apply to that of other languages too, just as vexing questions arising about them can likewise apply to non-creole languages. A broader, more integrative conception of genetic linguistics can certainly help us develop a better understanding of the ecological motivation of language restructuring and speciation, and of some messy aspects of language change, which I have not addressed here. These have to do especially with the range of variation within the prevailing language and whether specific structural changes are better accounted for as A —»B in the environment of C (mutation or transformational evolution) or as A was preferred over B in the environment of C (variational evolution, by competition and selection). To the extent that contact, situated at the idiolectal level, is acknowledged as a critical ecological factor in the actuation of change, the distinction between externally and internally motivated change becomes simply sociological, and the distinction between changes induced by contact and those independent of contact becomes misguided. Moreover, like evolutionary biology, genetic linguistics (which could also be called evolutionary linguistics) has everything to gain from being interested in issues of language vitality. It can then deal with the demise of languages by structural erosion (as associated with the putative debasilectalization of Creoles) or by language shift and with the emergence of new varieties. In all these cases of language evolution, the action of competition and selection among competing variants and/or systems is evident.

How Population-wide Patterns Emerge in Language Evolution: A Comparison with Highway Traffic1 Chapter Outline 4.1 Introduction 4.2 Highway traffic and language evolution 4.3 Conclusions

4.1

59 60 72

Introduction

The purpose of this Chapter is primarily heuristic. It is an attempt to articulate details of some aspects of the dynamics of language evolution through comparisons with something non-linguistic, the flow of motor vehicle traffic in this particular case. I could have made the comparison with a stream or a river (Mufwene 1998, 2001a), but I would not be able to highlight the role of individual speakers as "unwitting agents of change," which is central to the position I am developing in this book. The river metaphor has its own merits, such as regarding substrate influence and blending inheritance, but it does not lend itself easily to the identification of counterparts of idiolects, on which I wish to focus here. I wish to highlight similarities between the way individual speakers influence the evolution of their communal language and how the actions of individual drivers on a highway influence traffic flow. This comparison is in keeping with my assumption that the way in which a communal language changes is a function of how individual idiolects change under each other's influence and as a cumulation of adjustments that take place during their individual speech acts. Likewise, the way traffic flows on a highway is largely a function of how individual motor vehicles proceed. It is also true that what individual vehicles can do depends on the overall traffic itself, such as its volume. Likewise, the changes that idiolects can undergo are largely constrained by the language they are part of. That is, what individual speakers can produce and how they can modify their idiolects to express new ideas depends largely on what other speakers of the same language variety do or can understand. It is on the dynamics of this mutual dependence between idiolects and language, or between individual speakers and the language community, just like between motor vehicles and the surrounding highway traffic, that I focus below. I say nothing about the restructuring

ul

60

Language Evolution process itself qua system reorganization (Mufwene 200la), neither at the level of idiolects nor at that of language. This requires a separate essay in its own right (such as Chapter 7), to flesh out much of what has been proposed in the literature on, for instance, the development of Creoles. The present Chapter picks up from some thought-provoking comparisons that Keller (1994) makes between aspects of language change with aspects of automobile traffic in the city. It is also inspired by a term paper that Heidi Elston wrote for my "The Ecology of Language Evolution" class (Winter 2001) in which she critiques and elaborates on Keller's discussion of pattern emergence in the form of footpaths.2 Keller explains how a footpath emerges from the overlap of footsteps of different pedestrians walking repeatedly on more or less the same trajectory. Both discussions led me to realize that Keller may have scraped only the tip of an interesting heuristic iceberg. I show below that much more can be learned from these comparisons. I argue that, in language evolution, patterns emerge from the cumulation and convergence of what individual speakers do either when they innovate or when they deviate from established practices. The present essay is intended to raise our awareness of those seemingly fortuitous acts of speakers - during their communicative acts - which Keller identifies collectively as the "invisible hand" that drives language change.3 While it is not clear how much this discussion contributes to a better understanding of the "actuation question" discussed in Chapter 2, the thesis defended here is a natural extension of my observation in Mufwene (2001a) that speakers are the "unwitting agents of [language] change." They bring this about through repetitions of their own or other speakers' innovations or through repetitions of the recent modifications of older patterns in their communicative acts.

42

Highway traffic and language evolution

4.2.1

Keller's original examples

Keller (1994) provides several interesting analogies between non-linguistic and linguistic phenomena having to do with the role of the "invisible hand" that drives change and pattern emergence. One of these, discussed by Elston (2003), regards the paths that emerge across campus lawns not by design but as a consequence of pedestrians taking shortcuts from one building to another and avoiding the longer, paved, and geometrical paths provided by the landscapes The emergence of the paths illustrates the fact that repeated similar acts of individuals in a population wind up producing common patterns, unless some action is taken against such an evolution. Such a convergence of behaviors is also motivated by the particular ecology to which the relevant individuals respond, for instance, the need to reach a particular destination in the shortest possible time and with the least effort. The "invisible hand" which Keller invokes as the agent in the emergence of footpaths is the cumulation of

Emergence of Population-wide Patterns activities of individual pedestrians following more or less the same trajectory at different times and reinforcing each other's footprints. No single trip alone produces the footpath, while the succession of such trips by one pedestrian and/or several eventually does. The footpath is the result of the way a particular population of pedestrians has chosen to reach a specific destination, using always the "same" trajectory. In the case of language, the "invisible hand" amounts to the cumulative, though typically uncoordinated, actions of individual speakers which bring about change, this being typically an unintended outcome. The focus in this case should be both on the individual and on the role of imitation in group behavior. As some speakers adopt innovations or deviations produced by other speakers, new structural and/or pragmatic patterns emerge.4 As more and more new linguistic behaviors similar to the initial innovation occur, new norms emerge at the communal level and are identified as changes relative to an earlier stage of the relevant language. Norms emerge from the convergence of speakers' linguistic behaviors or from some speakers copying innovations or deviations produced by others in more or less the same way that footpaths emerge from the footsteps of individual pedestrians following more or less the same trajectory. To be sure, speakers do not necessarily behave in identical ways nor in a concerted fashion (Mufwene 200la), any more than pedestrians' feet can be expected to hit the same spots on the ground. However, similarities in the ways speakers communicate and exapt5 aspects of their linguistic knowledge to cope with new communicative challenges wind up producing communal patterns through the power of the "invisible hand." For instance, phenomena known as grammaticization evolve this way, with innovations by some speakers spreading through copying within a population.6 Keller makes another comparison, this time between the pace of language evolution and the speed of traffic. He invokes the case when, for reasons unknown to most drivers, the traffic jams and vehicles approaching a location must stop. Every driver applies their brakes a little bit heavier than the preceding driver, in order to avoid rear-ending the latter's vehicle. He wanted to illustrate with this example the fact that changes spread gradually, and faster and faster in a population. This is similar to the amplification of waves in wave theory, with the outer rings becoming bigger and bigger when one drops some mass (say a rock) into a pond of stagnant water. However, in the case of language change, things do not proceed only in the direction and way that caught Keller's attention. Just the opposite happens when vehicles start again after stopping. The time it takes a vehicle to start off again is largely a function of its distance from the first vehicle in front (and apparently also a function of the degree of patience of drivers in the other vehicles that precede). The farther back a vehicle is in the line, the more it seems that nobody is moving. Some linguistic changes proceed faster among speakers closer to the innovators than among others, depending in part on whether the latter can identify with the innovators in social class or in age group, among other factors. For instance,

61

62

Language Evolution according to Bailey & Maynor (1987), extensions of invariant be to denote the PROGRESSIVE rather than the CONSUETUDINAL/HABITUAL aspect in the AAVE variety of Brazos Valley, Texas, progressed faster in the urban younger generation than in the other social groups. Heterostasis and homeostasis both apply in language evolution (Gilman 1993), and in either case the speed can be faster or slower among the followers. The picture emerging from these evolutionary phenomena is a complex one, in which changes are initiated by some, can be adopted by some sympathizers or trendy members, but can also be resisted by some conservatives. Thus different innovations or deviations do not spread at the same speed, nor are they weighted identically, with the relevant populations of speakers.

4.2.2 Beyond Keller's comparisons In this section, I focus on highway vehicle traffic, where increase and decrease of volumes; as well as acceleration and deceleration of speed, are considered more normal than complete stops, which need explanations, such as congestion caused by a car accident or road construction. In language, these initial images correspond, respectively, to increase and decrease of number of speakers - hence of idiolects - and the pace of language evolution. There are no living languages in frozen, motionless state. The only ones that conjure up such an image are highly codified ritual language varieties, whose status as living languages is questionable, since they are not, or no longer, spoken. Below, I develop more analogies between the flow of vehicle traffic and that of language evolution, highlighting a number of respects in which the two evolutionary complexes of processes are similar.71 use the more concrete case of vehicle traffic to hopefully shed light on the dynamics of various factors that bear on how evolution proceeds in a language. I maintain that the dynamics of language evolution are better understood if a language is analogized with a biological, parasitic species whose existence and vitality depend on the communicative activities of its speakers. In this paragraph and the next few, I restate the central tenets of this position that also apply to highway traffic: (1) a communal language is an extrapolation from the idiolects of its speakers - it can be conceived of only on the basis of experience with some idiolects; (2) a language is internally variable, in part because its idiolects are not identical with one another - neither the experiences on the basis of which they have evolved nor the minds that have developed them are identical; and (3) a language does not evolve in a uniform way, because each speaker's communicative network is identical with nobody else's and each speaker's communicative experience is unique. Evolutionary patterns do emerge, along age groups, gender classifications, educational/professional affiliations, and other social parameters, but they are just trends, not mathematical categories. Some idiolects simply do not participate in some of the changes that a language undergoes, and those that do participate in them do not necessarily evolve at the same speed. These peculiarities are typical population behavior. Although the parameters of variation are not the same, I show below that

Emergence of Population-wide Patterns similar evolutionary patterns can be observed in highway traffic, assuming individual vehicles to be counterparts of idiolects. I use the comparison to show that language evolution follows some of the same patterns and is driven by the same "invisible hand" principles observable in non-linguistic phenomena. Therefore, an understanding of these phenomena can give us a useful heuristic perspective. While dealing with specific languages, I find it more informative to think of them as SPECIES than merely as POPULATIONS. Like members of a species, the idiolects of a language resemble each other on the family resemblance model.8 While likeness of structural features (the counterpart of phenotypes in biology) is not a requisite for isolating a group of individuals as a population, it often is a necessary condition for grouping individuals into a species. Moreover, speakers of similar idiolects also acknowledge that they speak the same language variety9 and they often assume that it descends from an earlier form to which their idiolects can be genetically related. They also claim association with the same community, even if they do not fully understand one another.10 For the purposes of this analysis, idiolects correspond to individual vehicles. The latter's individualities are determined by whether they are cars, vans, trucks, buses, or motorcycles, and by their different makes, ages, and drivers, among a host of factors that bear on the flow of traffic. Whether or not one can claim that different kinds of vehicles (cars, vans, trucks, buses, and motorcycles) are one species is debatable - and the question is similar to that of whether different breeds of dogs belong to the same species, or whether two or more language varieties (e.g., African American English and Appalachian English) are the same language. If they are not, the fact that the different kinds of vehicles share the same roadway is still significant. The presence of different kinds of vehicles on the same highway is similar to the coexistence of dialects or languages in the same population, which is reminiscent of, for instance, different animal species sharing the same econiche and affecting each other.11 I focus here on the one-way traffic of a highway, so that I can more clearly show analogies with the evolution of a species as a one-way motion from one point in time to another (regardless of regressions in the health or size of a species).12 Although one can think of traffic in motion as a giant unbounded organism, traffic is a function of individual vehicles that make it. Its overall motion is a function of the motions in which vehicles are individually engaged. It is affected by the actions of individual drivers in the different vehicles and of how these are affected by actions of their neighbors. This characteristic is almost like that of the life of a species, which is a function of individual lives of member organisms. Likewise, the life of a language is also a function of the lives of individual idiolects. It changes because idiolects undergo changes individually, because their speakers have innovated something (even by deviating from the current norm) or have adopted changes innovated by other speakers. A language dies also when the last of its idiolects dies. Whether an idiolect dies or thrives is a function of the communicative activities and life of its speaker, because it has no autonomous existence and is a curious parasite that is both made by the speaker and depends

63

64 Language Evolution on him or her to remain in existence. In this respect, languages and vehicle traffic are very similar phenomena. Provided they are in good condition, whichever way vehicles move in traffic depends very much on the drivers who operate them. The comparison of language with vehicle traffic makes it difficult to continue ignoring the agency of speakers in language change, pace the traditional practice of historical linguistics. Just as vehicles on the highway cannot move without the agency of drivers, languages do not evolve without the agency of speakers, on whom their lives depend. Traffic moves faster or slower depending on how fast individual vehicles are driven on the highway. How it moves also depends very much on whether individual drivers decide to follow a pace-setter. This depends of course on many ecological factors, for instance, whether or not road conditions make it possible to drive faster, or whether there are police patrolling the traffic, or whether staying alone in a seemingly isolated portion of the highway makes a particular driver feel lonely, unsafe, or uncomfortable. The reasons why speakers change their idiolects are similar, including keeping up with one's peers, not feeling comfortable in speaking differently, and an opportune time for a change. Linguistic changes also meet the resistance of academies or self-proclaimed patrollers of their communal language. Both vehicle traffic and languages are unbounded phenomena, not because they have no beginnings nor ends but because they are typically experienced in the present. Individual drivers or speakers focus on interactions with their neighbors and on reaching their destinations or being successful communicators. In some ways, they participate in mass behavior, in which individuals are driven by the patterns that emerge from the group that they are part of. Changes in both how traffic flows and the way a language evolves are ideally observed by outsiders, although drivers and speakers realize in hindsight how and to what extent their positions have changed relative to fellow drivers or speakers, respectively. The changes are largely due to the fact that drivers and speakers often realign themselves, as they are affected, for instance, by those that join the traffic or the community of speakers (children or immigrants) and by those who leave them (speakers who die or migrate away). In both cases, the entries and exits may be temporary or indefinite. For instance, a vehicle may get on the highway only for a short drive and exit a few miles away, just like those speakers who join a community but for reasons of relocation or death do not stay long enough with the current speakers. If they had not left when they did, they may have had an impact on how the traffic flowed or a language evolved, if they had behaved in ways that affected other drivers or speakers. Thus, by leaving, they have deprived the latter of whatever pressure they would have exerted.13 Mass behavior is especially significant in regard to drivers who share the highway for a long time with other motorists, not knowing when these other travelers joined the traffic or will leave it, if they themselves do not leave it sooner. Speakers coexist for years with each other not worrying about when the other speakers started using their communal language or will stop (due to death or to some form of impairment, or because they have shifted to another language), if they themselves do not stop speaking it before the others. Of all

Emergence of Population-wide Patterns those who participate in traffic or speak a communal language, very few are those who see it start or die. The vast majority see it in process and could not care less when it started, or if and when it will die, as long as they enjoy the company of other drivers or speakers. From a species perspective, traffic really flows in some ways that are similar to the evolution of a language. Traffic stays alive even while some vehicles are leaving and new ones join in. To say that traffic continues, or stays alive, does not mean that exactly the same vehicles make it at the beginning as at the end, wherever the beginning or the end happens to be. Typically no vehicle stays in the traffic from the initial time or location to the finish time or line. The vehicles that proceed at the same speed during a particular period do not maintain the same speed or positions throughout the flow of the traffic. Some drivers become tired and slow down, some become impatient and speed up, or they join other vehicles that are moving at a faster speed. The fear of being all alone between vehicle clusters may cause a driver to slow down so that they can join the following vehicle cluster, or to speed up until they catch up with the preceding cluster. And there are those drivers who must exit the highway to refuel or just rest, notwithstanding the unfortunate ones that are crippled by breakdowns or die in accidents. One must also recognize those drivers that break the momentum one way or another because they want to avoid particular company. All these happenings affect ways in which patterns qua some internal organization emerge in traffic. More or less the same thing is true of the way a communal language evolves. It stays alive independent of the fact that several idiolects disappear and new ones emerge, and there is no idiolect that proceeds all the way from the beginning to the end point. Speakers move around and associate with different speakers at different times. Participation in different networks (often in overlapping relations) affects the evolution of individual idiolects differentially, so that speakers find themselves in different evolutionary positions relative to each other. These dynamics are the "invisible hand" that can produce new patterns, part of the evolution process. However, not everything in a language works the same way as in traffic. For instance, unlike vehicles in traffic, evolving idiolects accommodate each other in giving up their own features or simply adding features observed in other speakers (see the "actuation question" in Chapter 2), so that even a long time afterwards, an idiolect may share features with another with which it is no longer in contact. Speakers of nonstandard varieties who, after a certain amount of schooling and learning a standard variety, move up the socioeconomic ladder and operate in different classes do not only acquire a second vernacular but often also modify their native idiolects. On the other hand, they speak the second vernacular or lingua franca under the influence of their first vernacular, and if they have not completely given up that first one, they are still similar in some respects to fellow speakers from the same background.14 This particular aspect of the evolution of idiolects is a function of Mufwene's (200la) observation that linguistic species are Lamarckian, as their features change several times

65

66

Language Evolution over during their life times. Because variation is possible even within the same idiolect (a respect in which the linguistic organism, the idiolect, is different from its biological counterpart, at least among animals), the principle of FAMILY RESEMBLANCE in the coexistence of idiolects in a communal language becomes even more critical. The idiolects of a language resemble each other for reasons that are not identical from one pair of idiolects to another. Nonetheless, like a biological species, both traffic and a communal language stay alive while they continue to have members, although these members need not be the same. The size of the population of vehicles or idiolects may decrease or increase - in fact it fluctuates but the traffic and language remain healthy as long they maintain a critical mass of vehicles and idiolects, respectively. As a matter of fact, saying that traffic or a species (biological or linguistic) grows has to do with the size of the membership rather than (necessarily) with age in its life. In the case of a communal language one can think of growth when it spreads in a population of potential speakers, which is different from conceiving of it as developing a more complex morphosyntax, for instance, in the way claimed by Schleicher (1863).15 In any case, part of the changes undergone by a language and by vehicle traffic are a function of all the exits from, and entries into, their collective bodies as they evolve. Focusing on individual members of traffic can actually shed light on some of the changes that affect a language. Much of this idiolect-and-language or vehicle-and-traffic interdependence falls in the category of the dichotomy between individual and group selection in biology, with pressure on individuals to make selections that are beneficial or adverse to the interests of their groups, while speakers or drivers aim at maximizing individual benefits a la Bourdieu (1991). The convergence dynamics involved in the process is the "invisible hand" that Keller (1994) invokes. An interesting example in the case of highway traffic is how speedy drivers determined to overtake every other vehicle in their way can change the speed of the overall traffic. Other drivers often join the speedy trail. When there are no troopers to patrol the highway and there are no particular obstacles in their path, a significant proportion of the traffic may accelerate. However, as pointed out above, there are also cases when individual drivers singly drive at a speed that is too fast or too slow and constitute a danger for the rest of the traffic. Their impact on the rest of the traffic may amount to nothing more than transitory nuisance. There are several speakers whose peculiarities are not copied by others. One of the interesting aspects of language evolution is that there are a lot of innovations or deviations that do not spread within the language community. As a matter or fact, successful innovations or deviations, which are copied by other speakers, are quite few in number compared to numerous others that are produced daily and die immediately or are short-lived within a network of speakers. This state of affairs accounts for why communal languages are assumed to be generally stable, changing very little during (parts of) the lives of some speakers, especially with regard to grammar and phonology (consistent with Dixon's 1997 claim of punctuated

Emergence of Population-wide Patterns equilibrium). For instance, the divergent chain shifts in the vowel systems of Northern and Southern White vernaculars as well as in AAVE, discussed respectively in Labov (1994) and Bailey & Thomas (1998), are unusual stochastic punctuations in the history of American English, just like the few cases of further grammaticization of invariant be (as in we be minding our own business) to express the PROGRESSIVE meaning, reported by Bailey & Maynor(1987). What should be emphasized here is that in language, changes are usually initiated by individuals. They spread to the communal level only after they have been copied by other speakers (not necessarily all of them), although the spread is determined by various aspects of the population structure, such as profession, level of education, age, gender, and socioeconomic class. Some of the changes spread rapidly and some take a longer time, depending on what advantages speakers find in adopting them, consistent with Bourdieu's (1991) notion of "linguistic capital." As far as traffic is concerned, it is in some cases simply the wisdom of not being too slow or too fast in the company of, conversely, fast or slow drivers that prompts drivers to adjust their speed. In the case of language, where the pressure of sounding like one's cohorts is often hard to resist, changes wind up largely being contained in a particular social group. Thus, rare are adult Americans who use thingy for thing or quotative like to report speech with some pretense of quoting a particular speaker. In both cases of traffic and language, the personality of the driver or the speaker matters. There are always individuals who are not afraid of being unique, regardless of whether or not they recruit followers. For instance, some particular professors or politicians are marked by their idiosyncratic vocabularies or syntactic constructions. However, the vast majority of speakers feel more comfortable remaining within the dominant norm. By the same token, there are also conventional-speed drivers in any traffic, and they in turn may keep the overall traffic from going too fast. These are the counterparts of those who resist changes in a language. They are not necessarily academy members who have claimed the prerogative of legislating on language, but the influential cautious speakers who are often last to catch up with changes or who, in their newspaper columns or some other fora, remind other speakers of what good usage is supposed to be. And there are always the older generations who stay conservative and different, just like old vehicles that either cannot pick up enough speed or can do so only at the risk of breaking down.

4.2.3

No cross-species analogs are perfect

No study of evolution in a language or in any species is complete without a discussion of speciation and contact-induced change. Both phenomena are also common in vehicle traffic, although the homologies here have serious limitations, consistent indeed with Mufwene's (2001a: 17, 193) observation that how a species evolves is in part a function of its ontogenetic properties and its species-specific modes of transmission.16

67

68

Language Evolution When two highways merge into one, their traffics do indeed merge in a way that illustrates some form of blending inheritance at the population level. While preserving their individual properties, their respective vehicles reposition themselves and adjust their speeds in ways that cause new patterns to emerge. The moment and point of encounter create a stochastic event that causes these adjustments. This change is more obvious when the new, merged traffic is to fork again into new streams that do not reproduce the pre-merger traffic bodies. This particular process is reminiscent of two populations speaking different languages that meet in the same territory. However, the traffic merger invoked here suggests only a rather unusual, if not unknown, situation where none of the members of the relevant populations is interested in learning the other language, although they interact with each other. That would be a situation where there is societal bilingualism without individual bilingualism, often mentioned in the literature on bilingualism but rarely illustrated with actual examples. The situation would be possible if, unlike what is suggested by the traffic merger, the populations remained isolated in their own respective neighborhoods and did not interact with each other at all.17 When populations mix in the way that traffics from two separate highways do, speakers interact with their neighbors and members of at least one group will eventually try to speak the other group's language. No such thing really happens in traffic, because vehicles do not have ontogenetic properties similar to those of idiolects. While traffics are osmotic, this property does not affect the ontogenetic structure of vehicles, unlike what is observable in languages, where both the communal and the idiolectal varieties are osmotic. Thus, although drivers accommodate each other during the merger, often repositioning their vehicles and adjusting their respective speeds to avoid collisions, none gives up their pre-merger vehicle to drive another one. Vehicles in traffic are like Darwinian organisms, with their mechanical makeups (analogous to genotypes) remaining the same from the time their drivers started moving them and/or got them into the traffic. On the other hand, idiolects are more like Lamarckian organisms whose structural features change several times during the lives of their speakers (Mufwene 200la, Chapters 1 and 6). The change occurs not only by substitution of features but also additively (by copying features from other idiolects and introducing intra-organism variation) or subtractively (by getting rid of some features and thus reducing intra-organism variation). This is a consequence of the accommodations speakers make to each other during their interactions.18 Moreover, speakers can use more than one language. Under various pragmatic constraints that need not be discussed here, they can alternate from one to another while the communal languages continue to be spoken in the settings in which they evolve. Drivers can only be in one vehicle at a time. They cannot alternate from one vehicle to another in the same traffic, not without abandoning the vehicle they step out of and causing problems to the

Emergence of Population-wide Patterns traffic flow. In contrast, bilingual speakers do not have to give up one language while speaking the other. In addition, the languages spoken by one speaker can influence each other, because a speaker can borrow materials from one language while speaking another. Such influence is of a different nature from speed adjustments that drivers make under the influence of others in traffic. I know of no counterpart of linguistic osmosis and mutual influences of idiolectal (sub)systems on each other in vehicle traffic. Such dissimilarities between the flow of traffic and language evolution are consequences of an important ontogenetic difference between a language as a population of idiolects and traffic as a population of moving vehicles. Languages are very much like viral species; they can copy features (analogs of genes) from each other and integrate them into a new whole. For instance, speakers who have been exposed to different dialects may have alternative pronunciations for the same word (e.g., [i:da\| ~ [aidcH for either, or [ru:t] ~ [rawt] for route), alternative terms for the same denotation (e.g., pail - bucket, coke ~ soft drink ~ soda, or rippled ~ grooved pavement), or alternative syntactic constructions (e.g., ask him not to come - ask him to not come or there are - there's - it's a lot of people there). Vehicles in traffic cannot alternate between competing mechanical properties, largely because the relation of a vehicle to its driver is not the same as that of an idiolect to its speaker. The driver "inherits" the vehicle, whose existence is independent of him/her, whereas the speaker is also the "creator" of his/her idiolect. The relation between him/her and the idiolect is symbiotic. Moreover, drivers on the highway have no experience comparable to that of language shift. This peculiarity accounts for an important difference between some consequences of language contact and those of traffic merger. As noted above, when two traffics merge into one, they literally blend into a new traffic in which vehicles from both gradually reposition themselves into the new one. However, when two languages or dialects come in contact with each other, they do not necessarily merge. One of them can gradually disappear from the scene, through the process of language shift, whereby one of the coexistent populations of speakers gives up their language for the language of the other population.19 Many immigrants to settlement colonies have lost their languages to that of the economically and politically dominant group. For instance, English has prevailed in the United States, in most of Canada, and in Australia at the expense of several European languages that either have been given up or are moribund now. The endangerment and loss of numerous indigenous languages in the same territories illustrate the same contact phenomenon, as explained in Part 3 of this book. The closest thing to such an evolution in vehicle traffic would be a situation never attested, though possible, in which drivers from one of the highways would gradually dispose of their vehicles and drive others coming from the other highway. One would have to explain how they dispose of their older vehicles (without inhibiting the traffic flow) and of drivers in the host vehicles, in ways that are not required to account for language shift. A language that one gives up simply vanishes from their repertoire of language varieties, by virtue of no longer

69

70

Language Evolution being spoken. If it is given up by all its pre-contact speakers and their offspring, it vanishes from the setting in which it used to compete with the language that has prevailed. Abandoned vehicles do not likewise vanish, though they are usually removed from the roadway. Also, languages that disappear often affect the prevailing one, with what has been identified as substrate, adstrate, or superstrate influence, depending on the contact setting. As explained in Chapter 3, several cases of language diversification, including those that have traditionally concerned both genetic linguistics and the development of Creoles, are due to language contact. In genetic linguistics, the development of Romance languages seems to illustrate this evolutionary pattern most incontrovertibly. More generally, the history of mankind, marked as it is by population movements and contacts, suggests that the actuation of the speciation of Indo-European and Bantu languages (among other language families), or of the development of Old English, lies in language contact. Linguists can also take advantage of the comparison with the flow of highway traffic to show that languages do not evolve in either unilinear or rectilinear fashions. Materials that affect the evolution of a language, or what used to be the same one, enter its evolution at different junctures, early or later in its evolution, before or after it has bifurcated into offspring varieties. This latter alternative makes it possible for a daughter variety to be influenced by a particular language or dialect while others do not undergo those particular influences. Thus, for instance, American English varieties have been subjected to influences from continental European and African languages in ways not experienced by British varieties. Even in the United States, these non-English influences have not been experienced in uniform ways from one colonial setting to another. African substrate influence has been stronger in coastal South Carolina and in Georgia, among Gullah speakers, than elsewhere, just like Scandinavian influence on White American English has been stronger in the Minnesota area than elsewhere. Differences in migration histories and in population structure become as important as ever to account for such diversity in the evolution of the same language. It is quite true that, as noted by Kurath (1928), dialectal differences are largely traceable to original patterns of migrations from England. However, one must remember that the English did not live in (total) isolation from nationals of other countries. Their interactions with those other populations influenced the ways in which English was being restructured as it was being appropriated by locally born children and other immigrants who would "pass it on" to their offspring. Thus, the evolutionary path of a language is like that of a highway traffic that is joined by other traffic and eventually bifurcates into more than one traffic stream, with the only difference (noted above) that languages do not really merge.20 If traffic did not have to flow from the start on a prebuilt highway in the context of this reflection, the comparison between language evolution and traffic flow would be more informative. Bifurcation is one of those respects that make the comparison difficult. Other than drivers wanting to go to different destinations (following pre-established paths), it is difficult to explain why highway traffic bifurcates. The apparent similarity to language

Emergence of Population-wide Patterns speciation in this case applies well especially to the envelope of the two evolutionary phenomena, viz., to the bifurcation or forking, and to the differential patterns that can emerge after the separation. Vehicle drivers have destinations and the topographical distribution of these destinations physically requires bifurcation in the interest of time and practicality. These considerations do not usually apply to language. Speakers contribute to the diversification of their languages unwittingly, despite their commitment to speaking the same language wherever they take it. Regional dialects correspond to language speciation only to the extent that the specifics of the processes of competition and selection vary from one geographical setting to another, depending in part on what particular languages have come in contact, the proportions of their respective speakers, and the kinds of interactions allowed by their population structures. Patterns of interaction are a significant component of the dynamics of language coexistence. However, speakers do not engage themselves consciously in the negotiations, accommodations, or innovations that bring about language change and divergence. The advantage of the comparison with vehicle traffic, rather than with a river that splits into a delta (Mufwene 1998), is useful in highlighting the way in which acts of individual drivers, like those of individual speakers, cumulate to produce, on the one hand, bifurcation of traffic and diversification in language evolution, and, on the other, different traffic and linguistic patterns. This is indeed very relevant to understanding Keller's (1994) adaptation of Smith's (1776) metaphor of the "invisible hand." We are thus also better prepared to address the subject matter of drift, "linguistic" first discussed by Sapir (1921). Note that he used the word drift quite differently from the usage of the term in biology for divergence that is not caused by natural selection. Sapir intended by drift no more than regular divergence caused by changes that a language has undergone in a particular community, i.e., when the way they speak it has become different from the way it is or used to be spoken by a related population or in an earlier community. He invoked it to account as much for the speciation of a protolanguage into subfamilies and individual languages as for diversification of a language into dialects. He invokes both internal and external factors to account for how it proceeds, although one must be surprised to see him rule out some linguistic evolutions attested in Creoles. For instance, he speculated that the development of the construction me see him from I see him would be a drift toward "horrors" (166), but it is attested in several Creoles, a function of the selection of tonic pronouns over their atonic counterparts (according to a normal principle in the assignment of markedness values). As is evident from Chapter 3, several linguists have followed Sapir's mistake by stipulating or suggesting that Creoles have not developed by the same natural processes that have produced language speciation. Yet a closer comparison between language evolution and the flow of traffic should at least prompt us to re-examine this position. The development of

71

72

Language Evolution Creoles cannot be excluded from regular cases of language drift, under particular colonial contact conditions. The process is in fact very similar in kind to the contact-induced speciation of Vulgar Latin into the Romance languages, the divergence of Old English from Old West Germanic, and later diversification of English, among several similar cases around the world. It is an instantiation of Sapir's own conception of how drift, or divergence, works. Also relevant to this discussion are some other remarks by Sapir. For instance, he observes that "What significant changes take place in [a language] must exist, to begin with, as individual variation" (155). He adds, ... it by no means follows that the general drift of a language can be understood from an exhaustive descriptive study of these variations alone. They themselves are random phenomena, like the waves of the sea, moving backward and forward in purposeless flux. The linguistic drift has direction. In other words, only those individual variations embody it or carry it which move in a certain direction, just as only certain wave movements in the bay outline the tide. This drift of a language is constituted by the unconscious selection on the part of its speakers of those individual variations that are cumulative in some special direction. (155)

Sapir certainly makes it obvious that a language is as heterogeneous as a species is and its evolution is a function of the acts of its speakers, though they do not all behave the same way. This is all consistent with the "invisible hand" metaphor. Sapir's use of "direction" must be interpreted in the sense of trend that emerges out of what evolves in individual idiolects, not as some evolutionary direction that has been planned by speakers. The elaborate comparison in this essay underscores the importance of understanding the dynamics of the coexistence of idiolects and how they influence each other, regardless of whether they are native or xenolectal, without any predetermined plan about what patterns of traffic flow are likely to emerge. The actuation of language evolution, including what Sapir identified as "drift," lies nowhere else but in the behaviors of individual speakers, whose primary concern, always in the present (without any foresight of how their language may be affected), is communication. In this context, Sapir's use of the term "unconscious selection" is similar to "natural selection" in biology, meaning prevalence out of a range of competitors. No speaker deliberately plans a particular restructuring process; the ecology of their language rolls the dice and patterns emerge spontaneously from changes which occur separately, consecutively, or concurrently. Evolution follows from the replacement of older patterns with new ones and/or the addition of new patterns to older ones.

4.3

Conclusions

The overall picture from the above considerations is that traffic is heterogeneous and does not flow at a uniform speed. The flow itself is subject to the interaction of various factors that largely reflect or affect the motions of individual vehicles. This is also true of language

Emergence of Population-wide Patterns evolution, which does not proceed at uniform pace and reflects the cumulative actions of individual speakers during their speech acts. Thus the evolution of a language is a function of how the individual evolutions of its idiolects balance out. Overall, the comparison between traffic flow on the highway and language evolution corroborates my position that a communal language is more like a biological species than like an organism. It exists only to the extent that there are coexistent idiolects whose speakers interact with each other and sometimes modify their "systems" either to meet new communicative needs or to accommodate other speakers. The focus on individual speakers, like on individual drivers on a highway, makes it possible to account for population contact and the various ways in which they coexist, depending on whether or not their individual members interact with, and influence, each other's behavior. Interaction and all it entails among the interactants is where the action that brings about change lies. That is where the "invisible hand" operates. One way or another, individual actions of drivers in traffic and speakers in a language community cumulate to produce what is later identified as evolution, i.e., the long-term changes that are observable in the behavior or characteristics of a species or, more generally, a population.

73

a

What Do Creoles and Pidgins Tell Us About the Evolution of Language?1 Chapter Outline 5.1 5.2 5.3 5.4 5.5

5.1

Introduction Why Creoles have not developed from pidgins Why Creoles were not made by children What Creoles do not tell us about the evolution of language in mankind What Creoles tell us about the evolution of language

74 75 78 83 88

Introduction

Bickerton (1990) and Givon (1998) claim that the development of Creoles and pidgins can provide insights about how language has evolved in mankind. This extrapolation has been encouraged by the position (disputed below) that Creoles were typically developed from erstwhile pidgins by children who transformed them from syntaxless protolinguistic means of communication to full-fledged languages (endowed with complex syntactic systems). Underlying this position in the literature which assumes that Creoles have pidgin ancestors is the unarticulated assumption that systems evolve from simpler to more complex structures. It has mattered very little that over the past few millennia the inflectional systems of many Indo-European languages (e.g., English and French) have likewise evolved from rich to poor ones, and their syntactic structures into increasingly analytical ones in which the position of syntactic constituents is critical to determining their functions. I argue that what little the development of Creoles and pidgins tells us about the evolution of language in mankind is largely not what has been claimed by Bickerton and Givon. It has to do with competition and selection during the said evolution, with how gradual the process was, and with how communal norms arise through the action of the "invisible hand." The histories of the development of Creoles and pidgins in, respectively, the European plantation and trade colonies of the seventeenth to nineteenth centuries present nothing that comes close to replicating the ecological conditions under which modern language emerged in the hominid-to-human phylogeny. Nor are there any conceivable parallels between, on the one hand, the early hominids' brains and minds that produced the protolanguage posited by Bickerton (1990, 2000) and Givon (1998) and, on the other, those of both the

Creoles, Pidgins, and the Evolution of Language modern adults who generated (incipient) pidgins and the modern children who produce child language, even if one subscribes to the ontogeny-recapitulates-phylogeny thesis. To be sure, Givon (1998) makes some correct observations regarding gradualness in the evolution of language, the coevolution of language and the cognitive infrastructure necessary to carry it, and the centrality or primacy of some aspects of language. These are precisely some of the hypotheses defended by Li (2002a, 2002b) and Slobin (2002), to which I also subscribe. Relying largely on my own longitudinal study of my daughter's child language (Mufwene 1999a), I capitalize both on Tomasello's (2002) "cut and paste" model of language acquisition, which suggests that learners develop the grammars of their idiolects incrementally, and on Slobin's observation that the order in which child language develops is largely also influenced by the kinds of primary linguistic data (PLD) to which the learners have been exposed. My arguments regarding Creoles corroborate Slobin's other observation that where a full-fledged language is already in usage, children (at child-language stage) are typically not the innovators of the new forms and structures that spread in the language of a population. However, I also agree with DeGraff (1999a, 1999b) that they contributed to the development of Creole vernaculars by selecting some of the adults' innovations (often associated with substrate influence), just like any other features that become part of their idiolects, thus making them available to future learners (Chapter 10).

5.2 Why Creoles have not developed from pidgins Most of the arguments summarized below from Mufwene (200la, 2005a) and Chapter 3 are intended to provide a notional background to the discussions in the following sections.2 I will elaborate only things that are not discussed in these references, starting with the following. It is surprising that the pidgin-to-creole evolutionary scenario has hardly been disputed from Schuchardt (1914), Jespersen (1922), and Bloomfield (1933) to the present day. A simple look at the geographical distribution of our heuristic prototypes of Creoles and pidgins (Map 1, Chapter 3) suggests already that the alleged pidgin ancestry of Creoles is questionable at best. Most pidgins are concentrated on the Atlantic coast of the African mainland and on Pacific islands, whereas most of the Creoles that have evolved from European languages, the only ones that matter for our purposes, are concentrated on Atlantic and Indian Ocean islands (including places such as Cape Verde and Sao Tome) and on the Atlantic coast of the Americas.3 They are the only ones that are associated with the earlier application of the term Creole as a noun or as an adjective to non-linguistic entities, although we must remember that Creole populations have not necessarily spoken Creole. For instance, the Bekes (or White Creoles) of Martinique do not consider themselves as Creole speakers, neither do the Louisianan White Creoles.

7 5

76

Language Evolution The European colonization of the coast of Africa and of the Pacific islands started on the trade model, characterized initially by egalitarian and sporadic contacts with the Natives, whose exposure to the European trade languages was limited. Rather than anything having to do inherently with adult L2 learning, the sporadicity of the contacts is the primary reason why incipient pidgins have been characterized as "broken." It is worth pointing out that the initial contacts of Europeans and the Natives depended on a handful of non-European interpreters, who spoke non-pidgin varieties (Mufwene 2005a). As the contacts increased, more and more non-Europeans who had no training and no access to interpreters would attempt to speak the trade language. As the number of such speakers grew, the structures of the trade languages became more divergent from the relevant European vernaculars and apparently more "broken."4 The direction of the divergence is thus similar to that of the basilectalization process associated with the emergence of Creoles, as shown below. On the other hand, as the pidgins' communicative functions increased (such as in the cities that emerged from erstwhile trade factories), these "contact varieties" became structurally more complex, and regularity of use gave them more stability. These additional characteristics changed them into what is known as expanded pidgins, like Tok Pisin and Nigerian Pidgin English, which for some speakers function also as vernaculars, rather than as lingua francas only. History suggests that children had no privileged role to play in this structural expansion (see, e.g., Miihlhausler 1997 in the case of Tok Pisin), though they certainly helped vernacularize the varieties. The evolution of some pidgins in Africa and the Pacific into expanded pidgins bears no consequence on Chaudenson's (1992,2001,2003) position and mine on the development of Creoles, as becomes obvious below. Creoles have typically developed in plantation settlement colonies, in which nonEuropeans formed the majority of their respective populations (Chaudenson 2001, 2003; Mufwene 200la, 2005a). Although some non-plantation settlement colonies, such as those of North America and Australia, also developed with European majorities that were nonnative speakers of English or French (which count among the languages that produced "classic Creoles and pidgins"), their new nonstandard vernaculars have not been identified as Creoles. As shown in Chapter 3, this is a distinction that has to do more with a social bias in the practice of genetic linguistics and creolistics than with actual differences in the structural processes that generated new Creole and non-creole varieties of European languages outside Europe (Mufwene 2001a and Chapter 6). In any case, Creoles developed in those settings where interactions between Europeans and non-Europeans had been regular during the initial, homestead phase of the colonies. Communication in almost all cultural domains was then (intended) in the European language, since, on average, nonnative speakers hardly had anybody else to speak their own ethnic languages with. If one were the only non-European in a homestead relatively isolated from others, there was nobody else to speak his/her language with. If there was another slave in the homestead or in a neighboring one, he or she probably did not speak the same language.

Creoles, Pidgins, and the Evolution of Language If they had a common language, they probably did not interact regularly enough to maintain and pass it on to children they could have had. Or the children probably did not find the command of such non-European languages particularly advantageous to their daily lives. The earliest vernaculars commonly spoken and appropriated by non-Europeans were approximations of the European colonial languages forged in part also by the nonnative European indentured servants with whom the slaves interacted regularly. All Creole children who grew up in the same homestead and spent their days together while their parents were at work spoke alike, regardless of how their parents spoke the local vernacular (Chaudenson 1992, 2001, 2003). The experience of such children would not have been different from that of Black middle-class kids growing up in integrated neighborhoods in American cities today, whose linguistic features typically reflect those of the larger community rather than those of their parents, especially if these are immigrants. It is during the plantation phase that the local European language gradually evolved into a different variety among the non-Europeans. The population majorities consisted then not only of non-Europeans but also of nonnative speakers (Baker 1993a), thanks to rapid population turnovers and to demographic growth made possible by importations rather than by birth. Although segregation played a role in fostering the divergence of speech varieties of Europeans and non-Europeans, the increasing demographic dominance of nonnative speakers among non-Europeans communicating primarily among themselves in the new vernacular also favored a greater role of non-European substrate influence. Bickerton (1988) agrees with part of the above position, as he admits that Creoles did not develop from erstwhile pidgins and that their basilects developed later than their mesolects. The reason is that the intimate living conditions shared by Europeans and non-Europeans alike during the homestead phase of settlement colonies made no allowance for the development of pidgins as structurally reduced language varieties associated with sporadic contacts. As we rethink the colonial history of the New World in particular, note that the Europeans colonized it in two concurrent ways. All along the Atlantic coast and on its barrier islands - including the Caribbean, Bermuda, and the Bahamas - they developed settlement colonies on land concessions which they would eventually expand at the expense of Native Americans. At the same time, the Europeans also developed trade relations with the Natives, before they eventually drove them westwards and into reservations and eventually absorbed large proportions of the survivors of this invasion into new, European-styled socioeconomic systems, especially since the nineteenth century. Pidgins in the Americas developed from those originally sporadic trade contacts between Europeans and nonEuropeans (Prudent 1980; Chaudenson 2001, 2003). We just must address the enigma of why "classic pidgins" in Africa and the Pacific developed from European languages but their counterparts in the Americas did from indigenous ones, for instance, Chinookan for Chinook Jargon, Delaware for Delaware Pidgin, Choctaw for Mobilian Jargon, and Tupi for Liengua Geral.5

77

78

Language Evolution Structural similarities between expanded pidgins and Creoles reflect the fact that they were developed largely by linguistic adults interacting regularly among themselves, using materials from typologically related European and/or substrate languages to meet diverse and complex communicative needs, and thus needing complex grammatical structures. As noted in Chapter 3, substrate influence seems to have been greater in colonies that Chaudenson (1979-2001) identifies as endogenous and where there was relatively more ethnolinguistic homogeneity in the substrate population (Sankoff & Brown 1976, Sankoff 1984, Mufwene 1986a, Keesing 1988, Singler 1988). It was less significant in exogenous colonies, where the continuous critical mass of speakers of closer approximations of the European languages offset the impact of substrate influence, despite the increasing role of the bozal slaves in the basilectalization process. Both Creoles and pidgins developed gradually, from closer approximations of the initial targets to varieties more and more different from them. As pointed out by DeGraff (1999a), they are creations no more of children exclusively than they are of adult L2 learners exclusively. Actually, substrate influence in Creoles would be difficult to account for if the role of adult nonnative speakers as carriers of xenolectal features were not factored in our hypotheses. The role of children in the development of Creoles involved selecting some of those substrate features into their idiolects and making them available to future learners. (See also Chapter 10.)

53

Why Creoles were not made by children

Kegl & McWhorter (1997), Kegl et al (1999), and Goldin-Meadow (2002) argue convincingly that children could develop elaborate sign languages. The fact that, in the case of Nicaraguan Sign Language (NSL), the new "system" is largely a systematization of materials that were already available to the children in their respective native communities (Senghas & Coppola 2001, Kegl et al 1999, Morford 2002) reduces nothing of the important role that children played in elaborating a long-lasting communicative system. In this respect, they can very well be compared to our hominid ancestors at various stages of the evolution of mankind, who would develop a more elaborate and systematic communicative means from what had been used by the earlier generation of hominids. Goldin-Meadow's evidence can even be used to argue that systematicness developed not because speakers/signers had to communicate with each other but because they had to be individually consistent (Mufwene 1989a). Among the relevant questions at the population level are the following two: How do communal norms develop? And does the development of communal norms entail the elimination of interindividual variation? The available evidence from child language and Creoles militates for a negative answer to the second question. MacWhinney (2002a: 254) even observes that "we should not be surprised to find large individual differences in the neuronal basis of higher-level dynamic control of language."

Creoles, Pidgins, and the Evolution of Language In any case, the evidence from home sign language and NSL is not comparable to that of the development of Creoles and pidgins. More interesting about incipient pidgins is the fact that they represent simplifications, reductive developments of some sort, from full-fledged languages. Evolutionarily, they have evolved in the opposite direction of the putative protolanguage, which started from non-linguistic means of communication.6 At best, what they teach us about the evolution of language is that not all structural components of modern linguistic systems are deeply entrenched, at least not to the same extent of time depth. Those morphosyntactic components that survive the "breakdown," so to speak, which produces incipient pidgins may be the most deeply entrenched in the architecture of language. The same is true of these incipient varieties' heavy dependence on the pragmatic context for the interpretation of utterances, as highlighted by the language disorder cases discussed by MacWhinney (2002a). The development of more complex structures would thus have streamlined the interpretation of utterances and reduced dependence on non-linguistic context. MacWhinney (2002a: 250-1) also argues that the ability to use articulate sounds to communicate linguistically developed between 200,000 and 50,000 years ago, and thus brought the increase of brain size in hominids to its full communicative potential, such as to develop a larger lexicon and, later on, to combine words into larger strings.7 Interestingly, the phonemic systems of incipient pidgins reflect mostly interference from languages previously spoken by the speakers. They do not reveal the kind of "breakdown" evidenced by the morphosyntactic components, which dispenses with some lexical and grammatical categories. Thus incipient pidgins support the hypothesis that the ability to produce phonemic sounds is one of the most deeply entrenched components of spoken languages. Regarding Creoles, we must recall that it has all along been misguided to define Creoles as nativized pidgins. Neither the geographical distribution of Creoles and pidgins nor the respective socioeconomic histories of the territories where Creoles developed support this position. Assuming as in Chapter 3 that contact has played a central role in the evolution of, say, Indo-European languages outside and within Europe, Creoles developed by the same normal restructuring processes, although the role of language contact must have been made more obvious by the non-European composition of the populations appropriating the European languages (Chapter 6). Below, I adduce structural arguments against the position that Creoles were made by children, although, in communities where adult L2 speakers and children use the same vernaculars, children produce utterances that apparently vary less in their structures from one speaker to another (p.c., Givon, 11 June 2001). Senghas & Coppola (2001) show that the signers exposed to NSL as children (below ten years of age) sign more systematically, uniformly, and fluently than adult signers. Children did indeed play a non-negligible role in the development of these new vernaculars, but it was not that of creating a grammar where their parents would presumably have failed. It is not true that incipient pidgins have no grammars, although these are

79

80

Language Evolution internally variable.8 Rather than creating new grammars for the overall community, children participated in the development of Creoles by selecting from the feature pools to which they were exposed particular subsets of features (including xenolectal ones) in manners that made their idiolects less different from each other than their nonnative parents'. Functioning thus as agents of normalization, they helped those features prevail over other alternatives, reducing the extent of variation observable in their xenolectal parents (DeGraff 1999a, 1999b). Creole children did this in the same way children everywhere normally contribute to changes and to divergence from their parents' language varieties, minimally in most cases except in exogenous settings, where the model speakers are the native speaking population if they have full access to it.9 Creole settings made variable access to native speakers of the colonial vernacular. First language development is a protracted process that reaches maturity after the speaker passes the critical period, i.e., by the time the speaker may be considered linguistically adult. Structures of English Creoles are not identical with those of English child language, despite some similarities between them. For instance, English-speaking children do not produce the kinds of serial verb constructions attested in Saramaccan or Sranan, for example, where a serial give conveys a DATIVE function, as in Feti gi mi kondre "Fight for my country" (lit., "fight give my country") in Sranan (Sebba 1987: 50). Nor do they distinguish between the kinds of aspectual nuances attested in Gullah and Guyanese Creole, in which the preverbal marker don(e) (from nonstandard/colloquial English PERFECT verb done "finish" rather than from the standard English past participle of do) conveys a different PERFECT meaning from the postverbal marker don, at least to those who use both constructions. While me done talk [mi dAn/dOn tO:k/ta:k] simply means "I have spoken," me talk done means more than that in these Creoles, viz., "I have said all I had to say and don't intend to talk again." Aside from the PROGRESSIVE construction with preverbal de, as in mi de taak "I am talking," Guyanese Creole also has a more specific, composite construction with preverbal de a, as in mi de a taak "I am busy talking." No contrasts similar to these have been documented in English child language. Also, almost all Atlantic Creoles have Predicate-Clefting, as in duh [da] talk he duh talk, illustrated here from Gullah and meaning "he is/was really talking." This can also be heard in nominalizations such as you shoulda hear da [da] talk he duh talk "you should have heard how he was talking." Like regular cleft constructions in English, PredicateClefting is produced by speakers whose linguistic competence is beyond child language, certainly past age three, beyond the period that seems to have concerned Bickerton (1990). If children innovated these structures in Creoles, then they must have innovated them after they had become linguistically adults. We must thus wonder why their adult parents would have waited for them (the children) to innovate for the community when they (the parents) could have done so themselves. The answer can of course be similar to that provided by Judy Kegl and her associates about the role of children in the development of NSL, on which I comment above. However, the fact that only English creole-speaking children,

Creoles, Pidgins, and the Evolution of Language but not their Anglophone counterparts, acquire these distinctions by the end of the critical period is a reflection of the influence that adult speech exerts on child-language development. It also shows that, by the principle of least effort, Creole children, like children everywhere, develop their idiolects from the PLD available to them from adult speech, even if this happens horizontally through the mediation of other children. We could also extrapolate that where the PLD lead to variable "systems," children will also make allowance for variation in their idiolects. This extrapolation explains why the variation mischaracterized as "(post-) Creole continuum" was not eradicated by children who participated in the development of Creoles, any more than any variation, or speech continuum, in any other language community would have been eliminated by children. Thus, we should not confuse variation in the structures produced by adult L2 speakers with the alleged inability on their part to develop a grammar or to innovate in the direction of a new full-fledged vernacular. There is no compelling evidence for the conjecture that Creoles owe to creations by children the structures that distinguish them from the European languages from which they have evolved. The case is even less convincing where the innovations can be related to features of substrate languages. Bickerton's (1981, 1999) language bioprogram hypothesis (LBH) precludes children from transferring elements of substrate languages into the emergent Creoles, because they had no prior knowledge of a language before the one they are misguidedly claimed to have created for their communities. The fact that not all Creoles have the particular constructions discussed here also suggests that, in the first place, there is no particular, uniform Creole grammar that is replicated from one Creole to another. It also shows clearly how futile it is to invoke children as the primary or exclusive makers of Creoles - at least not at the child-language stage - because these vernaculars contain some structures that have not been attested in child-language versions of the relevant European languages. Most of the arguments for claiming that Creoles were developed by children have had to do with the poverty, or absence, of inflections in these vernaculars. First, as argued by Slobin (2002), whether or not child language lacks inflections depends on what the target language is and how significant the role of inflections is in it. As Chaudenson (1992, 2001, 2003) points out, aside from the fact that the relevant European lexifiers are not so rich in inflections, they are also marked by inconsistencies that make them difficult for L2 learners to acquire. On the other hand, DeGraff (200la, 200Ib) shows that Creoles are not as deprived of inflections as has been claimed. In the very least, the old myth is not true of Haitian Creole, which displays some, both inherited from French and innovated thanks to patterns emerging from the new "system." What all these observations point to is that like language development among children, the development of Creoles is subject to structural and ethnographic factors in the relevant linguistic communities. My daughter's child language (Mufwene 1999a) suggests that the kinds of over-generalizations from regular morphological patterns (such as goed for went

81

82

Language Evolution and failed for fell) that occur when English-speaking children are three to four years old are transitional. At a younger age, my daughter had fell in contrast with falling and went in contrast with going, though there was no evidence that the pairs were grammatically related. While acquiring negation before age three, my daughter also produced didn fell, didn took, didn saw, and didn woke, instead of didn't fall, didnt take, didn t see, and didn'twake, and she abandoned these deviations as soon as she became aware that adults around her do not typically use these PAST TENSE forms with didnt. Note, however, that the origins of the forms and constructions themselves are in adult speech, consistent with the Tomasello (2002) "cut and paste" model of LI learning assumed here. From a developmental perspective, children are more conformist and imitative than has been suggested, or claimed, in some child-language literature, although they do indeed construct gradually the grammars of their idiolects by inference from the PLD that are accessible to them. They abandon their deviations quickly to conform to adult norms, including variation within those norms, as is evident from language communities anywhere, creole and non-creole alike. Given the way plantation settlement societies developed, we have no evidence for assuming that, linguistically and socially, slave children behaved differently from other children and did not just learn the vernaculars around them, as emergent as these were, especially if, as suggested by history, their foreign-born parents did not speak pidgins, influenced as their varieties must have been by substrate elements. Creole children must have simply contributed to the normalization of the plantation vernacular. They determined (albeit non-deliberately, through the population-level distribution of their individual selections) how much xenolectal element would become part of the systems these vernaculars now have. The children never were the majority, nor did they form sub-communities of their own that were isolated from adult communities. There is no particular non-structural, ethnographic reason why they would have imposed norms of their own over those of the adults around them. They mostly perpetuated variants of the vernaculars that were already normalized or normalizing around them. We must recall that creole children were not in situations similar to those of Nicaraguan deaf children, because there always was a full-fledged language of the same modality around them, regardless of the extent of population-level variation in it. According to Mufwene (200la), children actually slowed down the basilectalization process during the development of Creoles. This is because during the most prosperous period of the plantation colonies (i.e., the eighteenth century in the Atlantic and Indian Ocean colonies, and the nineteenth century in the Pacific), their populations grew more by importation than by birth. While the children "acquired" natively the local vernaculars that they heard around them, adult nonnative speakers continued to restructure them, influencing them with their xenolectal features. Children thus kept some xenolectal features from spreading to the next generation of learners, at least in reducing their statistical frequencies in the PLD accessible to the latter. In any case, there is still a lot to be learned about the

Creoles, Pidgins, and the Evolution of Language dynamics of competition and selection among structural features of the European and substrate languages in the plantation settlement colonies (Chapter 7). Still, we must remember that in this respect the difference between plantation societies and other communities is only a matter of degree. Variation is everywhere. Nowhere else have children been led by the circumstances to create a new language which would replace their parents' variable system. Slobin (2002) is correct again in holding adults as the primary innovators that should matter in language change and the development of Creoles. Children play an undeniable role in helping determine which of those innovations become part of the communal language and which of the extant variants become recessive and may eventually die out of the ever-evolving language. So far we have no reason for assuming that the development of Creoles is not the outcome of normal language change and diversification (Mufwene 200la). Thus, there is no empirical support for the LBH or the way Bickerton (1981ff) hypothesized Universal Grammar (UG) to work in the development of Creoles, crediting children almost exclusively with the emergence of their grammars.

5.4 What Creoles do not tell us about the evolution of language in mankind Bickerton's (1990) arguments that pidgins and child language can give us insights about how language evolved in mankind are partly based on the assumption that ontogeny recapitulates phylogeny. They are also based on his mistaken hypothesis that Creoles were made by children from erstwhile pidgins. As explained above, his position both lacks the support it would need from the socioeconomic histories of creole-speaking societies and the logical justification that would make it plausible in the first place.101 focus here on the ontogenyrecapitulates-phylogeny aspect of Bickerton's arguments and derive my support largely from Givon (1998), Li (2002a, 2002b), and MacWhinney (2002b). We must start with the fact that Creoles' grammars show no evidence of having started from scratch or of having developed according to typologically unmarked parametric settings of UG (cf. Bickerton 1981-1999). If these vernaculars had been produced by children, independent of the PLD available in their alleged syntaxless pidgin ancestors or, more realistically, independent of the data available in the colonial varieties of the European languages from which they have evolved, their grammars would not vary at all from one Creole to another.11 Creoles' grammars do indeed vary, reflecting the extent of inheritance from the European language and/or that of substrate influence. Contrary to what has been suggested at least by the earliest versions of Bickerton's LBH (1981,1984b), speakers do not develop grammars of their language varieties independent of the "acquisition" of their vocabularies. As suggested by Bolinger (1973), grammars are largely generalizations over the behaviors of individual lexical items. According to Wang (2007:291),

83

84

Language Evolution "most working sentences are cut-and-pastes of prefabricated parts," which need not be accounted for with the kinds of rules linguists claim grammars to consist of. As a matter of fact, I submit that grammars are more projections on the part of the linguist, in their attempt to account for how we communicate (as we infer the existence of systems consisting of units and of combinatoric rules) than they exist of necessity. (See also Le Page & Tabouret-Keller 1985: 332.) The connection between, on the one hand, Creoles and, on the other, the relevant European and substrate languages is unmistakable once one compares them not so much with the standard varieties of European languages but with the nonstandard vernaculars actually spoken by the European yeomen and indentured servants with whom the non-European labor interacted regularly. Alternately, Creoles' structures should be compared with those of the other vernaculars that have evolved from the koines spoken by the proletarian European settlers, the typical founder populations of European colonies, with whom non-Europeans lived fairly intimately during the homestead phase. One is hard-pressed to find in Creoles any grammatical features that have not been selected from the nonstandard varieties of the relevant European vernaculars or in their substrate languages, although these have not been replicated faithfully.12 There is in nonstandard English evidence of most of the features associated with English Creoles, for instance, NOMINAL PLURAL with them (as in dem boys "the boys"), COPULA absence (as in Mary sick/home "Mary [is] sick/home"), periphrastic marking of TENSE-ASPECT (viz., unstressed HABITUAL do [da] and does [daz], stressed REMOTE PHASE been [bin] - though this occurs typically with a contracted have - CONTINUATIVE do/duh [da],13 PERFECT done (as in Faye done gone/go "Faye has gone/left"), invariant relativization with what (as in everything what he say "everything that he said") or a null complementizer (as You say Faye done gone "You said Faye had left"), and reported speech introduced by say (as in / hear say Faye done gone "I heard that Faye had/has left"). Not only have all these markers been selected from the nonstandard English varieties to which the slaves were exposed, they also have similar functions in these "lexifiers." For instance, one often hears dem boys "those boys." although the meaning is both PLURAL and DISTAL. As explained in Mufwene (1996b), quotative say is commonly used in nonstandard English, although Creoles have extended its distribution to indirect reported speech. What is also attested as an invariant relative pronoun, as much as HABITUAL does in some nonstandard British English dialects. And there is similar evidence in Romance Creoles (Chaudenson 1992,2001,2003). Even serial verb constructions have partial models in the relevant European languages. (See, for instance, Pullum 1990 regarding how they work in English.) Unfortunately the literature on the development of Creoles has ignored all the structural features that most Creoles of the Atlantic and Indian Oceans share normally with their "lexifiers," e.g., the SVO pattern of clauses, the fact that the determiner and adjective precede the head noun, and the adjective

Creoles, Pidgins, and the Evolution of Language follows the determiner, that a relative clause starts with one of the relativizers in the lexifier, that preposition-stranding in English Creoles reflects the dominance of this strategy over pied-piping in vernacular English, and the comparative with MORE (however over-generalized it may be) is also the unmarked option in the same vernaculars. What has been modified from the European languages' grammars may have been exaggerated compared to what has been preserved in the relevant Creoles. The literature has also been misguided in lumping the Melanesian expanded pidgins in the same bag as these Creoles. The contact ecologies that produced the latter varieties were not identical to those of the former. As clearly shown by Keesing (1988), the Melanesian varieties evolved in ecologies that favored substrate influence. See also Faraclas (1988b), Sankoff (1984,1993), and Sankoff & Brown (1976) on the development of Tok Pisin. The key to understanding why Creoles are different from their non-creole kin that evolved from similar colonial varieties of European languages is that, as explained in Mufwene 200la (see also Chapter 10), "language acquisition" is a reconstruction process, which is sensitive to the variants in competition in the pool of features available to individual learners. The contact ecology of the appropriation of the European languages varied from one colony to another and from one period to another, which accounts also for why each Creole is somewhat different from others that developed from what has been identified, for convenience sake, as more or less the same language. Moreover, linguistic features often get modified during the "acquisition" process, especially during "L2 acquisition," as every learner analyzes the PLD from which they develop their idiolects on their own, without particular explicit training. In the case of the development of Creoles, congruence between structural features of the relevant European and substrate languages has been a critical factor, as clearly articulated recently in Corne (1999), Chaudenson (2001, 2003), Mufwene (2001a, 2005a; see also Chapters 7 and 8), Siegel (2006), and Aboh (2007). Thus, as also argued by Chaudenson (1992,2001,2003) and Mufwene (2001a, 2005a), Creoles have developed by gradual restructuring away from structures of the earliest colonial koines of the relevant European languages in the direction of their basilects, which are actually the latest to have evolved. None of this evolutionary process is remotely suggestive of how language evolved in mankind, originally from prelinguistic means of communication to protolanguage, concurrently with changes in the structure of human brain (Li 2002a, 2002b; MacWhinney 2002a). The relation of the development of Creoles to "language acquisition" deserves more discussion. We must first of all dismiss the myth that Creoles have diverged from the relevant varieties of European languages because there was a break in the "transmission" of the latter to non-European groups on the plantations (e.g., Polome 1983a, Thomason & Kaufman 1988). As explained in Chapter 3, even the lexicon of Creoles would not have been inherited so predominantly from the relevant European languages (at least 90% on average) if there

85

86

Language Evolution had been a break in the "transmission" of the latter. And one could not possibly imagine a group learning the vocabulary of a language, even under the conditions of sporadic contacts that produced pidgins, without also learning some of the grammar associated with it.14 Admitting substrate influence does not entail ignoring numerous basic structural similarities (not due to any universals) which obtain between Creoles and the relevant European languages from which they developed. The following randomly cited features will suffice to illustrate my point: (1) the phonologies of Creoles largely reflect how the colonial varieties of the relevant European languages were spoken in the seventeenth and eighteenth centuries (e.g., the palatalized pronunciation of cat as [kyat] and pear as [pys] in Jamaican Creole); (2) the extensive use of adjectives and their prenominal use in English Creoles (as noted above) reflect on English structures;15 (3) English Creoles have definite articles where most substrate languages use a distal demonstrative;16 (4) the postposed determiner la in Haitian Creole does not lack affinity with a morpheme which has similar morphosyntax, semantics, and pragmatics in nonstandard colloquial French varieties; and (5) the fact that only (Atlantic) English Creoles use a form evolved from the verb say as a complementizer must have something to do with the fact that colloquial and nonstandard English has more uses of say to report speech quotatively than French (Mufwene 1996b), etc. An important question to recall here is: Does SUBSTRATE INFLUENCE mean the same thing as ORIGIN of a structural feature? (See Chapter 8.) Language acquisition is of course imperfect. As noted by Lass (1997), imperfect replication is a normal condition in "language acquisition." Otherwise, there would be no language change in the first place, regardless of whether it is claimed to be internally or externally motivated. Even in ethnographic ecologies where no contact of significantly different dialects or separate languages is involved, a language or dialect is usually appropriated by other speakers with minor modifications. Most of these coincide with variants already available in the communal language or dialect. However, sometimes new ones creep in, and/or the dynamics of the coexistence of variants in the population of speakers may change in such ways that some of the variants become stronger and may even drive others out. Such accretions of modifications are what linguists later identify as language change. It is also useful to underscore the fact that "language transmission" and "language acquisition" are convenient misnomers for processes that are much more complex (DeGraff 1999a, 1999b, 2001a; Mufwene 2001a, Chapter 10). Speakers of a language provide only the PLD to the learners. Nobody ever transmits an integrated linguistic "system" to other speakers, and no speaker ever passively inherits such a ready-made idiolect from previous speakers. "Acquisition" as a reconstruction process proceeds piecemeal, with the language learner developing competence in the target language only gradually. There is indeed some language-building activity on the part of the learner, though it is not clear how the construction proceeds. It is not evident that the learner is focused on developing a "system" and is not just trying to communicate successfully. This does not, of course, rule out their ability

Creoles, Pidgins, and the Evolution of Language to capitalize with analogies to what they have already learned to say or heard other speakers say. Patterns, which linguists like to call rules, would just emerge from the learner's effort to speak systematically. Also, the patterns need not emerge as generalizations of the kind linguists describe as "significant," covering more or less exhaustively classes relevant to a linguistic analysis. Rather, they can be, and probably are, just partial and overlapping (McCawley 1976) and expanding gradually in the way Lakoff (1987) describes the development of metaphors from basic uses. In the case of a child, misperception or inaccurate analogy with something previously learned may account for deviations. In the case of an adult learner, aside from these particular reasons, previously established linguistic habits (i.e., xenolectal influence) also account for such deviations. One thing is certain, learners try to speak like those speakers of the language that they target, but they are not targeting a grammar or "system" in the way that a linguist would do by collecting a body of data, analyzing them, and producing a hypothesis of the system that can be inferred. McCawley's (1976) observation that a child should not be thought of as a mini-linguist can be generalized to say that a naturalistic language learner is not a linguist, especially since they approach their communicative challenges bit by bit, without waiting until enough data have been collected. The process of generalizing from previously learned cases appears to be simply analogical. Tomasello (2002) provides just the right kind of language development framework to account for this natural phenomenon of deviation or divergence from the target. His approach makes it possible for us to realize that the difference between deviations in LI development and deviations in L2 development lies not in how these deviations happen but in the additional causes for the deviations in L2 development. Naturally the additional causes increase the potential for deviations and produce in part what is known as nonnative accent. Otherwise, we see in both cases a confirmation of Meillet's (1929) and Hagege's (1993) observation that language "acquisition" involves both inheritance from the target and recreation by the learner. Aside from the restructuring cum recombination process discussed in Mufwene (200la and Chapter 2), the recreations involve innovations by the learner, regardless of whether they are made possible by analogies perceived in the target language itself or are caused by knowledge of another language. It is irrelevant whether at the communal level such innovations produce new features and divergences from the target communal language. When they do, we say that a language has evolved into another state, such as from Old to Middle English, or from English to English Creoles. As Creoles appear to be normal outcomes of language appropriation by new populations under contact conditions in which substrate influence applies, Bickerton's (1990, 2000) claim that they can inform us about how language has evolved in mankind is not justified. If they do, it must be in respect to gradualness, competition, and selection among variants, and the development of norms in populations of speakers. I turn to these determinative factors of evolution in the next section.

87

OO

Language Evolution

5.5 What Creoles tell us about the evolution of language We can now return to Givon's (1998) thesis that the evolution of language in mankind is an adaptive process. In child language, the development of Creoles, and other cases of language change, semantic and morphosyntactic innovations are especially exaptive, responding to (new) communicative needs of speakers beyond the meanings and uses they have already "acquired" of particular expressions. One thing that is especially noteworthy about Creoles is that, despite their divergence from their non-creole kin, they have preserved both several features of the relevant European languages and complexities similar to those attested in several other languages spoken by modern humans. If Creoles had really developed from erstwhile pidgins, progressing from simpler to more complex grammatical structures, their development would share with the evolution of language as hypothesized by Givon the fact that every later stage exapted materials in the earlier stage. The closest analog to this evolution would be home sign language starting in part from the gestures used by their speaking parents and innovating on their own, gradually developing a communicative "system" with rudimentary grammar. However, this observation remains guarded, because Goldin-Meadow (2002) does not discuss whether such home sign language had developed into adult language. Neither does she discuss it as a populationlevel process similar to the NSL case, in which one can observe the development of a communal norm. One particular characteristic that all the above cases of language development and evolution share is that they are all gradual processes. Givon (1998: 105) submits that "human language ... arose from the co-evolution of cognitive, neurological, communicative and sociocultural patterns of pre-human hominids." Complementing this, Li (2002a, 2002b) and MacWhinney (2002a) hypothesize that (proto)language evolved from gestural means of communication (about 5-6 million years ago) to vocalizations and eventually to phonetic linguistic systems (between 200,000 and 50,000 years ago), concurrently with the emergence of the specific physiological and mental infrastructures required to support the complexthought-processing capacity required to manage and process modern human languages. Perhaps McNeill (2005, Chapter 8) accounts for it accurately, arguing that the same Broca's areas 44 and 45 were co-opted for coordinated gestures and vocalizations as part of language and for processing the meanings associated with them. Gestures need not have emerged before speech and the whole evolutionary process was protracted, contrary to Bickerton's (1990) conjecture of an abrupt, "big bang mutation."17 LI development, from child to adult grammar, is not only gradual but also correlated with cognitive maturation/sophistication. The development of Creoles and pidgins is similar only in being a gradual process, not at all in allegedly evolving from simpler to more

Creoles, Pidgins, and the Evolution of Language complex morphosyntactic structures. Note that the problem pointed out by Slobin (2002) and Li (2002a, 2002b) with Bickertons (1990) hypothesis that early child language provides a window into the evolution of language in mankind remains, viz., that human infants today are born both with a brain infrastructure and in language ecologies that already distinguish them from our hominid ancestors. Therefore the ontogeny-recapitulates-phylogeny assumption does not apply at all. Modern children are typically born to social environments in which full-fledged languages are being spoken, unlike those settings where hominid and modern human ancestors would be innovating individually and negotiating tacitly for gestures, symbols, and expressions that would become part of modern languages, by the basic processes of competition and selection discussed below and in Chapter 7. Studying LI development while children socialize among themselves may be quite informative about how they influence each other's development. Otherwise, the order in which children develop competence in their native languages, starting with basic and simple structures, reflects the maturation of their cognitive capacities, similar to how phylogenetic changes in the physiology and brain structure of the human species predisposed them to develop more complex communicative systems. Although language "acquisition" is a reconstruction process, modern children cannot be credited with inventing a language in the same way that our hominid ancestors gradually invented language. If it is true that an individual's genotype determines his/her biological life trajectory, then, by the natural selection process that favored the modern human over other hominid alternatives, modern infants are born prewired to "acquire" the modern languages of the communities to which they are born. In this respect, Creole children are not different from other children. The development of Creoles and pidgins as communal languages also suggests another rarely discussed parallelism with the evolution of language in mankind as a population process, viz., inter-individual variation and the competition and selection that follow from it in the emergence of communal norms. We must remember that communal languages are only extrapolations from idiolects and language "acquisition" is an individual-based process (Mufwene 200la). Inter-individual variation is a consequence of the fact that no explicit teaching is involved in naturalistic language "transmission" and humans are not equally gifted in learning social skills by inference. Current speakers only provide the PLD from which the learners can construct their idiolectal "systems," which need not be identical with each other, though they are similar. As remarked by Chomsky (2000: 100), "We need not assume shared pronunciations or meanings to account for [successful communication], any more than we assume shared shapes to account for people who look alike." We must, however, ask why inter-idiolectal differences among the members of a language or dialect community are not as great as they could be. Part of the answer lies in the Cartesian assumption that UG, also identified as the biological endowment for language, is the "same" for all modern humans. According to Chomsky (2000: 30), "The only (virtually)

89

90

Language Evolution 'shared structure' among humans is generally the initial state of the language faculty" (i.e., UG). Consistent with the fact that language development is gradual and protracted over years, another part of the explanation lies in the mutual accommodations that speakers make to each other, bringing their "systems" closer to each other, at least with respect to the structures of the utterances they produce. These mutual accommodations are part of what Mufwene (200la) characterizes as competition and selection, which operate in the feature pool to which all speakers make contributions. As explained in Chapter 7, competition is a convenient misnomer for the coexistence of variants associated with the same, or similar, functions in a "system," in which they are not equally weighted. Selection refers to the advantage conferred to some of the variants that either prevail alone or are simply preferred (in more contexts) over other alternatives. In settings where several language varieties are spoken, competition and selection apply also to the different varieties (dialects and/or languages) in contact. We can say that they compete for speakers, if we overlook the fact that, through varying kinds of interaction settings, society influences the speakers' choices of codes qua language varieties. The cumulation of the choices (dis)favors some varieties. It can reasonably be surmised that the processes of competition and selection, which obviously apply in language "acquisition" and language change, must have also applied in the evolution of language in mankind, favoring innovations by some but not by others. Even from a phylogenetic perspective, action in the evolution of language must have lain in individual communicators' ingenuity and copying by others, not in any kind of explicit consultations within hominid groups. (Influences across languages are possible because languages are not selected wholesale with their features integrated, but rather because speakers develop their competences piecemeal, selecting features incrementally, often regardless of their origins.) Notwithstanding the fact that ecology rolls the dice to resolve competition, even in maintaining variation, we can imagine that spoken language was generally favored over signed language for precisely some of the reasons articulated by Givon (1998: 89), such as the following: (1) "freeing the hands and the body... so that communication may now proceed simultaneously with manual activities, and can in fact support them"; and (2) "transcending the visual field" so that "auditory-oral communication may proceed in the dark, in thick bush, over physical barriers that prevent eye contact."18 Typological structural variation among the world's languages is also a characteristic of modern languages that calls for an evolutionary explanation. It can be addressed from the same perspective as inter-individual variation above. Trans-local and trans-regional variation appears to be but a consequence of how variably selection resolved competition in different social ecologies. It is not far-fetched to speculate that competition and selection must have played an important role in reducing the range of wave frequencies used by humans in their phonetic inventories. They must also have played a role in leading members of individual communities to agree on the specific phonemes they use, on the ranges of allophonic

Creoles, Pidgins, and the Evolution of Language variation in the production of these phonemes, and on the way they combine them into longer meaningful utterances (words and sentences). Thus, they can tell which variants have normally been generated by their communal language and which ones have not. Thus, what theoretical linguists have referred to as "parameter setting" is nothing more than how selection, both at the idiolectal and communal level, resolves competition among variants in a particular ecology of language development. From an evolutionary perspective, I surmise that selection as a concomitant of normalization (as the emergence of norms) did not proceed the same way in different populations, thus leading to present-day typological variation. The ways in which incipient pidgins dispense with some structural characteristics of the languages from which they evolved suggest that they preserve the components of language architecture that are the most robust and perhaps the most deeply entrenched, which Lieberman (2002) identifies as the most "primitive." Some of these (such as dependence on vocal oppositions and the linearity of sounds) are shared by other primates, suggesting that the origins of language lie farther back in our phylogeny than in Homo erectus.19 For instance, incipient pidgins remain languages because they remain discrete and combinatorial spoken "systems" (relying on limited phonetic inventories), make use of lexical items (which even the earliest Homo erectus, if not Homo habilis, must have needed to survive), and preserve the symbolic referential function of language (which need not be associated with a phonological system).20 If, as Deutcher (2005) argues, demonstratives and terms for concrete objects and actions are also more deeply entrenched than abstract nouns and verbs as well as function words such as prepositions and conjunctions, it is informative to note that it is typically function morphemes (free and bound ones) that pidgins tend to dispense with. Not having a complex grammar accounts for why pidgins are said to rely heavily on pragmatic context for rich semantic interpretation, which seems to support Givon's (1998: 92) hypothesis that grammar has provided "speeded-up, more automated language processing" making full-fledged languages less dependent on pragmatic context. From the point of view of child language, one can argue that modern human infants are already hardwired for symbolic communication, in the spirit of Chomsky's UG (though details remain to be worked out) and that learning a native language entails starting with those aspects of communication that are phylogenetically the most deeply entrenched as determined in part by the state of the infant's cognitive capacity. However, we must remember that, unlike our hominid ancestors, modern children do not normally develop their own separate communal languages from some embryonic language, independent of what the adults who nurture them do to communicate. Thus the above observations do not lead to the traditional conclusion that ontogeny recapitulates phylogeny. Returning to Creoles, it is also noteworthy that the structural heterogeneity that has aptly been characterized in the literature as the "creole continuum" largely reflects the fact that those who developed these vernaculars were not engaged in the process as a team.

91

92

Language Evolution Rather, each one of them was personally trying to communicate and in the process developed their own individual idiolect, though they exerted mutual influences on each other - which accounts for the family resemblance that obtains among the idiolects of a communal language (Chapter 8). The divergence of Creoles from the relevant European languages is simply a selective accumulation (through the "invisible hand") of divergences that took place convergently in the idiolects of speakers, just what happens in any case of language evolution. I surmise that our hominid ancestors who developed various ancestors of modern languages did not proceed like teams either, as much as members of every community must have wanted to communicate successfully with each other and converged their "systems" through mutual accommodations.21 There must have been variation at all stages of the evolution of language, which fostered competition and selection, hence continuous evolution. Thus, at every stage of the evolution of mankind, speakers modified the language they learned from the preceding "generation of speakers" (regardless of age group), with children and adults all engaged in the process. If there was one protolanguage, which must be situated farther back than Homo erectus (pace Bickerton 1990), inter-individual variation is where the answer may lie for today's typological variation among the world's languages. Future research should shed more light on the social, population aspects of the evolution of language in the phylogeny of mankind.

Race, Racialism, and the Study of Language Evolution in America1 Chapter Outline 6.1 6.2 6.3 6.4

6.1

Preliminaries Race and ethnicity in American history Race in linguistics Conclusions

93 96 99 110

Preliminaries

The notion of RACE has often been invoked in linguistics either to dispel social biases or to explain sociolinguistic behavior. In the first case, the typical assertion is that the language variety one speaks is determined by their social environment, not by their "race" constructed as a biological notion. For instance, an African American child who has been raised in a European American family or has grown up in a predominantly European American middle-class neighborhood speaks like this population with which he/she has interacted regularly, despite the "racial" differences that distinguish him/her from them. Cases have also been reported of European Americans socializing the most with African Americans and speaking like the latter (see, e.g., Hatala 1976, Labov 1980, Jacobs-Huey 1997). The second case is consistent with the first in that in the United States, for instance, communicative networks are largely determined by the members' ethnic/racial affiliation, making it possible to speak of ways of speaking that are particularly African American or European American, although there is more obvious regional variation in the latter case.2 In this Chapter, I question the preconceived notion of RACE as biologically based, which is shared by both cases: African Americans speaking like European Americans and vice versa. Unlike biologists and physical anthropologists, linguists have hardly questioned RACE as a biologically defined notion. They have paid very little attention, if at all, to the fact that RACE has been treated as a social construct by other professionals. I wish to argue that the practice of both sociolinguistics and genetic linguistics can improve by adopting the biologists' and anthropologists' conception of RACE. I defend the thesis that RACE as a social construct has everything to do with the way an individual winds up speaking, especially in multilectal or

Ul

94

Language Evolution multilingual societies in which linguistic variation is correlated with cultural and racial or ethnic variation. In societies such as the United States, race affects population structure, determining which other members of his/her society an individual interacts with and therefore which other speakers he/she wishes to identify with. This explains why African Americans and European Americans who do not speak varieties associated with their races are considered as atypical as individuals who do not speak varieties associated with their socioeconomic classes. My position is simply a consequence of another common and well-justified working assumption in linguistics, viz., while humans are genetically endowed with a language faculty - also identified as Universal Grammar - competence in a particular language is a learned aptitude. A speaker works actively to develop it, largely influenced by his/her social ecology. Independent of his/her personal skills, the tacit knowledge that a speaker develops in his/her language is largely determined by the particular varieties he/she has been exposed to. Thus, we cannot expect a speaker who has been exposed only to Midwestern White middle-class English to have features of African American English in his/her idiolect, unless these are also peculiarities of the particular vernaculars spoken by the White speakers he/she has interacted with. Neither can we expect an African immigrant or his/her children to adopt African American English peculiarities unless they have socialized regularly with speakers of this sociolect. I am also concerned in this Chapter with positions defended in linguistics over the past few decades that raise the following question: To what extent have we emancipated ourselves from the dominant nineteenth-century ideology in Europe that considered European languages and cultures as superior, more evolved, or more refined than their non-European counterparts? Overall, linguists have presented themselves as fully emancipated from social biases that, for instance, posited a distinction between "languages of culture" and those that supposedly convey no "civilization" or were not considered sophisticated enough to express Western European cultures. Then, anticipating Darwin (1871), Schleicher (1863,1869), for instance, went as far as to declare that the structures of some languages, especially of those with isolating morphosyntax, were less evolved than the fusional morphology of (Indo-) European languages. His position reflected another then common assumption that some populations were biologically and/or culturally less evolved than others. It thus became customary to speak of "primitive languages," in the same way some races were considered evolutionarily inferior to others.3 To be sure, linguists no longer subscribe to such outrageous views today. However, I wonder whether hypotheses such as the following are not a legacy of the same nineteenthcentury assumptions about non-European populations, their minds, and their languages: (1) Creoles were made by children and thus represent grammatical structures the closest to specifications of Universal Grammar (Bickerton 1984a, 1984b); and (2) from a phylogenetic perspective, (incipient) pidgins give us insights into the nature of the human protolanguage (Bickerton 1990). If their structures are not protolinguistic, they are the closest thing to it.4

Race, Racialism, and Language Evolution Taking into account some of the arguments in Chapter 5, it is hard to ignore parallelisms between Bickerton's hypotheses and the views expressed by some nineteenth-century French philologists-cum-creolists who associate the emergence of Creoles with "the phonetic and grammatical genius, so to speak, of an inferior race" (Vinson 1882, my translation). Along with others such as Bertrand-Bocande (1849), Baissac (1880), and Adam (1882, 1883), Vinson assumed the minds of African slaves to be child-like and their languages considered as the reflection of these minds - to be "primitive," "instinctive," "in their natural state," and "simple." An important difference is that Bickerton does not assert that the minds of "creators" of Creoles are primitive, they only produced, according to his language bioprogram hypothesis, grammatical systems that are essentially protolinguistic. Another is that while the French scholars favored substratist accounts, Bickerton has favored a universalist account based on a language bioprogram putatively accessible only to children. One can thus see why, as I observe below, Bickerton shares with substratists some of the same mistaken working assumptions that have been inherited from nineteenth-century linguistics. Against such questionable working assumptions, DeGraff (2003,2004,2005) has done an impeccable job of documenting double standards in the ways the development and structures of Creole language varieties have been accounted for in comparison with other language varieties that emerged around the same time. He has basically espoused Mufwene's (200la) uniformitarian position in arguing that Creoles have evolved as naturally as noncreole language varieties, by the same restructuring processes, subject to peculiarities of the ecologies of their emergence. Below I adduce arguments that support DeGraff's position against Bickerton's (2004) response to it. RACE also comes to bear in studies of language evolution in the United States, especially regarding differences in the ways that English has evolved among African Americans and European Americans, a differential phenomenon dubbed since Labov & Harris (1986) as "divergence" of White and Black vernaculars, regardless of how old we think the process is (cf. Bailey & Maynor 1987). The process has been made more obvious by the Northern Cities Vowel Shift, which, according to Labov (2001), has affected only urban European Americans, indicating that Americans continue to socialize along color or race lines. That RACE bears differentially on language evolution has also been made obvious by Wolfram's (2000) study of the African American youth in Hyde County, North Carolina. Members of this group prefer to identify with urban African Americans rather than with their (grand)parents, whose speech they find too close to that of their European American neighbors.5 In all these cases the reason does not seem to be other than what was articulated above, viz., RACE (under any interpretation, to be clarified below) can play an important role in a population's structure, determining who can socialize and identify with whom; it can therefore be an important factor bearing on language speciation. While none of the above approaches appears to be misguided, we would be remiss in not reexamining the ways in which we have accounted for differences in patterns of language

95

96

Language Evolution evolution between descendants of Africans and Europeans in the New World. For instance, we have seldom questioned why only the evolution of English among African Americans has typically been studied as contact-based but not its evolution among European Americans. Yet the vast majority of European Americans today descend from immigrants from continental Europe and usage of English as a vernacular among them is as much the result of language shift as it is among African Americans. Even the following observations raise questions about the extent to which linguists, the majority of whom are of European descent, have really emancipated themselves from nineteenth-century racialism, the tendency to correlate behavioral differences within a population with racial variation. Thus, like lay people, linguists have typically considered English among African Americans as a less regular evolution than among European Americans. Yet a closer look at variation in North American English suggests that African American English varieties are as much instances of the varying divergence of colonial English from British English varieties as other North American English varieties. To be sure, some of these, such as Appalachian and Newfoundland vernaculars, have now attracted (more) attention, but the volume of the scholarship on them is very small compared to that on African American varieties, too small to shed light on the overall evolution of English in North America. Generally, very little has been said about the appropriation of English by indentured servants who did not immigrate from England compared to the attention that the appropriation of English by African slaves has received. No justification has been provided for measuring the extent of the divergence of African American English varieties by comparing them with standard English, whereas, as suggested by Krapp (1924) and Kurath (1928), these vernaculars evolved from the nonstandard varieties the African slaves had been exposed to, those spoken particularly by the low-class colonists with whom the slaves interacted the most regularly. Also, as pointed out by Lippi-Green (1997: 184-5), something is fundamentally wrong with saying that the "language convergence" expected by some linguists to take place in the United States must amount to the "assimilation of European American linguistic patterns" by African Americans. Although the term convergence does not preclude this particular kind of evolution, the expectation on the part of linguists suggests that something would be anomalous with an evolution in the opposite direction. A number of other embarrassments can also be pointed out. However, I need not be exhaustive here. I must first explain why linguists should not continue assuming the notion of RACE to be biologically based.

6.2

Race and ethnicity in American history

As noted above, the notion of RACE is far from being unequivocal. Two main interpretations emerge from the literature today: one biological and the other social. Overall, most biologists

Race, Racialism, and Language Evolution and physical anthropologists have avoided defining RACE biologically (see below), although, as pointed out by the Encyclopedia Britannica (2002 electronic edition), the term was originally adopted to identify subspecies and applied even to non-humans. For both biologists and anthropologists, no particular genotypic features can be used reliably to define RACE, a reality that Darwin (1871) was already confronted with.6 There can be more genotypic variation among individuals grouped in the same race than between members of different races. Moreover, while members of different biological species rarely interbreed, members of different races can and often do, as they use the "same" reproductive mechanisms. The principal constraints on mating practices are cultural. Thus, Cavalli-Sforza & Cavalli-Sforza (1995) find it more plausible to speak of the human race than to attempt to justify any number of races into which the human population can be categorized. This is basically the position endorsed by Marks (2002), who deplores the way some social Darwinists have misguidedly invoked race to promote racist agenda, for instance, by claiming that particular populations are biologically predetermined to have superior intellectual skills or athletic fitness (see Sarich & Miele 2004 for a more recent example). He points out that this is a kind of variation that can as easily be accounted for by the social ecologies of the populations which have been investigated. His position is shared by several other anthropologists, as evidenced by the American Association of Anthropology's (1999) and the American Association of Physical Anthropology's (1999) statements on race and by articles in the same issue of the American Anthropologist. The current anthropologists' position is that RACE is fundamentally a social notion which is unfortunately deeply embedded in social Darwinism. This tradition has posited competition and selection as the state and mechanism that account for ranking some populations as superior to others. The latter are claimed to be less advanced or "evolved" in the evolutionary trajectory of the human species. Nowadays, like biologists, anthropologists reject the biological significance of phenotypes such as COMPLEXION, EYE or NOSE SHAPE, EYE COLOR, SIZE OF LIPS, HAIR TEXTURE, and BODY SHAPE and SIZE, on which racial distinctions have typically been based, because they cannot reliably be correlated with race-specific genotypes. There is a lot of variation within every accepted race along these particular definitional criteria. Although variation in human populations has been noted since antiquity, it is especially in the eighteenth century that it was exploited by some Europeans to justify colonization. Since then, it has been manipulated and redefined in various ways in the interest of assigning citizenship discriminatively particularly in the former settlement colonies of the Americas and Australia, which were considered to be extensions of Europe.7 As noted above, promoters of such agenda have indeed also associated some stereotypic behavioral characteristics with some races, disregarding socioecological factors that affect particular evolutionary directions by natural selection and can clearly also be invoked to account for these peculiarities. The approach has rightfully also been characterized as "scientific racism," a more negative designation than the original "social Darwinism." The latter name is due to the fact that,

97

98

Language Evolution although he was opposed to racism (as pointed out by Gould 1993), Darwin (1871) also spoke of populations, races, and languages that were more evolved than others, which he considered "primitive" or "savage." (See also Radick 2002 for an informative discussion of the issue.) Smedley (1999) argues that although older civilizations knew of geographical variation between human populations, they did not use RACE as the basis of social discrimination. According to him, The term race had been used to refer to humans occasionally since the sixteenth century in the English language but was rarely used to refer to populations in the slave trade. It was a mere classificatory term like kind, type, or even breed, or stock, and it had no clear meaning until the eighteenth century. During this time, the English began to have wider experiences with varied populations and gradually developed attitudes and beliefs that had not appeared before in Western history and which reflected a new kind of understanding and interpretation of human differences.8 (1999: 694)

As noted above, the term race was thus used in biology to designate "subgroups of an individual species with distinctive phenotypes" (electronic Encyclopedia Britannica 2002). It is with the subsequent European colonial usage of the term to rank different populations that the ideologies of mission civilisatrice or the white mans burden would emerge, along with the pretension by Europeans to attend to the "needs" of populations they considered to be inferior races.9 Although in today's North America, racialism typically applies to relations between Whites and non-Whites, according to a gradation of "races" associated with lightness of complexion, one need not go farther back than the nineteenth century to notice that today's "White" race (also identified euphemistically as Caucasian) has not always been as unified or integrated. For instance, southern and eastern Europeans did not then count as Whites (Guglielmo & Salerno 2003), nor did the Irish (Harrison 1995). As the criteria changed over time and it became possible to include these populations, questions arose about whether populations of mixed Native American and European descent ("half-breeds") should count as Whites (James 2003:243). To date, the question of the classification of Hispanics remains equally unresolved, with the term White Hispanic suggesting that a person so designated is less white than his/her non-Hispanic White counterpart (James 2003: 242-3). The above concerns reflect the old preoccupation with distinguishing a putatively pure White race from others. As the term race became controversial, the term ethnicity was introduced, curiously to be applied more to non-Whites than to Whites (Tabouret-Keller 1999, Brutt-Griffler 2005), as evidenced by designations such as ethnic foods, ethnic cuisine, and ethnic neighborhood, which are typically associated with non-European customs. Thus, the concept of ETHNICITY has often functioned as a euphemism for that of RACE (Harrison 1995), ironically also used in a more exclusionary fashion. Fenton (2003: 22) tries to articulate the difference, claiming that "race makes explicit reference to physical or Visible' difference as

Race, Racialism, and Language Evolution the primary marker of difference and inequality" thus to phenotypes, in contrast with ETHNICITY, about which "the point of reference of difference is typically culture." However, he also adds that often "the group referred to is 'other' (foreign, exotic, minority) to some majority who are presumed not to be 'ethnic.'" As also pointed out by Harrison (1995), uses of the terms race and ethnicity overlap, suggesting we must be cautious when we use them to explain linguistic behavior or differential language evolution. Below I expose some problems endemic to linguistics.

6.3

Race in linguistics

6.3.1 James (2003:244) makes a distinction between "scholars who study race and racial dynamics" and "those who routinely use the concept of race in their studies" without questioning it. The former "study race and ethnicity as social phenomena [and] understand race as dynamic and situational," whereas the latter "tend to treat [race] as a primordial or fixed characteristic." They use it "as a cause [or basis] of a myriad of social processes and distinctions." Linguists appear to fall in the latter category. The question is how much needs rethinking. For this we must return again to the nineteenth century. Marked also by the beginning of the exploitation colonization of Africa and Asia (sanctioned by the 1884-5 Berlin Treaty),10 the nineteenth century is indeed the period when racialism becomes particularly strong among Europeans. The practice finds support in Darwin's extension of his natural selection to human races, claiming that some are less evolved than others. This can be correlated with Schleicher's (1863) ranking of morphosyntactic patterns, suggesting that the isolating type is the most primitive and the European fusional type is the most advanced. It is in this context that nineteenth-century French philologists such as Adam (1882,1883) and Vinson (188) could claim that the structures of sub-Saharan African languages reflected the child-like mental state of their speakers, whose speech organs were considered too "clumsy" to produce the "refined" systems of European languages.11 Creoles were then considered as aberrations, because they are not pure, being mixed, and Hugo Schuchardt's interest in them was scorned. All these academic developments were taking place, ironically, after Sir William Jones had conjectured, in 1786, that lexical similarities between, on the one hand, Sanskrit and, on the other, English, Greek, and Latin were due to their common ancestry, not to a polygenetic coincidence nor to mutual borrowings. His discovery would lead to the hypothesis that several Indo-Iranian and most European languages belong to the same genetic family now identified as Indo-European. However, as Hutton (2000) clearly shows, this hypothesis was met with resistance among (amateur) philologists, because, as is evident from Darwin (1871), language, race, and culture were then supposed to coevolve, hand in hand.

99

100

Language Evolution Europeans generally believed themselves to be so superior to other races, especially those that they had colonized, that they could not consider being classed together with some of them. Philologists such as Maine (1861) and Freeman (1881,1886,1892) could not conceive of the possibility that, as implied by Jones' hypothesis, Aryan and Indie populations may be of the same racial stock. They considered the latter too "barbaric" (similar to the "negro") to be associated with the "civilized" Europeans, even by "adoption."12 Freeman argued that there was no indication of cultural or racial assimilation of the Indians to the Europeans, the Indians having presumably remained distinct in their customs and languages. He remarked that where some mingling of Europeans and non-Europeans had taken place, such as in the "West Indies," this had produced a "grotesque imitation of the English ways," suggesting that "real assimilation [of the kind that had obtained in Europe relative to Greek civilization] is impossible" (quoted by Hutton 2000: 65). Referring to Hispanicization in the Americas, he said, according to Hutton (ibid.), that "the Europeans have sunk to a lower level, in contrast to the Greeks' ability to raise the inhabitants of Sicily and Southern Italy to theirs." Interestingly, European travelers to the southern parts of the United States and the Caribbean in the late eighteenth and in early nineteenth centuries claimed that similarities between Whites' and Blacks' English varieties were due to the corruption of the former by the latter (Brasch 1981). The main culprit was suspected to be the influence of Black nannies on White children. The success of writers such as Ambrose Gonzales, who had spent his childhood in destitution among former slaves on Edisto Island (South Carolina), in representing Gullah faithfully did not help sway opinions from this misrepresentation of language evolution in North America. His introduction to The Black Border (1922), in which he blames the "quaintness" of Gullah and divergence from White American English on the "clumsy tongues" and the "mental inferiority" of Black Africans, did not help the situation. On the contrary, it drew more attention to (his rendition of) it. As argued in Mufwene (200la), history suggests that the European indentured servants, with whom the slaves interacted regularly, played no smaller part than the Africans in the development of American Southern English. As suggested by Krapp (1924), Kurath (1928), and other early dialectologists, the particular way the plantation colonies evolved from the earliest homesteads limited the influence that African substrate languages could have exerted. As explained in Chapter 3 (see also Mufwene 1999b regarding African American English), the homesteads were integrated communities consisting of low-class Europeans (free and indentured servants) speaking nonstandard varieties and of a minority of Africans. The colonial populations then grew more by birth than by importation. Creole children then learned to speak the emerging nonstandard koines of these initial communities as their vernaculars, regardless of race. The adult African-born captives were in full-fledged communicative settings that precluded the development of pidgins as reduced means of

Race, Racialism, and Language Evolution communication, although they must have spoken second-language varieties of the colonial European koines. Limited importations of more slaves created situations in which they remained minorities and could impact the dominant models spoken by the European and Creole populations only minimally, although the possibility of substrate influence would increase dramatically by the time of the plantation societies, when the Bozal slaves, Africanborn, would become the majority and the newcomers would often learn the local vernacular from seasoned slaves, who were also African-born. Thus, the Creole slaves of the homestead phase became the strongest counterforce slowing down the spread and significance of substrate influence in the structures of the emergent Creole vernaculars. By continuing to make available native, non-creole linguistic models to the Bozal slaves, they provided the latter with variable opportunities to decrease the proportion of xenolectal elements in their competence (see Mufwene 2001a and Chapter 10). History shows indeed that they could not stop the inevitable, but for quite some time, they must have kept it to a minimum. The earliest Creole vernaculars in the New World do not seem to have emerged before the eighteenth century, perhaps the late seventeenth century in places such as Suriname. The earliest references to slaves' speech peculiarities date only from the early eighteenth century in Suriname, though we cannot deny the existence of Bozal slaves' interlanguages (i.e., transitional varieties) throughout the colonial and post-colonial periods until the end of the slave trade. In any case, what we cannot deny at all about the same period is that the linguistic inputs from the European and Creole populations of the homestead phase, who spoke no creole vernaculars then, consisted of nonstandard vernaculars. Variants peculiar to standard varieties were hardly part of the linguistic picture for the slaves who would develop Creoles. The Gonzales style of account for peculiarities of Gullah and similar varieties, an account akin to the baby-talk theory, had no historical backing at all. The question is why the theory got any attention in linguistics at all. By the time we get to Boas (1894), the myth of inferior races and of the coevolution of language and race has been partly put to rest,13 although an alternative would persist in "scientific racism" (e.g. Sarich & Miele 2004) which would associate specific races with specific skills. (See Caspar! 2003 for an informative historical perspective.) Unsurprisingly, the darker races would be associated mostly with somatic skills, whereas the lighter ones with higher intellectual fitness.

6.3.2 Although, to my knowledge, modern linguistics has played no role in scientific racism, the question remains of whether its practitioners have really emancipated themselves from some of the working assumptions that were typical of philology and the early stages of genetic linguistics in the nineteenth century. Linguists do not seem to have given up the

101

102

Language Evolution myth of "pure" languages little influenced by others with which they came in contact. This is clearly what is suggested by accounts of the emergence of the Romance languages that do not invoke the contact of Latin with the indigenous Celtic languages to explain why and how it speciatiated. As shown in Chapter 3, the same is true of traditional accounts of the diversification of Indo-European as owing to internally motivated change and their refusal to recognize Creole vernaculars (products of the contact of European languages with especially African languages) as the latest outcomes of the still-ongoing speciation of Indo-European. Has race not been a disfranchising factor in the ways linguists have accounted for varieties of European languages developed and spoken primarily by non-European populations? I address such questions below, without necessarily suggesting that the linguists associated with these particular research areas are racist. I simply wish to show that, whether or not we acknowledge it, today's linguistics still bears a legacy of the nineteenth century's ideology of language purity. As innovative as he was, Ferdinand de Saussure did not trash out negative aspects of the study of language evolution still held in the early twentieth century. Some of the nineteenth-century social biases against non-Europeans have survived to date. Revolutions are typically partial and remain in some ways continuations of the traditions against which their authors argue. It is not unwarranted to reexamine our working assumptions every now and then, before we stray too far on mistaken paths. The traditional stipulation of Creoles as separate languages, rather than as dialects of the European languages they have evolved from, is in fact a convenient starting point. To be sure, we no longer think that these recent colonial vernaculars are aberrations or adulterations of European languages by inferior minds and speech organs. On the other hand, the literature has been ambivalent over whether they are natural phenomena. Linguists have accounted for their emergence in ways that have privileged language contact and mixing over the inheritance of structural materials from their "lexifiers."14 Creoles appear to be crying out loud and clear that we reexamine our working assumptions on language evolution, as done in Chapter 3. Yet we seem to have refused to learn from them their "side of the story" and have chosen a priori to disfranchise them as unusual recreations from a prior "pulverization" of European languages (McWhorter 2001) or as non-ordinary communal "creations" by children, who allegedly assigned syntax to the "syntaxless" pidgins of their parents (Bickerton 1981, 1984a, 1999). Moreover, as discussed in Chapter 5, their putative pidgin ancestors have been claimed by Bickerton (1990) to offer us clues about the protolanguage putatively invented by our hominid ancestors about 200,000 to 50,000 years ago. Too few studies, such as Botha (2006a, 2006b), have questioned this loaded speculation. As explained in Chapter 3, none of the traditional assumptions regarding the development of Creoles is supported by the socioeconomic histories of the territories where Creoles have evolved. Noteworthy about the alleged pidgin ancestry of Creoles is the geographical complementary distribution of the places where the two kinds of language varieties have evolved. In Hawaii, where both varieties are claimed to have evolved, pidgins are

Race, Racialism, and Language Evolution associated with plantations, whereas Creoles are associated with the city (Roberts 1998, 1999), contrary to the plantation settlement colonies of the New World and Indian Ocean (Chapter 10). In the latter, Creoles typically developed on the plantations, and closer approximations of the European languages developed in the city and in smaller rural estates, where the non-Europeans were minorities, in equal numbers to, or just slightly more numerous than, the European settlers and indentured servants with whom they interacted regularly. There was no (rigid) residential race-based segregation between the Europeans and non-Europeans in non-plantation settings. One important factor which the settings where pidgins developed share with those that produced Creoles is that they had populations with overwhelming non-European majorities and at some point the proportion of non-fluent speakers (the Bozal slaves on the plantations of settlement colonies) surpassed that of fluent speakers. The latter consisted of the interpreters in the case of trade colonies and of the early Creole slaves in settlement colonies. Another shared factor of Creoles and pidgins is that, because of the putatively unusual critical role played by language contact in their emergence, they have both been considered evolutionarily less normal and natural, and therefore without the usual genetic ties to their ancestors, than non-creole languages. I provide counter-arguments to these positions in Chapter 3. More relevant to the focus of this Chapter is the question of the (relative) purity of a language. Although Creoles (and pidgins) are no longer considered as aberrations, they have been misconstrued as too mixed and therefore less "pure" than other languages. If it is true that language contact has played a certain role in the evolution of, for instance, English and the Romance languages, Creoles must, according to the received doctrine, have surpassed the threshold of the amount of xenolectal influence tolerable by a particular language in order to remain in the same genetic family. The problem is that nobody has articulated in any way what that threshold is. Nor has anybody shown to what extent the evolution of Creoles such as Gullah and Louisiana Creole is more discontinuous from nonstandard English and French, respectively, than that of, say, Amish English and Louisiana French. Nobody has shown in what way these particular Creoles are younger than their non-creole counterparts, nor in what ways the latter are less "restructured." As a matter of fact, nobody has told us what "restructuring" actually means; Mufwene (1996a, 200la) is exceptional in proposing a definition. Nobody has operationalized the notion of RESTRUCTURING in a way that enables us to tell the extent to which the evolution of non-creoles owes nothing, or less, to this feature-recombination process.15 In fact nobody has offered a yardstick for measuring variation in extent of restructuring even among Creoles themselves. History suggests that the input itself cannot have been identical from one setting to another! It is striking that a discipline can have propounded such strong hypotheses about the language varieties of particular populations without sticking to the same empirical

103

104

Language Evolution standards of proof or plausible justification as for the non-creole colonial varieties. Linguists have resisted considering the relevant sociohistorical evidence seriously, to verify whether Creoles have really been "created" de novo under unusual conditions of language contact that prevented their "creators" from having access to their "lexifiers." The discussion in Chapter 3 suggests otherwise. It is not clear why linguists have assumed that segregation entailed end of access to the "lexifiers" as if only people of European descent had learned them during the homestead phase. Nor is it clear why it was taken for granted that vernaculars of the Euro-African interactions during the homestead phase of settlement colonies must have been pidgins, like in trade colonies. Does the condition of being a slave or contract laborer affect the mechanism of language "transmission" in a way that can justifiably be characterized as a "break in the transmission of the lexifier"? How could there have been a break when the term lexifier itself suggests normal transmission of the vocabulary?16 Because the overwhelming proportion of the lexicon was inherited from the lexifier, there cannot have been a "break in the transmission" of the "lexifier," unless creolists suggest that the break involves only the "transmission" of grammar/structure. However, what does it mean to have a situation in which the vocabulary of a language is normally "transmitted" but the grammar is not? Does it mean that one can learn a language in a naturalistic setting paying attention only to the vocabulary but (almost) none at all to patterns of pronunciation and usage? Is this consistent with the layman's notion of what learning a foreign language is? Or do linguists expect grammar normally to be "transmitted" intact or with little modification? These questions and more bear on the practice of genetic linguistics and historical dialectology, especially on whether or not changes that have produced a particular variety constitute an evolutionary anomaly and what defines the anomaly. To date, if these questions have been addressed at all, they have in ways that can more easily be associated with socially inherited racialism than with any convincing linguistic arguments. There seems to have been more commitment to validating distinctions stipulated in the nineteenth century than to verifying them. In a somewhat different vein, it is also bizarre that descriptions of Creoles have focused more on features which distinguish them from their "lexifiers" (however poorly construed these are) than on features they share with them. Aside from the fact that the significance of structural differences between Creoles and the related European languages may have been blown out of proportion, one also develops from the usual publications, even book-length discussions, the impression that these colonial vernaculars are incomplete language varieties. Several creolists have also chosen to favor the "language bioprogram" or "substrate influence" to account for "creole features" that these vernaculars nonetheless share (partially) with their "lexifiers." This bias is undoubtedly a consequence of the assumption of the discontinuity hypothesis in relation to the "lexifier." It is otherwise not consistent with the expectation that the learner of a language will make an effort to "acquire" target patterns in the best way they can. Nor is it consistent with the observations that in L2 "acquisition,"

Race, Racialism, and Language Evolution target structures that are congruous with LI structures will tend to be favored over those that are different if these convey (more or less) the same meanings (Chapter 8). These comments do not, of course, deny the fact that the learner's perceptions and analyses are sometimes incorrect. As I argued in Chapter 5, the Language Bioprogram Hypothesis is not consistent with the history of the territories where Creoles have developed, as attractive as it has been. The relexification hypothesis, like any other hypothesis that downplays inheritance from the "lexifier" in the grammars of Creoles, is at odds with theories of naturalistic L2 "acquisition," although no adequate account can ignore the role of substrate influence in the evolution of these vernaculars. Why then have linguists subscribed to hypotheses that common sense should have advised them to question seriously, especially as our knowledge of colonial history, more precisely of the peopling of the colonies and of the ensuing population structures, has improved? Or, despite our rejections of the baby-talk hypothesis, have we kept a legacy of the nineteenth century in assuming that some races must have their own peculiar ways of appropriating languages of the dominant populations? Note that the same literature that has professed all or some of the above disputable assumptions about Creoles has not sufficiently invoked variation in the ecologies of language "transmission" as part of the explanation for why Creoles diverge in some, but not all, of their structures from their "lexifiers." "Imperfect language learning" is no explanation, since this applies to some extent, however variably, to all cases of language "acquisition" and no particular yardstick has been proposed to tell one kind of "imperfect learning" from another. Segregation is not a sufficient explanation either, especially if, as history suggests, model speakers of the colonial European languages need not have been European. Since variation must have been rampant in colonial varieties of the "lexifiers," segregation entails no more than fostering divergent evolutions, because the segregated communities must have inherited different patterns of variation, in which the same variants would not keep the same strengths even if they were attested in all the communities. Such disruptions of the earlier "balance of power" among the variants would trigger divergent evolutionary trajectories. We add to these observations the fact that Creoles evolved in those settings where the non-European populations became the overwhelming majorities and they developed especially at times when the African-born became demographic majorities. We can thus also factor into our accounts the natural fact that, under varying conditions, substrate languages have influenced the selection of the particular structures that make Creoles different from other varieties that have evolved from the "same" European languages (Chapter 8). In the same vein, our preoccupation with the divergence of varieties spoken by descendants of non-Europeans also raises some concerns. For instance, consider polities such as the United States and Australia, where the vast majority of European populations today have parents or ancestors who originated from continental Europe and therefore did not speak English. Why is language contact not assumed to have been a central factor in the way that

105

106

Language Evolution English has evolved among them? Why has it not been invoked to account as much for the divergence of American and Australian Englishes from British Englishes as for the loss of their heritage languages? Why has there been more emphasis on separation by the oceans separating North America and Australia from the British motherland than on contact? Let us assume, like Chaudenson (1992, 2001, 2003), that the (nonstandard varieties of the) Western European languages that came in contact in the colonies were structurally more similar to each other than they were to African languages. Let us also assume that such typological kinship would have made it easier for Europeans to learn each other's language more faithfully. Then the minimal divergence assumed to have obtained between the colonial varieties and their metropolitan counterparts does not entail lack of restructuring.17 As explained in Mufwene (2001a), all cases of language change involve some restructuring. The distinction between "restructured" and "non-restructured" varieties suggested by, for instance, Holm (1989) just does not make sense, especially in cases where new language varieties have emerged, regardless of whether they are considered Creoles or non-creoles. This kind of opposition is in fact also questioned by the notion of "degree of restructuring" that appears in the title of Holzschuh-Neumann & Schneider (2000), which is simply a consequence of the assumption that different contact settings exerted varying kinds of influences, at least in strength, on the restructuring process. What remains bothersome even in the latter case is the fact that the proposed measure of "degree of restructuring" has so far been limited to disfranchised European language varieties spoken by populations of non-European descent. Moreover, colonial history defies positing a common terminus a quo for all contact settings. In fact, it is not far-fetched to correlate the above racial bias with the failure of sociolinguistics and historical dialectology in North America to have learned more from creolistics than could have been possible. While the advancement of Labovian quantitative (socio) linguistics owes tremendously to the study of structural variation in AAVE and in Caribbean English Creoles, it is striking that the notions of SPEECH CONTINUUM, BASILECT, ACROLECT, and MESOLECT, which are associated with variation, have not readily been applied to non-creole communities, where there is indeed no sharp line demarcating nonstandard from standard varieties. The situation has largely to do with the fact that basilectal Creoles have been stipulated as separate languages, instead of nonstandard varieties of the corresponding acrolects. Since Schuchardt (1914), they have been assumed to change in the direction of the acrolect, in communities where they coexist with their "lexifiers." The debate continues to date on whether AAVE is not the best exemplar of this "decreolization" qua debasilectalization trend, as speculated by Schuchardt (in comparison with Saramaccan) and conjectured by DeCamp (1971) for Jamaican Creole. Yet, the notions of BASILECT, MESOLECT, and ACROLECT offer adequate alternative ways of discussing the traditional stratification of dialects between standard and nonstandard dialects, bridging a wide range of productions between them that do not quite fit in these polar varieties. Rickford's (1990) observation that the dominant

Race, Racialism, and Language Evolution norm in Caribbean Creole communities lies in the mesolect can certainly apply to non-creole communities too, where spoken acrolect is considered bookish and extreme nonstandard varieties are stigmatized. Linguists have been too wedded to the stipulation that the ecologies in which Creoles developed and/or continue to function are so unusual that little can be learned from them about even the social dynamics of language. As noted above, this indifference to possible lessons from Creoles is related to the stipulation that Creoles are separate languages from their "lexifiers" and from their non-creole kin which are considered as colonial dialects of the same European languages. Interestingly, linguists have ordinarily left it up to speakers to determine whether or not they speak the same language as another group, regardless of the extent of structural differences between the varieties involved (such as between classical, Koranic Arabic and vernacular varieties of Arabic). They have also typically rejected mutual intelligibility between speakers of the relevant dialects as unreliable. However, there are typically exceptions to the rule, and Creoles fall indeed in this category, as pointed out by DeGraff (2003, 2004, 2005). Yet, as often as Cockney has been cited as an English variety that is largely unintelligible to speakers of standard English (an ill-defined group), nobody has ever claimed it to be a separate, nonEnglish language. So, why have linguists been so dogmatic regarding Creoles, disputing their speakers' assertion that they speak one of the new varieties of the relevant European language? For instance, why do they insist that Gullah and Jamaican Patwa are not English while their speakers insist that they are?18 We may ask more questions. For instance, why have we been so eager to associate the Creole continuum with the decreolization-qua-debasilectalization hypothesis? Why haven't we considered the alternative that the socioeconomic histories of the relevant territories and the reality of language "transmission" around us have suggested, viz., that variation in the ecologies of individual speakers' language "acquisition" would have produced the same continuum of lects? With every individual speaker exposed to a different subset of PLD (produced by the particular group of speakers they have interacted with), it is impossible for all speakers of a language or dialect to develop identical grammatical systems. This is the reality that justifies the notion of "idiolect." As Wolfram (2000) clearly shows, even members of the same nuclear family are bound to vary idiolectally. Siblings of different ages do not have access to the same primary linguistic data, as they may not have been talked to in identical ways, do not have identical play groups, and do not have the same classmates nor necessarily the same teachers in the same schools. Notwithstanding variation in their abilities to learn and use language, they have not had access to identical inputs for the construction of their grammars. In situations where Creoles developed, things are compounded by the fact that the linguistic models for Creole children and Bozal slaves alike must have varied as noticeably as in situations where English has now indigenized, like in India, Nigeria, and other former British exploitation colonies of Asia and Africa.

107

108

Language Evolution Thus, the synchronic facts of variation alone present a situation that is simply analogous to that of the "half-empty vs. half-full bottle" dilemma. Some social bias that linguists may not have harbored consciously seems to have favored the decreolization hypothesis. Neither DeCamp's (1971) seminal paper on the subject matter, nor its forerunners by Schuchardt (1914) and Bloomfield (1933), nor even its follow up in, for instance, Bickerton (1973) adduce diachronic evidence to prove "decreolization." As exceptional as his study is, Rickford (1987) shows that Guyanese Creole has changed over time, not that it had debasilectalized. Only Rickford & Handler (1994) demonstrate that debasilectalization may have happened in Barbados, not necessarily also in other creole-speaking polities. As a matter of fact, Lalla & D'Costa (1990) argue against the debasilectalization of Jamaican Creole, and so does Mufwene (1994) for Gullah. Nevertheless, many are linguists who continue to associate the Creole continuum with decreolization. We could actually have considered an alternative arising from Irvine (2004), viz., both the basilects and their current acrolects have evolved from the same colonial feature pools, and they are related by the continua that have interested us simply because of the particular ways they selected structural features differentially from these pools. The received doctrine seems to have assumed unjustifiably that language practice and characteristics among descendants of Africans must be fundamentally different from their counterparts in European-majority communities. Yet, colonial history does not support this view. Regarding the "(post-)creole continuum" and the putative "decreolization," linguists appear to have given up the conventional research philosophy that new kinds of data should enable us to question some of our working assumptions.

6.3.3 The consequences of this social bias in the practice of linguistics are far-ranging. As shown in Chapter 3, linguists have denied Creoles any genetic connection to the European languages they have evolved from, allegedly because of the central role that contact has played in their emergence. They have also rejected completely, baby and bathwater along, Bailey & Maroldt's (1977) and Schlieben-Lange's (1977) studies that compare the evolution of (Middle) English and the Romance languages, respectively, with that of Creoles. Yet, there is so much that can be learned from the comparison, as much as it is justified to reject the extension of the term creole to English and the Romance languages. As a matter of fact, we should refrain from identifying Creoles and pidgins beyond the historical context of European plantation and trade colonies of the seventeenth to nineteenth centuries, typically between the tropics (Chaudenson 1992,2003). Bickerton (2004) appears to have overreacted to DeGraff (2003) in arguing that creolistics is not racialist. In many respects, linguists have provided no reliable criteria for the distinction between creole and non-creole language varieties, starting with not acknowledging

Race, Racialism, and Language Evolution Creoles as offspring of their "lexifiers" or even as (nonstandard) dialects of the same languages, in the way that Bickerton (2004: 831) misrepresents both DeGraff and Mufwene (2001a, 2003). The received doctrine appears to have prevented linguists from having a closer look at non-creole varieties of the same European languages and acknowledging contact as an important ecological factor in their evolution. It is doubtful that the "extraordinary socioeconomic conditions" (Bickerton 2004: 831) under which Creoles evolved "temporarily disrupted normal language transmission." Even in assuming the most rigid form of segregation in the plantation colonies, there was no shortage of native/fluent speakers of the "lexifier" among the slaves or contract laborers to "transmit" the European languages normally. We cannot ignore the role played by Creole slaves, speakers of noncreole and non-pidgin varieties, during the transition from the homestead to the plantation phases. Nor can we afford to ignore the role of interpreters in trade colonies, about which we are reminded, among others, by Naro (1978) for West Africa, Bolton (2000,2002) in relation to Chinese Pidgin English, Samarin (1982) for central Africa, and Reinecke (1969) and Beechert (1985) for Hawaii. To date, the discourse on language diversification has been reminiscent of that on racial variation in the nineteenth century, which used to disfranchise Mediterranean Europeans as less white or less pure because of mixing with non-European populations during the Hellenic and Roman Empires. The ideology of race purity, rather than today's assumption of racial variation in terms of a continuum, was so strong that even scholars ruled out by fiat the strong possibility, if not a fact, that Nordic Europeans must have mixed with the pre-Indo-European populations who preceded them where they migrated. History suggests successive contacts of Indo-Europeans not only with the pre-Indo-Europeans who preceded them (such as the Samis and the Basques) but also among themselves (Chapter 3). Mutatis mutandis, there are arguments against the assumption of mono-parental speciation associated with the neat Stammbaums of genetic linguistics. For some strange reason, Indo-European linguistics has made little room for language contact and has treated language convergence in the Balkans as an anomaly. Creole studies have developed as a byproduct of this particular ideology, which has led scholars such as Bickerton (2004) to suggest that these vernaculars evolved in an unusual way, because the language faculties of their "creators" were confronted with non-ordinary circumstances of population and language contacts. Apparently, Bickerton did not bother checking the socioeconomic histories of creole-speaking territories to find out that the cruelties of slavery and contract labor - no more "exceptional" than some cases of European indentured servitude - did not impede normal language "transmission" from one group of speakers to another. Although the nature of the adaptive responses to communicative challenges varies from one ecology to another, among all human populations, the particular processes involved (e.g., sound substitution or rule extension) do not, at least not in kind. This is the essence of the uniformitarianism that DeGraff (1999ff) and Mufwene (2001ff) have been arguing for. If Bickerton did check the

109

110

Language Evolution relevant colonial history, why has he placed so much emphasis on the alleged exceptionality of the development of Creoles? Now, consistent with some of the discussions above, it is natural to ask whether Bickerton is worse than the relexificationists and substratists who have privileged some form of substrate influence over inheritance from the "lexifier." Don't the latter accounts likewise presuppose or entail exceptionality in the development of Creoles? Studies of naturalistic language "acquisition" such as Klein & Perdue (1992), Perdue (1995), and Giacomi et al. (2000) prove as untenable the basic claim that one can learn a language by "acquiring" only its vocabulary and applying to this only, or mostly, the grammar of languages previously spoken. They provide no evidence of naturalistic language "acquisition" that supports relexification. Apparently, normal interference from a previously spoken language does not entail indifference to the structures of the target language. As a matter of fact, Creoles share many features with their "lexifiers" which are not attested in the relevant substrate languages, such as the prenominal position of determiners and adjectives (a category that is marginal in many of them), as well as preposition-stranding in English Creoles (Chapters 5 and 8). It also seems evident that features shared by varieties of the "lexifier" and some substrate languages were favored, by congruence, to become part of the relevant Creoles' structures, for instance, usage of a preverbal marker to mark tense or aspect. After all, substrate languages did not differ in all respects from the superstrate languages. What has so easily lent plausibility to hypotheses that should have been disputed from the start? The answer to this question may lie in the state of the art regarding language change, such as in the deeply entrenched assumption that the outcomes of internally motivated change are significantly different from those of externally motivated change. However, we have not demonstrated that the emergence of other, non-creole colonial vernaculars such as European American and Australian English varieties have not likewise arisen thanks to language contact. Trudgill (1986, 2004) suggests they have and Mufwene (2001a, 2006) argues that the same competition-and-selection mechanisms were involved in them as in the development of Creoles. So, have linguists inherited from the social background of the majority and of most of the leading experts too much of the "inferior" race ideology to see problems with the received doctrine?

6.4

Conclusions

Legacy has often not been questioned even by the most significant revolutions. For instance, despite all the shifts from the structuralist and descriptivist research paradigms, the assumption that languages are institutions whose "systems" have, in native communities, normally been "transmitted" almost intact from one generation of speakers to another has hardly been questioned. A consequence of this in theoretical linguistics is that it has become customary to work with one informant or by introspection and extrapolate one's findings

Race, Racialism, and Language Evolution to a whole community of speakers. In sociolinguistics, inter-group variation has received more attention than inter-individual variation. In quantitative sociolinguistics, individuals that diverge from group patterns have usually been tossed aside as "outliers" who skew the statistics, despite the fact that sometimes outliers also initiate changes that can spread in a community (Chapter 2). Overall, the notion of IDIOLECT has been under-exploited, despite all the contributions it could make to highlighting the complexity of communal grammar, especially regarding (mis)matches between, on the one hand, the knowledge that individual speakers have of their language and, on the other, the collective, communal knowledge that the community as a whole has been projected to have of the same dialect or language. Likewise, unlike anthropologists (see, e.g., the 1999 special issue of the American Anthropology on this question), linguists have hardly questioned the notion of RACE. Since the nineteenth century, it has been assumed to be biologically based. Thus, apparently in reaction to social Darwinism, linguists have usually claimed that a child normally "acquires" the language of their social environment, regardless of race; the latter does not predetermine what language variety a child will have as his/her mother tongue. So a Japanese child, of Japanese parents, growing up in a socially integrated neighborhood in the United States will speak American English of the same nature as his/her non-Japanese neighbors, and an African American child growing up in a predominantly White neighborhood (said to be "integrated") will speak White middle-class English rather than AAVE. Unfortunately this position in linguistics has not been in tune with the fact that definitions of the White race have changed several times in North America since the late eighteenth century. Nor has it kept up with the fact the current definition of the Black race in North America, based on the one-drop rule of hypo-descent (Taylor 2004), has been at variance with the Caribbean practice, where a "Brown" category has been interposed between the "White" and "Black" ones. This variation in time and space shows clearly that RACE is socially defined. The price of ignoring this history has been the failure to state that within some population structures - those in which race or ethnicity determines where one can reside and who he or she can socialize with - race, or ethnicity, also determines what particular language variety one is most likely to have as his/her (native) vernacular. Thus, it is certainly inaccurate to continue claiming that race does not bear on a speaker's idiolect, because race as a social construct does, in determining what other individuals a speaker can interact with, be influenced by, and can align his/her idiolect with. Thus, RACE and SEGREGATION are factors that shed light on the differential evolution of Creoles and AAVE compared to their colonial non-Creole kin spoken by populations of European descent. They are important parts of the explanations for the divergence hypothesis regarding European American and African American English vernaculars. As pointed out in Chapter 2, Labov's findings on the Northern Cities Vowel Shift, viz, that it affects only White urbanites, is a reminder that race as a social construct is an important component of the American population structure and it bears on patterns of language evolution.

111

112

Language Evolution The reason why African Americans have typically not participated in the Northern Cities Vowel Shift is that race barriers have prevented them from socializing (regularly) with European Americans and have discouraged them from identifying linguistically with members of other races. In this Chapter, I have obviously dwelled on what can be interpreted as race-based prejudice in the current professional practice of linguistics. It is far from my intention to accuse the practitioners themselves of racism. One can be trapped in a racial ideology without wanting to be racist. Any such victim can emancipate him-/herself. I wanted to show the extent to which we linguists have not yet (fully) emancipated ourselves from some of the social biases of the nineteenth century. This Chapter is an invitation to question ourselves on some of the most tacit of our working assumptions. I hope we will be able to discuss these issues and more in a frank dialogue.

Pa rt Two

Competition, Selection, and the Development of Creoles

This page intentionally left blank

Competition and Selection in Language Evolution1 Chapter Outline 7.1 7.2 7.3 7.4 7.5 7.6

7.1

Preliminaries Hybridism in the normal and natural development of Creoles Variational evolution: development and change by selection More on hybridism in language Constraints on selections Conclusions

115 121 124 127 128 131

Preliminaries

This Chapter capitalizes on my position that a language is a heterogeneous population of idiolects, often availing to every learner a range of variants for the same communicative function. They can be alternative pronunciations of the same word such as Tuesday with or without palatalization of t, soot with a long or short /u/ (thus with or without a contrast with suit), either pronounced with an initial /i:/ or /ay/, or part pronounced with or without a preconsonantal /r/ (what is known as rhoticity). The variants also often consist of alternative words denoting (more or less) the same thing, e.g., belly vs stomach vs gut (though beer gut/belly are more acceptable to most speakers than beer stomach). They may even be alternative syntactic structures such as it's too many people here vs there's too many people here vs there are too many people here. Depending on the particular social ecology in which an individual learns a language, he/she develops an idiolect in which one of the variants is dominant (rarely exclusive) and the other one is available for interpreting other speakers who use it (predominantly) or for cases when it may become handy to make oneself understood by speakers of other varieties. Usually members of the same communication network, or even the same speech community (subsuming several networks), select the same (dominant) variants. A convenient explanation for this convergence, as for lack thereof across networks, is that the variants are not equally weighted by speakers, some being preferable (or less marked) to them than others. The weighting can vary from one network to another, accounting for

7

116

Language Evolution variation across groups. This situation of unequal weighting of variants is what I describe in Mufwene (200la) as competition, with selection corresponding to the result of the preferences that speakers show for the less marked variants, for reasons that are not always obvious. At the population level, that of the communal language, where we must ask ourselves how selections made by individual speakers converge into selections of the whole community of speakers, or of most of them, the process conjures up again the notion of the "invisible hand" discussed in Chapter 4. A central question we were invited to address at the ISHPSSB forum was whether selection plays a role in language evolution. The invitation I had received did not specify the interpretation of "selection" that applies to language, let alone whether the term selection itself is applicable. The answer to the latter question is affirmative. The changes that affect a communal language are the cumulation and convergence of selections made repeatedly by individual speakers during their communicative acts. Note also that although speakers often make conscious choices of words, of pronunciations, and of grammatical rules when they speak a language (which they have already mastered by then), the selections that form their idiolectal habits are generally unconscious ones. The choices are influenced by where speakers "acquired" the language and who they have interacted with. These factors - the social environment and network of communication - are part of what I identify as the "ecology of language." Ruling out the typically unsuccessful attempts of academies to influence vernacular speech, speakers are normally unaware of how communal norms emerge from idiolectal habits to form dialectal or language-specific peculiarities. Several ecological factors - having to do with the ethnography of communication (the "external ecology") and the internal structure of the language at the time of its "acquisition" (the "internal ecology") - apply at this communal level. All these considerations justify speaking of "natural selection," especially in the sense that the agency of speakers is not a conscious one and the "invisible hand" that drives the convergence of the speakers' selections operates in more or less the same cumulative way as it does in biological evolution. It favors some genes while disfavoring others in nature, through the organisms, and indirectly the species, which carry them. Consistent with Sober (1984, see also Chapter 2), selection works directly on organisms (idiolects in the case of language). It is its cumulative effects that show up at the communal level of languages or species. In the introduction that Professor Hull (2002) had prepared for our panel (on "Language Change as a Selection Process") and subsequently published in the special issue of Selection, he also addressed the question of whether hybridism (in this case the state in which a system has integrated elements of different origins, i.e., from different language varieties) plays a role in language evolution. This enticed me also to express my position on the matter. Consistent with especially Chapter 3, I consider hybridism as a normal condition not just of those languages that have been disfranchised as "mixed" ones, chiefly Creoles, but also of the other, putatively unmixed ones. I am alluding here not just to the mistaken hypothesis about

Competition and Selection in Language Evolution Creoles that most of their vocabularies originate in their "lexifiers" while their grammatical features have been selected from (an)other language(s), nor to the possibility of a 50:50 ratio of contributions by different languages to the makeup of (the modules of) a language. I have in mind primarily the more common and natural condition which I associate with a feature pool from which individual speakers select materials specific to their idiolects. This Chapter revolves around this "feature pool" notion as explained in Mufwene (200la). An important caveat about this analogy is that the "acquisition" of linguistic features is typically with modification. In animal biology, the default condition of the transmission/acquisition of individual genes is perfect replication. Innovations arise from how genes are recombined into new genotypes and DNAs. As experimental phonetics makes it obvious, no speaker pronounces a sound exactly like another speaker, at least from the physics perspective of formant frequencies, if not perceptually. Likewise, as becomes more and more obvious in empirically grounded syntax, speakers often disagree on the ways and/or the extents to which constraints affect particular rules, for instance whether one can say I heard your prediction the Democrats would prevail, without a complementizer before the subordinate clause, in the same way that one can say You predicted the Democrats would prevail or, with less variation in acceptability judgments, You said/I heard the Democrats would prevail.2 All this follows largely from the fact that no speaker literally inherits a ready-made system transmitted to them; instead, as explained in Chapter 2, everyone builds their idiolectal "system" by approximating other speakers without being told explicitly how they must configure their speech organs nor what particular "rules" they must follow or apply. "Rules" are in fact constructed by analysts to capture patterns that emerge from extrapolations that the speaker applies based on partial analogies that he/she perceives among items that are related in meaning and/or in form, while these are "acquired" not wholesale but piecemeal, on different occasions, for communication. This process leaves a lot of room for inter-individual variation. According to the feature-pool approach, members of a speech community, especially those who interact with each other, contribute different models in pronunciation, lexical materials, grammatical models, and pragmatic constraints to the feature pool. The models that are pronounced more or less the same ways are inter-individually similar only on the family-resemblance model. Otherwise, they can differ in a number of ways, such as in the choice of words or in their preferences for particular grammatical variants, morphologically or syntactically. Various ecologies of interactions affect learners (and current fluent speakers willing to accommodate others) variably, selecting different subsets of variants into their idiolects. Variation among individual speakers' personalities and histories of social interactions (see Wolfram 2000) produce inter-idiolectal variation of the kind where no speaker replicates anybody else's idiolect, although those of speakers who evolve in the same communication networks are more similar to each other. That is, although all speakers/learners select their features (albeit with some modification) from the same pool, they do not select

117

118

Language Evolution exactly the same variants. Since selection is not always in exclusive terms, even if they select the same features, these are not recombined the same ways in their idiolects, because those that wind up as dominant are not always the same. In those cases where the contributors have more or less the same (kinds of) features, it is pointless to try to identify a single (dominant) donor. We then speak of congruence (Corne 1999), i.e. a situation where similar features reinforce each other. The features are thus favored to be dominant by the ecology in which the new idiolects develop. According to Chaudenson (2001,2003), congruence is an important reason why the new colonial varieties of European languages spoken by descendants of Europeans have not diverged as much from their lexifiers as Creoles have. Western European languages are typologically more similar among them than any of them is with many, or most, of the African languages they came in contact with. Although such colonial varieties have in fact diverged from their metropolitan varieties, due indeed to contacts with other European languages and, of course, also the new contacts of their own dialects, they are less divergent than Creoles are. In the case of Creoles, features of the substrate languages that new appropriators of the target language had spoken before were also contributed to the feature pool. The presence of these xenolectal features in the pool changed the "balance of power" among the native variants, making it possible for the selection into Creoles of features other than those selected into the other, non-creole varieties. It also made it possible for features from the substrate languages to make their way into the idiolects of Creole children, who were more concerned with communicating with the people they socialized and interacted with than with producing an ideal communicative system. As is evident from Chapters 3 and 6, the divergence between the Creole and non-creole varieties that emerged in the colonies is largely a consequence of race segregation, which reduced the extent of social and therefore verbal interactions between Europeans and nonEuropeans on the plantations. Thus, segregation fostered the emergence of different feature pools in which the variants from the lexifier no longer had the same "balance of power," notwithstanding the fact that the feature pools of slaves also contained non-European variants. This explanation provides the matrix in which DeGrafFs (1999a) position that children too played a role in the retention of substrate features in Creoles makes a lot of sense. Clearly, quite a few xenolectal features were present in the feature pools from which Creole children were developing their respective idiolects. The ecologies in which this process took place did not enable them to distinguish original native features of the homestead phase from xenolectal ones of the plantation phase during which not only did non-Europeans form the overwhelming majority but also, increasingly, most of this majority were African-born adults (Baker 1993a). Through their selections, Creole children became critical agents who gave long life to some of the substrate features or, in most cases, the native features that were (partly) congruous with them. This scenario is more consistent with history and language

Competition and Selection in Language Evolution "acquisition" theories than most of the hypotheses conjectured to date about the development of Creoles, especially the language bioprogram and the relexification hypotheses. I have thus far highlighted differences in contact conditions between the development of Creole and non-creole vernaculars in the colonies, not the absence or presence of contact. As explained in Chapter 3, language change has been contact-based everywhere, even if this involved only the contact of different idiolects in a population (as in the New Zealand communities discussed by Trudgill 2004 - see also Mufwene 2006), especially where there have been changes in population structure. These typically produce changes in patterns and dynamics of inter-individual interactions. Therefore, they too produce changes in the "balance of power" among competing variants. As explained in Mufwene (2001a, Chapter 2), even in the case of language or dialect contact, contact always boils down to that between idiolects, including xenolectal ones. Populations and their languages do not get in contact like armies or sport teams with explicit rules of engagement; nor is there contact when they only share a geographical space but are segregated and do not interact with each other. Rather, they are in contact when their members interact individually with each other, negotiating their individual roles in relation to each other and accommodating each other accordingly. Although the "invisible hand" question arises again, regarding how individualbased patterns spread into population-wide patterns, we must really focus on inter-idiolectal patterns to understand how selection works, notwithstanding the fact that from a historical perspective, we cannot avoid extrapolating to populations and communal "systems." Congruence, which favors features shared by the varieties in contact (languages, dialects, or just idiolects), is itself a factor that drives selection. Non-congruent xenolectal variants are often left out, especially at the level of grammar and phonology, notably if the populations that do not speak the target language natively are socially integrated. In this case, the xenolectal features tend to be stigmatized and avoided by their carriers' own children, especially if the parents are also demographic minorities. Chronologically, the settings of the emergence of Creoles are special mostly in having critical masses of xenolectal speakers, in ways that are in fact comparable to those where indigenized Englishes and African French varieties have emerged and where native models can easily be ignored. In any case, whether or not there is (extensive) congruence, every new colonial variety of a European language is contact-based, by the same competition-and-selection mechanisms that drive language evolution, subject to the specific contact ecologies of their emergence, as made more evident by sound research on the development of Creoles. However, as Robert Chaudenson has repeatedly reminded us, one should not overlook the role of the "lexifier" itself in the emergence of the new varieties (1973ff). I argue that selection is made possible largely thanks to the nature of the target language itself both as a species and as a "complex adaptive system". As explained in Mufwene (200la), language is a species because, at least in the way it is conceived of by linguists, it exists only as an extrapolation from similar but varying idiolects, just like a biological species is an extrapolation

119

120

Language Evolution from the existence of individual organisms. It is also a "complex adaptive system" because its structure is multi-modular. There are in almost every module units (viz. sounds, morphemes, words, and idioms) and usage principles (production and combinatorial ones) which compete or alternate with each other, at least in some contexts of use. All of these are units to which selection applies while an individual develops competence in a language and learns, for example, whether the word bucket or pail is the one they should normally use in their vocabulary; whether they will pronounce the first syllable of the word neither with the diphthong /ay/ or the monophthong /i:/; whether they will say / asked you not to go or / asked you to not go (interchanging the positions of to and not); whether they will say we was young or we were young', and whether they will say he was a-huntin or he was huntin(g). The selections an individual makes are similar to those occurring by gene recombination at the conception of an offspring, for instance, whether he/she will inherit brown or blue eyes, fat or thin lips, brown or blond hair, etc. from (one of) his/her parents, in the sense that the genes responsible for the phenotypes are only dominant but have not eliminated the recessive alternatives from the gene pool. In just the same way that another offspring may very well have any combination of the alternative genes as dominant and therefore different phenotypes, another language learner may select any alternative combination of the linguistic features into their idiolects, keeping their respective recessive variants handy for those occasions when they can become useful. A question that has received some attention in the study of the development of Creoles is what principles or constraints bear on such unconscious selections into speakers' idiolects. More challenging yet is the following question: Under what particular conditions do such individual selections become group selections and part of the communal speech norm? I return to these questions below, though I must alert the reader not to expect much about the second. Selection is also made possible by the fact that a language is not "acquired" wholesale, with the full set of its structural features, in the same way that a biological organism - especially an animal - inherits its genotypes, at conception, from its parents. As explained in Mufwene (2001a), a language is "acquired" piecemeal. Every individual speaker builds his/her idiolect gradually, as he/she matures cognitively, in the case of children, and as his/her communicative experience in the target language increases. Thanks to population structure and to polyploidy (which enables more than one speaker to exert convergent influences on him/her), selections in each learner/speaker largely reflect their interaction histories, hence the particular subset of the E-language he/she has been exposed to and variants that are dominant in it. We can, of course, not ignore the role of the learner's/speaker's personality where the selections involved are conscious, such as in academic or political language, and the learner/ speaker is also guided by a particular ideology. These are some of the same factors that account for hybridism in language, interpreted here as the mixing into the same "system" of elements from different ones. An idiolect is of necessity a hybrid from various idiolectal inputs.

Competition and Selection in Language Evolution

7.2 Hybridism in the normal and natural development of Creoles My concern with the questions of hybridism and selection started with my research on the development of Creoles, whose emergence has almost alone been associated unjustifiably with language contact (Chapter 3). Although Thomason & Kaufman (1988) and Thomason (2001) stipulate that, being mixed languages, Creoles cannot be assigned genetic classifications, creolists do actually speak of "lexifiers" of Creoles. These are languages from which Creoles have inherited the largest proportions of their vocabularies, often upwards of 90 per cent. This is why they are usually identified, for instance, as English, French, or Portuguese Creoles. I reject the traditional stipulation that Creoles have developed their grammars from sources that either rule out inheritance from the lexifiers or treat it as only marginal, suggesting that they are made de novo by children, thanks to the language bioprogram (Bickerton 1984b, 1999), or that they are made of rules selected from particular substrate languages (e.g., Alleyne 1980, 1996; Holm 1988, 2004; Lefebvre 1998, 2004). I am more sympathetic with approaches that underscore both inheritance from the lexifier and influence from substrate languages in all aspects of their structures. However, I submit that like in other cases of language recreation during the "acquisition" process, inheritance from the "lexifier" is modified under the influence of substrate languages or of the emergent "system" itself. For instance, the preposition for can be used as an OBLIGATION modal auxiliary in English Creoles, as in the following Gullah example: Jean binfuh [fa] come "Jean had to come" or "Jean was expected to come." This extension of its usage in English was made possible in part by the fact that adjectives and prepositions can be used predicatively without a copula, as in Lorry (very) pretty "Lorry [is] (very) pretty" and Fonzo in the house "Fonzo [is] in the house." One must also clearly distinguish between ORIGIN of a feature and INFLUENCE on the use of that feature. Often these notions are not coextensive, as usage of a feature from the "lexifier" has been influenced by the way a similar feature is used in substrate languages. That is, selection applies in more than one way, one of which is the particular way in which the external ecology provided by the substrate languages influences the selection of features from the target language into the learner's idiolect. From the point of view of language contact, where the concerns of diachronic linguistics have focused, hybridism arises from those cases where the substrate languages introduce into the learner's idiolect features or structural elements not attested in the lexifier. Much of this is illustrated in Chapter 8. My position is motivated by the following considerations: The grammar of a language is not independent of the morphemes and words that are used in it. It is a function of how these units are used in a language (Bolinger 1973). It is a set of generalizations over their respective behaviors in combination with like units. For instance, in English as in other languages, all words must have a base morpheme, such as code in encode and decode.

121

122

Language Evolution While a word can consist of a base morpheme only, it cannot consist of derivative morphemes only. Thus, *ende (intended to be pronounced as [endi]) or *deen (intended to be pronounced as [dien], with two syllables) are not acceptable words in English if they represent combinations of the prefixes de- (pronounced [di]) and en-(pronounced [en]). Many of these basic grammatical constraints vary from language to language and each module has constraints specific to itself. For instance, not counting AAVE, English syntax does not generally allow main clauses in which adjectives are used predicatively without a copula. One must say John is taller than Mary but not *John taller than Mary. However, Chinese and Kwa languages (of West Africa) allow such copula-less main clauses. In a language-learning setting where only one language is targeted and the learner is not exposed to a competing strategy provided by another language, he/she would normally approximate the pattern provided by speakers of the target language. In language contact settings, variation is often provided by all the competing linguistic "systems" that the learner is exposed to, which fosters competition and selection during the development of their idiolect. This makes it possible to mix in one's idiolect elements from different communal varieties. When such mixed idiolects reinforce each other, the evolution produces at the population level the kinds of structural divergence observable in Creoles. Because such variation often also exists within the same language, between alternatives provided by different dialects, speakers "acquiring" a language in a naturalistic setting experience basically the same kind of competition and selection. Thus, after living in the American South for a good while, an American English speaker who doesn't normally speak southern English may be heard using yall and yall s for the PLURAL of you and yours, respectively. Children of northern parents growing up in the American South in households where parents still scorn southern features often have idiolects that include subsets of southernisms even if they claim that they do not speak southern English. Likewise, a person who does not otherwise speak nonstandard English may be heard using airft (as an emphatic variant of isntlis not) or double negatives. The pressure from one's parents and/or from one's teachers to avoid nonstandard features is usually not sufficient to counteract the need to sound like those speakers one interacts and/or wants to be identified with. Somehow, in the effort to want to please both parties, the speaker/learner develops an idiolect that is an interesting linguistic hybrid. As one can imagine, if such speakers formed their own separate community or became the dominant members of a new one, a new English variety might emerge that is neither entirely southern nor really northern, which is perhaps the reality in transitional areas from the American South to the North, as dialect boundaries are normally fuzzy. From a diachronic perspective, the emergence of Creoles as normally mixed systems need not be conceived of differently from the above discussion. An important exception is that the conditions under which their "creators" targeted the "lexifier" led them to modify its units and principles in more conspicuous ways than one can observe in other cases of "language acquisition," when the feature pool contains elements attributed to only one language. In the

Competition and Selection in Language Evolution case of Creoles, the multiple; native and xenolectal origins of the feature pool to which speakers were exposed sometimes enabled influences from languages other than the "lexifier" to prevail and therefore produced more extensive modifications of the target language. Thus, although English provides the model construction Katie is (very) pretty, in which the copula is required before the predicative adjective pretty, most English Creoles have dispensed with the copula and produce something like Katie (very) pretty, a construction that is partly patterned on some substrate languages.3 Such mixing of grammar is often also facilitated by a perhaps minor feature of the "lexifier" itself, which has prompted creolists such as Chaudenson (1992,2001,2003), Come (1999), and myself since Mufwene (1986b) to invoke the role of the (partial) congruence of the competing linguistic systems. That is, the divergence of the emergent grammar is a consequence not just of the sztrong xenolectal pressure from the substrate languages but also of the ways that xenolectal speakers recognized similarities between the target structure and a substrate pattern. By the principle of least effort, they capitalized on it, usually extending it even to areas where it does not apply in the target language. For instance, nonstandard English provides a possessive use of me in me child. Speakers of languages in which POSSESSION is expressed by preposing the POSSESSOR noun or pronoun (with an invariant form in this case) to the head noun easily interpreted this POSSESSOR + POSSESSED pattern in a more general way. Thus, they also produced constructions such as you chile "your child," he chile "his/her child," we chile "our child," and deh/dem chile "their child," just like Sara chile "Sara's chile," which are attested in Gullah and other English Creoles. These vernaculars' pronouns also tend to have the same forms for the SUBJECT, OBJECT, and POSSESSIVE functions, though Gullah shows some variation (Mufwene 1993c). As a matter of fact, a congruence of many factors can be invoked to account for some of the more common cases of structural divergence in Creoles that have been attributed almost exclusively to Creoles. For instance, as shown above, the selection of copulaless constructions such as Katie/She pretty "Katie/She [is] is pretty" and Kim the talles'"Kim [is] the tallest" should be attributed not only to the fact that several substrate languages are copulaless but also to the following other factors: (1) the absence of a meaning associated with the copula; (2) the absence of a category ADJECTIVE in several non-European languages (Dixon 1982) and the fact that the predicative meaning conveyed by an adjective in European languages is conveyed by a verb in the RESULTATIVE aspect in many non-European languages; and (3) the fact that English allows tenseless, secondary predication in which the adjective is not linked to the subject by a copula, as in He swam naked and The chief wanted him alive.4 Another example is the selection of done as a PERFECT marker in English Creoles. Regardless of its actual evolution, the morpheme selected in this case was not the past participle form of the verb do but rather the adjective done, with the meaning "finished," as in the nonstandard English construction I'm done talking to you. Thanks to the previous paragraph, we can tell why the copula has not been retained in the construction, especially because the emergent basilectal grammar generally dispensed with the copula (if we agree that de [de]

123

1 24

Language Evolution in Kyari de huom "Carie is at home" in Jamaican Creole functions as a locative verb rather than as a simple copula). This evolution, which favored the be done PERFECT construction over the alternative with have + PAST PARTICIPLE, was facilitated by the fact that many substrate languages, especially those of the Kwa family, use the same grammatical strategy to express the meaning of PERFECT, modifying the main verb with one meaning "finish." Many concurrent factors obviously influenced feature selection during the development of Creoles. This is in no way different from the congruent influences that bore on the evolution of non-creole languages, regardless of whether or not contact is acknowledged as a factor. For instance, the development of relative clauses such as in the man -with "whom Joyce went out in standard English is a congruent legacy of both Old English, on the one hand, and Latin and Norman French, on the other. Since Old English, this pattern, involving piedpiping the preposition with the relative pronoun, has alternated with the more common alternative of colloquial and nonstandard English the man (that) Joyce went out with, in which the preposition is stranded, the relative clause starts with the complementizer that, and the complementizer itself can be omitted. This latter option does not exist in either Latin or French. The nature of the contact between, on the one hand, English and, on the other, Latin and French, in which the latter languages were prestigious and used by the intellectual elite and the upper class, explains the kind of impact that they had on its evolution. The pattern of the language of the lower class has continued in colloquial and nonstandard English, whereas that of the upper and elite classes, influenced by foreign languages, was selected for standard English. This bifurcated evolution is also consistent with other specializations in English, where a word such as perceive is part of standard and educated speech, whereas see and feel, whose meanings it conveys, are more general and the only alternatives in colloquial speech. Perceive would be out of place in vernacular speech. There is actually evidence of such socially stratified selections in English Creoles. Relative clauses start with weh (which behaves more like a complementizer than a relative pronoun) or a null complementizer, as in de man (weh) Joyce go out wid "The man (that/who) Joyce went out with". (One cannot say *\de man wid weh Joyce go out.) Weh appears to have been modified from the nonstandard English relativizer what (one of whose pronunciations was [waet], at least in some varieties), as in everything what you said. Its selection into English Creoles reflects the fact that their lexifiers were not the standard varieties of English. These are certainly not what the indentured servants that the slaves interacted regularly with spoke. They spoke vernacular, nonstandard English varieties.

7.3 Variational evolution: development and change by selection It may not be unnecessary to point out that it is because of selection that a particular language can now be singled out as the "lexifier" of a Creole. Pace Baker (1990), socioeconomic

Competition and Selection in Language Evolution pressures on the populations that were brought in contact in the relevant colony had the effect of them targeting a particular language, that of the population controlling the contact setting economically and politically, as their vernacular. The main reason is that the socioeconomically powerless members of the contact communities, starting with children of the homestead societies, would have found it more advantageous to be fluent in the language of the economically dominant group. Slaves of the plantation phase would simply follow the example of the Creole slaves in adapting to the new setting (Chapter 11). Bearing in mind that the appropriation of the colonial vernacular was also a restructuring process, we could thus argue that the development of Creoles started with the selection of one of the languages in contact as the vernacular. This is similar to selections made routinely by most speakers around the world in multilingual and/or multidialectal communities, viz. they select as their vernacular and target (under ethnographic conditions specific to their daily lives) the language variety that appears to be the most advantageous to them.5 In those socioeconomic ecologies where Creoles developed, a nonstandard "lexifier," more precisely, a heterogeneous cluster of several vernaculars concurrently converging toward a colonial koine, was that particular target. The fact that a language is "acquired" gradually and piecemeal, making its features the actual units of selection, made possible the selective restructuring which, during its appropriation by the Africans, produced varieties that reintegrate, albeit with modifications, not only features of the "lexifiers" but also some xenolectal ones selected from the substrate languages (Mufwene 2001a,2005a). When one "acquires" a language as a nonnative variety, the "acquisition" of every unit and principle is actually in competition with what one already knows if they are not the same. Interference from the language one already speaks, which can cumulate into substrate influence at the population and communal language level, reflects those cases where options previously known to the speaker prevail and are selected into the emergent "system," for instance, when an English speaker fails to aspirate voiceless stops in a stressed prevocalic environment (as in the word put) or omits a copula before a predicative adjective in the PRESENT tense. Although such influence happens despite every effort the learner makes to speak and to command the target language in the best way they can, it is an instantiation of selection from among competing alternatives in a feature pool including both native and xenolectal options. It is one of the forms of selection in language development, one of those that account for communal language change when such modifications spread across the population, beyond the level of just a few idiolects.6 Thus, the spread of innovations (as mere imperfect replications, as exaptations of extant materials, or as introductions of new elements) from a few idiolects to the communal language also represents another aspect of selection in the life of a language. Only some innovations, not necessarily by the same speakers, spread across the population while most others do not. This particular line of evolution follows from the fact that although a language is used by a community, it is "acquired" and spoken or signed individually. It is also

125

126

Language Evolution spoken or signed differently by every individual speaker or signer, which accounts for the fact that the idiolects of a language resemble each other only on the Wittgensteinian familyresemblance model. A consequence of this inter-idiolectal variation is that every time a new member joins a community by birth or by immigration, they face choices to make from among the individual models they are exposed to. The choices need not be conscious, but they occur nonetheless, dictated in part by what particular speakers or signers a learner wants to be associated with. Making the matter more complex is the fact that different speakers/signers are targeted for different features, so that, even under conditions of perfect replication, no particular speaker/signer would replicate another. Selection also takes place through accommodations that speakers make to each other in their respective communicative networks. As speakers/signers interact, they also try to align themselves with each other through a series of mutual accommodations. Although these are not necessarily symmetrical, the processes account for the particular ways in which the idiolects of speakers who belong in the same network converge and can function as group identity markers. Regardless of whether the features that are salient in such convergences are endogenous or exogenous, the accommodations that account for their spread within the relevant networks are tantamount to selections that speakers/signers make favoring them over their competitors. The reasons, which we need not get into here, are diverse, ranging from more effective communication to being a la mode. Since networks vary among themselves, very often alternatives that one abandons are still preferred by others and variation is maintained within the larger population of speakers. However, due to particular factors in the ecology in which a language is spoken or signed, some variants may disappear, new ones may be introduced, and the overall pattern of variation changes. Hence, there is evolution, whose ultimate origin lies in the selections that speakers make during their communicative acts, while accommodating other speakers or innovating new structures or uses. In the case of Creoles and other so-called "mixed" languages, these selections are more obvious because one can more easily tell when there is xenolectal influence. However, even in such cases, language "acquisition" is still individually based and the mind of every individual learner/speaker is an arena where different options from the same target language and/or from the languages known to him/her compete with each other. While Creoles make competition a more obvious factor in the life of a language, they also highlight the reality of what takes place even in those cases where presumably only one language is targeted. In the end, it all boils down to a speaker being driven by some ecological pressures to integrate only particular features as dominant in his/her idiolect. Thus, while idiolects of a particular language variety are similar, the subsets of units and principles that prevail in them as dominant vary from one another, reflecting the individual personalities of their speakers/signers and their particular interaction histories. For the student of language evolution, what matters is whether the total population of features in competition and/or

Competition and Selection in Language Evolution their distributional patterns within the community of speakers/signers change. This is really what evolution in a communal language amounts to. A challenging and so far elusive question is: What is the relationship between the individual selections made by particular speakers/signers and the group selections that map the evolutionary trajectory of a language?

7.4

More on hybridism in language

The foregoing helps me return to the question of hybridism in language. Creoles make it more obvious because they are quintessential cases that illustrate how influences from different languages can produce a new language variety. However, they also show clearly that it is never a simple matter of the "lexifier" putatively providing the vocabulary and the rest of the system originating elsewhere. Much of the grammar also originates in the "lexifier" although it is hardly maintained intact. Like in gene recombination, units and principles selected from different varieties of the "lexifier" are restructured into a new "system." This explains why Quebecois is quite a new French variety different from other nonstandard French varieties and even from Louisiana French, which developed under similar sociohistorical conditions. In the case of Creoles, the substrate languages often contributed by congruence, although they have also contributed peculiarities not attested in the "lexifier." In the case of Atlantic Creoles, the more indisputable evidence can be found in Berbice Dutch (Smith etal 1987, Robertson 1993, Kouwenberg 1994), to which I return in Chapter 8. However, as pointed out in Chapter 3, one can posit similar restructuring processes in the development of the Romance languages, from the contact of Latin and Celtic languages, and in the development of English, originally from the contact of the Germanic languages that were brought to England among themselves and with the Celtic languages, too, before Old English came in contact with Old Norse and Norman French. Without selection-guided restructuring under contact conditions, it is difficult to explain how these languages have evolved in the specific ways they have. So, the Romance languages and English are as mixed as Creoles are, in that they have developed by the same language evolution formula, with of course different values assigned to variables in every case. This is true of the evolution of other languages, too, where much less is known, or admitted, regarding language contact. The history of population movements and contacts in Eurasia and Africa alone over the past 5,000-6,000 years suggests that we cannot adequately account for the speciation of languages within their respective families (e.g. Indo-European or Bantu) without factoring in language contact. In all these cases, selections were made from among competing alternatives. Those selections account for differences between the Romance languages and Vulgar Latin and among the Romance languages themselves. They also account for differences between English and continental European Germanic languages and among the different English varieties. Much of the same explanation applies to differences between Bantu and Proto-Bantu and other

127

1 28

Language Evolution languages of the Niger-Kordofanian family, as well as among the different Bantu languages. In the case of the Romance languages, we can invoke contact of Vulgar Latin with the different Celtic languages and subsequent contact of the emergent varieties (Old French and Old Iberian in particular) with different superstrate languages, e.g., Prankish in Gaul and Arabic in Iberia. In the case of English dialects, we must take into account layers of contact between the different languages brought by the Germanic invaders into England and the subsequent contact of several stages of English with the insular Celtic languages, with Old Norse and Danish, with Norman French, and with various languages outside Europe. The case of Bantu languages conjures up contacts between speakers of pre-dispersal Bantu with those of Pygmy, Khoekhoe, and San languages, as well as contacts among the Bantu languages after they had already diversfied. All of these outcomes depended on the specific natures and conditions of contacts in particular settings and the particular ways that selection resolved or failed to resolve the relevant competition. The whole process of speciation in genetic linguistics is a byproduct of competition and selection in different contact settings, as a protolanguage was moved from one location to another and thus brought in contact with other languages. Population movement and contact have usually played a role in language speciation, and the particular dynamics of idiolect contacts account for the paths of language change and the development of new dialects and/or languages. The actual arena of competition and selection in all such cases, including the development of Creoles, lies discontinuously in individual speakers. Unless they interact with each other (and thus have the options of accommodating each other), there is no language contact. Part of the accommodation is made possible in the speakers' disposition to learn the other's language or even shift to it, thereby influencing its system. Thanks to geographical or social isolation, an accumulation of such influences can lead it to speciate into separate language varieties. If Creoles instantiate such outcomes, so too do the numerous colonial varieties of European languages that emerged around the same time and later.

7.5

Constraints on selections

The contribution of this section to research about language evolution lies more in formulating questions that cannot continue to be dodged than in proposing conjectures worth considering as probable answers. The question of what constrains the selections made by speakers remains an important one. One fold of this, and the one that has received the most attention to date, is what favors a particular variant over its competitors. Another, which has typically been overlooked but should be informative about the compatibility of various typological options, is whether the extant structures of the emergent language variety bear on what else can be allowed in to modify it.

Competition and Selection in Language Evolution Traditional accounts in linguistics addressing the first question have invoked markedness principles (reviewed in Mufwene 1991b, 2001a), which determine what particular options were unmarked, i.e. (likely to be) favored or preferred, under what specific conditions of competition. The competition of course takes place where there is variation along a particular phonological, morphological, syntactic, or semantic parameter. For instance, given a choice between a PAST tense construction with an inflection on the verb (e.g. -ed in English) and one with a free periphrastic marker (e.g. been) that combines syntactically with any predicate (be it verbal or non-verbal), which one is likely to prevail and why? Did the fact that an English creole allows non-verbal predicates to head predicate phrases (as in Don (ben) taller 'an me "Don is/was taller than me") bear on the selection of ben/bin as a PAST/ ANTERIOR tense marker? Or was it only a question of what marker was more salient?7 If it were just a matter of salience in the perception of the markers, why have forms such as dead, gone, and broke (which used to also function as a past participle and had an adjective-like behavior) survived in Creoles over the variants die, go, and break, which are not perceptively less salient but also have a wider morphosyntactic distribution?8 Note that the English "system" allows these variants in the INFINITIVE, in the PRESENT and FUTURE tenses, and in the IMPERATIVE. In an ecology-free setting, this distribution would make these variants more frequent and favor them, as they do not appear to be less salient.9 Are there nonstructural reasons that have favored these particular suppletive forms that can also explain why the past tense alternative went was not favored over gone? The same question arises about why the suppletive forms ran and brought were not selected over the base variants run and bring, which are the norm in basilectal English Creoles. These kinds of questions, which arise from the study of English Creoles in particular, should be informative about the mechanisms that drive evolution both in idiolects and in communal languages. They boil down to the following question formulated in Mufwene (1991b): What criteria can one propose for determining markedness values without being trivial and uninformative, e.g. Creoles have selected the particular structural options they have because those options are unmarked, with the proof lying in the selections themselves? Such an explanation would of course be circular, and one must resort to factors which can explain these specific choices. Regardless of whether or not they are addressed in an ecology-sensitive approach to markedness (Mufwene 1991b, 2001a), or in the framework of optimality theory (Fill 2004), the subject matter calls for more inquisitive research. Typologists, such as Croft (1990), have suggested that there are universally unmarked parametric options. It is important to clarify that the primary criterion for this position is the statistical distribution of particular structural options among the world's languages. For instance, there seem to be fewer languages that use numeral classifiers (e.g. head in two heads of cows/children in some English dialects) than those that do not. However, the approach has an inherent problem, viz. that the proportion of languages does not necessarily correspond

1 29

130

Language Evolution to the proportion of people using those options.10 A case in point: Chinese language varieties fall in the category of those languages that use numeral classifiers. If one adds some Indie languages to this category, it is no longer so certain that the numeral classifying system is more marked than the singulative option common in (Western) European and sub-Saharan African languages, expressing numbers without a classifier (as in one cow and two cows). By headcount of speakers, there are definitely more users of numeral classifying systems than of singulative systems. Hence, we could also argue that the numeral classifying option is less marked than the singulative system. There is, however, another kind of question that is more relevant to this Chapter: How relevant is such typological ranking to a person "acquiring" a language who does not have one or some of the typological options available to them? Is the fact that Creoles of the Atlantic do not have numeral classifiers a function of the assumption that singulative "systems" are typologically unmarked? Or is it rather a function of the fact that the numeral classifying option was hardly part of the competition in the ecologies at the time of their formation? Note that, like their "lexifiers," they still use constructions such as hand/bunch of bananas. If a universal, rather than an ecology-sensitive, scale of markedness is what applies in any setting of language "acquisition," what should we make of the fact that Melanesian expanded pidgins, which are structurally as complex as Creoles but developed in settings where numeral classifying is the norm, have a fossil numeral classifying system? They use pela (from English fellow) with numerals and demonstratives, regardless of whether the noun would be considered COUNT (e.g. "cow") or MASS (e.g. "water") in English. Did not their "creators" simply favor an option more familiar to them, just like the "makers" of Atlantic Creoles did? Did not they also select an option simpler to them, rather than one that would translate classifiers that vary from one language to another? The reduction of the number of classifiers, compared to the more indigenous languages, need not be interpreted as evidence that the numeral classifying system is more marked than the Western European singulative alternative. The "creators" of Melanesian pidgins could as well have done without a classifier at all. As will become more obvious in Chapter 8, evidence for the hypothesis that the "creators" of Melanesian pidgins must have been guided by an ecology-specific scale of markedness comes from the fact that they have in fact complexified, rather than simplified, the morphosyntax of the "lexifier" by introducing distinctions such as DUAL vs. PLURAL and INCLUSIVE vs. EXCLUSIVE, as well as combinations of these distinctions in their pronominal systems. In this context, it helps to compare the linguistic situation with a biological one. If an individual has been conceived in a society where the pigmentation correlated with race varies almost exclusively within the Negroid range, does it matter that there are other options in other societies? Would not black and kinky (rather than brown and straight) hair be the normal, common, hence unmarked, expectation? Is it not really the options available within a particular community that matter? In cases of hybridization, what principles determine

Competition and Selection in Language Evolution whether an offspring will have blue or brown eyes, especially when the parents have different phenotypic traits on this particular parameter? Such considerations have led me to submit that markedness factors must be assessed relative to the ecology in which a language is spoken, articulating those factors that matter to speakers, such as what option is the most common, or the most transparent, or the most regular, the most salient, not semantically empty, etc. (Mufwene 1991b, 2001a). Thus, what may be unmarked in a particular ecology may be marked in another, and only subsets of these factors apply in different cases. It is such considerations that lend plausibility to the often-repeated claim since Bickerton (1981) that Creoles have selected unmarked parametric options. One can say more accurately that languages often evolve by favoring unmarked options over marked ones where such competition is the case. These observations do not rule out the fact that some of these markedness values obtain in many typologically different languages and may be claimed to be universals. However, even universals must be verified by language-specific phenomena and perhaps the apparent universality of such cases maybe grounded in the cognitive infrastructure of language itself. The question remains open. Still, what matters the most in this case is that there is often cross-linguistic variation in the ways these values are assigned to competing variants, so much so that even Creoles that developed in different contact ecologies from more or less the same lexifier (e.g. nonstandard colonial English in the Caribbean and the Pacific) have not always selected the same variants. (See also DeGraff 1999b for a slightly different take on this subject.) This is not really the right juncture to carry on the discussion with which this Chapter is ending. I have mentioned markedness here only to show that in a way linguists have long dealt with alternatives competing for the same functions, hence with competition and selection. It is not obvious that markedness considerations account for all the selections. There are also many cases where there is no competition, like the fact that a particular language may present no variation regarding the position of the determiner in a noun phrase. However, one cannot deny the fact that competition and selection play a role in language "acquisition" and evolution.

7.6

Conclusions

Idiolects are "complex adaptive systems." So are communal languages as extrapolations from idiolects. In their evolutions, they reflect selectively adaptations undergone by idiolects. Because they exist by virtue of being spoken by individuals, communal languages are inherently variable. Variation in their "transmission" and "acquisition" is amplified in contact settings, because the target language and the other languages that it is brought in contact with make concurrent contributions to the feature pool from which the learner creates their idiolect. (See Aboh 2007 for a similar position, focused on syntax.) Although the units and

131

132

Language Evolution principles are predominantly selected from the target, there is usually (congruent) influence from the other languages, and there are often borrowings. In monolingual settings, there is competition and selection simply because individual speakers contribute variably to the communal pool from which the learner draws the materials for his/her idiolect. The new idiolect is a replica of none of the extant ones. It is a hybrid produced by allopolyploidy, with congruent influences coming from several speakers but with each speaker impacting variably the new idiolect as it develops. Competition and selection are thus inherent in the dynamics of language evolution. Assuming a feature pool from which every idiolect and, by extrapolation, every new state of the communal language draws its units and principles and recreates a new "system," every language is naturally a hybrid of some sort. Those that have recently been influenced by other languages, or emerged from the contact of several languages, are more obvious cases than others. Thus, as shown in Chapter 3, the history of modern mankind shows that every language spoken today has either been influenced by others or emerged from the contact of other languages spoken before it.

Transfer and the Substrate Hypothesis in Creolistics1 Chapter Outline 8.1 8.2 83 8.4

8.1

Introduction The substrate hypothesis Transfer and the substrate hypothesis Conclusions

133 136 149 158

Introduction

This Chapter was published much earlier than all the others in this book. A revised version of it was to be published in The Ecology of Language Evolution (Mufwene 200 la) but was removed because the book was becoming too big. It is included here because it contains data that further illustrate how, under particular ecological conditions, competition is often resolved in favor of substrate languages. Like the other Chapters, this edition is substantially different from the original 1990 publication in Studies in Second Language Acquisition and even from what would have been published in The Ecology of Language Evolution (2001), especially because it does not repeat those views that I have abandoned or articulated in more informative ways since then. For instance, I no longer believe that research on L2 acquisition (SLA) can tell us much about how Creoles evolved, at least for the following reasons: (1) as well argued by DeGraff (1999a), both adult L2 learners and LI learning children must have contributed to the gradual divergences that produced Creoles, the latter especially by selecting from the feature pool some of the substrate features contributed by their parents (Chapter 10); (2) research on SLA has focused on individual learners, often showing variation from one learner to another without focusing on the possible emergence of group norms, whereas genetic creolistics has focused on norms, therefore on the communal patterns of the new vernaculars called Creoles; and (3) the interlanguages on which research on SLA has focused are transitional idiolectal varieties toward closer approximations of the target langugage (TL), whereas Creoles are communal evolutions in the opposite directions, away from the "lexifier," by basilectalization (Chaudenson 1992, 2001, 2003; Mufwene 2001a, 2005a). For more discussion on the subject matter, see Mufwene (to appear).

8

1 34

Language Evolution The original purpose of this Chapter was two-fold: first, to highlight the development and various interpretations of the substrate hypothesis in creolistics, and, second, to identify ways in which findings in research on transfer can enrich that on substrate influence, and vice versa. It was then motivated by an assumption common among creolists that Creoles are possible outcomes of SLA, based especially on Alleyne (1971, 1980), Thomason (1980), Thomason & Kaufman (1988), and Holm (1988).2 The reason for including this revision of the essay in this volume is the quest to account for how some features from some substrate languages, but not necessarily all of them, made their way into the emergent norms of Creoles. Here, more than in Chapter 7, I show a number of substrate features that were selected over other structural alternatives and attempt to articulate the ecological factors that favored them. Although the evolutionary trajectories of expanded pidgins are not quite the same as those of Creoles, especially since the latter did not start as pidgins (Chaudenson 1992,2001, 2003; Mufwene 2001a, 2005a), they are alike in evolving in the direction of their basilects, with grammars that diverge most extensively from their "lexifiers." They are also alike in functioning as vernaculars, rather than as lingua francas, in the communities that speak them, although expanded pidgins still share the same geographical space with some of the substrate languages that influenced their evolution. I will thus not discriminate between these varieties and will select examples of substrate influence from both. I will adduce examples as easily from Melanesian pidgins as from any of the Atlantic Creoles, with which I am more familiar. The term transfer is used here in the same way that Weinreich (1953: 1) uses the term interference for deviations within individual speakers from the TL norms that are due to influence from languages they "acquired" previously. For convenience sake, I follow the literature in referring to these as substrate languages. Substrate influence is thus used in reference to influence on the structures of Creoles that originated in such languages. For the purposes of this Chapter, the connection between transfer and substrate influence is assumed in terms of cause and effect. Transfers apply in the speech of individual multilingual speakers and/or at particular stages of SLA. Substrate influence is the outcome of the cumulation of transfers in a communal variety that has developed its own separate norm. It may be due to the fact that several speakers sharing the same substrate language or typologically related ones transferred the same features into what would evolve into a new language variety. It may also be due to the fact that, by the Founder Principle or for other ecological reasons conferring a privileged status to a minority of speakers, some features of a particular substrate language or of a handful of them spread disproportionately in a particular population. Much of the action of the "invisible hand" is involved in the process on which even the competition-andselection approach cannot shed much light yet, since little is known about the dynamics that convert features of some idiolects (and different idiolects for different features) to communal characteristics. Nonetheless, it is after transfers have been replicated by different speakers

Transfer and the Substrate Hypothesis in Creolistics over a period of time and thus become established as part of the emergent norm (even as variable features) that they can be characterized as substrate influence. Once the new vernacular has stabilized and normalized, substrate influence need not be associated with multilingual speakers and/or SLA. Some of the best evidence for substrate influence can actually be collected from monolingual or monolectal speakers, especially in the case of Creoles and expanded pidgins. Today's Creole speakers have no knowledge of the substrate languages and many speakers of the expanded pidgins have little or no knowledge of the relevant substrate languages, although their idiolects still include substrate features. Like the original edition of this Chapter, the present version includes a brief synopsis of how the substrate hypothesis has developed in creolistics. However, some issues are not addressed in exactly the same ways. For instance, I now take more to heart some of the problems associated with proving substrate influence, especially in the sense of what Allsopp (1977) identifies as apport, i.e., elements not attested in the "lexifier" that have been selected from a substrate language or a group thereof. I am convinced that outside the Melanesian islands (per Keesing 1988) and the unusual colonial situation in Berbice (Smith etal. 1987, Robertson 1993, Kouwenberg 1994), most of the substrate influence can be associated with congruence and need not be interpreted as "apports" from substrate languages (Corne 1999, Mufwene 200la). I therefore articulate more forcefully the difference between those cases where an influence originated unequivocally from outside the "lexifier" and those where substrate influence was congruous with a variant that was already in the "lexifier" and favored the selection of the variant over another or others. In the latter case, the selected variant may have been influenced to the point of evolving additional characteristics not attested in the "lexifier," as in the often-cited case of serial verb constructions, to which I return below. Worth considering here would also be those cases where the "lexifier" was replaced by another "superstrate" language, such as Dutch in Suriname. The fact that the Dutch, who replaced the English in this colony, adopted English as the lingua franca for communication with their slaves cannot just be glossed over. English was not native to the Dutch, and there is no particular reason why they can be assumed to have "acquired" it perfectly, especially in the low ranks of the colonial European society who interacted regularly with the slaves. The situation is reminiscent of that in many parts of the world today where indigenized English varieties have evolved largely through the transmission of nonnative features of the linguistic models to L2 learners and not necessarily, or not exclusively, from the introduction of new ones into the emergent varieties by the learners themselves. The influence of Dutch on Surinamese Creoles, especially Sranan, such the selection of sa as a FUTURE marker where other English Creoles have selected go, gwine, or a go, is worth investigating. Bunting (to appear) raises the interesting question of whether distinctions among verba dicendi ("verbs of saying") in (nonstandard) Dutch may have contributed to the fact that Sranan is rare in adopting taki, rather than say (as in other English Creoles), as the basic verbum dicendi.

135

136

Language Evolution Unfortunately, to keep the Chapter manageable, I do not address below this interesting fold of the feature pool in settings where Creoles evolved.

8.2 The substrate hypothesis 8.2.1 A brief historical survey Over the past 30 years or so, the substrate hypothesis has been articulated mostly in contrast with Bickertons (1981, 1984b, 1999) language bioprogram hypothesis (LBH). According to the latter, grammatical features of "radical" or "prototypic" Creoles (e.g., Saramaccan) which deviate the most extensively from those of their presumably similar "lexifiers" are due to innovations by Creole children, who brought order and syntax to the allegedly "structureless" or "syntaxless" pidgin spoken by their parents. Assuming that Creoles evolved from pidgin ancestors, children putatively made maximal use of an innate language blueprint (the bioprogram) to produce a grammar with the parameters set to their default options (or "preferred settings" (Bickerton 1984: 178)). All deviations from these settings are accordingly due either to decreolization or to subsequent influence from substrate languages. During this period, substratists, especially those working on Atlantic Creoles, have mostly disputed Bickerton's claims. They argue that similarities between these vernaculars and particularly the Kwa and Mande languages could not simply be accounted for by coincidence with the work of the bioprogram, as suggested by, e.g., Givon (1979). A lot of energy has been invested into defending Sylvain (1936) and Turner (1949), drawing support from either Turner himself or Herskovits (1941) and adducing further data on African languages from Christaller (1933) and work from subsequent scholars. With substantial funding from the Canadian National Research Council, Claire Lefebvre led a team of mostly graduate students to gather information about Fongbe for systematic comparisons with structures of Haitian Creole, arguing that structures of this particular vernacular have been virtually all selected from this group of African languages spoken in Benin. Therefore, according to her, Haitian Creole differs from Fongbe especially by the fact that the Fongbe lexicon has been replaced by that of French, the "lexifier." The findings of this research program, associated with her relexification hypothesis, have been synthesized in Lefebvre (1998,2004). This is an extreme version of the substrate hypothesis extensively disputed by Chaudenson (1992, 2001,2003) and DeGraff (2002), as well as questioned on the basis of better-informed analyses of Gungbe by Aboh (2006), although it has surprisingly aroused interest in such fieldoriented scholars as Migge (2003) and Siegel (2003).3 The substrate hypothesis is actually as old as creolistics itself. Schuchardt may very well be considered the forerunner of what is now called the complementary hypothesis. According to this position, Creoles owe their structural features variably to both substrate influence

Transfer and the Substrate Hypothesis in Creolistics and the bioprogram, as well as to superstrate influence (Mufwene 1986a; Mufwene & Oilman 1987; Baker 1993a; Baker & Corne 1986; Corne 1999, and Hancock 1986,1993).4 Schuchardt's discussions of Melanesian pidgins (1883, 1889) and Saramaccan (1914) contain numerous references to substrate "systems," in contrast with his study of lingua franca, which suggests universals of SLA as the mechanisms which produced its "pidgin features." (See Gilbert 1980 for a brief discussion of the development of Schuchardt's position on Creole "genesis") A contemporary of Schuchardt, Lucien Adam, is credited by Holm (1988:28, citing Kihm 1984) with the first exclusively substrate hypothesis, at least based on the following quotation: ... the Guinea Negroes, transported to those [Caribbean] colonies, took words from French but retained as far as possible the phonology and grammar of their mother tongues ... Such a formation is surely a hybrid ... The grammar is no other than the grammar of the languages of Guinea.5 (Adam 1883: 4-7)

Adam thus predates by 53 years Sylvain's (1936) celebrated conclusion that Haitian Creole is Ewe couched with a French vocabulary.6 It is not clear where Dirk Christiaan Hesseling, a Germanist who published several papers comparing Afrikaans with other Creoles between 1897 and 1933, fits relative to the substrate hypothesis. He denies (any significant) influence of African languages on Afrikaans (1923) and yet claims in 1933 that: The slaves were less in a position than their masters to conceive of a different manner of communication other than their own; thus they had to content themselves with imitation. So they learn the surface structure of the European languages, although they make them suitable for their own manner of thinking. Something therefore emerges which satisfies both parties. The masters hear their own words, however truncated or misshapen, while the slaves employ the foreign material in a way which is not in complete conflict with their inherited manner of expressing themselves, (trans. 1979 by Markey & Roberge, p. 69)

Hesseling's attribution of the misshaping of European languages to the slaves' "inherited manner of expressing themselves" suggests African substrate influence. In fact, his discourse may be found similar to that of the proponents of the relexification hypothesis. However, nothing clearly precludes relating it to Bloomfield's (1933) "baby talk," which prevailed in the early twentieth century. Ultimately, these interpretations boil down to suggesting that the African slaves were incapable of learning linguistic structures beyond those of their native languages or what their "inferior," "child-like" minds could enable them to "acquire." The substrate hypothesis dies somewhat after Hesseling and Sylvain, being replaced especially in the USA by what may aptly be characterized as the dialectologist position (identified as the superstrate hypothesis in creolistics). It is resurrected by Turner's Africanisms in the Gullah Dialect (1949, much influenced by Herskovits 1935-41), which claims that Gullah owes much of its vocabulary, phonology, and morphosyntax to African languages. Although Turner's book has been criticized especially for invoking mostly proper names to prove

137

138

Language Evolution African linguistic influence (see, e.g., Hair 1965, Mufwene 1985, Swadesh 1951), it is easily the first large-scale endeavor, after Sylvain (1936), to demonstrate that Africans have contributed to the shapes of Creoles of the New World, at a time when dialectologists such as Krapp (1924) were arguing that the Africans had learned the European colonial language so well that there was little trace of their heritage languages left in their speech.What distinguished it from that of White Americans were simply archaisms, which the latter's speech had allegedly evolved from. However, Turner's hypothesis hardly replaced the dialectologist position; it only drove proponents of the latter to recognize the most obvious of African influence such as in the vocabulary. Aside from the fact that most critics have discounted the significance of proper names (about 87% of the lexical evidence Turner adduced), it was observed that most of the other words are restricted to ritual language (e.g., songs) and are therefore not part of the vernacular spoken on a daily basis (Cassidy 1983). Thus, from a dialectologist perspective, the contribution of African languages to Gullah remains minimal, although this conclusion disregards the role of congruence. Nonetheless, Turner has inspired considerable research on African substrate influence in New World's Creoles, including Alleyne (1971, 1979, 1980, 1986, 1993), Allsopp (1977), Boretzky (1983, 1993), DeBose & Faraclas (1993), Faraclas (1987, 1988a), Oilman (1986), Holm (1980a, 1980b, 1986, 1988, 1993), Holm & Oyedeji (1984), Koopman (1986), Kouwenberg (1994), Lefebvre (1986, 1993, 1998, 2004), Manessy (1985a, 1985b, 1988), Maurer (1987),Migge (2003), Robertson (1993),Singler (1984,1993), Smith etal (1987),Holloway & Vass (1993), Wade-Lewis (1988) and Warner-Lewis (1996). Some of these have shown little improvement over Turner's methodology (see Section 8.2.2.2), a weakness which has contributed not only to the stagnation of the substrate hypothesis (at least until recently) but also to its being slow in countering convincingly the exaggerated significance of Bickerton's language bioprogram hypothesis. Since the 1970s, many genetic creolists have preferred to acknowledge substrate influence in different ways. For instance, assuming that Creoles evolved from pidgin ancestors, Muhlhausler (1980, 1981, 1983, 1986a, 1986b, 1987) claims that substrate languages may influence a Creole after, rather than during, its formative stage, which is controlled by the bioprogram. On the other hand, supposing that the expanded pidgins of Melanesia can now be considered as Creoles, Bickerton (1984b) actually makes a one-time concession to substrate influence as an explanation for many structural features which they do not share with Atlantic Creoles and which his LBH could not either treat as unmarked or account for in terms of "decreolization" qua debasilectalization, e.g., the DUAL/PLURAL and the INCLUSIVE/ EXCLUSIVE distinctions. He posits a "pidginization index" which would exceptionally select some substrate features into a Creole under conditions where the Creole would be spoken for an extended period of time and the superior proportion of nonnative speakers of the lexifier would exceed the putative 80 per cent ratio. In the same vein, Baker (1993a), Baker & Corne (1986), Hancock (1986, 1993), Koopman (1986), Mufwene (1984, 1986a, 1993a),

Transfer and the Substrate Hypothesis in Creolistics Sankoff (1979,1984,1993),Sankoff & Brown (1976),Singler (1988), and Thomason (1980) make allowance for substrate influence favored by such ecological factors as the typological kinship of (most of) the substrate languages. Oilman (1986) invokes areal features, in reference to features that spread across language families in sub-Saharan Africa and need not be attributed to specific languages, for instance the phrase-final position of the determiner la/an in French Creoles and serial verb constructions, although, as explained below a congruence-based explanation may be more convincing. Over the last few years the substrate hypothesis has certainly improved and found a more sympathetic audience. Nonetheless, it still has to defend itself against hard-core dialectologists and universalists, who have used the shortcomings of earlier substratist accounts and those of some current extreme positions (especially the relexification hypothesis) to their advantage. I attempt to account for this interpretation of facts below.

8.2.2 Toward an understanding of the substrate hypothesis 8.2.2.1 Substratists maybe classed into several subgroups, depending on the criteria which are considered relevant. First, those like Adam (1883) and Lefebvre (1986ff) who attribute features of Creoles' grammars exclusively or primarily to substrate influence may be distinguished from those who allow more room for other kinds of influence. The main problem with the exclusive or primary substratist explanation is that it does not account for those Creoles' features which are not exactly matched by the relevant African languages. For instance, the Haitian Creole question phrase ki mun "who" (lit., "what person") is not an exact match to the Fongbe me(-te) (lit., "person what") in which the interrogative constituent follows the head noun. From an evolutionary perspective, the fact that the constituent order is not the same calls for an explanation. A similar question arises about the order of adjectives at least in phrases such as ti mun "child" (lit., "small person"), in contrast with Fonge's yokpo vi (lit., "person small"). In both of these examples, the word order is French-like, which lends more support to Aboh's (2006) argument for "hybridization" in the development of Creoles. This is just a variant of the Complementary Hypothesis articulated in terms of competition and selection (also adopted by Aboh), according to which structural variants that are (partially) congruent are favored over those that are not, where the learner has a choice. Aboh's work provides more examples from Gungbe, a language genetically and typologically related to Fongbe, that complement the above. They suggest that the question must arise more often to whoever addresses the comparison of Haitian Creole, Fongbe, and nonstandard French structures in more detail. DeGraff (2002) is particularly informative on the subject matter, notwithstanding Chaudenson's (2001) question of whether there is any evidence from SLA that supports the relexification hypothesis. Siegel (2006) replies unequivocally that there is none. As a matter of fact, research by, for instance, Perdue (1995)

139

140

Language Evolution highlights inter-individual variation among L2 learners of the same ethnolinguistic background. This suggests that the approximations of colonial French by the African Bozal slaves on Haitian plantations must also have varied inter-individually even among those who had spoken Fongbe and other Gbe languages. The exclusive substratist hypothesis is weakest in cases such as Adam (1883), whose statement refers to mother tongues which may have been typologically different. It is worthier of attention in the cases which refer either to a specific African language (Lefebvre 1986) or to a group of genetically and typologically related African languages (Lefebvre 1993ff). In the latter case, allowing for "founder effect/principle," what has been missing the most from the studies is not only evidence from SLA that supports the relexification hypothesis (pace Sprouse 2006) but also sociohistorical documentation that backs it. Singler's (1993, 1996) findings are not particularly consistent with Lefebvre's claims. One may assume that Lefebvre's invocation of "leveling" to account for the emergence of a putatively uniform Haitian Creole out of the expected relexifications of diverse substrate languages addresses the question of why the Fongbe population has been privileged in a setting that was highly multilingual and included critical masses of speakers of Bantu and other Kwa languages. If this means that Gbe speakers were the dominant population during the most critical period in the formation of Haitian Creole, then one must also explain whether, even in assuming "Target Shift" (Lumsden 1999, Mather 2006, and Kouwenberg 2006), the later Bozal slaves did not also learn the local colonial vernacular among the Creole slaves by relexification and why they wouldn't have produced structures divergent from those of Gbe. Likewise, Lefebvre's (1986) argument, based on Bastide (1967) and Herkovits (1975), that her linguistic findings are correlated by similarities between Haitian and Gbe cultures does not address the question of what happened to influence from other African cultures, especially where they are not congruent. After all, typologically speaking, sub-Saharan Africa is no more culturally homogeneous than it is linguistically. Even if it were, does the evidence entail, contrary to Sapir's (1921) words of caution, that cultural substrate influence is a concomitant of linguistic substrate influence? What arguments can be used against the claims by Ans (1987) and Chaudenson (1992,2001) that in some domains, such as religion, African cultural influence was the strongest toward the end of slavery and afterwards? This is the period during which more and more slaves were being imported from the same region, in response to preferences expressed by some plantation owners, which would have favored larger aggregates of slaves speaking the same or related languages. There's also the fact that the slaves who arrived soon before Emancipation would have been more likely to retain their African customs than those who arrived earlier, especially if they were only slowly integrated by the Creole (former) slaves. These conjectures are indeed supported by Bilby's (2003) position on the development of Kumina and similar religious syncretisms in Jamaica. The ecological factor of greater revalorization of Africa, and, in the case of the Anglophone Caribbean, the social isolation of

Transfer and the Substrate Hypothesis in Creolistics the African contract laborers from the Creole former slaves, would have favored more retentions of African traditions. That the social ecology bore on the selection of particular African elements is made more evident by the fact that the drum, an instrument so fundamental in the music of most sub-Saharan Africa, made its entrance quite late in musics of people of African descent in the New World and Indian Ocean. It was forbidden and alternative devices such as rhythmic clapping and stamping were enhanced to regulate the beat.

8.2.2.2 It may also be convenient to distinguish those who invoke substrate influence from single languages or groups of typologically related languages (e.g., Alleyne 1993; DeBose & Faraclas 1993; Faraclas 1987, 1988a; and Lefebvre 1986ff) from those who refer to, for instance, all the languages of the coast of West and Central Africa. One of the problems on which Bickerton (198Iff) has capitalized in arguing against substrate influence, particularly as presented in such major works as Alleyne (1980), Holm (1988), Jourdain (1956), and Turner (1949), is random reference to various African languages, down-playing the typological differences between them. As is evident from Corne (1999), there are simply no general African linguistic features which would not be shared, at least partially, by the European "lexifiers" of Atlantic Creoles. Works such as Faraclas (1987, 1988a), Gilman (1986), Manessy (1985a, 1985b, 1988), Sankoff (1984,1993), and Sankoff & Brown (1976), which associate substrate influence with areal features, are a small antidote to the deep-rooted trend especially among Atlantic Creole substratists to invoke African linguistic features randomly from languages sometimes belonging in different typological groups. This dominant research style has seriously undermined the substrate hypothesis, despite tendencies such as in Migge (2003) to adduce historical demographic evidence to justify the privileging of particular languages over others. One often comes across references to languages of West Africa or of the Niger-Congo. These make the substrate hypothesis neither more explicit nor less confusing than the language classifications are at times. There is variation even within the Niger-Congo family regarding morphosyntax. Likewise, attempting to account for the origins of some Creoles' morphosyntactic features by invoking features sometimes from the Mande family, at other times from the Kwa group (disregarding internal variation within this group, too), and at yet other times from the Bantu system is disappointing when no explanation is given for why those particular features would have been favored. This is not an argument against what has been bashed as the "Cafeteria Principle." As explained in Mufwene (1996a, 200la), it is just an objection to the fact that no attempt has been made in most of the literature to submit any principle regulating this seemingly haphazard selection of substrate features into the Creoles. It is not that such a situation is impossible, but rather that principles accounting for the selection of the features have not been proposed. Thus, with few exceptions to be discussed below, much of the research on substrate influence has been in the

141

142

Language Evolution direction of sensationalism in noting more similarities between various African languages and New World's Creoles. Little research has been in the direction of making the substrate hypothesis qualitatively more plausible. In a different vein, since Atlantic Creoles' morphosyntactic features are similar mostly with western Kwa languages, one would have expected some research about whether or not some founder-effect phenomenon was involved in the selection of substrate influence. This possibility would have been largely supported by the fact that the transatlantic slave trade started on the coast of West Africa and in time expanded inland and southward to Central Africa. However, there are to my knowledge no significant studies in this direction. Considerations such as cited above from Lefebvre (1986) seem to be accidental. In addition, most such works have addressed the relevant questions rather superficially or partially. The true exceptions to this observation are discussed below. The question of sociohistorical justification is equally significant when single languages are involved. Aside from Lefebvre (1986), the only other justifications are provided in Smith et al. (1987) and Kouwenberg (1994): the eastern Ijos are said to have worked with the Dutch in the slave trade and putatively formed the majority in Berbice, Guyana. The only problem is that Smith et al. and Kouwenberg have no explanation for those features of Berbice Dutch that do not follow the Ijo "system." For instance, it has SVO surface constituent order while Ijo is claimed to have SOV order, and it has no gender, whereas Ijo does. As noted above, similar questions arise from Lefebvre's (1986) claim that Haitian Creole is a relexification of Fongbe with French lexical items. Accounting for constituent order differences between Haitian and Fongbe within the noun phrase is as important as accounting for major constituent order differences between Berbice Dutch and Eastern Ijo. These structural differences question genetic accounts of these Creoles that invoke influence from single languages. In the case of Haitian, it may simply have been a matter of not expressing things right. To be sure, Lefebvre (1986) also acknowledges, with reference to Koopman (1986), that Fongbe is typologically akin to the neighboring substrate languages involved in the contact with French; Lefebvre (1993ff) speaks of influence from the whole Gbe group, treating Fongbe explicitly as a representative. Nonetheless, she has not addressed the question of what proportion of the relevant structural features were favored by the fact that they are shared by the grammars of other African languages brought to the plantations of Haiti, aside from that of the extent to which, as objected by Chaudenson (1992, 2001, 2003) and noted by Corne (1999), most or many of these features are attested, partially at least, in colonial nonstandard French varieties. Thus, the adult "creators" of Haitian Creole would have selected, from their French target, structural variants that resembled those of the languages they had spoken in Africa. The rest would be the kind of account provided by DeGraff (1999a, 1999b, 2002). The state of the art has, however, not been as hopeless as may have been suggested by the above comments. For instance, Ferraz (1979, 1984) clearly indicates that the large

Transfer and the Substrate Hypothesis in Creolistics proportion of Bantu slaves on the Gulf of Guinea islands, particularly Sao Tome and Principe, accounts for the significant number of Bantu grammatical features such as the following in their Creoles: the phrase-final position of the determiner, the post-nominal position of the modifier, usage of a BE WITH construction for "have," and sentence-final negation. These features distinguish them not only from other Atlantic Creoles but also from the Portuguese Creoles of Upper Guinea, where the presence of Bantu languages has been negligible. Maurer (1987), too, gives more plausibility to the substrate hypothesis, at least as it applies to Palenquero in Colombia. To support his claim that western Bantu languages have influenced the grammar of this Creole, he indicates clearly that Bantu people constituted the majority of the first slaves brought to San Basilio. The following features attributed to Bantu influence may thus be attributed to the founder effect, especially since, compared to other Atlantic Creoles, they appear to be uncommon: sentence-final negation, suffixation of the ANTERIOR tense marker ba, agglutination of the pronominal direct object between the verb and the ANTERIOR tense marker, and use of a separate marker se for the HABITUAL aspect. From the ecological perspective developed in this book, the relative ethnolinguistic homogeneity of the slave populations during the critical formative period of Palenquero conspired to transfer into the "stable" feature pool quite a few Bantu elements that would be selected by subsequent learners, including children. In the absence of (partial) congruence with Portuguese features, such xenolectal elements must have been favored by especially the sheer importance of their statistical frequency among the extant speakers of the colonial vernaculars. In a setting that did not develop into a remarkable plantation colony or a slave depot, the population could not grow in a way likely to offset the founder effect, unlike in territories such as Jamaica, Guyana, coastal South Carolina, Cape Verde, and Curasao, where rapid population replacement was a critical factor in the development of Creoles. In San Basilio, the small increments of newcomers constantly remained in the minority and just learned with minor modifications the language of the speakers who preceded them. As with immigrants to North America and Europe, the adults would die with most of their divergent features and the children would acquire the local ones natively. More support for the substrate hypothesis comes from Singler (1993, 1996), especially from his correlation of the socioeconomic history of tobacco, cotton, rice, and sugarcane plantations in the New World with particular outcomes of language restructuring. Singler associates the intensification of the slave trade with the emergence of large plantations and their lucrative role in the colonial economy. Contrary to the traditional belief that plantation owners did not want people from the same ethnic groups, a predilection for slaves from some specific parts of Africa putatively did ensue from this system. Corroborating Singler (1988), he also highlights the typological kinship between especially the Kwa and Mande groups affected by the slave trade and thus lays the long-awaited groundwork for the plausibility of variable African substrate influence in several parts of the New World.

143

144

Language Evolution Singler's work is consistent with the spirit of Baker (1993a), Baker & Corne (1986), Goodman (1985), Mufwene (1986a), and Corne (1999), viz., the complementary hypothesis. As explained in Mufwene (200la) and in Chapter 3, this position should not be misconstrued as a mere compromise solution to the controversy about Creole genesis. It capitalizes on the osmotic nature of linguistic "systems" and the fact that every idiolect is constructed selectively from various competing models to which the learner has been exposed. It is also inadequate to characterize influences on Creole structure as splitting three ways, between the bioprogram, the substrate, and the superstrate languages. The bioprogram should simply be interpreted as a body of regulatory constraints on the restructuring that the "lexifier" undergoes. An impact of Singler (1988) is to have made more obvious the principle of least effort in the formation of a new language: why would the speakers of the substrate languages have followed the putative bioprogram or structural variants of the "lexifier" that were strange to them when they could adopt other variants that were similar to those of the substrate languages they had "acquired" earlier? In a few cases, substrate influence simply amounted to modifications that the substrate languages imposed onto the "systems" of the "lexifiers," especially when they did not impede communication within the community of speakers. A good example from English Creoles is usage of done without the auxiliary be or have and followed by the base form of the main verb as a marker of PERFECT, thus NP done tell "have told," instead of NP be/have done told (with the auxiliary usually contracted). The above position is amply corroborated by Keesing (1988). Citing data from the Melanesian pidgins, he shows that the majority of Oceanic languages, especially those of the central Pacific, belong to the same morphosyntactic type. He also shows that the features which distinguish Melanesian pidgins as a group from, e.g., Atlantic Creoles, are precisely some of those which characterize the Oceanic substrate languages as members of the same family, for instance, the "predicate marker," the classifier, the DUAL/PLURAL distinction, the INCLUSIVE/ EXCLUSIVE distinction, and the TRANSITIVE-CAUSATIVE verbal derivation. One can add to this list Sankoff & Brown's (1976) account of bracketing of relative clauses with ia and Sankoffs (1993) discussion of focus constructions in the same language varieties. Moreover, all of these features are considered "marked" by typologists, i.e., as not commonly attested among the world's languages. In his detailed sociohistorical account, Keesing also shows that the influence from the Oceanic languages predates the expansion and stabilization of these new languages, having started on board the trade ships where he claims the original pidgin was born.7 These observations not only make substrate influence quite plausible, with its significance varying from one contact ecology to another, but also question the position discussed in Section 8.2.2.4 below that substrate influence occurs (as of necessity) after the new linguistic "system" has crystallized. 8.2.23 Works such as Daelman (1972), Huttar (1975, 1985), and Wade-Lewis (1988), which are mostly concerned with the African etymologies and models of some terms and caiques,

Transfer and the Substrate Hypothesis in Creolistics respectively, should also be distinguished from those others, such as discussed above, which attempt to determine major influences on the formation of individual Creoles, particularly their grammars. Daelman (1972) is particularly interesting because his main concern is establishing certain and possible Kikongo etymologies, as opposed to improbable ones, in Saramaccan vocabulary. The certain and possible etymologies represent a small fraction of the vocabulary examined. On the other hand, Wade-Lewis relies heavily on the notion of semantic fields but cares little about the trajectories of the specific borrowings or "apports" into English, e.g., whether or not an African word came into a New World's English Creole or into AAVE indirectly through English itself or directly during the contact between English and the African languages on the plantations. The wide range of African languages invoked seemingly by chance also provides more ammunition to those who bash the "Cafeteria Principle," especially since no sociohistorical facts are invoked to justify their claims. In any case, this etymological aspect of the substrate hypothesis brings to mind germane questions. For instance, since much of the support for grammatical substrate influence in works such as Turner (1949) has been adduced from lexical and cultural influence, do these kinds of influences entail grammatical influence? Could the lexical-cultural and the grammatical influences come from different and conflicting sources, as noted by, e.g., Huttar (1985) about the role of Kikongo lexical items in Ndjuka? Could lexical elements have been introduced into a Creole after its structures had already developed? In the case of Gullah, one must note here the significance of a phonology that is essentially English and has influenced the forms of the African "basket names" discussed by Turner (1949) (Mufwene 1986d). Could cultural influence from some areas of Africa be associated with later migrations, after slaves from (a) different part(s) of Africa had already influenced the "system" of the emergent vernacular? Wouldn't this be the way to address the question of whether the Yoruba contract laborers in Trinidad could have influenced the development of Trinidadian Creole the same way as any demographically dominant ethnolinguistic group before Emancipation? Has substrate influence occurred in all Creoles in the same fashion? Is there a particular stage in the development of a Creole when substrate influence may (not) occur? Part of the evidence for linguistic substrate influence has also been adduced from the survival of (words and phrases from) African languages in some religious rituals of the Caribbean and South America. The question is whether such restricted usage of some languages (e.g., Kikongo in Jamaica and Yoruba in Trinidad and Brazil) is evidence for the continued usage of African languages as vernaculars on plantations. Or, as argued by Chaudenson (1992,2001) and Singler (1993,1996), is this evidence of the changing ecological conditions toward the end of slave period which made it possible for relatively critical masses of slaves from the same area in Africa to find themselves on the same plantations and speak their heritage languages among them? As Warner-Lewis (1996) has made so evident, the fact that the post-Emancipation contract laborers generally came from the same ethnolinguistic area and would be integrated by the former, Creole slaves only gradually

145

146

Language Evolution made it possible for them to maintain their languages for a while, before they too shifted to the local European vernacular (Chapters 10 and 11). Genetic creolistics must definitely address such ecological questions to explain why the evidence for African substrate influence on the Creoles of the Caribbean-American area and of the Indian Ocean is not as extensive as in the expanded pidgins of Melanesia and West Africa.

8.2.2.4 It is also informative to distinguish between those such as Baker & Corne (1986), Bickerton (1984b), and Mtilhausler (1980ff), who allow for substrate influence after the formation of the Creole and those, such as discussed in Section 8.2.2.2 (most substratists), who claim that it applied during the formative stages. Although Miilhausler's hypothesis on Tok Pisin is disputed by Keesing (1988), the question actually arises of whether substrate influence counts as formative only in relation to the initial formative stages of Creoles and the rest counts as borrowings. Is there compelling justification for restricting "creole genesis" (interpreted as the development of Creoles) to one single generation, as claimed by Bickerton (1981, 1984b, 1999), Thomason (2001), Thomason & Kaufman (1988), and others who still believe that there are Creoles that have evolved abruptly?

8.2.2.5 The term substrate hypothesis thus applies to a variety of overlapping, but non-identical, positions on the formation of Creoles. Much of its limited and slow success is due to its methodology, the style of the arguments supporting it, and, perhaps inaccurate interpretations of the structures of the relevant African and European vernaculars involved in the contact. The relevance of nonstandard varieties of the "lexifiers," as pointed out by Chaudenson (1973ff) and Mufwene (1996ff) cannot be overestimated. In the same vein, Hagege (1985: 39) remarks that features such as economy/simplicity, analyticity, and motivation, which have been associated with Creoles, are also observable in the spoken (i.e., colloquial and nonstandard) varieties of the European "lexifiers." This observation complements Chaudenson's (1985) comment that impoverishment of the morphology in French Creoles is the ultimate outcome of a "morphological erosion" process that was already underway in nonstandard French varieties (les fran$ais populaires) when colonization started. Much of the resistance to the substrate hypothesis may also be attributed to a misconception of the role that the Bickertonian language bioprogram can play in the restructuring process that would result in Creoles. As noted by Thomason (1983) and Hagege (1985), most of the creole features attributed to the bioprogram are actually features that are shared by many of the languages in contact (substrate and superstrate); demonstrating the role of the bioprogram would thus be tantamount to establishing conclusively that structural congruence between these languages would not have favored the selection of those particular features, by the principle of least effort, as argued by Mufwene (2001a). Chaudenson (2001,

Transfer and the Substrate Hypothesis in Creolistics 2003) explains that some structures of the nonstandard "lexifiers" targeted by the slaves are typologically different from those of their standard counterparts that have too often been compared, unjustifiably, with those of Creoles. As argued by Corne (1999) for French Creoles, features shared by the superstrate and several substrate languages seem generally to have been favored in Creoles' structures. It is this kind of congruence that has largely undermined the appeal of the relexification hypothesis, notwithstanding the shortcomings discussed above (Mufwene, to appear). In a slightly different vein, Mufwene (1986b) also notes that there are no formal features, nor any combinations thereof, which distinguish Creoles from other languages. Otherwise, a language such as Mandarin Chinese would be mistaken as a prototypic Creole; it exhibits most of the putative "creole features" with a systemic regularity (and communal uniformity) unmatched by the Creoles themselves. Central distinctions such as the following are equally significant in Mandarin: REALIS/IRREALIS, INDIVIDUATED/NONINDIVIDUATED (i.e., Stewart's 1974 inadequate SPECIFIC/NONSPECIFIC), STATIVE/NONSTATIVE (with regard to verb delimitation), the absence of the copula before a predicative adjective or preposition, and serialization of predicates.8 The question of how much structural materials the "lexifier" and the substrate languages have contributed to the emergent creole can simply be recast in terms of ecologically driven competition and selection between features of "lexifier" and substrate languages. Much of what has been associated with the bioprogram can be reinterpreted as some form of Chomskyan UG-based constraints on the emergent structures of Creoles. It is not clear that one must even invoke some "universals of second language acquisition" (Thomason 1980ff; Alleyne 197 Iff), because, as pointed out by DeGraff (1999a, 1999b), both children and adults were continually involved in the gradual restructuring process (involving competition and selection) that produced Creoles. (See also Chapter 10.) The bioprogram, which must be accessible to, and operational in, speakers and language learners of any age who have been normally socialized with a language, must also specify which combination of formal features and rules produces a language and which would produce a non-language.9 Likewise, in those cases where some features of a language are being dispensed with, the bioprogram must make the speaker/learner able to tell which particular kinds of components and oppositions are central and critical to the architecture of a language and must be maintained (through "acquisition" by new learners) even during the kinds of extensive cumulative restructuring that produced Creoles (Mufwene 1999a). The linguist can thus determine what particular linguistic elements (e.g., lexical categories) are, from an evolutionary perspective, so deeply entrenched and generative (Wimsatt 1999) that no language can do without them. They are very likely the components on which the rest of the complexities that vary cross-linguistically are scaffolded. An interesting phenomenon in this particular case that should not be overlooked is also how easily materials from diverse languages (what Chaudenson 1979f calls materiaux de construction) can be recombined and

147

148

Language Evolution integrated into a new linguistic "system." Thus, as observed above, we should focus more on principle in the Cafeteria Principle, instead of rejecting it offhand and uncritically. Osmosis, which is the general phenomenon that substrate influence must be associated with, seems so significant in the evolution of particular languages that it is worthwhile investigating how it is constrained. Provided the issues raised in Sections 8.2.2.1-8.2.2.4 are addressed, any version of the substrate hypothesis other than the relexification hypothesis seems plausible under the right ecological conditions. For instance, the special conditions of the development of Berbice Dutch, with the majority slave population originating in the Eastern Ijo-speaking area of Nigeria, accounts for presence of structural peculiarities also originating in or just favored by their common substrate language, viz., postpositions, the sentence-final negator ka(ns), the focus marker sa/so> aspectual suffixes such as the PERFECTIVE t£, a rich morphological system in the pronominal system, and a plural suffix that is not etymologically related to the third person plural pronoun (Kouwenberg 1994). Along with Keesing (1988), Kouwenberg thus shows that substrate influence (even in the sense of "apport") is the strongest where the substrate population was relatively homogeneous ethnolinguistically (Mufwene 1986a, Singler 1988) and the learners likely to reinforce each other's deviations from the TL during the critical period in the development of a Creole. Where this was not the case and a Creole exhibits structural features quite divergent from those of its nonstandard "lexifier," typically a koine that was also emerging concurrently in the colony, one must not only consider the possibility of areal substrate features (Gilman 1986) but also look into the patterns of population growth and how competition and selection appear to have proceeded, at different times, within the substrate population. Unfortunately studies on the development of Creoles' structures have focused more on a handful of features that distinguish them from the relevant European languages than on possibly the majority of others that these emergent vernaculars share with the latter. Even studies such as Keesing (1988) and Kouwenberg (1994) have not proved that Creoles' features point only to divergence. Chaudenson (1992,2001,2003) and Corne (1999) show that the evidence for retention (with modification) of elements and structures of the "lexifiers" is still very strong. For instance, it is significant that the TENSE-ASPECT markers of French Creoles have all been selected from morphemes with similar TENSE-ASPECT functions in nonstandard French, that relativizers are etymologically related to the French relative pronouns quilque, that (like in French and unlike in English Creoles) French Creoles do not strand prepositions in relative clauses and in questions, and that they exhibit no possessive constructions in which the POSSESSOR noun precedes the POSSESSED one, and they have no nominal compounds on the model of man-hunt in which the OBJECT precedes the ACTION syntactically. Many of the relevant substrate languages do not exhibit these syntactic peculiarities. It is also hard to assume that all Creoles emerged according to the same fixed restructuring "recipe." A more adequate alternative is the ecological approach submitted in Mufwene

Transfer and the Substrate Hypothesis in Creolistics (1996a, 200la, 2005a), according to which particular features are selected into the new vernacular from a feature pool to which the languages in contact have contributed variably from one colony to another and from one phase of colonization to another. Once one factors in the nature of the "lexifier" (typically nonstandard), the founder principle, patterns of population growth, and the timing of changes in the population structure, one can see why structures of the "lexifier" would not have been disfavored as much as has been assumed. As argued in Chapter 3, Creoles must have developed by a normal gradual process of divergence in which substrate influence was facilitated selectively under particular ecological conditions. Accordingly, different kinds of substrate influence would be effective in different situations and call for a different kind of explanation based on the kinds of languages in contact and the kind of interactions among speakers. One of the things most successfully shown by Thomason & Kaufman (1988) is that the development of Creoles differs from that of other language contact phenomena (such as the "indigenization" of European languages in the Third World) more in degree than in kind. To account for this variation in outcomes, one must of course factor in differences in the nature of the TL (nonstandard vs. scholastic varieties) and in the medium of "transmission" (natural interactions vs. guided teaching in the classroom), as explained in Mufwene (200la, 2005a, and Chapter 3). The better we understand these other outcomes of language contact, the harder it becomes to make a sharp distinction between them and some instances of the development of Creoles, and the more compelling it is to adopt a uniformitarian approach, according to which all language "transmission" entails some restructuring of the "system" on the part of the learner, with allowance made for a variable dosage of xenolectal elements and/or influence. This is as true of language varieties which, for non-linguistic reasons, have been disfranchised as Creoles, as of others.

8.3

Transfer and the substrate hypothesis

This part of the Chapter identifies ways in which research on transfer in SLA and that on the substrate hypothesis in creolistics may benefit each other. The focus here is on what has not received enough attention in both fields. Generally, transfer theory and the substrate hypothesis seem to share several assumptions and differ minimally. I focus below only on two points which are related to markedness and on a third one regarding misconceptions of the similarities between the development of Creoles and SLA. The headings of the first two sections state my concerns clearly.

8.3.1 What creolists could learn from the literature on SLA Invocations of markedness in the SLA literature (e.g., by Meisel 1983a, 1983b; McLaughlin 1987; and White 1987) have generally been associated with two assumptions: (1) typologically

149

1 50

Language Evolution unmarked features of the TL are the easiest for the L2 learner to acquire, and (2) unmarked features or structures of languages previously known to the learner are likely to be transferred to the learner's interlanguage, especially if the targeted forms and/or structures are "marked." The basic idea behind these assumptions, which is about the only thing that I share with those who hold them, is that transfer applies in a selective way. My position on markedness in the development of Creoles is clearly articulated in Mufwene (1991b, 2001a) and need not be repeated here. Basically, accounts which claim that a particular parametric or typological option has been selected because it is unmarked explain nothing if they do not identify the particular factor(s) that make(s) the option unmarked. The determination of the option as (un)marked is even more uninformative if it is based on typological considerations of the statistical attestations (in type rather than in token) of the relevant feature among the world's languages rather than relative to the specific ecology and feature pool in which the selection was made. That is, it is pointless to claim that a speaker or a particular population has selected a particular variant or typological option because it appears to be unmarked relative to what is attested in the world's languages if the speaker or the population is not aware of the existence of some of the variants or options, thus if the features are not part of the feature pool from which their particular selection has been made (see below). As in my previous work, the term unmarked applies in this Chapter to the fact that in a particular setting of language "acquisition" and/or change (including the development of Creoles) an option has been favored by a variety of structural and other, external ecological factors, for instance, SIMPLICITY, GENERALITY, FREQUENCY, SEMANTIC TRANSPARENCY, SALIENCE, and CONGRUENCE between features of the "lexifier" and of the substrate languages. Although the notion of MARKEDNESS is frequently invoked in genetic creolistics, it has seldom been used by substratists who advocate exclusive or primary substrate influence. As shown above, numerous substratists are too easily content with simply noting that a particular Creole feature is also attested in some substrate language which was involved in the contact situation. Although the feature may have been in competition with other features of substrate languages, little effort has been made to explain what particular factors may have favored the selected alternative. Like contrastive analysts in the early days of research in SLA, many substratists have given the impression (perhaps false) of assuming that any feature of substrate languages may find its way into a Creole. Reality shows that many such features have been selected out, as evidenced by the fact that most Creoles are neither tonal nor agglutinating, despite the fact that most substrate languages are tonal and a significant proportion of them are agglutinating (particularly the West Atlantic and Bantu languages, and to some extent some eastern Kwa languages). Perhaps some of the best defense to date for the substrate hypothesis is to be found in Ferraz (1979), Maurer (1987), Keesing (1988), Kouwenberg (1994), and Erhardt (1993) because they describe varieties that emerged in ecologies where there was little typological diversity and competition among the substrate features, therefore more reinforcement by the learners of each other's deviations from the TL during the language "acquisition" process.

Transfer and the Substrate Hypothesis in Creolistics The above shortcoming of the substrate hypothesis is the result of considering genetic influence not only in (quasi-)exclusive terms (e.g., predominant or only sub- or superstrate influence) but also as a three-way competition. Enriching research on the substrate hypothesis with the ecology-specific notion of MARKEDNESS submitted in Mufwene (1991b, 2001a) can actually cast things in a different light. For instance, it must lead to a more charitable interpretation of Thomason's (1983) and Hagege's (1985) hypothesis, according to which the features found in most Creoles and inaccurately attributed to the putative bioprogram are actually features shared by many of the languages in contact (see also Chaudenson 1992, 2001,2003). This is an ecological situation that favored congruence and therefore made features such as the periphrastic expression of TENSE and ASPECT, inflection-poor uses of verbs and nouns, and invariant relativizers unmarked. Shared features became preferred options, which could be learned with the least effort. As shown in Mufwene (1989b), the features need not be shared by most of the languages; they must simply appear preferable or safer to other, statistically less common alternatives in the contact situation. Such a situation could have easily emerged also when the earliest nonnative speakers of the colonial vernaculars had spoken languages whose relevant structures showed congruence with the targeted structures and the incremental growth of the population hardly created a situation where the other options gained a critical mass. The presence of children, Creole and African-born, who could learn the dominant options among the current adult speakers, just favored the gradual selection of the extant speakers' features into the emergent Creoles, subject to the dynamics of the "invisible hand" discussed in Chapter 4. These considerations do not preempt the fact that for each individual learner, markedness is first of all determined by factors such as the TYPOLOGICAL KINSHIP of the languages in contact, STATISTICAL DOMINANCE, SALIENCE, SEMANTIC TRANSPARENCY, SIMPLICITY (regarding form or combination of features encoded in a form) of particular features, and a number of other relevant factors discussed in the literature on markedness. When it comes to the emergence of Creoles as communal varieties, one must make allowance for variation in what individual learners of the colonial vernacular found (un)marked, which accounts for continuum-related phenomena discussed in the literature on "decreolization." The above considerations explain the conditions under which the options selected by some speakers would spread within the relevant population and often displace the competition and therefore prevail as the communal selections. They also make it hard to speak of ex-nihilo innovations in the sense of Bickertons (1981, 1984b, 1989,1999) LBH, which has not made allowance for generalizations (constrained by UG) from features of any of the languages in contact, except from the "lexifier" (with the lexical learning hypothesis). As far as I can tell from my study of the development of Creoles embedded in their socioeconomic histories, learners have typically proceeded by the principle of least effort, with the innovators often extrapolating from similarities perceived between the TL and a language previously "acquired." Leaving room for reinterpretation of some forms and/or constructions (in the sense of extensions as allowed by the prototype theory espoused by Lakoff 1987), formal

151

152

Language Evolution features of Creoles seem generally to have been selected from those of the languages in contact; this is the essence of particularly Thomason (1980, 1983), though the explanation at the level of Creoles as communal varieties does not boil down to adult SLA only (DeGraff 1999a, 1999b, and Chapter 10). Where most of the substrate languages (as in Melanesia) are typologically similar, many of their common structural features were selected by the relevant Creole or expanded pidgin. These are situations similar to those where a new language variety, such as Kituba and Lingala (though I do not think they must be called Creoles), evolved out of the contact of genetically and typologically related languages. Despite Owens's (1998) observation that they may qualify more easily for "prototypical Creoles" (by McWhorter's 1998 criteria), these vernaculars remain quite Bantu in a number of structural features. They have preserved a noun class system that is little impoverished, though they have lost much of the SubjectVerb and Head-Modifier agreement "system" canonically associated with Bantu morphosyntax. They have also preserved the verbal extension (i.e., the derivational part). Although they have lost much of the inflectional system to mark PERSON and NUMBER, their TENSE-ASPECT systems have remained partly inflectional, and they have not lost the copula in any tense. Moreover, Lingala has preserved both grammatical and lexical tones, while Kituba has maintained only some lexical tones, even if they are less extensive than in Lingala (Mufwene 1989d, 1997a). Kituba's periphrastic marking of some aspects is very much a legacy of options that have been available in (Kikongo-)Kimanyanga, its "lexifier." In the case of Melanesian expanded English pidgins, although much of the already impoverished morphology of the "lexifier" was lost, new distinctions were introduced into the pidgins' morphosyntax that reflect common grammatical distinctions in the substrate languages, such as the DUAL/PLURAL and INCLUSIVE/EXCLUSIVE distinctions, as noted above. Particular morphosyntactic strategies from the substrate languages were also introduced into them that are not attested in the "lexifier," such as bracketing relative clauses with ia (< English here), usage of pela (< English fellow) as an indiscriminate classifier, and suffixing im to the predicate to mark it as transitive (Sankoff 1984,1993; Sankoff & Brown 1976; Keesing 1988). All such cases support the assumption that in the contact situation, a mixed feature pool emerges from which speakers/learners can select many options from the substrate languages if there are no particular ecological/societal pressures against them. Indeed, Melanesian pidgins evolved in relatively endogenous colonies without earlier homestead societies nor Creole populations speaking less modified approximations of the "lexifier," thus in ecologies that made the options selected into them relatively unmarked. The selections enabled the speakers to preserve grammatical distinctions and strategies that they found normal to express and use, respectively, although some of these are somewhat simplified. From the point of view of communication, speaking the English way would have been less natural to them; those particular English morphosyntactic options must have been marked. Based on their uncommon attestations among the world's languages, the selected features may be considered marked from a typological perspective. However, the contact

Transfer and the Substrate Hypothesis in Creolistics situation, which defines which specific options are available to the speakers/learners, is what is relevant to determining what is marked and what is not to them. The speakers'/learners' particular selections are made unmarked especially by mutual reinforcement of their productions on each other's. An important factor in this case is also what Chaudenson (1992, 2001, 2003) characterizes as the absence of "normative pressures," referring to the absence of institutional mechanisms which, in native-speaking populations (for instance), stipulate for the speakers what is acceptable or unacceptable. Where conflicts arise from the markedness values assigned to some features (e.g., when different factors identify a particular linguistic feature as both unmarked and marked), a weighting of factors may come into consideration. It may thus favor one of the factors, e.g., SEMANTIC TRANSPARENCY (Seuren & Wekker 1986), SALIENCE (Mufwene 1989b, 1991b, 2001a), or INVARIANCE of form and/or free morphemes (Andersen 1983). This seems to have been the case with TENSE-ASPECT delimitation strategies in the majority of Creoles: they have generally selected the periphrastic strategy, which is more salient than other equally transparent systems such as agglutination or synthesis. For instance, Kituba has selected Kikongo's seemingly marked periphrastic alternatives (e.g., mene "finish" + V for PERFECT and kele/vanda "be/stay" + V for DURATIVE) over the more common and apparently unmarked agglutinating options, although it has maintained the oka suffix for PAST (probably because there was no competing alternative even from the other languages that Kimanyanga came in contact with). In English Creoles, the PERFECT construction done "finish" + V has typically prevailed over the alternative have + Past Participle, probably because have is typically contracted and \e can be inconspicious to a nonnative speaker once cliticized on the verb, which makes the tense-aspect distinction between the PRESENT and the PERFECT more difficult to make while that between the past participle and the base verb form is also being lost. For the same reason, the DURATIVE de/& + V construction is preferred over the be + V-m construction, largely while the weak -in suffix is being lost. An emergent pattern of regularity also favors marking TENSE and ASPECT preverbally over a "system" that expresses them postverbally and/or with a verbal suffix, creating some internal inconsistency. In varieties such as urban Lingala, which have preserved an agglutinating TENSE-marking "system" (e.g., a/ofc-/"he spoke-NEAR PAST," alob-d "he spoke-REMOTE PAST," alob-aka "he speaks-HABiTUAi"), the significance of both lexical and grammatical tones contributes to the salience of the TENSE suffix.10

8.3.2

What students of SLA could learn from genetic creolistics

Most studies in SLA, as in theoretical linguistics, discuss markedness from a typological perspective, as if features or constructions that are unmarked in one language or setting must be unmarked in other languages or settings world-wide. This may be illustrated with two studies on preposition stranding and pied-piping which struck me as particularly interesting when I wrote the original version of this Chapter for the (1990) publication: Bardovi-Harlig (1987) and White (1987), which, ironically, assume contradictory positions

153

1 54

Language Evolution on the subject matter. Stranding a preposition (as is permitted in only some languages like English) is considered by White to be less marked than pied-piping it with the fronted Wz-phrase; just the opposite is assumed true by Bardovi-Harlig, on the grounds that the latter option is cross-linguistically more common. Not considered in these studies and the theoretical linguistic literature on which they are based are languages which allow neither syntactic strategy but resort to other strategies discussed below. While many languages do not allow pied-piping at all, the system selected by a Creole depends apparently on what is unmarked or less marked in the contact situation. For instance, in the case of Atlantic Creoles, it is noteworthy that many African languages (Kwa and Bantu) show a predilection for a resumptive pronoun in the relativized NP's position, which is preferably gapped in English. They share with, e.g., nonstandard English, to which the Africans were the most exposed, the absence of pied-piping (*Dis'ere the boy with weh you go to d'city "This [is] the boy with whom you went to the cirf'/Dis'ere the boy you go to d'city wid "This [is] the boy you went to the city with"). Fronting the preposition along with Wh-phrase is not typical of the basilects of Atlantic English Creoles; the preposition is typically stranded in English Creoles as in colloquial and nonstandard English varieties. It is sometimes gapped together with the relativized NP (as in T s only ting weh covetin happier "It's the only thing that coveting is happier than") or used with a resumptive pronoun (as in Teddy da man (weh) everybody taller than him/0 "Teddy is the man (that) everybody is taller than"). (All these examples are from Mufwene 1986c.) Interestingly, Haitian Creole allows resumptive pronouns too, but without a relative pronoun, as in the following construction cited from Valdman (1978: 278): Moun ou rive ak li (lit. person you arrived with him/her) "the person you arrived with." It is interesting that such phenomena that may reflect substrate influence (possibly by congruence with nonstandard English and French) affect relativized NP's in the lower levels of the Keenan & Comrie NP Accessibility Hierarchy (1977). The possibility of substrate influence from African languages is made stronger when we pay more attention to French Creoles. What they do is closer to what most African languages do than to the French "system," in which pied-piping is required (as in the following construction from Valdman 1978:278: Se boug-la ar ki mo koze "The man with whom I talked") but preposition stranding is disallowed. However, as Valdman also points out (279), some of the seemingly divergent constructions are attested in nonstandard French, as in Cest legars queje te cause "It's the man I am telling you about" and Cest le type queje suis venu avec lui. "It's the man I came with." In any case, nothing so far has suggested that alternatives to what they do (especially pied-piping) are marked among the world's languages. Because the Africans were the majority in practically all the settings where Atlantic Creoles emerged, we can surmise that congruence may have been a significant factor. The non-pied-piping strategy with or without a resumptive pronoun (depending on the syntactic function of the relativized NP) must have been more natural and less marked to the adult L2 learners

Transfer and the Substrate Hypothesis in Creolistics of colonial English, being the strategy which most, if not all, of them shared and probably produced the most frequently in their initial interlanguages. In English Creoles, the fact that preposition stranding is the dominant strategy in nonstandard English just made the non-pied-piping option a stronger preference in the relevant contact situation. The above does not prove, of course, that the non-pied-piping option is not the unmarked setting in UG; a more extensive cross-linguistic comparison would help in addressing this question. Nor is it even necessary to assume that UG specifies which parametric options are marked and which ones are not. From an evolutionary perspective, it would be difficult to explain why some populations would have preferred marked options over unmarked ones. Assuming that UG just specifies the parametric options without ranking them leaves it up to chance to determine which particular options a population would select. Even if not pied-piping turned out to be the unmarked parametric setting regarding movement of a W7z-constituent which is a prepositional object, the proposed ecology-specific explanation (invoking African substrate influence) would apparently be the first thing to consider, as learners are naturally inclined to select strategies familiar to them when there is congruence. The principle of least effort still applies even though the learner focuses on how the TL differs from the language(s) they already speak. Because the adult Bozal slaves targeted languages that they expected or knew to be different from their own, and because they made an effort to learn them the best way they could, much of the cross-creole variation may be attributed to differences that existed already between the "lexifiers," e.g., pied-piping in Papiamentu, as in Portuguese and Spanish, as opposed to preposition stranding in English Creoles, as in English. However, several divergences from the "lexifier," e.g., absence of pied-piping and usage of a resumptive pronoun in French Creoles, can plausibly be accounted for by substrate influence, if, as explained above, there were in the contact situation factors which made the relevant substrate alternative less marked. The fact that Melanesian expanded English pidgins have selected several features which may be considered typologically marked among the world's languages is among the strongest pieces of evidence for the thesis that Creoles select grammatical features which were less marked in the respective feature pools of their "creators." (The application of this observation to SLA is given below.) I conjecture that whether or not a feature selected by a particular Creole is statistically unmarked among the world's languages is only secondary, if relevant at all, to the fact that it is unmarked in the contact situation. The point made above about conflicts in markedness values assigned by different factors is also worth elaborating on here. For instance, as shown in Mufwene (1988c, 1991b) and noted partly above, Kituba has selected for its time reference and pronominal systems the free periphrastic strategy with free aspectual markers and with only independent pronouns (without the clitic pronouns which are canonical in Bantu). In Kikongo-Kimanyanga, its "lexifier," as in other Bantu languages, the periphrastic strategy is less common than the agglutinating strategy for time reference. Also, their independent pronouns are used

155

1 56

Language Evolution primarily for emphasis or contrast; and they may be considered marked in these "systems ," in which pro-drop is typical. However, the same marked strategies are more salient, and this factor seems to have favored the independent pronouns and the periphrastic strategies as unmarked, or less marked than the agglutinating affixes, in the contact ecology in which the inflections did not always have consistent position relative to verbal root but also varied from one language to another, aside from the fact that TENSE-ASPECT is sometimes expressed only by tone on ethnic Kikongo varieties. The above situation suggests that the factors which determine markedness values must be weighted differently (just in case there is a conflict). The proposed weighting of factors which determine markedness can help explain why only one alternative is transferred in those cases where two or more L2 features appear to have equal markedness values. For instance, preposition stranding and pied-piping seem to be equally salient in English. However, the fact that preposition stranding is the dominant pattern in nonstandard and colloquial varieties tips the scale against pied-piping. This may explain why preposition stranding is transferred more readily by native English speakers "acquiring" a foreign language. The same weighting also explains why it is "acquired" first by those learning English as an L2. One need not invoke typological markedness to account for such common phenomena in language "acquisition."

8.3.3

Points of mutual interest

Many linguists have misguidedly analogized SLA and the development of Creoles in ways that suggest absolute parallelism. The patterns of substrate influence in Creoles suggest that group SLA (as it partially applies to Creoles) must be different and certainly more complex than individual SLA (the focus of most of the literature on SLA, even when the relevant studies contain generalizations about populations of immigrants, such as the migrant workers in France and Germany). While in both cases, failure to correct deviations from the TL is an equally important factor, it is also noteworthy that during the development of Creoles the TL is most used for communication among nonnative speakers or speakers of already divergent varieties than with native speakers of the original variety of the homestead period. This situation tends to reinforce nonnative characteristics, leading them to become the norm in the emergent variety. Just the opposite is the case in the context of migrant workers, for instance. They live in communities which are relatively homogeneous ethnolinguistically and isolated from the host population. As they socialize mostly among themselves, they use their ethnic languages as their vernaculars. They use their approximations of the TL typically to communicate with its native speakers and occasionally with immigrants from different nations and/or outside their communities, though such interactions are limited to necessary and non-social encounters and a few accidental ones. Frequency and nature of interactions with native speakers, as well as individual skills in L2 learning, can lead them to

Transfer and the Substrate Hypothesis in Creolistics reduce their deviations from the TL. There is no particular communal variety emerging from their interlanguages, because their communal vernacular is their ethnic language. The similarities observed between features of individual speakers' interlanguages and some structural characteristics of Creoles should not distract us from one important question: How do communal norms emerge from the individual behaviors of speakers who are not always behaving in identical ways? Communal characteristics emerge not because every speaker innovated the same features but more typically because many speakers accommodate others by copying the latter's patterns, for any number of reasons, even if the patterns are recognized by the copiers as divergences from the TL. Ultimately, the reasons why some of the features are widespread in particular communities are not necessarily the same as those that account for their presence in the interlanguages of L2 learners. Thus, the context of the development of Creoles adds a communal dimension to the factors which determine markedness; for instance, it maximizes the significance of FREQUENCY, since this may be amplified by the sheer number of speakers rather than by structural distribution of a morpheme or construction, or by the extent to which the languages in contact are typologically related. Even if a particular feature is not dominant relative to the different languages in contact, it may well become dominant if the speakers of a particular language using it are a significant majority in the contact situation at the formative stage. In this situation, frequency of use may either make a feature unmarked or create a markedness value conflict with another factor which makes the same feature marked but an alternative unmarked. This can account for why periphrastic marking of TENSE and ASPECT has prevailed in all Creoles, notwithstanding other factors such as SALIENCE of a FREE MORPHEME (as opposed to inflections) and REGULARITY. We must bear in mind that the TL varieties did not systematically mark all tenses and aspects periphrastically, and the substrate languages did not all have isolating or analytical morphosyntax. The emergence, under such contact conditions, of a dominant feature that diverges from the dominant pattern in the "lexifier" accounts for some structural peculiarities not only of Creoles but also of nativized versions of European languages in former European colonies. Conflicts and therefore variation in the grammatical strategies utilized in an emergent variety arise in those situations where more than one strategy is used for the same function, e.g., the different ways of expressing the PROGRESSIVE in Gullah, viz., with duh [da], V-mg, or duh V-ing, as in im (duh) talkin and im duh talk "he's talking." The same is true of alternative constructions such as Uh tell umfuh [fe] go and Uh tell um go "I told him/her to go/leave," or he gwine go home and he guh [ga] go home "he/she will go home," or whah make im duh cry and why im duh cry "why is he/she crying?" As shown in Mufwene (1992a), Creole "systems" are, like those of non-creole languages (and in an amplified manner), not monolithic. We are just scratching the tip of the iceberg regarding what we might learn about the development of new linguistic "systems." LI features attested in the speech of an individual L2 learner are not necessarily typical of substrate

157

1 58

Language Evolution influence in contact communities where this individual's first or primary language is only one of the many substrate languages.

8.4

Conclusions

In this Chapter, I have attempted to give an informative historical synopsis of the substrate hypothesis, to present various versions of it, to assess their strengths and shortcomings, and to compare research on substrate influence with that on transfer in SLA. With this in mind, the following main points may be highlighted. The main shortcomings of the substrate hypothesis have been methodological, with most of the claims being based on insufficient evidence. Otherwise, any of the versions discussed in Part 2 are conceivable, provided the right sociohistorical conditions backing the relevant claims are adduced. On the other hand, some changes, though limited, were noted toward more convincing arguments in which sociohistorical and formal aspects of transfer of features were addressed which up to the 1970s had generally been lacking (e.g., Maurer 1987, Keesing 1988, and Singler 1988). Developments such as the complementary hypothesis are also a welcome change, most notably in Corne (1999). However, more work is needed with regard to determining what factors favor transfer of what features in what sociohistorical and linguistic settings. This is where both the suggested founder effect and markedness considerations (as summarized below) become particularly relevant. While markedness has come to play an important role in this research, its values appear to be determined relative to specific contexts of language contact and not relative to UG. It is not even clear why such considerations would shed any light on how languages have evolved, with some populations selecting options that are marked, therefore less preferred. Random probabilities seem to have determined how different typological options have evolved in different language varieties. Nor does it seem relevant at all to invoke typological markedness to explain the selections made by particular speakers/learners in various language contact settings. How would some options matter that speakers/learners do not even know? Only ecology-specific scales of markedness seem to make sense. Moreover, since more than one factor is often involved in determining what typological option is less marked and since the factors can assign conflicting values to the same option, it appears necessary to weight them. For instance, SALIENCE can be more heavily weighted than FREQUENCY alone, or FREQUENCY can tip the scale for one out of two features that appear to be equally SALIENT. Also, the specific ecologies in which particular languages evolve can vary in the ways the different factors are weighted, for instance MORPHOLOGICAL REGULARITY may be favored over TRANSPARENCY in one language, while just the opposite is true in another. Contrary to what I thought in the 1990 ancestor of this Chapter, it has become obvious to me that while research on SLA can inform us about conditions that favor transfer during L2 "acquisition," it cannot inform us on how substrate elements influence the development

Transfer and the Substrate Hypothesis in Creolistics of Creoles or any other languages for that matter. Substrate influence is a population-level phenomenon that results from both the repeated occurrences of xenolectal elements in some idiolects and their spread within the population of speakers when other speakers simply copy them. To date, research on naturalistic SLA has focused on interlanguages, with the expectation that they will, or may, evolve into closer approximations of native linguistic "systems." It has also focused on individual learners. Despite frequent reports of tendencies observed in groups of learners, there have also been reports of variation in the phenomena observed among learners of the same ethnolinguistic background (e.g., Perdue 1995). Among other things, this underscores the need not to expect all speakers of the same African languages in plantation colonies of the New World and the Indian Ocean to have produced the same deviations from the European languages they targeted. On the other hand, one must remember that research on SLA has not studied the emergence of communal norms. Thus, research on (naturalistic) SLA offers nothing that can be compared to the interidiolectal mechanisms of competition and selection that led to the emergence of communal norms in Creoles. Creolistics must rely on itself to figure out the ways in which the "invisible hand" of evolution has driven the structural divergence of Creoles from their "lexifiers."

159

9

Grammaticization in the Development of Creoles1 Chapter Outline 9.1 9.2 9.3 9.4 9.5

9.1

Introduction The Bruyn-Plag position: how justified is it? How "creolization" proceeded in Suriname Grammaticization in Creoles Conclusions

160 163 171 172 178

Introduction

Until about the mid-1990s it was taken for granted that grammaticization is among the restructuring processes that produced Creoles from their "lexifiers ," in that they extended forms or constructions of the latter to new grammatical functions. Such developments include the following examples: (1) usage of a serial verb meaning "give" after a head verb to convey a DATIVE or BENEFACTIVE meaning, viz., [NP1 "buy" NP2 "give" NP3] in, for example, Saramaccan and Haitian Creole, which is interpreted as "NP1 buys NP2 for NP3"; (2) predicative usage of the preposition [f9-fi/-fu/pu] "for" before a verb to express an OBLIGATION modal meaning in several English and French Creoles, e.g., Sara bin fuh [fa] come in Gullah and Sara te pou [pu] vini Haitian "Sara had to/was expected to come"] (3) usage of the verb stay before a main verb in Hawaiian Creole, or stop in Tok Pisin, as a DURATIVE marker; or (4) extending the function of past participle been in English Creoles, or the inflected copular forms ete/ etait in French Creoles, as an ANTERIOR tense marker.

Except for identifying the origins and/or models of the constructions in the "lexifier" or in the substrate languages, nobody has seemed particularly concerned by whether such evolutions were internally or externally motivated. In the way that Bruyn (1996) reviews the state of the art (see below), the latter alternative involves cases where a Creole has selected and/or generalized a strategy that was already developing in the "lexifier," for instance, using gwine or a go in Jamaican, or go [ga] in Gullah, as a FUTURE marker, on the model of be going to/

Grammaticization in the Development of Creoles gon(na) + Verb in English (Mufwene 1996b). It also includes cases where a Creole's construction is patterned on some substrate language(s), such as the use of a verb meaning "say" as a complementizer after verba dicendi and verbs of perception, consistent with similar uses of "say" in several West African languages, in the case of Atlantic English Creoles. Bruyn (1996) and Flag (1999) seem to have taken to heart the discontinuity hypothesis, according to which, Creoles are de novo creations from whatever was left from their lexifiers after breaks in their "transmission" to nonnative speaking populations (see Chapter 3). Putatively, the grammatical patterns they share with their "lexifiers" and some substrate languages may be considered as borrowings. Thus, cases of grammaticization in Creoles that are patterned on similar developments in the "lexifier" or the substrate languages only instantiate "apparent grammaticization." Accordingly, true cases of grammaticization should strictly be internally motivated. As shown in Chapter 3, the discontinuity hypothesis has no grounding in the histories of the territories in which Creoles have developed. Like other language varieties, Creoles are outcomes of "transmission" with modification, the default in language "transmission" (Mufwene 200la). Continuity in the "transmission" of a language does not presuppose (constant) presence of native speakers in the community that uses it, as has been clearly demonstrated by the histories of pidgins and indigenized varieties of European languages all over the world. Likewise, the traditional distinction between internally and externally motivated change is assumed here to be simply sociological, justified only by factors that do not have to do with language structure per se. The nature of the negotiations and accommodations that speakers make to each other do not change, regardless of whether these processes are prompted by idiolectal differences that are xenolectal or independent of influence from other languages. Therefore, although it helps us determine whether a change was motivated by pressures from within a particular community of speakers or resulted from contact with another community, the distinction between internally and externally motivated change sheds no particular light on the mechanisms of language change themselves. Despite the fact that they are not exactly the same processes from one language-specific case to another, the mechanisms are of the same kind both in those changes assumed to be internally motivated and in those which are said to be externally driven. The "invisible hand" that drives evolution always lies in the ecology, which is external, regardless of whether the trigger lies in speakers' attempts to reduce differences among their idiolects and/or between their dialects, or whether it lies in influence from another language. Even needs to develop some symmetry or fill gaps within a "system" are triggered by factors external to the language itself, bearing in mind that all changes reflect actions of speakers during their communicative acts and the speakers are indeed external to the language they speak (Mufwene 200la). These considerations underlie my arguments below against Bruyn's (1996) and Flag's (1999) position that many cases of grammaticization in Creoles are only apparent.

161

162

Language Evolution Grammaticization is a diachronic process. At the communal level of a language, it is subject to the observations articulated above. Bruyn's and Flag's position is disputed at least by the alternative conception of the evolution of Creoles developed in this book, which assumes no break in the "transmission" of the "lexifier." However, Bruyn and Flag provide me an opportunity to emphasize a point made in Mufwene (1996b), viz., that research on the development of Creoles can contribute to scholarship on grammaticization by showing that such processes need not be unilinear nor rectilinear. I show below that this position is consistent with that of Traugott (1999:210), who, contrary to Flag (1999), observes that "the study of contact languages can be particularly instructive for grammaticalization theory." As made more evident by Heine & Kuteva (2005), the model for grammaticization can lie in another language. The process itself is typically gradual. Even if the speed appears to be more rapid in the case of Creoles than in other language varieties, speed in itself does not dispute the gradual nature of its development (Mufwene 200la, 2003a). As pointed out in Chapter 3, the empirical foundation of the claim that Creoles have evolved faster than other language varieties is quite dubious. In the particular case of North America, for instance, Gullah has developed within the same time period as other American English varieties, especially in the east. In any case, as far as Creoles are concerned, it does not matter whether a grammaticization process originated in the "lexifier" or was influenced by some substrate language(s). This is a natural consequence of how competition and selection proceed in the feature pool provided by the contact ecology. Consistent with the uniformitarian position developed in this book, the particular way grammaticization proceeds even in the case of Creoles is by co-opting morphosyntactic behavior that is already part of the extant grammar. For instance, as shown in Mufwene (1996b), the FUTURE construction with a go + V in Jamaican Creole is a caique of the be going to/gonna + V construction in English. It differs from the latter by co-opting the established way of forming the PROGESSIVE/DURATIVE by anteposing the locative marker a to the main verb, as in im a run "he is running." Likewise, usage of the preposition [fb/fo/fi/fu] and [pu] "for" as an OBLIGATION modal auxiliary in English and French Creoles is quite consistent with the fact that, unlike their "lexifiers," the grammars of these vernaculars allow predicate phrases that are not headed by verbs, as in Mary tall "Mary [is] tall" and this book for you "this book [is] for you." Such developments are indeed among the restructuring processes that have shaped Creoles as structurally different from their lexifiers. Thus, it should come as no surprise that, having evolved from their "lexifiers" through the same kinds of restructuring processes that are observable in other languages (Mufwene 200la), Creoles would normally share many structural features with them or with some of the substrate languages. Contrary to the fundamental claim of the language program hypothesis that children would have developed the syntax of Creoles using a blueprint provided by the bioprogram (Bickerton 198 Iff), very few, if any, grammatical strategies of significance appear to have been worked out without (partial) models in the languages that

Grammaticization in the Development of Creoles came in contact. In many cases, it is a simple matter of congruence between structures of the nonstandard European "lexifiers" and those of some substrate languages (Corne 1999). Although Bruyn (1996) and Flag (1999) may have intended for research on grammaticization to focus on those cases that can be attributed to the internal dynamics of Creoles' "systems" after they had been stipulated as separate languages, there is no particular reason for ruling out the other, earlier cases that are part of their normal evolution as only "apparent." The mechanics of the process are in no way different. The evolution has been continuous. The division between the formative and post-formative periods is arbitrary.

9.2 The Bruyn-Plag position: how justified is it? 9.2.1 Bruyn (1996) starts her discussion quite strangely with the suggestion that there is a synchronic notion of grammaticization, which is applicable to the development of Creoles. In her own words: Although grammaticalization is generally conceived as a diachronic process, it is possible to approach it from a synchronic point of view. (29) With regard to Creole languages, however, there are reasons to question the assumption that synchronic phenomena reflect a diachronic development in the way in which grammaticalization is normally conceived, namely proceeding gradually and language-internally. (30)

Clearly, according to her, things that look like grammaticization in the development of Creoles are quite distinct from similar processes that have been observed in other, non-creole language varieties. She then proceeds to observe: The fact that ben does not function as either a particle or a lexical verb in the early stages of Sranan implies that there has been no language-internal development of a lexical item into a functional one. Rather, there must have been a kind of short-cut from the English participle, which has already some grammatical value, to the function of tense marker in the Creole language. To regard such a development as grammaticalization without further qualification would be unrevealing to the extent that in this case the process must be assumed to have taken place from one language into another, such that patterns of usage in English provided the starting point for a grammatical form in Sranan or other Creoles.2 (30-1)

Clearly, Bruyn subscribes to the discontinuity hypothesis, discounting the evolution of ben into an ANTERIOR marker as a mere "short-cut from the English participle" been and suggesting that the internal dynamics of the emergent system of Sranan and other English Creoles had no role to play. Although colloquial and nonstandard English varieties avail constructions such as he s BEEN ready/gone in reference to states of affairs that took place a long time

1 63

164

Language Evolution ago (in the subjective perception of the speaker), their been functions more as a REMOTE PHASE marker, like in AAVE, than as an ANTERIOR marker. While the contracted has/is is variably missing in AAVE, it is never used in Creoles. The evolution of ben as an ANTERIOR marker in English Creoles appears to have proceeded like other cases of grammaticization. It involved, subject to other constraints imposed by some of the previously developed aspects of the emergent "system," an extension of the usage of a lexical item to a new, grammatical function, regardless of whether the older uses are strictly lexical or grammatical. Why should a language such as Sranan, which is identified as an English Creole, be expected not to have patterned some of its grammatical constructions on the "lexifier" itself in order for these to count as internally motivated? Does any normal speaker ever learn words of a language without paying attention to their patterns of usage, although its "acquisition" need not be perfect (Chapter 8)? One of the diachronic questions arising from Bruyn's comments is the following: How clearly different was the ancestor of modern Sranan, or of any other Surinamese Creole, from the colonial nonstandard English spoken locally by a sizeable proportion of European nonnative speakers in the late seventeenth century? The end of the seventeenth century is when large plantations began to develop and the African-to-European demographic disproportion became drastically lopsided (over 10:1) in favor of Africans, creating an ecology favorable to (more) substrate influence from substrate languages. As explained in Chapter 3, Creoles have not developed from erstwhile pidgins, contrary to an assumption that seems to underlie Bruyn's position. Instead, Creoles have evolved continuously by basilectalization from closer approximations of their "lexifiers." Even Suriname did not provide the conditions that would back the discontinuity hypothesis. According to Arends (to appear), some English traders remained in Suriname after it became a Dutch colony. We must also recall that the shift from English to Dutch colonization took place during the homestead phase, before any plantations had developed, let alone large ones, thus during a period when some of the slaves would have been speaking closer approximations of English nonstandard "lexifier." This position is in fact consistent with the scenario that Bruyn (1996) herself presents, and has in other studies of eighteenth-century texts, about the development of Sranan. There is no evidence of an initial pidgin. The earliest texts are less basilectal than the later ones. Thus, even after segregation had been instituted, the "lexifier" continued to be "transmitted" among the slaves as much by the Creole slaves as by the Dutch colonists, as well as by some English colonists who continued to do business in Suriname. If these are the ecological conditions that have been interpreted as breaks in the "transmission" of the "lexifier," the analysis is obviously misguided, since English need not have been spread everywhere by its native speakers nor, more specifically, by the English colonists. A lesson can indeed be learned from all over the world since the mid-twentieth century, and since earlier from the

Grammaticization in the Development of Creoles British exploitation colonies, where English has been spread more by nonnative than by native speakers. Then Bruyn observes: In Keesing's (1991) discussion of grammaticalization in Melanesian Pidgin, he concludes that lexical borrowings from English can acquire grammatical functions that correspond to those in substrate languages. (31)

As noted in Section 9.1, there is nothing in grammaticization that inherently precludes borrowings from introducing a new structural pattern or function in a language. On the other hand, what makes the position to which Bruyn subscribes questionable is also the fact that forms or constructions selected and extended from the "lexifier" itself should be treated as borrowings. This is tantamount to claiming that any grammatical forms and structures that French has selected and grammaticized from Latin - such as its INDEFINITE SINGULAR and DEFINITE articles, its "passe compose" with avoir "have" (from Vulgar Latin habere "have" + Past Participle), and its synthetic FUTURE construction (a coalescence from Vulgar Latin infinitive/gerund + habere) - must be treated as borrowings, because the grammaticization processes themselves started in Latin. This position is also questioned by all the evidence which suggests that Creoles have evolved like other modern European languages by gradual divergence from the structures of their ancestors, for instance, Old English in the case of modern English. It all sounds as if the myth that Creoles have developed in their own "exceptional" way must be preserved, contrary to the increasing evidence that prompts us to re-examine at least some of our traditional working assumptions. Even if one assumed, without justification in fact, that language boundaries were clearly delineated in contact settings, it is not clear why legacy from the "lexifier" or influence from substrate languages would entail that such cases of grammaticization in Creoles are synchronic processes. "Borrowings," as Bruyn identifies these cases, are normal parts of language change, reflecting the linguistic contacts of a population. They are among the natural ways in which speakers modify their communicative habits, and they remain bona fide concerns of historical linguistics. Thus, they are diachronic data, more or less like fossils in material culture. Since Creoles have been disfranchised a priori as "mixed languages," which contain elements from more than one parent, restricting grammaticization to those restructuring processes that are subsequent to their formation raises the question of why (partial) continuities from the parent languages do not count. These continuities are in fact what accounts for the classification of the new vernaculars as "mixed." Moreover, since most of the recent scholarship on the development of Creoles has been more in support of gradual, rather than abrupt, evolution (Chapter 3), how does one determine the particular time at which a Creole may be considered to have completed its formation? Are Creoles not like other

165

166

Language Evolution languages in a constant state of flux and therefore in the process of being reshaped as they are being spoken, variably, by their users? Bruyn also notes: Since patterns in substrate languages may have arisen through universal grammaticalization developments, it will often be difficult to establish whether something that looks like grammaticalization is the result of internal development in the Creole language, or of calquing of substrate patterns on lexifier forms, or of the convergence of universal conceptually based developments and features of the languages present in the contact situation. (31)

Those universal paths of grammaticization which are presumably independent of languagespecific systems are yet to be determined. There must definitely be some difference(s) between, on the one hand, evolutionary paths that languages tend to follow (almost) uniformly because they are based on some principles considered universal3 and, on the other, some developments that are consequences of traits that are specific to them. Even in this case, one must wonder why some languages select a locative preposition or adverb, or a verb meaning "be at a place," for the PROGRESSIVE function (basically what Kiyansi, one of my native languages, Bantu B85, does with NP wii "here'Ywaa "there" + Verb, depending on the location of the speaker relative to the agent), whereas some languages select a verb meaning "stay" or "sit" (what Kikongo-Kituba, my second native language, and Hawaiian English Creole do, the former only in the past tense only) but not a verb meaning "lie." Nor is it clear yet why some languages select a verb meaning "want" but some others select a verb meaning "go," and yet a few others select a verb meaning "come" (Kiyansi is among these), to express the FUTURE. We cannot in any case dodge the question of whether there can really be universal paths of grammaticization which are not simply statistical reflections of what languages of any, or some, typological affiliation tend to do. Otherwise, why would some languages stray from these putative universals that must guide all of them?4 Bruyn also seems to muddle the debate about the development of Creoles as "mixed languages" whose structural features have been selected from both the "lexifier" and substrate languages. If she intended to prove that Creoles have evolved by restructuring processes that are different from those of non-creole languages, the putative cases of "apparent grammaticalization," as she identifies them prove just the opposite. As shown in Chapter 3, every language that has undergone some major structural change at some point or another in its history has done so under contact conditions. For instance, Old English and Old French developed under conditions of population and language contact. In the former case, different Germanic languages that need not have been more mutually intelligible than modern Dutch and German came in contact with each other and with Celtic languages in a setting exogenous to their cultural heritage. In the latter case, (varieties of) Vulgar Latin came in contact with Gaulish, Celtic languages in settings endogenous to the new speakers.

Grammaticization in the Development of Creoles The case of the evolution of Vulgar Latin into the Romance languages is even more comparable to that of the emergence of Creoles, because it too involved language shift, although, as explained in Chapter 11, this was probably more gradual, or less rapid, than in the case of the formation of Creoles.5 Even the developments of English and the Romance languages from the Old-X stage to Middle-X stage were triggered by other layers of population and language contact. The fact that historical linguistics has not addressed the actuation question (Chapter 2) but has been contented with formulating rules of the form X—> Y under particular structural conditions is not evidence that there were no ecological triggers of the relevant changes. Pace Thomason & Kaufman's (1988) argument against Old Norse and Norman French influence on Old English, one cannot say for sure that the development of, for instance, the Great Vowel Shift and relative clauses introduced by WH and involving piped-piping would have been ineluctable in standard English without the contact of Old English with both Norman French and Latin. Although one can argue that usage of WH in relative clauses is also evident in German and Dutch, which would make Norman French influence just an accident of history, the fact that English alone uses pied-piping in this case suggests that there was much more role in the congruence of Norman French than Thomason & Kaufman admit. French influence on the English language is far from being contained only in its lexicon and phonemic system. Even the prevalence of nominal PLURAL marking with -s, unlike in German and Dutch, may be a reflection of this particular contact with French. Bruyn's stipulation to exclude from the category of grammaticization processes in Creoles all those cases where either the model came from substrate languages or the new pattern developed thanks to some congruence between the lexifier and some substrate languages seems unjustified. She then proceeds to consider the following data: The patterns exemplified by (1-3) occur in the eighteenth century Sranan sources as well as in those from the later periods. What is particularly relevant is that PPs of the type na baka fu NP do not occur relatively more frequently in the eighteenth century - as would be expected to be the case if there had been diachronic development. This indicates that the occurrence of the various types of complex PPs in present day Sranan is not the result of a gradual process of grammaticalization of items such as baka from nouns in the direction of adpositions. (1) na

a

oso

baka

[ 18th-2Oth century]

LOC the house back 'behind the house' or 'at/in the back part of the house' (2) na LOC

baka

fu a

oso

[18th-2Oth century]

back of the house

'behind the house' or 'at/in the back part of the house'

167

168

Language Evolution (3) na baka a

oso

[18th-20th century]

LOG back the house 'behind the house' (4) baka

a

oso

[20th century only]

behind the house

(34)

It is not clear why gradualness in the development of these strategies would presuppose a more frequent attestation of na bakafu NP in early Sranan text. The data presented here just illustrate layering (Hopper 1991) and competition among various structural alternatives. It is even debatable whether, in the first place, post-nominal baka in such Sranan constructions as in (1) really functions as an adposition. It has syntactic characteristics of a noun modifying another noun, as in house front-front of a/the house or desk top - top of a/the desk. The fact that the eighteenth-century texts do not show a higher frequency of post-nominal uses of baka for LOCATION is perhaps further evidence of the fact that Creoles did not start from "systems" that were more different from their "lexifiers" and that the features that have caused them to diverge as much from their "lexifiers" as they do today are really later developments (Chaudenson 1992,2001,2003). We must also be cautious, because the earlier texts mostly prove that a Creole was already emerging or in place; and they show what particular features were identified by a writer as particularly representative of a colonial vernacular. They are not really evidence of the statistical distribution of variants in earlier Creole speech. I am still not convinced that a process is less diachronic or less gradual because it has occurred in less time than another (Mufwene 200la). Finally Bruyn concludes: The historical investigation of certain aspects of Sranan suggests that there are two main dimensions along which developments in Creole languages may differ from ordinary grammaticalization in languages with a longer history: they may proceed at a higher rate, and they may be crucially determined by influence from other languages. (39)

Thus, ignoring the particular contact conditions of rapid language shift for slaves, which were conducive to rapid convergence in the changes to which they subjected the target language, she just jumped to the conclusion that the evolution is not "usual" and must therefore amount to "apparent grammaticalization" (40). I think the argument in Mufwene (200la) that rapid change does not deny gradual evolution already undermines this conclusion. Grammaticization in Creoles appears to have proceeded as normally and naturally as it does in other languages. Casting it aside as a process that is only similar but does not quite fit in what has been observed in other languages may lead us to miss the opportunity that the study of Creoles is providing us to learn more about the mechanisms of grammaticization as a diachronic process and as a facet of the restructuring that brings about change in a language. Certainly, true answers to questions about language evolution cannot lie exclusively

Grammaticization in the Development of Creoles in non-creole languages. Sometimes Creoles just suggest that the old order of business is not necessarily the right one.

9.2.2 Endorsing Bruyn's position, Flag (1999:206) also claims that "grammaticalization is generally conceived as a unidirectional and gradual process." Although this claim has been disputed (Janda et al. 2000) - not so convincingly - I know of no case in the development of Creoles or any other language variety where grammaticization has proceeded upstream, reverting processes that took place earlier in the lexifier or other (substrate) languages. At least there is nothing that Flag brings up in his review to illustrate his observation. He may have intended "unilinear" development and thus, like Bruyn, he may have wanted to exclude substrate influence as a legitimate contribution to grammaticization. He may also have wanted to argue against Mufwene's (1996b) argument that grammaticization can proceed in bifurcated paths, the case of the parallel evolution of the preposition for in English Creoles both as a complementizer and as an OBLIGATION modal auxiliary, as in the following Gullah examples: 5 a. This bag (bin) fuh BB. "This bag is/was for BB." (PREPOSITION) b. We tell BB fuh come. "We told BB to come." (COMPLEMENTIZER) c. BB (bin) fuh come. "BB must/had to come" or "BB is/was expected to come" (MODAL)

As shown in Mufwene (1989a, 1996b), both grammatical extensions are concurrent developments from the PURPOSIVE function of the preposition for in the "lexifier," a function that Creoles have also maintained (consistent with Hopper's 1991 principle of layering). Such bifurcated evolution would not be limited to Creoles either. In English both the complementizer that and the definite article the have evolved from the same distal demonstrative pronoun that, which still functions as such to date. Like many other linguists who claim that Creoles have developed abruptly from their "lexifiers," Flag too must have confused the speed of a process with gradual evolution either in terms of the several stages that the process must go through or in terms of how long it takes a process to pass from initial attestations in some idiolectal systems to widespread usage in a communal language. I am not sure that this question has been sorted out in studies of other languages, for instance, how long it really took the Old English ancestors of the verbs will can, shall and the like to individually grammaticize as auxiliaries.6 If linguists had not been too eager to identify Creoles primarily with basilectal varieties (and then assumed, contrary to historical evidence, that basilects developed first), we probably would not have much motivation for assuming that grammaticization, from the point of view of spread throughout a population of speakers, was less gradual than in other languages. For instance, since not every Gullah speaker expresses the PROGRESSIVE with the preverbal marker duh [da],

169

1 70

Language Evolution there is no particular reason for assuming that its evolution is complete already. Nor is there a better one for assuming that in Jamaican Creole preverbal a go as a FUTURE marker is an evolution more complete than gwain or vice versa. Flag proceeds to argue as follows: Assuming the correctness of Bruyn's analysis of the historical development of the Sranan article wan as a true case of instantaneous grammaticalization, it is unclear how the notion of gradualness can be reconciled with her finding. If grammaticalization can indeed happen instantaneously, how can it be a gradual process? (207)

Independent of the fact that, to begin with, wan may not be an article at all but only a quantifier that happens to be used where the "lexifier" sometimes uses an indefinite article, one-stage developments would not be unique to Creoles. The transition from the Latin numeral quantifier unus to the French indefinite article un "a(n)" cannot have involved several intermediate structural steps, no more than there are in the evolution of the English indefinite article a(n) from the quantifier one or of the definite article the from the distal demonstrative that. Also, there are no communities where all speakers participate concomitantly in the innovation of a new form, construction, or process. Such developments spread gradually within a population, being truly innovated by one or a few speakers and then copied by others (Mufwene 200la). It is not clear what Flag and Bruyn have noticed in the evolution of Creoles that make them different from non-creole languages.

9.2.3 The reservations that Bruyn (1996) and Flag (1999) have expressed about grammaticization in Creoles are contrary to the interest that veteran students of this restructuring process such as Traugott (1999) and Heine (in progress) have. Traugott (1999: 210) observes: The usual assumption is that these changes [i.e., grammaticization processes] are internal and embedded within a relatively homogeneous history - this is where the study of contact languages can be particularly instructive for grammaticalization theory.

Indeed, as shown in Chapter 3, the assumption that there are any languages whose structures have evolved independent of any contact among its dialects and/or with other languages appears to be unjustified. The fact that there is idiolectal and dialectal variation within a language is already ground for assuming that contact plays an important role in the evolution of any language, especially since the locus of any mental representation of a language is the mind of every individual speaker. Anybody who interacts with other speakers and often has to accommodate their idiolects is an arena of lectal contact at the idiolectal, dialectal, or language level.

Grammaticization in the Development of Creoles Traugott then proceeds to observe that "different uses will crystallize in different ways, even in a homogeneous situation, a process called 'polygrammaticalization' (Craig, 1991)" (211). Regarding gradualness, she adds: The difficulty may be exaggerated. Reanalysis (innovation) itself is abrupt and not gradual (Hopper & Traugott, 1993, p. 36), and sometimes individual changes may follow one upon another in fairly short order, even in standard languages (e.g., the development of individual modal auxiliaries in the history of English). From the point of view of grammaticalization theory, what is gradual is typically accretion of properties that lead up to reanalysis (e.g., the processes that led up to the development of auxiliaries in English) and there may even be periods when the structure is somewhat indeterminate, especially as exemplified by blends (Tabor, 1994, Chapter 6). (211)

This statement is so well articulated that I will not even try to add anything to it. It will, however, help to recap briefly the respects in which I disagree with the conception of the development of Creoles that Bruyn and Flag suggest and why I therefore reject their position on grammaticization in these vernaculars.

9.3 How "creolization" proceeded in Suriname The issues discussed above stem primarily from how Bruyn and Flag seem to understand CREOLIZATION. Although they are not quite explicit about it, they seem to espouse the traditional assumption that Creoles have evolved from erstwhile pidgins, a view that they need to support the discontinuity hypothesis. This scenario is contrary to the history of Suriname and other creolophone territories, as shown below. The literature on Creoles and pidgins have led both Robert Chaudenson since the early 1970s and myself since the early 1990s to endorse the traditional characterization of pidgins as reduced second-language varieties that developed in trade colonies, where they functioned as lingua francas between, in my view, economically egalitarian partners who met sporadically for trade. (Shifts in the "balance of power" would eventually lead to the development of exploitation and settlement colonies from these trade relations.) The trade partners interacted minimally and could therefore be contented with the minimal communication that a pidgin enabled to establish between them. After all, they had their native, or primary, languages to communicate with members of their respective ethnolinguistic groups. Defined as above, pidgins as lingua francas cannot plausibly be situated historically in the homestead communities of settlement colonies in which plantation societies associated with Creoles eventually developed. Having generally started as homestead societies - with family-size population clusters (see Dunn 1972) in the case of the Anglophone Caribbean social interactions in these colonial communities were intimate (despite race discrimination),

171

172

Language Evolution relatively integrated, and marked by regular verbal exchanges. Their Creole children - of European, non-European, and mixed backgrounds alike - "acquired" their colonial vernaculars together and developed similar full-fledged idiolects, an asset which, according to Berlin (1998), would enable some of them to become power-brokers in colonial societies. Although some of their parents (non-Europeans but also some Europeans) did not speak the vernacular natively, at least all the locally born residents did, variable as the colonial vernacular was. The homestead communities also grew more by birth than by importation of labor, since by then most of the wealthy European residents had not accumulated enough capital yet to develop large plantations. Assuming pidgins to have developed in such settings of regular interactions is contrary to the working notion of PIDGIN articulated above. The position is also inconsistent with the histories of the development of plantation settlement colonies summarized in Chapter 3, according to which plantation societies developed later than, and gradually and non-universally from, the homestead societies.7 These histories suggest continuity in language "transmission" and the later development of Creoles during the time when, as explained in Mufwene (200la), the mortality rate was high, life expectancy short, and the plantation populations grew more by importation of labor than by birth. It was also the time when the local vernacular was being "transmitted" to the newcomers as much by nonnative, "seasoned" slaves as by native speakers, the Creoles. This demographic development fostered ethnographic ecologies in which, as one can observe today in the indigenization of European languages in former exploitation colonies of Africa and Asia, structures of the vernacular diverged farther and farther away from the original "lexifier." The reason is that more and more of its "transmitters" were nonnative speakers who themselves had "acquired" it from other nonnative speakers. With this misconception corrected, we can now focus on grammaticization as part of the development of Creoles and on how, consistent with the expectation stated above from Traugott (1999), our adequate understanding of the process in this context can enrich discussions of it in the broader literature.

9.4

Grammaticization in Creoles

Arguments for the gradual development of Creoles were also presented in Chapter 3. It was shown that even pidgins must have developed by gradual divergence from their "lexifiers." They appear to have started from the approximations spoken by interpreters during the earliest periods of contacts between Europeans and the indigenous populations to the more and more divergent approximations produced by almost anybody engaged in trade and similar interactions. This evolution proceeded all the way to some point when a communal norm would emerge in some long-lasting contact settings. The reason for underscoring the gradual development of these colonial varieties is to show why it makes little sense to claim that grammaticization proceeded in a different way in the development of Creoles than it did in the evolution of other languages.

Grammaticization in the Development of Creoles There is yet an unexpected gain from Bruyn's and Flag's discussions, viz., they unwittingly provide support for continuity in the development of Creoles, showing that these vernaculars have not been created ex nihilo and their structures contain forms and/or functions that have been selected intact or slightly modified from their "lexifiers" and/or the substrate languages. Mufwene (200la) accounts for such developments by arguing that the populations that produced Creoles selected their structural features, according to ecologysensitive markedness principles, from a feature pool to which both the "lexifier" and the substrate languages made contributions. In the vast majority of cases, the forms originated in the "lexifier," since it was already selected by the new socioeconomic ecology as the local lingua franca and later as the vernacular, and it was clearly targeted by whoever intended to adapt to this new ecology, pace Baker (1990). The substrate languages easily influenced selections of structural options which the "lexifier" shared with them. When particular modifications took place either because speakers did not replicate the targeted forms or constructions perfectly, or because they had to extend them to new communicative needs (e.g., expression of HABIT while the "lexifier" availed no specialized form/construction for it), substrate influence also helped shape the evolutionary trajectory of the "lexifier." As is well known in research on SLA misidentification of forms and structures of the target language with those of the source language leads to reanalysis or misinterpretation of the targeted forms and structures, i.e., to restructuring of the system. It is precisely in this context that grammaticization as adaptation of an extant form or construction to new uses is to be situated. In Mufwene (200la), I liken such grammatical adaptations to exaptations in biology or kludges in computing. They are unplanned, occurring on the spur of the moment when a communication need must be met. Only through repeated successful uses (call them "copies" at the population level, where they spread gradually) are they identified as instances of grammaticization in a language.8 In reality, however, the processes have been initiated at the idiolectal level, where they could very well be considered "idiogrammaticizations" (by analogy to idiolects). I need not go into the question of the relation between INDIVIDUAL SELECTION (at the idiolect level) and GROUP SELECTION (at the communal language level) in this Chapter. Suffice it to know that it is an important component of grammaticization, as of any diachronic process, driven by the "invisible hand" that drives evolution at the communal language level. There are no group selections without individual selections and therefore no grammaticization processes at the communal language level without initial processes of idiogrammaticization. Some of the individual selections converge or reinforce each other but some others do not. Those that converge or reinforce each other define some of the evolutionary paths of a communal language. The non-convergent ones, like other deviations and/or innovations at the idiolectal level, often die out without any consequence for the communal language. The few that converge into viable alternatives settle in the language, without necessarily driving the older forms or constructions out of the language but sometimes producing specializations by register or setting. They account for internal variation, which is very

173

1 74

Language Evolution real and an important characteristic of any living language. Examples of this state of affairs can be adduced from English, which has more than one strategy for relativization (viz., the complementizer strategy versus the relative pronoun strategy, with the latter attested most often in written registers); more than one strategy for expressing FUTURE (viz., the auxiliary verb will + Bare Infinitive versus the complex "semi-auxiliary" be going +to + Infinitive, with the latter specialized for INTENTION, among some speakers); and more than one modal construction for ABILITY (can + Bare Infinitive versus able + to + Infinitive versus capable + of+ Gerund). Though some partial grammatical and/or pragmatic specialization may justify having the alternatives, none of them is especially correlated with any change in progress. We cannot be surprised to find similar variation in Creoles, not only because the "lexifier" itself was heterogeneous and often amplified variation in metropolitan settings but also because the feature pool was made more heterogeneous by the xenolectal productions of some speakers of the substrate languages. As explained in Mufwene (1994, 2001a, 2005a), the "creole continuum" need not be associated with debasilectalization, not any more than a Creole need be equated with the basilect. We would be better off interpreting the whole continuum as the creole language variety, accepting its variation as normal as in the nonstandard-to-standard continua of non-creole languages.9 Studies of grammaticization based on non-creole languages have emphasized the importance of language-internal evolution (see, e.g., Hagege 2001). Their cross-linguistic comparisons have suggested "universal" paths of grammaticization, i.e., those that various languages, some of them genetically and/or typologically unrelated, have tended to follow, for instance, the tendency for PROGRESSIVE constructions to have developed from LOCATIVE ones, or for TEMPORAL markers to have been extended from LOCATIVE ones. Among the things that the hitherto limited research on grammaticization in Creoles has revealed is that even these vernaculars follow more or less the same processes, with their creators making choices from among the structural options then available in the "lexifier." For instance, as shown in Mufwene (1996b), most Atlantic English Creoles have selected go as the FUTURE marker, owing apparently to the option of expressing FUTURE with be going to (or be gonna) + Infinitive in English. Interestingly, they have done it in different ways. The makers of Jamaican Creole selected a DURATIVE construction of go, viz., a go + Verb, which translates the English be going + to + Verb (there being no clear FINITE/NONFINITE distinction in these Creoles, Mufwene & Dikhoff 1989), whereas those of Gullah grammaticized go by weakening it to [ga] (sometimes reducing it only to the velar consonant [g]) and combining it with the verb stem. Speakers of both Jamaican Creole and Gullah also use an alternative form of this construction that was available in the seventeenth and eighteenth centuries, viz., gwine + Verb, but neither of them uses the alternative gon(na) + Verb, which is attested in present-day nonstandard English vernaculars and AAVE. The reason is so far not obvious. What is especially interesting here too is the fact that these Creoles have chosen constructions that are consistent with the development of the relevant aspects of the rest of their grammars.

Grammaticization in the Development of Creoles Jamaican Creole has remained faithful to the fact that go itself must be used in the PROGRESSIVE, although it has not maintained a constraint often associated with the original English construction, viz., the requirement for the subject to be HUMAN or ANIMATE, hence capable of intentions. Its major change lies in the fact that it is perusing its own pattern of DURATIVE/ PROGRESSIVE construction in this evolution, with the preverbal marker a instead of the suffix -ing. The makers of Gullah saw instead the fact that go conveys the basic meaning of FUTURE, while also ignoring the INTENTIONALITY or HUMAN/ANIMATE SUBJECT presupposition associated with this auxiliary in some English varieties. Applying the phonological principle that weakens the form of an auxiliary verb, they changed its form to [ga]. In this particular case, there is no clue yet whether chance or some still unknown factor(s) in the sociolinguistic ecologies of Jamaica and coastal South Carolina bore on these divergent evolutions. It remains, however, that the makers of both varieties worked with materials that were available to them, following constraints imposed by the rest of their emergent grammars. We must bear in mind that loss of inflections and predicate serialization are among the restructuring processes that produced Creoles as different "systems" from their "lexifiers." In the above cases, the Creoles seem to have evolved without using any influence other than from their "lexifier." One may thus be tempted to ignore the fact that several substrate languages use a verb meaning "go" to express FUTURE and this congruence must have influenced the selection of go as a FUTURE marker. However, contacts in which some of the substrate languages do not express FUTURE with "go" sometimes produced a different strategy when there was enough critical mass among speakers of substrate languages with their particular alternatives. Tok Pisin is especially interesting in this particular case, as it uses bai (< by and by in the "lexifier") + Verb (Romaine 1988), consistent with the fact that several of the Melanesian and Papuan languages express the FUTURE with an adverb rather than with an auxiliary verb or an inflection. This is to say that structures of Atlantic English Creoles could have been quite different, had the substrate languages been typologically different.10 Frajzyngier (1984) also shows convincingly that the "lexifier" has exerted a strong deterministic influence in the development of Creoles. He illustrates this with the development of a complementizer for NON-PURPOSIVE subordinate clauses from the verb "say" in Atlantic English Creoles. He argues convincingly that the explanation cannot be wholly substrate, since Atlantic French Creoles do not use a morpheme from either dire "say" or parler "speak, talk" for the same grammatical function. Indeed, colloquial and nonstandard English discourse avails common usage of say to report speech quotatively as in the following examples (quoted from Chase 1943: 4-5 and Dance 1978: 33) which are repeated from Mufwene (1996b,exx.4&9): (6) a. A man came to the door, says, "Hello, stranger what'IIye have?" b. Jack says, "Well, I'll stop a little while, I reckon."

175

1 76

Language Evolution c. "Well, now, King/' says Jack, "hit [it] looks like you'd be a-needin' somebody with all your land..." d. The old King sort of looked at Jack and how little he was, says, "Well, now, Jack, I have got a little piece of new ground I been tryin' for the longest to get cleared..." e. King says, "Think ye can clear that patch, Jack? (7) "This mine... that's yours... You take this one; that's mine ... that's yours." He [the Deacon] went up there; he say, "Reverend, don't you know something. Judgment Day comin' fast, 'cause the Lord and the Devil down there in the graveyard dividing out souls." Reverend say, "Brother, you know you wrong." "Come on go down there." So they went on down there; these kids were still countin' 'em: "This mine... that's yours." He [one of the boys] say, "Now, but it's two at that gate; let's go get them." They [Minister and Deacon] say, "Ah, naw you don't!" [Makes a gesture to indicate a hasty departure.]

There is no evidence of such extensive usage of dire (least of all purler) in French. Thus, although several, if not most, substrate languages of West Africa use a verb meaning 'say' to report speech quotatively in serial verb constructions, no Atlantic French Creole has developed this particular strategy. This suggests indeed that grammaticization in Creoles, hence the development of their grammars, has depended largely on options available in the "lexifier" itself and that, as explained in Chapter 8, the case of substantial model inputs from the substrate languages such as in Berbice Dutch (Kouwenberg 1994) and Melanesian pidgins (Keesing 1988; Sankoff 1984,1993; Sankoff & Brown 1976) call for special external ecological explanations. An examination of relative clauses in Atlantic English Creoles may shed light on how the availability of models in the "lexifier" and substrate influence balanced their acts, bearing in mind that the learners' determination to speak the "lexifier" (despite their concurrent restructuring of its "system") already weighs things in its favor in the vast majority of cases. As shown in Chapter 8, both French and English Creoles have selected relative clause options available in their nonstandard "lexifiers," with the English Creoles selecting the prepositionstranding option and the French Creoles selecting a construction akin to pied-piping. In the case of basilectal English Creoles, genesis in nonstandard vernaculars is more evident in the fact that generally no relativizers other than the invariant weh [ws] (< English what/where11), which acts more like a complementizer than like a relative pronoun (Mufwene 1986c), were selected to start relative clauses. The following examples are from Gullah: (8) a. Da man (weh) Uh tell you come here laas week The man who/that I told you came here last week

Grammaticization in the Development of Creoles b. Every word (weh) Pa say Everything that/which Pa said. c. Da man (weh) I meet he son laas week The man whose son I met last week d. Da manj (weh) everybody taller than him. The man that everybody is taller than.

Besides, weh alternates with the null complementizer and shows exclusive predilection for preposition-stranding, the dominant, if not the exclusive, pattern in colloquial and nonstandard English. There are no cases of pied-piping in Atlantic English Creoles. One may conclude again these are all cases of inheritance from their "lexifiers," with or without congruent influence from the substrate languages. This observation is largely true and is indeed confirmed by the fact that Atlantic French Creoles exhibit no preposition-stranding. The reason is simply that no particular French vernacular other than Quebecois, which developed concurrently with French Creoles but in an ecology of contact with English, allows preposition-stranding. Interestingly, however, French Creoles do not pied-pipe prepositions, either.12 Yet, pied-piping must have been widespread in colonial nonstandard French. It is the only option French allows in relative and interrogative clauses in which the object of a preposition must be fronted. To date, relative clauses in some dialects still start with the complementizer quey even when the relative NP is the subject. In Haitian Creole, relative clauses start with ki when the relative NP is a subject. However, when it is an object, including that of a preposition, the relativizer ki (< French qui), then yields to the null complementizer, as shown in the following examples provided by Michel DeGraff (p.c., August 2001): (9) a. Men pwofese ki te ekri liv sa a Here professor WH ANTER write book this DET Here's the professor who wrote this book b. Men moun^ mwen te al nan sinema ak //. a Here person 1sc ANTER go LOCATIVE cinema with SSGJ DET Here's the person I went to the movie with c. Men kaye. mwen te ekri ladan //. a.13 Here notebook 1sc ANTER write inside 3sG. DET Here's the notebook I wrote in

Substrate languages appear to have influenced these particular structural selections. Although some Bantu languages, such as Swahili, have relative pronouns, which agree in noun class with the antecedent head noun, the vast majority of West African languages start their relative clauses with a complementizer and use a resumptive pronoun where the object of a preposition has been relativized. This trend is attested in several of these Creoles.

177

178

Language Evolution The conclusion drawn in Mufwene (2001a, Chapter 2) is that a lot of "recycling," in the sense of using in novel ways or contexts materials and/or principles available to speakers, contribute to language evolution. The FEATURE POOL concept makes it possible to make language boundaries more osmotic than linguists have made them to be, at least as porous as they are to speakers who easily draw "materiaux de construction" (as Chaudenson 1992, 2001,2003 calls them) from all the resources available to them. There is no particular reason why we should expect grammaticization trajectories to be unilinear, limited to influence from only one language, when in reality language contact is a major and common factor in language evolution, as explained in Chapter 3. Just because other studies of language evolution and grammaticization have ignored this factor does not entail that we must dismiss what Creoles are bringing to our attention. Au contraire, this may be an opportunity to reexamine what role contact has played, for example, in the development of the auxiliary verbs and article and gender distinctions in the Romance languages.

9.5

Conclusions

It should be useful to conclude this discussion with a few last comments on Bruyn's invocation of "universal grammaticalization developments" as an argument against identifying some restructuring processes as part of grammaticization in Creoles. If one approaches grammaticization primarily from a typological perspective, comparing strategies used by various languages to express specific grammatical functions (e.g., Bybee et al. 1994, Heine 1993, Heine & Kuteva 2002, Hagege 2001), it is not really clear that one can identify particular "universal grammaticalization developments" (other than tendencies) which do not amount to basic language universals. Indeed languages can be organized into types relative to their grammatical strategies, for instance those that use "go," or "come," or "want" to express FUTURE. Since Creoles are natural languages, there is no particular reason why they should be excluded a priori from such considerations, independent of the fact that there is nothing that would make grammaticization particularly interesting in this case. They can certainly fit in some of these types. At the synchronic level, which has retained Bruyn's attention, true universals of language, those properties that are shared by all, have received less attention than typological variation. Even research in theoretical linguistics that focuses on properties of UG has had to turn to parametric variation. Grammaticization has been of particular interest because it subsumes a subset of diachronic processes that account for how some grammars express some functions in the specific ways they do and how the emergent typology may shed light on the way the linguistic mind guides structural exaptations to meet the varying communicative needs of speakers. (Titles such as Heine 2001, Cognitive Foundations of Grammar, underscore the significance of "universal" processes that guide grammaticization, a phenomenon also highlighted by Hagege 2001.) In this regard, Creoles as some of the most obvious recent evolutions from

Grammaticization in the Development of Creoles their lexifiers are especially interesting because they highlight the inventiveness of their makers, who have "recycled" what they found in the "lexifiers" that express grammatical meanings significant to them. This is what makes Heine (in progress) particularly interesting. In any case, even from the synchronic perspective in which Bruyn and Flag prefer to frame their discussions, there is a certain amount of interesting variation in the ways that languages express particular grammatical functions. For instance, languages that do not use inflections to express FUTURE vary depending on whether they use a verb meaning "want," "go," or "come," or any other marker. A typological classification of languages according to which particular strategy they use and how it evolved is informative. Since one cannot rule out a priori whether or not contact played a role in the evolution of the relevant languages, there is no particular reason why similar evolutions in Creoles should be ruled out by fiat as "apparent grammaticalization[s]" simply because the model was available in the "lexifier" or in (any of) the substrate languages. One would have to determine beforehand whether the rules that must be posited to account for such exaptations in the development of Creoles are different from those that have been, or must be, posited for those adaptations attributed to language-internal evolution in non-creole languages. If there are no (significant) differences in the restructuring formulae, then one must consider whether there are (significant) differences in conditioning ecological factors that justify treating similar processes in Creoles as only "apparent grammaticalization[s]." So far there seems to be no compelling evidence for treating as different restructuring processes the development of grammatical strategies in Creoles from forms and constructions available in their "lexifiers" or the substrate languages. That would preclude discussing the development of a classic topic such as the "passe compose" in French (avoir "have" + Past Participle or etre "be" + Past Participle) as an instance of grammaticization, since it has evolved from Vulgar Latin, under the kinds of ecological conditions that Bruyn and Flag would like to preclude as making the development of Creoles different. One would have to construct a different history of the evolution of French to deny the role of contact and language shift. These are not arguments in favor of invoking "creolization" in the evolution of French. This would be just the opposite of one of the main theses of Chapter 3, viz., we could do without this disfranchising term in order to focus more on those restructuring processes which Creoles share with the seemingly contact-based evolution of all other languages.

1 79

10

Multilingualism, "Creolization," and Indigenization1 Chapter Outline 10.1 10.2 10.3 10.4

Introduction Language "transmission" and "acquisition" The development of Creole and indigenized varieties Language shift, socioeconomic status, and language diversification 10.5 Language shift and loss in creole-speaking communities and elsewhere 10.6 Some clarifications 10.7 Conclusions

10.1

180 182 184 193 196 199 200

Introduction

My primary goal in this Chapter is to dispel some myths about the development of varieties identified as creolized and indigenized. I need not go over notions such as CREOLIZATION and INDIGENIZATION, as these have been reviewed in Chapter 3. It was shown that they are misnomers that reflect negative colonial biases toward non-European populations and their ways of speaking languages whose genetic classification should for all intents and purposes be Indo-European (Germanic or Romance). The notions have no solid grounding in how languages evolve and speciate. We should dispense with the biases if we care to learn from the development of these particular varieties some important lessons about language evolution in general. I focus below on SOCIETAL MULTILINGUALISM and say very little of substance about INDIVIDUAL MULTILINGUALISM. (Readers interested in the distinction should check references such as Romaine 2000.) Nonetheless, individual multilingualism will always be assumed, because, as explained in Mufwene (200la, 2005a), the real locus of language contact is the individual speakers (Weinreich 1953), the makers and hosts of a communal language qua species.

"Creolization" and Indigenization The main reason for this focus on societal multilingualism is simply the fact that, as pointed out in Chapter 3, histories of languages as accounts of their evolutions have focused on communal languages. They have disregarded I-languages (or idiolects), although the immediate causes of changes that cumulate to produce evolution (at the communal level) lie nowhere but in the communicative acts of individual speakers. These behave linguistically not as members of an organized team with a common goal of winning a game but rather as interconnected individuals focused on solving their respective instantaneous communicative problems in particular communicative acts of a moment, without any specific plans for future communicative acts. This observation does, of course, not discount the fact that their memories help them draw from past solutions to solve new instantaneous challenges. The observation also underscores the significance of the "invisible hand" in driving evolution. We will always wonder how come separate spontaneous innovations or deviations of individual speakers behaving without foresight, at different places and times, so often converge toward a communal norm. Social interconnectedness, which facilitates mutual influence on, or copying from, each other is indeed what has made speakers produce similar idiolects, which converge into varieties that non-linguists have typically characterized as the particular way a given population speaks. This can be inferred from names of languages in, or associated with, particular communities, e.g., Kiswahili as the language of Waswahili "Swahili people" and Japanese as the language of the Japanese people. Through imitations of each other, speakers who interact regularly wind up speaking alike, though not necessarily in identical ways (see below). It is also from the same communal perspective that diachronic processes discussed as CREOLIZATION and INDIGENIZATION have been identified as evolutionary. Populations who appropriated European languages in particular ecologies in which they were isolated from, and interacted little with, the original native speakers developed new ways of speaking that also reflect influence from their substrate languages. Consequently, I discuss MULTILINGUALISM here from the point of view of the coexistence of this or more languages within the same community in which members of the relevant ethnolinguistic groups interact with each other (as individuals rather than as team members) and resolve their instantaneous communicative challenges by using one language or another. According to the ecological approach to language evolution adopted in this book, multilingualism is part of the external ecology of a language to the extent that its structures and vitality are affected by other languages spoken in the community in which it is used. In the particular context of "creolization" and "indigenization" (not as development toward a well-defined state but rather as divergence from a Western-based yardstick), the affected language is the one that was once targeted by speakers of other languages, which influenced how they spoke it and therefore partly determined the direction of its divergence from what they were exposed to.

181

182

Language Evolution

10.2 Language "transmission" and "acquisition" The processes of language "transmission" and "acquisition" are central to language evolution. Quite fundamental in this case are the following questions: Does any speaker or group thereof ever pass on an integrated "system" of units and rules to any learner? Is such a transmission process conceivable even in a classroom setting where learners are drilled several grammatical rules and constraints over them to apply in their interpretation and/or production of discourse? What kind of understanding must we develop of the concepts TRANSMISSION and ACQUISITION in order to figure out whether multilingualism makes secondlanguage "acquisition" (by adult speakers) fundamentally different from first-language "acquisition" (by children), and, if so, how? Both the reality of inter-idiolectal variation (well underscored by Paul (1880/1891), Chomsky 2000, and Mufwene 200la - see also Chapter 2) and the fact that a speaker develops competence in either LI or L2 only gradually suggest that no language is ever "transmitted" wholesale from one group to another. No group of speakers passes on a ready-made grammar to a new group of speakers, no individual speaker does to any other speaker. As pointed out by Meillet (1929), Hagege (1993), DeGraff (1999a, 1999b), and Mufwene (2001a), language "acquisition" is a (re)construction process, which indeed proceeds piecemeal. As with the development of any social competence, this (re)construction process - actually the development of one's I-language or idiolect - depends both on the learner's individual skills and on the particular network that he/she has participated in. The network has consequently determined the particular E-language qua collection of utterances to which the learner has been exposed (the primary linguistic data) and from which he/she could infer the grammar that he/she puts into use. For whoever subscribes to a distinction between the structural peculiarities of internally and externally motivated change, it would be difficult to explain why internally motivated change occurs in the first place if language "transmission" proceeded wholesale among native speakers and there were no room for imperfect replication. This is a question that recurs in Weinreich et al. (1968) as the "actuation problem." The acquisition-qua-gradual-(re)construction conception of the development of linguistic competence makes it easier to understand Lass's (1997) observation that there is no perfect replication in language "transmission'Tacquisition." As shown in Chapter 2, it also makes it easier to understand where the partial causes of internally motivated change lie, viz., in every speaker's/learner's inability to replicate others and in their attempts to accommodate each other toward the production of some norm applicable either to their network of communication, to some cluster thereof, or to the overall language community. An implication of situating the answer to the actuation question in the communicative activities of the speakers is that there are no internally motivated language changes. While language

"Creolization" and Indigenization 183 makes communication possible, communication itself is not part of language, although it helps shape it. Accommodations as performance-related behavior reflect influences that speakers exert on each other's idiolect. They are conceivably small changes that can eventually cumulate, through little-understood dynamics of group selection, into norms of the communal language (identifiable with Paul 1880/1891 as "language custom," per Weinreich et alls 1968 translation). They instantiate the level at which language contact really operates, viz., that of idiolects (Paul 1880/1891, Weinreich 1953, Mufwene 200la). Recall that I have in mind interactions of individuals who are just interconnected by their social practices but are not organized as members of a sport team playing for a common goal. It really makes little difference whether the idiolects in contact are native and xenolectal, because the whole game of influences is grounded in efforts that speakers make to communicate successfully. How individual speakers' selections may translate into their community's overall selections (i.e., spread and stabilize within the community in some form of norm) is a function of still obscure dynamics of the interaction of various internal and external ecological factors that bear on competition and selection in the community's feature pool. Only future research will shed light on this facet of competition and selection in language evolution. (For more on this subject matter, see Milroy 1996, 1997; Mufwene 2001a, and Pargman 2002.) The dynamics of group selection amount to the "invisible hand" invoked by Keller (1994) and elaborated in Chapter 4. Our challenge is to identify specific ecological factors that determine whose innovations or deviations, and which particular accommodations, spread to become (parts of) communal norms, and which particular ones do not survive. The same challenge arises as much from the study of monolingual communities as from that of multilingual ones. An important difference to remember in the case of multilingual communities is that xenolectal deviations or innovations are added to the range of variants that already obtain in the monolingual feature pool produced by native speakers only. However, this difference is merely quantitative, not qualitative, and the particular ethnographic ecology in which speakers interact determines whether xenolectal deviations or innovations stand a chance of being accepted and perhaps prevailing in a multilingual community. The challenge for the linguist interested in language evolution in settings where Creoles and indigenized varieties of European languages have emerged is to figure out what particular principles regulate competition and selection during language use and therefore affect its evolution. The proposed approach dispels the myth that only Creoles (and pidgins), and apparently also indigenized varieties, owe their development to imperfect learning. The reality remains that imperfect replication in language appropriation is the norm and accounts for all cases of language evolution. One thing that the proposed approach makes more obvious is the need to better understand how languages coexist in the individual speaker's mind, how competition and selection operate among languages and among their respective structural features. As explained in Chapter 7, COMPETITION is nothing more than the unequal relations that hold between languages and among their features (not always in tandem) when they coexist in the speaker's

184

Language Evolution mind and in a speech community (hence in the feature pool they generate), such that one feels more compelled to speak a particular language rather than an alternative or prefers particular features over others. SELECTION amounts to no more than the cumulation of such choices into a preferred language (variety) and preferred features that a speaker uses in communicating with others. Understanding how competition and selection work thus amounts to finding out under what specific conditions particular features of languages already known by a learner influence the appropriation of structures of the target language (TL), a phenomenon that is facilitated by the fact that linguistic systems are naturally osmotic. It also entails examining under what particular conditions the accumulation of such restructuring processes (natural even in cases where dialects of the same language are in contact) lead to the development of new "systems" (idiolectal or communal) and thus produce evolutions of all kinds (structural and pragmatic). We can thus recast the difference between LI and L2 "acquisition" in terms of whether or not there is another language, previously known by the learner, whose features come to compete with those of the TL. Since there are individuals who grow up bilingual, perhaps it is too biased to present things this way. An interesting question is therefore also the following: If two languages are appropriated concurrently as native "systems," to what extent do their respective "systems" influence each other? In the context of the present Chapter, this question is secondary, because it is not clear how extensive child bilingualism was in the communities where Creoles developed, nor that children had a particular role to play in the introduction, rather than the spread, of xenolectal features. If the development of Creoles is correctly associated with an ethnographic process of language shift, adult L2 speakers are important (though not the exclusive) agents of the divergence of these vernaculars from the European colonial koines from which they emerged. They not only participated in the relevant language shifts but also introduced xenolectal features into their respective idiolects of the emergent vernaculars. On the other hand, indigenized varieties are byproducts of the alternations that the colonized populations make between their indigenous vernaculars and the European languages they have appropriated as lingua francas. As well explained by Kachru (2005), Kachru & Nelson (2006), and Shneider (2007), no shift has been involved in this case, at least not a generalized one among the speakers of these varieties. Nonetheless, both Creoles and indigenized varieties can help us better understand linguistic osmosis and the ways in which xenolectal influence works.

10.3 The development of Creole and indigenized varieties Creoles and indigenized varieties are definitely interesting to investigate because they are largely, though not exclusively, byproducts of L2 appropriation by adults, although one

"Creolization" and Indigenization must always remember differences in the ecological conditions that produced them. For instance, in plantation settlement colonies where Creoles developed, children appear to have "acquired" language in the same way as in any other community with speakers around them providing the E-language from which they selectively worked out their I-languages. They reconstructed the current vernaculars of their plantations with fewer deviations than the adult Bozal slaves around them, generally favoring the lects that were less stigmatized as xenolectal. Needless to add that the stigmatization must have varied with time and from one plantation to another. Eventually, some of the xenolectal features produced by the adult learners became part of the feature pool from which the Creole and non-Creole children developed their idiolects. Creoles' structures would have diverged more significantly from the relevant European languages if it had not been for the role of children as a stabilizing factor. They helped balance adults' higher rates of deviations and innovations with their own higher rates of closer approximations of the local vernaculars. Part of the variation in Creole speech continua is a consequence of this coexistence of (largely) overlapping individual speakers' idiolects produced by the different language "acquisition" outcomes of children and adults. In this particular case, we cannot miss an important lesson from indigenized varieties, whose norms are set by the older speakers. Rather than metropolitan native speakers, local teachers and parents have functioned as model speakers to the younger learners. These circumstances of language appropriation are somewhat reminiscent of the racially segregated plantations on which children targeted the local colonial vernacular as spoken by the people they lived and interacted with, not the varieties spoken by the privileged settlers with whom they hardly had any significant contacts. Their models were mostly older Creole or seasoned slaves and the TL was literally what these adults spoke. This was a continuum of approximations of, and divergences from, the original homestead koine vernacular, which never fully disappeared in the first place (Chaudenson 1992,2001, 2003). What children recreated relatively more faithfully than the adult learners around them was simply what they were exposed to and recognized as non-xenolectal, depending on who they interacted with, as in any social setting. They were not in any privileged position to develop a Creole for the growing slave community. They never had parents who were deprived of a vernacular or who relied exclusively on some pidgin qua reduced language of minimal communication in a setting of sporadic contact. What we cannot overlook, however, is the high degree of xenolectal features to which the children must have been exposed, which, as explained above, made more complex (than in "non-contact" communities) the way in which competition and selection proceeded in the development of Creoles as communal "systems." Some xenolectal features eventually became locally normal and could naturally be selected into the children's idiolects. The selection of these by the children would make them less stigmatized and more likely to be selected by other child learners, subject to the vicissitudes of the "invisible hand."

185

186

Language Evolution We must also recall that Creoles developed through naturalistic language appropriation, with no language classes involved in the process, despite frequent invocations of the "seasoning" of Bozal slaves by Creole and "seasoned" (i.e., more acclimated) ones in the process. Given the multitude of idiolectal, native, and xenolectal varieties of the local colonial vernacular involved, there could not be a more propitious setting for testing competition and selection in the development of idiolectal competence and of a new communal system. However, unlike "creolization," "indigenization" starts with relatively controlled "transmission" of a preselected "system" (a scholastic, non-vernacular, primarily written variety of the European language), although the involvement of nonnative speakers as "transmitters" of the language affects the interpretation of the TL itself. On the other hand, as in the development of Creoles, restructuring away from the original model is accelerated by uses to which (imperfect) internalizations of the TL are put when the learners interact not only with more experienced speakers but also among themselves. In the same vein, another important difference between "creolization" and "indigenization" must be articulated, viz., in the case of indigenized varieties, speakers are putting into use a model that was "acquired" primarily from a written medium. The process amounts to naturalizing an artificial means of communication. We may also want to think over the question of whether the fact that indigenized varieties typically function as lingua francas bears on how they have been restructured away from the original scholastic models. In any case, from the point of view of language evolution, hence of speciation, Creole and indigenized varieties had different starting points. In measuring the extent of their divergence from the relevant European language, we must remember to compare them each with the right starting point, unlike the poor job done in the literature that has compared Creoles typically with the standard varieties of the corresponding European languages instead of comparing them with the nonstandard colonial koines from which they have diverged gradually towards their basilects. Curiously, contrary to the received doctrine, there is no evidence that a Creole ever developed on Barbados or Martinique during the earlier colonial period when their colonists cultivated tobacco rather than sugarcane. Likewise, the Virginia colony, which chose to grow tobacco (rather than rice, as in South Carolina, or sugarcane), developed AAVE, which is not considered to be a Creole, rather than a Gullah-like variety. This brings us back to a question addressed in Chapter 3 which is worth revisiting here, viz., why did Creoles evolve in some slave colonies, such as Cape Verde and the Netherlands Antilles which did not develop any (significant) agricultural industry, while no vernaculars recognized as such emerged in large plantation settlement colonies such as Brazil? In the case of Brazil, the main reason seems to lie in a different pattern of colonization and its pace of population growth. Its homestead phase (see below) lasted long, and sugarcane cultivation did not boom until the early seventeenth century, almost 100 years after the colony was founded in 1500 and a little over half a century after the first slaves

"Creolization" and Indigenization were imported in 1538. Although by the eighteenth century populations of (partial) African descent constituted close to two-thirds of the overall population - with 80 percent of them in the northeastern region - the gradual pattern of population growth and freer miscegenation between populations of European and non-European descents (compared to Dutch, English, and French colonies) did not favor the development of varieties that are particularly Africanized. The admitted social and regional variation in Brazilian Portuguese (Naro & Scherre 1993) has to do more with the time of settlement and origins of the European immigrants than with the presence of Africans. More or less the same explanation holds for the scarcity of Spanish Creoles in the Americas and the rest of the world (Chaudenson 1992,2001,2003; pace McWhorter 2000). However, the question arises now of why Portuguese Creoles did emerge on Cape Verde, in Casamance (southern Senegal) and Guinea Bissau, and on the Gulf of Guinea islands. What particular colonization differences obtained between them and the American territories? These are questions worth exploring in the future. For instance, though we know that in the seventeenth century the Gulf of Guinea islands turned largely into slave depots (Le Page 1960 - like the Netherlands Antilles in the eighteenth century, where Papiamentu developed), it is not clear how quickly they shifted from the homestead to the plantation phase, which may account for why their basilects are reported to be more divergent from (nonstandard) Portuguese than the Cape Verdian varieties.2 It would also help to find out how much miscegenation took place here, although, as reported by Le Page, most of the Portuguese settlers left for Brazil the moment sugarcane cultivation became more lucrative there. (Lorenzino 1998 reports that Portugal encouraged race-mixing during the early, pre-seventeenth-century period.) Did Sao Tomense, Angolar, Principense, and Fa d'Ambu (also known as Annobonese) develop after the abolition of the slave trade in the nineteenth century, under conditions similar to those of Suriname in the seventeenth century, or (much) earlier, as claimed by Ferraz (1979: 19) or Post (1995: 191)? Ferraz also reports that in the sixteenth century at least the free Africans spoke Portuguese rather than anything identifiable as pidgin or Creole (17). Lorenzino (1998) observes that the departure of Portuguese planters for Brazil during the seventeenth century and the relative isolation of the islands from Portugal fostered a contact more favorable to the development of Creoles. The post-Emancipation recruitment of contract laborers from Angola and Cape Verde during the second half of the nineteenth century and the first half of the twentieth century (for about 100 years) adds more to the complex history of the islands. A comparison of the above history with that of Suriname is informative. In the second half of the seventeenth century local colonial English, from which Sranan, Saramaccan, and the like largely developed (see below), was retained as the dominant vernacular among the non-Europeans and as the lingua franca between them and the Europeans during the Dutch rule, after the English colonists had left. Note that at the time the English traded Suriname against New York, Suriname was still in its homestead phase, the fur trade being its principal

187

188

Language Evolution economic activity, and the slave population was demographically smaller than the European population, a situation aggravated by the fact that many slaves escaped to become Maroons. The linguistic evidence suggests that, like in Suriname, some local colonial Portuguese remained the vernacular among the Africans on the Gulf of Guinea islands even after the Portuguese developed more interest in Brazil and neglected or abandoned their central African plantations. (The islands did remain their colonies.) Does this explanation also apply to Cape Verde, which also played the role of slave depot during the seventeenth and eighteenth centuries? I have no more to say on this than in Chapter 3, which suggests that the patterns and speed of population growth, combined with that of population replacement, must have played a more important role than the plantation industry, which has received more attention. In any case, pidgins based on European languages are concentrated on the Pacific islands and on the coast of West Africa - where the most common one until the early nineteenth century was Portuguese-based, according to Huber (1999).3 During that time period the Europeans developed no more than trade colonies in both regions and originally communicated with the Natives through interpreters and middlemen (grumetes4) who had been trained in a European language. It is only during the nineteenth century that the Europeans would transform them into exploitation colonies. While Portuguese Pidgin has vanished, some English pidgins became expanded pidgins, corresponding to the progression of interactions from sporadic to more regular ones and to the evolution of their functions from lingua francas to urban vernaculars. This coevolution of structure and function occurred in the urban centers of the exploitation colonies of Africa and in the Pacific sugarcane plantations to which large numbers of contract laborers were imported from neighboring Pacific islands. Whether or not a trade colony developed into a European exploitation colony in the nineteenth century may explain why Chinese Pidgin English which developed in Canton in the late eighteenth and early nineteenth centuries and bequeathed us the very term pidgin has not survived.5 Expanded pidgins are recognized to be structurally as complex as Creole vernaculars (de Feral 1989; Jourdan 1985, 1991). Because adult L2 learners are more likely than children to introduce xenolectal features into a TL, the massive incontrovertible evidence of "substrate" influence in Melanesian pidgins discussed by, among many, Sankoff & Brown (1976), Sankoff (1984,1993), and Keesing (1988) supports the hypothesis submitted earlier by Sankoff (1979) for the significant agency of adults rather than children in their developments. Rather than NATIVIZATION (as acquisition of native speakers), VERNACULARIZATION as usage of a language variety as a vernacular appears to be the explanation for the structural expansion of these pidgins into stable and more complex varieties identified as expanded pidgins. We learn from their case that the agency of Creole populations (to which I return below) was not as critical to the development of Creole vernaculars as traditionally assumed in

"Creolization" and Indigenization 189 creolistics since the nineteenth century. In fact, once we factor in the fact the pidgins too (including the expanded ones) have evolved by divergence away from their "lexifiers" and the closer approximations spoken earlier by the interpreters, the explanation for the differential evolution of Creoles and pidgins boils down to an ecological one. Creoles evolved in ecologies in which they have always functioned as vernaculars and their gradual basilectalization is a function of how it was appropriated by successive generations of speakers/learners exposed to already divergent varieties, but without ever undergoing a period of structural "pulverization" (pace McWhorter 2001). Pidgins evolved in settings of sporadic trade contacts, even during the peak of the slave trade, in which the European language functioned as a lingua franca. These settings favored the kind of structural impoverishment and extensive substrate influence associated with them. The latter was made even more possible when the pidgin, adopted as an urban vernacular, had to complexify structurally without any further input from the original "lexifier." Hawaii, which has figured prominently in the debate on the development of Creoles, was not a typical plantation settlement colony. It did not develop on the model of the Caribbean and Indian Ocean settlement colonies presented in Chapter 3. Colonized in the nineteenth century by Americans, rather than by the English, it was developed partly on the model of plantation exploitation colonies of the Pacific, with an abrupt transition from a whaling/ trade colony into a plantation alternative, without an initial homestead phase during which proletarian European and non-European populations would have been integrated. According to Masuda (1995:276-7), the first non-European laborers, the Chinese, were not brought to the islands until 1853. The Japanese arrived apparently after 1878 (25 years later, during what Roberts (1998: 6) identifies as a "major plantation expansion" phase), and the Koreans and Filipinos apparently after 1900 (a period during which the labor population also increased by birth). One factor is especially significant in this case: Although the plantation laborers came from different parts of Asia, they do not seem to have been ethnically mixed, at least not in the overwhelming, if not Babelic, way that the African slaves from more diverse linguistic backgrounds were in the plantation settlement colonies that developed a century or two earlier around the Atlantic and in the Indian Ocean. Aside from the fact that the different ethnic groups were not brought in during the same period, Asian laborers lived in separate quarters, consistent with their waves of immigration. This was more or less in the same way that Singapore (and perhaps other Strait of Malacca colonies) was peopled, with the different ethnic groups kept in their quarters and their members socializing more among themselves than with outsiders. Overall, pressure to socialize (on a regular basis) with speakers of other languages was less intense on Hawaiian islands than in the New World and Indian Ocean plantations. Thus, at least before 1900 (see below), most Asian laborers in Hawaii did not have to use English as a vernacular, only as a lingua franca, more or less like in Pacific exploitation colonies. An important difference is that Hawaii also became a settlement

1 90

Language Evolution colony and English quickly evolved into the vernacular needed by those who crossed ethnic boundaries. Interestingly Roberts (1998) reports that the structural features associated with Hawaiian Creole (henceforth "creole features") do not seem to have developed until the early twentieth century and apparently off the plantations, where "the majority of foreign-born individuals were [then] living" and "where English speakers were more common" (34). She credits nativization for the development of Hawaiian Creole (35), but she also reports that the relevant "creole features" occurred in the speech of immigrant adults (19). Although the features became much more common in the speech of the locally born, it is debatable whether their role really extended beyond what Mufwene (1996a, 200la) characterizes as "slowing down the basilectalization process." In any case, Roberts also observes that none of the "creole features" she discusses is attested in the pidgin spoken earlier on the plantations, that "what Hawaiians spoke early on was a second-language register of English" (14), and that her early texts "exhibit more stylistic sophistication and structure than" her pidgin texts (33). The historical scenario suggests that Hawaiian Pidgin English and Hawaiian Creole English developed under different ecological conditions of contact and at not such different times, regardless of whether or not their developments were influenced by Pacific pidgin and Atlantic and creole varieties that had developed earlier. (For more information on the evolution of English in Hawaii, see Roberts 2005, to appear.) As explained in Chapter 3, the socioeconomic histories of the Atlantic and Indian Ocean plantation settlement colonies suggest that their Creoles developed neither from pidgin ancestors, nor concurrently with any pidgins spoken by some Africans in the same settings. I am deliberately overlooking pidgins (such as Delaware Pidgin, Chinook Jargon, Mobilian, and Lingua Geral) which developed from the concurrent trade interactions of Europeans and Native Americans, and which are based on indigenous languages and appear to have had no significant bearing on the development of New World Creoles. (However, see Viada & Faraclas 2006 for a contrary position.) The same is also true of the varieties that the French colonists in the Caribbean identified as baragouins (Prudent 1980, Wylie 1995, Chaudenson 1992, 2001, 2003) which seem to be based on French (and other European languages) but did not survive the expansion of settlement colonization in the Americas. The Africans of the initial, homestead phase of these colonies interacted more regularly with the European colonists than with Native Americans outside the settlements.6 African slaves in the Atlantic and Indian Ocean settlement colonies are not reported to have spoken Pidgin (under either the name jargon or baragouin), though they had to go through interlanguages while developing competence in the local colonial vernaculars of the time of their arrival. There are also several reports of these earlier slaves becoming fluent in the European colonial koines of the homestead phase. (See especially Chaudenson 1992, 2001, 2003 for reports of the slaves' language practices in French colonies.) We can safely conclude, along with Alleyne (1971) and Chaudenson (1979ff), that African slaves did not develop their Creoles from antecedent pidgins of any kind.

"Creolization" and Indigenization What the socioeconomic histories of the former plantation settlement colonies suggest about the evolution of European languages from the sixteenth to the nineteenth centuries is particularly relevant to the hypotheses submitted in this Chapter, viz., that Creoles developed according to the same restructuring processes as other languages, diverging gradually from their termini a quo. One must remember that colonization was an economic venture involving cautious investments of large amounts of capital that typically were no more easily available than they would be in today's globalized economic world order. The colonies developed only gradually into large plantation settlement economies, starting on average with a 15- to 50-year homestead phase during which the main industry consisted of farming and trade with the neighboring indigenous populations. (Exceptions to this generalization include Mauritius, Reunion, Cape Verde, and the Gulf of Guinea islands, which had been uninhabited before the Europeans arrived.) Interactions with the indigenous populations were reserved to trade and wars, so that concessions which functioned as homesteads experienced socioeconomic developments of their own, which were more or less independent from the non-coastal lands where the Natives continued their traditional economies or adjusted to the threat of the invaders. On the homesteads, the Europeans were numerically either superior or equal to the nonEuropeans. The small demographic size of the colonies and the greater familiarity of the non-Europeans with these generally tropical physical ecologies produced societies that were generally integrated and in which captive and free residents interacted regularly with each other and even had and raised their Creole children together. Regardless of their status and ethnic affiliations (or political classification of their race), these children all spoke the same vernaculars. During the homestead phase, when the economy was still too weak to shift to the plantation phase, the population grew more by birth than by importation, providing the local European colonial koine some stability and making non-European Creole populations an important proportion of the future "transmitters" of the European language during the plantation phase. There was thus never a period of "break in the transmission" of the European language, as its "transmission" no longer depended on contact with Europeans even after race segregation had become the norm in especially Dutch, French, and English colonies, and the non-Europeans had become the overwhelming majority. With respect to language evolution, there has all along been something to learn from the ongoing "indigenization" of European languages in former exploitation colonies, especially the fact that those who spread the former colonial language, even more fervently after their countries had gained independence, were members of the indigenous populations. The divergence from the original scholastic model of the European language also intensified during that time, as more and more non-Europeans have used it to communicate more among themselves than with the Europeans, notwithstanding the fact that some of the Europeans with whom they have communicated have not been native speakers anyway. Keeping in mind that Creoles actually developed from nonstandard varieties of European languages, the evolutionary paths of both Creole and indigenized varieties were quite similar. In the case of Creoles, the stochastic event was when various estates grew much bigger,

191

192

Language Evolution in ways that produced overwhelming non-European majorities without a common nonEuropean language they could all speak for communication among themselves. It marked the shift to the plantation phase, when the new, non-European majorities were driven by the circumstances of power and adaptation to the local socioeconomic ecologies to adopt the relevant European colonial language as their vernacular. Its "transmitters" were either the Creole slaves or the seasoned ones, who had come a few years earlier and had already been acclimated linguistically and in other cultural ways. As the proportion of nonnative speakers grew, sometimes to the majority within the non-European population (Baker 1993a), there was room for more and more xenolectal influence on the structure of the evolving vernacular and for divergence from the European varieties. It is thus that the relevant Creole vernaculars developed, by gradual restructuring away from the original colonial European koines, without any evidence of break in the "transmission" of the European languages, contrary to the myth fostered in much of the relevant literature. Even in places like Suriname, where the English left about 15 years after founding their colony and took most, but not all, of their slaves with them (Richard Price, p.c., November 2001), there was always a cohort of non-Europeans who knew the local colonial English vernacular and "transmitted" it to the slaves brought later on by the Dutch. The evidence of post-English European influence on that vernacular lies in the extensive Dutch influence on Sranan and Portuguese influence on Saramaccan. Sranan developed on the coast, where the Dutch were the most heavily represented, whereas Saramaccan developed in the interior, where the planters were predominantly Sephardic Jews who had been expelled from Brazil and had spoken Portuguese. These evolutions suggest that accounts of the development of Creoles must always factor in the varieties spoken by the Europeans who interacted with the slaves. A critical factor in the particular case of Surinamese Creoles is that the Dutch let a language variety that they found in place remain the vernacular of the non-Europeans in their colony. Otherwise, no English pidgin seems to have developed there either, because Suriname was still in its homestead phase, with hardly any plantations, when the English traded it against the Dutch North American colony of New Netherland, and the conditions of sporadic contact associated with pidgins never obtained. Hard to discount in this particular case is the possibility that, aside from the Africans as the central agents of change, the Europeans themselves influenced the restructuring of the vernacular left by the English colonists into Surinamese Creoles.7 It is helpful to distinguish pidgins as stable communal varieties associated with minimal and irregular contacts from interlanguages, which are individual and transitional stages during the development of competence in an L2. Vernacular usage of interlanguages among nonnative speakers definitely had a lot to do with the restructuring of European languages into the divergent vernaculars called Creoles. However, nothing in this evolutionary pattern makes these new vernaculars different from the equally new contact-based vernaculars spoken by European-majority populations in former settlement colonies such as Canada,

"Creolization" and Indigenization the United States, Argentina, and Australia.8 Nor are their evolutionary patterns different from what produced, say, the diversification of Vulgar Latin into the Romance languages. In both cases, history suggests that L2 "acquisition" in settings where more and more nonnative speakers became models for new learners was critical to the speciation process. Thus, rather than being an exception to the rule, creole vernaculars, like indigenized varieties, illustrate how language diversification has proceeded in the history of mankind, as languages were taken to different places and were appropriated by new speakers. Differences in the extents of divergence from the earlier and "sister" varieties can be attributed to variation in external and internal ecological factors.

10.4 Language shift, socioeconomic status, and language diversification I have deliberately avoided unnecessarily (over)using the terms slave and slavery above, because I assume that slavery was one of the many reasons in the history of mankind that have brought populations in contact and have created particular socioeconomic conditions under which speakers of different languages have come to interact with each other. The development of creole vernaculars "lexified" by European languages and other varieties to which the term creole has been extended was not exclusively associated with slavery. In places such as Sri Lanka and Macau, there emerged, without the involvement of slavery and out of the seemingly intimate coexistence of Portuguese and Asian populations, some colonial Portuguese vernaculars that have also been called Creoles (Holm 1989, Smith 1995). The same may be said of English vernaculars that developed in Queensland and in the Torres Straits (north of Australia) among non-Europeans who have shifted to English. To be sure, even to those like me who believe that CREOLIZATION is the social act of disfranchising a divergent European colonial language variety appropriated primarily by a nonEuropean population, it is not evident that all these vernaculars to which I have given no, or marginal, attention should be called Creoles. Note that many of these varieties have been branded this label by linguists, not by their speakers, to whom the term creole itself is often unknown. This is especially true of Gullah, for which no particular local name is recorded in its history, and of Caribbean English Creoles, which the locals identify as Patwa, from the French derisive term patois for a rural nonstandard language variety. What is true of all of them is certainly the fact that they reflect particular ecological conditions of language shift concurrent with the divergence of structures of the adopted vernacular, under xenolectal influence, away from its European ancestor or congeners. However, based on the discussion in Chapter 3, how unique is this particular evolution? Other developments around the world have also shown that the language shift and appropriation of a new vernacular with modification, in the way that has been associated

1 93

194

Language Evolution with Creole vernaculars of the Atlantic and the Indian Ocean, are not limited to cases where the relevant populations were underprivileged. For instance, being underprivileged or enslaved had nothing to do with the development of Baba Malay among the Peranakans of the Straits of Malacca (Mufwene 2005a, Ansaldo etal. 2007). Biologically, the Peranakans are partly descendants of the Chinese merchants who developed trade colonies in the Malayspeaking Southeast Asia (particularly Penang, Melaka, and Java). They came as single males, married low-class Malay-speaking women, and developed a Chinese-influenced culture (especially in religion) in which the male children (Baba) could marry in the Malay population but the female children (Nyonya) had to marry Chinese men, until the sexes were more or less equal numerically and then they became a closed community. The Peranakans are, like the Creole populations of the Atlantic and Indian Ocean, locally born Chinese, many of them of mixed ethnic/racial ancestry. They all speak Malay as their vernacular, and until recently those who spoke Chinese had to learn it as an L2 variety, with the TL now shifting from Hokkien and other, minority Chinese varieties to Mandarin. What is more relevant to this Chapter is of course the fact that they developed a particular Malay variety called Baba Malay, after the local name for the male Peranakans, who were involved in trade with the non-Chinese, more indigenous populations. The variety was predetermined by the conditions under which the men came in contact with the latter, viz., market (and trade), also identified as bazaar, hence its common confusion with Bazaar Malay. It is closer to nonstandard Malay spoken by the commoners in the Malay-speaking population and diverges from it by xenolectal influence from primarily Hokkien Chinese (Collins 1984, Pakir 1986, Ansaldo & Matthews 1999). The lesson is quite consistent with that from the socioeconomic histories of the territories where Creole vernaculars have developed, viz, the particular kinds of contacts maintained with speakers of a language and the particular variety spoken by that population have largely determined the kind of variety that developed from the contacts. The status of the learner may not be a primary factor in language shift and diversification, whereas the need to communicate with the population within which one evolves is. Indeed, language shift had a significant role to play in the development of Creoles, unlike in the case of "indigenization." In Atlantic and Indian Ocean plantation settlement colonies, the shift was accelerated by both the linguistic heterogeneity of the places where the Africans were captured and the size of the homesteads, in which integrated Africans, isolated from each other, could hardly speak their native languages. Even if two or more slaves found themselves in the same homestead, the chances that they spoke the same language were very much reduced, in part because they may not have arrived at the same time or from the same place. Even if two or more slaves living in neighboring homesteads (often miles apart) spoke the same language, they did not socialize often enough to maintain it. Like in today's sub-Saharan African cities, where children have preferred the urban vernaculars to their (grandparents' ethnic languages, Creole children may have found it more advantageous to speak the European colonial vernacular than their parents' languages, to the extent that these could still be "transmitted" to them. As discussed in the next section,

"Creolization" and Indigenization the same socioeconomic history suggests that African languages could have been spoken more easily under the segregated conditions of the late plantation phase and during the post-Abolition contract-labor period than during the earlier stages of colonization. Africans were then brought from more ethnolinguistically homogeneous areas, as in the case of the Yorubas in Trinidad, who maintained their language until the first half of the twentieth century (Warner-Lewis 1996). Moreover, the fact that nineteenth-century African contract laborers were not immediately absorbed by the former slaves favored the longer maintenance of the African languages. More or less the same thing happened to the languages of indentured Indian contract laborers (identified as "East Indians" in the Caribbean) in all former plantation colonies in the Indian Ocean, South Africa, the Atlantic, and the Pacific. They too went through a similar gradual assimilation or integration process, although things have been more complex and varied in the latter case. In places like Jamaica, where the proportion of Indians has been very small (about 1.3%, Jamaican 1991 census) and the original contract laborers had few women among them, the integration within the overall non-European and creole-speaking population proceeded fairly rapidly. In places like Trinidad and Guyana, where their numbers came to equal or surpass those of the Creole populations of African descent who preceded them (40% East Indians vs. 41% Africans in Trinidad; over 50% East Indians vs. 43% Blacks or mixed in Guyana), the pressure to mix with other ethnic groups was not greater than in places like Singapore, where most of the immigrants have resided in their respective ethnic quarters and only a few have chosen to mix. While the newcomers have adopted the local Creole as their language of wider communication, they have not necessarily had to give up their ethnic Indian languages, at least not until a couple of decade ago. Yet the structural attrition of the Indian languages under the influence of the local Creole cannot be ignored (Mohan 1990, Mohan & Zador 1986, Bhatia 1988). As in South Africa (see below), the gradual loss of East Indian languages in the Caribbean has largely depended on factors having to do with multilingualism among the East Indians themselves and what advantages the vernacular of the larger, host population, also the language the economy in the predominantly manual labor system, had to offer them. More specifically, according to Mohan & Zador (1986), the majority of the earliest contract laborers came from Eastern India, where Bhojpuri was a major vernacular. The next numerically important group consisted of Hindi-speakers. Structural similarities between Hindi and Bhojpuri, compounded with the sentiment that Bhojpuri was one of the nonstandard varieties of Hindi, led to the prevalence of Bhojpuri in a koineized form as the East Indian vernacular in Trinidad and standard Hindi as their language of literacy. Communicative pressures from the larger, creole-speaking community led Hindi to lose its acrolectal function among East Indians to the local English acrolect and Bhojpuri to the local Creole vernacular. The agency of this demise of East Indian languages is primarily that of children, or the younger generation, who found it less and less useful to use their ancestral language varieties and to speak like the descendants of Africans who preceded them.

195

196

Language Evolution Mesthrie (1991, 1992, 2006) gives an elaborate account of this process of language shift in South Africa, which first also favored a koineized Bhojpuri and then South African Indian English as an ethnic language variety. In Mauritius, which is closer to the motherland and where the Indians have become the majority (two- thirds of the population, according to Chaudenson 2001: 31), Bhojpuri and perhaps some ethnic Indian languages have survived, although Creole has also been claimed as the common national language, in opposition to attempts by Creole populations (largely of African and Malagasy descent) to claim it as their own group language (Chaudenson 2001). It is equally important to highlight the fact that the situation of Indians in South Africa is not the same as in the other former British colonies, because South African Indians have really been the counterparts of Africans in bondage in plantation settlement colonies or of other Asian contract laborers in the Pacific. (Black South Africans are the counterparts of Native Americans, especially in having been dispossessed of their lands and having long been marginalized from the economic developments of the new nations.) South African Indians were ethnolinguistically as diverse as the East Indians of the Caribbean, were more dispersed in Kwa-Zulu Natal than the migrant workers from Southern Africa (e.g., Mozambique), were relatively favored over other marginalized populations in the education system, and therefore produced children who would quickly shift to English and encouraged their parents to learn it too (through adult education). Although they learned other languages, by the 1970s, English had become the dominant language among East Indians, articulating the demise of their ancestral languages. One must remember that in former plantation settlement colonies, the assimilation or integration of non-Europeans since after the homestead phase has applied relative to the Creole population. The particular status and privileges that Creole populations (as opposed to the Bozal slaves) acquired during slavery must have contributed to the attrition of nonEuropean languages if they did not totally die (see next section). The survival, or slower decline, of East Indian languages in the Caribbean is largely due to the fact that they arrived later in the region, to their numerical size which enabled them to quickly build a critical mass relative to the creole-speaking populations that preceded them, and to their slow integration in the culture that the latter had developed. It seems that an understanding of the conditions of language attrition and loss during the development or appropriation of Creoles in former plantation settlement colonies should contribute to a better understanding of the subject matter of language endangerment and extinction in general. I turn to this question below.

10.5 Language shift and loss in creole-speaking communities and elsewhere Because Chapters 11-14 are devoted to this subject matter, I will keep this section very brief, focusing on those aspects of the language loss that will hardly be discussed in them.

"Creolization" and Indigenization 197 In Mufwene (2001a, Chapter 6), I submit that language shift is inversely correlated with socioeconomic integration. I could not then explain to my own satisfaction why (descendants of) Africans in European settlement colonies were among the first to lose their languages while they have been the most conspicuously isolated residenttially, especially during the later stages of those colonies. Segregation, hence fewer interactions with the socioeconomically privileged group(s), favors the maintenance of the languages of the underprivileged groups. This is a plausible explanation for the post-Abolition maintenance of AAVE and Gullah in the United States, despite the fact that they are highly stigmatized. It is also a partial explanation for the maintenance of Makhuwa in South Africa (where it is spoken by Mozambican migrant workers) as opposed to the demise of the ancestral languages of South African Indians (Mesthrie 2006). A simple explanation for the loss of African languages in European settlement colonies is that, like the European indentured servants, the Africans of the homestead phase were under more pressure to suddenly give up their languages, regardless of whether or not they were prohibited to speak them. They hardly formed critical masses that would help them continue to speak them, certainly not at the ratio of one or two domestic slaves per homestead (in which they were socially integrated though discriminated against). As noted above, even if by accident they knew another African speaking the same language in a neighboring homestead, a few miles away, they would not have interacted regularly enough between them (as opposed to Europeans of their respective homesteads) to maintain (fluent command of) their languages. For those who produced Creole offspring, the conditions were not favorable to the "acquisition" of these African languages as vernaculars by the latter, although some of them undoubtedly learned something of their parents' languages. As noted above, the conditions for societal multilingualism must have been different during the plantation phase. Sheer numbers alone increased the probability that some slaves found themselves together with some others who spoke the same African language natively or as a second language. The preference expressed by some planters for slaves from particular African regions must have made it more likely for several slaves in (the second half of) the eighteenth century to have found cohorts speaking the same African language, especially those that had also served as lingua francas on the continent. This would explain why, as reported by Ans (1996), leaders of the Haitian Revolutionary War were able to use African languages as secret codes. Thus, although slaves are reported to have typically been forbidden to speak African languages, it is not obvious how rigorously such a law would have been followed or enforced, especially during activities that Chaudenson (2001) identifies as "communal" (i.e., restricted within the slave population) as opposed to those interactions that he characterizes as "transcommunal" (i.e., involving both Europeans and non-Europeans). Here, one is again easily reminded of colonial and post-colonial sub-Saharan African cities, where, although most of the different ethnic groups have had enough critical mass to maintain their respective languages - especially when most of them have lived in their

198

Language Evolution respective ethnic neighborhoods - they have still generally and gradually lost their languages to the non-European urban vernacular. The relevant groups have owed the loss of their languages to the agency of children who have generally found it more advantageous to communicate either exclusively or mostly in the urban vernacular. Interethnic marriages and the integration of some neighborhoods have also precipitated the process, typically favoring the urban vernacular over the ethnic languages. Newcomers to the city have done their best to speak the urban vernacular, at least as a lingua franca outside home and with relatives who are ignorant of their ethnic languages. Their children have typically selected their vernaculars in the same ways as those born in the city. One can imagine Africans of the plantation phase to have experienced similar pressures for abrupt language shift. The pressures must have been intensified by being mixed with speakers of other African languages that were not intelligible to them. The pressure to shift vernaculars was thus universally experienced, regardless of how likely it was for some of the Africans to meet somebody who knew their language, especially during the (late) eighteenth century, when some plantation owners expressed preference for slaves from particular parts of Africa. Like in sub-Saharan African cities, it was pressure to assimilate to, or be integrated in, slave communities (in which the Creoles and seasoned slaves set the norms of conduct) that led to the loss of African languages. The pressure then came more from within the slave community itself than from the European master. And this pressure was undoubtedly stronger in Atlantic and Indian Ocean settlement colonies than in Hawaii, where the non-European contract laborers came only from four major linguistic groups and lived in quarters that were ethnically segregated. These conditions left room for the development of a pidgin whereas the former Atlantic and Indian Ocean plantation settlement colonies did not. They also favored the maintenance of traditional ethnic boundaries whereas this was impossible in the New World and Indian Ocean colonies. This particular history also sheds more light on the spread of indigenized varieties of European languages, although language shift (in the sense of giving up one's vernacular in favor of another) played no particular role in the process. Before independence, indigenized varieties of European languages were regularly used, especially in written form, primarily by handfuls of colonial auxiliaries, who had been taught these varieties for the specific purpose of serving the local colonial administration. (As accurately noted by Brutt-Griffler 2002a, the exploitation colonial intention was less to educate the colonized populations in general as bilinguals in the European languages than to reduce the cost of administering the colonies by using some Natives in the local administration.) The indigenous colonial auxiliaries served as interpreters between minorities of European administrators and the masses of the colonized populations. Although they represented some intermediate colonial authority, the socioeconomic status and power of these elite increased after independence, as they ascended to positions vacated by the European colonizers and to new ones, using the same colonial language to rule and administer their countries. The new, politically independent

"Creolization" and Indigenization states also prescribed universal teaching of the European official language in all schools, while more and more educated Natives adopted it as a lingua franca and as a social status emblem distinguishing them from the less educated masses. Competition for better positions in the economic system increased the desire by many more non-Europeans to learn these otherwise colonial languages. More extensive usage of them by and among nonnative speakers whose xenolectal features reinforced each other produced more extensive divergence from the original scholastic model. Eventually, the scholastic variety itself would be indigenized, as it has been taught more and more by nonnative speakers (Bamgbose 1992, Bolton 2003, Kachru 2005, Kachru & Nelson 2006, Schneider 2007). INDIGENIZATION and CREOLIZATION undoubtedly differ from each other regarding whether the European language variety to which the non-Europeans were exposed was or was not standard, and whether it was primarily spoken or written, as well as whether the setting of language appropriation was naturalistic or scholastic. Creoles and indigenized varieties are, nonetheless, similar outcomes of the nonnative appropriation of a language by populations which have influenced it with features from languages they had spoken previously. "Creolization" on the plantations of the Atlantic and Indian Ocean can very well be characterized as "Africanization" of the European languages appropriated by the African slaves. We must bear in mind that in places like Mauritius, Trinidad, and Guyana the Indian contract laborers learned and shifted to a local Creole that they found in place, although they have also minimally modified it with their own later contributions, at least by way of new structural variants and, more obviously, lexico-semantic additions. The fact that today's Indian members of these communities have the same phonological peculiarities as the descendants of Africans is consistent with the following observation: Like in any other established community (European countries or today's North American polities) where the immigrant population has increased incrementally, the adult immigrants have died away with their xenolectal features while their children have "acquired" the local vernacular relatively faithfully, regardless of where they were born (Chapter 3). Overall, the more extensively a language has been used almost exclusively by and among nonnative speakers, the more xenolectal elements have influenced the structural direction of its "indigenization" (see also Schneider 2007). Otherwise, divergence is less extensive.

10.6

Some clarifications

Some clarifications are in order now. My position that Creoles of the Atlantic and Indian Ocean can be considered as Africanized varieties of the European languages from which they developed does not presuppose some of the myths propounded in the literature on the development of these vernaculars. For instance, I do not at all assume that Creoles are varieties with vocabularies inherited overwhelmingly from their "lexifiers" and their grammars from the substrate languages. As explained in Chapter 8, even extremist accounts such as the

199

200

Language Evolution relexification hypothesis (e.g., Lefebvre 1998, 2004) have not been able to explain mismatches between Creoles' structures and the alleged models in the relevant non-European languages. Nor could they account for the fact that critical masses of other non-European populations speaking different languages would have equally contributed to the development of the same Creoles, because they were in the same plantation settings during the same critical periods of the developments of these new vernaculars. Overall, hypotheses advocating such unnatural macaronic mixing have partly invoked the absence of a TL, because the "lexifier" was heterogeneous (e.g., Baker 1990,1997; Thomason 1997). They have generally ignored the fact that in any setting of naturalistic language acquisition the TL has been heterogeneous (since, as explained above, idiolects are not identical) and "acquisition" itself is a selective (re)constructive process. Thomason (2001) and Thomason & Kaufman (1988) have also claimed that Creoles cannot be classed genetically. Arguments against this position are presented in Chapter 3. The colonial varieties of European languages spoken primarily by descendants of European colonists are no less contact-based than Creoles, the extent of "substrate" influence being variably determined by the ecological constraints on language shifts, such as the time of the communal shift to the new vernacular, the relative integration of the new speakers by the nativespeaking population, and the pattern of population growth (primarily by birth or by immigration as adults) among the new speakers. Aside from the original restructuring that was induced by new patterns of contacts among native dialects, contacts with other European languages influenced the way selection operated over the variants in competition in the contact feature pool. White American English varieties are different from each other not only because of new contacts among British English dialects in the colonies and because of isolation from the British Isles (Algeo 1991) but also because of the contacts of native English dialects with xenolectal varieties spoken by other European immigrants even in the later stages of their evolution since the nineteenth century. The possibility of the latter's influence on colonial English increased as the European populations became more and more integrated. The divergence of American and British English varieties cannot be fully understood without an understanding of the way in which such xenolectal varieties (e.g., Italian and German Englishes), therefore L2 "acquisition," affected the evolution of English in North America.

10.7

Conclusions

It is hard to conclude this Chapter without generally repeating some of the conclusions of Chapter 3. By way of transition to Part 3,1 focus here on the fact that, in the case of language evolution, it seems inadequate to treat multilingualism as an undifferentiated ecological factor that remains uniform from one contact setting to another. Its effects on the evolution of a language vary depending on how it obtained in a particular speech community and on

"Creolization" and Indigenization the kind of population structure that hosts it. It is as important to tell whether only two languages or more languages were in contact with each other, with what ethnographic functions. There is competition between languages when they have the same ethnographic functions (vernacular or lingua franca) and therefore vie for the same populations of speakers. It is equally important to determine whether or not the new speakers used it to communicate only with the native speakers or also among themselves. The latter situation increases the potential for language to diverge significantly from the native varieties, as the new speakers are more likely to reinforce each others' deviations. It is just as important to determine whether the new speakers are residentially integrated, which increases the possibility of ethnolinguistically different individuals interacting regularly with each other and therefore developing a communal variety that is not marked exclusively by one particular ethnolinguistic group. If they are residentially segregated by ethnolinguistic origins, as was the case on Hawaiian plantations, the common lingua franca is used rather sporadically or just less frequently. The integration/segregation factor affects not only the nature of the emergent variety (pidgin, Creole, or otherwise) but also whether or not the relevant substrate languages are doomed to disappear, as in the plantation settlement colonies of the Atlantic and the Indian Ocean. In the case of North American colonies, how long it took social integration (hence some sort of societal and perhaps individual multilingualism) to replace geographical and/or residential parochialism (thus spatial multilingualism) among Europeans determined how long it took before some immigrants' languages died. It also determined the extent to which the population of new speakers could influence the overall colonial varieties of the European language. Overall, the particular dynamics of population growth and interactions across language boundaries and on the vernacular level, bear on how one language affects the structures of the language that prevails and the vitality of the competing languages. We cannot of course ignore the market values of the languages in competition, especially in regard to their economic usefulness, because they determine which languages are likely to be given up. Most speakers evolve only within their speech communities and socialize within limited networks; therefore, they don't worry about whether an additional language will enable them to communicate with people outside their speech community (pace Swaan 2001, Calvet 2006). The development of Creoles has usually been treated as not following the normal course of language evolution. Reversing the tables, I have discussed multilingualism-related topics in this Chapter to show that a better understanding of the many evolutionary processes involved in the development of Creoles and indigenized varieties can shed light on certain ecological factors that have influenced the evolution of other languages. Comparisons of language evolutionary processes in various colonial contexts, including those of language "indigenization" and urban multilingualism in sub-Saharan Africa, appear to be particularly informative, especially about the concurrence of language shift and language loss.

201

This page intentionally left blank

Part Three

Globalization and Language Vitality

This page intentionally left blank

Language Birth and Death Chapter Outline 11.1 11.2 11.3 11.4 11.5

11.1

Introduction The meanings of BIRTH and DEATH applied to languages Questioning some usual accounts of language birth and death A historical perspective into language birth and death By way of conclusion

205 208 209 213 221

Introduction

Linguistics publications on language endangerment and death have increased since Dorian's (1989) Investigating Obsolescence and more so since the publication of a special issue of Language on the subject matter in 1992. Books alone include the following: Fishman (1991), Robins & Uhlenbeck (1991), Brenzinger (1992, 1998), Hagege (1992, 2000, 2006), Miihlhausler (1996,2003), Cantoni (1997), Dixon (1997), Grenoble & Whaley (1998,2006), Hazael-Massieux (1999), Reyhner et al. (1999), Crystal (2000, 2004), Nettle & Romaine (2000), Skutnabb-Kangas (2000), Hinton & Hale (2001), Maffi (2001), Mufwene (2001a, 2005a), Swaan (2001), Dalby (2002), Harmon (2002), Joseph etal (2003), Maurais & Morris (2003), Phillipson (2003), Batibo (2005), and Ostler (2005). Experts will undoubtedly notice some omissions, but one cannot help noticing the strong interest the subject matter has aroused among linguists over the past two decades. Research and publications on new language varieties have interested linguists in a less dramatic way, despite the high visibility of Bickerton (1981, 1984a). This asymmetry may reflect the concern among linguists - stated in numerous publications - about the increasing loss of materials that should inform us about typological variation. It may also be due to the following: Although genetic linguistics has always been about speciation, researchers have typically focused on whether or not particular language varieties descend from the same ancestor and can thus be claimed to be genetically related. As noted in Chapter 3, research on the development of Creoles, pidgins, and indigenized varieties - which is obviously on the birth of new language varieties - has hardly been connected to genetic linguistics.

11

206

Language Evolution Thus, because of the way contact is thought to have exerted a rather unusual influence on these cases of language divergence, linguists seem to have been more interested in showing how the development of Creoles, pidgins, and the like deviates from what they take to be the "normal," or "usual" kind of language change and speciation than in explaining the process of language birth itself. Overall, the way that scholarship on language loss and birth has developed reflects in some ways the fact that, as explained in Chapter 3, genetic linguistics has assumed scenarios in which language contact has played an incidental, rather than catalytic, role in the process of speciation. While so much work has been done to retrace genetic kinship through sound and lexical correspondences, less has been done to explain how the speciation proceeded, nor what triggered it. Even much less attention has been paid to the fact that language loss was a concomitant of this process, despite mentions of the terms substratum/substrate and super stratum/superstate, which allude to the fact that in many places some languages have replaced others on the model of geological strata. Paying more attention to this aspect of language vitality would of course have entailed discussing those cases where the languages of the subordinate, colonized populations prevailed over those of the colonizers/invaders, such as with Prankish in Gaul, Norman French in England, and Arabic in Iberia. The research orientation would have prompted linguists to reflect over those conditions under which the superstrate or colonial language prevails versus those under which it does not. We could even have thought over those cases where the relation between the coexistent languages is not hierarchical and multilingualism is sustained (such as in many parts of the Third World), as opposed to those under which the populations in contact have evolved toward monolingualism (as with immigrant languages in European settlement colonies since the fifteenth century). Typically, research on language loss has focused primarily on the indigenous languages of European ex-colonies and to some extent on minority languages of the European Union such as Breton, Occitan, Basque, Sami, and Gaelic, which are still endangered by the official and dominant languages of their nations. The almost exclusive association of language death and birth either with the emergence of modern European nation states united by single national languages or with the colonization of most of the world by Europe effectively since the sixteenth century has led to the illusion that both processes might be recent developments in the history of mankind. Rare are perspectives such as in Ostler (2005), and to some extent Hagege (2000), which paint the vitality histories of primarily major languages of the world from the point of view of colonial/imperial expansion. Likewise, the overemphasis on world-wide economic globalization as the primary cause of language loss has prevented fruitful comparisons between, on the one hand, recent and current evolutions and, on the other, what occurred, or must have, during the earliest political and economic hegemonies in the history of mankind. There are, however, some historical studies of globalization such as Cowen (2001) and Osterhammel & Peterson (2005) which offer just this perspective, which is the one adopted below.

Language Birth and Death Unfortunately, as a good review of the literature will reveal, linguists interested in the subject matter of language endangerment have largely based their reflections on a myth of globalization interpreted as Westernization, Americanization, or McDonaldization of the world made more and more uniform largely by the Hollywood movie industry. As noted in Mufwene (2005a), the exclusive usage of the term mondialisation in French linguistics - in contrast with that of the term globalisation by French economists - reflects this particular practice in which linguists have failed to connect two important and related aspects of language evolution. On the one hand is the fact that the Stammbaums of genetic linguistics suggest a continuous process of diversification, therefore an increase in the number of languages. On the other is the fact that loss of diversity and the feared drastic reduction of the number of languages within a century or so are not such recent developments. It has been a concomitant of language speciation, as the prevailing language wins only a Pyrrhic victory: it speciates under the influence of the languages that it replaces (its substrates). Throughout the history of mankind, since the first pastoralists/farmers colonized hunter-gathering populations, language loss and language speciation have been byproducts of colonization and therefore of population movements and language contacts. A disappointing aspect of the recent literature on language endangerment is that virtually no attempt has been made to understand what globalization really means, how it is related to colonization as a form of political and/or economic domination of a population by another, and how it can represent the replacement of one form of diversity by another. Noteworthy is also the fact that the same literature has conflated the European colonization of the Americas and Australia with that of Africa and Asia as if they were of the same kind, yielding identical linguistic consequences. Linguists have failed to distinguish the expansion of European languages as vernaculars in the former from their spread as lingua francas in the latter, which disputes the common claim that English is the "killer language" par excellence driving all languages to extinction the world over. Although current research on language birth and death is well grounded in population contacts, the relevant literature has hardly highlighted the fact that these processes have usually occurred concurrently, under the same or related socioeconomic conditions. For instance, as pointed out in Chapter 10, the birth of Creoles in the plantation settlement colonies of the Atlantic and Indian Ocean is a concomitant of language shift among the African populations who developed them. Likewise, the emergence of American English(es) is a concomitant of both the gradual loss of traditional English varieties brought over from England and, later, of other European languages that came in contact with English in North America.1 Below, I elaborate on the above observations, focusing especially on the concern that scholars such as Nettle & Romaine (2000), Skutnabb-Kangas (2000), and Maffi (2001) have expressed about loss of "biodiversity" applied to the coexistence of languages. I historicize both colonization and economic globalization to show how they are related and provide differential ecologies for language birth and death. I highlight the role that speakers, considered

207

208

Language Evolution individually rather than lumped indiscriminately into populations, play as the agents of these phenomena. I also question the adequacy of terms such as language war, killer language, and linguicide. As in the rest of this book, I treat languages as viral-like species of idiolects rather than as organisms (Chapter 1), assuming that their vitalities depend largely on the socioeconomic activities of their speakers acting individually but often converging in their repeated selections of particular languages at the expense of other competitors. Thus, we need to understand how the "invisible hand" works (Chapter 4) in order to understand how languages die and new ones come into existence. Situating the speakers in their respective socioeconomic ecologies will also enable us to tell which languages are driving which other languages out, showing that the "killer" language is not always a European one and PRESTIGE is not always a relevant factor, least of all LACK OF PRIDE in one's cultural heritage. This Chapter is not an indictment of pleas and efforts by some linguists to revitalize the endangered languages, ironically more in the interest of linguistics rather than speakers, so that we may not lose data that can inform us on language universals and typological diversity. I simply wish to shed more light on the phenomena, so that we may know what linguistics can and cannot do, and why in many cases we must be satisfied with documenting the moribund languages, because there is no way we can (help) revitalize them, as even political institutions cannot control the factors that have weakened their vitality.

11.2 The meanings of BIRTH and DEATH applied to languages The notions of BIRTH and DEATH are used here as they apply to species rather than to organisms. Languages are unlike organisms in the way they are born or die. As well noted by Chaudenson (1992,2001,2003) in the case of Creoles, and Szulmajster-Celnikier (2000) regarding Yiddish, languages as communal phenomena cannot be issued birth or death certificates. The relevant processes are protracted, spanning several generations. The concept of LANGUAGE BIRTH is in fact a misnomer of some sort. The birth involves no pregnancy and delivery stages, and the term refers to a stage (not a point in time!) in a divergence process during which a variety is acknowledged as structurally different from its ancestor and/or other related varieties. For instance, there is no particular point in time that can be associated with the emergence of Creoles as separate vernaculars from the colonial European languages they have evolved from and/or from the other colonial varieties that emerged around the same time they did. Unlike in the case of organisms, but like in the case of species, language birth cannot be predicted; it is determined postfactum, retrospectively. The recognition of separateness is made possible by a cumulative accretion of divergence features relative to an ancestor language, regardless of whether or not contact with other languages is factored in.

Language Birth and Death LANGUAGE DEATH is likewise a protracted change of state.2 Used to describe communitylevel loss of competence in a language, it denotes a process that does not affect all speakers at the same time nor to the same extent. Under one conception of the process, it has to do with the statistical assessment of the maintenance versus loss of competence in a language variety within a population of speakers. Total death is declared when there are no speakers left of a particular language variety who use it to communicate with one another.3 An important question nowadays has also been whether Latin, whose standard variety (Classical Latin) is still the lingua franca of the Vatican and whose vernacular, non-standard variety (Vulgar Latin) has evolved into the Romance languages, is really a dead language (Hagege 2000). If so, what is the most critical criterion in identifying a language as dead? Is language death predicated on the absence of native speakers and on its "transmission" from one generation of speakers to another through the mediacy of the scholastic medium, as in the case of Classical Latin? How does this criterion apply to indigenized varieties of European languages in Africa and Asia, where they are typically "transmitted" through the scholastic medium? Does this position imply that pidgins do not become new languages until they also acquire native speakers while complexifying into expanded pidgins? In the case of the evolution of a language into a new variety (such as that of Vulgar Latin into the Romance languages), what is the relationship between language death and language birth? Can these processes be considered as two facets of the same coin? We return to these questions and their derivatives below.

11.3 Questioning some usual accounts of language birth and death As noted above, the birth of new language varieties has been central to creolistics, to the study of indigenized varieties of European languages, and to historical dialectology. The list of titles is too long to include here and any choice of a representative list would be biased. Consistent with the tradition in genetic linguistics, there has been little interest in the birth process itself, except that in the case of Creoles and indigenized varieties, language contact and the influence of non-European languages on the European targets have been acknowledged as important ecological factors. As in the case of Creoles, the emergence of the new dialects of European languages in former settlement colonies has hardly been correlated with the concurrent erosion and death of other European languages that did not become the official languages of the specific colonies, for instance, French, Dutch, Danish, Norwegian, and German, among a host of others in the United States. Works such as Haugen (1953) and Clyne (2003) - to cite only two chronological extremes - are more on language obsolescence than the emergence of new varieties of the dominant language. The fact that language contact is seldom invoked to

209

210

Language Evolution account for the divergence of these new, colonial dialects - which has implicitly contributed to making the development of Creoles so curious - remains an intriguing matter, reflecting largely the bias discussed in Chapter 6. The topic of their birth itself, which can very well be discussed in relation to that of indigenized varieties of the same languages, has generally been overlooked (see, e.g., Thomason & Kaufman 1988, Thomason 2001, and Winford 2003), although Schneider (2007) is a refreshing change. One important exception to the above bias insofar as the evolution of English is concerned is the growing literature on Hiberno-English, as represented, for example, by Kallen (1997), Filppula et al (2002), Rickey (2005), and Tristram (2006). Putting things in a longer perspective, Thomason (2001) correctly notes that history provides several instances of language death. However, the linguistics literature of the past two decades on language endangerment has done little to cast this aspect of language evolution with such an enriching perspective. As noted above, the growing scholarship on language loss has focused primarily on the recent and ongoing attrition of the indigenous languages of former European colonies. In the vast majority of cases, these languages (especially those of the Americas and Australia) have certainly played a marginal role in the evolution of the European varieties that either have driven them to extinction or are threatening them, as their speakers have been marginalized and participated little in the earlier interactions that shaped the colonial evolutionary trajectories of the structures of the European languages.4 It would thus be unjustified to expect the relevant literature to have related the topic of language death with that of language birth. On the other hand, it is definitely justified to expect similar scholarship about Europe to have related these processes with the experience of several European languages that have become langues minorees. Unfortunately, as noted above, such scholarship is rather scant regarding Europe, especially from the perspective of colonization and of the concurrent population movements that brought about the relevant language contacts. As I show below, capturing these parallel evolutions would have enabled us to better understand why languages have been dying so rapidly since the nineteenth century and how what some populations have experienced recently differs from what some Europeans experienced (much) earlier.5 The literature has generally also invoked globalization to account for the loss or endangerment of several non-European languages. Unfortunately it seldom articulates what GLOBALIZATION means and it often confuses it with the phenomenon identified as McDoNALDIZATION, i.e., the spread of McDonald stores around the world (see, e.g., Nettle & Romaine 2000). It even says nothing about whether globalization is a novel phenomenon and how it is related to colonization. This is quite critical because the related literature on the revitalization of endangered languages seldom refers to the ecology that would be the most favorable to the revitalization process and would therefore guarantee success in restoring communities where the relevant languages would be naturalistically "transmitted" from generation

Language Birth and Death to generation without the mediation of schools. Would commitment on the part of the relevant linguistic communities alone do? Or would any conditions other than the precolonial ones, under which most of these languages had thrived previously, be supportive of the revitalization efforts? The vitality of languages cannot be dissociated from the socioeconomic interests and activities of their speakers. It is not true that Native Americans have been shifting from their native vernaculars to those of the European colonists because they have lost pride in their traditions - not any more than the Celts did in giving up their indigenous languages in favor of Vulgar Latin and, later, of the Romance languages in continental Europe, or in favor of English in the British Isles. The reasons for these shifts cannot be (so) different from those that led numerous Europeans to give up their heritage languages in favor of the dominant ones in the Americas and Australia. Linguists have focused so much on languages as "systems" that in this particular case they have omitted thinking of them as tools produced by humans and constantly being reshaped by their users and adapted to their communicative needs. As much as they reflect the world views of their speakers, the cultures that produce these world views are far from being static; the changes affecting the languages reflect changes undergone by the relevant cultures. In a way, language shift is part of that continual adaptive process, often at the expense of multilingualism when the ambient socioeconomic ecology does not foster this.6 Since, to date, language endangerment and loss have not been equally devastating nor proceeded at the same speed in different parts of the world, it is relevant to ask whether globalization has been uniform. Why are Native American languages more endangered in North America, where English has been so dominant, than in Latin America? Is there any correlation between this discrepancy and the fact that in the Americas Creoles have developed more in former French and English plantation colonies than in Portuguese and Spaniard ones? Are the reasons for all these cases of language loss different from those that led to the loss of African languages, or of the vernaculars of politically non-dominant European populations, in the New World? Like among other populations, language shift among Native Americans seems to be an adaptive response to changing socioeconomic conditions, under which the indigenous languages have been undervalued and driven to the margins of the European-style world order. Native Americans have recognized the economic/market values of the European colonial languages supported by the new, global economies. It also appears that Native Americans have been doing recently what several European immigrants had done before them, shifting from heritage languages with lesser market values to economically dominant ones, such as English in North America and Portuguese in Brazil. This is evident in the ongoing demise of French varieties in the state of Louisiana and the endangerment of French in Quebec in the nineteenth century and first half of the twentieth. Indeed, French

211

212

Language Evolution in Quebec would not have been revitalized without the mobilization of the economic machinery in the process, particularly the requirement that all industries and businesses function (also) in French in this Province. This economics-based explanation is also consistent with why indigenous languages in former exploitation colonies of Africa and Asia have been losing ground not to European colonial languages but to (new) indigenous vernaculars (former lingua francas) that are associated with new urban life, such as Swahili in much of East Africa, Town Bemba in Zambia, Lingala in parts of the Democratic Republic of Congo (DRC) and of the Republic of Congo, Wolof in Senegal, Malay in Indonesia and Malaysia, and Hindi in India. In their respective parts of the world, these urban languages and regional lingua francas (some of which happen to be major ethnic languages) are also the only relevant languages of the slow-growing economies that are accessible to the disfranchised majority populations. As I show below, more often than not, the language varieties that have spread and prevailed at the expense of others are not necessarily those associated with prestige; they are common vernaculars of the working populations, such as Vulgar Latin in the southwestern part of the former Roman Empire and Creoles in former plantation settlement colonies (which, as explained in Chapter 3, have evolved from colonial nonstandard varieties of European languages). Globalization has not affected former exploitation colonies in the same way it has former settlement colonies. The high rate of illiteracy, the scarcity of jobs requiring command of European languages, the fact that other jobs are accessible with command of a lingua franca (which is acquired by oral interaction with speakers of the language), and the fact that the inspiring urban culture is expressed in the same urban varieties have all converged to make these new indigenous languages more realistic goals than the European colonial ones. I return to this topic below. Invoking lack of pride or prestige to account for the loss of minority languages and of the langues minorees fails to explain why the Romance languages have evolved from Vulgar Latin (the nonstandard variety) rather than from Classical Latin; why, where Latin prevailed as the language of scholarship, it was not offset by Ancient Greek, despite the higher prestige of the latter even among the Roman elite; and why Sanskrit is dead, or dying, despite all the prestige it has carried relative to other Indie languages. There are also other ecological factors that we should endeavor to identify, for instance, why part of the Western Roman Empire Romanized but the eastern part did not, despite the extended presence of the Romans in the territory (by another 1,000 years); or why in parts of the same eastern part of the Roman Empire Arabic managed to impose itself as an important language (at the expense of both Greek and Latin) but nothing like this happened in the western part of the Empire, where the Arabs and Moors colonized Iberia from the seventh to the fourteenth century. Answers to such questions, parts of which may now be surmised from Ostler (2005), should help us understand what ecological factors are particularly conducive to language endangerment and death.7

Language Birth and Death Efforts to revitalize some of the endangered languages have devoted a lot of energy to developing writing systems for them and generating written literature, a solution that is ironic in a discipline that has always emphasized that human languages are primarily spoken.8 Noble as they are, most of these endeavors have also confused REVITALIZATION, which promotes usage of a language in its community, with PRESERVATION, which does not do more than preserve texts in (and accounts of) a language more or less as museum artifacts. As a matter of fact, the term language preservation conjures up the lifeless preservation of, for instance, fruits in jars or of brains in bottles in research labs. Note that Classical Latin and Ancient Greek, among others, are cited as dead languages, despite the abundant literature that is available in them. As noted in Chapter 10, this pronouncement is intriguing in the case of Classical Latin, which continues to be spoken today as the lingua franca of the Vatican. On the other hand, it is noteworthy that the absence of a writing system has not led to the extinction of nonstandard varieties of the same European languages that have endangered non-European languages, as stigmatized as they have been for centuries now.9 It is in fact from the contact of many of these varieties among themselves - not of standard varieties that the new Indo-European vernaculars spoken in European settlement colonies of the Americas and Australia have developed, which shows that lack of prestige does not equate with lack of vitality or lack of attraction. Likewise, despite numerous predictions of their imminent death, unwritten nonstandard vernaculars such as Gullah and AAVE have shown a lot of resilience (Mufwene 1994 and Chapter 14).10 Overall, speakers keep up with their neighbors or the other people they interact regularly with, not worrying about whether or not the varieties they use in such interactions are prestigious. It is the dynamics and the frequency of speakers' interactions that influence the vitality of their languages and, in multilingual settings, these factors (among others) determine whether or not a language (variety) will prevail at the expense of others, simply thrive, or die.

11A A historical perspective into

language

birth and death 11.4.1 The recent past The correlation of language death with globalization as an economic network of production and consumption interdependences is only partly correct. The whole world is not equally affected by this phenomenon, at least not to the same extent, especially when it is conceived of at the scale of multinational corporations that run the economies of the most industrialized nations of North America, Western Europe, Japan, Australia, and some city states such as Singapore and Hong Kong. We should also bear in mind that globalization as a network

213

214

Language Evolution of economic interdependences can be very local, what students of multiculturalism call localization (as in the case of many aspects of American industry, such as food production); it often develops into regional organizations (such as the European Union, the North American Free Trade Agreement, and the Association of South East Asian Nations) when partners set up privileged trade and/or production relations. It is mostly at the local level that globalization seems to have exerted the greatest impact on languages, and the impact has been more disastrous to indigenous languages in former settlement than in exploitation colonies. This differential evolution is largely a reflection of the fact that the European colonists sought to create "new Europes" outside their metropoles (Crosby 1986), and inherited from them the ideology of nation states ideally unified by one single language. As the European populations became the majority in their new nations, they gradually also evolved from societal multilingualism among themselves to monolingualism, with one language emerging as dominant for all branches of their governments, in the emergent global industries, in their religious practices, and in their school systems. The prevailing language gradually penetrated the private domains of citizens' lives to the point where it became almost everybody's vernacular. To be sure, the shift from other languages to the dominant colonial vernacular was not sudden, nor did it affect all nationalities at the same time, except for the slaves and indentured servants. The slaves were everywhere the first to lose their ancestral languages, not so much because they were forbidden to speak them or were always put in situations so multilingual that they could not do so, but because of the way the plantation societies developed from earlier homesteads. As explained by Chaudenson (1992, 2001), the latter, farm-size dwellings, in which the slaves were the minority and well integrated in family units, did not favor the retention of African languages. The Creole children not only "acquired" the colonial European languages natively but also adopted them as their vernaculars even if they also learned an African language.11 As some of the homesteads grew into large plantations, in which African-born slaves would gradually become the majorities, the Creole and, later, the seasoned slaves speaking modified varieties of the colonial languages (be they Creole or other nonstandard varieties) became the linguistic models, just like city-born children in Africa have been the models for rural-born children. This ethnographic state of affairs played a central role in favoring language shift, and therefore loss, in the settlement colonies. Language shift and loss among post-Abolition contract laborers from Africa and Asia who gradually assimilated to the Creole communities (Chapter 10) speaks more strongly in support of the above explanation. Their relative ethnolinguistic homogeneity did not prevent the change of vernaculars, as they all felt the pressure to operate in the language that brought them income and enabled them to compete with their Creole models, viz., the Creole vernacular. (For more information, see, e.g., Mohan & Zador 1986 and Bhatia 1988 regarding the Indian contract laborers, Ferreira 1999 regarding the Portuguese, and Warner-Lewis

Language Birth and Death 1996 regarding the Yoruba, all of these cases applying to Trinidad. Mesthrie 1991,2006 provides useful information about language shift among Indians in South Africa.) While the experience of European indentured servants especially in the sixteenth and seventeenth centuries is comparable to that of the slaves, that of the free European populations that spoke languages other than the dominant one was rather different. The process of language shift was rather protracted among the latter. In some cases, it did not come to completion until the twentieth century (see, e.g., Haugen 1953 regarding the Scandinavians and Salmons 2003 regarding the Germans). The main reason is that the Europeans, especially those who did not immigrate as indentured servants, were nationally segregated (see, e.g., Fischer 1989). American cities are nowadays not only still segregated into White and Black neighborhoods but have also inherited from the pre-World War II period names such as Irish, Italian, and German neighborhoods as the legacy of the way European immigrants were also segregated among themselves. The preservation of these national identities and the ability within them to run local business in their ancestral vernaculars (Salmons 2003)12 slowed down the language shift process. These developments suggest that in losing their indigenous languages Native Americans have followed the evolutionary trajectory already taken by the immigrants to their land, who were developing an economic system that made their indigenous one obsolete or maladaptive. Thus, ecological pressures for survival forced Native Americans to adapt to the new world order, which entailed some command of the local dominant European language in order to earn one's living. The language shift proceeded faster where miscegenation with the newcomers was either encouraged or allowed, as in especially southern Brazil. It was otherwise made possible by exodus to the city and other places for jobs. Basically language loss among Native Americans has largely been a concomitant process of their Americanization in the sense of shift from their ancestral socioeconomic systems to those of the European immigrants.13 The fact that fewer Native American languages have vanished in Latin America than in North America suggests that both COLONIZATION (in the sense of settlements in new places, especially by taking land away from the Natives) and GLOBALIZATION (in the sense of the development of European-style economic systems with interdependent components) have not proceeded at the same speed and to the same extent in settlement colonies. The differential evolution seems to reflect variation in the kinds of economies that the European settlers developed or the kind of physical ecological challenges they faced in spreading from the Atlantic coast. For instance, the Amazon forest has been difficult and slow to penetrate, and that is precisely where the highest concentration of Native American languages is to be found today. It is not by accident that deforestation and the immediate impact of this exploitation on the local biotas and on the indigenous populations have awakened our awareness of language endangerment, on the model of species endangerment in macroecology. The deforestation has made it obvious that changes in the habitats and economic activities of the

215

216

Language Evolution indigenous populations bear on the vitality of their languages and cultures as their members adapt to the new lifestyles into which they are forced. More languages have died in North America because changes in the socioeconomic world order have been more advanced and more pervasive in its overall population. More or less the same explanation applies to the extensive loss of indigenous languages in Australia. The less marginalized the Natives are from the local global-economy structure, the more likely they are to lose their heritage. As becomes more obvious in Chapter 13, the above explanation does not quite apply to former exploitation colonies of Africa and Asia, where relatively fewer indigenous languages are threatened and where they are typically endangered not by the European colonial languages but by other indigenous languages. In some cases, the latter have stood in the way of the demographic expansion of their European competitors. For instance, urban Lingala now competes in many domains with French in the DRC, and Swahili with English in Tanzania, although both languages maintain their prestige as world and scholarly languages. In the case of the DRC, there is a vibrant urban vernacular culture, epitomized by its popular music, which has co-opted Lingala and brands anybody lacking competence in it as muvila "villager," regardless of their level of schooling. Unlike in the colonial and early post-independence period, intellectuals who have migrated to Kinshasa, the capital, from regions speaking lingua francas other than Lingala can no longer afford to invoke the higher status of French as an excuse for not speaking Lingala in the street. More and more, fluency in Lingala functions as a marker of indigenity, distinguishing not only the Kinois "permanent residents of Kinshasa" from the non-Kinois but also the Natives from the expatriates. One may remark that, in the above case, the indigenous urban vernacular is claiming prestige with some vengeance, so to speak, despite the limited institutional support the evolution has received from the government, such as in allowing a number of programs in French to be shown on the local TV.14 However, the reality is that local dynamics of social interaction in the capital since the colonial period has fostered a two-tiered cultural structure, viz., one culture, essentially European but adapted to the colonial context for the expatiates, a small minority with disproportionate socioeconomic privileges, and the other for the Natives, with the overwhelming majority marginalized from the colonial European lifestyle. The new, colonial elite fell in between, while the majority developed a new, urban and non-traditional lifestyle little influenced by the European culture. Now, with the legitimization of indigenity since independence, and the erosion of the prestige of the ruling class (due largely to the disintegration of the European-style economy), practice of the European language outside the business environment appears to be negatively marked as pretentious. In many settings, it is disadvantageous to the speaker. There are numerous ecological reasons that account for this differential evolution of European languages in former settlement and exploitation colonies:

Language Birth and Death (1) The European colonizers hardly intended to settle permanently in Africa and Asia, although many of them wound up doing so; thus, they were not interested in implementing a socioeconomic structure that would Europeanize the Natives. (2) Determined to exclude the masses of the populations from sharing the economic profits of the new world order, the exploitation-colony system hardly intended to share its languages with them; only a small elite class of auxiliaries that would serve as intermediaries between the colonizers and the colonized were taught these colonial languages (see, e.g., Brutt-Griffler 2002a, 2002b). (3) Unlike in the earlier cases that produced pidgins (such as in Nigeria, Cameroon, and Papua New Guinea), the European languages were introduced in the exploitation colonies as lingua francas based on scholastic inputs rather than as vernaculars naturalistically "transmitted" outside the school system. This "transmission" strategy placed a de facto cap both on the domains in which the European colonial languages would be used and on the proportion of the Natives that could "acquire them." Thus, although they have indigenized, these world lingua francas have remained "languages of exclusion" (Mazrui & Mazrui 1998), still associated with power of domination and exploitation, and restricted to minorities of national elite classes. (4) However, despite the higher status they gained from Western-style education and the association with the now indigenized varieties of the European colonial languages, most of the elite have not severed their ties with their ancestral traditions and have continued to use their indigenous languages as their vernaculars and/or as necessary lingua francas for communication with their less affluent relatives and with the other members of their ethnic groups. In socioeconomic systems where the majorities have typically been marginalized, this communicative practice has continued to recognize the value of the indigenous languages, associating them especially with the maintenance of traditions. The bottom line is that exploitation colonization introduced a socioeconomic structure that was intended to function concurrently with the traditional ones, which it did not intend to change for the masses of the populations, despite its repeated attempts to legitimize its "mission civilisatrice."

One must remember that while settlement colonization has gradually reduced ethnic identities and languages among populations of the new polities (except between Europeans and non-Europeans, as shown in Chapter 10), exploitation colonization has retained them, thus preserving the function of most indigenous languages as markers of ethnic identity. Only the city, in Africa at least, has come close to reducing them, acting like sugarcane plantations and rice fields of the Atlantic and Indian Ocean settlement colonies. The gradual obliteration of ethnic boundaries, caused in part by interethnic marriages, has been an important factor in the loss of ethnic languages.

11.4.2 A dialogue between the recent and distant pasts Languages have been dying since far back in human history (e.g., Hagege 2000, Mufwene 200la, Thomason 2001, Ostler 2005). Assuming the perspective on population movements presented in Chapter 3, human languages must have been dying since pastoralists and agriculturalists started colonizing hunter-gatherers far back in human history. Although linguists have correctly noted that language death has proceeded at an unprecedented pace

217

218

Language Evolution during the last century, they have still not fully explained why languages die and what or who kills them. As shown above, language death has not proceeded uniformly in different parts of the world either. This may also have been the case in the distant past. It should thus be rewarding to establish a heuristic bridge between the distant and the recent pasts in order to learn what they can tell us about each other. I won't speculate about human prehistory beyond the above remarks. Readers of Ostler (2005) will also find in it a gold mine of information about how often languages have died while new varieties have emerged from the prevailing ones in our documented history. They will also realize that those that have benefited from various "imperial" expansions have not always been the most prestigious ones, such as Aramaic from the Assyrian Empire to the Roman colonization of the eastern Mediterranean area. On the other hand, languages of the economically and militarily dominant groups have not prevailed consistently everywhere either. For instance Arabic has not prevailed in Iberia, despite seven centuries of the colonization of the region, while it has evolved into the vernacular of North Africa and a good deal of the Middle East and into the religious language of the Islamic world. Meanwhile, little is left from the language of the Moors, the Berber-Arab mixed populations from North Africa who carried on much of the Arabic colonization of Iberia. Below I focus briefly on language evolution in the western side of the Roman Empire, especially the Romance countries and the British Isles. To be sure, the Romans do not seem to have colonized Europe and the rest of the Mediterranean world on the model of recent European exploitation or settlement colonies.15 There is also little evidence that they claimed full nationwide geographical spaces as their colonies. According to Polome (1983b), the Romans seem to have taken more interest in developing trade and military centers, a practice that leads him more or less to equate the Romanization of these colonies with urbanization. This interpretation is also what emerges from Janson (2004), according to whom the Empire consisted essentially of a network of towns built and controlled by the Romans. Indeed, from the point of view of military and administrative domination, the Romans created networks of towns interconnected by good road and water transportation infrastructures, all ultimately leading to Rome.16 According to Garnsey & Sailer (1987), the Romans formed alliances with local rulers, whom they coaxed to administer their territories in the Roman style, assisted with their technical expertise (including military), and got to work in the economic interest of Rome. These leaders were taught Classical Latin, their children were sent to Roman schools, they were granted Roman citizenship, and they could compete with the Romans themselves for offices as high as generalship, provincial governorship, and the Roman Senate. Some of them, such as Marcus Ulpius Traianus (born in Spain), Lucius Septimus Severus (born in North Africa), and Marcus Aurelius Antoninus Caesar (born in Gaul) even rose to become emperors.17 Indirect evidence that the Romans did not colonize Europe on the settlement model comes from the fact that they left the Western Empire, or were driven out of it (Janson 2004), in the

Language Birth and Death fifth century. However, the local rulers, who had Romanized had already, adopted Latin as the language of their administration, while Vulgar Latin continued to be spoken among those who had interacted regularly with the legionaries, whose children with indigenous women took advantage of their knowledge of the colonial language to access important offices. Latin would continue to be used by the missionaries and the intellectual elite, but the linguistic evidence suggests that this standard and scholastic variety had little, if anything at all, to do with the development of the Romance languages. What is more significant is that Roman colonies were not fully Latinized in the fifth century. When the Romans left, Celtic languages continued to be spoken by the lower classes (the majority), especially in rural areas (see also Janson 2004). According to Polome (1983b), the non-ruling classes were largely multilingual in a Celtic language, Latin, and sometimes also Greek. This explains to some extent why it would take up to the twentieth century before the Celtic populations of France would become fully francophone. It is not clear whether the migration of the Bretons from England to northern France in the fifth and sixth centuries slowed down the process of language shift at the nationwide population level. In any case, the process was protracted and did not affect all segments of the population either uniformly or simultaneously. The fact that no Romance language developed at all in England confirms two points: First, in the former Roman Empire just as in the former European colonies since the sixteenth century, the dominant language has not spread uniformly within the colonized populations. Second, and indirectly, for the masses of the populations in today's Romance countries, the real shift to Latin as a vernacular took place only after the Romans had left, concurrently with the diffusion of the Roman-style socioeconomic system from the urban centers and Christian missions (Landa 2000). The apparent exceptionality of England is so much the more striking that Latin continued to be used there by the missionaries and the intellectual elite until the eighteenth century (even longer as a ritual language), as well as by the Hanseatic League until the seventeenth century. On the other hand, the protracted development of the Romance languages under the substrate influence of the Celtic languages is correlated with the gradual loss of the latter, as fewer and fewer children found it useful to acquire the Celtic languages (identified pejoratively as patois in France) and learned instead the derivatives of Latin now identified as Romance languages. Today most of the Celtic and other more indigenous languages similar to Basque formerly spoken in the same territories have vanished. If we wish to learn more about language vitality, more specifically how some languages die and some others survive, then it would also help to figure out the particular socioeconomic conditions that have helped these languages survive to date and others such as Breton to have taken so long before losing the competition to French. The "Germanicization," rather than Romanization, of England is more curious if we assume that the Romans ruled their Western Empire in more or less the same way and Latin became the vernacular of the people in the Romance countries only after they had left in the

219

220

Language Evolution fifth century. It is perhaps justified to assume that the Germanic colonizers (Jutes, Angles, and Saxons) who replaced the Romans used their non-indigenous languages in their military and political institutions, while the intellectual elite and the Hanseatic traders continued to use Latin. Although Janson (2004) claims that the insular Celts never fully accepted Latin, note that throughout the Western Empire Latin was contained typically in the urban centers and the missions by the fifth century. As noted above, Latin spread within the masses of the populations and in the rural areas only after the Romans had left. It is at the same time difficult to avoid asking, for the sake of comparison, why Iberia and Gaul continued to Romanize despite the colonization of the former by the Arabs and Moors from the seventh to the fourteenth centuries and that of the latter by the Franks from the fifth to the ninth centuries. I will not address this question and other related ones in detail here. Suffice it to note, however, that it is not quite accurate to argue generally that the new colonizers revered the Roman cultural heritage in the conquered territories. While the explanation is perhaps true for Gaul, it does not seem to be so true for Spain, where the Arabs and Moors did not rule in Latin. Here, hatred of the former colonizers after the Spaniards regained control of their land and the zeal for the spread of Christianity, accompanied by the practice of Inquisition, may have reversed the potential spread of Arabic. So much remains unknown. These differing evolutions underscore the need to distinguish between different colonization styles and the different ways in which the colonizers and colonists interacted with the indigenous populations. It now seems necessary to also distinguish between different styles of settlement colonization. The Franks may have settled in Gaul in a different way from the Jutes, Angles, and Saxons in England, and apparently also from the Arabs and Moors in Iberia. The history of England seems to suggest that the language of the Angles prevailed even among the Germanic invaders and subsequently the Celts were gradually Anglicized. The Franks apparently gave up their Germanic traditions, as they embraced the language and religion of the indigenous rulers, Latin and Catholicism. The Arabs and Moors were eventually driven out and whatever Arabization process may have started appears to have been largely reversed, although Spanish bears influence from Arabic. More questions arise now than I can answer about these differences in colonization styles and their fates. Future scholarship will shed light on the relevant questions. The above cases also show that language birth and death often proceed concurrently, though the balance sheet of gains and losses in almost all these territories seems to have typically been negative for most indigenous languages, in the past as in the present. This conclusion is plausible, especially when one doesn't factor in the fact that the languages that prevailed have evolved into several varieties and we know nothing about their future.18 The histories of England, France, and Iberia also show that the language of the colonizers/ colonists does not always prevail. Like the colonization of Gaul by the Franks, that of England by the Norse, the Danes, and the Norman French did not produce language shifts

Language Birth and Death of any consequence among the "Natives" in the history of this territory, although the Anglicization of the Normans (Liidtke 1995) led to the development of a standard English which is largely influenced by French and to some extent Latin, which functioned as a scholarly language until as late as the early eighteenth century and as the lingua franca of the Hanseatic League until the seventeenth century. No German is spoken as a vernacular in France outside Alsace, and no Arabic is spoken as a vernacular in Spain or Portugal today. And yet, Arabic is today the vernacular of North Africa and most of the Middle East, which was also colonized in the seventh century, while it is primarily a religious language in South and Southeast Asia, as in West and East Africa. Does this mean that the Arabs applied different colonization styles in different parts of the world, thus perhaps applying a Roman kind of exploitation colonization in Iberia, settlement colonization in North Africa and the Middle East, and only some sort of trade colonization in South and East Asia? The linguistic consequences of Arabic colonization are clearly different in these different parts of the world, although the commitment to Islam is just as fervent everywhere.

11.5

By way of conclusion

It is difficult to summarize this Chapter at this point without leaving out a number of other important considerations that are relevant to its subject matter. Since I return to them in the next Chapters, I will be selective in this section, focusing mostly on how we must situate language endangerment and death heuristically in the study of language evolution. As noted in Mufwene (2001a), there are some parallelisms between, on the one hand, language evolution in England since its settlement colonization by the Germanics and, on the other, language evolution in North America since the European colonization. In both cases the invaders came to settle new homes. Oversimplifying things somewhat, note that in England the languages of the Germanics koineized into a new variety called "Old English" (although the name English suggests that the Angles largely prevailed) and its subsequent transformations (Middle and Modern Englishes) gradually displaced the indigenous Celtic languages. In North America, as in other recent settlement colonies, the varieties brought from the major colonial metropoles likewise koineized into new colonial varieties and prevailed over the languages both of other European nations and of the indigenous populations. The Celtic languages have died as gradually as the Native American languages are dying now, regardless of the difference in speed. There also appear to be some similarities between, on the one hand, former European exploitation colonies of Asia and Africa and, on the other, southwestern Europe as a former constellation of Roman town colonies. The most significant of these similarities may lie in the fact that in both cases the language of the former colonizer has (initially) been retained after independence as the language for the ethnographically high functions of their societies and as the lingua franca, if not the vernacular, of the intellectual elite. Thus arises the

221

222

Language Evolution following question: Are the indigenized varieties of European languages likely to displace the indigenous languages in the same way as the "indigenization" of Latin has in the now Romance countries? This is really an open question, as a great deal depends on how the economies of the former exploitation colonies fare and how the masses of the population are engaged in it. An evolutionary path similar to that of the Romance countries seems possible in economically successful, though small, polities such as Singapore, where political leaders have promoted English as the main language for the overall population. On the other hand, one cannot ignore the different language evolution course followed by another successful city state, Hong Kong, where usage of English in the white-collar sector of the economy has had no negative impact on Cantonese. The demographic dominance of the Cantonese in Hong Kong, facilitated by the geographical proximity of Canton, makes the city an endogenous contact setting more favorable to the retention of its major indigenous language. Note that although the Chinese are an overwhelming majority in Singapore, they have spoken several Chinese varieties that are not necessarily mutually intelligible, and they are descendants of colonists who have always been surrounded by Malay-speaking populations in Malaysia and Indonesia, a situation that makes the city an exogenous contact setting more likely to thrive with a colonial language. Moreover, it has a significant Peranakan population whose ancestral language is Malay, rather then Hokkien or any other Chinese variety (Chapter 10), and which shifted to English as a vernacular during the nineteenth century (Ansaldo et al. 2007). Although Malay has been a major vernacular and lingua franca in Singapore, one cannot ignore the extent to which English is spreading as a vernacular, largely thanks to the agency of this Peranakan population, which has played a central role in the Singaporean economy and politics. In any case, theories of evolution are not about predicting the future and only the future will rule on these speculations. For the vast majority of former exploitation colonies, one important factor bearing on the fate of the European languages is the concurrent development of indigenous lingua francas which function also as urban vernaculars of the overwhelming proletarian majorities and are associated with modernity for the masses of the populations. These nations actually share this linguistic peculiarity with former plantation settlement colonies, which were indeed ruled like exploitation colonies after the abolition of slavery and where Creoles, rather than the local acrolectal varieties of the European languages, are also the primary vernaculars of their proletarian majorities. To the extent that Creoles are considered separate languages by linguists, we can generalize that the economies of all these places (Africa, Asia, the Caibbean, and Indian Ocean) have typically functioned in more than one language: the indigenized, local standard variety of the European language (including the acrolect in creole continua) for the white-collar sector of the economy, and an indigenous lingua franca or some variety of Creole (mesolectal or basilectal) for the other sectors of the economy.

Language Birth and Death With the unemployment rate quite high and most of the available jobs limited to the non-white-collar sector, most of the populations have no incentive for speaking the European language, even if they learned it at school. The elite continue to speak any of the indigenous languages or some variety of Creole - as is obvious in Haiti (Dejean 1993) and in Jamaica (Mufwene 2002b) - in order to remain in touch with less fortunate members of their societies. There is thus an ethnographic division of labor that does not necessarily make European languages threats to indigenous ones. However, in the same way that, thanks to the urban lifestyle associated with it, Vulgar Latin was attractive to rural populations of Gaul, Iberia, Italy, Romania, and parts of today's Switzerland, the urban vernaculars are now attractive to rural African populations in particular. Although one can argue that the prestige of urban lifestyle is having a negative impact on rural lifestyle, the notion of prestige itself, which has often been invoked as an important factor in language attrition, needs to be reconsidered in rather complex and relative terms. One would have otherwise expected European colonial languages to have given a fatal blow to the relevant urban vernaculars, as they are the ones used in the economically most powerful jobs associated with the elite. Although there are undoubtedly several African and Asian languages that are endangered, it is not evident that the erosion of these languages is proceeding as fast in Africa and Asia as in Europe and its former settlement colonies. In the next two Chapters, I will show that predictions of the imminent extinction of non-European languages around the world need to be framed in a perspective that reflects the complex and non-uniform nature of the competition and selection now taking place among the world's languages. We must also reassess the adequacy of terms such as language war, linguicide, and killer language in our academic discourse. They seem to be misnomers for states of affairs in which languages have no agency at all. Their authors have also overlooked the agency of speakers as those who actually select or give up particular languages (although they are not necessarily aware of their acts), allowing some to thrive and dooming others to extinction, more or less like animals that would refuse to mate and procreate with their conspecifics. Languages do not engage in wars either, though they coexist in competition, like biological species sharing an econiche and vying for the same resources. The species that has the less access to the resources is endangered. Languages can be favored or disfavored by particular speakers depending on whether they are found advantageous or disadvantageous. We must also note that languages are more endangered when the populations speaking them interact peacefully with each other, although we cannot overlook the statistically minority cases of language loss through genocides and through prolonged expatriate life as refugees or as slaves. Noteworthy also is the fact that populations do typically not engage in the activities which endanger some languages in a concerted way, certainly not like sport teams with

223

224

Language Evolution explicit rules of engagement which anticipate some specific outcomes. In this respect, language loss proceeds like language speciation, gradually and cumulatively, through the individual practices of speakers, from which communal patterns emerge. There is one last thing that we certainly can learn from language endangerment in England and the Romance countries. It has everything to do with colonization and globalization, with the former interpreted as the political and economic domination of a population by another and the latter interpreted as an economic network of production and consumption interdependencies. I discuss this fold of the phenomenon in Chapter 12. As is obvious from Ostler (2005), we should beware of discussing language loss today without reference to several previous experiences of the process in the history of mankind. This certainly had something to do with the fact that only 3 per cent of the world's languages are spoken in Europe today (Mayor & Binde 2001), although it is one of the most densely populated areas in the world. The fact that in places such as the United Kingdom one can no longer easily tell the Germanics and the Celts apart by their phenotypes and/or linguistic features alone should tell us a great deal about the impact of social integration on the vitality of competing languages. However costly language endangerment is to typological research in linguistics, we cannot forget the fact that speakers shift languages as part of their adaptive responses to changing socioeconomic ecologies (see also Pandharipande 2002). Linguists concerned with the rights of languages must ask themselves whether these prevail over the right of speakers to adapt competitively to their new socioeconomic ecologies, including speaking the language they find the most advantageous to them. Advocates of the revitalization of endangered languages must tell us whether the enterprise is possible without restoring the previous socioeconomic ecologies that had sustained them or whether linguists can help create alternatives that are equally advantageous.19 Like cultures, languages are dynamic, complex adaptive "systems" that cannot be considered independent of the adaptive needs of their speakers. In fact they are constantly being reshaped by those who speak them, precisely what the indigenization of European languages illustrates. It's certainly not outrageous to counter the current practice in linguistics by claiming that, like features that are associated with them, languages and cultures at any given point in time are commodities with "market values" - "linguistic capitals" according to Bourdieu (1991) - which are subject to competition and selection. Speakers decide what is useful to them, and they determine history relative to their current needs without any foresight of the ultimate consequences of the present behaviors. Such has been history in population genetics, and such it is among humans, despite our consciousness of it.

Globalization and the Myth of Killer Languages: What's Really Going On? Chapter Outline 12.1 12.2 12.3 12.4 12.5 12.6

12.1

Introduction Globalization and colonization Precursors of modern colonization and globalization Will there be an English-only Europe? The economics of language maintenance vs. shift Conclusions

225 226 234 241 245 250

Introduction

This Chapter is about some misgivings I harbor over the ways that GLOBALIZATION has been invoked to account for language endangerment and loss. I am particularly concerned by lopsided claims that the world is becoming more and more uniform. They have apparently overlooked the new forms of diversity that are concurrent with the apparent homogenization of the world. Equally intriguing is the claim that English is the linguistic vehicle of McDonaldization and dissemination of Hollywood movies around the world, hence of the Americanization of our planet, and ultimately the almighty "killer language" that has been driving all languages to extinction. This is, incidentally, a myth that La Francophonie, the intergovernmental organization of nations that claim to be Francophone, have exploited to develop a "partnership" between French and the indigenous languages of the relevant Third World nations to stand together against the world-wide expansion of English. Very little of the literature refers to the impact of colonization on the vitality of the languages of the colonized territories. When they do, it is as if no such process of population relocation to new territories and/or of the political and economic domination of the indigenous populations of the latter ever took place before the European "Great Explorations" of the late fifteenth century. Only a few exceptions, such as Hagege (2000) and Ostler (2005), show that language endangerment and loss as consequences of language shift have occurred several times in the history of mankind. However, they do this without explaining that neither globalization nor colonization are recent phenomena and that this historiographic approach

12

226

Language Evolution to the subject matter of language endangerment and loss should help us better understand what is going on today or what went on in the past. My intention here is precisely to articulate how this can be accomplished. A special metalanguage has also emerged, in which languages are treated as agents and speakers of the endangered or dead languages only as victims. Thus, it has been acceptable to speak of LANGUAGE WARS (Calvet 1998) and of KILLER LANGUAGES and LINGUICIDE (Price 1984, Nettle & Romaine 2000), as well as to conceive of LANGUAGE CONSTELLATIONS (Swaan 2001, Calvet 2006) within which alliances can be formed to resist the spread of more powerful languages. False concerns have likewise arisen about English endangering continental European languages, because it is now the dominant lingua franca of the European Union (Phillipson 2003, Hagege 2006), the one that most bureaucrats and delegates are likely to speak in settings that do not require ideology-based use of national languages, for instance, in the less formal settings of offices and corridors, outside parliamentary sessions and ministerial meetings (Swaan 2001). The distinction between the VERNACULAR and LINGUA FRANCA functions of languages has completely been overlooked in such cases of language competition, critical though it is to determining whether or not a language is endangered. In the same vein, the rhetoric has been less about the rights of speakers than about the RIGHTS OF LANGUAGES to survive (Skutnabb-Kangas 2000, Crystal 2004) and, in much of the linguistics literature since Krauss (1992) and Hale (1992), about the benefits of LINGUISTIC DIVERSITY to linguistics (especially the extent to which research on language universals and typology is negatively affected by the lost languages). The literature has generally underscored the cultural impoverishment of the affected population - not withstanding issues arising from such a static notion of 'culture' - but little has been said about whether it has all been losses and no gains among the relevant populations that have somehow adapted to their changing socioeconomic ecologies. In this Chapter, I submit that globalization in the way it has been invoked in linguistics, in association with multinational companies and with the process of Americanization, is an aspect, if not a product, of colonization. While this facet of it can be traced as far back in time as the forerunners of the colonization (Cowen 2001, Osterhammel & Petersen 2005), there is another aspect of globalization, which is primarily local but has to do with interconnectedness in a complex system, which bears the most significantly on the vitality of languages.

12.2

Globalization and colonization

Globalization is not synonymous with what is immediately suggested by the French word mondialisation (roughly, "universalization" as covering the whole globe).1 It can be local or regional, having to do with interdependencies in a system or a network (such as in the industrial world). Global economic systems in this local sense are actually the unmarked

Globalization and the Myth of Killer Languages structure in polities such as the United States and Australia, where indigenous languages have massively been replaced by European colonial languages over the past couple of centuries. They have evolved from a European tradition in which numerous vernaculars have been replaced by national, official languages. Globalization at the world-wide level is a function of how faster and more reliable communication, both in terms of telecommunication and transportation of people and goods, has made it possible to establish interdependent industries and resources that are thousands of miles apart (Cowen 2001). It is also a function of how economic colonization by multinational corporations has replaced the political and economic colonization of some territories by specific metropolitan nations, such as in the former British Empire.2 Globalization need not be associated with uniformity either, as many of the diffusions associated with globalization acquire local characters and therefore reflect some cultural hybridization (Pieterse 2003, Tomlinson 1999). The McDonald menu is partly adapted to the local diet, in the same way as Chinese cuisine in restaurants is produced differently in different parts of the world (e.g., Hong Kong vs. France vs. the USA). Like Chinese restaurants, McDonald eateries also serve their customers in the local language and adapt their menus to the local demands, for instance, serving beer and/or veggie burgers in some countries but not in others. Major prestigious hotel chains such as the Hilton and the Hyatt, which lure American businessmen abroad, are typically also decorated in ways that reflect local customs, up to the size of the room, although these are bigger by local standards. And they too operate in the local lingua franca and currency, although special accommodations are made to American guests.3 We should normally examine things beyond the epiphenomena that strike our initial perceptions and, as observed by Tomlinson (1999), we should resist over-generalizing from the illusion of uniformity given by the "globalized [business] spaces" of air terminals and five-star international hotels to the larger "cultural spaces" from which these economic islands stand out. We should also determine the extent to which these manifestations of globalization qua mondialisation are integrated in the local cultural ecologies, instead of assuming a priori that only the local cultural system is affected by the imported culture. And we should think beyond usage of English at a Hilton in Tokyo or in Paris to ask ourselves why this foreign language is being used in such a setting and whether the setting is representative of other domains of language use in the larger society (in this case Tokyo or Japan, or Paris or France). Does English function as vernacular in these places, or is it used only in some settings, by a (small) subset of the local population, as part of the accommodations made to the travelers? Do such accommodations affect domestic communication among the locals? It is likewise informative to determine whether the lingua franca used by airline employees and by custom and immigration officers at airports also functions as a vernacular among themselves. As becomes obvious below, multilingualism does not necessarily endanger those languages that are not associated with economic or political power.

227

228

Language Evolution It is difficult to understand globalization as a form either of partnership or of colonization without making sense of the notion of COLONIZATION itself. Outside population genetics, colonization has typically been associated with the political and economic domination of a territory by another in the interest of the latter. However, the history of mankind suggests that we not overlook the meaning of COLONIZATION as the settlement of a territory by a population originating in another, such as in the development of Diasporas. One can also argue that the more common political and economic domination interpretation of COLONIZATION is a consequence of the resettlement interpretation. Both interpretations bear on the following discussion of the differential effects of globalization. One way or another, colonization started the first time agriculturalists and pastoralists encroached on the territories of hunter-gatherers and they either assimilated these residents or drove them away. The earliest instances of language competition must have started then, although we cannot imagine much of what happens until we think of, for instance, the Bantu populations dispersing into territories formerly inhabited by the Pygmies and the Khoisans in sub-Saharan Africa, gradually driving the indigenous languages to extinction. Linguists have seldom addressed the following relevant questions from the point of view of language evolution: Have the Pygmy and Khoisan languages not contributed to the speciation of the Bantu languages through their substrate influence? What can linguists tell us about the regional balance sheet from the contacts of Bantu, Pygmy, and Khoisan languages in sub-Saharan Africa? That is, how many languages have died and how many new varieties have emerged? From the perspective of language birth and death, have things always evolved in the direction of fewer and fewer languages being spoken?4 Were there (really) more languages spoken on earth when the overall human population was smaller and our ancestors lived in smaller communities? Is there paleontological evidence that supports such an assumption? I argue below that colonization is related to globalization to the extent that both have led to the establishment of asymmetrical economic interdependencies between territories or parts of their industries, to unilateral control of resources, to attempts for uniformity in the modes of production and in the transportation of goods, and to inequities in the ways that costs and benefits from the relevant industries are distributed.5 In both cases, they have increased the sphere of economic interaction. Recent colonial history, that of the European expansion since the fifteenth century, also suggests that at the macro-level we distinguish between three basic economy-driven colonization styles, viz., trade colonization, settlement colonization, and exploitation colonization (Mufwene 200la). It also shows that the latter two generally emerged from, or concurrently with, the first. One way or another, trade colonization did not last forever. The economically or militarily more powerful partner wound up subjugating the weaker or weakened one, thus imposing the terms of trade and/or of the exploitation of the coveted resources. While it lasted, trade colonization had no negative impact on the indigenous languages,

Globalization and the Myth of Killer Languages unless one subscribes to a purist position that is opposed to lexical borrowings and other xenolectal structural influence on their language. The typical balance sheet from these contacts is that trade colonization has often produced new language varieties called "pidgins." Curiously, they are based on European languages in Africa and the Pacific but on indigenous languages in the Americas. Expanded forms of pidgins based on European languages include Nigerian and Cameroon Pidgin Englishes, Babu English in India, Tai Boi in Vietnam (the only documented French pidgin), Tok Pisin and Bislama in, respectively, Papua New Guinea and Vanuatu.6 Trade pidgins based on Native American languages included the Mobilian Jargon in the southeastern USA, Delaware Pidgin in the area from New Jersey and New York to Delaware, Chinook Jargon in the northwestern part of the United States and British Columbia, and Lingua Geral or contact-based Tupi and Tupinamba in Brazil. Overall, contrary to the dominant literature on language loss, trade colonization has enriched the linguistic landscapes of the relevant territories. The vitality of the indigenous languages would not be affected until a trade colony evolved into a settlement or an exploitation one. Whether or not the pidgin died subsequently has varied, depending on the particular dynamics of the subsequent colonial regime. Here is where one notices again that the colonization of Hawaii, discussed in Chapter 10, does not neatly fit in the classic settlement style, while it is not certainly in the exploitation style either. We must consider fitting mixed models into this typology, as with the way in which the Romans colonized the world around the Mediterranean. One important characteristic of European settlement colonization is that the European colonists typically maintained their respective languages as vernaculars, up to when one of them gradually prevailed as the exclusive vernacular of the citizens of the new polity, such as English in the USA and Portuguese in Brazil. An important reason for this evolution is that, unless they were indentured laborers, the colonists lived in nationally segregated communities (rural towns or particular city neighborhoods) often until the mid-twentieth century. Although languages such as Italian and German did not die out until the twentieth century (and pockets of them continue to be spoken in Brazil and Argentina), settlement colonies have generally caused a demographic attrition of languages other than the dominant European languages. This was largely due to the fact that most of the more lucrative industry and the education system associated with it would develop in the language of the politically dominant group and in the city, attracting the younger generation to learn and eventually function only in the language associated with it. Eventually, the same process would affect several indigenous people, being the most advanced in North America and in Australia (based on Nettle & Romaine 2000). As suggested in Section 11.4.1, note 13, there are several reasons for both the endangerment and the loss of indigenous languages. One of these, especially during the colonial period, was the demographic attrition of the Native populations due to low immunity to the ills brought over from the Old World. Another reason is migrations of the Natives who

229

230

Language Evolution escaped enslavement and Old World ills to places where other languages where spoken and it became difficult to maintain their own. Later, those Natives who were driven to reservations also faced multilingualism, as many of them came in contact with other groups speaking different languages (Banner 2005). Some of the languages did not survive the competition. Since the nineteenth century, the gradual exodus of the Natives from the reservations and their integration in the new, mainstream socioeconomic world order, which requires mastery of the relevant European language, has probably been the most important factor. Thus, as with many immigrant populations, urbanization and monolingual schooling in the economically dominant European language has been the most important factor.7 However, things have not evolved the same way in all settlement colonies. They must be subdivided into at least two categories: (1) those whose primary industry consisted of plantations of sugarcane (the dominant kind) or rice, relying on slave labor since the seventeenth century or non-European contract labor since the second quarter of the nineteenth century, and whose majority populations now consist of non-Europeans (e.g., Anglophone and Francophone Caribbean islands and coastal states south of the Caribbean Sea); and (2) those whose industry was diversified and relied marginally or not at all on plantations, whose farming industry used a significant proportion of European indentured servants, and whose majority populations consist of descendants of Europeans (e.g., North America and the South American mainland).8 In both cases, we must understand that in shifting to the European languages of their polities Native Americans typically followed in the footsteps of African slaves and many European immigrants before them. These newcomers to their lands were absorbed in the new economic world order before Native Americans could compete in it. As explained in Chapter 11, the Africans were typically first to give up their languages in part because they were initially integrated (though certainly discriminated against) in the homesteads, where they were minorities. However, another important reason is the gradual way in which the relevant colonies evolved into segregated plantation societies in which the Creole slaves no longer spoke African languages. Yet, it is by them that the Bozal slaves, those coming freshly from Africa, were "seasoned" to the new society. For the Bozal slaves, who were societally multilingual, this acculturation process entailed learning the Creoles' vernacular, especially since they (the Bozals) did not have a common group language to speak among themselves. Under the virtually Babelic conditions of the growing plantations, the slaves' subordinate socioeconomic position made it impossible for one African language to prevail as a lingua franca on any particular plantation, least of all in any particular polity. Undoubtedly because the European colonial languages enabled the slaves also to function in the socioeconomic systems in which they were exploited, the Creole varieties into which their approximations evolved would become their vernaculars. By the time African and East Indian contract laborers were brought, after Emancipation, to replace the former slave laborers on the plantations, African languages and cultural practices were too stigmatized to

Globalization and the Myth of Killer Languages be maintained as vernaculars, even within groups that were relatively homogenous ethnolinguistically, such as the Yorubas of Trinidad (Warner-Lewis 1996).9 As noted above, the Europeans too gave up their heritage languages under the pressure of an emergent dominant national socioeconomic system that functions only in one language. Subscribing to the same ideology of national monolingualism as their counterparts in Europe, most of those in North America, for instance, who did not speak English as a vernacular gave up their heritage languages as they gradually assimilated to the dominant Anglo-Saxon regime in order to feel more integrated or simply to experience less discrimination. New patterns of social interaction must simply have gradually made the ancestral language redundant, especially after everybody else in the (extended) family, or the relatives one interacted the most regularly with, could communicate in the dominant vernacular. The shift seems to have proceeded faster when the immigrants were also stigmatized, or excluded by the nineteenth-century ideology of race from the category of (PURE) WHITE, as in the case of the southern Italians and the Irish (Fenton 2003, Guglielmo & Salerno 2003). Thus, the processes of language endangerment and loss in settlement colonies cannot be understood adequately without showing, as I have attempted here, the extent to which the speakers of the affected languages themselves were involved in the socioeconomic activities that produced these negative results. Languages were not warring with each other. Speakers simply adapted to changing socioeconomic ecologies, with the noticeable distinction that they have responded to changes imposed by forces outside them. The agency of the speakers of the relevant languages is more obvious in the case of the diffusion of especially English and French in former exploitation colonies. As is evident from the literature on indigenized Englishes and lesfranfais africains, those spreading them today are primarily the Natives themselves, who have endorsed the hegemonic status that the European colonizers have conferred to the colonial languages, having associated political power and to some extent also economic development with them, and having promoted them in the name of democracy (equal rights of the citizens to the dominant language of the economy and of power) in their school systems. In former settlement colonies, where the European languages have not faced much resistance and have become popular vernaculars, one can also say that indigenous languages have suffered attrition or death by the "neglect" by their own (would-be) speakers, although, as explained in Chapter 11, it is debatable whether they could resist the shift at all. If something must be refuted, it is the oft-repeated claim that Native Americans and the Australian Aborigines have given up their ancestral language for lack of pride in their heritage (Mufwene 2003b). Their reason for doing so does not appear to be different from what induced the demise of Irish, Italian, Dutch, French, German, and a host of other European languages in the United States, for instance. Moreover, the resilience of several highly stigmatized varieties, such as basilectal AAVE, Gullah (Chapter 14), Appalachian English, and Old Amish English suggests that lack or loss of pride in one's heritage may not have much to do with

231

232

Language Evolution the plight of many languages. Speakers stick loyally to the group identity that such languages assign them. Adaptation to changing socioeconomic ecologies, provided the local population structure allows the relevant groups access to the economic opportunities, is the primary reason for language loss. This occurs through the cumulation of decisions taken individually by the members of the relevant population as they face, individually, the communicative challenges of their socioeconomic environments. Language loss is noticed only retrospectively and language endangerment is noticed when the cumulation of such individual responses has progressed so far that proportionally fewer and fewer people (can) speak their ancestral language. On the other hand, interested linguists must address the question of why the loss of Native American languages has been more extensive in North America than in Latin America. Given the geographical ecology of Brazil, for instance, we can ask whether the size and density of the Amazon forest has not been a physical obstacle to the European expansion and the spread of their languages.10 Equally important among the questions to address is whether variation in the style and speed of economic development is not a relevant factor. Could the following facts also account for the difference in the number of surviving Native American languages: urbanization is more limited in Latin America (despite the huge size of cities such as Sao Paulo, which is Lusophone). Moreover, its capitalist economic system is less developed than its North American counterpart, and larger proportions of its population are still rural and therefore not incorporated in the new socioeconomic world order. As a matter of fact, the socioeconomic structure of many Latin American countries is hardly better off than that of African and Asian countries in the Third World bloc. Note, incidentally, that Native American languages have survived the most in countries such as Paraguay and Bolivia, which are the least industrialized. Also, since relatively more Native American languages have survived in Canada than in the United States, could variation in population density and in the proportions of hospitable land unclaimed by the European immigrants be part of the explanation? After all, Canada is more sparsely populated, while its socioeconomic structure is fundamentally the same as that of the United States. Socioeconomic structure had a lot to do with the endangerment of French in Quebec; targeting this particular structure was central to the success of its government's intervention to revitalize the language. The above remarks lead us to consider again the case of exploitation colonies, in which, to adopt the language of Crosby (1986), the Europeans did not plan to build new Europes or homes but came to work on short terms for particular colonial companies or in the colonial administration, hoping to retire back in the metropole.11 The colonizers were not interested in sharing their vernaculars with the whole indigenous populations (Brutt-GrifQer 2002a, citing from the "Macaulay Minute of 1835" for India). They taught scholastic varieties to an elite class of colonial auxiliaries who, by design, interfaced between them and the Natives. The scholastic varieties of their languages - which have evolved into what are now identified as indigenized Englishes, lesfranfais africains, and the like - have been associated

Globalization and the Myth of Killer Languages with novel and special communicative functions introduced with the colonial regimes. The settings for these include higher education, higher echelons of the administration, and higher strata of the industrial sector, at which the local structure interfaces with the metropolitan and now also with other international structures. The indigenous elite interfaced between, on the one hand, the top level of the European colonial administration and of its economic system and, on the other hand, the lower level, in which almost only the indigenous populations were exclusively involved. The European official language ensured communication with the former group whereas the new indigenous lingua francas now functioning as urban vernaculars permitted communication with the masses of the colonized. The system was thus set up in such a way that the European languages would not directly endanger the indigenous languages, especially since the domains of usage of the European and indigenous languages were different and complementary. The colonial rule generally produced triglossias or, according to Chaudenson (2003), "overlapping diglossias," in which the indigenous lingua francas and urban vernaculars became associated with modernity. They have functioned as the languages of trade and salaried labor. In the city, they are what children acquire natively, often at the expense of the ethnic languages of their parents. They are also the languages that many adults who migrate from the rural areas adopt as their new vernaculars, especially if the city is not ethnically segregated. It is thus the urban vernaculars, not the colonial European languages, which have competed with the ancestral, ethnic languages and endangered them, more so in the city than in the rural areas. Rural populations have largely kept to their traditional, pre-colonial ways. Many of them have learned the lingua francas only minimally for the occasional contacts that they have with outsiders. As pointed out in Chapter 11, one must also note in this context that most of the elite who are fluent in the European languages either do not speak these as vernaculars or have not given up their indigenous languages. They need them to communicate with their relatives who have not been equally successful or lucky with their formal education and/or economically, just as they need to maintain communication with the masses of their populations at the market place and in other domains of their social and political lives who operate only in the indigenous languages. They stick to their ethnic affiliations too, in relation to which the indigenous languages are emblematic. They likewise make sure that their children remain fluent at least in the relevant urban vernacular, as it can be a liability to be (totally) uprooted from one's parents' cultural heritage. Where mixed marriages have made ethnic identification difficult, competence in the urban vernacular guarantees that one is not identified as uprooted from what can be claimed as indigenous. As noted above and explained in Mufwene (200la), exploitation colonization has modified the linguistic landscape of Africa and Asia essentially in introducing European languages for novel communicative domains and causing the emergence and geographical expansion of new, contact-based, regional lingua francas that have become urban vernaculars. While these have endangered traditional, ethnic languages, they are not the only threat.

233

234

Language Evolution Wars along ethnic lines and "refugeeism" are taking their toll too. In the latter case, the fate of the refugees' languages depends largely on whether or not the relevant populations are integrated in the host territory, thus whether their children are under pressure to be assimilated within the local populations and find it unnecessary to speak their parents' ethnic languages. This phenomenon is reminiscent of a factor discussed above that must have played an important role in the loss of Native American languages, viz., relocation in the land of other indigenous populations or in multiethnic reservations. It is certainly misguided to attempt to account for all the above evolutions unidimensionally, as in much of the linguistics literature, simply invoking globalization and/or colonization, without invoking ecological specificities. Population contacts around the world have not been uniform. As much as colonization and globalization are implicated in all these cases of the expansion of some languages at the expense of others, the evolutions are primarily local, depending largely on how the relevant populations in contact interact with each other and respond to local economic pressures. It should be informative to find out how much is new in the recent cases of language endangerment and loss that have got linguists concerned. I focus on this question in the next section.

12.3 Precursors of modern colonization and globalization Since a large part of Chapter 13 is on this history of colonization/globalization and language loss, I focus here on discussions that complement the above sections, focusing on some welldocumented examples from the past that shed light on the present or raise interesting issues. Two interesting cases are the colonization of Western Europe by Rome and that of England by the Germanics as already discussed in Chapter 11. Two similarities deserve attention in the context of this book: (1) the diversification of the prevailing language has been a concomitant of language replacement (i.e., language death); (2) the evolutions have been differential, calling for ecology-specific explanations for both language diversification and language loss. For instance, it is curious why England or the eastern part of the Roman Empire did not Romanize. The greater association of the Eastern Empire with Greek also raises the question of why the Hellenic Empire did not produce a phenomenon similar to Romanization, especially when Greek carried even more prestige than Latin in the Roman Empire. To begin with, as also noted in Chapter 11, it is not clear whether the Romans colonized Europe in the exploitation or the settlement style. Trade colonization would apply only to the initial stages, just like in the recent colonization of Africa, Asia, and the Americas by Europeans. That the regime had an element of exploitation style is suggested by the fact that the Romans left their colonies when the area closer to Rome was threatened by their

Globalization and the Myth of Killer Languages Germanic neighbors in the north. Yet, the fact that elite members of the colonies were not only granted Roman citizenship but also allowed to hold important offices in the Roman government and imperial administration suggests an assimilation policy that is not typical of recent exploitation colonies. That retired generals were given land in the colonies, where they settled, is also evidence of some form of settlement colonization. On the other hand, the Roman Empire was more of a network of colonial sources of wealth, amounting to specific interconnected towns (Polome 1983b), than a continuous mass of foreign land subjected to the economic, political, and military control of a metropole. From this perspective, the phrase "all roads lead to Rome" makes a lot of sense, especially if analogized to today's ramifications of world-wide global economic networks that benefit the powerful members more than the weaker ones. Latin was introduced initially to the indigenous upper class, which the Romans supported militarily, advised to rule their territories in the Roman style, enticed to govern in the interest of Rome, and rewarded with powerful positions in the Roman political system. Thus, Latin became associated with power. Although interactions between the masses of the indigenous, Celtic populations and the Roman legionaries exposed the former to Vulgar Latin, the language itself would not spread among them until after the Romans had left the Empire in the fifth century. It is largely thanks to the retention of Latin by the local rulers as their vernacular, as the language of their administration and of the school system that the indigenous populations saw the powerful market value of their colonial language and gradually shifted to it. Just as English and French have indigenized in the former exploitation colonies of Africa and Asia, so did Latin in Gaul, Iberia, Italy, Romania, and part of Switzerland, as more and more nonnative speakers used it to communicate among themselves. The more it served as a vernacular in the urban environment, the more it became associated with the new lifestyle of the Middle Ages (the counterpart of modernity in today's Africa and Asia), and the more ground the Celtic languages lost, although the whole process would be protracted over centuries. Usage of Latin in all sectors of the economic system appears to have been a critical factor, because it made Latin useful to the lives of the masses of the Celtic populations. The fact that the Romance languages evolved from Vulgar Latin, rather than from Classical Latin, suggests that the school system must have been no more successful than the boarding schools of Ireland and the United States in spreading the colonial language. Informal and naturalistic language "transmission" must have played an important role in the spread of Latin and its Pyrrhic victory over the indigenous languages. We cannot conclude from the above whether European languages will have similar evolutions in Africa and Asia, because they are not being used yet in the lower sectors of the economy, where the masses of the populations can find jobs. Much depends also on whether most of the economies of Africa and Asia will emerge from their present condition of stagnation or disintegration.

235

236

Language Evolution We can, however, speculate on why the Romanization of England was aborted and the country "Germanicized." Old English emerged within two centuries, about the same time Old Romance varieties emerged, and before the colonization of Iberia by the Arabs and Moors (from the seventh to the fourteenth centuries). Subsequent colonization of England by the Danes, Norsemen, and the Normans had no similar linguistic consequences. The question is so much the more interesting because Germanic populations colonized both England and Gaul, apparently in the settlement style, around the same time. Emperor Charlemagne was Prankish, but embraced both the Roman cultural heritage and the Catholic faith, while his counterparts in England did not. Even the Norman rule from the eleventh to the thirteenth centuries did not convert them to the Roman style, while the Intelligentsia evolved in the same style as their continental counterparts, writing in Latin. Was the fact that the insular Celts were weakened by wars among themselves a factor that favored the Germanicization process? A similar question about differential language evolution arises when we note that the most significant evidence of language Romanization in the Eastern Empire is to be found in Romania and in some parts of the Balkans (Friedman 2001), among populations that seem to have emigrated from Romania. The rise of the Ottoman Empire in the fifteenth century is certainly a factor that cannot be overlooked, though it raises two interesting questions: (1) Why did the colonization of Iberia by the Arabs and Moors not stop the spread and vernacularization of Latin among the Natives? (2) Why did Turkish not evolve into a vernacular in southeastern Europe? Yet the Ottoman Empire was more successful than the Arab colonization of Europe in spreading Islam and Arabic as a religious language. While the attitude or disposition of the colonized population to the rulers is undoubtedly an important factor to consider, a more significant one may be the particular way in which the colonizing group ruled the colony and the extent to which they integrated the Natives in the economy that operated in their language. Moreover, as is evident in the case of the Roman Empire, the colonizers may not have applied a uniform linguistic policy during the economic exploitation of the colony. Thus, whether or not European languages will endanger indigenous languages in Africa and Asia is an open question. Right now it is the indigenous lingua francas that do; and their effect has been far less extensive than that of European languages in former settlement colonies. One would not even want to speculate on the feared "lethal" effect of English, identified precociously by Crystal (1997, 2004) as a "global language." After all, it is spoken only by small proportions of national populations in former British exploitation colonies and in most of the countries where its usage has increased over the past few decades, even in Japan and Taiwan, which have maintained extensive economic relations with the USA since the end of World War II. Besides, its strictly lingua franca function is in competition with none of the indigenous vernaculars, especially those with important local economic market values such as Japanese and Putonghua. The fact that a language acquires prestige from

Globalization and the Myth of Killer Languages functioning as a world language does not necessarily situate it in the kind of ecological dynamic that would make it dangerous to indigenous vernaculars. Not all ecologies favor the spread of world languages at the expense of indigenous ones. I return to this question below. Returning to Europe, though it is only part of the Roman Empire which Romanized, it is nonetheless curious that no phenomenon similar to this occurred in the Hellenic Empire that preceded it. Yet, Greek had so much prestige that the cultivated Romans associated themselves with it and allowed it to be used as the lingua franca of their administration and economic transactions in their Eastern Empire (e.g., Martin Goodman 1991).12 How different was the Hellenic colonization style from the Roman? Is there a correlation between lack of linguistic Hellenization and the fact that Greek colonies around the Mediterranean were primarily coastal, restricted to trade centers? Or did Greek endure the same fate as Latin in the Eastern Roman Empire because of the expansion of the Ottoman Empire? What's the significance of the fact that Greek was largely contained in the elite class, without any association with the development of the local industry, in a way that made it unattractive to the masses of the population? After all, it is Vulgar Latin that prevailed, not Classical Latin, the variety learned by the intellectual and political elite. The suggested account of the evolution of the vitality of Greek is not so unusual. For instance, we note in modern times that the association of Sanskrit with high culture has not kept it from dying. The development of Irish English seems to be more the result of the introduction of potato farming and the participation in this new industry of the migrant workers who learned English naturalistically than the consequence of teaching English in boarding schools (Odlin 2003). The term Irish English is associated more with the nonstandard vernacular than with standard Irish English, whose morphosyntax is closer to that of British standard English. The fact that indigenized Englishes have remained primarily lingua francas spoken only by a small elite fraction of the populations of former British exploitation colonies is indeed quite telling. They did not develop through the same kinds of ethnographic transformations that are associated with "native Englishes" of the British Isles, the Americas, and Australia. Indigenized Englishes are more comparable to Latin at the Vatican, with its speakers putting into use knowledge acquired in school. The growing divergence of indigenized Englishes from "native Englishes" is a function of the xenolectal settings in which English is being spoken, settings in which the extent of divergence is a function both of the ethnographic and demographic dominance of the substrate languages - which most speakers acquire natively- and of the level of education of the speakers. All these combined histories suggest the resilience and vitality of colloquial and nonstandard varieties, especially when they are associated with some economic and social advantages to their speakers, compared to the exclusionary connotation of the standard varieties associated with the elite. And we could also speculate whether their exclusionary connotation (Mazrui & Mazrui 1998) is not one of the reasons why the foreign official languages of former European

237

238

Language Evolution exploitation colonies do not constitute a danger to the indigenous vernaculars. It can very well be the reason why Classical Latin had little influence on the development of the Romance languages and why Classical Latin had to yield to educated varieties of European national languages in the domain of scholarship. Less exclusionary languages would thus have more vitality than the more exclusionary ones. Although one may invoke some form of globalization - more, or less, sophisticated, depending on the relevant time in history and the relevant part of the world - a host of other ethnographic ecological factors must be considered to understand the dynamics of language vitality, endangerment, and loss. Note that there is no particular reason why colonization must always be discussed in relation to European expansion. We can redress this common bias, at least slightly, in directing attention to Black Africa, for instance, where the Pygmy and Khoisan populations were colonized, assimilated, and/or displaced by Bantu populations and have lost most of their languages. About 5,000-6,000 years ago, the latter dispersed from their homeland in the Eastern Nigeria/Western Cameroon area. Initially, they dispersed southward along the Atlantic Ocean, eastward on the fringes of the equatorial forest (in the Congo basin), and then southward along the Indian Ocean (Vansina 1990, Newman 1995). From these outlying areas of territories where the hunter-gathering Pygmies and Khoisans preceded them, the invading Bantu, mostly agriculturalists, also penetrated the interior. Becoming majorities, they imposed their economic systems and most of their cultural ways, including their languages, which accounts for the near-total loss of the Pygmies' languages and the ongoing extinction of the Khoisan vernaculars in places such as Namibia, South Africa, Botswana, and Tanzania. Technological superiority is not a prerequisite for a culture, including its language, to replace another, especially in an endogenous setting. Changes in socioeconomic ecologies will suffice, especially when the populations in contact maintain relatively egalitarian relations. This is precisely what Tosco (1998) reports about language loss among the Elmolo and the Yaaku in Kenya. The Elmolo are a very small fishing population on the southeastern shores of Lake Turkana. They spoke a Cushitic language related to Dhaasanac and Arbore near southwest Ethiopia. In the late nineteenth century, the drought, famine, and bovine pleuro-pneumonia which hit the Samburu population, their pastoral neighbors, very hard also intensified neighborly contacts between the two groups. The Elmolo were gradually assimilated by the majority and shifted to their language, Maa. The Yaaku, also a Cushitic population, were bee-hive keepers. They had very little contact with the Maasai before the second half of the nineteenth century. Absorption within the majority population, whose language, Maa, functioned also as a lingua franca with huntergatherers of the region, and exogamy led to the loss of the Cushitic language. During the same drought and famine period, it is the pastoralists who depended on the hunter-gatherers for their survival, but it is the latter, demographically much smaller, who gave up their languages.

Globalization and the Myth of Killer Languages The situation is somewhat different but just as informative in Botswana, where the Khoisans have been giving up their language even in places where they are the majority. According to Batibo (1998,2005), Yeyi (Khoisan) has been losing grounds to Tswana (a dialect of Setswana) in Ngamiland simply because the Khoisans have been subjected to some sort of serf status relative to the Bantu. What is equally informative from the above cases of language shift is that English, the official language in both Kenya and Botswana, hardly figures as a contender, despite its political and economic prestige in these countries. An important reason is undoubtedly the fact that it hardly functions as a vernacular, aside from the fact that it is probably spoken only by negligible proportions of these rural populations. Language competition is more local than linguists have shown it to be, and the fact that English does not function as a vernacular has certainly kept it out of the arena in which these indigenous vernaculars have been competing. This supports my position in Mufwene (2003d and Chapter 10) that European colonial languages are generally no threat to the indigenous languages of Africa and Asia. In some cases, as noted above, the expansion of the European official languages is hampered, instead, by the indigenous urban vernaculars, which are associated with an aspect of modernity that is more tangible, being closer to indigenous cultures. The above examples of language vitality in sub-Saharan Africa also make it opportune to discuss something else that seems to have been elusive to Western scholars, viz., that nonEuropeans, especially in former exploitation colonies, do not really want to be (exactly) like Europeans, though colonization has led them to aspire to new socioeconomic values. Invocations of the mass media, for instance, by Crystal (2004) to account for the spread of European languages in Africa and Asia and therefore for the potential threat the spread constitutes to indigenous languages reveal ignorance of the fact that in the Western world, where they are more accessible, the same mass media have failed to drive out scores of nonstandard varieties that are highly stigmatized. For instance, AAVE and American White Southern English have proved to be very resilient. Even Gullah, which is far more stigmatized, has "hung in there" and is probably there to stay for another while (Chapter 14). Overall, bilingualism and bidialectalism do not entail language or dialect shift, and the ecologies of language death must be more complex and diversified around the world than Nettle & Romaine (2000), Crystal (2000, 2004), Skutnabb-Kangas (2000), among a host of others, lead us to believe. Lest the whole issue of language endangerment is also trivialized, it helps to distinguish, as does Calvet (2002, 2006), between language competition at the national level and competition at the imperial level. The reason why French is losing imperial ground to English as a world language has to do more with differences in, inter alia, the international economic and military practices of the most important nations associated with these languages (viz., the USA, the United Kingdom, Australia, France, and Canada/Quebec) - as well as their culture diffusion strategies - than with a more aggressive or assertive introduction of English to non-Anglophone

239

240

Language Evolution populations, as alleged by writers such as Phillipson (2003) and Hagege (2006). Technological eminence alone is not enough lure. Otherwise, German would have played, and could still be playing, a greater role as a lingua franca in Europe than French does. Spanish has emerged as a world language more because of the legacy of its settlement colonization in densely populated parts of the world, a large part of Latin America, than because of the economic prestige of Spain.13 France performed poorly in the imperial venture of settlement colonization and would have to include French Creole speakers to inflate the number of French speakers. Overall, speakers of English and French as official or foreign languages in territories other than former settlement colonies are not particularly preoccupied by whether France, the United Kingdom, or the United States is economically or militarily more powerful. They are more concerned by the tangible, mostly local opportunities that their ability to speak one and/or the other language offers them to improve their professional profiles or increase their job prospects. Although many speakers capitalize on their competence in any or all of these languages in deciding one day whether or not they can travel to any of these countries or others, being able to use this competence to such an advantage was not a primary consideration, at least not for the majority of foreign language learners, when they decided to take classes in a particular language. For a large proportion, if not the vast majority, of such speakers, it was simply a matter of what foreign language was accessible or required during their formal education. More relevant to the imperial competition among the world languages is how they were introduced to a particular territory. There is a certain irony here, because La Francophonie have been quite aggressive in developing all sorts of strategies to promote and spread French around the world, but it also looks like the harder they try the more it is evident that French cannot keep up with English. On the other hand - if I may use a phrase favorite to Francophones when they refer collectively to the Britons and Americans - "Anglo-Saxon" companies do not make English a requirement in developing branches outside the UK, the USA, and Australia. Except for the highest management levels of their business ventures, the language of the relevant industry is the local lingua franca (especially at the lowest level of the work structure) or the official language.14 Invocations of McDonaldization to illustrate the cultural (including linguistic) Americanization of the world have typically overlooked the fact that McDonald eateries function in the local lingua franca, if not the local vernacular, wherever they are in the world. In keeping with the globalized business ambition of opening or keeping up the widest markets for one's products, Hollywood movies are now released in several world languages, so that only countries that either prefer them in the original language or do not have enough economic clout receive them in English. In either case, the economic principle of offer and demand applies. Major, multinational corporations that shape regional world-wide economic globalization do not make themselves missionaries of the world languages, because they believe in using strategies that achieve the most profit.

Globalization and the Myth of Killer Languages Very often this approach entails using the language of the local workforce and the local market, except in (most) senior managerial positions. (For similar perspective, see Marling 2006.) Though the corporations do provide opportunities through the basic fact of representing particular nations in the host countries where they have developed branches, it is the aspirations of local (potential) employees that create the market for English (or any other language of power), which then takes a free piggyback ride on the industry. Hopes to rise to high or top levels of the firms' managements incite the locals to learn English (or any other relevant language). Still, we must remember that this language operates in domains where it competes not with the indigenous vernaculars or lingua francas but with other imperial languages. Hence, however "global" English becomes (in the way Crystal 1997,2004 defines it or how Phillipson 2003 uses the term), it endangers only other European imperial languages with which it competes in the prestigious and potentially lucrative ethnographic domains allocated to them. The rapid and insidious spread of English thus owes much more to the success of the "Anglo-Saxon" (mostly American) business and military interventions around the world than to the aggressive language-teaching policy claimed by Phillipson (2003). If languageteaching could do so alone, the vitality of Irish in Ireland would have been restored by now. Contrary to what the government of Quebec has done for French, the Irish government has not enticed the local/national economic sector to operate in Irish. English has remained the language of the mainstream economy in Ireland. The commitment that the government of Ireland has made to Irish through the school system and the Gaeltachts over nearly a century now has been only symbolic, contributing to what can be called a "museum culture," which cultivates fond memories of the past but does not foster practice. Yet, the vitality of the language depends on those day-to-day practices in which speakers are enticed by various interactional conditions to speak a particular language rather than an alternative. The same explanation also accounts for why European colonial languages have not recruited more speakers than they could have in former exploitation colonies. They have been restricted to domains that do not involve, or exclude, the majority of speakers. Except to purists, languages do not particularly need schools to spread or remain viable. Linguistics based on Western practices has misled us into this illusion.

12.4 Will there be an English-only Europe? Invocations of language practice in institutions such as the European Union and the United Nations to prove either the displacement of other languages by English or mere competition among major languages oversimplify the complexity of language endangerment issues. They ignore the fact that not all cases of individual multilingualism or executive monolingualism are dangerous to the lives of the less prestigious languages. Highly stigmatized

241

242 2

Language Evolution minority varieties such as Gullah, AAVE, and Amish English in the United States tell us that such languages can be quite resilient. Recent studies such as Swaan (2001), Phillipson (2003), and Hagege (2006) confuse usage of lingua francas in specialized domains bearing little on communication in vernaculars with competition among vernaculars in domains most likely to affect their vitality. Increased usage of English in the offices of the United Nations in New York and Geneva, or in the offices of the European Union in Brussels and Strassburg, do not in any way jeopardize the naturalistic transmission of French as a vernacular from generation to generation in France, in Francophone Belgium and Switzerland, or in Quebec. From an ethnographic perspective, a distinction must be made between language use in a professional domain and vernacular language practice. Societally multilingual settings have naturally produced countless individuals, such as the white-collar elite in Africa and Asia, who routinely use languages that are not their vernaculars in their professional lives but continue to speak their ancestral languages in vernacular settings. The prestige associated with their professional languages has not at all deterred them from "transmitting" their indigenous languages (ethnic or urban vernacular) as mother tongues to their offspring. On the other hand, the recognition of Breton in the European Union as an ethnic or minority language will not necessarily revitalize it unless changes in language practice are adopted in Bretagne, where it faces the competition of French as a vernacular. To date, local efforts to revitalize Breton through Breton-language schools have produced only some bilingual children who in their adult lives may still wonder what practical advantages the language provides them, just like their counterparts in the Irish Gaeltachts who eventually leave these isolated communities for better economic opportunities in the city or elsewhere. While these efforts have definitely promoted the symbolic value of Breton, they have done little to enhance its market value in France's socioeconomic structure. As the above comparison between Quebec and Ireland shows, it is an enhanced market value that will really revitalize a language in its vernacular function, guaranteeing that it is "transmitted" naturalistically from one generation to another, with marginal dependence on the school system. The latter guarantees its preservation, not its maintenance, as is obvious from the Irish experience. I should perhaps emphasize that competition between English and French as imperial languages affects their ethnographic function as lingua francas, not their function as vernaculars. Just in case somebody takes seriously La Francophonie's propaganda that French is endangered by English, it is not clear to me how the vitality of vernacular French in France, Quebec, Belgium, and Switzerland could be endangered by the fact that Africans in Francophone Africa are increasingly interested in learning English as a foreign language. Isn't it ironic that an organization that is promoting partnership between French and the indigenous languages of its Third World members is promoting a policy that should protect a colonial linguistic hegemony in the same polities? Assuming that the imperial expansion of some European languages is really dangerous to the indigenous languages of Third World countries, why should speakers of the latter be more receptive to French than to

Globalization and the Myth of Killer Languages English in the first place? And why do the Western members of La Francophonie not forbid their citizens to learn English as a lingua franca? Phillipson (2003) and Hagege (2006) are also concerned that English has increasingly been used as a scholarly language in Western Europe. It is more and more accepted as a medium in higher education in Holland, Denmark, and Norway, among other polities. The French national center for scientific research (CNRS) now encourages its scholars to publish in English and the College de France now publishes an English version of its newsletter. Hagege sees in these developments a betrayal of the French glorious cultural heritage and condemns them as contributing to the endangerment of French. Unfortunately, he fails to distinguish between vernacular French, which is very vibrant (with some unavoidable "contaminations" from primarily English), and academic French as a lingua franca. In the post-colonial style of globalization, in which scholarly ideas compete, with relatively more liberalism, from different parts of the globe, one must ask why in the first place French, Dutch, Danish, and Norwegian scholars (among other Europeans) prefer to publish in English. Could it be that the academic market value of their languages has decreased in a post-colonial global world in which good scholarship has increasingly been recognized regardless of the national origin of ideas? Like vernacular speakers, scholars will naturally select languages which are the most beneficial to them. Here, the question of language rights clashes with that of human rights, particularly the right to express oneself in the language of one's choice and to reach the widest readership in one's research area. Should any institution inhibit the academic ambitions of some scholars by preventing them from publishing in the language that is the most beneficial to them? Granted, French is losing some of its ethnographic domains to English, just like Dutch, Norwegian, and Danish. However, it is also remarkable that the same scholars who publish in English (including the language advocates themselves) have not given up their national vernaculars. Those of us from the Third World who have evolved in polities that foster usage of different languages in different ethnographic settings have known all along that multilingualism in imperial languages is not necessarily dangerous to the indigenous lingua francas and vernaculars. Under particular ecological conditions, it is possible to accommodate the socioeconomically more powerful without losing one's language. As a matter of fact, the history of the prestigious Classical Latin tells the same story, as it lost to the competition of the European national vernaculars that had hitherto been kept in lower status. An important question is how to tell the ecological conditions which are dangerous from those which are not, and whether anything realistic can be done, consistent with the rights of the speakers and free of personal linguistic ideologies. Linguistic advocacy should clearly be distinguished from linguistics as a discipline that endeavors to explain, among other things, how language works. And the way to fight the preeminence of English as a scholarly language today lies apparently not in discouraging authors from publishing in this language but in changing the market ecology that makes it more advantageous to them.

243

244

Language Evolution Monolingualism in most European countries is only indirectly a consequence of policies promoting one language per nation. It is the direct effect of economic systems that have operated only in some dialect of the official language, at the exclusion of all other language varieties. It is this utilitarian aspect of language practice, through the potato plantation rather than through the boarding school, that spread English in Ireland. It is also what spread English in America, making it more and more difficult for immigrants from continental Europe to sustain parochial education in their respective national languages. It is the same practice that has endangered Native American languages, not the Jesuit fathers' Irishstyle boarding school system, although this contributed to the demographic attrition of speakers/"transmitters" of the indigenous languages by producing citizens who would live in the city and raise families speaking the dominant colonial language. Because most boarding school graduates did not go back to live on the reservations, they could not spread the English there. It would have been too artificial for the few who returned not to learn their heritage language back. As becomes more obvious in Chapter 13, Africa has not had a similar experience wTith European languages, because these have typically functioned as lingua francas for a small elite that is a legacy of the colonial auxiliary system and because the industry that would promote the spread of the European languages is very small. Nonetheless, its linguistic experience suggests that English is not about to endanger Dutch, Danish, and other continental European languages as long as the latter remain the vernaculars, the primary languages of the local governments, and the primary languages of the industry targeted by the vast majority of the populations, even if they are losing grounds to English in the academic world. The claim that the spread of English as a world lingua franca is endangering the vitality of indigenous languages outside the United Kingdom, the United States, Australia, and New Zealand seems to ignore the naturalness of egalitarian multilingualism as a way of life in many parts of the world. Worse of all, it ignores the fact that in many of the former exploitation colonies where English is spreading, this imperial language, like the other European world languages it is endangering, is spoken only by no more than 20 per cent of the local population. The remaining 80 per cent of the local populations who are practically disfranchised by the world-wide global economy are not at all affected by whether or not English or French is prevailing as a/the foreign language that interfaces their nations with the rest of the world. The relevant literature has also ignored some other basic things, including the following: Africans and Asians still hold to many of their traditions, despite the non-negligible cultural influence they have received from the West. Even the elite who speak European languages, as lingua francas and of course nonnatively, still claim that indigenous identity, which has made them loyal to the indigenous languages. Indigenous urban vernaculars and lingua francas such as Lingala, Swahili, Wolof, Hausa, and Town Bemba, which are locally associated with urban life and modernity, are appealing to rural populations, whereas European languages are still treated as foreign and often also with distrust, especially by those with

Globalization and the Myth of Killer Languages little or no formal education. The association of these European languages with the power of abusive and exploitative post-colonial governments has not at all helped their image with the masses of the disfranchised populations, even among those who learned enough to make themselves understood in some "broken" form. Besides, the high rate of illiteracy in socioeconomic systems that are far from recovering from either stagnation or deterioration also constitutes a major obstacle to the spread of any language that relies, unnaturally, on the school system for its "transmission." Invocations of globalization by Crystal (2004) and other linguists to forecast the obsolescence and replacement of indigenous languages around the world by English have simply been too simplistic to help us understand the complexity of competition and selection in the coexistence of languages.

12.5 The economics of language maintenance vs. shift As noted above, linguists' interest in language maintenance, often embarrassingly confused with language preservation,15 has been spurred largely by environmentalists' concern with the loss of biodiversity in the world's ecologies. Advocacy on the subject matter has increased since, among other concerns, the deforestation of rain forests and the discovery that the relevant vegetation contained medicinal plants that can contribute significantly to modern pharmaceutics. The fact that the industrial exploitation of such medicinal plants starts with folk knowledge of the indigenous populations, which is most accessible in their own heritage languages, has revalorized the Sapir-Whorf hypothesis. According to this, every language structures the world, more accurately its speakers' universe of experience, in its own peculiar way (Couto 2007), sometimes storing knowledge that is foreign to speakers of other languages. This particular benefit has prompted language rights advocates (e.g., Skutnabb-Kangas 2000, Maffi 2001, Miihlhausler 2003) to speak louder for the maintenance or preservation of indigenous languages in various former colonies. The advocates have argued that through the loss of their languages the affected populations are also losing cultural knowledge that had helped them survive in their traditional ecologies, an experience that endangers the future of mankind in the same way that deforestation does. All this can putatively be lumped together with numerous other cases of ecological mismanagement, which endanger the future of our planet, including our own as humans. A global examination of the whole situation prompts several questions, the first of which is whether language advocates have a global understanding of the complexity of the issues. Are the language advocates and environmentalists approaching the subject matter of endangerment with the same or similar concerns?16 Two concerns have been the most prominent among language rights advocates: (1) as noted above, loss of languages entails loss of traditional knowledge that can be helpful to

245

246

Language Evolution mankind (e.g. Nettle & Romaine 2000; Crystal 2000, 2004; Maffi (2001); and (2) loss of languages entails loss of linguistic diversity, which would be harmful to linguistics, because linguists would have fewer data for a more accurate understanding of the nature of typological variation and language universals (Krauss 1992; Hale 1992,1996). Starting with the first argument, bear in mind that it is the indigenous populations themselves that gave up their own languages, understandably under new socioeconomic pressures beyond their control. The literature has not articulated under what specific conditions the victims could have maintained them as vernaculars. In other words, what are the alternative ecologies that would have favored the victim populations' commitment to their indigenous vernaculars? We must note that the affected populations' lifestyles had changed or were changing, along with the socioeconomic ecologies that induced language shift. Therefore it is not clear how useful the knowledge that is embodied in the endangered or lost languages would have been to their speakers, except in their memories of the past, because it could no longer apply. Since languages also serve as markers of identity and can be used against their speakers, what is the point of maintaining diversity if it appears to be adverse to one's adaptation to the new socioeconomic ecology? Aren't linguists focusing on an epiphenomenon and simply being paternalistic without making an effort to understand what has led speakers to give up their languages? Regarding the loss of linguistic diversity, genetic linguistics informs us indirectly that it would be too simplistic to assume that languages have always evolved in the direction of loss. Its cladograms have always shown language diversification (as a matter of fact, with no reference whatsoever to language casualties). The different xenolectal varieties that evolved out of the Celts' gradual shift from their indigenous languages to Vulgar Latin have now become separate languages with their own numerous dialects. Thus, it is not clear what the future of Creole and indigenized varieties of European languages will be, as they continue to diverge, contrary to the misguided literature on decreolization (Mufwene 1994, 2005a). Although Creoles have already been disfranchised by linguists as separate languages (against the sentiments of many of their speakers), nothing precludes indigenized Englishes and the like from being identified some day as separate languages. As much as language rights advocates decry (rather inaccurately) the Anglicization of the world, it is also obvious that English is concurrently speciating into diverse varieties that may some day no longer be recognized as the same language. Besides, it is not likely that globalization will drive Chinese varieties and Japanese, among many other major or powerful vernaculars, to extinction, at least not in the near future. So, as in the past, linguistic diversity will always be self-regenerating, although the patterns of typological variation will hardly always be the same - and why should they? To the extent that in affecting each other - even in causing some other populations to lose their languages - human populations are enacting natural processes of evolution, would it not be equally informative to try to understand the evolutionary mechanisms and conditioning factors that are involved?

Globalization and the Myth of Killer Languages And wouldn't the next step be that of finding out whether there are victim populations that are maladaptive, in the sense of not doing the right things in their struggle to survive in their changing socioeconomic ecologies? To what extent are the solutions now proposed by language rights advocates really beneficial to populations whose languages are endangered? How can these solutions help the victims stop the endangerment when they themselves have little, or no, control over the socioeconomic changes that affect them and their languages? Language rights advocates have dodged the above questions. To be sure, they have also advocated multilingualism but have not articulated the ecological conditions under which this practice is sustainable. Could it be maintained as easily in former European settlement colonies as in Africa? In the former, European ideology of one language per nation has prevailed at the expense of ethnic identity since the late eighteenth century, whereas in the latter, the vitality of ethnic identity, which is primarily associated with language, has made difficult the implementation of this European ideology. Can most of the indigenous languages be maintained without changing the current socioeconomic world order among both the victims and those who control it? The answer to this latter question is obviously negative. The embarrassment is that language rights advocates have given little thought to the "revolution" (to borrow an ironic term from the title of Crystal 2004) that is entailed by their discourse. They have provided no answer to the implicit question of what alternative socioeconomic world order must be recommended to the victims to meet their new material and spiritual aspirations, which depend in part on languages of the workforce. Equally embarrassing yet is the fact that language rights advocates keep preaching to the victims without sensitizing the victimizers, those who run the socioeconomic and political machineries that affect population structures. This machinery is not necessarily the government, which is also one of its products. It is true that funding has been raised to document and analyze endangered languages. In practical terms, one must ask whether it is linguists or would-be speakers of the relevant languages who are benefiting (the most) from the exercise. If one had to rate the socioeconomic needs of the relevant populations in a world that is admittedly materialistic - nonetheless one that has shaped their aspirations - the question is whether funding to linguists earning a living and building a career on the preservation of endangered languages is helping create more jobs and improve living conditions for the relevant populations. Are linguists helping the humanity or helping linguistics? In what ways is the fact of documenting these languages as important as, if not more than, teaching literacy to several children in the language of the workforce so that, everything being equal, they may be more competitive in their socioeconomic ecologies? The phrase "everything being equal" is quite critical here, and the question is whether what it so obviously refers to is not what should call for more advocacy. (See Brutt-Griffler 2005 for a similar argument.) Regarding losing heritage knowledge concurrently with the loss of languages, even Edward Sapir, who has been associated with the linguistic relativity hypothesis, has made it clear that language and culture are not wedded like two sides of a coin. One can change

247

248

Language Evolution without engaging the other in the process (Sapir 1921). The facts presented above about deforestation and language endangerment also suggest that the only knowledge that a language would preserve about a destroyed ecology would be archival, available to experts. That knowledge would not be any more helpful to the average speaker than knowledge of medieval English cultural practices is to an average speaker of modern English. On the other hand, the preservation of the ecology and of the cultural practices it has traditionally nurtured would entail transferring the relevant conceptual distinctions to the language that came to replace the indigenous one. This is made obvious particularly by the indigenization of European languages in former European exploitation colonies. Even in settlement colonies, where the impact of European languages has been generally catastrophic, these have been enriched with indigenous terminologies in domains that have been assimilated into European cultures. European cultures and languages in the Americas and Australia are new phenomena influenced by the local physical and cultural ecologies, despite the heavy heritage from Europe. The most annoying part is perhaps the static notion of culture that the language rights advocacy literature has suggested, as if traditional culture had to survive in its pristine form and as if European cultures were unaffected by the cultures of the populations that adopted them. Another aspect of the economics of language shift is what opposes linguists to economists. The latter have found the multiplicity of languages to be inhibitive to economic development. For instance, Alesina et al. (2003) argue that, although there are exceptions, societally or regionally multilingual nations have a higher ethnicity index than monolingual ones.17 Countries in the latter category, especially those of the Third World, tend to have too many conflicting ethnic interests, which inhibit the implementation of efficient national economic development programs. One can of course argue that poor political leadership rather than multilingualism and multiethnicity- is the real source of the problem. However, limited financial resources and socioeconomic systems often inherited from poor colonial rules that were insensitive to local traditions have made it almost impossible for uncorrupted political regimes to evolve that would not make ethnicity and the ensuing multilingualism non-issues. Post-colonial governments committed to solving their development problems have thus shown a strong interest in the European ideology of one language per nation, especially under the pressure of Western lending institutions which stipulate unilaterally the terms of their loans. Ironically, most of the multiethnic nations concerned by this discussion are former exploitation colonies, precisely where indigenous languages are generally less endangered, least of all by European languages. Whether or not the economists' advice is sound, it has the merit of treating languages as factors that affect human economic development, therefore making the adaptability of people to the current economic world order central to the debate. To be sure, language rights advocates have always claimed to protect the interests of the speakers of endangered languages. However, as noted above, they have hardly explained

Globalization and the Myth of Killer Languages how their solutions should help the relevant populations adapt (more) competitively to their ever-changing socioeconomic ecologies. While it is clear that past socioeconomic world orders in which the indigenous languages used to thrive cannot be restored, especially in former settlement colonies, language rights advocates have not articulated the alternative ecologies in which the endangered languages can be revitalized. Nor have they articulated in what ways competence in the endangered languages would make them competitive with those speaking languages to which they have been shifting. Because socioeconomic integration has been an important factor in the process of language loss, one must wonder whether the advocates are not promoting an ethnolinguistic segregation that can be disadvantageous to the affected populations. To be fair, what linguists have advocated is bilingualism that should enable the affected populations to remain loyal to their ethnolinguistic traditions while being fluent in the language of the workforce. My question is whether this is still possible when the ambient socioeconomic ecologies have typically led to the disintegration of the affected communities, which have become minorities in their own homelands and whose members have increasingly been led to abandon them and join the majority populations. Noticing that a language is endangered is largely a retrospective outlook on evolution, particularly because language loss is the cumulative result of individual decisions that have been made independent of each other during isolated communicative events. Few individuals if any have the foresight of anticipating that their communicative practices will cumulate into populationwide language shift and loss. It is not clear whether such individuals react with guilt to the cumulative outcomes of their communicative acts when asked to think retrospectively about them, especially why they did not "transmit" them to their offspring. Chances are that in the latter case the conditions were not favorable to, or supportive of, the "transmission" of languages other than the non-indigenous vernaculars. In this case, it is also worth asking whether indigenous populations in former settlement colonies have shifted languages under conditions different from some of the immigrants, such the Italians and Germans in North America, Australia, and Brazil, for instance. Would the relevant Natives want to go back to what is left of their isolated traditional communities? Perhaps language rights advocates should find out why most of them have not. Yet, it is likely that those left behind in the traditional communities who have given up their indigenous languages are following the example of those who have moved out. Even if the latter come back to visit every now and then, their competence in the heritage language may be in such a bad shape that they would prefer not to speak it poorly or they may prefer to show their success by speaking the vernacular of the majority community. Language practices are easier to copy from speakers one can associate with than from others. Could language attrition in the affected communities reflect evolution toward a solution of the lesser evil? If so, then language rights advocates, who typically happen to be of European descent, should really explain why they advocate imposing on "indigenous populations"

249

250

Language Evolution solutions that they would not impose on descendants of European immigrants, most of whom are indeed monolingual in a language other than those of their ancestors. Overall, things boil down to the following: People react adaptively to changes in their socioeconomic ecologies. Some of these involve giving up their languages, and sometimes their cultures, for those of the prevailing group while also maintaining their identities through the ways they adapt the new languages and cultures to their communicative needs. From the point of view of culture, this is a process of indigenization, which winds up producing a new and hybrid culture that continues to define their ever-adaptive identities. Although it is useful to document the past, it seems equally important to ask whether, from an evolutionary perspective and from the point of view of populations which define their own identities, language rights advocates should bemoan normal adjustments that people continue to make to changing ecologies. They should ask themselves whether they are approaching the question of language endangerment in the same ways, or with the same objectivity, as environmentalists are addressing that of species endangerment.

12.6

Conclusions

Language endangerment is certainly not a recent process, any more than globalization is. Globalization itself cannot be dissociated from colonization as a process whereby one population not only relocates to some other place but also dominates the latter's indigenous population and imposes on the latter its own style of economic and political organization. The recent form of globalization differs from earlier ones by the scope of the world involved and by the speed of communication and transportation between "nodes" in its networks of interdependencies, as well as by the complexity of its organization. Globalization need not be of world-wide scope, and in fact most of its forms are local and regional, amounting simply to complex networks of interdependencies. This is in fact its basic form in Western Europe, in its former settlement colonies, and in Japan. It has developed the most in those nations that have espoused the ideology of one language per nation, a language that in its standard and vernacular varieties has emerged as the language of the industry and of the workforce. These are countries whose dominant populations have adopted largely monolingual practices and have marginalized other languages and/or reduced them to the status of minority ones. These are also countries where socioeconomic pressures, rather than repeated political stipulations (as in the case of France since the seventeenth century or so), have the most successfully driven speakers of the minority or marginalized languages to shift vernaculars in favor of the dominant one. Some of these territories, such as Europe and Japan, are ironically among the most densely populated places in the world but have the fewest languages. It is not evident that modern local or regional globalization accounts for all countries that can be characterized as societally monolingual. Settlement colonization history and the geographical size of the nation states can account for societal monolingualism in

Globalization and the Myth of Killer Languages countries such as Burundi, Rwanda, Thailand, and Korea. A combination of these reasons may account for the situation in Japan. A long succession of imperial regimes may account for the situation in China (despite pockets of its minority languages), although it remains an open question whether "Chinese" consists of several dialects of one single language or of a cluster of genetically and typologically related languages. World-wide globalization since the late 1980s maybe considered as a new form of exploitation colonization, one in which multinational corporations headquartered mostly in the Western world, rather than specific nations, are colonizing Third World nations without ruling them politically. While European settlement and exploitation colonization since the fifteenth century had expanded the Indo-European dispersal and spread major European languages around the world, the latest form of world-wide globalization has spread English beyond the political boundaries set up by the earlier colonial regimes. It is this latter form of economic globalization that has shaken the imperial position of French and has made it obvious that Spanish and Portuguese (and, to some extent, Russian?) although major languages are not real world languages. The situation is in fact a challenge for linguists to articulate more clearly the distinction between MAJOR and WORLD languages. Although they may have more speakers world-wide, major languages such as Mandarin, Cantonese, and Hindi have traditionally been spoken each in one geographical area, are associated primarily with single ethnic populations (though there are some minority populations who have shifted to them), and continue to function as vernaculars even in their diasporas. On the other hand, world languages are spoken in various places around the world, by ethnically diverse populations, a large proportion of which use them only as lingua francas. As a matter of fact the majority of English and French users are nonnative speakers, who are incidentally the major agents of the current spread of these languages, a phenomenon that is reminiscent of the way Latin was spread in the European Romance countries, originally by its legionaries and during the post-Roman period by the local elite (Chapter 11). The recent rapid spread of English around the world is also comparable to that of major European languages during the exploitation colonization of Africa, Asia, and the Pacific since the nineteenth century. It is spreading among the same kind of intellectual and/or politicaleconomic elite who had not only appropriated and indigenized European languages as their lingua francas but also maintained their indigenous languages as their vernaculars. Thus, while English has shaken the position of French and other colonial European languages in these former exploitation colonies, it is far from endangering the indigenous ones. English is not really the "killer language" that non-global approaches to language endangerment have painted it to be, certainly not in relation to the indigenous languages of former exploitation colonies. Moreover, it does not have the kind of agency that would have made it a "killer," because it is (would-be) speakers who "kill" their languages by opting to speak another language, which amounts to what is known as language shift.

251

252

Language Evolution In the case of sub-Saharan Africa, for instance, it is important to take account of the expansion of the indigenous lingua francas that also function as urban vernaculars and are spreading at the expense of traditional ethnic languages and apparently also of the colonial European languages which continue to function as official languages.18 Non-negligible is also the still-ongoing effect of indigenous population contacts that started before European colonization, such as those of the Bantu with the Pygmy and Khoisan populations, at the expense of the languages of the latter groups. History also reminds us of other settlement contacts between Bantu and other non-Bantu populations, such as in the northern part of the Democratic Republic of Congo (DRC) and in Ruanda and Burundi. The latter case raises, for instance, the interesting question of why and how the Tutsi in both countries have "Bantuized" linguistically whereas the Maasai (also Nilotic) in Kenya have not. Recent history marked by genocidal wars and refugee resettlements have also produced new intimate contacts between diverse ethnolinguistic populations. Time will tell what the outcomes of these new population contacts will be. There are also many places where small Bantu populations are found surrounded by larger Bantu populations speaking different languages. Such enclave populations have nonetheless maintained their languages as markers of their separate ethnic identities. Linguistics ought to articulate the specific conditions under which this has remained possible, especially in a history marred by periods of interethnic wars, even though some of these may not have been local. All the above considerations suggest that a global approach capable of discriminating a wide range of variation in how language competition and selection have proceeded around the world is what linguistics needs in order to understand language endangerment. What is true of Europe and its former settlement colonies is not necessarily true of the rest of the world. While it is certainly regrettable that some languages are vanishing, it is undeniably evident that the socioeconomic ecologies in which speakers have evolved have also changed to points of no return. We cannot address the question of language loss without concurrently addressing that of costs and benefits to would-be speakers of the lost and endangered languages. A problem as "wicked" and "vexed" as this one cannot be addressed unidimensionally nor non-globally.

Myths of Globalization: What African Demolinguistics Reveals Chapter Outline 13.1 13.2 13.3 13.4 13.5

13.1

Introduction What GLOBALIZATION means and what it does not Language vitality in sub-Saharan Africa: a historical perspective Lessons from South Africa Conclusions

253 254 258 265 269

Introduction

I have borrowed the term demolinguistics from French linguistics, to situate the discourse on language endangerment and loss where it really belongs, viz., at the level of population and society conceived of as constructs from individuals who interact with each other or have the potential to do so in particular geographical settings. This is indeed the level at which language evolution, which endangerment and loss are aspects of and which results from the cumulation of the behaviors of individual speakers, must also be situated. I also focus on the instrumental function of language, thus on the agency of speakers assessing their languages as tools intended to help them adapt to changing socioeconomic ecologies rather than on the "systems of units and rules" that enable them to function as means of communication. Multilingual speakers constantly determine which language is more appropriate on every particular occasion of interaction. Ultimately, they develop "market values" of the languages in their repertoires, based on the choices they are repeatedly forced to make in various communicative settings. In a way they realize which language is more advantageous to them, although in reality they only notice retrospectively that, by some sort of atrophy, (former) multilingual adults are less, or no longer, competent in one or more of the languages. Others simply realize that they are hardly invested in some other language they could have learned in their multilingual setting. None of these outcomes of the competition among coexistent languages has anything to do with their structural features. All languages are adequately equipped to serve the traditional communicative needs of their speakers and to adapt to new ones. It is simply a matter of whether a language is deemed more advantageous to serve

13

254

Language Evolution not only a speaker's communicative needs but also his/her rapport with other speakers than other alternatives in various interaction settings. Using Bourdieu's (1991) notion of "linguistic market" at its fullest, one can consider language choice in terms of "costs" and "benefits" to the user. This sort of "language economics" approach makes it more justified to invoke "globalization" as a possible reason for language endangerment and loss, although, as discussed in Chapters 11 and 12, the question is which of the many interpretations of GLOBALIZATION matters and how. Since most of the limited theorizing on the subject matter has been based on the particular experiences of especially the Americas and Australia, I thought it would be rewarding to question some of the relevant generalizations with more facts from sub-Saharan Africa. The demolinguistics perspective also helps me validate a position of mine about language evolution since Mufwene (2001a), viz., ecology works directly on individuals and it is patterns of interactions among individuals that account for how changes spread, typically discriminately and at varying speeds, within the overall population and their communal language.

13.2 What GLOBALIZATION means and what it does not As is obvious from Chapters 11 and 12, globalization has often been invoked (too casually, I think) as the main cause of the ongoing and imminent death of "indigenous languages." It has been blamed as the mechanism that has allegedly transformed English into a "killer language" par excellence. What globalization actually is, with its many facets, has hardly been explained by the linguists who invoke it. Contexts of its invocation suggest, however, that they mean by it any of the following interpretations or combinations thereof: (1) the fact that various and sometimes very distant parts of the world today are interconnected by complex networks of rapid international telecommunication and transportation; (2) the claim that national boundaries between neighboring countries are apparently fading thanks to the emergence of multinational companies that have developed webs of economic interdependencies; and (3) the impression that the world has become culturally less and less diverse, thus more and more uniform, as the same kinds of industrial goods can now be found almost everywhere. Linguists have routinely adduced evidence for this latter impression from the wider and wider diffusion of Hollywood movies and the increasing presence of McDonald's stores around the world. Often enough, they have equated GLOBALIZATION with AMERICANIZATION but failed to note that, like the spread of English around the world, both McDonaldization and the world-wide diffusion of Hollywood movies are results of the globalization process itself (under the above interpretation), not the means of its implementation.

Myths of Globalization: What African Demolinguistics Reveals It is interesting to note that McDonald's menus have diversified around the world and been adapted to local dietary customs, that the language of the menus is no more English than the staff speaks the language everywhere in the world, and that Hollywood movies are typically released simultaneously in different languages around the world and sometimes in locally modified versions (see also Marling 2006). These practices clearly show that businessmen are not language missionaries and their primary interest lies in making profits, which often entails communicating with the local population in their lingua franca, if not in their vernacular. In the "foreign" branches of multinational corporations headed by Americans, Britons, or Australians, English is used only in the top tiers of the administration, alternating with the local official language. The local official language and lingua franca of the masses of the population are used in the other levels of production. This demolinguistic structure applies even in the diplomatic missions of these countries, when local personnel are also employed. Members of the local population learn English for the purpose of moving up in their careers and earning a better living. While we can admit that American, British, and Australian businesses provide motivation for members of the indigenous population to learn English, they are far from making a deliberate effort to spread their language (see also Schneider 2007). Little has changed from the colonial tradition that restricted the colonial language to a small, privileged, and elite class of colonial auxiliaries (Brutt-Griffler 2002a). Even when their metropolitan language is widely spoken in a particular territory, multinational executives still prefer a local operator who can speak the local official language fluently and insist on having a structure in which the masses of employees in the bottom tiers are addressed in a language that they can understand the best, a local one. As explained in the previous two Chapters, for the local population, the additional, foreign language hardly calls for giving up their local lingua franca, least of all their ethnic vernaculars, chiefly because the languages have different ethnographic functions. Local speakers of the European languages, albeit their indigenized varieties, still must participate in the local culture, or buy from the local market, in the local urban vernacular or in their regional lingua franca. And they still must interact with their grandparents, parents, and the economically less affluent or less educated members of their families or native places in their ethnic vernaculars. The same applies to recent urbanites who normally do not want to sever their cultural links with childhood friends often left in the rural areas, regardless of whether they now evolve just in the urban vernacular and/or in the European colonial language. Despite the increasing number of interethnic marriages, adults who have recently migrated to the city like to maintain their authenticity by using whatever opportunity they find to speak their ethnic vernacular, however much it may be influenced by the urban vernacular and regardless of the fact that their children may not speak it at all. If anything endangers the ethnic vernaculars, it is not the "global" languages used at the top echelon of the multinational companies; rather it is the urban vernaculars and regional lingua francas (such as Lingala in DRC, Swahili in East Africa, Town Bemba in Zambia, and

255

256

Language Evolution Wolof in Senegal) that do. They are the languages associated with modernity for the masses of the population (Spitulnik 1999), and they are the ones in which the Western-style economy (outside the white-collar sector) is run, and popular culture (epitomized by modern music) is produced. They are the emblems of indigenous nationhood, in opposition the European intrusion and domination associated with the European languages, regardless of the nationality of their speakers. One must remember that, with the exception of South Africa, sub-Saharan Africa was generally colonized in the exploitation style, which fostered socioeconomic segregation between the colonizers and the colonized, typically interfaced by an elite class of colonial auxiliaries, and assigned only a lingua franca function to the European languages. What linguists have failed to note is the local form of globalization identified by cultural studies experts as GLOCALIZATION, in reference to local interdependencies that regulate the local, grass-roots economic system everywhere. It has borne the most significantly on the vitality of languages. It is the factor that determines which variety of the dominant language will spread (typically not the standard or the most prestigious one) and within which part of the population. Within the local population structure, it is this "glocal" economic structure which has endangered indigenous languages. Even in the Western world, it is that dominant language variety at the level of jobs associated with the masses of the population which has endangered the other vernaculars, indigenous and non-indigenous alike. In the American and Australian settlement colonies, in which the language of the blue-collar economy has also evolved to be the dominant vernacular, my explanation accounts for the disappearance not only of the Native American and Aboriginal languages, but also of the European vernaculars that once competed with the now dominant language. The timing of the disappearance can be correlated with the time when a particular population was forced by the circumstances to assimilate to and function in the dominant economy. As explained in Chapter 12, those who were the weakest economically were first to be affected. It is also noteworthy that Native Americans, like African immigrants of the twentieth century, are among those populations that do not have their separate ethnic neighborhoods in North American cities. The same is true of many European immigrants since the second half of the twentieth century who had to shift from their ancestral languages to the dominant one, despite current belated efforts among some to revive their ethnic cultures. This urban immersion pattern has made it difficult for both Native Americans and recent, non-segregated immigrants to maintain their languages, as their children have found it more advantageous to learn English, which they use to socialize not only with Americanborn children but also with others from ethnolinguistic backgrounds different from their own. What linguists have poorly understood is the general trend, viz., that the shift has usually been an adaptive response of the populations that are weaker economically or socially, regardless of race, provided they are integrated in or have access to the dominant socioeconomic structure.

Myths of Globalization: What African Demolinguistics Reveals In the case of Africa, we would be remiss to overlook the effects of interethnic wars and refugeeism on ethnic languages. Like natural disasters, ethnic genocides endanger languages. Fortunately these have not been as devastating as in the colonization of the Caribbean, where Carib languages were "killed" through the relocation and extermination of their speakers by the colonists. Refugeeism has led to population contacts in which the migrants are linguistically disadvantaged, being under pressure to interact with the host population in their language and often thinking it more advantageous not be identified as refugees when such concealment is possible. The longer the relocation lasts, the longer children often find it even more advantageous to speak the host, rather than the ancestral, language. If there is no chance to return to the homeland and the children are integrated in the host population, the ancestral language will die with their (grand)parents. This is in part how the European languages died in settlement colonies. As the dominant socioeconomic structure prevailed and other European minorities were integrated and functioned in it, the children migrated socioeconomically and linguistically into it, even if they did not necessarily move out of their ethnic communities. Eventually the physical migration from the neighborhood followed, reinforcing and/or speeding the process of language shift. More or less the same kind of explanation accounts for the loss of Native American languages. It is often said (e.g., by Crosby 1992) that it is not the European weapons that killed Native Americans during the early stages of colonization, it was the Old World ills that the immigrants brought with them, to which Native Americans were not immuned, which reduced their population sizes. As well shown by Boyle (2007), the ills endangered or "killed" some Native American languages, through the reduction of the demographic sizes of speakers and by causing the survivors to relocate in territories where their languages came to compete with those of their hosts. It was certainly not the occasional contacts with the European languages before the nineteenth century that "killed" them. Part of the explanation for the reportedly more rapid loss of Native American languages since the ninteenth century certainly has to do with migrations from the reservations to the city, in which they have not lived segregated, which has taken the same toll on their ancestral languages as on those of immigrants. My explanations are perhaps too speculative and they certainly do not account for all the reasons why Native American languages are so seriously endangered today, but somebody will have to explain why these factors have typically not been invoked to account for the endangerment and loss of Native American languages. There is more to the story than the alleged lack of pride on the part of Native Americans toward their heritage or the prestige of the Western lifestyle and languages. Adaptation to the new socioeconomic world order seems to be the bottom line, regardless of whether some individuals find the solution in running away to places and populations less affected by the newcomers and their ills or in seeking to partake in the new socioeconomic system. These are all aspects of glocalization, at least when one considers how ancestral traditions are weakened in the competition by the system that prevails.

257

258

Language Evolution

13.3 Language vitality in sub-Saharan Africa: a historical perspective Black Africa has always been geographically multilingual, due to the multitude of languages, some of them demographically very small, spoken there. In precolonial times, limited geographical mobility, due to a host of physical ecological factors and to highly localized subsistence economy, did not enable extensive societal multilingualism, restricting this to royal courts, trade centers, adjacent villages (especially in the case of those consisting of a main "monoethnic" village with an ancillary community from a different ethnic group). While there were trade languages even during that period, such as Swahili (very much exploited by Arab traders in East Africa; Nurse & Spear 1985, Fabian 1986), KikongoKimanyanga (in the western part of the present DRC; Fehderau 1966), Songhay (across the Sahel; Nicolai 1990), Lingala (along the north-south stretch of the Congo River; Hulstaert 1974), and Haussa (in northern Nigeria and Chad), no regional or nation-wide lingua francas associated with functions of the state or with ordinary cross-ethnic communication seem to have emerged. Kingdoms relied heavily on the service of interpreters. As observed by Samarin (1982), travelers just learned the relevant neighboring language, unless they followed a trade route and knew the relevant lingua franca. The socioeconomic setup simply did not favor the emergence of extensive individual or societal multilingualism before the colonial period. Societal multilingualism increased with the emergence of colonial administrative posts, the forerunners of today's cities, where indigenous populations from different ethnolinguistic backgrounds were brought to serve the colonial administration in different capacities and therefore live with each other. Many so-called "contact varieties," such as Sango, Lingala, Kikongo-Kituba, and Town Bemba, are the products of this colonial expansion, and they have indeed evolved into urban vernaculars and regional lingua francas. By today's standards, the precolonial setup was very rural, marked by relative isolation of the different ethnic groups surviving on localized subsistence economies. It favored the maintenance of ethnolinguistic diversity. The question is whether the colonial rule really disrupted much of this state of affairs. The European colonizers did indeed modify the African "linguascape" by introducing their own languages and by producing new socioeconomic structures that favored the emergence of new language varieties, such as Kikongo-Kituba in DRC and the Republic of Congo, Kinubi in Sudan, Cameroon and Nigerian Pidgin Englishes, Fanakalo in South Africa, Town Bemba in Zambia, and Sango in the Central African Republic. In other cases, the new socioeconomic structures transformed some indigenous languages such as Wolof in Senegal, Swahili in East Africa, Hausa in Nigeria, and Lingala in DRC and the Republic of Congo1 into urban vernaculars, while also spreading them as regional lingua francas. To the extent that the birth and/or expansion of these languages are associated with colonization, one

Myths of Globalization: What African Demolinguistics Reveals must also understand the way in which the colonial rule of Africa proceeded. Most of the colonial administrators remained in the cities or towns that evolved from their administrative posts. They visited the surrounding rural areas of their respective administrative provinces or districts occasionally, limiting the visits to the minor administrative posts manned by indigenous colonial auxiliaries. For most administrators, their competence in the indigenous languages was limited to the urban vernaculars and regional lingua francas. Those who did more became the forerunners of Africanist linguists, totaling very few. Overall they did not use their European languages to communicate with the masses of the indigenous populations either; they limited their usage to interactions with the indigenous elite of their colonial auxiliaries (Brutt-Griffler 2002a). While they expanded the precolonial royal courts' interpreter system, the European colonizers also introduced the practice of language stratification in Africa. This stratification was controlled through the school system and through the geographical structure of their colonies, as usage of the European languages was limited to urban centers and the Christian missions, just like Latin in the Western Roman Empire. The school system controlled the spread of the European languages (i.e., their scholastic varieties), limiting it to a few who could advance fairly high in their education and could work as certified teachers or colonial auxiliaries. In a colonial system where the rate of scholarization remained very low and only a very small proportion of pupils progressed to, or past, high school, mastery of the European languages remained the privilege of a few. That system has almost continued to date. Competence in the European languages has become emblematic of education through high school, if not past it. Moreover, very few acquire them as mother tongues or use them as vernaculars. Used in the post-primary education system, in the government and higher echelons of public administration as well as in the white-collar sectors of the new socioeconomic world order, the European languages have hardly encroached in the communicative domains of the indigenous ethnic vernaculars. Thus the European lingua francas and the indigenous vernaculars have hardly competed with each other, if at all. The lesson from the above is, simply, that languages can be in competition only when they share communicative domains, where they would compete for the same speakers and topics. They can coexist when they do not compete with each other, at least not for the vast majority of the relevant population, just like species that share the same econiche but thrive on different resources, e.g., giraffes and antelopes in the savannah. What has happened in the West, where varieties of the same language have been in usage for all the communicative functions, has hardly occurred in sub-Saharan Africa. It is thus misguided to extrapolate hastily, without an accurate comparison of the relevant ecologies of language practice, from the Western experience with language competition to another socioeconomic structure. Language coexistence does not unavoidably lead to the extinction of minority or marginalized ones. Bearing in mind that vernaculars compete with other vernaculars and lingua francas with other lingua francas, the geographical structure of Africa, with poorly developed

259

260

Language Evolution transportation systems, has not made it possible for the urban vernaculars to endanger the ethnic languages of the rural areas, where the vast majorities of the populations still live. Outside the city, the urban vernaculars typically function as regional lingua francas, therefore they do not compete for speakers in most of the domains of vernacular communication. Very few affluent urban vernacular speakers return to settle back in the rural areas. When they do, it is often for reasons that make them more likely to communicate with the home population in the local ethnic vernacular. Among such reasons is the fear to appear snobbish or disdainful of their ancestral culture. So, speakers of ethnic varieties feel no serious motivation to give them up for the urban vernaculars, though they accommodate outsiders in making an effort to be bilingual (see below). Outside the city, where ethnic languages are definitely giving way to urban vernaculars, danger for the ethnic languages in rural areas lies in, for instance, students who attend boarding schools and experience the stigmatization of being rural. Affected by this attitude toward their backgrounds, they often return home showing off and pretending not to be able to speak their ethnic vernaculars. Those who proceed to leave the village and live in the city may give them up altogether, though many grow to keep them as markers of ethnolinguistic identity and as a means by which they can connect more strongly with those who share their ethnic backgrounds. This is, however, no guarantee that the children of these new urbanites will learn to speak their ancestral languages, as they face the pressure to blend with the "native" urbanite population whose exclusive or primary linguistic identity is associated with the urban vernacular. On the other hand, those school graduates or dropouts who remain in the village soon give in to the criticism of being snobbish or confused. They return to the tradition of using their ethnic vernacular for day-to-day communication, reserving their knowledge of the regional lingua franca for communication with outsiders. The fact that the city and the village are two worlds apart is an important factor that continues to protect the indigenous ethnic languages. In the very least, this factor slows down the threat from the regional lingua franca, especially through the media. Whether or not several ethnic languages in sub-Saharan Africa are endangered by the urban vernaculars or regional lingua francas depends largely on whether it will urbanize (soon), on the model of the Western world. There are a number of reasons why the regional languages used in the media do not necessarily endanger the ethnic vernaculars. The first is that, in the rural areas, they serve only as lingua francas with the occasional outsider (public servant or trader) and therefore do not endanger the local vernaculars. The second reason has to do with economics. The print media is not well developed and does not circulate in the villages. Even if it did, the dominant mode of rural, subsistence economy is almost tantamount to unemployment in the national modern economic structure, thus to lack of extra cash to afford the media and books. Even literate villagers do not read past the school experience and a few random occasions of reading correspondence; therefore they cannot be affected by the print media.

Myths of Globalization: What African Demolinguistics Reveals Moreover, in several parts of sub-Saharan Africa, the indigenous urban vernaculars and regional lingua francas are allocated only small proportions of time on the radio (and television), which remains quite limited, even when all the programs in the indigenous languages are counted together, compared to the time allocated to the official language. Since this medium is not interactive, there is no reason to fear that the listeners' ability to understand the relevant lingua francas can endanger their communicative competence in their local vernaculars. So, it appears that, on their own, the media cannot endanger other languages. They can if they reinforce usage of the relevant lingua francas in face-to-face interactions in domains that encroach on those of the local vernaculars. There are a few things that can be learned from this particular interpretation of the dynamics of language coexistence and competition in sub-Saharan Africa. The first has to do with the factors that determine the language choices that speakers make on a day-to-day basis when they interact with other people. It is these individual choices that cumulatively bias the fate of a language toward its maintenance or extinction in a particular community. Pace Swaan (2001), they have very little to do with considering a language as a common good or resource that speakers either should look after or can neglect, more or less consciously. While language choices can be discussed from the perspective of Bourdieu's (1991) MARKET VALUE, this has almost nothing to do with a "Q-value" (Swaan 2001) determined by how many other people speak a particular language or how large the population is that an individual has the potential to communicate with in the relevant language. If this were the case, the total population of our planet would be aiming at speaking Chinese or Hindi today. Nor do speakers' choices of particular languages have anything to do with identifying what languages have world-wide or "global" currency as lingua francas (or "(super)central languages," in Swaan's terminology). Most Africans would otherwise be endeavoring to develop some competence in the official languages of their countries. They know how much economic and political power these lingua francas have, but they also know that even if they spoke them, their economic conditions would not necessarily improve, because there aren't that many jobs that are worth the investment that one makes in time and money to develop that particular competence.2 The decisions which speakers make to speak language LI or L2 are typically very local and circumstantial. They are determined by the interactive settings in which a speaker finds him- or herself, what language the other participants or potential interlocutors use, what advantages they gain in speaking it, what social price one is likely to pay if they did not conform to the local norm, whether the price is worth the alternative gains, what one decides to do economically, and where one decides to live the rest of their life. The benefits from such language investments are not typically seen in the very long term either. Although we must admit that schools have long-term visions in providing language courses to students, the value of that investment in time and effort for the student is co-determined by his/her

261

262

Language Evolution socioeconomic environment, where he/she assesses in practical terms whether it is worthwhile. In this case, the local job markets bear as much on their decision as the students' personal ambitions (which remain local for the majority) and the particular individuals they interact with. The greatest danger to indigenous African languages lies more in the rural exodus toward the city than in the potential any "global" or "colonial" language has to displace them. I explain this below with an example from South Africa. In former settlement colonies, European languages have endangered their indigenous counterparts because of urbanization, which has reduced the rural populations to small minorities, and because the prevailing European-style socioeconomic structure has spread geographically and socially. This structure functions only in the European language, and gradual geographical and social assimilation of the relevant populations (starting with the Europeans themselves) has let it encroach onto the domains of the traditional vernaculars. These developments are not happening in Africa, not only because the European languages have remained sectional but also because the non-traditional economy has not spread outside the city. Evidence for this may also be found in Latin America, where the Western-style economy has not spread as fast, or as pervasively, as in North America. Native American languages have died the most within and around cities such as Sao Paulo and Rio de Janeiro than in the more rural areas, and more in the southern and eastern parts where sugarcane cultivation, mining, and ranching have prospered as modern industries than in the northern part where the dense Amazon forest has resisted the rapid expansion of the European colonization and its socioeconomic structure until recently. It is precisely where the European presence is the least felt that Native American languages have survived. (See, e.g., Nettle & Romaine 2000 for statistics.) Interestingly, as pointed out in Chapter 12, Phillipson (2003) and Hagege (2006) express strong concerns that English may be evolving into the dominant vernacular of continental Western Europe. Among the main reasons they adduce are the following: (1) English is prevailing as the main working language of the European Union. Despite strong support for multilingualism within (the administration of) the Union, more and more representatives find it more expedient to communicate directly in English than to wait for the translation. (2) Universities in Scandinavian countries and in Holland allow courses to be taught in English rather than in the local, official language. (3) More and more European academics prefer to publish in English than in their respective national academic language (e.g., French, Dutch, German, and Norwegian), aiming at a wider readership. CNRS, the dominant funding agency for scholarly research in France, encourages its grantees to publish their findings in English, to ensure wider dissemination and the recognition of these in the international market of scholarship. (4) English has become the dominant language of international (academic) conferences, even those held in France and Germany. (5) More and more European private companies are requiring knowledge of English by their senior and/or white-collar employees. Consequently, more and more Europeans select English as their second language, over French, German, and Russian.

Myths of Globalization: What African Demolinguistics Reveals Although no two situations are identical regarding language coexistence and competition, these authors describe an evolution quite similar to language practice in sub-Saharan Africa since the colonial period. A European colonial language is typically the dominant medium of communication in high school and in higher education, in the government, in the higher echelons of public administration, in the legal system, and in all prestigious white-collar jobs. However, as explained above, the European colonial languages have typically remained lingua francas and have therefore not endangered the vitality of the indigenous languages. Their main negative impact may consist in the following: The continued colonial practice of associating them, almost exclusively, with prestigious socioeconomic functions has prevented the indigenous lingua francas from rising to more or less the same ethnographic status. Even in Tanzania, where Swahili has been promoted for almost half a century as a co-official language, English remains very prestigious (see, e.g., Billings 2006). In Kinshasa, DRC, where Lingala remains the dominant language of popular culture, French still claims exclusively the status of language of scholarship and world affairs. Nonetheless, despite all this ethnographic stratification, the indigenous languages are generally not endangered by the European colonial languages. Like most students of language endangerment, both Phillipson (2003) and Hagege (2006) fail to distinguish vernacular and lingua franca functions of languages, as well as the fact that languages are in competition mostly when they are used for the same ethnographic functions, making one or some of the competing languages redundant. In the particular case of the European Union and the other cases discussed above, the European languages are competing as lingua francas but not as vernaculars. There have been no reported cases of Dutch or French children in Holland and France, respectively, opting not to acquire their national vernaculars as mother tongues, even those who are speakers of minority languages. If there are some such cases, they are isolated ones that are far from setting a particular trend among the relevant child populations. In the particular case of France, Breton activists' children who are raised in their heritage language still must learn French, so that they are not disadvantaged in comparison to the overwhelming majority of French monolinguals and on the job market (Tomlinson 2006). Discussions of the potential for language shift such as Phillipson (2003) and Hagege (2006) overlook a critical condition for language endangerment, viz., speakers no longer use their traditional language for day-to-day communication or do so less and less routinely. A language is not really endangered unless it loses its vernacular function. There are many languages in the world whose usage has been limited to vernacular communication, while the other aspects of their speakers' communications are assumed by other languages. In fact, this has always been true of the nonstandard varieties of the same European languages, which have not typically had lingua franca functions but have been quite resilient despite both their stigmatization by speakers of the "standard" or educated varieties and the relentless campaign of the school system to eradicate them.

263

264

Language Evolution Also worth noting in this particular extension of the concern with language endangerment are invocations by La Francophonie, and to some extent Hagege (2006), that French is endangered by English in Francophone territories other than France (including its overseas departments), Belgium, Switzerland, and Quebec. It is true that usage of English as a lingua franca has spread all over the world since the mid-twentieth century, especially since the independence of former exploitation colonies of Africa and Asia in the 1960s and 1970s, and also since the collapse of the Eastern European bloc. English has been penetrating even former French and Belgian colonies, where it is being taught as a foreign language, in addition to the official language. French has not expanded to the same extent in former English colonies. In places such as Brazil, there has even been a shift from French to English as the preferred foreign language. In Central Europe, English has made faster and more expansive strides than French has. French may have maintained its status as "language of culture" but it has not been as much identified with modernity in technology or with economic and military might as English. France has not emerged as much as a world power as the United Kingdom (with its huge commonwealth of former colonies) and the United States, singly or combined. As observed in Chapter 12, La Francophonie has reacted to this "threat" by initiating "language partnerships" especially with the African former colonies of France and Belgium, allegedly so that they can all defend themselves better from the English expansion (Vigouroux, in press). The theoretical model for this planned resistance to language endangerment is provided by Swaan's (2001, and Calvefs (2006) "gravitational" model of language coexistence. According to them, languages in different nations or regions can form galaxy-type "constellations" in which "(super)central," official languages (French in this case) enable them to connect with the world on the international level, while the indigenous languages facilitate contacts within the masses of the population. La Francophonie more or less uses this model to argue that the commitment of speakers of these languages to French would protect their heritage languages from the expansion of English, although it may not be able to stop it. La Francophonie can in return fund the promotion of the (demographically more important) indigenous languages nationally, especially in the area of education. As Vigouroux aptly asks, what do African languages really have to gain from this "partnership"? It is not obvious that the indigenous African languages have much to gain from the proposed "partnership." The level of funding is usually only symbolic and too small to solve practical issues associated with the teaching of even a privileged subset of the indigenous languages and with the production of linguistic teaching materials. Insofar as the maintenance of the indigenous languages is concerned, schools have had a marginal role to play, although adequate education in indigenous languages would produce qualitatively more fluent literacy. Note that the ethnographic marginalization of the indigenous languages during the colonial rule and by most of the post-independence administrations (a process aptly captured by the French term minoration Hnguistique) has not reduced their vitality.

Myths of Globalization: What African Demolinguistics Reveals Any threat to them has come from elsewhere and, as shown above, potentially no more from English as an imperial lingua franca than it has come from French, the official language. As shown in Chapter 12, the "gravitational" model itself has been conceived of on a fundamental ignorance of important cases in Africa itself that speak against it. Two such counterexamples can be found in Tanzania and Cameroon, where English and French (in eastern Cameroon in the latter case) replaced German as the "central language" but the shift has affected neither the vitality of the indigenous languages nor the geographical and demographic expansion of Swahili in Tanzania. Language coexistence and competition are certainly complex states of affairs and such oversimplified approaches to them do very little that sheds light on the process of language endangerment as one of the possible outcomes, least of all about the conditions that produce it.

13.4

Lessons from South Africa

With an overwhelming Black majority, 76.7 per cent of the total population according to the 1996 census, which is largely disfranchised from the modern economic system, South Africa is very similar to other sub-Saharan African countries. Yet, it is also so different from the same nations, not only because it now has the most prosperous economic system in subSaharan Africa and is emerging as a major player in international diplomacy but also because it has had a different kind of colonial history. Unlike the other sub-Saharan African countries, which evolved from trade to exploitation colonies and hosted a moderate number of European colonizers, South Africa experienced an overlay of settlement and exploitation colonizations, which both introduced a sizeable proportion of Europeans (close to 10%) and Asians (as slaves or contract laborers up to and since the beginning of the nineteenth century, respectively). Quite early in the mid-seventeenth century, Dutch subjects came to settle in South Africa, joined later on by some French Huguenots and Germans, and together they would later identify themselves, in the late nineteenth century, as Afrikaners. This self-identification was in resistance to the British exploitation colonization since the late eighteenth century, under which they were disfranchised politically, economically, and linguistically and decided to distinguish themselves more clearly as African Whites from the European Whites, more or less like White French Creoles in the Caribbean (such as the Beke of Martinique), in Louisiana, and in the Mascarene islands (Indian Ocean). Toward the end of the nineteenth century, they would also declare the political and structural autonomy of their language variety, henceforth known as Afrikaans, from metropolitan Dutch. The British colonizers also introduced English, which the Afrikaners felt obliged to "acquire" as a lingua franca, while keeping Dutch as an exclusionary vernacular and the dominant business language in parts of South Africa where they reached substantial proportions of the total population (now close to 20% in Western Cape, close to 5% in Eastern Cape, around 6% in KwaZulu-Natal,

265

266

Language Evolution and around 20% in Gauteng, estimated from the 1996 census) or became economically the most powerful ones. To be sure, Afrikaans also became the vernacular of many Khoisan populations who, also facing the southwards expansion of the Bantu populations, were apparently the first to be enslaved by the European colonists (those who would become the Boers and, later, the Afrikaners) and worked mostly as their domestics. It was likewise appropriated as a vernacular by most of the Coloreds (products of mixed unions between the Europeans and the non-Europeans), the majority in Western Cape (54.2% of the total population) and in Northern Cape (51.8%, according to the 1996 census). Of equally limited access to the Bantu people until the collapse of the apartheid system, English remained the vernacular of a minority of British colonizers/settlers and South African Indians. Under the apartheid regime, the Bantu languages remained largely contained in the rural areas, since their speakers were prohibited by the law of the land to live with or near the urban White populations. Those who worked there lived in the neighboring townships, the home of most Colored and Asian populations. The most destitute of them lived in squatter camps, which have now mushroomed into huge South African urban peripheries of the economically disfranchised. The country thus entered the twentieth century with a unique population structure that defines the dynamics of language coexistence and competition differently from elsewhere, with peculiarities that are nonetheless quite informative about ecological factors that bear on the processes of language endangerment and loss. The first Asian populations brought to South Africa were Malay speakers. They came as slaves and, like African slaves in the New World and Indian Ocean, they too lost their language to the masters' language: then colonial Dutch and now Afrikaans. Living conditions, under which they did not form critical masses locally, must have prevented them from maintaining their language, almost like the African slaves of the homestead phase in the New World and the Indian Ocean. Those populations are nowadays included among the Coloreds. The contract laborers from the Indian Peninsula had an experience similar to those who were taken to the Caribbean (as in Guyana and Trinidad, where they became Creole speakers, as explained in Chapter 10). Multilingualism among them and the work conditions in the mines forced them to shift from their ancestral languages. Social isolation or distinction from the Bantu populations and the economic destitution of the latter kept the newcomers from investing in any of the Bantu languages, none of which is widely spoken throughout the country. They appropriated Fanakalo, a new contact-based variety based primarily on Zulu, for work in the mines. Gradually, primarily through the schooling of their children (Mesthrie 2006), they would shift to English as a vernacular, which has evolved into what Mesthrie (1992) calls South African Indian English. It is to this new English vernacular that Indie languages have been losing the competition in South Africa, as children have found it less competitive to stick to their parents' language traditions.

Myths of Globalization: What African Demolinguistics Reveals As noted above, most of the Khoisan populations, whose land was being settled by both the Bantu and the Boers populations (the Dutch and other Europeans who had settled before the British colonization), lost their languages to those of both groups of colonists. As the Bantu populations were barred by the apartheid system from living in the city and from socializing with the Europeans, few of them learned either English or Afrikaans as a vernacular, although the better educated of them acquired these European languages as lingua francas. As a matter of fact, the anti-apartheid riots of the 1970s had a lot to do with how Black South Africans claimed their rights to English, which held one of the keys to the modern economic world order.3 The containment of the Bantu languages in the rural areas, where they have also remained geographically and socially separate from each other as vernaculars, has kept them from competing with each other and as a group with the European languages. Recent migrations to the squatter camps have been largely within the same language areas, as poverty limits the distance that their speakers can travel. Although English has been gaining more currency in the city, even among Bantu populations, it will be a long time before it spreads as a vernacular in the townships and the squatter camps, certainly not before the current socioeconomic structure has changed. The fact that at least half of the South African Bantu population is still rural is the best protection that their languages have got, though it sadly also reflects the extent to which the same populations continue to be marginalized economically.4 South African Blacks are still where Native Americans were in the nineteenth-century United States, when isolation on some reservations protected their languages while marginalizing them from the economic prosperity enjoyed by the majority of the European immigrants. Quite interesting about language vitality in South Africa is also the evolution of the two European languages. The balance of power between them has shifted significantly since the end of the apartheid regime. English has strengthened its position as the lingua franca of predilection among South Africans, at least among those who have graduated from high school. More and more affluent families of all races in the city are raising their children as native English speakers. Although most Afrikaners are still raising their children as Afrikaans native speakers, more and more of them are also raising them as bilingual native speakers of Afrikaans and English. According to Breeder et al. (2002), there are also a few Afrikaner families who are raising their children as monolingual or dominant-English speakers. Particularly significant in this trend is the fact that Afrikaans seems to be returning to its nineteenth-century status as an ethnic language, especially with more and more affluent Colored parents and members of the Bantu elite making sure that their children grow to be fluent (native) English speakers ready to compete with other English speakers in the economic sector where English is the dominant language (Giliomee 2003). On TV and the radio, there are more programs in English than in Afrikaans. On TV, many Afrikaans programs are shown with subtitles in English, while the converse is not true. English is

267

268

Language Evolution definitely growing as the dominant lingua franca of the educated and/or affluent segment of the population at the expense of Afrikaans. Overall, Afrikaans was the mother tongue of close to 6 million South Africans (14.4% of the total population; most of the speakers being Afrikaners and Colored) in 2000, while English was the mother tongue of only 3.5 million (8.6% of the total population; 40% of the speakers being Whites; Giliomee 2003). Despite these statistics, the current dynamics of language practice suggest that Afrikaans may be "at risk" (if I may borrow this term from language rights advocates), especially with more and more Colored children acquiring English only, with fewer and fewer non-Afrikaners showing interest in speaking Afrikaans, and also more and more Afrikaner children acquiring both English and Afrikaans natively or only English. As observed by Giliomee (2003: 21), English is emerging as "the only language that is spoken by virtually all the literate people" in South Africa. The situation is reminiscent of the process that drove several "minority" European languages to extinction in North America. Loss of protection from the government (unlike the practice during apartheid) and the association of English with the more lucrative sectors of the economy and connection with the outside world maybe sounding the toll of the demise of Afrikaans, a process that is already also obvious in the academic sector. It is generally dangerous to hazard predictions about evolution, as it is subject to various ecological vicissitudes that may change unpredictably. The revitalization of Afrikaans in the late nineteenth and early twentieth centuries speaks clearly against such gloomy prognostication. The main reason why I chose to speculate on the possible demise of Afrikaans is the fact that linguists have relied too heavily on demographics to predict which particular languages were about to die. They have usually claimed that the languages that have few speakers are likely to die soon, despite the fact that the sizes of some of the relevant populations of speakers are not decreasing, and the relevant languages may not be threatened by "globalization" as long they remain isolated or marginalized from the dominant cultures of the polities in which they are situated. I submit that it is the evolutionary trajectory experienced by a particular language (rather than the relative demographic size of its speakers compared to that of users of another language) which determines its future. Statistics alone without the history of trends do not mean much. Many indigenous languages in South Africa with fewer speakers (such as siSwati, Tshivenda, and isiNdebele, each spoken by less than one million people) are not following this evolutionary trajectory of Afrikaans, largely because most of their speakers are not in the same socioeconomic situation as those of Afrikaans and they do not experience any particular competition from English. It also appears that speakers do indeed target particular languages locally, in the geographical spaces they share directly with their neighbors (i.e., other people they interact with), and their day-to-day decisions are based on the benefits they hope to derive immediately or in the near future, if they must take classes. Local economic considerations figure prominently among the relevant factors.

Myths of Globalization: What African Demolinguistics Reveals South Africa also draws our attention to the role that urbanization can play in driving language endangerment and loss. As elsewhere in Black Africa, South African cities are contact settings in which languages experience new forms of coexistence and competition. One of the legacies of the apartheid system is that indigenous populations of most cities in South Africa are predominantly from the ethnolinguistic regions where they are located, producing no real competition among the indigenous languages. The competition between the indigenous and Asian languages has also been preempted by the apartheid regime, which has made obvious the fact that the geographical coexistence of languages does not necessarily entail competition among them, especially when the political system has imposed a clearcut social complementary distribution in the practice of the languages. The competition that matters the most in South African cities is that between the European and the other languages. In the case of the Bantu languages, South Africa makes it obvious that a whole lot depends on how successful and accessible the economic system is. As noted above, the new political and economic regime since the collapse of the apartheid system has made English the most important language in the urban economy, especially outside the squatter camps and the townships. The proportion of non-Whites who speak English in a city such as Cape Town is very high, compared to speakers of French in cities such as Kinshasa, Brazzaville, and Bangui in Central Africa. Could the economic "boom" of South Africa herald the spread of English even within the rural population once it spreads to the countryside too and the unemployment rate (now at 33%, lower than elsewhere in Black Africa) drops dramatically? This particular evolution is actually reminiscent of that of Abidjan, in Cote dTvoire, where, as an urban vernacular or lingua franca, Dioula has not spread to the same extent as Lingala in Kinshasa or Wolof in Dakar and le fratifais populaire ivoirien (FPI) has emerged as an alternative lingua franca or vernacular among the less educated, separate from the local standard variety of French used by the more educated and affluent urbanites. Abidjan is one of those few places in colonial Africa (other than Gabon and Mozambique) where no indigenous language has prevailed as the universal vernacular or lingua franca of the masses of the urban population. It is not clear what the reason is for this and why Wolof has evolved into the urban vernacular of Dakar but not Dioula in Abidjan. Both cities are capitals of former French colonies. Although FPI is associated only with the local informal economy, not with the mainstream sectors of national economy, its position does not seem that different from that of Creoles in the Anglophone Caribbean and in Mauritius when postEmancipation contract laborers were brought to work there in the nineteenth century (Chapter 10). The future is open to all sorts of speculations.

13.5

Conclusions

The experience of language endangerment and loss has certainly not been the same in Africa as in North America and Australia, which the literature has presented as prototypical.

269

270

Language Evolution Fewer languages have been endangered or reported dead in Africa than in these latter parts of the world. Where they have, other indigenous African languages had more to do with the processes than the European languages, with the exception of some of the Khoisan languages in South Africa. These exceptions also show that the African experience has not been uniform everywhere either, which is in fact reminiscent of the fact that the experience of language endangerment and loss has been different in North America than in Latin America, where it is less advanced (Nettle & Romaine 2000). Although European colonization has been the main trigger of these evolutions around the world over the past half millennium, it did not proceed the same way in North America and Australia as it did in Africa. The consequences of exploitation colonization differ from those of settlement colonization regarding the kinds of economic systems that have developed in the different colonies and the population structures that ensued. One important factor to consider is the extent to which the new socioeconomic world order replaced the old one. That determined whether or not (most of) the indigenous populations found it necessary to acquire the colonial language in order to adapt to their changing ecologies. Regardless of the extent to which the colonial language produced by its new speakers has diverged from the metropolitan varieties, an important question is whether the Natives have found it necessary for their adaptation. Although it is important to ask whether, in the first place, the colonial system made it accessible to the Natives and how, the history of anti-apartheid riots in South Africa makes it obvious that the colonized populations can claim their right to important economic languages of their nation; they can fight for access to the relevant languages. Quite significant here is also the question of whether the indigenous populations speak the colonial languages as vernaculars or as lingua francas. It appears that indigenous languages have been threatened by European colonial languages the most in those places where the latter have prevailed as cross-ethnic vernaculars, not in places where they have functioned primarily, if not almost only, as lingua francas. This is not to say that no indigenous languages in former European colonies have been threatened by other languages. It only means that while European colonization can be the main reason why the socioeconomic world order has changed in many parts of the world, the European languages have not always been the ones affecting the vitality of indigenous languages. Sometimes it is some indigenous languages that have displaced other indigenous languages especially when they share vernacular functions for the same populations. The variation in the evolutions summarized above can be correlated with differences in colonization styles. As noted in Chapter 3, trade colonization has generally not disturbed the traditional patterns of language practice. For most speakers the repertoires of languages have remained the same. Those involved in the trade with the foreigners learned to speak the foreign lingua franca or developed a special variety of one of their own languages associated with trade.

Myths of Globalization: What African Demolinguistics Reveals It can thus safely be observed that trade colonization added new language varieties to the local repertoires. The main reason for this particularly positive "balance sheet" is that the new varieties did not compete in communicative function with the traditional languages or language varieties. They came along with the new communicative functions that they served. This particular evolutionary pattern can also be recognized in the introduction of scholastic varieties of European colonial languages as elite lingua francas, associated with non-traditional communicative functions and therefore competing not even with the new indigenous lingua francas and certainly not with the indigenous ethnic vernaculars. Instead, it is the urban vernaculars associated with the dominant, blue-collar sector of the limited modern economy that have become a threat to the vernaculars, certainly in the city. However, only the future will tell whether the urban vernaculars will displace the rural ones and when. A significant factor in this case is whether the African economies will prosper from their present stagnation, if not disintegration, which bears heavily on whether language practice in the city will spread into the rural areas. Exceptions to the emergence of indigenous languages as urban vernaculars in Black Africa are to be found in cities like Abidjan (Cote d'lvoire), Libreville (Gabon), and Maputo (Mozambique). As in other African countries, only the future will tell whether, with some of the current changes in national language ideologies, the European languages will really spread to rural areas in the way they have in former settlement colonies around the world and therefore displace (most/many of) the indigenous regional lingua francas and/or ethnic vernaculars. To date, there is no indication that they are, not even in South Africa, where the European languages compete primarily with each other in the city and seem to coexist with the indigenous languages in some partly topographic complementary distribution. Most of the traditional speakers of the indigenous languages are still in the rural areas or have joined the growing squatter camps around the townships. In connection with the above, the (non) uniform future of indigenous languages in South African cities can teach us a great deal, as it can highlight the connection that language spread has with economic development. Although South Africa already underscores, along with other sub-Saharan African countries the significance of variation within a population's geographical and social structure, it is special because it is one of the few places where settlement and exploitation colonization have overlapped. It may help us articulate more clearly what particular factors make trade, settlement, and exploitation colonizations so different from each other with regard to language endangerment. The present Chapter helps us explain what linguists have misunderstood about the role of globalization in the endangerment of indigenous languages around the world. It makes obvious, I hope, that they have been misguided in singling English out as the "killer language" par excellence, aside from the mistake of attributing to it agency that it does not have. If one understands by GLOBALIZATION the world-wide diffusion of industrial and other cultural goods from particular parts of the world through a better infrastructure of

271

272

Language Evolution transportation and telecommunication, then it is noteworthy that European languages, especially English, have not spread around the world to assume the same communicative functions everywhere. In many places, these languages have spread only as lingua francas and have therefore not affected the indigenous languages that continue to function as vernaculars, especially in rural areas, where the majorities of Third World populations continue to live. Some forms of spatial and societal multilingualism do not necessarily create the kind of situations where one language must prevail at the expense of others. There can thus be multilingualism without competition of vernaculars, therefore without language endangerment, at least not immediately. If one means by GLOBALIZATION the interdependencies that hold among the different parts and sectors of a territory's or community's economic system, then it is clear that this state of affairs is far from being uniform around the world. The economies of Western Europe and its former settlement colonies are more globalized than those of its former exploitation colonies. This is where language loss and endangerment are the more advanced. North America is more globalized than Latin America, and more indigenous languages have survived in the latter than in the former. Overall, the factors determining language endangerment and loss are more local and complex than has been shown in the literature. Things have certainly not evolved in a uniform way around the world. Finally, the impact of wars and refugeeism on the vitality of some languages is still to be explored. In this respect, Africa may be experiencing something that is decreasing elsewhere, but it is an aspect of language endangerment and loss that deserves attention, especially if we care to understand what happened in the past, in the history of mankind.

A Case Study: The Ecology of Gullah's Survival1 Chapter Outline 14.1 14.2 14.3 14.4

14.1

Introduction Fears of Gullah's death: myths and facts Why hasn't Gullah died yet? Conclusions

273 273 275 283

Introduction

In this Chapter, I discuss the kind of ecology that, on the one hand, has helped Gullah stay more or less the same since it stabilized apparently in the nineteenth century and, on the other hand, has kept it from debasilectalizing by assimilation to varieties of English spoken either by descendants of Europeans or by the educated middle class. I do not, however, claim that Gullah is not endangered. I argue below that it is not endangered by "decreolization" qua debasilectalization but by factors that are largely ecological, such as the exodus of its speakers to Charleston and Columbia, South Carolina, and other cities, particularly Atlanta, New York, and Washington, DC, for better job opportunities. Before discussing all this, however, it will help to provide a short historiography of the speculations and debates about Gullah's "imminent" death.

14.2

Fears of Gullah's death: myths and facts

Gullah has been claimed to be dying since the end of the late nineteenth century, when the Reverend John Williams claimed in the Charleston, S.C., Sunday News (1885): ... as the "befoo de wah" negroes die out, and the present generation of young ones get "lahnin out the book" and scatter off from the plantations, the old time Gullah will die out.

This fear was echoed by Stoney & Shelby (1930), to whom Reinecke (1937/1969: 494) attributes the following quotation, which anticipated the discourse of today's language activists:

14

274

Language Evolution It may not be long before Gullah is a lost tongue; and the point of view and tales that go with it will have vanished. Therefore the fiction writer performs a semi-historic function when he sympathetically records all three.

Fourteen years later, Stoddard (1944: 188) warns again: ... until very recent years the dialect remained nearly the same. No doubt there has been more change during the past twenty years than during the century preceding it. Gulla [sic] as a language is passing out rapidly. Even the older Negroes on the sea islands speak the dialect now if excited, and with their passing it will become obsolete.

The observation that "the older Negroes ... speak the dialect now if excited" is related to one expressed more accurately earlier by Smith (1926) and Johnson (1930), according to whom Gullah was spoken mostly as an in-group variety. This practice is still true today. Highly stigmatized, this English variety, disfranchised by linguists as a Creole and as a separate language (see Chapter 3), is seldom spoken before outsiders. Accessing it requires patience and being accepted by the Native community. The claims of Gullah's "decreolization" are discussed extensively in Mufwene (1994), where I question the empirical foundation of this speculation and argue that Gullah is pretty much alive, probably with as much vitality as it had in its beginnings. They are based on the following incorrect assumptions disputed in Chapter 3: (1) as a Creole, Gullah must have evolved from an erstwhile pidgin; (2) there was a time when all field hands, the vast majority of African slaves in coastal South Carolina and Georgia, must have spoken it, as opposed to English, presumably spoken only by Whites and very few slaves who have privileged contacts with them; (3) Gullah must have originally had a monolithic structure corresponding to the basilect projected in linguistic analyses; and (4) the variation attested in it today must reflect influence from standard English, to which African Americans have been exposed in school and through increased interactions with European Americans. These assumptions also underlie Jones-Jackson's hypothesis that Gullah was decreolizing (i.e., debasilectalizing). I addressed all these assumptions and claims in the 1994 essay. I maintain that the socioeconomic history of coastal South Carolina and Georgia supports the hypothesis that Creoles evolved by basilectalization, diverging away from their "lexifiers," a thesis that is also supported by Mille (1990). Basilectalization is a process consistent, though not identical with Labov & Harris' (1986) and Bailey & Maynor's (1987) divergence hypothesis on the development of AAVE, according to which the vernacular has been diverging from White varieties of English.2 In the case of Creoles, it consists of the consolidation of basilectal features into a basilect. Putting things in the broader picture of North American colonial history, I submit that the hypothesis that Gullah evolved gradually, basilectalizing away from the initial colonial English koine to which the slaves of the homestead phase had been exposed, is more

A Case Study: The Ecology of Gullah's Survival consistent with the way the plantation industry developed. As these evolved from smaller and relatively integrated homesteads to larger and more-segregated plantation communities in which (full descendants of) Europeans and (descendants of) Africans interacted less and less, life expectancy dropped quickly, and the labor population kept growing more by importations from Africa and Europe than by birth, and faster, which created conditions of rapid population replacement (Mufwene 2001a, Chapter 2). This entailed a situation of language "transmission" in which many of the model speakers developed idiolects that were becoming more and more divergent from the closer approximations of the colonial koine of the homestead phase. These conditions, which favored substrate influence (not necessarily "apports" from the substrate languages, as explained in Chapter 8), also favored the divergence of Gullah from the other emergent American English varieties. Note that together they were all diverging from British English(es), under differing contact conditions. As in Mufwene (1994, 1996a), I conjecture that Gullah must have basilectalized gradually in the course of the eighteenth century and that this process must have continued up to around the abolition of slavery in the mid-nineteenth century.3 After that period, Gullah must have been as stable as any other new variety of English that developed in North America, although it need not have consolidated concurrently with them. Gullah has been under no more pressure to change than any other nonstandard variety of English in North America. As noted above, the above position finds partial structural support in Mille (1990), who shows, through a judicious period-based comparison of Ambrose Gonzales' texts, that his Gullah of the late nineteenth century is very similar to Patricia Jones-Jackson's and Mufwene's field data collected in the 1970s and 1980s respectively.4 With liberties typical of a literary and fiction writer, he apparently embellished his early-twentieth-century Gullah into a more basilectal style that is hardly matched by any native speaker's vernacular today regarding the density of basilectal features. This, of course, misled some linguists into hypothesizing a debasilectalization hypothesis that has not been documented or supported diachronically (Mufwene 1994).

14.3 Why hasn't Gullah died yet? Gullah is undoubtedly one of the most stigmatized American English varieties. Smith (1926: 22) speaks of "one author" (apparently John Bennet, Esquire) who characterizes it as "the worst English in the world." Bennett (1908-1909) considers it "so singular in its sound as constantly to be mistaken for a foreign language" (332), as "the quaint outlandish jargon of the swampy regions of the coastal plain" (336), and as "the quite logical wreck of once tolerable English [of the English yeomen and indentured servants], obsolete in pronunciation, dialectical [sic] in its usage" (338). The stigmatization is so strong that its speakers prefer switching to some other variety or not speaking before outsiders, a practice that partly accounts for why many scholars have assumed it to be dying. If the conditions articulated by

275

276

Language Evolution DeCamp (1971) for decreolization qua debasilectalization were correct, they would have certainly conspired with the linguistic insecurity of its speakers to guarantee its death by transformation if not by language shift. Neither of these changes appears to be happening, at least not in the immediate future. Given these fears of its demise, expressed for over a century, its amazing resilience deserves some explanation, which I attempt below, invoking all sorts of factors from the ecology of its evolution and practice. As we are concerned with language vitality, the ecological factors I invoke are essentially external, including GROUP IDENTITY, SOLIDARITY, LOYALTY, RESIDENTIAL PATTERNS, and ABILITY TO SWITCH CODES, although the internal ecological factor of INTRA-DIALECTAL VARIATION also appears to bear on this aspect of Gullah's evolution. I start with ecological factors that led the colonial English appropriated by (descendants of) Africans in coastal South Carolina and Georgia to diverge into Gullah. Contrary to the literature on the development of Creoles, the disproportion of (descendants of) Africans to that of (descendants of) Europeans is not the primary reason why Creoles have developed. Otherwise, AAVE, a genetically related variety, would not have developed in communities where (descendants of) Africans have always been minorities. In most of the American hinterland they have been minorities, even if they may have reached higher proportions on cotton and tobacco plantations in coastal Virginia (AuCoin 2002) and North Carolina. The primary reason for the divergence of English among African Americans into AAVE is the same as that advanced by proponents of the divergence hypothesis: reduced interaction between African and European Americans. Since segregation precluded African and European Americans from socializing routinely across racial boundaries, the restructuring of English into North American varieties was subject to influences that varied in some respects. This situation cleared the way for more, or fewer, divergent developments depending also on (dis)similarities of the initial conditions. Thus AAVE remains close to White nonstandard varieties in the American southeast, from which it has diverged only since the late nineteenth century, based on Bailey & Maynor (1987, 1989), Bailey & Thomas (1998) and Schneider (1995), among others. Gullah is more divergent because social segregation in coastal South Carolina was institutionalized in the early eighteenth century (Wood 1974) and the African slave population continued to grow primarily by importations, which created the conditions discussed in Chapter 3 as favorable to the emergence of a creole. They were more favorable to substrate influence and favored more marked divergence from the "lexifier." It must be remembered that, by the Founder Principle (Mufwene 1996a, 200la), the earlier colonial models spoken alike by Africans and Europeans on the farms, trade posts, and the early plantations were the targets that were gradually restructured by each generation of newcomers. Thus most features of the later, today's nonstandard vernaculars may be traced to those earlier models, as well observed by Krapp (1924), Kurath (1928), Johnson (1930), and several other dialectologists since then. However, the reason why Gullah has evolved to

A Case Study: The Ecology of Gullah's Survival be so different from the nonstandard English vernaculars spoken by especially Whites in the American hinterlands was not what these scholars proposed. African Americans did not preserve an archaic form of English outgrown by European Americans; the varieties just developed concurrently, diverging in some respects but also retaining several common features whose frequencies differ ethnically. Though constituting only indirect evidence for Gullah, this scenario is consistent with studies such as Wolfram (1974), Schneider (1989), Poplack (2000), and Poplack & Tagliamonte (2001), in addition to the dialectological ones cited above, which argue that the origins of most features of AAVE lie in their colonial "lexifier(s)." However, as explained in Chapter 3, we must remember not to confuse origins with the influence that the substrate languages must have exerted on the particular selections attested in Gullah, as in AAVE, either in simply increasing the statistical frequencies of some features or in favoring their reanalyses. In short, segregation fostered the development of separate group identities, including separate language norms, even where differences are minimal. It's common knowledge that de facto residential segregation has not disappeared in the USA, despite its abolition de jure. Thus, the vast majority of African Americans socialize among themselves, just like the vast majority of European Americans do among themselves (and likewise with other ethnic groups). Although members of the different ethnic groups meet at school, at work, and at other public places and institutions, the sense of ethnic identity developed in the African American community has not favored abandoning features of "the community."5 Shedding it and the language variety that marks it is typically associated with moving out of "the community," which only very few have been able to do. Remaining loyal to their backgrounds, most highly educated African Americans have learned to switch to the educated variety of English (with, in most cases, some features that still distinguish them from other Americans), but they can easily revert to some form of African American English (AAE, which I use to subsume both AAVE and Gullah) when they want to. I have been told twice by non-linguists that middle-class African Americans who attend (predominantly) African American colleges or universities, such as Howard and Atlanta Universities, often work on increasing their AAE features just in order to reassert their identity. Otherwise, most African Americans who grew up speaking AAE have generally retained their variable linguistic features (less so in the case of Gullah, given how more stigmatized it is outside coastal South Carolina and Georgia), just as much as they stick to other communitybased norms, such as cooking, singing, dancing, and religious styles.6 In the case of Gullah, group identity has been an important factor in preserving the variety. (One can see the strength of this ecological factor more convincingly in the offshoots of this population that migrated to Texas and Oklahoma in the nineteenth century, per Hancock 1980b, as they have continued to speak it.) Although those who go to school outside "the community" - especially the few who go to college in the American hinterland return home not speaking the same way, a significant proportion of them learn to codeswitch successfully. The reason for retaining Gullah, as stigmatized as it may be outside

277

278

Language Evolution "the community" is that, as one high-school senior on Wadmalaw Island once explained to me, they would not like to be considered snobbish to the peers they left behind (Mufwene 1988b, 1993). If they spoke Gullah before they left coastal South Carolina or Georgia, people who return after often long periods of residence in New York and similar cities typically revert back to this vernacular, because for them it is one of the ways to identify with home. As reported in Mufwene (1993), some of Gullah's most basilectal features are attested in such speakers, after they have idealized it according to a stereotype that is not quite an accurate picture of the variable vernacular spoken by the community members they left behind.7 Nonetheless, one can see here the strength of community identity as an important factor in helping Gullah thrive against those fears of its extinction. It must be remembered that according to the social history of the region, there was never any time when (basilectal) Gullah was spoken universally by (descendants of) Africans in the region, let alone in the United States or North America, contrary to suggestions in some of the literature that have hypothesized debasilectalization to account for the development of AAVE. One may also wonder whether group identity is the only reason why Gullah has survived although it is spoken by a tiny minority (estimated under 500,000 speakers) relative to the general American population. Group identity is an important reason to the extent that it has helped other varieties stay alive despite the fact that they are also stigmatized. For instance American Southern English as spoken by European American Southerners has continued to be spoken in the South, its homeland, and does not seem to be endangered at all, despite all the derisions it has suffered from American Northerners and dedicated attempts in the school system to eradicate it. So it is with Appalachian and Ozark Englishes, and a host of other varieties. Stigmatization can work only if the community's sense of identity has been eroded and this is more likely to happen if a non-colonized population has been relocated. Gullah speakers stop speaking this vernacular when they are away from home (such as in the city or in the armed forces), especially when they interact with people of different backgrounds, but not when they are at home. One observes this behavior also in many American Southern Whites who have relocated from the South or evolve in social enclaves, such as a university campus faculty circle in the South, where Southern speech is not the norm or competes with one from outside the area. In addition to the above, however, geographical and social isolation has contributed to helping Gullah, like other stigmatized varieties, survive any outside pressure for change. Pressure to assimilate to majority groups is non-existent for those who do not venture outside "the community." This is actually one of the reasons why American Southern English still thrives, to the point where children of in-migrants from the North learn to speak it, to assimilate with their new peers, much to the horror of their parents. In the case of Gullah, despite population movements since WWII, there has always been a critical mass of the population that has remained well rooted in coastal South Carolina and Georgia that has maintained it as their vernacular and has seen no reason to shift to an alternative. As they

A Case Study: The Ecology of Gullah's Survival have remained socially segregated from the more affluent in-migrants from the hinterland and have interacted no more regularly with the guests in the coastal vacation resorts than maids in hotels have with the rest of us when we travel, they have been little influenced if at all by the newcomers' varieties. Racial differences, which still bear significantly on patterns of socialization in the United States, have fostered rather distant social relations between Gullah speakers, who are overwhelmingly African Americans, and the in-migrants, who are predominantly European Americans (as explained below). This is a different contact situation from that on Ocracoke Island, whose traditional residents are almost exclusively European Americans. On the latter island, race has not been an issue and there have been inter-group marriages between the locals and the newcomers, despite class differences. So, contrary to the literature on "decreolization," Gullah has not changed any more, or faster, than any other language variety that has coexisted with another, more affluent one, in the same geographical area but whose speakers have not been integrated in and/or assimilated to it. A case in point is that of some Old Order Amish communities (such as in Holmes County, Ohio), which have not yet given up Pennsylvania Dutch, despite business relations that their speakers have maintained with the neighboring dominant European American communities. As Keiser (2003) points out, a great deal depends on the critical mass of each community and whether they have a sustainable economic system, traditional or new, that is independent enough from that of the dominant community. Thus, for instance, the circumstances for the erosion of Sutherland Gaelic (Dorian 1983) are not identical to those surrounding Gullah. On the Sea Islands of Georgia and South Carolina, the traditional African American and the in-migrant communities are separated not only by highways passing between them, but also by stretches of undeveloped land, the grocery store on the highway being the meeting point (see below). Patterns of social interaction more or less replicate those of American mainland cities. For instance, the brief exchanges between, on the one hand, the African American cashier or bagger and, on the other, the European American customers do not prompt the former to modify their vernaculars when they return home. The traditional residents and the newcomers seldom socialize together. One hardly sees them in the same "joints" or social clubs, and they barely talk to each other in the grocery store, not any more than strangers do in city grocery stores. It is a mistake to expect every stigmatized population to change even when it only shares geographical space with a non- or less-stigmatized population but does not socialize with it. Curiously, since DeCamp's (1971) conjecture, the literature on "decreolization" has typically also invoked migrations of (the more affluent) outsiders into "the community," tourists' traffic, the media, and the school system as factors bringing about the putative disappearance of Gullah. I will address these reasons one by one, starting with in-migrations. While the newcomers, mostly European Americans, may have absorbed some of the local Whites (who may also be Gullah speakers) among them, they have usually settled in parts

279

280

Language Evolution that had not been (heavily) inhabited and have then formed communities that remain separate from where the "Native" populations have lived to this day. The principal meeting point between them and the latter has typically been the grocery store. Otherwise, their social networks of interaction have remained separate. As noted above, such ethnographic conditions make it difficult to imagine why anybody would expect the presence of the in-migrants in the area to influence local speech other than by adding to intra-local variation. That is, one may hear more varieties of English now in this part of the United States; while the addition of the new ones does not necessarily lead to the loss of the earlier ones. Linguistic features do not spread like cold germs; they require more intimate and/or regular interaction between carriers and non-carriers. Note, incidentally, how little, if at all, American Southern English has changed despite the influx of Northerners to the South. Although one may identify communities, such as in coastal Florida, where Northern linguistic norms have become prominent, these are novel ones. In several southern cities, the local norm has not quite been displaced, nor are there any good reasons for assuming it to be threatened under the present conditions of the coexistence of dialects. To be sure, in smaller cities, the newcomers' speech is tolerated, not necessarily emulated; hence it does not threaten the local norm. As for the presence of tourists, one wonders how many times they stop to interact with the local populations in the first place. Many of the presumed interactions between the locals and tourists on the resorts of South Carolina and Georgia are associated with those who work there. Typically they are maids and janitors; managerial positions are generally assumed by European Americans, from any part of the United States. Contact has also been associated with the basket ladies who sell their products along the highway or at the market points at the resorts. We may start by asking ourselves whether the few occasional exchanges of greetings we have had with maids in hotels are likely to influence local speech. As for the basket ladies, first, only a small proportion of the population is engaged in basket-weaving. Second, the few markets at the resorts are from a linguistic point of view reminiscent of those at the Charleston market. Here, the basket ladies, who are little-educated and not necessarily fluent in the middle-class variety, use their native vernacular to speak among themselves and an approximation of the variety spoken by populations of a comparable socioeconomic class in the city to speak to tourists. This is somewhat reminiscent of diglossia in many Third World countries, where one variety (often a different language) is used to accommodate the stranger and another is reserved for communication among the locals, as discussed in Chapters 12 and 13. We may call it "bidialectalism" or "bilectalism" in the case of Gullah; it may be bilingualism elsewhere. The names for this ability to manipulate more than one system are socially motivated; the linguistic strategies are of the same kind. The real question is when such alternations are likely to produce change. They do not seem to have had such an effect on Gullah, based on the diachronic studies of Mille (1990) and Mufwene (1991a, 1994).

A Case Study: The Ecology of Gullah's Survival On the Sea Islands of South Carolina and Georgia, the location of the vacation resorts is especially interesting, as they have typically been developed at the beach, isolated from the traditional residential neighborhoods, which are far from the beach. The highways leading to the resorts do not cross the residential communities; they bypass them. Few tourists stop to buy baskets on the highway and the locals need a pass to access the resorts. Why change should be expected when the tourists are not absorbed in the local communities or vice versa is quite intriguing. I once asked a traditional local resident working at one of the resorts whether, after work and back home, he ever felt compelled to talk like his patrons during the day. Thinking I was a fool, he asked me what the motivation would be for this shift at home. He wanted to talk like everybody else at home. The reply is also akin to what I learned from a high-school senior who would not speak Gullah in my presence but spoke it freely in my absence, not knowing at that time I had left my tape-recorder running when I left the room. Confronted with the evidence that he could speak it and asked why he would not stop speaking it altogether if he was embarrassed to use it in the company of outsiders, he said that he would otherwise be treated as a snob in "the community" and would not have many friends (Mufwene 1988b,1993). Thus, stigmatizing a language variety does not seem to automatically augur its death. There is no documented evidence of unequivocal linguistic influence on the structure of Gullah from tourists or in-migrants. Most of what has been associated with its debasilectalization, such as the putative replacement of /hana/ by /yO:l/ for the SECOND PERSON PLURAL pronoun or the use of gender-based THIRD PERSON SINGULAR pronouns he and she in alternation to a genderless subsystem of pronouns (in which im is used indiscriminately to refer to males, females, and non-humans), cannot be proved without a careful diachronically based comparison with neighboring nonstandard varieties of English which developed around the same time and from similar inputs. History does not suggest that those who made Gullah would have selected only features distinguishing them from their neighbors, without variation among them. No evidence of influence from outside the South has ever been invoked. These remarks are somewhat similar to Labov's (1963) observations on language change on Martha's Vineyard Island: not only does the centralization of /ay/ and /aw/ mark some ancestral islanders off from outsiders and reveal their loyalty to their heritage, but the features themselves are also among those that the island used to share with neighboring mainland communities. In a way Wolfram and Schilling-Estes' (1995) study confirms the relevance of social factors such as INTEGRATION and RECEPTIVENESS TO OUTSIDERS: the Ocracoke brogue is dying because there has been less resistance to the more numerous in-migrants and the latter have in turn absorbed the ancestral islanders among them. Among the new pronunciations of/ay/, traditionally centralized on the island, is the monophthongized [a:], which is common in southeastern mainland communities. These developments show that one must pay careful attention to the ethnographic ecology of language use in order to

281

282

Language Evolution understand its fate. So far nothing guarantees that the ethnographic ecologies have been identical for all islands to which mainlanders have migrated, or that what happened in one must have been replicated in the others too. The role of the media and the school system seems to have been misinterpreted. Because the locals can watch, listen to, and understand programs on TV and the radio, as well as read articles in newspapers, there is no doubt that they have developed competence in the variety used in the media, at least enough to interpret what they perceive. It is also true that the more educated locals can speak (some approximation of) that variety. Does this mean that such speakers also speak it as their vernacular while continuing to reside in the Gullah community? My field observations do not suggest this. I have often noticed that those who wanted to conceal their Gullah speech from me would speak it when they were not aware of my presence, or forgot that I may be listening, or simply did not mind me any more. I have seen people watch TV and then turn around to comment on it in their traditional vernacular, either what we linguists identify as Gullah (some basilectal variety) or something closely related to it that we would characterize as mesolectal. But then we must remember that history does not suggest that there was ever any time when all members of the community spoke our academic stereotype of Gullah. One must also wonder why the media would have produced in the Gullah community what they have failed to do in the rest of the South, where the population, in majority European American, has had earlier and greater access to the media (Mufwene 1987). We may also ask ourselves whether our ability to read English varieties that are rated higher ethnographically, or any other scholarly language, has led us to speak the same written varieties. What motivation would a non-colonized native speaker have to shift to a vernacular that is external to their domestic environment?8 In sum, the history of English in North America has been marked with assumptions and expectations that do not seem to be justified. Several questions of language evolution have too often been approached as if the ecology in which a language variety has been used was irrelevant to its fate. In the case of Gullah, the reality of glass-wall isolation, which has replaced that of geographical isolation, has been overlooked too easily because of land developments which have brought newcomers that tend to become demographic majorities to the coastal areas of South Carolina and Georgia. Patterns of social interaction have not been taken into account. Persistent neglect of the fact that even speakers of stigmatized varieties have a sense of linguistic/community identity and loyalty to their culture have too easily led linguists to speculate that Gullah speakers must feel some social pressure to shift to "mainstream" American values. We have simply overlooked the fact that humans have adaptive skills that enable them to alternate between two or more systems, according to some division-of-labor norms that they devise. (See also Poplack and Tagliamonte 1991 for a similar argument.)

A Case Study: The Ecology of Gullah's Survival Most of the above discussion has explained why the structure of Gullah is not being threatened by the other systems that its speakers have come in contact with. It says nothing against the fact that Gullah may be endangered by the economic system. Emigration of coastal African Americans to cities in South Carolina, Georgia, and outside these states, e.g., New York and Washington, DC, where they are under real pressure to assimilate to the local population and to avoid being stigmatized, reduces the total population of speakers. Not taking into account maid and janitorial service in the neighboring resorts, lack of any kind of industry that would motivate the more economically ambitious young islanders to remain in the area erodes the reproduction of the language, as the number of potential new speakers may not grow, nor be sustained, in some of the communities. The depopulation of some islands, such as Cumberland in Georgia and Daufuskie in South Carolina,9 is also a more serious threat than the in-migration of continental Americans to, for instance, Hilton Head in South Carolina or St. Simon Island in Georgia. In the former case, the islanders who migrate to the city succumb to fear of stigmatization and adopt the majority's norms. In the latter case, the locals sometimes feel more pressure to assert their own identity and to speak in ways that will continue to distinguish them from the newcomers. This is happening on Hilton Head, where Gullah is attested with more and more basilectal features (Tometro Hopkins, p.c., 1988). The more Gullah seems to be an underground phenomenon, the more resilient it may be to outside pressure. So the real threat to this vernacular seems to be potential demographic erosion of its speakers. One of the factors favoring its survival now is the fact that its speakers are not integrated in the population of the more affluent newcomers.

14.4

Conclusions

To sum up, in the case of Gullah, one must distinguish the question of debasilectalization from that of endangerment. Although the former process may contribute to the loss of a language by transformation rather than by shift (Mufwene 2005a), there is no evidence for it, any more than in several other nonstandard varieties of English in North America. The de facto segregation of ethnic groups (as highlighted especially by residential patterns) seems to have helped foster a sense of linguistic and ethnic identity that has contributed to the maintenance of Gullah's structural features. However, endangerment cannot be denied by fiat, even if Gullah's death may not be predicted with certainty. Economic rather than traditional sociolinguistic factors seem to bear the most critically on this condition. They will largely determine how much longer Gullah will be spoken, if it is moribund at all. Overall, what takes place in one community need not be expected to happen in another, regardless of apparent similarities such as between, on the one hand, North Carolina's Ocrakoke Island and, on the other, Georgia's

283

284

Language Evolution and South Carolina's Sea Islands, on which Gullah-speaking populations still live in segregated communities. Differences in ethnographic ecological conditions are likely to produce different outcomes of events that seem similar only on the surface, in this case migrations of outsiders to island communities.

Notes Chapter 2 1. I am grateful to David Hull for feedback on an earlier version of this essay. We disagree on whether viruses can be grouped into species, on a par with animal species. I see no reason to change my position, as I also assume that the evolutionary peculiarities of various species are largely determined by their respective ontogenetic properties and their particular modes of gene or feature transmission. I am solely responsible for the shortcomings of the positions submitted here. 2. Paul (1891, Chapters 1 and 2) says something that amounts to this position. For him languages are species and communal norms are no more than averages across convergent idiolects. 3. Assuming mistakenly that Creoles have evolved from pidgins, Bickerton has argued since his Roots of Language (1981) that pidgins have no syntax or grammar, having been made by adults, who presumably no longer had (full) access to Universal Grammar (UG) or the biological endowment for language. Accordingly, children, guided predominantly by this language organ, would have imposed a UG-based grammar onto their parents' syntax-less pidgins, which would account for world-wide structural similarities among them. The same UG must account for the putative absence of syntax in the protolanguage spoken by Homo erectus from about 1.5 million years ago to about 200,000 years ago. He assumes that its protosyntax must have been similar to that of child language (by two years of age), which allegedly shares features with that of incipient pidgins. (Anybody who pays attention to cross-creole structural variation would speak of Creole grammars, in the plural, instead!) Every one of these claims is highly disputable, supported by tenuous evidence, but it would be digressive to discuss this issue here. See, Chapter 4, Mufwene (1991c), Lieberman (1991, 2002), McNeill (1992), and Pinker (1992) for informative discussions of Bickerton's thesis as well as Mufwene (200la) and Chaudenson (2001) for alternative interpretations of the facts that constitute the sociohistorical ecologies of the evolution of Creoles. DeGraff (1999a, 1999b, 2005) is equally informative, from a language acquisition perspective. Suffice it to note here that Bickerton (1984a: 141,157-8) also suggested that Creoles instantiate languages in a primitive developmental stage. 4. Linguists such as Franz Bopp and August Schleicher seem, however, to have used the term organism with the meaning of species, a term that also occurs in the work of Schleicher. They speak of inter-individual, now known also as interidiolectal, variation within a language, to which Darwin's evolutionary principle of natural selection could also be invoked to apply (Schleicher 1863). As observed in note 2, one important nineteenth-century linguist whose views foreshadowed some of those expressed in this Chapter is Paul (1891). 5. Even Chomsky (2000) now adheres to this particular view of a language, explaining that lumping idiolects into the same language is a matter of likeness, not of sameness or identity. On the non-monolithic nature of language

286

Notes architecture - i.e., how the different structures coexist, with the functions of some of them overlapping or sometimes conflicting with each other - see Mufwene (1992a). As explained in Mufwene (200la), the piecemeal way in which a language is "acquired," with features selected incrementally from diverse inputs/idiolects, accounts at least partly for this state of affairs. 6. Similar ideas can be seen in Paul (1891, Chapters 1 and 2). 7. As Thomason (2001) points out, the common characterization of language death that I just repeated may be oversimplified, because the lone speaker would not have anybody else to speak it with and his/her knowledge would probably have fallen into attrition by then. One sense in which the characterization may be correct is when a language is thought of as knowledge rather than as practice. Then arises the question of the death of a language by transformation. What proportion of its original features must it preserve in the last speaker or in the last group of speakers in order to be considered the same and alive? We return to these issues in Part 3. 8. Those who are not impressed by the distinction I make between evolution of language and language evolution should note that in biology the term evolution is also ambiguous between the two senses that I highlight here. The advantages that the term evolution offers over the traditional term change in broadening the scope of genetic linguistics with topics which are germane to language speciation outweigh the inconvenience of an ambiguity that is easily resolved by the context of the discussion. 9. One clarification is needed here. Although the number of variants at the population level is a set-theory union of variants at the level of individuals, the strengths of the variants at the population level are a function of population structure and different kinds of dynamics of interaction and influences among individual speakers. This is also the level where it becomes more obvious that selection does not necessarily eliminate disadvantageous variants. As observed by Kretzschmar & Tamasi (2003) in American dialectology, populations have long memories of variation. Even the most marginal features can be resilient. 10. This otherwise useful statement is problematic. The contents of UG - that part of human predisposition to cognition that putatively specializes in the development of linguistic and communicative competence - has not been explicitly articulated, at least not in terms of how what it is claimed to do cannot be accomplished by a general learning module of the mind. It is not evident either that, while he/she develops competence in a particular language, the learner really purports to develop a "system," as systematic as he/she wants to be. One can also argue that a "system" simply emerges to the analyst - or is projected by him/her - out of the routines that an individual develops while purporting to communicate with particular sequences of spoken or signed gestures. Fortunately, these issues do not bear (significantly) on the main theses of this Chapter. 11. Clearly, languages and idiolects cannot have genotypes, because they are not biological systems. What is meant here is that they have nothing similar to a genotype in a biological organism or species. I also deliberately ignore here the fact that there is no isomorphism between phenotypes and genotypes, as well as the fact that ecology plays an important role in determining the phenotypes of a population. It is nonetheless true that only some genotypes can be associated with particular phenotypes, for instance the kinds of noses, hairs, lips, ears, limb morphology, and complexion that are found only among humans but not among other mammals. 12. I speak of "copying" tongue in cheek here. As Sperber (1996:141) observes, "1) De facon generate, les representations ne sont pas copiees, elles sont transformers dans le processus de la transmission. 2) Les representations se transforment par 1'effet d'un processus cognitif constructif." That is," 1) Generally speaking, representations are not copied, they are transformed during the process of transmission. 2) The representations are transformed by a constructive cognitive process." David Hull (p.c., 2004) remarks that selection as an evolutionary mechanism is restricted to vertical, not to horizontal, transmission. I wonder whether the difference does not lie in the complexity of the

Notes process, rather than in whether or not selection applies. After all, it is the outcome that shows whether a gene or feature has been preserved in a population, not how it has been preserved or eliminated. 13. These remarks do not deny the fact that adults continue to exert some linguistic influence on children. As Zuzana Tomkova has reminded me (p.c., 7 February, 2007), children who interact more often with adults than with other children tend to have some linguistic mannerisms that are not so typical of children. What I want to point out is that once children start socializing also outside the family, especially with other children, they become more selective and align themselves more with their peers than with others. 14. The process is very similar, if not identical with, what Sperber (1996:147) identifies as "synthesis." The bearer of, for instance, an idea or a story in the case of culture, or of a feature or even an idiolect in the case of language, receives input from several individuals, some of whom influence him/her more than others. The inputs contribute selectively and differentially to the formation of the new version of the original idea or story, or to the formation of the new speaker's feature or idiolect. 15. We should remember here that selection does not necessarily operate in terms of total exclusion; it operates in terms of dominance. As explained by Sober (1984), it works more like in a golf tournament, in which scores are added up for every player. While the winner is the one with the best score, the other players are still part of the overall competition and each occupies the position determined by their respective scores. The competition of linguistic features proceeds more or less the same way, making allowance for those features that are not dominant to coexist with the dominant ones in particular idiolects and to be used every now and then. Because non-dominant features in some speakers' idiolects can be dominant in other idiolects, languages can have longer memories than an oversimplified approach to language evolution, especially one conceiving of languages as organisms, may lead us to expect. The lives of linguistic features are thus to some extent comparable to those of genes, subject to mating/interaction practices. 16. Following Dawkins (1976), Hull (1988) opts for the term replicator to identify units of selection, with basically the same meaning as what I express here with linguistic feature, viz., units and principles that the learner of a language, in the present case, would endeavor to replicate. Croft (2000) follows both, alternating between replicator and lingueme "linguistic meme." I am sticking to the traditional term in sociolinguistics, because the others represent no improvements over it. They merely identify features as the elements that can be "replicated," albeit imperfectly which is just a difference of perspective. Actually, the term lingueme conveys as much vagueness and inaccuracy as meme, especially because a language is not a body of utterances or texts (pace Croft) but rather the "system" that produces them. 17. Although Lewontin (1970: 14) makes allowance for populations to be selected, his primary position is that "many [population] adaptations turn out to be explicable by simple selection at the individual level" (13). This takes us back to idiolects and the piecemeal pattern of their emergence, through the cumulative selection of individual features and their (re)integration into a new "system." This process justifies the position I defend here. Another way of explaining it is that particular units, such as words (which also implies sounds and morphemes), and particular combinatoric rules (such as in syntax) are associated with specific languages. In practice, however, it is individual units and combinatoric rules which are perceived and can be selected in or out of the emergent idiolects. In multilingual or multidialectal contexts, units and rules from different systems (languages or dialects) are often mixed and can lead to the emergence of significantly different language varieties, such as the Romance languages (in relation to Latin and the Celtic substrate languages) or Creoles (in relation to Western European and African languages in European plantation settlement colonies of the New World and Indian Ocean). As should also be obvious from the literature on code-switching or mixing, what language or dialect a speaker claims to be speaking is sometimes determined more by the speaker's intention than by the actual text of his/her discourse.

287

288

Notes 18. The terms are convenient misnomers because, in the case of lexifier, it is inconceivable that one would naturally target only the vocabulary of a language and ignore the grammar associated with the words, thus hoping to get it from some other source (Chaudenson 2001,2003; Goodman 1999). The terms superstratum and substratum are also inaccurate in the case of Creoles because they are based on the social stratification of the populations in contact but not on the time their languages arrived in the settings where Creoles developed (Goodman 1993). Technically, in parts of the New World and the Indian Ocean where Creoles developed, the relevant African languages arrived later than the European languages that the slaves appropriated as vernaculars and modified. 19. This phenomenon is a consequence of both polyploidy and the fact that populations meet not like armies or sport teams engaging in contact at the same time and in an organized way, but rather as individuals engaging in particularistic contacts at different times and in different settings, all of them having the potential to influence the development of each relevant speaker's idiolect in some respect. How competition and selection of features apply within a population, under specific constraints determined by the relevant ecologies, accounts for the specific kinds of mix and heterogeneity that distinguish one language or dialect from another. Although the transmission mechanisms by which a viral species can display such genetic mixing and heterogeneity are not the same (see Section 2.4), the similarities in outcomes are too close to escape notice. 20. This term refers to the emergence of a new, compromise variety out of the contact of dialects of the same language or, according to Siegel (1987), languages that are closely related in the same cluster, such as the Bantu cluster of languages identified in the literature as (ethnic) Kikongo. As explained in Mufwene (2001a), they have evolved by the same competition and selection process as Creoles, not by simplification to a common denominator (pace the received doctrine). 21. I assume successful communication to be the closest counterpart of biological interbreeding. Differences between these notions are quite consistent with the fact that idiolects have no genotypes and develop gradually. An individual can learn only a limited number of linguistic features from a particular communicative event. As utterances share many features, there are many communicative events from which little or nothing new is learned, even during the early stages of the development of idiolects traditionally identified as "language acquisition." 22. This is of course in the context of species, with the process of evolution premised on variation, inheritance, and differential reproduction, although layman's language allows individuals to evolve. As explained in Mufwene (200la, Chapter 1), inheritance need not be interpreted in strictly biological terms of gene transmission. It can be extended to "information copying," as in culture and language, pace Fracchia & Lewontin (1999). 23. It would be misguided to assume that the actuation question does not apply to language or dialect choice, as indirect as this process is. It is important to examine what particular changes in the ethnographic settings of communication would have speakers prefer to use one but not (an)other language(s) or dialect(s). 24. The term applies to chain-style changes undergone by vowels in urban American English dialects. 25. Paul (1891) often invokes shift in the frequency of variants to account for language change. It is curious why this factor has figured so little in twentieth-century qualitative historical linguistics. 26. I borrow this term from Goodnight & Wade (1999) and Wimsatt (1999), though they have not articulated its meaning quite explicitly. It is clearly used in reference to the internal social organization of a population, as suggested by another term used by Goodnight & Wade: population subdivision. I use it below for a variety of social factors, including how a population is variously sub-categorized (often in overlapping terms, such as by gender, age, race, education, profession, and socioeconomic class), how it is stratified, and who interacts or socializes with whom. My working assumption is that linguistic features spread along speakers' patterns of socialization. Needless to say that I am rearticulating, from a population perspective, parameters that sociolinguists have traditionally invoked to

Notes account for language variation. My primary intention is to highlight not only ecological similarities between the linguistic and viral species but also methodological ones between linguistics and biology (too often ignored in traditional, twentieth-century genetic linguistics) that must be used to advantage in order to better understand certain general mechanisms of evolution. 27. As explained in Chapters 10 and 11, the reasons for the extinction of the African languages in the plantation settlement colonies of the New World and Indian Ocean are not identical to those for the extinction of the indigenous languages, although the institution of the new socioeconomic world order by European settlers was a major stochastic event in both cases. The particular way in which the plantation societies developed played a central role in an evolution that could otherwise be described as "chaotic" (from the perspective of chaos theory). They started from homesteads on which the Africans were minorities, integrated, and their children acquired colonial varieties of European languages as their primary, if not exclusive, vernaculars. These Creole, locally born slaves would become cultural, and especially linguistic, models for the Bozal, incoming African-born slaves of the larger plantation societies, in which African languages were underrated. Societal multilingualism among the slaves just compounded the problem, leading to the abandonment of African languages in the same way that many of them are now losing ground to urban vernaculars in African cities. On the other hand, most Native American languages would be endangered much later by European languages, when European Americans claimed more of their ancestral lands and forced them to relocate to places where it was not economically rewarding to evolve in the indigenous languages. The lure of the young to city life, which requires communication in the European vernacular, contributed significantly to the decline of the indigenous languages. 28. In the case of the United States, segregation by nationality among the European colonists during the colonial period must have reduced, by retardation, influence from continental European languages. Gradual language shifts from continental European languages must have taken place after the American English koines had "crystallized" primarily from the contact of metropolitan English dialects, in the same way as on the Falkland Islands. Continental European linguistic influence must have been kept in check in the same way that influence from the xenolectal speech of today's waves of immigrants is. By the Founder Principle, children acquire the extant variety natively and their parents die with their xenolectal features identified as such and largely not accepted by the host populations. 29. Labov (2001: 514) captures something similar with his "Golden Age Principle," according to which "At some time in the past, language was in a state of perfection," consistent with the common reaction among purists that one's language is degenerating or decaying. 30. Saussure (1916) may have been a little mistaken in analogizing a language, interpreted as a system, to a chess game, in which players follow communal conventions. Regarding languages, the putative conventions are more like emergent patterns (i.e., partial regularities emerging to the analyst) than something whose existence is independent of the speaker's knowledge. Those partial regularities do not entail that a person who "acquires" a language naturalistically learns particular rules. Linguistic regularities are like beauty, in the eyes of the beholder. As McCawley (1976) points out, it is not evident that the generality of the rules by which speakers putatively operate coincides with those that the linguist seeks or claims to capture as "significant." This observation follows from the fact that a speaker or a naturalistic language learner is not a linguist. He/she does not process the materials from which his/her competence gradually develops in the same way as the linguist who gathers a (significant) corpus of (varying) data, analyzes the corpus, and seeks to capture the broadest generalization(s) applicable to the data that interest him/her. 31. This question is connected to the disputable assumption among linguists that native speakers of the same variety use the same grammar (identically represented in all their minds) to process and produce utterances (Mufwene 1992a).

289

290

Notes 32. This social attitude accounts for the resilience of AAVE, as stigmatized as it is, because it also functions as an identity marker. It also accounts for why the Ocracoke brogue, spoken by White islanders on the coast of North Carolina, is endangered by the vernaculars of White in-migrants from the neighboring mainland (Wolfram & Schilling-Estes 1995), whereas Gullah, the Creole vernacular spoken by African Americans in coastal South Carolina, has so far been endangered by the exodus of its speakers, not by the White mainlanders' "invasion" (Chapter 14). White mainlanders have mixed with White islanders but not with coastal African Americans. In coastal South Carolina and Georgia, the residential communities remain racially segregated, just like parts of many American cities. 33. By puberty, the cognitive infrastructure that enables this linguistic development is assumed to have reached an acceptable level of maturity, and the "linguistic system" - as inferred by the linguist - is considered full-fledged, although the vocabulary will continue to grow and more stylistic variation will emerge. 34. The physiology of a speaker is part of the immediate external ecology of language "acquisition." The variation alluded to here has to do with factors such as the following: specific shape and size of the oral cavity, length and width of the throat, size and length of the nasal cavity, alignment of teeth, and width and thickness of the tongue. 35. This is a position consistent with part of the conclusions of Hull (2002: 27), when he states, "Of course the flow of information may just as well go in the other direction. Biologists may find the solutions that linguists have suggested useful in biology."

Chapter 3 1. I am grateful to Peter Bakker for helpful comments on a distant ancestor of this Chapter, especially for recommending it for publication in the Journal of Language Contact despite our fundamental differences on whether or not Creoles are genetically or evolutionarily unique. I hope that, overall, this book proves my position that they are not. 2. DeGraff (2003,2005) provides a detailed critique of the "exceptionalism" paradigm. 3. Readers need not be distracted here by the question of whether these new vernaculars are dialects of their lexifiers or separate languages. This is a question that cannot be settled with arguments based on structural features. There are no particular features that can help linguists tell languages from dialects. What matters here is whether or not these vernaculars have evolved by the same evolutionary processes observable in other language varieties and whether the ecological conditions that are relevant, especially those having to do with population contacts, also apply to the evolution of other languages (Mufwene 2007). 4. I return below to the distinction between "endogenous" and "exogenous" evolutions, proposed by Chaudenson (1992,2001,2003), which some may consider relevant here. 5. Campbell (1998: 286-7) supports this distinction, considering as internally motivated those changes that are caused by speakers' physiological and perceptual peculiarities. Actually, such factors which shape particular language varieties are external to language. They add to the complexity of external factors that influence the evolution of a language. They too have contributed to the emergence of Creoles. 6. It is absolutely important to conceive of "competition" here relative to a speech, rather than to a language, community, because the competing elements can amount not only to regional and/or social varieties but also to different languages used by the same population. This perspective bears on the subject of language endangerment and loss to which I return briefly below but elaborate in Part 3. The meanings of the terms competition and selection are explained in Chapter 7. They have to do with the differential way the competing elements are weighted by their users. 7. But even such changes can spread in the communal language only through inter-idiolectal contacts. In some communities, especially colonial ones, some of the influential idiolects are xenolectal, associated with L2 "acquisition "

Notes 8. The folk term for the vernaculars in the relevant communities has been Patois (now nativized as Patwa) in Anglophone and Francophone territories. It has borne more or less the same meaning as during the colonial period and before, viz., "rural or provincial, especially nonstandard variety of a language, usually derided by speakers of the urban or standard variety as bad and/or unintelligible." For discussions of the term Creole and how linguists extrapolated it, see Chaudenson (1992,2001,2003) and Mufwene (1997a), as well as the sources cited in both publications. 9. Tremblay (2005) criticizes Meillet, however, for not taking this evidence to its normal conclusion, viz. to discourage genetic linguists from relying excessively on the comparative method. 10. Interestingly, the structures of these Creoles appear to be quite close to those of Gullah and thus not as divergent from their other nonstandard colonial kin as the basilects of Guyanese and Jamaican Creoles, for instance. This phenomenon raises interesting questions on the progression of the geographical continuum that Alleyne (1980) posited for English Creoles as extending from Suriname to North America. Note that the problem arises only if one assumes, as in the relevant literature on "decreolization," that the continuum is linear, with basilectal and acrolectal poles, with partial implicational scales of structural features. I disagree with that particular view of the continuum, according to which some Creoles are allegedly more basilectal or conservative than others, as if there had ever arisen a common basilect once spoken everywhere in territories where Creoles have evolved from the same "lexifier" (Mufwene 1994, 2001a). See also note 14. 11. Note that this factor accounts negatively for why AAVE, unlike Gullah, is not significantly different from American White Southern English. As noted above, segregation in the hinterland of the American South was institutionalized only in the late nineteenth century, with the passage of the Jim Crow laws, almost two and a half centuries after (former) slaves and (former) White indentured servants had lived intimately and developed a common American Southern English (Schneider 1995). Bailey & Thomas (1998) are right in situating the divergence of White Southern and AAVE in the late nineteenth to twentieth centuries. 12. One must also note that there are other colonies, such as Madeira and the Canary Islands, which also engaged quite early (in the fifteenth century) in sugarcane cultivation and slavery but did not produce Creoles either. Size cannot be such an important factor, because of what happened in Cape Verde and other small islands or archipelagos. The industry did not last long in these colonies (Schwartz 1985), although it is not clear whether they had population structures similar to that of Brazil. On the other hand, it is noteworthy that the structures of Cape Verdean Crioulo are said to be less divergent from those of Portuguese than those of Sao Tomense and Principense in the Bight of Biafra. It has also been pointed out that the tense-aspect and number marking systems of Papiamentu are not exactly like those that are attested in most Creoles (Andersen 1990). Could it be that these Creoles developed primarily from the varieties of permanent Creole populations that spoke what Chaudenson identifies as "approximations" of the European language and underwent less influence from Bozal speech than elsewhere? 13. We need not assume that the Europeans spoke to the slaves like to children (Adam 1883), but it is quite conceivable that the Europeans became more tolerant of deviations from their languages and reproduced some of them in their own speech. Recall that they were themselves developing colonial koines out of the metropolitan varieties they had brought with them. 14. As explained in Mufwene (2001a, 2005a), there cannot really be a common yardstick for measuring degrees of divergence, because the "lexifier" was hardly the same from one colony to another. Assuming that the European languages were internally variable, the different colonies hardly constituted identical representations of metropolitan variation patterns, especially in terms of the demographic strengths of the different variants. From an evolutionary perspective, this cross-setting variation entailed different dynamics of coexistence, different markedness values associated with the variants, and thus the possibility that different variants could be (dis)favored in different settings. As in

291

292

Notes much of the literature, the present discussion oversimplifies a colonial situation in which metropolitan languages evolved differentially from one setting to another. Besides, the evolutions were primarily local (Kretzschmar 2002); one must indeed remember that polities are in effect "metapopulations" of localities interconnected by a transportation infrastructure (traditionally roads and rivers) and by dispersing individuals (Hanski 1996). 15. This is an improvement over what he originally misformulated as "generations of Creoles" on the models of generations of computers in Chaudenson (1979, 1992, 2001). The colonial language variety brought to a colony from an earlier colony need not have been a Creole already. In some cases, such as when Mauritius was settled from Reunion or Suriname and Jamaica from Barbados, it is doubtful that a Creole had already emerged in the mother colony. What matters the most in this particular case is the fact that second-generation colonies did not start from the same kind of language contact situation as their first-generation counterparts. The slave populations from the earlier colonies already spoke some variety of the colonial language and provided a less diffuse linguistic infrastructure from which the Creole would develop, perhaps faster. 16. Readers more interested in an ecological approach to the development of "colonial Englishes" should read Schneider (2007). 17. As explained in Chapter 10, the experiences of societal multilingualism and language shift in Hawaii, the New World, and the Indian Ocean were not identical. The concurrent emergence of a Creole population that never spoke a pidgin or a creole and served as model to the Bozal slaves contributed to the continuous obliteration of ethnic distinctions, even during the post-Emancipation importation of contract laborers from regions of Africa and India that were ethnolinguistically less diverse (Mufwene 2005a). This is another interesting manifestation of the Founder Principle, showing how some of the practices of the founder population can have far-reaching consequences generations later. 18. However, as Hookoomsing (2000) points out, in the case of Mauritius, this is not so true of French colonization. The Dutch had discovered the island in 1598, colonized it for the first time in 1638, and then finally abandoned it in 1710 to pirates (Encyclopedia Britannica, electronic edition, 2000), leaving some Maroon slaves behind (Hookoomsing 2000:149). It is not clear what language varieties those Maroons spoke, if they survived the pirates and/or the island's colonization by the French in 1721. If they did, what impact did they exert on the development of the French colony and of the new colonial vernacular that was to become Mauritian Creole? 19. I submit that languages are quite osmotic (see also Chaudenson 1992, 2001) and their boundaries are less rigidly defined than linguists have suggested. Although some institutions try to erect official "language borders" and impose themselves as immigration and/or customs officers (as in the case of the Academic Fran$aise), the average speaker just ignores these artifacts and feels free to let xenolectal elements into their language variety. Even if they subscribe ideologically to a particular language they intend to speak, the evidence of varieties derided asfranglais or Spanglish suggests that practice holds "open borders" and is more open to cross-language "migrations." Such varieties represent, in the extreme, how competition and selection work in multilingual communities. Needless to say, there are ecologies, such as those where Creoles developed, which strongly favor one particular language. 20. As explained in Chapter 11, the colonization of much of Europe (and of the world east and south of the Mediterranean) by the Romans was not really on the settlement model. It was perhaps closer to the exploitation model that applied to Asia and Africa, although the Romans were more interested in benefiting economically from the imperial enterprise than in claiming political and economic control rights over every square inch of every colony. Rather, they developed networks of city colonies administered by Romanized indigenous aristocrats, whom they rewarded well. This all shows that language dispersal does not proceed everywhere through a single, uniform population structure, suggesting that even today's pidgin and indigenized varieties of European languages may very well count as new offspring of Indo-European languages.

Notes 21. My use of the term adstrate here is just an unfortunate legacy of creolistics that could be done away with. It continues to suggest that substrate influence is from populations that are ranked lower while adstrate influence is from populations that are of more or less the same social stratum. Linguistically, the influences do not work in different ways, especially in the recent European colonial settings since the fifteenth century. Here, adstrate is used, for convenience sake, in reference to other non-indigenous languages that have not been identified as "substrate." 22. Although the phenomenon has emerged in another polity, London West Indian Patwa, which is not spoken by all offspring of Caribbean immigrants (even in those cases where the parents are Patwa-speakers), is just an instantiation of this selective process. London West Indian Patwa is in fact spoken primarily by those immigrants who do not feel integrated in the British social structure and/or use language as an identity marker (Sebba 1993). 23. The combination of contact and segregation also accounts, with a dosage of their own sociohistorical "ecological" peculiarities, for the emergence of mixed language varieties such as Mitchif (from French and Creek) and Copper Island Aleut (from Russian and Aleut). See, e.g., Thomason & Kaufman (1988), Golovko & Vakhtin (1990), and Bakker (1997) for informative discussions of their developments. Geographical and social isolation is a critical ecological factor bearing on the development of their structural peculiarities. 24. This utilitarian explanation can explain, conversely, why the spread of English as a vernacular has progressed in multilingual places like Singapore: its association with the polity's economic progress, favored by the small size of its population, has made English a more useful language. However, much more remains to be learned as Cantonese seems to have slowed down a similar spread of English in Hong Kong. Despite some general similarities, there are always local ecological specificities that must be factored in the language evolution equation.

Chapter 4 1. An ancestor of this Chapter was circulated in volume 3 of Contemporary Linguistics (working papers of the Department of Linguistics at the University of Chicago) in 2003 under the title "Analogs anywhere: The flow of highway traffic and language evolution." I am grateful to Michel DeGraff, Alison Irvine, and Rudi Keller for feedback on drafts of that essay. 2. That paper was also circulated in the same volume of Contemporary Linguistics in which the ancestor of this Chapter first appeared. It is identified below as Elston (2003). 3. Keller uses the phrase after Smith (1776), who showed in economics how the acts of individuals vying with each other, in their own personal interests, often result unwittingly in something good for their society, such as better prices and better salaries. Thus, although the competition that drives such evolution is basically selfish, its outcomes maybe (mis) interpreted as altruistic. In some ways, language behavior is indeed individualistic, intended to serve the interests of the speaker, despite the patterns that emerge in language evolution. Like biological evolution, this one has no purpose either; it just happens. 4. To be sure, repetition is an important factor in the emergence of new patterns, especially at the level of individual speakers. As noted above, a path can well emerge from the footprints of a single pedestrian who takes the same shortcut over and over again. However, it emerges faster if several other pedestrians take the same shortcut. At the communal level, changes in a language are more manifest, even if they consist only of new (patterns of) variation, when several speakers are involved. 5. The verb exapt is used here as a backformation from Stephen Jay Gould's term exaptation, intended for those adaptations that things undergo that had nothing to do with their original designs, such as the use of vocal cords to produce speech or of hands for sign language.

293

294

Notes 6. There are indeed frequent cases when the same expression was innovated independently by different speakers isolated from each other in space and/or time. However, the process of copying, which accounts for the spread of the innovation, applies the same way in their communicative networks and through them within the larger language community. 7. I could have chosen to compare language evolution with the flow of pedestrian traffic too, but things are much more complex in this case and the rules regulating it apparently much looser, making its emergent patterns more "chaotic." It seldom proceeds in one direction and there are no rigid lanes in the pedestrian two-way traffic. There are undoubtedly interesting things to be learned from such a comparison, for instance, the fact that in the same traffic some people walk at what may be considered a normal pace (though not at exactly the same speed), some people are almost running, while others are strolling instead, making it difficult for even normal walkers following them to proceed ahead, when traffic is heavy. That kind of comparison will have to wait for another study. Readers will also notice that the comparison undertaken here has its share of limitations. One of these has to do with making mistakes. In vehicle traffic, mistakes are often crippling and sometimes fatal. There is no such corresponding price to pay for linguistic mistakes. Although I highlight some other differences more relevant to this discussion below, I focus on aspects of highway traffic that can help us articulate some aspects of the interactive dynamics of idiolects, which are characterized by Mufwene (200 la) as part of the internal ecology of a communal language considered as a species. 8. For an explicit expose of my position, see Mufwene (2001a: 1-3). A similar view is presented by Chomsky (2000). According to him, "speaking the same language is on a par with being-near or looking-like" (31). An important correlate of this is his other observation that "successful communication between Peter and Mary does not entail the existence of shared meanings or shared pronunciations in a public language" (30). Idiolects differ among themselves precisely in regard to such details as pronunciation, meanings of words or phrases, and sets of items affected by a grammatical rule. Chomsky's remark is similar to Mufwene's (1989a, 1992a) observation that speakers of a language communicate with each other not because they operate according to the same system but because they are capable of assigning to each other's utterances interpretations that are similar most of, but not all, the time. 9. This is analogous to members of the same species (especially among animals) recognizing which other individuals they can interbreed with - a criterion that keeps recurring in conventional definitions of SPECIES in biology. 10. See O'Hara (1994) for such a definition of a biological species based on claims of common ancestry rather than on sharing particular phenotypes. See Mayr (2001) about the position that a species is not an essentialist category but involves degrees of membership. Hey (2001) provides a survey of what he terms "the species problem," which need not concern us here, except to remind us that there is no unanimity on either the concept itself nor on the membership of individuals in a species. I return to the latter aspect of categorization below, just to show that the comparison is even more apt, because it raises the same population grouping problems about languages and traffic as about biological species. 11. Such comparisons are rarely perfect, of course. One cannot think of some kinds of vehicles or of some language varieties as preying on others in the way that carnivorous animals live at the expense of other animals. As I show in Part 3 of this book, the phrase killer language, so often used in the language endangerment literature, is inadequate. One should also realize that the highway is no more part of traffic than the community in which a language is spoken can be assumed to be part of it. Assuming that idiolects are to a language more or less what the individual vehicles are to a particular traffic stream, both the highway and the community are just the physical loci of the relevant species being compared here. If every kind of vehicle is a species in its own right, then traffic is a heterogeneous, multi-species population in the same way that the totality of language varieties spoken in a community can be claimed to form a linguistic population of many species, with every variety considered to be a species in its own right.

Notes 12. Despite some imperfections that have already been pointed out, and some others still to be explained below, this comparison is justified by another important reason: like life itself, species evolution does not backtrack. It is illegal to reverse, or drive backward, in one-way vehicle traffic, and the norm has been to avoid it, even at the cost of missing one's exit. 13. There is also an analog for drivers and vehicles that leave the highway traffic to refuel and/or take a break and then come back. These actions are reminiscent of speakers who leave their communities for others and return some time later. They may have lost some of their previous features or bring with them new ones that other speakers may adopt or ignore. 14. An interesting example comes to mind here. In the USA, the average White Southerner and White New Englander or Midwesterner do not speak standard English the same way. The late Professor Raven McDavid, a dialectologist, reported that while living in Chicago his colleagues sometimes noted his southernisms when he was relaxed and made no conscious effort to repress them. The southernisms showed up again when he returned to the South Carolina Piedmont to visit. 15. DeGraff (2001a, 2001b) provides a compelling refutation of Schleicher's view that languages evolve from a simple and primitive morphosyntactic type, allegedly the isolating type, to more complex and sophisticated ones. There is much more evidence of communal languages developing a simpler morphosyntax (e.g., English compared to other Germanic languages, or the Romance languages compared to Latin) than of the complexifying kind of evolution. At best, there have been evolutions in both directions, if one accepts Givon's (1971) argument that agglutinating temporal markers in the Bantu verb complex evolved from erstwhile verbs. As shown in Chapter 2, what has been assumed about Creoles as developments putatively from pidgin ancestors has turned out to be a misinterpretation of history. 16. It is not by accident that this discussion has been limited to a comparison of traffic flow and language evolution without mention of processes of "transmission" and "acquisition," relevant as these are to evolution (Mufwene 200la). It is not clear to me that there is a traffic counterpart to language "transmission" or "acquisition," although the whole process of traffic flow involves constant restructuring of its composition, as should be evident from Section 4.2.2. 17. Appel & Muysken (1987: 2) invoke "former colonial countries, where the colonizer spoke English, for instance, and the native people a local language." No such situation really obtained, because the colonizers endeavored to learn some form of the local language, which was necessary for their survival, and/or they trained colonial auxiliaries in a variety of their language so that this elite class could function as interpreters. Colonization could not have been effected without any form of communication between the colonizers and the colonized, although, more often than not, command of the others' language left much to be desired. 18. Intra-organism variation in language and in other cultural "systems" follows from the fact that they often integrate more than one feature for the same function. Some of the variants do not fit in the same paradigms, such the complementizer that and relative pronouns in English, although they both introduce subordinate clauses. As noted above, the mechanical structure of motor vehicles does not lend itself to this kind of function-sharing by their parts. 19. As explained in Chapter 3, when two or more populations come in contact, they do not behave like organized sport teams with common goals and well-orchestrated engagement rules. Members of the populations behave as individuals, with similarities in their behaviors following mostly from the common cultural backgrounds, which have not suppressed their individualities, as made evident by the maintenance of idiolectal differences. Responses to communicative pressures from the contact setting are individual, though speakers are influenced by each other. Some feel them sooner and/or more strongly than others and therefore shift earlier than others to the prevailing language or dialect. At the population level, the shift from one language to another is a gradual and often protracted process. This need not be so at the individual speakers' level.

295

296

Notes 20. It is not really clear to me that Sprachbund, or language convergence, can be analogized with traffic merger. Although languages in such contact settings become more and more similar to each other, they still maintain their individualities, at least in the minds of their speakers. Recent history of the Balkans since the collapse of the former Soviet Union has led to a reassertion of such identities (see, e.g., Friedman 1996.)

Chapter 5 1. I am grateful to Michel DeGraff, Alison Irvine, and Bertram Malle for feedback on earlier versions of this essay. All the remaining shortcomings are my sole responsibility. 2. Bickerton's notion of pidgins that bears on the present discussion is that they are syntax-less, which justifies comparing them to the protolinguistic ancestor of modern human languages. Among his central arguments is the interindividual variation observable in them. There is yet no evidence that the idiolects they consist of are not internally systematic. Neither can we overlook the inter-idiolectal variation that obtains in any language or dialect community (Paul 1891). I will assume in the rest of this Chapter that Bickerton must have had incipient pidgins in mind. 3. As shown in Mufwene (2000a, 2005a), linguists have provided no justification for applying especially the term Creole to language varieties that do not meet the necessary historical conditions, especially since there are no particular structural features nor any language restructuring processes or combination thereof that set them apart, typologically or genetically. 4. Bolton (2000, 2002) provides very informative accounts of the development of Chinese Pidgin English that is in agreement with the position submitted here. 5. Note that according to Keesing (1988), the birth place of Melanesian pidgins is not the plantations on which they have flourished. Rather, it is the whaling and trade ships on which some of the first plantation laborers had worked and spoken a proto-Melanesian Pidgin. This would account for structural similarities among Melanesian English pidgins, especially in those respects that distinguish them as a group from Atlantic Creoles. However, Baker (1993b) argues that a truer proto-Melanesian Pidgin would have evolved earlier in Australia, especially in Queensland, where several Melanesians had worked too. Moreover, Australia controlled much of that fishing and trade fleet in that part of the Pacific. The basic idea remains the same, viz. that pidgin was imported to the plantations, where it would eventually evolve into the vernaculars known today as Tok Pisin and Bislama, for example. 6. This is more or less in the spirit of Comrie (2000) and Kihm (2002), who observe that the makers of protolanguage did not have an antecedent (proto)language to derive materials from, although there must have been an earlier means of communication that would have paved the way for the evolution of the earliest linguistic systems. Jackendoff (2002) speculates that communication before protolanguage must have consisted of one-"word" signs, as among non-human primates. This still makes the case of Creoles and pidgins, which developed from fully developed languages, quite different. 7. Jackendoff (2002) comes close to this idea as he argues against "syntactocentrism" in favor of "parallel architecture" of phonology, syntax, and semantics. However, the evolutionary order he suggests on page 238 seems intuitively less plausible than MacWhinney's (2002a, 2002b) alternative, according to which syntax must have developed later than the aptitude to articulate sounds beyond vocalization (which made it possible to produce larger vocabularies) and the referential ability to use vocal symbols earlier than phonetic communication. For MacWhinney, the ability to form more words made it possible to convey more complex thoughts, which called for more complex syntax. 8. Insofar as the notion of'idiolect' is metalinguistically significant, each idiolect has a grammar to the extent that it is systematic, regardless of whether or not its "system" is identical with those of other idiolects in the relevant language

Notes variety. Like biological species, communal languages and dialects (as constructs of convenience) are internally variable (Mufwene 200la). Such variation can of course be more conspicuous in some varieties, such as incipient pidgins, than in others. Instead, what can be claimed to be lacking in such cases is what Chaudenson (1992, 2001, 2003) identifies as "normalization," i.e., the emergence of a communal norm that makes the different idiolects more similar to each other. 9. This fact does not of course preclude current children from producing innovations that can spread within their language communities once they are past the child-language stage. Creole children undoubtedly innovated too but not in ways that would uniquely make them alone the "creators" of Creoles. 10. As explained in Chapter 10 (see Mufwene 2005a for an elaborate discussion), the Hawaiian plantations, which have figured prominently in Bickerton's hypothesis on the development of Creoles, did not develop on the model of those of the Atlantic and Indian Oceans. Hawaii is unique in having had a Creole and a pidgin develop concurrently but in a complementary distribution, with pidgin spoken on the plantations and Creole in the city. 11. Independent of history, Bickerton's assumption that Creoles' grammars largely reflect UG with their parameters set in their unmarked options is more biased by considerations of typological markedness than by anything else. This is itself a function of a probability factor that need not be part of UG! The Cartesian conception of UG suggests that all parametric options are equivalent and only particular linguistic systems, which have evolved in specific ecologies of communication, would make some variants more, or less, marked than others. Thus the makers of Creoles are likely to have selected options that were less marked in the structural and ethnographic contact ecologies of their developments, not necessarily in UG (Mufwene 1991b, 200la). 12. This does not make Creoles exceptional at all, as the changes traditionally discussed in historical linguistics are evidence of imperfect replication, regardless of whether the changed items are considered individually or as parts of lexical domains or grammatical paradigms. 13. Gullah makes a distinction between the following constructions: How you duh [da] cook it? "How are you cooking it?" and How you duhz [daz] cook it? "How do you [usually] cook it?" 14. This is where Bickerton's (1989) "lexical learning hypothesis" still falls short of an adequate account, as he suggests that children would be inventing on their own the grammatical properties of the lexical items they nonetheless took from their parents. 15. In the case of Atlantic and Indian Ocean Creoles, we must bear in mind here the fact that the category ADJECTIVE is hardly part of the grammars of many African languages. 16. It is not evident that usage of the quantifier ONE in the stead of the indefinite article is exclusive substrate influence. 17. See also Lieberman (2002) and MacWhinney (2002b) for another informative discussion of the gradual evolution of language in response to adaptive pressures associated with the human physiology, mind, and social life all to the way to the emergence of phonological communication 200,000-50,000 years ago. They agree with McNeill that the evolutionary path from the most primitive forms of communication to the modern form covers no less than 5 million years. Noteworthy is also the fact that not all the modules of modern languages developed at the same time. For instance, semantic representation, associated with representations of the environment, seems to have evolved much earlier than phonology. According to Lieberman, even some primitive aspects of syntax and phonetics may have developed earlier. Homo erectus must have been able to communicate vocally (better than modern non-human primates), although not as the modern human. 18. In evolutionary terms, signed language (not to be confused with communication by gestures) has indeed not been eliminated; it has remained statistically an alternative to spoken language.

297

298

Notes 19. Wimsatt (2007) argues that the structures that are phylogenetically the most deeply entrenched are also the most resistant to change in biological and cultural systems, including research paradigms. See also Wimsatt (2000). 20. See Deacon (1997) for similar ideas about features that distinguish the earliest forms of human languages from communication systems among other animals, especially the significance of symbolism. 21. This communal behavior is undoubtedly cooperative but it differs from that of a team. In the latter case, the goal of the activity, the roles of the participants, and the rules of cooperation are explicitly articulated. The participants are often required to behave altruistically in the interest of the team. With successful communication as the goal, linguistic behavior is not altruistic. It serves the communicator's interests.

Chapter 6 1. I am grateful to Michael Picone and Catherine Davis, the convenors of the LAVIS III Conference, for inviting me to address the topic of this Chapter, about which I have learned more than I had expected. I am also indebted to Cecile Vigouroux and Michel DeGraff for helpful comments on its drafts. Needless to say I am solely responsible for all the remaining shortcomings. 2. Experts will undoubtedly recognize a gross oversimplification in this perspective, which is essentially an outsider's, non-European American one. It ignores rural varieties such as Appalachian, Amish, and Ocrakoke Englishes. It also overlooks the fact that, until relatively late in the mid-twentieth century, it was common to distinguish European American varieties by nationality, such as German and Italian Englishes. Consistent with the discussion in Section 6.2, the gradual disappearance of these distinctions can be correlated with changes in the ways RACE and ETHNICITY have been defined in the history of the United States and how the boundaries have collapsed between some groups but not others. However, this history does not deny the observation that Americans have traditionally socialized primarily along ethnic or national lines. 3. Darwin (1871) never clearly says that some languages are primitive. As a matter of fact, he observes that languages of "savage" or "barbarous" races (his terms) are much more complex than the rudiments spoken by our hominid ancestors. However, he speaks so much of, on the one hand, differential evolution, selection, and perfection, and, on the other, mental, intellectual, emotional differences among the races that it is difficult not to think of his theory as "ranking" languages of the different human races (qua geographical "subspecies" of mankind) at different levels of evolution then understood as "progress toward perfection." After all, he compares the highest, "anthropomorphous apes," such as the gorilla, to the "lower/inferior races," viz., the Australian aborigines, the "negro," and the "aborigines of S[outh] America." For similar assessments of Darwin's position, see Gould (1993: 266—9) and Radick (2002). 4. As observed in Chapter 5, Bickerton does not invoke pidgins alone in his quest for a window into the phylogenetic protolanguage. He also discusses child language (under two years of age), cases like those of Genie, who could not develop some significant syntax after being deprived for so long from participation in verbal communication, and apes trained to communicate with humans using some primitive symbolic code. What I find particularly disturbing is the hasty way in which he lumps all these cases together. At least since his (1990) Language and Species, sociohistorical evidence has increased that argues against assuming that Creoles evolved from erstwhile pidgins, especially Chaudenson (1992,2001,2003), Mufwene (1996a, 2001a, 2005a, and Chapter 3 in this volume), and Roberts (1998, 1999,2005). So has the literature in support of gradual development, led by Arends (1989, to appear). 5. Cukor-Avila & Bailey's (1996) longitudinal and apparent-time analysis of language evolution among their informants in Brazos Valley, Texas, had led to a similar conclusion, showing the younger generation aiming at urban variants which maximize differences between European American and African American varieties.

Notes 6. Quite aware of the fact that humans can interbreed across race lines, that races "graduate into each other" in ways that are independent of "intercrossing," and that members of different races share a lot of physical, intellectual, and emotional properties among them, Darwin answers the question of whether or not races are similar to species by identifying them with "sub-species" (p. 182). Surprisingly, he also claims that the "intellectual and social faculties of [primeval] man [before the exodus out of Africa] could hardly have been inferior in extreme degree to those possessed at present by the lowest savages," i.e., humans supposed to be less evolved (187). He simply ignored empirical evidence, choosing instead to endorse the social biases of his time. An interesting question is whether modern students of language evolution are not following Darwin's mistake in this respect. 7. Indeed Crosby (1986) explains settlement colonization as the European quest for territories where they could create better Europes than what they had left behind. 8. According to Tate (1965), it took Virginia's English colonists until the late seventeenth century, about 50 years after the arrival of the first Africans (then indentured servants), before they separated them from White indentured servants and enslaved them for life. 9. It is noteworthy that the term rank is excessively (mis)used by Darwin (1871) for dass(ify). He constantly "ranks" races even when he means 'class (ify)'. 10. The distinction between exploitation and settlement coloniest is discussed in Chapters 11 and 12. 11. As shown below, this account of the development of Creoles can also be found in Ambrose Gonzales's (1922) introductory comments in The Black Border, in his attempt to explain why Gullah is so different from (educated?) White American English varieties. I return below to the distinction. One can easily recognize in its language either Darwin's (1871) influence or, as suggested by Radick (2002), the general nineteenth-century social ecology that influenced Darwin's own references to "primitive" and "savage" races. 12. "Acquisition" of civilization by "adoption" was a concession that could be made for other presumably less evolved Europeans such as the Irish and southern, Mediterranean Europeans other than the Ancient Greeks. 13. This observation applies only to scholarly works. It is also dubious whether the myth of inferior races ever died completely. In the mid-twentieth century, Claude Levi-Strauss would still find it necessary to use the lectures he had been invited to give at UNESCO, "Race et histoire" in 1953 and "Race et culture" in 1971, to chastise Westerners for being ethnocentric and interpreting racial and cultural differences as meaning the superiority of the White race and Western civilizations. (See Levi-Strauss 2001.) 14. The term lexifier itself reflects that "exceptional" or "non-ordinary" paradigm used to explain these language varieties. According to Thomason (2001), for instance, they have [unusually] inherited most of their vocabularies from one source but the essence of their grammars from elsewhere, regardless of whether this is the bioprogram or some substrate languages. 15. Mufwene (200la, 2005a) analogizes restructuring to feature recombination in biology, making allowance for features from substrate languages to mix, in variable proportions, with those of the "lexifier" but mostly making it clear that features from the "lexifier" itself recombine often in novel ways, especially when, at the communal level, they originate from different dialects. Restructuring is the null hypothesis in the ontogenetic formation of idiolects, as every speaker develops their idiolect selectively from inputs availed by diverse speakers (Chapters 7 and 10). 16. Relexificationists may object to my remarks and claim that Creoles have selected only the lexicon, or rather the phonological forms of the lexical items, but have associated them with semantic and other grammatical properties of the dominant substrate languages. The whole relexificationist account of the development of Creoles has repeatedly been questioned in creolistics. I need not repeat the literature here. More interested readers should see DeGraff (2002) for the most elaborate and convincing refutation. Regarding the claim that the phonetic shapes of lexical items from the

299

300

Notes "lexifier" were generally retained intact, note that it is ironically at the phonetic level that interference from previously spoken languages is the most obvious. If the slaves learned the phonetic systems faithfully enough, then it is likely that they also learned other aspects of the seventeenth- and eighteenth-century nonstandard varieties of the European languages faithfully enough. Creoles may thus be among our best windows into colonial varieties of the European languages among the indentured servants and other European colonists. 17. The reader should realize that none of these scenarios is the full story. The colonial history of the New World also suggests that because of national segregation among them, free, non-indentured Europeans were not under immediate pressure to shift to the language of the politically and economically dominant group. For them, the process was more gradual and spread over several generations. This gradual process, correlated with the progressive integration of the European populations, would not become significant until the nineteenth century. In some cases, as with the Germans, the process did not become complete until the early twentieth century (Salmons 2003). This suggests that the transformation of the dominant language into its colonial varieties took place with marginal influence from speakers of other languages. The other Europeans would shift gradually after these new, colonial varieties had emerged already, with the adults' nonnative accents being stigmatized and avoided by their children. This scenario is similar to that of many recent immigrants, whose children have "acquired" the local varieties but who have not exerted any significant influence on either the varieties spoken, say, by African Americans or those spoken by European Americans. Thus, the Africans would have anglicized before the free immigrants from continental Europe. 18. To be sure, interested tourists in Jamaica can buy T-shirts and small books teaching them the "Jamaican language." One should not forget that the merchants who have produced these products respond to an appetite for exoticism, do not represent the general sentiment of the Jamaican Patwa speaker, and spit back to their potential customers what they have learned from linguists. As Irvine (2004) shows, much of the official Jamaican literature produced for foreign investors and tourists asserts that only one language is spoken in Jamaica, English, although its nonstandard variety has its own local idiosyncrasies. The same treatment applies to nonstandard varieties in other parts of the world that are difficult for outsiders to understand.

Chapter 7 1. This essay was originally presented at the meeting of the International Society for History, Philosophy, and Social Studies of Biology (ISHPSSB), held at Quinnipiac University, Hamden, Connecticut, 18-22 July, 2001.1 was invited to participate in the panel on "Language Change as a Selection Process," organized by David Hull, Professor Emeritus at Northwestern University. I am very grateful to him and Michel DeGraff for candid comments on the post-conference draft that was eventually published in the now defunct Selection (3.45-56). I am solely responsible for the remaining shortcomings, especially where I did not follow their advice. I am equally indebted to William Kretzschmar for suggesting that I explain somewhere what I mean by linguistic features "competing" with each other while I maintain that linguistic features have no more agency than biological genes in determining the evolutionary trajectory of a species. 2. An informal survey I conducted among 19 native speakers in the Department of Linguistics at Chicago actually revealed more about variation than I can discuss here. Although the sentence with prediction is generally considered ill-formed, six respondents rated it 3, rather than 4 (the lowest on a scale of 1 to 4). According to the rating, they would not be shocked to hear a fluent native speaker say it. Five respondents rated it 2, and one even rated it 1 (the rating for an unproblematic construction). The population was generally split on the acceptability of the sentence with predict, with 13 respondents assigning it a rating of 1, four a rating of 2, and only two respondents rating it 3.

Notes There was unanimity only in rating the sentence with heard as 1 and near unanimity with the verb said (one respondent rated it 2). Another sentence with wrote (You wrote the Democrats would prevail} displays just as much variation, with six respondents assigning it a rating of 1, three a rating of 2, six a rating of 3, and four a rating of 4. Although these are raw data, on a very small population, which call for a more sophisticated analysis, they are nonetheless informative about how grammar varies inter-idiolectally and about the danger of extrapolating too hastily from one speaker to a whole population. 3. Note that in this predicative function, the adjective pretty can still be modified by the intensifier very, which proves that it continues to function as an adjective and does not become a verb, contrary to what much of the literature on Creoles has claimed (Seuren 1987). Even in English Creoles one cannot modify a verb with the intensifier very. One cannot say *De man very/bery ron any more in English Creoles than one can say * The man very runs in any English dialect. These vernaculars really differ from their European lexifiers in not using a copula to form a verb phrase for predication in this context. More on this below. 4. In English, the copula carries no particular lexical meaning, although it carries tense. One can say that it does so simply because it is a verb morphologically and behaves like other verbs in tensed clauses. In infinitival clauses such as Holly wants to be a doctor, the copula carries no tense at all. So the real reason why it is used in English sentences is that this language requires a predicate phrase to be headed by a verb and this verb must be a dummy one when semantic head of the predicate phrase is not a verb. The exception to this observation is secondary predication. The copula carries tense in finite clauses simply because English also requires that in finite clauses tense be expressed through a verbal inflection. Some languages, such as Chinese and those of the Kwa family, do not have these requirements. See Mufwene (1992a) for such non-monolithism in English grammar and DeGraff (1992, 1998) for more on such secondary predications in English and Haitian Creole. See Mufwene (2005b) and references therein cited for more elaborate discussions of the copula in especially English. 5. I argue in Mufwene (2000a), contra Baker (1990) and Thomason (1997), that the linguistic heterogeneity of the contact settings in which Creoles developed did not preclude the identification of a particular language variety as a target. Actually, these settings are reminiscent of the heterogeneity of the primary linguistic data in those settings involving only one language, as each idiolect and each dialect have their own peculiarities and, as discussed above, the learner must often "make choices," or at least determine which variants are dominant or which ones he/she can use under what particular ethnographic conditions. There is no particular naturalistic setting where a language is made available to the learner in a homogeneous form. The reader should remember that the selections are not necessarily conscious; they are patterns emerging from nonce decisions made during particular speech acts. 6. Idiolects change all the time, as speakers innovate or accommodate each other. However, most of these changes have no effect on a communal language, as explained below. 7. Ben is considered phonologically more salient than the inflection -ed on the verb. In this particular case, it is an autonomous full syllable, is not used in a reduced form, and can bear an emphatic stress. The inflection has none of these peculiarities and is often affected by cluster simplification in "sloppy" speech. It is not uncommon even among native speakers to produce a form such as worked /wa^kt/ as [wa^k]. 8. The facts are actually more complex than this summary explains. One can indeed say in Jamaican Creole or Gullah he dead/gaan yeside "he died/went yesterday" or he broke im/um yeside "he broke it yesterday." On the other hand, while one can say he go dead/die/go/broke (with an object NP required for broke], one cannot say hegogaan. The variants he die/go/broke ... yeside are also acceptable, suggesting that selection does not operate in exclusive terms but rather by ranking the variants in order of preference, which can of course assign a null value to some variants.

301

302

Notes 9. One must note, however, that frequency did not prevent the selection of the copula out of Creoles' grammatical "systems," suggesting that selection is driven by more complex processes than one may suspect. 10. This problem is significant if one agrees with me that communal languages are mere extrapolations from speech/ signing habits of individuals. While social and political structures determine the variable demographic sizes of populations associated with such communal constructs as Chinese, Xhosa, or IXung, the number of "languages" displaying typological patterns of diversity reflect the boundaries imposed by the sociopolitical group distinctions. They tell us very little about the evolutionary processes that would make some typological options preferable to particular speakers.

Chapter 8 1. I am grateful to Albert Valdman for the challenging invitation to write the 1990 ancestor of this Chapter, which has motivated me to sit down and assess what has been going on in this research area. I also thank Philip Baker, Frank Byrne, Robert Chaudenson, Nancy Condon, Glenn Gilbert, the late Guy Hazael-Massieux, Silvia Kouwenberg, and Margaret Wade-Lewis for helpful comments on one or the other of the first two drafts. All the remaining shortcomings are my sole responsibility. 2. Recent publications such as in Kouwenberg & Patrick (2003) and in Lefebvre et al, (2006) seem to still subscribe to this older position. 3. Siegel (2006) has apparently changed his mind on this, as he observes that there is no evidence for relexification in the development of any Creole or in any literature on SLA. 4. Mufwene (200la, 2005a, and Chapter 7) is actually an elaboration and refinement of this position, articulated from the perspective of competition and selection from structural elements of both the "lexifier" and the substrate languages, situating in the ecology the factors that (dis)favor particular "contenders," and attributing to UG general constraints that guarantee that the outcome of the restructuring process is still a language. 5. Aboh (2006) also speaks of "hybridization" in the emergence of Creoles' structures, but his discourse is more in line with the complementary hypothesis (based on Mufwene's competition-and-selection approach) than with Adam's notion of hybridologie. Other than Adam, one must recognize other French philologists, especially Bertrand-Bocande (1849), Baissac (1880), and Vinson (1882, 1888), who invoked African substrate influence with racist overtones to account for the structures of Creoles. As explained in Chapter 6, the denigrating racist interpretation reflects colonial attitudes of the nineteenth century, which saw most non-Europeans as being physiologically and mentally less evolved than Europeans and therefore incapable of learning a European language well. Creoles were then treated as linguistic aberrations and the minds of their creators as "natural" and "primitive." 6. Note, however, that Sylvain is better associated with the complementary hypothesis, as suggested by her discussion in the body of the book, than with the relexification hypothesis (Mufwene 200la). 7. As noted in Chapter 5, note 5 (p. 296), Baker (1993b) disputes this particular conclusion, arguing that the originating point of these pidgins would be Queensland, from which the whaling and trade ships dispersed in Melanesia. 8. As argued in Mufwene (2003c), McWhorter's attempt to by-pass this problem by invoking a set of three features which define some putative "creole prototypes" (absence of tones, absence of inflections, and absence of derivations) led implicitly to the conclusion that most of the vernaculars hitherto identified as Creoles are less creole than they have been claimed to be, just like claiming, on the basis of structural features, that English is less Germanic than it has been claimed to be. Genetic classifications are not typological; they are literally about shared linguistic ancestors known as protolanguages. We must also note that the European "lexifiers" targeted by the slaves were not tonal.

Notes While the significance of tones in Creoles such as Papiamentu suggests that we look for an explanation regarding the influence of substrate languages, the absence of tones in these new vernaculars suggests that there is much more from the structures of the lexifiers that was faithfully learned by the African slaves. The absence of tones can thus be accounted for by the same kind of explanation as the position of articles, adjectives, and the relativizer in English Creoles that is more like in the "lexifier" than in the substrate languages. See also Section 3.2.3.4. 9. I am tempted here to compare the bioprogram, UG, or the biological endowment for language (whichever name suits the reader) with the basic machine software that every computer must have in order to run other software. Nothing runs when this basic software is absent or defective, though quite a wide range of well-designed or operational softwares can be used in one's computer when the basic one is installed. The constraints that the basic software imposes on the latter are limited. 10. Incidentally, an important difference between Kituba and Lingala, which evolved in similar contact situations involving primarily Bantu languages, is that the former is only partly tonal (limited to some words - Ngalasso 1989) and has a predominant pitch-accent that is borne by the penultimate syllable, whereas the latter is completely tonal, lexically and grammatically. It is not clear what ecological factors account for this divergence in the evolutionary trajectories of both Kituba and Lingala. Urban Swahili, which evolved under similar contact conditions, has only a pitch-accent that is also borne by the penultimate syllable (Mufwene 1989d). According to Kapanga (19991), Shaba Swahili, the contact variety, also has a richer TENSE-ASPECT "system" than ethnic Swahili, having been influenced by the Bantu languages of the region.

Chapter 9 1. I am grateful to Bernd Heine, Michel DeGraff, Annegret Bollee, and Sybille Kriegel for useful feedback on earlier versions of this Chapter. I assume alone full responsibility for the remaining shortcomings. 2. The conclusion that "there has been no language-internal development of a lexical item into a functional one" doesn't of course follow if one assumes that Creoles developed gradually from their lexifiers. The absence of ben from early Sranan texts (i.e., early eighteenth century) simply suggests that it had not yet emerged as an ANTERIOR marker or was simply not recognized as such by the authors of those texts. If Creoles have evolved from their lexifiers by basilectalization (the formation or consolidation of a basilectal sociolect) - which seems consistent with Bruyn's own observation about features missing from early Sranan texts - then the features considered particularly "creole" are later developments (Chaudenson 1992,2001). If we can extrapolate from Baker (1995a, 1995b), these features would have emerged at different stages of the development of the Creoles. 3. Unfortunately, she gives no examples of such "universal grammaticalization developments." 4. See DeGraff (200la, section 4.5) for a complementary discussion of UG, where true universals would be specified. According to him and several sources he cites, UG is largely underspecified and parameters are set in favor of some typological options provided by the primary linguistic data produced by speakers of the "lexifier," influenced as (several of) the constructions are by the other languages spoken by nonnative speakers. (This is my interpretation, based on my feature-pool approach, of his more elaborate discussion.) As Michel DeGraff (p.c., 13 September, 2001) has reminded me, "the crucial point... is ... that, in all cases, the language acquirer is exposed to utterances, not to grammars, thus the ineluctability of language (re)creation." 5. It is not clear what must be taken as the onset of the pressure on continental Celts to shift to Latin, the beginning of the Roman colonization or the beginning of the post-colonial period, after the Romans had left and indigenous monarchs and other, Germanic settlers (e.g., the Franks and Vandals) adopted Latin as their vernacular and the

303

304

Notes language in which they ruled the former Roman colonies of southwestern Europe. What I know is that the Germanics did not interact much with the insular Celts in England until after Old English had emerged, around the seventh century. There is not much mention of Celtic influence on English until the Middle-English period, although students of Celtic Englishes can tell us otherwise. This is just one more interesting question about (differential) language evolution in the former Western Roman Empire. 6. Warner (1993: sections 8.5, 9.6, and 10) is informative, although not with all the details one would want to see for the kind of position I am developing. The subject matter of the development of auxiliary verbs in English is a complex one, involving at least the following dimensions: (1) the basic grammaticization process in the usage of individual lexical items; (2) the widening of a class of verbs or verb-like items that already functioned as auxiliaries in Old English; (3) the spread of the defective-morphology peculiarity from the first specimen of this class to other members; (4) the geographical spread of the grammaticization process from some parts of England to others; and (5) the role of extant morphosyntactic properties of the relevant verb-like items in favoring their grammaticization as auxiliary verbs. Some of these considerations are part of the present discussion. The study includes no sociohistorical considerations. So we cannot conclude anything about the role of external ecology in this aspect of the evolution of English grammar. There is certainly nothing so far that inspires distinguishing in kind between grammaticization in English and what appears to be the same process in English Creoles. 7. The plantations did not universally replace the homesteads in the relevant communities, since not every farmer had enough money to become a planter. What should be understood by "transition from homestead to plantation societies" is that plantations came to be the overwhelmingly dominant employers of slaves and in some cases also of indentured servants. Thus there was no uniform language evolution in the colonies either, because patterns of social interaction on the plantations (where segregation was instituted early) were different from those on the homesteads (where there was no segregation, despite the discrimination which changed the status of Africans from indentured servants to slaves for life within 50 years in seventeenth-century Virginia). 8. Note that this explanation is at variance with how Hopper & Traugott (2003) paraphrase Lass (1997), rather incorrectly, interpreting EXAPTATION as an alternative definition of grammaticization. I understand this notion, instead, as characterizing the process by which a form or construction is grammaticized, in the way grammaticization was originally defined by Meillet (1912) and paraphrased in by now classic references such as Heine et. al (1991). 9. In a few publications, including Chaudenson (2001, 2003), Chaudenson has quibbled on this, speaking instead of "acrolectal Creoles" and never clarifying whether he intends by this term the acrolectal, standard varieties of the same European languages typically not considered Creoles in creolistics, or the upper mesolectal varieties (more commonly spoken than the acrolects themselves), which putatively approximate the acrolects. 10. Heine (1993:44) mentions a "Serial Schema" used by some African languages for FUTURE constructions. He cites the following example from Turkana, taken from Dimmendaal (1983: 136): ki-poni' dt-mdt-a we-go We-drink-PL We shall drink In principle, partial congruence between such a construction pattern and the be going to + Verb pattern in the "lexifier" would have favored the kind of go-based constructions attested in Atlantic Creoles. However, Turkana is spoken in East Africa. Unless there are such constructions among Western and Central African languages, one may have to be content with the usual kind of constructions with 'go' + Verb that have traditionally been adduced in the literature and are typologically closer to some of the Creoles' structures.

Notes 11. In some nonstandard English dialects, what continues to function to date as a non-locative relativizer, as in Are you the one what said it? (Trudgill 1983:189). One of its alternative pronunciations during the colonial period was [waet]. It is thus not surprising that it seems to have merged with where into [we] or [wSy] in Gullah and other English Creoles. It is not clear why that is not used in English Creoles' relative clauses. It may have to do partly with the fact that it is not used to introduce complement clauses in general. Instead, s£ (< say] is used for this function. However, like in the substrate languages, this function of the complementizer is limited to complements of verbs; it is not extended to relative clauses. Polyploidy in the inputs (i.e., congruence), in the way both the "lexifier" and substrate languages contributed to the formation of Creoles' grammatical features, deserves more attention. 12. According to Michel DeGraff (p.c, 13 September, 2001), pied-piping occurs in questions, as in (i), as opposed to (ii): (i) Ak ki moun mwen te ale nan sinema ? With WH person me ANTER go to cinema With whom did I go to the movie? (ii) *Ki moun mwen te ale nan sinema ak? "Who did I go to the movie with?" 13. According to Michel DeGraff, if he had to use an overt relativizer at all in constructions (9b-c), it would be ke, but not ki. However, he also makes the following cautious observation: "My own impression is that the use of ke instead of the null relativizer may be more common with French-Creole bilingual speakers, but this is only an impression (not based on any serious sociolinguistic statistics)." (p.c., 13 September 2001).

Chapter 10 1. This Chapter has benefitted from discussions with Michel DeGraff and Alison Irvine, to whom I am very grateful. I am alone responsible for the remaining shortcomings. 2. The number of extensive and integrated descriptions of these vernaculars has increased over the past few years, providing more materials for systematic comparative studies. They include Kihm (1994), Baptista (2004), and Quint (2000), among others. 3. It appears from Ostler (2005) that Portuguese was a major colonial language, which was even adopted as a diplomatic language in Asia until the eighteenth century. It prevailed as the primary trade language between Europeans and non-Europeans until the exploitation colonization of Africa and Asia started in the nineteenth century. All the expanded pidgins that survive today have evolved from pidgins that started in the nineteenth century, consistent with Huber's observation on the significant role of Portuguese as a trade language in West Africa. What Hancock (1980a, 1986) identifies as "Guinea Coast Creole English" was probably not a pidgin. Nor was what Dillard (1972, 1985) identified as "West African Pidgin English." These must have been nonstandard varieties spoken by the children of the "lancados" and the early interpreters, as explained in Mufwene (2005a). 4. This was apparently an extension of the term used for low-ranked auxiliaries on the trade ships. The "grumetes" were thus auxiliaries to the Euro-African trade between the fifteenth and the eighteenth centuries. 5. In addition to the hitherto plausible etymology provided by Baker & Muhlha'usler (1990), deriving pidgin English from business English, Smith & Matthews (1996: 146) observe that "a more likely origin might be the common expression [ bei chin] which means 'give money' or 'pay' in Cantonese." Congruence of forms, in a context of undoubtedly societal variable pronunciation, would have contributed to the emergence of the term pidgin, which is no closer phonological match to English business than it is to the proposed Cantonese etymology. On the other hand, Morris Goodman (p.c., 9 May, 2007) observed that business [blznis] was apparently also modified to [blztn(s)], which would have evolved into pidgin. We cannot rule out the convergence of multiple etymologies in this evolution, ruling

305

306

Notes out, of course, the most implausible ones reviewed in Miihlhausler (1986a, 1997), which are not backed by colonial history. 6. The fact that there is hardly any mention of the contribution of Native American languages to the structures of New World Creoles reflects two important factors: (1) contacts between Native Americans and the Africans probably did not take place in the European languages and were not significant enough, even during the homestead phase, to influence the evolution of the European languages among the Africans; and (2) the enslavement of Native Americans was avoided during the plantation phase and/or their numbers were not significant enough to exert any impact on the emergent vernaculars beyond the lexicon. 7. As Ostler (2005) corroborates, the Dutch colonists and colonizers had a peculiar linguistic practice of adopting the lingua franca they found in the colonies they took (away) from other Europeans, such as Portuguese and Malay in today's Sri Lanka, in coastal India, and in Indonesia, for trade and proselytizing among the Natives and for communication with their slaves. Dutch spread as a vernacular or lingua franca with the Natives and slaves only in the colonies they were first to settle, viz. the New Netherland (New Jersey, the southern part of New York state, and Delaware), some of the Leeward and Virgin Islands, Berbice (in today's Guyana), and South Africa (after the Dutch had taken Cape of Good Hope, a promontory south of present-day Cape Town, away from the Portuguese). Dutch-based Creole vernaculars developed only in Berbice and the Leeward and Virgin Islands. The status of Afrikaans remains controversial in creolistics, although, as we should remember from Chapter 3, the designation as Creole is, from a genetic linguistics point of view, a distraction from what we should be learning about language speciation, especially regarding the varying role of the socioeconomic ecology from one colony to another. 8. An important difference may lie in the fact that, as noted in Chapter 3, the European colonists maintained some parochialism during the colonial period, settling in national colonies in which they continued to speak the languages they had brought from their metropoles. Even after economic and political independence from Europe, they continued to live in nationally segregated communities. Some did not shift to the dominant European vernacular until the twentieth century. Such ecological peculiarities prevented such Europeans from exerting on the dominant colonial vernaculars the influence they could have if they had experienced language shift early in the colonial history, like the African slaves and the European indentured servants. American English would have been more divergent from British English, had the continental Europeans, whose descendants are now the majority, shifted earlier from their national or ethnic languages (Chapter 6).

Chapter 11 1. French in Quebec was saved by legislative and economic measures that recreated an ecology that has favored its usage and prevented it from dying in the same way it has in Nova Scotia and Maine, for instance, and is now in Louisiana. 2. To be sure, cases of sudden language death by genocide have been attested (see, e.g., Hagege 2000 and Nettle & Romaine 2000) but they are rare compared to the other cases most commonly discussed in the literature. They are not really part of natural evolution by competition and selection, as explained in Section 11.5, and they will not be discussed below. 3. It is less clear whether a language is still alive or just moribund when it is used by semi-speakers, individuals who claim they speak it but mix its vocabulary and grammar with the "system" of another language. Dorian's (1981) discussion of Scottish Gaelic has made such cases an important part of understanding language "obsolescence." 4. Hazael-Massieux (1999) reports an important distinction made by French sociolinguists between langues minoritaire ("minority language") and langue minoree ("undervalued and/or marginalized language"). The latter need not be

Notes spoken by minority populations. Like Haitian or Jamaican Creole, they may be spoken by the majority population of a polity but are relegated to ethnographically "low" communicative functions. If this view is taken literally, most indigenous languages in former European colonies fall in this category, as they are not associated with the "high(er)" communicative functions of their polities. (See Pandaharipande 2002 regarding such situations in India.) Consistent with seemingly precocious predictions that 50% - 90% of the world's languages will have vanished by the end of the twenty-first century, the approach misleadingly suggests the same perspective on the coexistence of languages of the powerless and of the powerful everywhere. However, the history of the world shows languages of the powerless have often been more resilient, or demonstrated more vitality, than those of the powerful. Pace Fishman (2002), there is much more ecological complexity and variability that must be factored in regarding this subject matter, as attempted, for instance, by Pandharipande (2002). 5. As noted in Mufwene (2002a), the strange way in which linguists have been using the term indigenous in the relevant literature would also mislead the uncritical reader into thinking that Europe has had no indigenous populations and the Indo-European expansion has had no particular negative effect on the languages of such populations. One gets much of the same feeling from, e.g., the way Hughes (2003) speaks of the "indigenous peoples," "aboriginal peoples," "indigenous societies," and "indigenous systems" all situated in former European colonies since the fifteenth century. 6. The informed reader will have noticed that much of the literature since Kraus (1992) has hardly articulated the ways in which the maintenance of such cultural heritage would help the would-be speakers of the relevant languages adapt to the socioeconomic changes they experience. Since everything else suggests that the relevant populations have shifted from their heritage languages because these disadvantaged them, one would expect the linguists' encouragements that the populations stick to their languages to address this particular question too. As noted above, nobody has articulated how societal and individual multilingualism, which in former settlement colonies has generally yielded to language shift, could be sustained where the European populations who had had a similar experience could not do it. Curiously, books such as Nettle & Romaine (2000) and Miihlhausler (2003) that largely rearticulate the Sapir-Whorf hypothesis do not address these important issues. This may in fact reflect the fact that several linguists have been trendy on the subject matter of endangerment without preparing themselves theoretically to address real-life issues for the relevant populations. We should realize what our expertise can or cannot (help) change. 7. A commonly advanced reason about the prevalence of Greek over Latin in the Eastern Roman Empire is that the Romans rated Greek as more prestigious than their own language. Therefore, they adopted it as their administrative language. That raises the question of why no modern languages other than Modern Greek have evolved from Ancient Greek, just like that of why no Romance language(s) evolved in England. Janson's (2004) account of the latter phenomenon with the attitudes of the insular Celts to Latin is worthwhile considering, though I suspect other factors having to do with the Germanic colonization may shed light on this differential evolution. 8. One can in fact argue that writing systems are technological inventions intended to record languages nonacoustically. The most efficient ways to revitalize languages should consist in providing the kinds of ecologies in which they will continue to be spoken and "transmitted" naturalistically from one generation to another. As should be obvious by the persistent moribund state of Irish in Ireland, the "transmission" of a language through the written medium is pointless if nothing else is done that guarantees its oral usage and its naturalistic "transmission." The same is also true of Classical Latin compared to Vulgar Latin. 9. One should remember here that it is in part because they are generally unwritten that terms such as patois and brogue (used in reference to nonstandard varieties of French and English, respectively - and the former also in reference to Celtic languages in rural France) have had negative connotations. Earlier uses of jargon, associated also with some contact language varieties, have similar negative connotations of "unintelligible," "meaningless," or "gibberish".

307

308

Notes 10. Since DeCamp (1971), who repeats a hypothesis conjectured earlier by Schuchardt (1914) and Bloomfield (1933), a number of Creoles have been assumed to die by a process misidentified as decreolization (Mufwene 1994, 200la). However, varieties such as Jamaican Creole (ironically the focus of DeCamp's own speculation) are not only still thriving but also developing more divergent varieties, such as "Dread Talk" (Pollard 2000). As I show in Chapter 14, Gullah maybe dying for a reason that is independent of "decreolization" qua debasilectalization: the massive exodus of its speakers to the city, where the variety is given up not only because it is stigmatized but also because it is economically disadvantageous in the urban environment and outside its traditional habitat. 11. During the homestead phase, the "acquisition" of an African language would probably have been the exception rather than the rule, as the demographics of the homesteads disfavored this possibility. As explained in Chapter 10, it must have been very difficult for the African parents integrated as small minorities in various homesteads and isolated from one another to find another person with whom they could speak their ancestral language. The difficulty stemmed from the fact that the slaves originated in different ethnolinguistic areas. Those who came from the same areas were more likely to live in different homesteads, separated by distances that discouraged regular contacts. Private parent-child interactions alone could hardly have favored the maintenance of African languages as vernaculars under the social-integration conditions of the homesteads. 12. Salmons reports that the Germans in Wisconsin owned parochial schools and published newspapers and other literature in their language until the readership waned, as more and more of the community members were attracted by the larger, urban American global economy. 13. As well articulated by Banner (2005), "Indian reservations" quickly turned into schemes of exclusion and "pockets of poverty" marked by hardship. Aside from the fact that the tracts of land allotted to Native Americans kept shrinking, the latter often found themselves in strange territories in which they could not continue their traditional farming or hunting practices. Such ecological pressures encouraged migrations out of the reservations, while the mixing of Natives from different nations on the same tracks undermined the maintenance of the integrity of their respective traditions. This evolutionary trajectory of cultural erosion was also aggravated by high mortality rates (see also Boyle 2007). I return to this in Chapter 12. 14. The government continues to promote education in French after elementary school, where the language has been taught as a subject since the fourth grade. 15. Trade colonization is not worth considering in this case, because the structure of the Empire seems to rule it out, although the exploitation of the Empire through its indigenous rulers is reminiscent of the European trade with the interiors of Africa and Asia through intermediaries before the onset of the exploitation colonization of these places in the nineteenth century. An important difference from the latter is that the Roman presence in their Empire was permanent and they held regular, not sporadic, contacts with the local populations in the cities they built or inherited from the Greeks both on the coast and in the interior. 16. In a still unpublished manuscript, Robert Chaudenson emphasizes the role of waterways before the invention of trains, planes, and telecommunication, in spreading languages outside their homelands. I return to the significance of geographical interconnectedness in economy below. To be sure, the Roman Empire's economic system instantiates some of the early stages of globalization. 17. One must realize that Rome was too small to provide all the legionaries needed in the Roman army and the manpower necessary to staff its colonial administration. Vernacular Latin was spread outside Rome largely by the foreign mercenaries in the Roman legions, just like English is spreading today as a world lingua franca significantly by nonnative speakers using and teaching it to others in their respective countries.

Notes 18. At the world-wide scale we have no idea whether 1,000 years from today the different varieties of English, for instance, will still be considered the same language. The fact that some of them have already been disfranchised as "indigenized" raises the question of whether they may not some day be considered separate languages altogether in the same way that Creoles already are. 19. This is not discourse against the documentation of endangered languages for the purpose of preserving records of their structures. It is against the paternalism of dictating which particular languages particular populations must speak, as if the languages were deterministically more suitable for them, regardless of the fact that the speakers find them disadvantageous for one reason or another. It is also discourse against the reification of cultures and their conception as static entities, whereas they are practices that are being constantly reshaped by their practitioners.

Chapter 12 1. Authors such as Dollfus (2001) and Carroue (2002) actually make it clear that French linguists should have used globalisation, the correct translation of globalization, instead of mondialisation as the unmarked term. The reasons are the same as those I give below regarding English. 2. The difference can be explained as follows: In the traditional colonial system, one particular nation had the monopoly of control over a particular territory and of access to its raw materials for the development of the metropolitan industry. In the globalized form of colonization, one company holds privileged relations with another company or many others, and it derives preferential profits from the connection(s). The common colonial feature is that in both cases the exclusionary relation is asymmetrical, favoring the exploitation of the weaker by the stronger. 3. In this case too, one can point out the kinds of asymmetries that have become all too common in today's global world "system." However, I argue below that not all such power asymmetries lead to the endangerment of indigenous languages by Western European world languages. 4. Note in connection to this question that the cladograms of genetic linguistics normally show only diversification, hence continual increase in the number of languages that have evolved from a common ancestor. These inaccurate illustrations, which do not factor in the languages that a photo-language at any level of the cladogram came in contact with, conjure up a scenario of language evolution that is contrary to the gloomy one portrayed in the literature on language loss. As explained in Chapter 3, they say nothing about either the concurrent loss of languages or the role that these played in the speciation process through substrate influence. 5. This is basically also the position developed by Cowen (2001), who interprets colonization as a quest for more resources for subsistence and sees in the history of the world "patterns of exchange and exploitation, patterns of conquest and military rule, and patterns of military conviction and endeavor" (9). All this is also correlated with "tension between movement and settlement," from a perspective that shows globalization as a process or condition that has always been inherent in the history of mankind but has consistently increased in complexity, speed of expansion, and the size of the territories involved. 6. As explained in Chapter 3, pidgins as reduced language varieties spoken in settings of sporadic contacts either died or evolved into expanded pidgins once the contact situations evolved into settlement or exploitation colonies. Thus we can say that they have not outlived trade colonies. 7. Note that the boarding schools in which Jesuit priests subjected their indigenous students to humiliating punishments if they were caught speaking their native languages did not have a more significant effect than their counterparts in Ireland, where the potato plantations were more successful in spreading (nonstandard) English. The students

309

310

Notes could not speak the European language with their uneducated relatives once they came back home; and the graduates went to work in the city rather than on the reservations. Their indirect role may have lay in motivating more Native children to go to school and migrate from the reservations. 8. Latin American countries fit only more or less in the second category. Their socioeconomic histories and developments are not quite like those of North America and Australia. They have larger proportions of Native American populations, who are still on the margins of the new socioeconomic world order, and their European style socioeconomic world order has developed ways partly reminiscent of former exploitation colonies. I return to this below. 9. As could be inferred from Chapters 10 and 11, the history of plantation settlement colonies makes implausible the myth that the slaves gave up their languages because their masters forbade their usage. The masters could not have controlled what the slaves spoke and taught their children in private. Another myth not to entertain is that the slaves' African languages died as soon as their Creoles developed. Up to the time of Emancipation, there always were among the Bozal slaves those who were fortunate enough to meet somebody who spoke the same language (native or not), especially at times when some planters preferred slaves from the same region. The main question is whether their children were interested in speaking their African languages, even in addition to Creole. 10. It is only in the second half of the twentieth century that the economic colonization of the forest has intensified, that its ecosystems have been disturbed, and that the relocated Native populations and others who have recently come in contact with outsiders have been obviously negatively affected by the changes. In fact, it is thanks to these particular changes that we have become more aware of the erosion of our ecosystems and of species endangerment, and ultimately of the endangered languages. Today's arguments for the maintenance or preservation of languages (see, e.g., Nettle & Romaine 2000, Maffi 2001, Miihlhausler 2003, Crystal 2004) are in fact often cloned on the discourse of environmentalists, though they do not articulate what particular adaptive advantages to the new socioeconomic ecologies the revitalization of the dead or moribund languages would confer to their speakers. They all speak of knowledge encoded in these languages but do not say how useful it would be and in what particular ecologies. 11. I will ignore here cases such as in Zimbabwe, where a settler class emerged and segregated itself, with a host of socioeconomic privileges, from the indigenous populations. The case of South Africa, to which I return in the next Chapter, must also be treated as a case of dual colonization, viz., that of the Afrikaners, on the settlement model, and that of the British rule, on the exploitation model. Needless to say that the two regimes clashed, which accounts for many of the social woes of South Africa since the institutionalization of the Apartheid regime in 1948. 12. For those who assume that "global languages" necessarily endanger the indigenous ones in the territories where they have been adopted, it is worth noting that Greek did not endanger Latin even in the Eastern Roman Empire, where the Roman elite spoke it as a lingua franca. Latin died in the Western Empire through its own geographical expansion and transformation, whereas in the east, as in England and North Africa, it simply does not seem to have had a chance to evolve into a vernacular, "global" though it may be considered thanks to its association with the Roman rule. 13. There is a misguided tradition that confuses the notions of "world language" and "major language." For instance, Chinese varieties and Hindi are major languages confined primarily to where they originated, despite the fact that they are also spoken by Chinese and Hindi populations in the relevant diasporas. However, they are not world languages, because they are hardly targeted as lingua francas by non-Chinese and non-Hindi populations outside China and India. In this respect, Spanish is not as much of a world language as French is, although it has more native speakers than the latter does. The vast majority of Spanish speakers are those native ones, from Spain and its former settlement colonies in Latin America. The situation is just the opposite with English and French, most of whose speakers are nonnative.

Notes 14. Phillipson (2003) must refer to this same management level when he alleges that there are asymmetries in the ways American and European firms function abroad, with the former being less accommodating to the local official language than the latter, owing largely to the monolingualism of American CEOs. (Marling 2006 reports otherwise; American firms prefer representatives that are familiar with, if not fluent in, the local official language. They often hire local senior administrators to manage foreign branches of their industries.) Regardless of whether or not his observation is accurate, this is not the level of population contact that affects the vitality of a language in a polity. The alleged behavior of American senior managers is comparable to what Phillipson reports of senior administrators in the European Union, in contrast with the individual multilingualism of mid-level bureaucrats. I am sure lowerranked employees can function monolingually in the official language of the city where the office is located. 15. Language maintenance has to do with the activities of a population that help them keep their language alive. It is related to language revitalization, which literally means "bring life back" to a language that may die without the relevant intervention. Language preservation amounts to no more than documenting a language through writing texts and/or grammars and through compiling dictionaries, interventions that do not necessarily get its would-be speakers to use it. The closest analog I can think of is suggested by the term preservation itself which conjures up lifeless organisms, plants, or body parts in a jar (Chapter 11). 16. It may be informative to clarify here that linguists have approached the issues without as much scholarship as environmentalists. The latter are generally biologists well trained in macroecology, in the dynamics of the coexistence of species, and in the ways that biotas balance themselves out in ecosystems. They are usually well informed in biotic disequilibria that follow from mismanagements of the physical ecologies, such as through urban and industrial developments. Zealous though concerns with non-human species sometimes appear to be, the future of humans typically emerges at the center of controversies. The question can often be recast in terms of (ultimate) costs and benefits to mankind, if not to the whole ecosystem, in developing a project one way rather than another. To date, linguistics has had no specialization area comparable to macroecology. We have a vague notion of what constitutes the ecology of a language, how it affects language competition and selection - in spite of vague invocations of globalization and/or colonization - and what conditions are most (dis)favorable to languages. 17. Monolingual countries that question the generalization include Ruanda and Burundi, which have maintained ethnic divisions and are not more developed than their multilingual neighbors. Multilingual countries that also dispute the rule include Belgium and Switzerland. 18. In Kinshasa, for instance, more and more members of the government who have migrated from non-Lingalaphone parts of the DRC have found it imperative to develop fluent competence in the local vernacular in order not to be treated as outsiders. This is a good incentive for letting their own children develop native competence in Lingala (very often at the expense of French), which entitles them to claim cultural roots in the capital.

Chapter 13 1. Some earlier variety of Lingala, apparently a form of Bobangi, had already functioned as a trade language along the Congo River before the arrival of the Europeans. 2. Because the official languages are transmitted primarily through the school system, the development of competence in them is financially costlier than that of the indigenous lingua francas, which are acquired naturalistically even by adult L2 learners. The investment is therefore weighed against the expected profit. 3. The Afrikaners, who once had kept Blacks from learning their language (from the late nineteenth century to the mid-twentieth century), now imposed it on them as the European language in the school system but prevented

311

312 Notes them from learning English, the language that would open the doors of economic success and the outside world to them. 4. The South African road infrastructure also leads to this conclusion. The Bantu villages are far from and not easily accessible from the country's first rate highway system. It is as if no particular effort has been made to connect the relevant populations with the modern economy that is still largely controlled by the White population, although not every SA White is affluent.

Chapter 14 1. I am grateful to Sali Tagliamonte for comments on a draft of the 1997 ancestor of this Chapter. I am alone responsible for all the remaining shortcomings. 2. According to Bailey & Thomas (1998), the divergence started in the late nineteenth century, which is more consistent with the history of the American South, especially if the onset of the process follows the passage of the Jim Crow laws, which institutionalized the residential segregation of White and Black populations. But see also Fasold (1987), Butters (1991), and Rickford (1992) for different perspectives on the divergence hypothesis. 3. An important factor here is the stabilization and growth of the African American population by birth rather than by importation or by immigration. On the other hand, the immigration process increased among European Americans, offsetting drastically the general race demographic disproportion but bringing no end to the dejure and later de facto segregation. Thus, marginal regular or intimate social intercourse between some European and African Americans hardly favored generalized influence of European American English on African American speech. 4. The former's data produced, among other things, Jones-Jackson (1978,1987). 5. I use community between quotes to reflect how African American leaders refer to the African American community, especially the less affluent and typically segregated part of it, with more or less the meaning of "neighborhood." Otherwise, the term retains its ordinary meaning in the linguistic literature. 6. One must remember that while we linguists have so typically focused on linguistic features of AAE, what matters to African Americans in terms of establishing one's linguistic identity is not so much speaking AAE but "sounding Black" (Kashif 1991, Mufwene 2001b), using features (structural and non-structural) that are typical of African Americans. Thus the vast majority of African American leaders "sound Black" but do not speak AAE, at least not by the criteria assumed in, for instance, the quantitative (socio)linguistics literature. 7. This phenomenon is similar to Labov's (1963) observation that on Martha's Vineyard Island the most "hypercorrect" Vineyarder features are attested in those who are the most strongly opposed to outside "invasion" and/or made a conscious decision to stay on the island while they could earn better living elsewhere. A similar phenomenon has likewise been observed among some traditional residents of Ocracoke Island, especially those opposed to the in-migrants who have changed their local lifestyle (Wolfram & Schilling-Estes 1995). 8. I have been told anecdotally that children left alone watching TV without interaction with the speaking community have failed to learn the language spoken on this medium. It would thus take more than simple exposure to the media to learn to speak any variety, except of course in cases where one deliberately selects such a medium as the source of linguistic information. Thus a few Africans in countries surrounding the DRC and the Republic of the Congo have learned to speak a little Lingala from radio and TV music programs by sheer determination to understand the music. On the other hand, there are also a lot of Congolese who sing those Lingala songs without understanding much of what the words mean. 9. In the case of Georgia, the depopulation was initiated by the State in order to develop wildlife reserves.

Bibliography Aboh, Enoch. 2006. The role of the syntax-semantics interface. In Lefebvre etal.y eds., 253-75. 2007. Le genese de la peripherie gauche du saramaka: un cas d'influence du substrat? In Grammaires Creoles et grammaire comparative, ed. Karl Gadelli & Anne Zribi-Hertz, 73-97. Paris: Presses Universitaires de Vincennes. Adam, Lucien. 1882. Les classifications, I'objet, la methode, les conclusions de la linguistique. Paris: Paris VII. 1883. Les idiomes negro-aryens et malayo aryens: essai d'hybridologie linguistique. Paris: Maisonneuve. Aikhenwald, Alexandra, & R. M. W. Dixon (eds.). 2001. Areal Diffusion and Genetic Inheritance: Case Studies in Language Change. Oxford: Oxford University Press. Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, & Remain Wacziarg. 2003. Fractionalization. Journal of Economic Growth 8.155-94. Algeo, John. 1991. Language. In The Reader's Companion to American History, ed. Eric Foner and John A. Garraty, 637-40. Boston: Houghton Mifflin Co. Alleyne, Mervyn C. 1971. Acculturation and the cultural matrix of creolization. In Hymes, ed., 169-86. 1979. On the genesis of languages. In Hill, ed., 89-107. 1980. Comparative Afro-American: An Historical-Comparative Study of English-Based Afro-American Dialects of the New World. Ann Arbor: Karoma Publishers, Inc. 1986. Substratum influences: Guilty until proven innocent. In Muysken & Smith, eds., 301-15. 1993. Continuity vs. creativity in Afro-American culture. In Mufwene, ed., 167-81. 1996. Syntaxe historique Creole. Paris: Karthala. Allsopp, Richard. 1977. Africanisms in the idioms of Caribbean English. In Language and Linguistics Problems in Africa, ed. Paul F. Kotey & Haig Der-Houssikian, 429-41. Columbia: Hornbeam Press, Inc. American Association of Anthropology. 1999. AAA Statement on Race. American Anthropologist 100.712-13. American Association of Physical Anthropology. 1999. AAPA Statement on the Biological Aspects of Race. American Anthroplogist 100.714-15. Andersen, Roger. 1983. Transfer to somewhere. In Language Transfer in Language Learning, ed.Susan Gass & Larry Selinker, 177-201. Rowley: Newbury House. Andersen, Roger (ed.). 1983. Pidginization and Creolization as Language Acquisition. Rowley, MA: Newbury House. Andersen, Roger. 1990. The Papiamentu tense-aspect system, with special attention to discourse. In Pidgin and Creole Tense-Mood-Aspect Systems, ed. John V. Singler, 59-96. Amsterdam: John Benjamins. Ans, Andre-Marcel d'. 1987. Haiti: paysage, societe. Paris: Karthala. 1996. Essai de sociolinguistique historique & partir d'un temoignage inedit sur 1'emploi des langues, notamment africaines, en Haiti, au cours de la guerre de liberation et des premieres annees de 1'independance. Etudes Creoles 19.110-22.

314

Bibliography Ansaldo, Umberto, Lisa Lim, & Salikoko S. Mufwene. 2007. The Peranakans as Southeast Asia's Creoles: A languagecentered history. In Deconstructing Creole, ed. Umberto Ansaldo, Stephen Matthews, & Lisa Lim, 203-26. Amsterdam: John Benjamins. Ansaldo, Umberto, & Stephen Matthews. 1999. The Minnan substrate and creolization of Baba Malay. Journal of Chinese Linguistics 27.38-68. Appel, Rene, & Pieter Muysken. 1987. Language Contact and Bilingualism. London: Edward Arnold. Arends, Jacques. 1989. Syntactic developments in Sranan: Creolization as a gradual process. Doctoral thesis. To appear. A demographic perspective on Creole formation. In Handbook of Pidgins and Creoles, ed. Silvia Kouwenberg & John Singler. Oxford: Maiden, MA: Blackwell. Arends, Jacques, Pieter Muysken, & Norval Smith (eds.). 1995. Pidgins and Creoles: An Introduction. Amsterdam: John Benjamins. AuCoin, Michelle M. 2002. The sociohistorical and linguistic development of African American English in Virginia and South Carolina. PhD dissertation, University of Chicago. Bailey, Charles-James N., & Karl Maroldt. 1977. The French lineage of English. In Meisel, ed., 21-53. Bailey, Guy, & Natalie Maynor. 1987. Decreolization? Language in Society 16.449-73. 1989. The divergence controversy. American Speech 64.12-39. Bailey, Guy, & Erik Thomas. 1998. Some aspects of African-American vernacular English phonology. In Mufwene, Rickford, Bailey, & Baugh, eds., 85-109. Baissac, Charles. 1880. Etude sur le patois Creole mauricien. Nancy: Imprimerie Berger-Levrault. Baker, Philip. 1984. Agglutinated French articles in Creole French: Their evolutionary significance. Te Reo 27.89-129. 1990: Off target? Column. Journal of Pidgin and Creole Languages 5.107-19. 1993a. Assessing the African contribution to French-based Creoles. In Mufwene, ed., 123-55. 1993b. Australian influence in Melanesian Pidgin English. Te Reo 36.3-67. 1995a. Motivation in Creole genesis. In From Contact to Creole and Beyond, ed. Philip Baker, 3-15. London: University of Westminster Press. 1995b. Some developmental inferences from the historical studies of pidgins and Creoles. In The Early Stages of Creolization, ed. Jacques Arends, 1-24. Amsterdam: John Benjamins. 1997. Directionality in pidginization and creolization. In Spears & Winford, eds., 91-109. 2001. Chinese Pidgin English: The making of a pidgin for all seasons. Paper presented at the University of Westminster Workshop on (the Definition of) Pidgins and Creoles. London: April. Baker, Philip, & Adrienne Bruyn (eds.). 1998. St. Kitts and the Atlantic Creoles: The Texts of Samuel Augustus Matthews in Perspective. London: University of Westminster Press. Baker, Philip, & Chris Corne. 1986. Universals, substrata and the Indian Ocean Creoles. In Muysken & Smith, eds., 163-83. Baker, Philip, & Peter Miihlhausler. 1990. From business to pidgin. Journal of Asian Pacific Communication 1.87-115. Baker, Philip, &Anand Syea (eds.). 1996. Changing Meanings, Changing Functions: Papers Relating to Grammaticalization in Language Contact. London: University of Westminster Press. Bakker, Peter. 1997. A Language of Our Own: The Genesis ofMichif,

the Mixed Cree-French Language of the Canadian

Metis. Oxford: Oxford University Press. Bamgbose, Ayo. 1992. Standard Nigerian English: Issues of identification. In The Other Tongue: English Across Cultures, ed. Braj Kachru, 148-61. Urbana: University of Illinois Press. Banner, Stuart. 2005. How the Indians lost their land: Law and power on the frontier. Cambridge, MA: Belknap, Harvard University Press.

Bibliography Baptista, Marlyse. 2004. The Syntax of Cape Vedean Creole: The Sotavento Varieties. Amsterdam: John Benjamins. Bardovi-Harlig, Kathleen. 1987. Markedness and salience in second-language acquisition. Language Learning 37.385^07. Bastide, Roger. 1967. Les Ameriques noires. Paris: Payot. Batibo, Herman M. 1998. The fate of the Khoesan languages of Botswana. In Brenzinger, ed., 267-84. 2005. Language Decline and Death in Africa: Causes, Consequences and Challenges. Clevedon, UK: Multilingual Matters. 2006. Language Decline and Death in Africa: Causes, Consequences, and Challenges. Clevedon, UK: Multingual Matters. Beechert, Edward D. 1985. Working in Hawaii: A Labor History. Honolulu: University of Hawaii Press. Bennett, John. 1908-1909. Gullah: A Negro patois. South Atlantic Quarterly 7.332-47 and 8.39-52. Berlin, Ira. 1998. Many Thousands Gone: The First Two Centuries of Slavery in North America. Cambridge, MA: Harvard University Press. Bertrand-Bocande, Emmanuel. 1849. Notes sur la Guinee portugaise ou Senegambie meridionale. Bulletin de la Societe de Geographie 12.57-93. Bhatia, Tej K. 1988. Trinidad Hindi: Its genesis and generational profile. In Language Transplanted: The Development of Overseas Hindi, ed. Richard K. Barz & Jeff Siegel, 179-96. Wiesbaden: Otto Harrassowitz. Bickerton, Derek. 1973. On the nature of a Creole continuum. Language 49.640-69. 1981. Roots of Language. Ann Arbor: Karoma Publishers, Inc. 1984a. The language bioprogram hypothesis and second language acquisition. In Language Universals and Second Language Acquisition, ed. William E. Rutherford, 141-61. Amsterdam: John Benjamins. 1984b. The language bioprogram hypothesis. Behavioral and Brain Sciences 7.173-221. 1988. Creole languages and the bioprogram. In Linguistics: The Cambridge Survey. Volume 2: Linguistic Theory: Extensions and Implications, ed. Frederick J. Newmeyer, 268-84. Cambridge: Cambridge University Press. 1989. The lexical learning hypothesis and the pidgin-creole cycle. In Wheels within Wheels: Papers of the Duisburg Symposium on Pidgin and Creole Languages, ed. Martin Piitz & Rene Dirven, 11-31. Frankfurt am Main: Verlag Peter Lang. 1990. Language and Species. Chicago: University of Chicago Press. 1999. How to acquire language without positive evidence: What acquisitionists can learn from Creoles. In DeGraff, ed., 49-74. 2000. How protolanguage became language. In The Evolutionary Emergence of Language: Social Function and Origins of Linguistic Form, ed. Chris Knight, Michael Studdert-Kennedy, & James R. Hurford, 264-84. Cambridge: Cambridge University Press. 2004. Reconsidering Creole exceptionalism. Language 80.828-33. Bilby, Kenneth. 2003. Performing African nations in Jamaica. Paper presented at the "Recreating Africa in America" conference, De Paul University, Chicago. Billings, Sabrina. 2006. Speaking beauties: Language use and linguistic ideologies in Tanzanian beauty pageants. PhD dissertation, University of Chicago. Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart & Winston. Boas, Franz. 1894. Human faculty as determined by race. Proceedings of the American Association for the Advancement of Science 43.301-27. Bolinger, Dwight. 1973. Getting the words in. In Lexicography in English, ed. Raven McDavid, Jr. & Audrey Duckert, 8-13. New York: New York Academy of Science.

315

316 Bibliography Bolton, Kingsley. 2000. Language and hybridization: Pidgin tales from the China coast. Interventions 5.35-52. 2002. Chinese Englishes: From Canton jargon to global English. World Englishes, 21.181-99. 2003. Chinese Englishes: A Sociolinguistic History. Cambridge: Cambridge University Press. Bopp, Franz. 1833. Vergleichende Grammatik des Sanskrit, Zend, Griechischen, Lateinischen, Litthauischen Gothischen und Deutschen. Berlin: F. Diimmler. Boretzky, Norbert. 1983. Kreolsprachen Substrate und Sprachwandel. Wiesbaden: Otto Harrassowitz. 1993. The concept of rule, rule borrowing, and substrate influence in Creole languages. In Mufwene, ed., 74-92. Botha, Rudolf. 2006a. On the windows approach to language evolution. Language and Communication 26.129-43. 2006b. Pidgin languages as a putative window on language evolution. Language and Communication 26.1-14. Bourdieu, Pierre. 1991. Language and Symbolic Power. Cambridge, MA: Harvard University Press. Boyle, John. 2007. Language eradication among the Native North Americans. Paper presented at the 11th African, African American, Native American, Caribbean and the Americas Heritage Conference. Northern Illinois University, Chicago. Brasch, Walter M. 1981. Black English and the Mass Media. Lanham: University of America Press. Brasseur, Patrice. 1997. Creoles a base lexicale francaise et fran^ais marginaux d'Amerique du Nord. In Contacts de langues, contacts de cultures, creolisation, ed. Marie-Christine Hazael-Massieux & Didier de Robillard, 141-66. Paris: L'Harmattan. Brenzinger, Matthias (ed.). 1992. Language Death: Factual and Theoretical Explorations with Special Reference to East Africa. Berlin: Mouton de Gruyter. 1998. Endangered Languages in Africa. Cologne: Riidiger Koppe Verlag. Breeder, Peter, Guus Extra, & Jeanne Maartens. 2002. Multilingualism in South Africa, with a Focus on KwaZulu-Natal and Metropolitan Durban. Cape Town: PRAESA, University of Cape Town. Brown, James H. 1995. Macroecology. Chicago: University of Chicago Press. Brutt-Griffler, Janina. 2002a. World English: A Study of Its Development. Clevedon: Multilingual Matters Ltd. 2002b. Class, ethnicity, and language rights: An analysis of British colonial policy in Lesotho and Sri Lanka and some implications for language policy. Journal of Language, Identity, and Education 1.207-34. 2005. Race and ethnicity in the English-speaking world. In English and Ethnicity, ed. Janina Brutt-Griffler & Catherine E. Davies, 277-99. New York: Palgrave Macmillan Ltd. Bruyn, Adrienne. 1996. On identifying instances of grammaticalization in Creole languages. In Baker & Syea, eds., 29-46. Bunting, Jacqueline. To appear. Give and Take: The Role of'give' In One Sranan Transitivity Shift. Butters, Ronald. 1991. The Death of Black English: Divergence and Convergence in Black and White Vernaculars. Frankfurt am Main: Peter Lang. Bybee, Joan, Revere Perkins, & William Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Calvet, Louis-Jean. 1998. Language Wars and Linguistic Politics. Oxford: Oxford University Press. 2002. Le marche aux langues et les effets linguistiques de la mondialisation. Paris: Plon. 2004. Essais de linguistique: la langue est-elle une invention des linguistes? Paris: Plon. 2006. Towards an Ecology of World Languages. Cambridge: Polity. Campbell, Lyle. 1998. Historical Linguistics: An Introduction. Cambridge, MA: MIT Press. Cantoni, Gina (ed.). 1997. Stabilizing Indigenous Languages. Northern Arizona University: Center for Excellence in Education. Carroue, Laurent. 2002. Geographic de la mondialisation. Paris: Armand Colin.

Bibliography Caspar!, Rachel. 2003. From types to populations: A century of race, physical anthropology, and the American Anthropological Association. American Anthropologist 105.65-76. Cassidy, Frederick G. 1980. The place of Gullah. American Speech 55.3-16. 1983. Sources of the African element in Gullah. In Studies in Caribbean Language, ed. Lawrence D. Carrington, 75-81. University of the West Indies, St. Augustine: Society for Caribbean Linguistics. Cavalli-Sforza, Luigi Luca, & Francesco Cavalli-Sforza. 1995. Great Human Diasporas. Cambridge, MA: Perseus Books. Chase, Richard. 1943. The Jack Tales: Folk Tales from the Southern Appalachians. Boston: Houghton Mifflin. Chaudenson, Robert. 1973. Pour une etude comparee des Creoles et parlers fran9ais d'outre-mer: Survivance et innovation. Revue de Linguistique Romane 37.342-70. 1979. Les Creoles francais. Paris: Fernand Nathan. 1985. Francais avance, franc, ais zero, Creoles. In Actes du XVIIe dCongres de Linguistique et Philologie Romanes. Vol. 5: Sociolinguistique des langues romanes, 165-80. Aix-en-Provence: Publications de 1'Universite de Provence. 1986. And they had to speak any way... Acquisition and creolization of French. In The Fergusonian Impact. Volume 1: From Phonology to Society, ed. Joshua A. Fishman, A. Tabouret-Keller, M. Clyne, B. Krishnamurti, & M. Abdulaziz, 69-82. Berlin: Mouton de Gruyter. 1990. Du mauvais usage du comparatisme: le cas des etudes Creoles. Travaux du Cercle Linguistique d'Aixen-Provence 8.123-58. 1992. Des ties, des hommes, des langues: essais sur la creolisation linguistique et culturelle. Paris: L'Harmattan. 2001. Creolization of Language and Culture. London: Routledge. 2003. La creolisation: theorie, applications, implications. Paris: L'Harmattan. Chaudenson, Robert, Raymond Mougeon, & Edouard Beniak. 1993. Vers une approche panlectale de la variation du francais. Aix-en-Provence : Institut d'Etudes Creoles et Francophones. Chaudenson, Robert, Andre Valli, & Daniel Veronique. 1986. The dynamics of linguistic systems and the acquisition of French as a second language. Studies in Second Language Acquisition 8.277-92. Chomsky, Noam. 1986. Knowledge of Language: Its Nature, Origin, and Use. New York: Praeger. 2000. New Horizons in the Study of Language and Mind. Cambridge: Cambridge University Press. Christaller, Rev. J.G. 1933. Dictionary of the Asante and Fante language called Tshi (Twi). Basel: Basel Evangelical Missionary Society. Clyne, Michael. 2003. Dynamics of Language Contact: English and Immigrant Languages. Cambridge: Cambridge University Press. Collins, James. 1984. Malaysian and Bazaar Malay: Polarity, continuity and communication. In National Language and Communication in Multilingual Societies, ed. Asmah Haji Omar, 151-74. Kuala Lumpur: Dewan Bahasa dan Pustaka. Comrie, Bernard. 2000. From potential to realization: An episode in the origin of language. Linguistics 38.989-1004. Come, Chris. 1999. From French to Creole: The Development of New Vernaculars in the French Colonial World. London: University of Westminster Press. Couto, Hildo Honorio do. 2007. Ecolinguistica: estudos das relacoes entre lingua e meio embiente. Brasilia: Thesaurus. Cowen, Noel. 2001. Global History: A Short Overview. Cambridge: Polity Press. Craig, Colette. 1991. Ways to go in Rama: A case study in polygrammaticalization. In Traugott & Heine, eds., 455-92. Croft, William. 1990. Typology and Universals. Cambridge: Cambridge University Press. 2000. Explaining Language Change: An Evolutionary Approach. London: Longman. Crosby, Alfred W. 1986. Ecological Imperialism: The Biological Expansion of Europe, 900-1900. Cambridge: Cambridge University Press.

317

318

Bibliography 1992. Ills. In Atlantic American Societies: From Columbus Through Abolition 1492-1888, ed. Alan L. Karras & John Robert McNeill, 19-39. London: Routledge. Crystal, David. 1997. English as a Global Language. Cambridge: Cambridge University Press. 2000. Language Death. Cambridge: Cambridge University Press. 2004. The Language Revolution. Cambridge: Polity Press. Cukor-Avila, Patricia, & Guy Bailey. 1996. The spread of urban AAVE: A case study. In Sociolinguistic Variation: Data, Theory, and Analysis, ed. Jennifer Arnold, Renee Blake, Brad Davidson, Scott Schwenter, & Julie Solomon, 469-85. Stanford: CSLI Publications. Daelman, Jan, S.I. 1972. Kongo elements in Saramacca Tongo. Journal of African Languages 11.1-44. Dalby, Andrew. 2002. Language in Danger. New York: Columbia University Press. Dalby, David. 1971. Communication in Africa and the New World. In Black-White Speech Relations, ed. Walt Wolfram & Nona H. Clarke, 99-138. Arlington: Center for Applied Linguistics. Dance, Daryl Cumber. 1978. Shuckin and Jivin: Folklore from Contemporary Black Americans. Bloomington: Indiana University Press. Darwin, Charles. 1859. On the Origin of Species. London: John Murray. 1871. The Descent of Man. Amherst, NY: Prometheus Books. Dawkins, Richard. 1976. The Selfish Gene. Oxford: Oxford University Press. Deacon, Terrence W. 1997. The Symbolic Species: The Co-evolution of Language and the Brain. New York: W.W. Norton &Co.

DeBose, Charles, & Nicholas Faraclas. 1993. Getting to the roots of tense/aspect and copula systems in Afro-American. In Mufwene, ed., 1993,364-87. DeCamp, David. 1971. Toward a generative analysis of a post-creole speech continuum. In Hymes, ed., 349-70. DeGraff, Michel. 1992. The syntax of predication in Haitian. In Proceedings of the 22nd Meeting of the North-Eastern Linguistics Society, ed. Michel DeGraff, 103-17. Amherst: University of Massachusetts. 1998. Non-verbal predication in Haitian and in Irish. In Proceedings of the Sixteenth West Coast Conference on Formal Linguistics, ed. Michel DeGraff, 113-28. Stanford: Center for the Study of Language and Information. 1999a. Creolization, language change, and language acquisition: A prolegomenon. In DeGraff, ed., 1-46. 1999b. Creolization, language change, and language acquisition: An epilogue. In DeGraff, ed., 473-543. DeGraff, Michel (ed.). 1999. Language Creation and Language Change: Creolization, Diachrony, and Development. Cambridge, MA: MIT Press. DeGraff, Michel. 200la. On the origin of Creoles: A Cartesian critique of neo-Darwinian linguistics. Linguistic Typology 5.213-310. 200Ib. Morphology in Creole genesis: Linguistics and ideology. In Ken Hale: A Life in Language, ed. Michael Kenstowicz, 53-121. Cambridge, MA: MIT Press. 2002. Relexification: A reevaluation. Anthropological Linguistics 44.321-414. 2003. Against Creole exceptionalism. Discussion note. Language 79.391-410. 2004. Against Creole exceptionalism (redux). Language 80.834-9. 2005. Morphology and word order in "creolization" and beyond. In Comparative Syntax, ed. Guglielmo Cinque & Richard S. Kayne, 293-372. Oxford: Oxford University Press. Dejean, Yves. 1993. An overview of the language situation in Haiti. International Journal of the Sociology of Language 102.73-83. Deutscher, Guy. 2005. The Unfolding of Language: An Evolutionary Tour of Mankind's Greatest Invention. New York: Henry Holt and Company.

Bibliography Dillard, J. L. 1970. Principles in the history of American English: Paradox, virginity, and cafeteria. Florida FL Reporter 8.32-3. 1972. Black English; Its History and Usage in the United States. New York: Random House. Dillard, J. L. (ed.). 1980. Perspectives on American English. The Hague: Mouton. Dillard, J. L. 1985. Toward a Social History of American English. New York: Mouton. Dimmendaal, Gerrit Jan. 1983. The Turkana Language. Dordrecht/Cinnaminson: Foris Publications. Dixon, R. M. W. 1982. Where Have All the Adjectives Gone? And Other Essays in Semantics and Syntax. Berlin: Mouton. 1997. The Rise and Fall of Languages. Cambridge: Cambridge University Press. Dollfus, Olivier. 2001. La mondialisation. Paris: Presses des Sciences Po. Dorian, Nancy. 1981. Language Death: The Language Cycle of a Scottish Gaelic Dialect. Philadelphia: University of Pennsylvania Press. 1983. Natural second language acquisition from the perspective of the study of language death. In Pidginization and Creolization as Second Language Acquisition, ed. Roger Andersen, 158-67. Rowley, MA: Newbury House. Dorian, Nancy (ed.). 1989. Investigating Obsolescence: Studies in Language Contraction and Death. Cambridge: Cambridge University Press. Dunn, Ernest F. 1976. The black/southern white dialect controversy: Who did what to whom? In Black English: A Seminar, ed. Deborah Sears Harrison & Tom Trabasso, 105-22. Hillside, NJ: Lawrence Erlbaum Associates. Dunn, Richard S. 1972/1992. Sugar and Slaves: the Rise of the Planter Class in the English West Indies, 1624-1713. Chapel Hill: The University of North Carolina Press. Elston, Heidi. 2003. Focusing on the "footpath": a critique of Keller's On Language Change. In Contemporary Linguistics 3: Perspectives on Language Evolution, ed. Salikoko S. Mufwene, 27-37. University of Chicago, Department of Linguistics. Erhardt, Sabine. 1993. Le Creole francais de Saint-Louis (le tayo) en Nouvelle Caledonie. Hamburg: Helmut Buske. Fabian, Johannes. 1986. Language and Colonial Power: the Appropriation of Swahili in the Former Belgian Congo, 1880-1938. Cambridge: Cambridge University Press. Faine, Jules. 1937. Philologie Creole: etudes historiques et etymologiques sur la langue Creole d'Haiti. Port-au-Prince: Imprimerie de FEtat. Faraclas, Nicholas. 1987. Creolization and the tense-aspect-modality system of Nigerian Pidgin. Journal of African Languages and Linguistics 9.45-59. 1988a. Nigerian Pidgin and the languages of southern Nigeria. Journal of Pidgin and Creole Languages 3.177-97. 1988b. Rumors and the demise of Descartes are premature: Review of Pidgin and Creole Linguistics, by Peter Muhlausler. Journal of Pidgin and Creole Languages 3.119-35. Fasold, Ralph (ed.). 1987. Are black and white vernaculars diverging? Papers from the NWAVE XIV Panel discussion. American Speech 62.1. Fehderau, Harold W. 1966. The origin and development of Kituba (lingua franca Kikongo). PhD dissertation, Cornell University. Fenton, Steve. 2003. Ethnicity. Cambridge: Polity Press. Feral, Carole de. 1989. Pidgin-English du Cameroun. Paris: Peters/SELAF. Ferraz, Luis Ivens. 1979. The Creole of Sao Tome. Johannesburg: Witwatersrand University Press. 1984. The substrate of Annobonese. African Studies 43.119-36. Ferreira, Jo-Anne Sharon. 1999. The Portuguese language in Trinidad and Tobago: A study of language shift and language death. PhD dissertation, The University of the West Indies, St. Augustine, Trinidad. Fill, Alwin. 2004. Review of Mufwene 2001. Journal of Pidgin and Creole Languages 19.199-203.

319

320

Bibliography Filppula, Markku, Juhani Klemola, & Heli Pitkanen (eds.). 2002. The Celtic Roots of English. Studies in Languages series, number 37. Joensuu: Joensuu University Faculty of Humanities. Fischer, David Hackett. 1989. Albion's Seed: Four British Folkways in America. Oxford: Oxford University Press. Fishman, Joshua A. 1991. Reversing Language Shift: Theoretical and Empirical Foundations of Assistance to Threatened Languages. Clevedon, UK: Multilingual Matters. 2002. Endangered minority languages: Prospects for sociolinguistic research. MOST Journal of Multicultural Societies 4.270-5. www.unesco.org/shs/ijms/vol4/issue2/art7 Forston, Benjamin W. 2004. Indo-European Language and Culture: An Introduction. Maiden, MA: Blackwell. Fracchia, Joseph, & R. C. Lewontin. 1999. Does culture evolve? History and Theory: Studies in the Philosophy of History 38, no. 4,52-78. Frajzyngier, Zygmunt. 1984. On the origin of say and se as complementizers in Black English and English-based Creoles. American Speech 59.207-10. Freeman, Edward Augustus. 1881. The Historical Geography of Europe. Vol. 1. London: Longmans, Green, & Co. 1886. Greater Greece and Greater Britain; George Washington: The Expander of England. Two lectures with an Appendix. London: Macmillan. 1892. Historical Essays. London: Macmillan. Friedman, Victor A. 1996. Observing the observers: Language, ethnicity, and power in the 1994 Macedonian census and beyond. In Toward Comprehensive Peace in Southeastern Europe: Conflict Prevention in the South Balkans, ed. Barnett R. Rubin, 81-105. New York: Twentieth Century Fund Press. 2001. The Vlah minority in Macedonia: language, identity, and standardization. In Selected Papers in Slavic, Balkan, and Balkan Studies (Slavica Helsingiensa 21), ed. Juhani Nuoluoto, Marti Leiwo, and Jussi Halla-aho, 26-50. University of Helsinki. Garnsey, Peter, & Richard Sailer. 1987. The Roman Empire: Economy, Society, and Culture. Berkeley: University of California Press. Giacomi, Alain, Henriette Stoffell, & Daniel Veronique (eds.). 2000. Approximations du franc,ais par des Marocains arabophones a Marseilles. Aix-en-Provence: Publications de PUniversite de Provence. Gilbert, Glenn G. (ed. & trans.). 1980. Pidgin and Creole Languages: Selected Essays by Hugo Schuchardt. London: Cambridge University Press. Gilbert, Glenn. 1986. The language bioprogram hypothesis: Deja vw? In Muysken & Smith, 15-24. Giliomee, Harmann. 2003. The Rise and Possible Demise of Afrikaans as a Public Language. Cape Town: PRAESA, University of Cape Town. Gilman, Charles. 1986. African areal characteristics: Sprachbund, not substrate? Journal of Pidgin and Creole Languages 1.33-50. 1993. Black identity, homeostasis, and survival: African and metropolitan speech varieties in the New World. In Mufwene, ed., 1993,388-402. Givon, Talmy. 1971. On the verbal origin of Bantu verb suffixes. Studies in African Linguistics 2.145-63. 1979. Prolegomena to any sane creology. In Readings in Creole Studies, ed. Ian E Hancock, in collaboration with Edgar Polome, Morris Goodman, & Bernd Heine, 3-35. Ghent: E. Story-Scientia. 1998. On the co-evolution of language, mind, and brain. Evolution of Communication 2.45-116. Givon, Talmy, & Bertram F. Malle (eds.). 2002. The Evolution of Language out of Pre-language. Amsterdam: John Benjamins. Goldin-Meadow, Susan. 2002. Getting a handle on language creation. In The Evolution of Language out of Pre-language, ed. Talmy Givon & Bertram F. Malle, 343-74. Amsterdam: John Benjamins.

Bibliography Golovko, Eugeni V., & Nikolai Vakhtin. 1990. Aleut in contact: The CIA enigma. Acta Linguistica Hafniensia 22.97-125. Gonzales, Ambrose E. 1922. The Black Border: Gullah Stories of the Carolina Coast (with a Glossary). Columbia, SC: The State Co. Goodman, Martin. 1991. Babatha's story. Journal of Roman Studies 81.169-75. Goodman, Morris. 1964. A Comparative Study of Creole French Dialects. The Hague: Mouton de Gruyter. 1985. Review article on Bickertons Roots of Language. International Journal of American Linguistics 51.109-37. 1993. African substratum: Some cautionary words. In Mufwene, ed., 1993,64-73. 1999. Review of Problemes de genese et de description, by Guy Hazael-Massienx. Journal of Pidgin and Creole Languages 14. 398-403. Goodnight, Charles J., & Michael J. Wade. 1999. The ongoing synthesis: A reply to Coyne, Barton, & Turelli. Evolution 54.317-24. Gould, Stephen lay. 1993. Eight Little Piggies: Reflections in Natural History. New York: W. W. Norton & Co. Grenoble, Lenore A., & Lindsay J. Whaley (eds.). 1998. Endangered Languages: Current Issues and Future Prospects. Cambridge: Cambridge University Press. Grenoble, Lenore A., & Lindsay J. Whaley. 2006. Saving Languages: An Introduction to Language Revitalization. Cambridge: Cambridge University Press. Guglielmo, Jennifer, & Salvatore Salerno (eds.). 2003. Are Italians White? How Race is Made in America. New York: Routledge. Hagege, Claude. 1985. L'homme de paroles: contribution linguistique aux sciences humaines. Paris: Fayard. 1992. Le souffle de la langue. Paris: Odille Jacob. 1993. The Language Builder: An Essay on the Human Signature in Linguistic Morphogenesis. Amsterdam: John Benjamins. 2000. Halte a la mort des langues. Paris: Editions Odile Jacob. 2001. Les processus de grammaticalisation. In Language typology and language universals, ed. Martin Haspelmath, Ekkehard Konig, Wulf Oesterreicher, & Wolfgang Raible, 1609-24. Berlin: Walter de Gruyter. 2006. Combat pour lefrancais: au nom de la diversite des langues et des cultures. Paris: Odile Jacob. Hair, P. E. H. 1965. Sierra Leone items in the Gullah dialect of American English. Sierra Leone Language Review 4.79-84. Hale, Ken. 1992. On endangered languages and the safeguarding of diversity. Language 68.1-3. 1996. Universal Grammar and the roots of linguistic diversity. In MIT Working Papers in Linguistics 28: Papers on language endangerment and the maintenance of linguistic diversity, ed. Jonathan David Bobaljik, Rob Pensalfini, & Luciano Storto, 137-61. Hall, Robert A., Jr. 1950. The genetic relationships of Haitian Creole. Ricerche Linguistiche 1.194-203. 1962. The life-cycle of pidgin languages. Lingua 11.151-6. 1966. Pidgin and Creole Languages. Ithaca: Cornell University Press. Hancock, Ian E 1980a. Texan Gullah: The Creole English of the Brackettville Afro-Seminoles. Perspectives on American English, ed. J. L. Dillard, 305-32. The Hague: Mouton. 1980b. Texan Gullah: The Creole English of the Brackettville Afro-Seminoles. In Dillard, ed., 305-32. 1986. The domestic hypothesis, diffusion, and componentiality: An account of Atlantic Anglophone Creole origins. In Muysken & Smith, eds., 71-102. 1993. Creole language provenance and the African component. In Mufwene, ed., 1993, 182-91. Hanski, Ilkka. 1996. Metapopulation ecology. In Population Dynamics in Ecological Space and Time, ed. Olin E. Rhodes, Jr., Ronald K. Chesser, & Michael H. Smith, 13-43. Chicago: University of Chicago Press.

321

322

Bibliography Harmon, David. 2002. In Light of Our Differences: How Diversity in Nature and Culture Makes Us Human. Washington, DC: Smithsonian Institution Press. Harrison, Faye V. 1995. The persistent power of "race" in the cultural and political economy of racism. Annual Review of Anthropology 24.47-74. Hatala, E. 1976. Environmental effects on white students in black schools. Master's thesis, University of Pennsylvania. Haugen, Einar. 1953. The Norwegian Language in America: A Study in Bilingual Behavior. Philadelphia: University of Pennsylvania Press. 1971. The ecology of language. The Linguistic Reporter, Supplement 25.19-26. Reprinted in The Ecology of Language (1972), ed. Anwar S. Dil, 325-39. Stanford: Stanford University Press. Hazael-Massieux, Guy. 1982. Ambiguite genetique et bipolarite dans le fonctionnement du Creole de Guadeloupe. Etudes Creoles 4.47-62. Hazael-Massieux, Marie-Christine. 1999. Les Creoles: I'indispensable survie. Paris: Editions Entente. Heine, Bernd. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. New York and Oxford: Oxford University Press. 2001. Cognitive Foundations of Grammar. New York: Oxford University Press. in progress. Creating new functional categories: The case of genitives in Creoles. Ms. Heine, Bernd, Ulrike Claudi, & Friederike Hiinnemeyer. 1991. Grammaticalization: A Conceptual Framework, Chicago: Unriversity of Chicago Press. Heine, Bernd, & Tania Kuteva. 2002. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Herskovits, Melville. 1935. The social history of the Negro. In A Handbook of Social Psychology, ed. Carl Murchison, 207-67. Worcester, MA: Clark University Press. 1936. The significance of West Africa for Negro research. Journal of Negro History 21.15-30. 1938. The ancestry of the American Negro. The American Scholar 8.84-94. 1941. The Myth of the Negro Past. New York: Harper & Brothers. 1975. Life in a Haitian Valley. New York: Octagon Books. Hesseling, Dirk Christiaan. 1897. Het Hollandisch in Zuid-Afrika. De Gids 61.138-62. 1923. Het Frans in Noord-Amerika en het Nederlands in Zuid-Afrika. De Gids 87.438-57. 1933. Papiamentu en Negerhollands. TNTL 52.265-88. 1979. On the Origin and Formation of Creoles: A miscellany of articles, trans. & ed. Thomas L. Markey & Paul T. Roberge. Ann Arbor: Karoma Publishers, Inc. Hey, Jody. 2001. Genes, Categories, and Species: The Evolutionary and Cognitive Causes of the Species Problem. Oxford: Oxford University Press. Hickey, Raymond. 2005. Dublin English: Evolution and Change. Amsterdam: John Benjamins. Hinton, Leanne, & Ken Hale (eds.). 2001. The Green Book of Language Revitalization in Practice. San Diego: Academic Press. Hock, Hans Henrich, & Brian D. Joseph. 1996. Language history, language change, and language relationship. Berlin: Mouton de Gruyter. Hoenigswald, Henry M. 1990. Does language grow on trees? Ancestry, descent, regularity. Proceedings of the American Philosophical Society 134.10-18. Holloway, Joseph E. & Winifred K. Vass. 1993. The African Heritage of American English. Bloomington: Indiana University Press.

Bibliography Holm, John. 1980a. African features in white Bahamian English. English World-Wide 1.45-65. 1980b. The Creole 'copula' that highlighted the world. In Dillard, ed., 367-75. 1986. Substrate diffusion. In Muysken & Smith, eds., 259-78. 1988. Pidgins and Creoles. Vol. 1: Theory and Structure. Cambridge: Cambridge University Press. 1993. Phonological features common to some West-African and Atlantic Creole languages. In Mufwene, ed., 317-27. 1989. Pidgins and Creoles. Vol. 2: Reference Survey. Cambridge: Cambridge University Press. 2004. Languages in Contact: The Partial Restructuring of Vernaculars. Cambridge: Cambridge University Press. Holm, John, & A. Coke Oyedeji. 1984. The Yoruba language in the New World. OSO, A Journal for Surinamese Linguistics, Literature, and History 3.83-9. Holzschuh-Neumann, Ingrid, & Edgar W. Schneider (eds.). 2000. Degree of Restructuring in Creole Languages. Amsterdam: John Benjamins. Hookoomsing, Vinesh. 2000. (Re)constructing identities: Creole and ancestral connections. In Globalization and the South-West Indian Ocean, ed. Sandra J. T. Evers & Vinesh Hookoomsing, 149-55. Leiden, Netherlands: International Institute for Asian Studies, and Reduit: University of Mauritius. Hopper, Paul. 1991. Some principles of grammaticization. In Traugott & Heine, eds., 17-35. Hopper, Paul, & Elizabeth C. Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press, Second edn. Huber, Magnus. 1999. Atlantic Creoles and the Lower Guinea Coast: A case against Afrogenesis. In Spreading the Word: The Issue of Diffusion Among the Atlantic Creoles, ed. Magnus Huber & Mikael Parkvall, 81-110. London: University of Westminster Press. Hughes, Lotte. 2003. The no-nonsense guide to indigenous peoples. Oxford: New Internationalist Publications, Ltd. in association with London: Verso. Hull, David L. 1988. Science as a Process: An Evolutionary Account of the Social and Conceptual Development of Science. Chicago: University of Chicago Press. 2002a. Species, languages and the comparative method. Selection 3,1.17-28. 2002b. Introduction. Selection in Language Evolution, special issue of Selection 3, no. 1, 5-6. Hulstaert, Gustaaf. 1974. A propos des Bangala. Zaire-Afrique 83.173-85. Huttar, George L. 1975. Sources of Creole semantic structures. Language 51.684-95. 1985. Sources of Ndjuka African vocabulary. De Nieuwe West-Indische Gids 59.45-71. Hutton, Christopher. 2000. Race and language. Interventions 2.53-72. Hymes, Dell (ed.). 1971. Pidginization and Creolization of Languages. Cambridge: Cambridge University Press. Irvine, Alison. 2004. A good command of the English language: Phonological variation in the Jamaican acrolect. Journal of Pidgin and Creole Languages 19.41-76. Jackendoff, Ray. 2002. Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford: Oxford University Press. Jacobs-Huey, Lanita. 1997. Is there an authentic African American speech community?: Carla revisited. University of Pennsylvania Working Papers in Linguistics 4.331-70. James, Angela. 2003. Making sense of race and racial classification. Race and Society 4.235-47. Janda, Richard, Brain Joseph & Lyle Compbell (Guest eds.). 2000. Language Sciences nos. 2-3: Special issue on "Grammaticalization." Janson, Tore. 2004. A Natural History of Latin: The Story of the World's Most Successful Language. Oxford: Oxford University Press. Jespersen, Otto. 1922. Language: Its Nature, Development and Origin. London: Allen & Unwin.

323

324

Bibliography Johnson, Guy. 1930. Folk Culture on St. Helena Island, South Carolina. Chapel Hill: University of North Carolina Press. Jones, Sir William. 1786. The third anniversary discourse, on the Hindus, Royal Asiatic Society. Published 1788, Asiatick Researches 1.422. Jones-Jackson, Patricia. 1978. The status of Gullah: An investigation on convergent processes. PhD dissertation, The University of Michigan. 1987. When Roots Die: Endangered Traditions on the Sea Islands. Athens: University of Georgia Press. Joseph, Brian, Johanna DeStefano, Neil Jacobs, & Use Lehiste (eds.). 2003. When Languages Collide: Perspectives on Language Conflict, Language Competition, and Language Coexistence. Columbus: Ohio State University Press. Jourdain, Elodie. 1956. Dufrancais aux purlers Creoles. Paris: Klincksieck. Jourdan, Christine. 1985. Creolisation, nativisation or substrate influences: What is happening in bae in Solomon Islands Pidgin. Papers in Pidgin and Creole Linguistics 4.67-96. 1991. Pidgins and Creoles: The blurring of categories. Annual Review of Anthropology 20.187-209. Kachru, Braj. 2005. Asian Englishes Beyond the Canon. Hong Kong: Hong Kong University Press. Kachru, Yamuna & Cecil L. Nelson. 2006. World Englishes in Asian Contexts. Hong Kong: Hong Kong University Press. Kallen, Jeffrey (ed.). 1997. Focus on Ireland. Amsterdam: John Benjamins. Kapanga, Mwamba Tshishiku. 1991. Language variation and change: A case study of Shaba Swahili. PhD dissertation, University of Illinois at Urbana-Champaign. Kashif, Anette. 1991. For a more inclusive definition of Black English. Paper presented at the Panel on What is Black English? 15th Annual Meeting of the National Council for Black Studies, Atlanta, 20-24 March. Position summarized in "What is Black English?" by Salikoko Mufwene, 1992, The Carrier Pidgin 20, #2,3-4. Keenan, Edward L. & Bernard Comrie. 1977. Noun phrase accessibility and universal grammar. Linguistic Inquiry 8.63-99. Keesing, Roger M. 1988. Melanesian Pidgin and the Oceanic substrate. Stanford: Stanford University Press. 1991. Substrates, calquing and grammaticalization in Melanesian Pidgin. In Traugott & Heine, eds., 315-42. Kegl, Judy, & John McWhorter. 1997. Perspectives on an emerging language. In Proceedings of the Twenty-Eighth Annual Child Language Research Forum, ed. Eve Clark, 15-38. Stanford, CA: Center for the Study of Language and Information. Kegl, Judy, Ann Senghas, & Marie Coppola. 1999. Creation through contact: Sign language emergence and sign language change in Nicaragua. In DeGraff, ed., 179-237. Keiser, Steve Hartman. 2003. Pennsylvania German and the "lunch pail threat": Language shift and cultural maintenance in two Amish communities. In When Languages Collide: Perspectives on Language Conflict, Language Competition, and Language Coexistence, ed. Brian D. Joseph, Johanna Destafano, Neil G. Jacobs, & Use Lehiste, 3-20. Columbus: Ohio State University Press. Keller, Rudi. 1994. On Language Change: The Invisible Hand in Language. London: Routledge. Kephart, Ronald F. 2000. "Broken English": The Creole Language ofCarriacou. New York: Peter Lang. Kihm, Alain. 1984. Les difficiles debuts des etudes Creoles en France (1870-1920). Langue francaise 63.42-56. 1994. Kriyol Syntax: the Portuguese-Based Creole Language of Guinea-Bissau. Amsterdam: John Benjamins. 2002. Langues Creoles et origine du langage: etat de la question. Langages 146.59-69. Klein, Wolfgang, & Clive Perdue. 1992. Utterance Structure: Developing Grammars Again. Amsterdam: John Benjamins. Koerner, Konrad (ed.). 1983. Linguistics and Evolutionary Theory: Three Essays by August Schleicher, Ernst Haeckel, and Wilhelm Bleek, with an introduction by J. Peter Maher. Amsterdam: John Benjamins.

Bibliography Koopman, Hilda. 1986. The genesis of Haitian: Implications of a comparison of some features of the syntax of Haitian, French and West African languages. In Muysken & Smith, eds., 231-58. Kouwenberg, Silvia. 1989. SVO patterns in Ijo. Paper presented at the Conference on Niger-Congo Languages, University of Tilburg. 1994. A Grammar ofBerbice Dutch Creole. Berlin: Mouton de Gruyter. 2006. LI transfer and the cut-off point for L2 acquisition processes in Creole formation. In Lefebvre et al.y eds., 205-19. Kouwenberg, Silvia & Peter L. Patrick (eds.). 2003. Reconsidering the Role of SLA in Pidginization and Creolization. Special issue, Studies in Second Language Acquisition 25, no. 2. Krapp, George Philip. 1924. The English of the Negro. The American Mercury 2.190-5. 1925. The English Language in America. New York: Century. Krauss, Michael. 1992. The world's languages in crisis. Language 68.4-10. Kretzschmar, William A. 2002. American English: Melting pot or mixing bowl? In OfDyuersitie & Chaunge ofLangage: Essays presented to Manfred Gorlachy ed. Katja Lenz & Ruth Mohlig, 224-39. Heidelberg, Germany: C. Winter. Kretzschmar, William A., & Susan Tamasi. 2003. Distributional foundations for a theory of language change. World Englishes 22.377-401. Kurath, Hans. 1928. The origin of dialectal differences in spoken American English. Modern Philology 25.385-95. Labov, William. 1963. The social motivation of a sound change. Word 19.273-309. Reprinted in Sociolinguistic Patterns, by William Labov, 1972. Philadelphia: University of Pennsylvania Press. 1980. Is there a Creole speech community? In Valdman & Highfield, eds., 389-424. 1994. Principles of Linguistic Change: Internal Factors. Oxford: Blackwell. 1998. Co-existent systems in African-American vernacular English. In Mufwene, Rickford, Bailey, & Baugh, eds., 110-53. 2001. Principles of Linguistic Change: Social Factors. Maiden, MA: Blackwell. Labov, William, & Wendell Harris. 1986. De facto segregation of black and white vernaculars. In Diversity and Diachrony, ed. David Sankoff, 1-24. Amsterdam: John Benjamins. Lafage, Suzanne. 1977. Lefrancais en Afrique noire: Elements de bibliographie. Abidjan: Publications de 1'ILA. Lakoff, George. 1987. Women, Fire, and Dangerous Things. Chicago: University of Chicago Press. Lalla, Barbara, & Jean D'Costa. 1990. Language in Exile: Three Hundred Years of Jamaican Creole. Tuscaloosa: University of Alabama Press. Landa, Manuel De. 2000. A Thousand Years of Nonlinear History. New York: Swerve Editions. Lass, Roger. 1997. Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Le Page, Robert. 1960. An Historical Introduction to Jamaican Creole. In Jamaican Creole, ed. R. B. Le Page & David De Camp, 1-124. Cambridge: Cambridge University Press. Le Page, Robert, & Andree Tabouret-Keller. 1985. Acts of Identity: Creole-Based Approaches to Language and Identity. Cambridge: Cambridge University Press. Lefebvre, Claire. 1986. Relexification in Creole genesis revisited: The case of Haitian Creole. In Muysken & Smith, eds., 279-300. 1993. The role of relexification and syntactic analysis in Haitian Creole: Methodological aspects of a research program. In Mufwene, ed., 254-79. 1998. Creole Genesis and the Acquisition of Grammar: The case of Haitian Creole. Cambridge: Cambridge University Press.

325

326

Bibliography 2004. Issues in the Study of Pidgin and Creole Languages. Amsterdam: John Benjamins. Lefebvre, Claire, Lydia White, & Christine Jourdan (eds.). 2006.12 Acquisition and Creole Genesis. Amsterdam: John Benjamins. Levi-Strauss, Claude. 2001. Race et histoire, race et culture. Paris: Albin Michel & Editions UNESCO. Lewontin, Richard Charles. 1970. The units of selection. Annual Review of Ecology and Systematics 1.1-18. Li, Charles N. 2002a. Missing links, issues and hypotheses in the evolutionary origin of language. In Givon & Malle, eds., 83-106. 2002b. On the evolutionary origin of language. In Mirror Neurons and the Evolution of Brain and Language, ed. M. Stamenov & V. Gallese. Amsterdam: John Benjamins. Lieberman, Philip. 1991. Review of Bickerton 1990. American Anthropologist 93.998-1000. 2002. On the nature of evolution and the neural bases of human language. Yearbook of Physical Anthropology 45.36-62. Lightfoot, David W. 2002. Myths and the prehistory of grammar. Journal of Linguistics, 38.113-36. Lippi-Green, Rosina. 1997. English with an Accent: Language, Ideology, and Discrimination in the United States. London: Routledge. Lorenzino, Geraldo A. 1998. The Angolar Creole Portuguese of Sao Tome: Its grammar and sociolinguistic history. PhD dissertation, City University of New York. Liidtke, Helmut. 1995. On the origin of Middle and Modern English. In Linguistic Change Under Contact Conditions, ed. Jacek Fisiak, 51-4. Berlin: Mouton. Lumsden, John. 1999. Language acquisition and creolization. In DeGraff, ed., 129-57. MacWhinney, Brian. 2002a. The emergence of language from body, brain, and society. In Proceedings from the panels of thirty-eighth meeting of the Chicago Linguistic Society, Volume 38-2, ed. Mary Andronis, Erin Debenport, Ann Pycha, & Keiko Yoshimura, 247-71. Chicago: Chicago Linguistic Society. 2002b. The gradual emergence of language. In Givon & Malle, eds., 233-63. Maffi, Luisa (ed.). 2001. On Biocultural Diversity: Linking Language, Knowledge, and the Environment. Washington, DC: Smithsonian Institution Press. Maine, Henry Sumner. 1861. Ancient Law: Its Convention with the Early History of Society and Its Relation to Modern Ideas. London: John Murray. 1875. Effects of Observation of India on Modern European Thought. The Rede Lecture, Delivered before the University of Cambridge, 22 May. London: John Murray. 1883. Dissertation on Early Law and Customs. New York: John Murray. Manessy, Gabriel. 1985a. La construction serielle dans les langues africaines et les langues Creoles. Bulletin de la Societe de Linguistique de Paris 80.333-57. 1985b. Remarques sur la pluralisation du nom en Creole et dans les langues africaines. Etudes Creoles 8.129-43. 1988. U extraposition du predicat verbal dans les Creoles a base lexicale anglaise et francaise. In Languages and Cultures in Honor of E.C. Polome, ed. Mohammed Ali Jazayery and Werner Winter, 401-17. Berlin: Mouton de Gruyter. 1994. Creoles, pidgins, varietes vehiculaires: proces etgenese. Paris: CNRS. 1996. Reflexions sur les contraintes anthropologiques de la creolisation: de 1'improbabilite du metissage linguistique dans les Creoles atlantiques exogenes. Etudes Creoles 19.61-71. Manessy, Gabriel, & Paul Wald. 1984. Lefrancais en Afrique noire, tel quon leparle, tel quon le dit. Paris: L'Harmattan. Marks, Jonathan. 2002. What It Means to be 98% Chimpanzee. Berkeley: University of California Press. Marling, William H. 2006. How "American" is Globalization? Baltimore: Johns Hopkins Press.

Bibliography Masuda, Hirokuni. 1995. TSR formation as a discourse substratum in Hawaii. Journal of Pidgin and Creole Languages 10.253-88. Mather, Patrick-Andre (2006). Second language acquisition and creolization : Same (i-) processes, different (e-) results. Journal of Pidgin and Creole Languages 21.231-74. Maurais, Jacques, and Michael A. Morris (eds.). 2003. Languages in a Globalizing World. Cambridge: Cambridge University Press. Maurer, Philippe. 1987. La comparaison des morphemes temporels du papiamento et du palenquero: Arguments centre la theorie monogenetique de la genese des langues Creoles. In Varia Creolica, ed. Philippe Maurer & Thomas Stolz, 27-70. Bochum: Studienverlag Dr. N. Brockmeyer. Mayor, Federico, & Jerome Binde. 2001. The World Ahead: Our Future in the Making. Paris: UNESCO Publishing; London and New York: Zed Books. Mayr, Ernst. 2001. What Evolution Is. New York: Basic Books, Perseus Books Group. Mazrui, Ali A., & Alamin M. Mazrui. 1998. The Power of Babel: Language and Governance in the African Experience. Oxford: James Currey; Nairobi: E.A.E.P.; Kampala: Fountain Publishers; Cape Town: David Philip Publishers; Chicago: University of Chicago Press. McCawley, James D. 1976. Some ideas not to live by. Die neueren Sprachen 75.151-65. McLaughlin, Barry. 1987. Theories of Second Language Learning. London: Edward Arnold. McMahon, April. 1994. Understanding Language Change. Cambridge: Cambridge University Press. McNeill, David. 1992. Review of Bickerton 1990. International Journal of Primatology 13.575-83. 2005. Gesture and Thoughts. Chicago: University of Chicago Press. McWhorter, John H. 1998. Identifying the Creole prototype: Vindicating a typological class. Language 74.788-818. 2000. The Missing Spanish Creoles: Recovering the Birth of Plantation Contact Languages. Berkeley: University of California Press. 2001. The Power of Babel: A Natural History of Language. New York: Times Books. 2005. Defining Creole. Oxford: Oxford University Press. Meillet, Antoine. 1900. Note sure une difficulte generale de la grammaire comparee. Reprinted in Linguistique historique et linguistiquegenerale, 36-43. Geneve: Slaktine, and Paris: Champion (1982). Meillet, Antoine. 1912. Devolution des formes grammaticales. Scientia (Revista di Scienza) 12, no. 26. Reprinted in Meillet (1951: 130-48). 1921. Linguistique historique et linguistique generale. Vol. 1. Paris: Champion. 1929. Le developpement des langues. In Continu et discontinu, 119ff. Paris: Bloud & Gay. Reprinted in Meillet 1951:71-83. 1951. Linguistique historique et linguistique generale. Vol. 2. Paris: Klincksieck. Meisel, Jiirgen M. (ed.). 1977. Pidgins - Creoles - Languages in Contact. Tubingen: Narr. Meisel, Jiirgen M. 1983a. Transfer as a second-languages strategy. Language and Communication 3.11-46. 1983b. Strategies of second language acquisition: More than one kind of simplification. In Andersen, ed., 120-57. Mesthrie, Rajend. 1991. Language in Indenture: A Sociolinguistic History ofBhojpuri-Hindi in South Africa. Witwatersrand University Press. 1992. English in Language Shift: The History, Structure and Sociolinguistics of South African Indian English. Cambridge: Cambridge University Press. 2006. Subordinate immigrant languages and language endangerment: Two community studies from Kwa-Zulu Natal. Language Matters 37.3-15.

327

328

Bibliography Migge, Bettina. 2003. Creole Formation as Language Contact: the Case of the Suriname Creoles. Amsterdam: John Benjamins. Mille, Katherine. 1990. A historical analysis of tense-mood-aspect in Gullah Creole: A case of stable variation. PhD dissertation, University of South Carolina. Milroy, James. 1996. Linguistic ideology and the Anglo-Saxon lineage of English. In Speech Past and Present: Studies in English Dialectology in Memory ofOssi Ihalainen, ed. Juhani Klemola, Merja Kyto, & Matti Rissanen, 169-86. Frankfurt am Main: Peter Lang. 1997. Internal vs external motivations for linguistic change. Multilingua 16.311-23. Mithun, Marianne. 1992. The substratum in grammar and discourse. In Language Contact: Theoretical and Empirical Studies, ed. Ernst Hakon Jahr, 103-15. Berlin: Mouton de Gruyter. Mohan, Peggy. 1990. The rise and fall of Trinidadian Bhojpuri. International Journal of the Sociology of Language 85.21-30. Mohan, Peggy, & Paul Zador. 1986. Discontinuity in a life cycle: The death of Trinidad Bhojpuri. Language 62.291-320. Montgomery, Michael. 1995. The koineization of colonial American English. In Sociolinguistic Studies and Language Planning: Proceedings of the XVIth Annual Meeting of the Atlantic Provinces Linguistic Association, ed. Catherine Philiponneau, 309-31. Moncton, NB: Centre de Recherche en Linguistique Appliquee. Morford, Jill P. 2002. Why does exposure to language matter? In Givon & Malle, eds., 329-41. Mufwene, Salikoko S. 1984. The language bioprogram hypothesis, Creole studies, and linguistic theory [commentary on Derek Bickerton's "The language bioprogram hypothesis"]. The Behavioral and Brain Sciences 7.202-3. 1985. The linguistic significance of African proper names in Gullah. New West Indian Guide 59.146-66. 1986a. The universalist and substrate hypotheses complement one another. In Muysken & Smith, eds., 129-62. 1986b. Les langues Creoles peuvent-elles etre definies sans allusion a leur histoire? Etudes Creoles 9.135-50. 1986c. Restrictive relativization in Gullah. Journal of Pidgin and Creole Language 1.1-31. 1986d. The linguistic significance of African proper names in Gullah. New West Indian Guide 59.146-66. 1987. Review article on Language Variety in the South: Perspectives in Black and White, ed. Michael Montgomery & Guy Bailey. Journal of Pidgin and Creole Languages 2.93-110. 1988a. English pidgins: Form and function. World Englishes 7.255-67. 1988b. Why study pidgins and Creoles? Column. Journal of Pidgin and Creole Languages 3.265-76. 1988c. Formal evidence of pidginization/creolization in Kituba. Journal of African Languages and Linguistics 10.33-51. 1989a. Equivocal structures in some Gullah complex sentences. American Speech 64.304-26. 1989b. Some explanations that strike me as incomplete. Column. Journal of Pidgin and Creole Languages 4.117-28. 1989c. Review ofMelanesian Pidgin and the Oceanic Substrate, by Roger Keesing. The Carrier Pidgin 16, 3.4-5. 1989d. La creolisation en bantou: les cas du kituba, du lingala urbain, et du swahili du Shaba. Etudes Creoles 12.74-106. 1991a. Is Gullah decreolizing? A comparison of a speech sample of the 1930's with a speech sample of the 1980's. In The Emergence of Black English, ed. Guy Bailey, Patricia Cukor-Avila, & Natalie Maynor, 213-30. Amsterdam: John Benjamins. 1991b. Pidgins, Creoles, typology, and markedness. In Development and Structures of Creole Languages: Essays in Honor of Derek Bickerton, ed. Francis Byrne & Thorn Huebner, 123-43. Amsterdam: John Benjamins. 1991c. Language genesis and human evolution. Review article on Bickerton 1990. Diachronica 8.239-54. 1992a. Why grammars are not monolithic. In The Joy of Grammar: A Festschrift in Honor of James D. McCawley, ed. Diane Brentari, Gary Larson, & Lynn A. MacLeod, 225-50. Amsterdam: John Benjamins. 1992b. The reviewer responds. Diachronica 9.149-51.

Bibliography 1993a. African substratum: Possibility and evidence. In Mufwene, ed., 1993,192-208. 1993b. Investigating Gullah: Difficulties in ensuring "authenticity." In Language Variation in North American English Research and Teaching, ed. A. Wayne Glowka & Donald M. Lance, 178-90. New York: MLA. 1993c. Are there possessive pronouns in Atlantic Creoles? In Atlantic meets Pacific: Papers from the Society for Pidgin and Creole Linguistics, ed. Francis Byrne and John Holm, 133-43. Amsterdam: John Benjamins. Mufwene, Salikoko S. (ed.). 1993. Africanisms in Afro-American Lnguage Varieties. Athens: University of Georgia Press. Mufwene, Salikoko S. 1994. On decreolization: The case of Gullah. In Language and the Social Construction of Identity in Creole Situations, ed. Marcyliena Morgan, 63-99. Los Angeles: Center for Afro-American Studies. 1996a. The Founder Principle in Creole genesis. Diachronica 13.83-134. 1996b. Creolization and grammaticization: What creolistics could contribute to research on grammaticization. In Changing Meanings, Changing Functions..., ed. Philip Baker & Anand Syea, 5-28. London: University of Westminster Press. 1997a. Jargons, pidgins, Creoles, and koines: What are they? In Spears & Winford, eds., 35-70. 1997b. The legitimate and illegitimate offspring of English. In World Englishes 2000, ed. Michael L. Forman & Larry Smith, 182-203. Honolulu: The University of Hawaii Press. 1998. What research on creole genesis can contribute to historical linguistics. In Historical Linguistics 1997, ed. Monika Schmid, Jennifer Austin, & Dieter Stein, 315-38. Amsterdam: John Benjamins. 1999a. The language bioprogram hypothesis: Hints from Tazie. In DeGraff, ed., 95-127. 1999b. Some sociohistorical inferences about the development of African-American English. In The English history of African American English, ed. Shana Poplack, 233-63. Oxford: Blackwell. 2000a. Creolization is a social, not a structural, process. In Degrees of Restructuring in Creole Languages, ed. Ingrid Neumann-Holzschuh & Edgar Schneider, 65-84. Amsterdam: John Benjamins. 2000b. Population contacts and the evolution of English. The European English Messenger 9.9-15. 200 la. The Ecology of Language Evolution. Cambridge: Cambridge University Press. 200 Ib. What is African-American English? In Sociocultural and Historical Contexts of African-American Vernacular English, ed. Sonja Lanehart, 21-51. Amsterdam: John Benjamins. 2002a. Colonization, globalization, and the plight of "weak" languages: A rejoinder to Nettle & Romaine's Vanishing Voices. Journal of Linguistics 38.375-95. 2002b. Colonization, globalization, and the future of languages in the twenty-first century. MOST Journal of Multicultural Societies 4.162-93. www.unesco.org/most/vl4n2mufwene.pdf 2003a. Genetic linguistics and genetic creolistics. Short note. Journal of Pidgin and Creole Languages 18.273-88. 2003b. The shared ancestry of African-American and American White southern Englishes: Some speculations dictated by history. In English in the Southern United States, ed. Stephen Nagle & Sara Sanders, 64-81. Cambridge: Cambridge University Press. 2003c. Preface. In Chaudenson, 1-28. 2005a. Creoles, ecologie sociale, evolution linguistique. Paris: L'Harmattan. 2005b. How many bes are there in English? In Polymorphous Linguistics: Jim McCawley's Legacy, ed. Salikoko S. Mufwene, Elaine J. Francis, & Rebecca S. Wheeler, 225-46. Cambridge, MA: MIT Press. 2006. "New-dialect formation" is as gradual as "creole-development": A response to New-Dialect Formation: The Inevitability of Colonial Englishes by Peter Trudgill. World Englishes 25.177-86. 2007. Les Creoles: de nouvelles varietes indo-europeennes desavouees? In Actes du collogue "Creolisation Linguistinque er Sciences Humanies? ed. Marie-Paul Ansie, 59-70. Paris: Presses Universitaires Hatiano-Antillaises.

329

330

Bibliography to appear. Second language acquisition and the development of Creoles. Studies in Second Language Acquisition. Mufwene, Salikoko S. & Marta B. Dijkhoff. 1989. On the so-called "infinitive" in Atlantic Creoles. Lingua 77.319-52. Mufwene, Salikoko S., & Charles Oilman. 1987. How African is Gullah and why? American Speech 62.120-39. Mufwene, Salikoko S., John R. Rickford, Guy Bailey, & John Baugh (eds.). 1998. African-American English: Structure, History, and Use. London: Routledge. Miihlhausler, Peter. 1980. Structural expansion and the process of creolization. In Valdman & Highfield, eds., 19-55. 1981. The development of the category of number in Tok Pisin. In Generative Studies on Creole Languages, ed. Pieter Muysken, 35-84. Dordrecht: Foris Publications. 1983. The development of word-formation in Tok Pisin. Folia Linguistica 17.463-87. 1985. The number of pidgin Englishes in the Pacific. Papers in Pidgin and Creole Linguistics No.l Pacific Linguistics, A-72.25-51. 1986a. Pidgin and Creole linguistics. New York: Basil Blackwell, Inc. 1986b. Bonnet Blanc and Blanc Bonnet: Adjective-noun order, substratum and language universals. In Muysken & Smith, eds., 25-40. 1987. Tracing predicate markers in Pacific Pidgin English. English World-Wide 8.97-121. 1989. Review article on Roger M. Keesing: Melanesian pidgin and the Oceanic substrate. Studies in Language, 13. 459-75. 1996. Linguistic Ecology: Language Change and Linguistic Imperialism in the Pacific Region. London: Routledge. 1997. Pidgin and Creole Linguistics, expanded and revised edition. London: University of Westminster Press. 2003. Language of Environment, Environment of Language: A Course in Ecolinguistics. London: Battlebridge. 2005. Review of Mufwene (2001). Language 81.265-8. Muysken, Pieter, & Norval Smith (eds.). 1986. Substrata versus Universals in Creole Genesis. Amsterdam: John Benjamins. Naro, Anthony Julius. 1978. A study on the origins of pidginization. Language 54.314-47. Naro, Anthony Julius, & Maria Marta Pereira Scherre. 1993. Sobre as origens do portugues popular do Brasil. Revista de Documentacao de Estudos em Linguistica Teorica e Aplicada (DELTA) 9.439-54. Nettle, Daniel. 1999. Using social impact theory to simulate language change. Lingua 108.95-117. Nettle, Daniel, & Suzanne Romaine. 2000. Vanishing Voices: The Extinction of the World's Languages. Oxford: Oxford University Press. Ngalasso, Mwatha M. 1989. Tons ou accents? Analyse des schemes intonatifs du kikongo vehiculaire parle dans la region de Bandundu (Zaire). Paper presented at Sixieme Colloque International des Etudes Creoles. Cayennne, French Guyane. Nicolai, Robert. 1990. Parentes linguistiques (apropos du songhay). Paris: CNRS. Nurse, Derek, & Thomas Spear. 1985. The Swahili: Reconstructing the History and Language of an African Society, 800-1500. Philadelphia: University of Pennsylvania Press. O'Hara, Robert J. 1994. Evolutionary history and the species problem. American Zoology 34.12-22. Odlin, Terence. 2003. Language ecology and the Columbian exchange. In Joseph, DEstafano, Jacobs, & Lehiste, eds., 71-94. Osterhammel, Jiirgen & Niels P. Petersen. 2005. Globalization: A Short History. Princeton: Princeton University Press. Ostler, Nicholas. 2005. Empires of the World: A Language History of the World. New York: HarperCollins. Owens, Jonathan. 1998. Representativeness in the data base: Polemical update for the twenty-first century. Language Sciences 20.113-35.

Bibliography Pakir, Anne Geok-In Sim. 1986. A linguistic investigation of Baba Malay. PhD dissertation, University of Hawaii. Pandharipande, Rejeshwari. 2002. Minority matters: Issues in minority languages in India. MOST Journal of Multicultural Societies 4.213-34. www.unesco.org/shs/ijms/vol4/issue2/art4 Pargman, Sheri. 2002. Internal and external factors in language change. PhD dissertation, University of Chicago. Paul, Herman. 1880. Prinzipien der Sprachgeschichte. Halle: Niemeyer. 1891. Principles of the History of Language. Translated from the second edition of the original by H. A. Strong. London: Longmans, Green, and Co. Perdue, Clive. 1995. L'acquisition dufrancais et du I'anglais par des adultes: former des enonces. Paris: CNRS. Phillipson, Robert. 2003. English-only Europe? Challenging Language Policy. London: Routledge. Pieterse, Jan Nederveen. 2003. Globalization and Culture: Global Melange. Lanham: Rowan & Littlefield Publishers, Inc. Pinker, Steven. 1992. Review of Bickerton 1990. Language 68.375-82. Plag, Ingo. 1999. Review of Baker & Syea (1996). Journal of Pidgin and Creole Languages 14.202-8. Pollard, Velma. 2000. Dread Talk: The Language ofRastafari. Mona, Jamaica: The University of the West Indies Press. Polome, Edgar. 1983a. Creolization and language change. In The Social Context of Creolization, ed. Ellen Woolford & William Washabaugh, 126-36. Ann Arbor: Karoma Publishers, Inc. 1983b. The linguistic situation in western provinces of the Roman Empire. Principat 29.509-53. Poplack, Shana (ed.). 2000. The English History of African American English. Maiden, MA: Blackwell. Poplack, Shana, & Sali Tagliamonte. 1991. African American English in the diaspora: Evidence from old line Nova Scotians. Language Variation and Change 3.301-39. 1994. -S or nothing: Marking the plural in the African American diaspora. American Speech 69.227-59. 2001. African American English in the Diaspora. Maiden, MA: Blackwell. Posner, Rebecca. 1985. Creolization as typological change: Some examples from Romance syntax. Diachronica 2.167-88. 1996. The Romance Languages. Cambridge: Cambridge University Press. Post, Marike. 1995. Fa d'Ambu. In Arends, Muysken, & Smith, eds., 191-204. Prescod, Paula. 2004. Une description grammaticale du syntagme nominal dans le creole anglophone de St-Vincent-etles-Grenadines. PhD dissertation, Universite Paris III - Sorbonne Nouvelle. Price, Glanville. 1984. The Language of Britain. London: Edward Arnold. Prudent, Lambert-Felix. 1980. Des baragouins a la langue antillaise: Analyse historique et sociolinguistique du discours sur le creole. Paris: Editions Caribeennes. Pullum, Geoffrey K. 1990. Constraints on intransitive quasi-serial verb constructions in Modern Colloquial English. In When Verbs Collide: Papers from the 1990 Ohio State Mini-Conference on Serial Verbsy ed. Brian D. Joseph & Arnold Zwicky, 218-39. Department of Linguistics, Ohio State University. Quint, Nicolas. 2000. Grammaire de la langue cap-verdienne: etude descriptive et comprehensive du creole afro-portugais des lies du Cap-Vert. Paris: L'Harmattan. Radick, G. 2002. Darwin on language and selection. Selection 3.7-16. Reinecke, John E. 1937/1969. Language and Dialect in Hawaii: A Sociolinguistic History to 1935. Honolulu: University of Hawaii Press. Renfrew, Colin. 1987. Archaeology and Language: The puzzle of Indo-European Origins. Cambridge: Cambridge University Press. Reyhner, Jon, Gina Cantoni, Robert N. St. Clair, and Evangeline Parsons Yazzie (eds.). 1999. Revitalizing indigenous Languages. Flagetaff, A2: Northern Arizona University: Center for Excellence in Education.

331

332

Bibliography Rickford, John R. 1987. Dimensions of a Creole Continuum: History, Text, and Linguistic Analysis of Guyanese Creole. Stanford: Stanford University Press. 1990. Number delimitation in Gullah: A response to Mufwene. American Speech 65.148-63. 1992. Grammatical variation and divergence in vernacular Black English. In Internal and external factors in syntactic change, ed. Marinel Gerritsen & Dieter Stein, 175-200. Berlin: Mouton de Gruyter. Rickford, John R., & Jerome S. Handler. 1994. Textual Evidence of the Nature of Early Barbadian Speech. Journal of Pidgin and Creole Languages 9.221-55. Roberts, Sarah Julianne. 1998. The role of diffusion in the genesis of Hawaiian Creole. Language 74.1-39. 1999. The TMA system of Hawaiian Creole and diffusion. In Creole Genesis, Attitudes and Discourse, ed. John R. Rickford and Suzanne Romaine, 45-70. Amsterdam: John Benjamins. 2005. The emergence of Hawai'i Creole English in the early 20th century: The sociohistorical context of Creole genesis. PhD dissertation, Stanford University. to appear. The Social Context of Language Birth: New Light on the Origin of Hawaiian Creole. Cambridge: Cambridge University Press. Robertson, Ian. 1993. The Ijo element in Berbice Dutch: The pidginization/creolization process. In Mufwene, ed., 296-316. Robins, Robert H., & Eugenius M. Uhlenbeck (eds.). 1991. Endangered Languages. Oxford: Berg. Romaine, Suzanne. 1988. Pidgin and Creole Languages. London: Longman. 2000. Multilingualism. In Handbook of Linguistics, ed. Mark Aronoff & Janie Rees-Miller, 512-32. Maiden, MA: Blackwell. Salmons, Joseph. 2003. The shift from German to English, World War I and the German-language press in Wisconsin. In Menschen zwischen zwei Welten: Auswanderung, Ansiedlung, Akkulturation, ed. Walter G. Radel & Helmut Schmahl, 179-93. Trier: Wissenschaftlicher Verlag. Samarin, William J. 1982. Goals, roles, and language skills in colonizing Central Equatorial Africa. Anthropological Linguistics 24.410-22. Sankoff, Gillian. 1979. The genesis of a language. In Hill, ed., 23-47. 1984. Substrate and universals in the Tok Pisin verb phrase. In Meaning, Form, and Use in Context: Linguistic Applications, ed. Deborah Schiffrin, 104-19. Washington, DC: Georgetown University Press. 1993. Focus in Tok Pisin. In Focus and Grammatical Relations in Creole Languages, ed. Francis Byrne & Donald Winford, 117-40. Amsterdam: John Benjamins. Sankoff, Gillian, & Penelope Brown. 1976. The origins of syntax in discourse: A case study of Tok Pisin relatives. Language 52.631-66. Sapir, Edward. 1921. Language: An Introduction to the Study of Speech. New York: Harcourt, Brace & World, Inc. Sarich, Vincent, & Frank Miele. 2004. Race: The Reality of Human Differences. Boulder, CO: Westview Press. Saussure, Ferdinand de. 1916. Cours de linguistique generate, ed. Charles Bally & Albert Sechehaye in collaboration with Albert Riedlinger. Paris: Payot (Trans. Wade Baskin. 1966. Course in General Linguistics. New York: McGraw-Hill Co.) Schleicher, August. 1863. Die darwinische Theorie und die Sprachwissenschaft: Offenes Sendschreiben an Herrn Dr. Ernst Hadte/. Weimar: Bohlau. 1869. Darwinism tested by the science of language. Translated by Alexander V. W. Bikkers. London: J.C. Hotten. Reprinted in Koerner, ed., 1983. Schlieben-Lange, Brigitte. 1977. L'origine des langues romanes: un cas de creolisation? In Meisel, ed., 81-101. Schneider, Edgar W. 1989. American Earlier Black English. University: University of Alabama Press.

Bibliography 1990. The cline of creoleness in English-oriented Creoles and semi-creoles of the Caribbean. English World-Wide 11.79-113. 1995. Black-White language contact through the centuries: Diachronic aspects of linguistic convergence or divergence in the United States of America. In Linguistic Change under Contact Conditions, ed. Jacek Fisiak, 237-52. Berlin: Mouton de Gruyter. 2007. Post-Colonial Englishes: The Dynamics of Language Diffusion. Cambridge: Cambridge University Press. Schuchardt, Hugo. 1883. Ueber das Melanesoenglische. Sitzungsberichte der philosophisch-historisch Klasse des kaiserlischen Akademie der Wissenschaften in Wien 5,105.151-61. 1884. Slavo-deutsches und Slavo-italienisches. Graz: Leuschner & Lubensky. 1889. Beitrage zur Kenntnis des englischen Kreolisch II: Melanesoenglisches. Englische Studien 13.158-62. 1914. Die Sprache der Saramakkaneger in Surinam. Amsterdam: Johannes Miiller. Translated in Hugo Schuchardt, 1979, The Ethnography of Variation: Selected Writings on Pidgins and Creoles, trans. & ed. T.L. Markey. Ann Arbor: Karoma, 73-108: "The language of the Saramaccans." Schwartz, Stuart B. 1985. Sugar Plantations in the Formation of Brazilian Society: Bahia, 1550-1835. Cambridge: Cambridge University Press. Schwegler, Armin. 2006. Bozal Spanish: Captivating new evidence from a contemporary source (Afro-Cuban "Palo Monte"). In Studies in Contact Linguistics: Essays in Honor of Glenn G. Gilbert, ed. Linda L. Thornburg & Janet Fuller, 71-101. New York: Peter Lang. Sebba, Mark. 1987. The Syntax of Serial Verbs: An Investigation into Serialisation in Sranan and Other Languages. Amsterdam: John Benjamins. 1993. London Jamaican: Language Systems in Interaction. New York: Longman. Senghas, Anne & Marie Coppola. 2001. Children creating language: How Nicaraguan Sign Language acquired a spatial grammar. Psychological Science 12.323-8. Seuren, Pieter M. 1987. A note on siki. Journal of Pidgin and Creole Languages 2.57-66. Seuren, Pieter M., & Herman Wekker. 1986. Semantic transparency as a factor in Creole genesis. In Muysken & Smith, eds., 57-70. Siegel, Jeff. 1987. Language Contact in a Plantation Environment: A Sociolinguistic History of Fiji. New York: Cambridge University Press. 2003. Substrate influence in Creoles and the role of transfer in second language acquisition. Studies in Second Language Acquisition 25.185-209. 2006. Links between SLA and Creole studies: Past and present. In Lefebvre et al., eds., 15-46. Silverstein, Michael. 1996. Dynamics of recent linguistic contact. In Handbook of North American Indians vol. 17, Languages, ed. Ives Goddard, 117-36. Washington, DC: Smithsonian Institution Press. Singler, John. 1984. Remarks in response to Derek Bickerton's Creoles and the universal grammar: The unmarked case? Winter Meeting of the Linguistic Society of America, Baltimore. 1988. The homogeneity of the substrate as a factor in pidgin/creole genesis. Language 64.27-51. 1993. African influence upon Afro-American language varieties: A consideration of sociohistorical factors. In Mufwene, ed., 235-53. Singler, John. 1996. Theories of Creole genesis, sociohistorical considerations, and the evaluation of evidence: The case of Haitian Creole and the Relexification Hypothesis. Journal of Pidgin and Creole Languages 11.185-230. Skutnabb-Kangas, Tove. 2000. Linguistic Genocide in Education - or World-Wide Diversity and Human Rights? Mahwah, NJ: Lawrence Erlbaum.

333

334

Bibliography Slobin, Dan 1.2002. Language evolution, acquisition, diachrony: Probing the parallels. In Givon & Malle, eds., 375-92. Smedley, Audrey. 1999. "Race" and the construction of human identity. American Anthropologist 100.691-702. Smith, Adam. 1776. The Wealth of Nations, vol. 1. Oxford: Clarendon Press. Smith, Geoff, & Stephen Matthews. 1996. Pidgins and Creoles. In The Atlas of Languages: The Origin and Development of Languages Throughout the World, ed. Bernard Comrie, Stephen Matthews, & Maria Polinsky, 144-61. New York: Facts On File. Smith, Norval. 1995. An annotated list of pidgins, Creoles, and mixed languages. In Arends, Muysken, & Smith, eds., 331-74. Smith, Norval, Ian E. Robertson, & Kay Williamson. 1987. The Ijo element in Berbice Dutch. Language in Society 16, 1.49-90. Smith, Reed. 1926. Gullah. Columbia: Bureau of Publications, University of South Carolina. Sober, Elliott. 1984. The Nature of Selection: Evolutionary Theory in Philosophical Focus. Chicago: University of Chicago Press. Spears, Arthur K., & Donald Winford (eds.). 1997. The Structure and Status of Pidgins and Creoles. Amsterdam: John Benjamins. Sperber, Dan. 1996. La contagion des idees. Paris: Odile Jacob. Spitulnik, Debra. 1999. The language of the city: Town Bemba as urban hybridity. Journal of Linguistic Anthropology 8.30-59. Sprouse, Rex A. 2006. Full transfer and relexification : Second language acquisition and Creole genesis. In Lefebvre et al, eds., 169-81. Stewart, William. 1974. Acculturative processes in the language of the American Negro. In Language in Its Social Setting, ed. William W. Gage, 1-46. Washington, DC: The Anthropological Society of Washington. Stoddard, Alber H. 1944. Origin, dialect, beliefs, and characteristics of the negroes of the south Carolina and Georgia coasts. The Georgia Historical Quarterly 28.186-95. Stoney, Samuel G., & Gertrude M. Shelby. 1930. Black Genesis. Macmillan: New York. Swaan, Abram de. 2001. Words of the World: The Global Language System. Cambridge: Polity Press. Swadesh, Morris. 1951. Review of Lorenzo Dow Turner's Africanisms in the Gullah dialect. Word 7.83-4. Sylvain, Suzanne. 1936. Le Creole haitien: Morphologie etsyntaxe. Wettern, Belgium: Imprimerie De Meester. Szulmajster-Celnikier, Anne. 2000. Un regard particulier sur le yidiche, 1'ivrit et 1'esperanto. Cahiers Bernard Lazare 204.15-19. Tabor, Whitney. 1994. Syntactic innovation: A connectionist model. PhD dissertation, Stanford University. Tabouret-Keller, Andree. 1999. Western Europe. In Handbook of Language and Ethnic Identity, ed. Joshua A. Fishman, 334-49. Oxford: Oxford University Press. Tagliamonte, Sali, & Shana Poplack. 1988. Tense and aspect in Samana English. Language in Society 17.513-33. 1993. The zero-marked verb: Testing the Creole hypothesis. Journal of Pidgin and Creole Languages 8.171-206. Tate, Triad W. 1965. The Negro in Eighteenth-century Williamsburg. Williamsburg, VA: The Colonial Williamsburg Foundation. Taylor, Paul C. 2004. Race: A Philosophical Introduction. Cambridge: Polity Press; and Oxford: Blackwell Publishing Ltd. Thomason, Sarah G. 1980. On interpreting 'The Indian Interpreter.' Language in Society 9.167-93. 1983. Chinook jargon in areal and historical context. Language 59.820-70. 1997. A typology of Contact languages. In Spears & Winford, eds., 71-88. 2001. Language Contact: An Introduction. Washington, DC: Georgetown University Press.

Bibliography 2002. Creoles and genetic relationship. Column. Journal of Pidgin and Creole Languages 17.101-9. 2003. A response to Mufwene's response [Mufwene 2003a]. Journal of Pidgin and Creole Languages 18.289-98. Thomason, Sarah G., & Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Tomasello, Michael. 2002. The emergence of grammar in early child language. In Givon & Malle, eds., 309-28. Tomlinson, John. 1999. Globalization and Culture. Chicago: University of Chicago Press. Tomlinson, Molly. 2006. Breton or Neo-Breton: Examining the generation gap. Honors BA thesis. University of Chicago. Tosco, Mauro. 1998. "People who are not the language they speak": On language shift without language decay in East Africa. In Brenzinger, ed., 119-42. Trask, Robert Lawrence. 1996. Historical Linguistics. London: Arnold. Traugott, Elizabeth C. 1999. Review of Baker & Syea (1996). Journal of Pidgin and Creole Languages 14.208-13. Traugott, Elizabeth C., & Bernd Heine (eds.). 1991. Approaches to Grammaticalization. Amsterdam: John Benjamins. Tremblay, Xavier. 2005. Grammaire comparee et grammaire historique: quelle realite est reconstruite par la grammaire comparee? In Aryas, Aryens et Iraniens en Asie Centrale, ed. Gerard Fussman, Jeans Kellens, Henri-Paul Francfort, & Xavier Tremblay, 33-180. Paris: Edition-Diffusion de Boaccard. Tristram, Hildegard. 1999. How Celtic is Standard English? St. Petersburg: Nauka. Tristram, Hildegard (ed.). 2000. The Celtic Englishes II. Heidelberg: Carl Winter. 2006. The Celtic Englishes IV: The Interface Between English and the Celtic Languages. Potsdam: Potsdam University Press. Trubetzkoy, Nikolai S. 1939. Gedanken iiber das Indogermanenproblem. Acta Linguistica 1.81-9. Trudgill, Peter. 1983. On Dialect: Social and Geographic Perspectives. New York: New York University Press. 1986. Dialects in Contact. New York: Basil Blackwell, Inc. 2004. New-Dialect Formation: The Inevitability of Colonial Englishes. Oxford: Oxford University Press. Turner, Lorenzo Dow. 1949. Africanisms in the Gullah Dialect. Chicago: University of Chicago Press. Valdman, alber. 1978. Le Creole: structure, statut et origine. Paris: Klinksieck. 1983. Creolization and second language acquisition. In Andersen, ed., 212-34. Valdman, Albert & Arnold Highfield (eds.). 1980. Theoretical Orientations in Creole Studies. New York: Academic Press. Vansina, Jan. 1990. Paths in the Rainforest: Toward a History of Political Tradition in Equatorial Africa. Madison: University of Wisconsin Press. Vennemann, Theo. 2001. Atlantis Semitica: Structural contact features in Celtic and English. In Historical Linguistics 1999: Selected papers from the 14th International Conference on Historical Linguistics, Vancouver, 9-13 August 1999, ed. Laurel Brinton, 351-69. Amsterdam: John Benjamins. Viada, Marta & Nicholas Faraclas. 2006. Matelots, Metis, and Maroons meet mass comparison and the matrix: Possible influences from the indigenous languages of the Caribbean on the grammars of Caribbean Creoles. Paper presented at 16th Biennial Meeting of the Society for Caribbean Linguistics, Dominica. Vigouroux, Cecile B. In press. The smuggling of Francophonie. Language in Society 37,4. Vinson, Julien. 1882. Creole. In Dictionnaire des sciences anthropologiques et ethnologiques, ed. Adolphe Bertillon et al Paris: Doin. 1888. La linguistique. La grande encyclopedie 22.286-96. Paris. Wade-Lewis, Margaret. 1988. The African substratum in American English. PhD dissertation, New York University. Wang, William S-Y. Talking and human evolution. Trends in Ecology and Evolution 22.290-1. Warner, Anthony. 1993. English Auxiliaries: Structure and History. Cambridge: Cambridge University Press.

335

336

Bibliography Warner-Lewis, Maureen. 1982. The Yoruba Language in Trinidad. PhD dissertation, University of the West Indies. 1996. Trinidad Yoruba: From Mother Tongue to Memory. Tuscaloosa: University of Alabama Press. Weinreich, Uriel. 1953. Languages in Contact: Findings and Problems. New York: Linguistic Circle of New York. Weinreich, Uriel, William Labov, & Marvin I. Herzog. 1968. Empirical foundations for a theory of language change. In Directions for Historical Linguistics: A Symposium, ed. Winfred P. Lehman & Yakov Malkiel, 97-195. Austin: University of Texas Press. White, Lydia. 1987. Markedness and second language acquisition: The question of transfer. Studies in Second Language Acquisition 9.261-86. Williams, Rev. John G. 1895. A study in Gullah English. The Charleston (S.C.) Sunday News, Feb. 10. Author's ms. in the University of South Carolina Library, Columbia, SC. Wimsatt, William C. 1999. Genes, memes and cultural heredity. Biology and Philosophy 14.279-310. 2000. Generativity, entrenchment, evolution, and innateness. In Biology Meets Psychology: Constraints, Connections, Conjectures, ed. Valerie Gray Hardcastle, 139-79. Cambridge, MA: MIT Press. 2007. Re-engineering Philosophy for Limited Beings: Piece-wise Approximations to Reality. Cambridge, MA: Harvard University Press. Winford, Donald. 1997. Re-examining Caribbean English Creole continua. World Englishes 16.233-79. 2003. Introduction to Contact Linguistics. Maiden, MA: Blackwell. Wolfram, Walt. 1974. The relationship of white Southern speech to vernacular Black English. Language 50.498-527. Reprinted in Verb Phrase Patterns in Black English and Creole, ed. Walter Edwards & Donald Winford, 60-100. Detroit: Wayne State University Press, 1991. 2000. On the construction of vernacular dialect norms. In Papers from the 36th Meeting of the Chicago Linguistic Society. Vol. 2: The Panels, ed. Arika Okrent & John Boyle, 335-58. Chicago: Chicago Linguistic Society. Wolfram, Walt, & Natalie Schilling-Estes. 1995. Moribund dialects and the endangerment canon: The case of Ocracoke Brogue. Language 71.696-721. Wolfram, Walt, & Erik R. Thomas. 2002. The Development of African-American English. Maiden, MA: Blackwell. Wood, Peter. 1974. Black Majority: Negroes in Colonial South Carolina from 1670 Through the Stono Rebellion. New York: Alfred A. Knopf. Wylie, Jonathan. 1995. The origins of Lesser Antillean French Creole: Some literary and lexical evidence. Journal of Pidgin and Creole Languages 10.77-127.

Author Index Aboh, Enoch 136,139, 302n. 5 Adam,Lucien 137,139,140 Alleyne, Mervyn C. 41,43,141 Allsopp, Richard 135 Andersen, Roger 153,291n. 12 Ans, Andre-Marcel d' 140 Ansaldo, Umberto 194,222 Arends, Jacques 164,298n. 4 Bailey, Charles-James N. 45-6,108 Bailey, Guy 26, 39,42,45,65,67,95,108,274,276, 291n. ll,298n.5,312n.2 Baissac, Charles 95,302n. 5 Baker, Philip 124-5,144,146,173 Bamgbose,Ayo 199 Bardovi-Harlig, K. 153-6 Batibo, Herman M. 205,239 Beechert, Edward D. 109 Bennett, John 275 Berlin, Ira 172 Bhatia, Tej K. 195,214 Bickerton, Derek 13, 74, 77, 81,83,87,88, 89, 94-5,102,108-10,136,138,141,146, 205, 285n. 3,296n. 2,297n. 14,298n. 4 Bilby, Kenneth 140-1 Binde, Jerome 224 Bloomfield, Leonard 41,75,108 Boas, Franz 101 Bolinger, Dwight 83,121 Bolton, Kingsley 54,109,199,296n. 4 Bopp, Franz 285n. 4 Boretzky, Norbert 138 Botha, Rudolf 102 Bourdieu, Pierre 66,67,224,254,261 Breeder, Peter 267 Brutt-Griffler, Janina 98,198,217,232,247,255,259

Bruyn,Adrienne 160-9,178-9 Bybeejoan 178 Calvet, Louis-Jean 264 Campbell, Lyle 290n. 5 Cassidy, Frederick G. 36,138 Cavalli-Sforza, Luigi Luca 97 Chaudenson, Robert 34,40,42,43, 52-3,81,85, 106,118,119,136,139,140,142,145,146, 151,153,168,185,304n. 9,308n. 16 Chomsky, Noam 89-90,294n. 8 Christaller, Rev. J.G. 136 Clyne, Michael 53,209 Collins, James 194 Comrie, Bernard 154,296n.6 Coppola, Marie 78, 79 Come, Chris 118,123,135,141,142,144,146, 147,163 Couto, Hildo Honorio do 38, 245 Cowen, Noel 50,206,226,227, 309n. 5 Croft, William 129-30 Crosby, Alfred W. 214,232,257,299n. 7 Crystal, David 236,239,245 Cukor-Avila, Patricia 26,298n. 5 Daelman, Jan, S.I. 144-5 Darwin, Charles 12,94,97-8,298n. 3 Dawkins, Richard 20,287n. 16 D'Costa,Jean 41,108 Deacon, Terrence W. 298n. 20 Debose, Charles 138,141 DeCamp, David 41,106,108,276,279 DeGraff, Michel 34,42, 74,95,107,108-9,118, 136,139,142,147,152,177,295n. 15,303n. 4, 305n. 12 Dejean,Yves 223

338

Author Index Dillard,J.L. 305n. 3 Dunn, Richard S. 35,171 Elston,H. 60 Erhardt,S. 150 Fabian, Johannes 258 Faine, Jules 30,48 Faraclas, Nicholas 85,138,141,190 Fehderau, Harold W. 258 Fenton, Steve 98-9 Feral, Carole de 188 Ferraz, Luis Ivens 142-3,150,187 Fill,Alwin 129 Filppula, Markku 14,45,210 Fischer, David Hackett 215 Fishman, Joshua A. 205,307n. 4 Forston, Benjamin W. 37 Fracchia, Joseph 288n. 22 Frajzyngier, Zygmunt 175-6 Freeman, Edward Augustus 100 Friedman, Victor A. 236,296n. 20 Garnsey, Peter 218 Giliomee, Harmann 267,268 Gilman, Charles 62,137,138,139,141,148 Givon, Talmy 74-5,83, 88,90,91,136 Goldin-Meadow, Susan 78,88 Golovko, Eugeni V 293n. 23 Gonzales, Ambrose E. 299n. 11 Goodman, Martin 1991 237 Goodman, Morris 1964 30,47,48,144,288n. 18 Goodnight, Charles J. 288n. 26 Gould, Stephen Jay 98,298n. 3 Grenoble, Lenore A. 205 Guglielmo, Jennifer 98,231 Hagege, Claude 146,151,206,225,242,243,262, 263,264 Hair,P.E.H. 138 Hale, Ken 205,226,246 Hall, Robert A., Jr. 30,47,48 Hancock, Ian F. 137,138,277,305n. 3 Hanski, Ilkka 25,292n. 14 Harrison, FayeV. 98,99 Hatala,E. 93 Haugen,Einar 209,215 Hazael-Massieux, Guy 302n. 1 Hazael-Massieux, Marie-Christine 205,306n. 4 Heine, Bernd 29-30,162,170,178-9, 304n. 10

Herskovits, Melville 136,137 Hesseling, Dirk Christiaan 137 Hey, Jody 14,294n. 10 Hinton, Leanne 205 Hock, Hans Henrich 16,29 Hoenigswald. H. 13 Holm, John 38,43,106,121,134,137,138,141,193 Holzschuh-Neumann, Ingrid 106 Hookoomsing, Vinesh 292n. 18 Hopper, Paul 168,169,304n. 8 Hulstaert, Gustaaf 258 Hull, David L. 116,285n. 1,287n. 12,287n. 16, 290n. 35 Huttar, George L. 144-5 Hutton, Christopher 99,100 Irvine, Alison 108 Jackendoff, Ray 296n. 7 Jacobs-Huey, Lanita 93 James, Angela 98,99 Janson,Tore 218,219,220,307n. 7 Jespersen, Otto 75 Johnson, Guy 274,276 Jones, Sir William 99-100 Jones-Jackson, Patricia 312n. 4 Joseph, Brian 16,29,205 Jourdain, Elodie 141 Jourdan, Christine 188 Kachru,Braj 54,184,199 Kachru, Yamuna 184,199 Kallen, Jeffrey 210 Kapanga, Mwamba Tshishiku 303 Kashif,Anette 312n.6 Kaufman, T. 29,35,36,40,121,134,146,149, 167,200 Keiser, Steve Hartman 279 Keesing, Roger M. 85,144,146,148,150,165 KegUudy 78 Keller, Rudi 60-2,71 Kephart, Ronald F. 39 Kihm, Alain 137,296n. 6,305n. 2 Klein, Wolfgang 40,110 Kouwenberg, Silvia 142,148,150 Krapp, George Philip 96,100,138,276 Krauss, Michael 226,246 Kretzschmar, William A. 32,286n. 9,292n. 14 Kurath, Hans 70,96,100,276 Kuteva,Tania 29-30,162

Author Index Labov, William 23, 26,95,111-12,281, 312n. 7 Lafage, Suzanne 54 Lalla, Barbara 41,108 Landa, Manuel de 46,219 Lass, Roger 182 Le Page, Robert 43,84,187 Lefebvre, Claire 136,139,140,142 Levi-Strauss, Claude 299n. 13 Lewontin, Richard Charles 287n. 17,288n. 22 Li, Charles N. 75,83, 85,88,89 Lieberman, Philip 91,297n. 17 Lightfoot, David W. 30 Lippi-Green, Rosina 96 Lorenzino, Geraldo A. 187 Ludtke, Helmut 221 Lumsden, John 140 McCawley, James D. 87,289n. 30 McLaughlin, Barry 149 McMahon, April 13,23 McNeill, David 88,285n. 3 MacWhinney, Brian 78, 79,83, 88 McWhorter, John H. 36,42,102, 302n. 8 Maine, Henry Sumner 100 Manessy, Gabriel 54,138,141 Marks, Jonathan 97 Marling, William H. 241,255,31 In. 14 Maroldt, Karl Bailey 45-6,108 Masuda, Hirokuni 189 Mather, Patrick-Andre 140 Matthews, Stephen 194,305n. 5 Maurer, Philippe 143,150 Mayor, Federico 224 Mayr, Ernst 22, 294n. 10 Mazrui, Alamin M. 217,237 Mazrui,AliA. 217,237 Meillet, Antoine 36-7, 87 Meisel, Jurgen M. 149 Mesthrie, Rajend 196,197,215,266 Miele, Frank 97,101 Migge, Bettina 136 Milroy, James 183 Mille,Katherine 274,275,280 Mohan, Peggy 195,214 Montgomery, Michael 21 Mufwene, Salikoko S. 11,13,14,16,17,29,31, 32, 34, 37, 39-41,44,47,48, 60, 65, 67, 75, 82, 84-5, 90, 95,100,103,106,108-10,116-17, 119-20,123,129,138,141,144,146-51,155, 157,162,168-9,172-5,178,180,182,190,

197, 205, 207, 221, 233, 239,254, 274-5,278, 280, 285n. 3,286n. 5, 288n. 20, 291n. 8,14, 294nn. 7, 8, 296n. 3, 298n. 4,299n. 15, 301nn. 4,5,302nn. 4,8,305n. 3, 307n. 5 Muhlhausler, Peter 138,146 Naro, Anthony Julius 38,109,187 Nelson, Cecil L. 184,199 Nettle, Daniel 207, 210, 226,229,239,270 Nicolai, Robert 258 Nurse, Derek 258 O'Hara, Robert J. 294n. 10 Odlin, Terence 237 Osterhammel, Jurgen 206,226 Ostler, Nicholas 206,212,218,224,225 Pakir, Anne Geok-In Sim 194 Pandharipande, Rejeshwari 224,307n. 4 Pargman, Shed 31,183 Paul, Herman 18,285n. 2,285n. 4,288n. 25 Perdue, Clive 139-40 Petersen, Niels P. 226 Phillipson, Robert 240,241,242,243,262,263 Pieterse, Jan Nederveen 227 Plag, Ingo 161-3,169-71,179 Polome, Edgar 36,85,218,219,235 Pollard, Velma 42, 308n. 10 Poplack, Shana 42,277,282 Posner, Rebecca 30,48 Post,Marike 187 Prescod, Paula 39 Price, Glanville 226 Prudent, Lambert-Felix 77,190 Pullum, Geoffrey K. 84 Radick, G. 12,98,298n. 3,299n. 11 Reinecke, John E. 109,273 Renfrew, Colin 50 Rickford, John R. 106,108 Roberts, Sarah Julianne 190 Robertson, Ian 127,135,138 Romaine, Suzanne 207,210,226,229,239,270 Salerno, Salvatore 98,231 Sailer, Richard 218 Salmons, Joseph 215,300 Samarin, William J. 109,258 Sankoff, Gillian 78, 85,139,141,144,152,176,188 Sapir, Edward 71-2,140,247-8

339

340

Author Index Sarich, Vincent 97,101 Saussure, Ferdinand de 289n. 30 Scherre, Maria Marta Pereira 38,187 Schilling-Estes, Natalie 281,290n. 32,312n. 7 Schleicher, August 12-13,66,94,99,285n. 4 Schlieben-Lange, Brigitte 46,108 Schneider, Edgar W. 41,46,210 Schuchardt, Hugo 41,75,106,108,136-7 Schwartz, Stuart B. 291n. 12 Schwegler, Armin 39 Sebba,Mark 80,293n.22 Senghas,Anne 78,79 Seuren, Pieter M. 153,301n.3 Siegel,JefF 136,139 Silverstein, Michael 55 Singler, John 143,144,145 Skutnabb-Kangas, Tove 205,207,226,239,245 Slobin,DanI. 81,83,89 Smedley,A. 98 Smith, Adam 71,293n. 3 Smith, Reed 274,275 Smith, Geoff 305n. 5 Smith, Norval. etal 127,135,138,142,193 Sober, Elliott 116,287n. 15 Spear, Thomas 258 Sperber, Dan 286n. 12,287n. 14 Spitulnik, Debra 256 Sprouse,RexA. 140 Stewart, William 147 Swaan, Abram de 242,261,264 Swadesh, Morris 138 Sylvain, Suzanne 137 Szulmajster-Celnikier, Anne 208 Tabouret-Keller, Andree 43, 84,98 Tagliamonte, Sali 42,277,282

Tate,ThadW. 299n. 8 Taylor,PaulC. Ill Thomas, Erik R. 39,42,67,276,29In. 11, 312n.2 Thomason, Sarah G. 29,35,36,44,121,146,149, 151,152,167,200,210,286n. 7 Tomasello,M. 75,82,87 Tomlinson, John 227 Trask, Robert Lawrence 30,48 Traugott, Elizabeth C. 162,170-1 Tremblay, Xavier 36,29In. 9 Trubetzkoy, Nikolai S. 49 Trudgill, Peter 110 Turner, Lorenzo Dow 136-8,141,145 Valdman, Albert 154 Vansina, Jan 238 Vennemann, Theo 14 Vigouroux, Cecile B. 264 Vinson, Julien 12,95,99,302n. 5 Wade, Michael J. 288n.26 Wade-Lewis, Margaret 138,144-5 Wald,Paul 288n. 26 Warner, Anthony 304n. 6 Warner-Lewis, Maureen 145-6 Weinreich, Uriel 22,134,182 Wekker, Herman 153 Whaley, Lindsay J. 205 White, Lydia 153-6 WiUiams, Rev. John G. 273 Wimsatt, William C. 298n. 19 Winford, Donald 43,210 Wolfram, W. 95,107,117 Wood, Peter 276 Wylie, Jonathan 190

Subject Index Abidjan 269,271 Abolition 187,222,275,277 Aboriginal 256 Aborigines 247,298n. 3 abrupt evolution 34-5, 88,146,165,169,189,198 see also gradualness: pidgin-to-creole evolutionary scenario Academie Francaise 292n. 19 academy 67,99,243,262 academies 4,116 accent 53, 87, 300n. 17 accommodation 15-16,32, 54,65,68,90,92,126, 157,183 accumulative process, change as 11,14-16,22,32, 59-73,92,116,134-5,184,224, 232, 249 see also gradualness "acquisition", language 17,18,27,90,130,131, 200,285n. 3,288n. 21,290n. 34 Creole development 85-7, 89,133,140,156, 184-93 "cut and paste" model of LI learning 82,84 imperfect learning 35,104-5,183 imperfect replication 3, 32,36,86,117,125-6, 182-3 multilingualism 182-4 naturalistic language acquisition 3, 56-7, 89, 104-5,110,117,120,122,182-4,186,199,235 primary linguistic data (PLD) 75,81-2, 85, 86-7, 89,107,182 reconstructive process 2-3,18,85-9,173,182-4 SLA (second language acquisition) 35,104-5, 125,133,140,149-59,173,182-93,290n. 7 acrolect, application to non-creole communities 106-7 actuation 7,32, 58,70, 72 actuation problem 22,182

actuation question 3,22-6,32,44,59-73,116, 128,182-3,223,251,288n. 23 see also change ADJECTIVE 123,297n. 15 adposition 168 adstrates 53,293n. 21 see also substrate influences; superstrate influences advantage, social (speakers' selection aimed towards) 48, 55,67,77,125,194-5,223,224, 242,243,253-4,261-2 advantageous 48,125,224,253,257 Africa African languages as substrates see substrates influences Bantu colonization 31,128,228,238-9,252, 266-9 globalization 253-72 indigenous languages, loss of 197,253-72 language evolution in 244,253-72 African "basket names" 145 African American 26,93-4,95-6,277,279,290n. 32, 298n. 5, 300n. 17, 312n. 3, 312n. 5, 312n. 6 African American Vernacular English (AAVE) 26, 39,41-2,43, 53,106-7,111-12,186,197,239, 274, 276-8, 290n. 32,291n. 11 Africanization 199 Afrikaans 54-5, 265-9, 306n. 7 Afrikaner 54-5,265,267,268,310n. 11,31 In. 3 Afro-American 41 agriculturalist 217,228,238 Aleut Copper Island Aleut 293n. 23 allopolyploidy 132 American Association of Anthropology 97 American Association of Physical Anthropology 97 American Revolution 52

342

Subject Index American Southern English 44,100,278,280, 291n. 11 Americanization 4,207,215,225,226,240, 254-5 Americas see Canada; Caribbean; Latin America; Native Americans; North America Amish 43,298n.2 Amish English 43,52, 53,103,242 Old Amish Order 231,279 analogy 19,26 analogies 3,12,60,62 ancestral language, loss of 194-9,211,214-15 see also indigenous languages Angles 14,15,45,52,220,221 Anglicization 221,246 Anglo-Saxon 231,240,241 Angolar 44,187 Annobonese 187 apartheid 266-70 Appalachian English 53,63,231 apport 135,145,148 Arab colonization 218-21 Arabic 40,107,128,206,212,218,220-1,236 Arabization 220 architecture of language 79 areal diffusion 13 areal features 139,141 Argentina 193,229 Aryan 100 Asia Minor 49 Association of South East Asian Nations 214 "at risk" languages 54-5,268 Afrikaans 54-5,265-9,306n. 7 Atlantic Creoles 80,127,130,134,136,141-4,154, 296n. 5 Atrophy 253 Attrition 195,196,229,231,244 Australia 24,33,49, 53,55,69, 76,105-6,110,193, 207,210-11,216,227,229-31,237,239,240, 244,248-9,254-6,270,296n. 5,298n. 3, 310n.8 Baba 194 Baba Malay 194 Babu English 229 baby-talk theory 40,101,105,137 Bajan 43 balance of power (competing variants) 234,32, 105,118-19,176-8,267 Balkans 236

Bantu Bantu colonization 31,128,228,238-9,252, 266-9 Bantu languages 127-8,143,152,155-6,166, 228,238, 252,266-9 baragouin 190 Barbados 186 basilect, application to non-creole communities 106-7 basilectalization 76-8,82,85,189,274-5 Basque 109,206,219 bazaar 194 Beke 75,265 Berbice Dutch 36,40-1,127,142,148 Bermuda 47,77 Bhojpuri 195-6 bidialectalism 239,280 bifurcated evolution 46,124,169 bifurcated paths 169 bifurcation 46,70-1,124,169 big bang mutation 88 Bight of Biafra 44,29In. 12 bilectalism 280 bilingualism 68-9,239,261-2,280 individual bilingualism 68 societal bilingualism 68 biodiversity 207,245 biological approach to language evolution 11-28 feature (gene) pool analogy 17,20-1,27,69,108, 117-32,149,152,173,178 idiolects as organisms analogy 14-15,26-7,116 species, languages as 11,14-15,18,21,23,25, 26-7, 63-73,69,119-20, 208-9,285n. 1 viral species, languages analogous to 11,18,21,23, 25, 26-7,69,208,285n. 1 biological endowment for language 89,285n. 3, 303n. 9 bioprogram birth, language 1,205-24 bioprogram (language bioprogram hypothesis) 13,81, 834,95,104-5,121,136, 144, H6-7,151 Black 6,34,46,52,77,95,111,215,238,258, 269,271 Black nannies 100 Black South Africans 196,267 blending inheritance 59,68 boarding school 235,237,244,260, 309n. 7 Boers 266,267 borrowings 146,165 see also apport

Subject Index Botswana 238,239 Bozal 34, 38-9,46-7, 52, 55, 78,101,103,107, 140, 155,185,186,196, 230,289n. 27, 291n. 12, 292n. 17,310n.9 Brazil, lack of Creoles 37-40,186-7 Brazilian Portuguese 35,43,187,192 Portuguese in Brazil 211,229 break in the transmission 104,191 Bretagne 242 Breton 219,242,263 British Isles 45-6,52, 53,200,211,218,237 brogue 281,290n. 32, 307n. 9 "Cafeteria Principle" 141-2,145,148 Cajun French 52 Canada 69,232 Canary Islands 29In. 12 Cantonese 222,251,293n.24,305n. 5 Cape of Good Hope 306n. 7 Cape Verde 39-40,187 Cariacou 39 Caribbean 37, 39,41,43,53, 55,77,100,106-7, 111, 131,145-6,189-90,193,195-6,236,257, 265,266,293n. 22 Creole evolution 39,41,43 indigenous languages, loss of 195-6,257 influence an American English varieties 53 race 111 Castilian Spanish 38 Caucasia 49 Caucasian 98 Celtic 52, 211,219-20,235, 307n. 7 Celtic Englishes 45, 304n. 5 Celtic languages 14,16,20, 31, 37,40,47,48,102, 127,128,166, 219, 221, 235, 307n. 9 central language 261,265 change agency of (actuation question) 3,22-6, 32, 59-73,116,128,182-3,223,251 as cumulative process 1,14-16,22,32, 59-73, 86, 92,116,134-5,184,224,232,249 see also gradualness external vs internal motivation 29-32, 58,110, 160-1,182-3 language change (as term) 16-17 children child language 80-2,91,298n.4 innovation 74, 80,297n. 9 LI development 87,88-9,184 see also "acquisition'

language bioprogram hypothesis (LBH) see language bioprogram hypothesis (LBH) local vernaculars 40-1, 53, 77,118-19,185,198 not being Creole creators 78-83 Chinese 147,188,251,261 Chinese Pidgin English 109,188,296n. 4 Chinook largon 55,77,109,229 Chinookan 77 Choctaw 77 Christian mission 46-7,219,259 cities 194,197-8,215,23-34,260,262,269 urban vernaculars 94-5,212,216,222-3,233-4, 244-5, 255, 258-9,260 urbanization 18,232,262,267,269 cladograms 49-50, 50/, 246,309n. 4 see also Stammbaum classic Creole 76 classic pidgin 77 classifier 129-30,144,152 cluster-bomb model 37,38*' Cockney 107 code-switching 277-8,280-1 coevolution 75,101,188 Colombia 143 colonial auxiliary colonial auxiliaries 45,198,232,244,255,256, 259,295n. 17 colonial Englishes 45-6,53,96,131,155,187,192, 200, 274,276,292n. 16 colonization compared to globalization 215-16,224,226-41, 250-1 differing styles of colonization 200,234-5,270-1 exceptionalism of Creoles theory 30-1, 33-45 exploitation colonies see exploitation colonies exploitation colonization 56, 99, 217,221, 228, 251, 265, 270, 271, 305n. 3, 308n. 15 homestead phase see homestead phase settlement colonies see plantation settlement colonies; settlement colonies settlement colonization 49, 55,190,217,220-1, 228-9,235, 240, 250,270, 299n. 7 trade colony 188-9,229 trade colonies 74, 103-4, 108-9, 171, 188, 194, 309n. 6 trade colonization 221,228-9,234,270-1,308n. 15 Colored 266,267-8 Colored people 4-5 communal communal grammar 111

343

344

Subject Index communal (Cont'd) communal language 2-3,14,15,32,41,45,59,62, 64-6,68, 73, 83,86,87,89,91,116,125,173, 290n. 7,294n. 7,295n. 15,301n. 6,302n. 10 communal norms, development 59-72,157, 173-4,163,201 communal systems, languages as 25-6,32,89,92, 126-7,181,208-9 communicative needs, languages as tools to meet 76-8,92,100-1,118,194,195,253 comparative method 13,29-30, 36-7,44 competition competition and selection 4,19-22,27-8,56,58, 66-7,90-2,115-32,150,183-4,200-1,259, 287n. 15 Complementary Hypothesis 29-57,136,139,144 complementizer 84,86,117,124,161,169,174, 175,176-7 complex adaptive system 119-20,131 congruence 118,119,123-4,135,151,163 congruent 124,132,139,140,177 congruous 105,118,135 conscious vs unconscious speakers' choices 116,126 constellation, language 226,264 constraints 3,128-31,147 contact contact-based approach 13-14,24-5,29-58,67-72 individual interactions as basis of 17,31-2,50,119 as overlooked factor 49-57,105-6,109,127-8, 209-10 continuum 41,109,185 Creole continuum 41,91,107-8,174 Creole continua 222 geographical continuum 41,29In. 10 geographical continua 43 speech continuum 106 contract laborer 6,51,104,109,141,145,187-8, 195-6,199,214,230,265,266,292n. 17 convergence, language 13,96,109,296n. 20 cooperative rather than team, speakers behaving as 92,181,183 copying 61-2,66-7 costs and benefits 47,56,228,252,31 In. 16 Coted'Ivoire 269,271 Creole see also named Creoles; English Creoles; Gullah; Guyanese Creole; Haitian Creole; Hawaii; Jamaican Creole; Kikongo-Kituba; Palenquero; Papiamentu; Portugese Creoles; Saramaccan; Sranan; Surinamese Creoles; Tok Pisin Creole continuum 91-2,106-8,174

Creole development 29-58,74-88,171-2, 184-93,199-200 "creole features" 41-2,104,146-7,190 creolization 29-58,171-2 decreolization see decreolization exceptionalism argument 29-31,33-58,71-2, 102,107,109-10,165,179,205,246 generations of Creoles 292n. 15 intermediate creole 43 language loss in creole communities 196-9,201 pidgin-to-creole evolutionary scenario 34,41, 74-92,102-3,138,164,171-2,190,298n. 4 "primitive" languages 12-13 semi-creoles 38,40-2,43-4 see also AAVE structural heterogeneity 36-7,91-2 creolization 40-2,46,171-2,179,180-201 Crioulo 39,291n. 12 critical mass 102-3,175,197-8,200,266,279 critical period 80-1,140,148,200 Cuba 38-9 cultural influences 145,227,244-5 cultural spaces 227 Curasao 143 "cut and paste" model of LI learning 82,84 Dakar 269 Danes 220,236 Danish 128,209,243,244 death, language 1,15,205-24,257 Debasilectalization 28-34 see also decreolization decreolization 41-2,107-8,138,151,291n. 10, 307n. 10 AAVE 106 Gullah 273-84 deforestation 215,245,248 Delaware Pidgin 55, 77,190,229 Democratic Republic of Congo (DRC) 216,258, 311n. 18 depopulation, as threat to language vitality 283 demolinguistics 253-71 deviation 2-3,61-2,66,82,87,134,136,148, 156-7,181,183,185,201,291n. 13 dialectologist position 137-8 see also superstrate influences Diaspora 228,251,310n. 13 diffusion 4, 29-30, 36,219,227,231,254,271 diglossia 233,280 disadvantageous 216,223,249,286n. 9, 308n. 10, 309n. 19

Subject Index discontinuity hypothesis 36,161,163-4 see also "transmission", breaks in Discrimination 39,98,231,304n. 7 disfranchised population 245 dispersal 30,37,46,49,292n. 20 dispersal point 49, 54 divergence degrees of divergence 29In. 14 divergence hypothesis 111,274,276,312n. 2 diversification 14,24,29-30,31,50,52,54,70-2, 83,102,103,109,193-6,207,234,246,309n. 4 diversity, linguistic 226,246 division of labor 223,282 Dominican Republic 38-9 Dread Talk 54,308n. 10 drift, language 71-2 dual colonization 54,31 On. 11 Dutch 5,14, 36,40-1,43,123,135,142,164, 166-7,187,191-2,209,231, 243-4,262-3, 265-7,279,292n. 18,306n.7 dynamics of coexistence 4,29In. 14 East Indian 195-6,230 Eastern Ijo 36,40,41,142,148 ecological motivation 58 ecology 4,16-26, 31,43,47, 85,117, 246-7,273 external ecology 3-4,20-21,23,116,121,181, 290n. 34, 304n. 6 internal ecology 20-21,23, 53,116,193,276, 294n. 7 social ecology 23,27,90,94,97,115,141,299n. 11 socioeconomic ecology 20-1,224,226,231-2,246 econiche 63,223,259 economic pressure 54,234,246,250 economics 211-12, 213-14,222,239-41,245-50, 271,293n. 3,307n. 10 see also colonization; globalization language economics (costs and benefits to speaker) 253-71 education 56-7,196,199, 230, 231,235,241,245, 259,264-5,277-8 E-language 120,182,185 elite 46, 54, 56,124,198,216,217,219-23,232-3, 235,237,244,251,255 Elmolo 238 Emancipation 140,145,230,269,310n. 9 emergence emergent "system" 221 emergent basilectal grammar 223 emergent Creoles 20,47, 81,101,147,151

emergent grammar 123,175 emergent norms 134-5 emergent pattern 153,289n. 30,294n. 7 pattern emergence 2,60 patterns emerge 2,4,25,59,301n. 5 patterns emerging 42 enclave population 252 endangered languages 205-84 endogenous 40,47,78,126,152,166,222,238, 290n. 4 English American English varieties see North America creole-like evolution 31,45-9,167, 219-22,236 dominant vernacular of Western Europe 225-50, 262,263 English Creoles 84, 86,110,122,123,129,144, 153,154,163-4,174-6 "global language" 236,239-41 in Hawaii 189-90 "killer language" 207,225-52,254,271-2 lingua franca 236-7,264 in South Africa 265-9 world language 222,239-40,241-5,264 English Creoles 84,86,110,122,123,129,144,153, 154,163-4,174-6 Ethnic background 93-112 indigenous languages as markers of 216-17,233 ethnicity 96-9, 111, 248,298n. 2 ethnic cuisine 98,256 ethnic foods 98 ethnic identity 26,217,233,247,252,277,283 ethnic language 55, 57,76,156-7,194,196,198, 217,233-4, 257, 260, 267, 306n. 8 ethnic neighborhood 98,198,256 ethnicity index 248 ethnographic function 201,242,255,263 ethnography of communication 116 ethnolinguistic homogeneity 40, 78,143,214 Eurasia 127 European American 93-6,110-12,274,276-80, 282,289n. 27,298n. 2,298n. 5,300n. 17, 312n.3 European Union 206,214,226,241-2,262,263, 311n. 14 evolution 1 differential evolution 42,44,45,54-5,111,189, 214,215,216-17,298n. 3, 307n. 7 evolutionary linguistics 58 evolutionary pattern 62-3, 70,192-3,271

345

346

Subject Index evolution (Cont'd) gradual (...) evolution 44,46,48,168,169 transformational evolution 58 variational evolution 124-7 evolution, language, as term preferred to language change 16-17 evolution of human language 74, 75, 79, 83-92, 94,296n. 6 exaptation 61, 88,125,173-8,293n. 5, 304n. 8 exapt 293n. 5 exceptionalism 290n. 2 exceptionalism of Creoles theory 29-31,33-58, 71-2,102,107,109-10,165,179,205,246 exclusionary language 237-8,265 exogenous 40-l,47-8,290n.4 expanded pidgin 76,134,188,229 Melanesian expanded pidgins 85,130,137,138, 144,152,155,165,296n. 5 exploitation colonies 188,212,216-17,222-3,228, 231-4,251,256 external vs internal motivation for change 29-32, 58,110,160-1,182-3 extinction 24,57,196,207,213,223,259,289n. 27 Fad'Ambu 187 Falkland Islands 24,49,289n. 28 family resemblance 7,14,26,63,66,92,117,126 Fanakalo 56,258,266 feature/gene pool 17,20-1,27,69,108,117-32, 149,150,152,173,178,183-4, 303n. 4 Filipino 189 Fongbe 136 footpath 60-1 foreign workers' German 40 founder founder colonist 36 founder effect/Founder Principle 134,140,142, 143,158,276,289n. 28,292n. 17 founder population 34,84,292n. 17 fran$ais africain 47,54,231,232 fran$ais populaire 146,269 fran$ais populaire ivoirien 269 Francophonie, La 225,240,242-4,264 franglais 292n. 19 Prankish 128,206,236 Franks 15,220,303n. 5 French see also Romance languages French Creoles 154-5 Quebec French 44,127,177,211-12,232,306n. 1 world language 225,239-40,242-5,251,264-5

Gaelic 206,279, 306n. 3 Gaeltacht 241,242 Gaul 5,128,206,218,220, 223,235, 236 Gaulish 166 Gbe languages 140,142 genetic linguistics 11-14,16,29,44,58, 70,205-6, 207, 209, 246, 286n. 8,289n. 26,291n. 9, 306n. 7, 309n. 4 genocide 223,257,306n. 2 Georgia 70, 274, 277-8,279, 280-1, 282-3,290n. 32,312n.9 German 14,15,37,40,43, 53,56,156,166,167, 200, 209, 215, 221, 229,231,240,249,262, 265, 298n. 2, 300n. 17, 308n. 12 Germanic 14,31,42,44,45,49-52,127-8,220-1, 236,295n. 15 Germanicization 219-20,236 Germanicize 236 "global languages" 236,239-41,310n. 12 globalization 4,206-7,213-14,225-52,254-7, 271-2 compared to colonization 210-11,215-16, 226-41,250-1 confused with McDonaldization 210-11 glocalization 214,239-41,250-1, 256-7 Golden Age Principle 289n. 29 gradualness 86-7,88-9,162,168-71,172-8,224, 300n. 17 grammar 83-4, 121-2,145 grammaticization 61,67,160-79,173-4 pidgin 79-80,91 grammaticization 61,160,162,169,178-9 apparent grammaticization 161 idiogrammaticization 173 instantaneous grammaticalization 170,181 paths of grammaticization 166,174 polygrammaticalization 171 gravitational model of language coexistence 264-5 Great Vowel Shift 45,167 Greek 237, 307n. 7 Ancient Greek 212-13,299n. 12,307n. 7 "gravitational" model 264-5 group identity markers 126,232,244-6,260, 276-8 group vs individual selection 66-7, 173,183 Guinea 143,229 Guinea Bisau 187 Guinea Coast Creole English 305n. 3

Subject Index Gulf of Guinea islands 143,187-8 Gullah 30, 36,100,107,108,121,123,137,145, 157,160,169,174-5,176-7,193, 239, 273-84, 297n. 13,307n. 10 Guyana 40-1,43,142,143,195,199,266, 306n. 7 Guyanese Creole 108 HABITUAL 62,84,143,153 Haiti 39,223 Haitian Revolutionary War, 197 Harden 44 Haitian Creole 42,48, 81,136,137,139-40,142, 154,160,177,301n.4,307n.4 half-breed 98 Hanseatic League 219,221 Hausa 56,244,258 Hawaii 6,102,109,189-90,198,229,292n. 17 Hawaii, pidgin and Creole development in 102-3, 166,189-90,198, 229, 297n. 10 hegemony 242 hegemonies 206 Hellenic 50 Hellenic Empire 49,109,234,237 Hellenization 237 heritage language (see also ancestral language) heritage languages, loss of 211 see also indigenous languages heuristic prototype 33, 75 Hiberno-English 210 highway traffic analogy 59-73 Hindi 195,212,251,261, 310n. 13 Hip Hop 54 Hispanic 98 Hispanicization 100 Hokkien 194,222 Hollywood 207,240,254-5 home sign language 79, 88 homestead 34, 38, 76-7,100,125,152,171-2,191, 194,214,230,304n.7 homestead phase 34,38-9,76-7,100-1,104,109, 164,171-2,185,186-7,190,197,214,230,274 hominid 74, 78-9, 88-9,90,91-2,298n. 3 Hong Kong 222 Huguenot 265 human language development (protolanguage) 74, 75, 79, 83-92, 94, 296n. 6 hunter-gatherer 207,217,228,238 hybrid 22,30,132,250 hybridism 116,120-4,127-8 hybridization 14,129,139, 227, 302n. 5

hypercorrect 312n. 7 hypo-descent 127 Iberia 5, 128,206, 212,218,220-1,223, 235-6 identity community identity 278,282 ethnic identity 26,217,247,277,283 group identity 126,232,276,277-8 identity markers 126,216-17,232-3,244-6,260, 276-8 idiogrammaticization 173-4 idiolect 3,14,19, 27, 59-60, 62-3,65-6,68-9, 107, 111, 116,119,122,125-6,131-2, 170,173 idiolects as organisms analogy 14-15,26-7,116 I-language 181,182,185 illiteracy 212,245 imperial language 35,104-5,183 imperfect replication 3,36, 86,117,125-6,182-3 indentured servants 21,33,52,84,96,100-1,109, 215,299n. 8 see also slaves India 107,195-6, 212,214-15,229,232,292n. 17, 306n. 7 Indian Ocean 6, 20, 24, 31, 33,47,49, 75, 82, 189-95,198-9, 201,207,217, 222,266, 287n. 17, 288n. 18,289n. 27, 292n. 17 Indian reservations 308n. 13 Indie 100,130,212,266 indigenization 56-7,184-93,198-9, 210, 217, 222, 232-3, 248, 250 indigenized Englishes 46,47, 54,119,135,231, 232, 237, 246 indigenized varieties 1,47,184-93,198-9, 209-10, 255 indigenous indigenous language 24,55-6,69,210-12, 214-17, 219-20, 222-3, 227, 228-9, 233, 239, 242, 244-7,248-9, 251, 254,263,264-5, 270-2,289n.27,293n.21 indigenous peoples 229, 307n. 5 indigenous societies 55,307n. 5 indigenous systems 307n. 5 indigenous languages vitality of 194-9,205-21,229-41,245-50, 255-6, 262 individual speakers agents of change 3,22-6, 32, 59-73,67,116,128, 182-3,223,251 as basis of contact situations 17, 31-2,50, 59-73, 119

347

348 Subject Index individual speakers (Cont'd) individual vs group selection 66-7,173,183 multilingualism 180-1 relationship with language community 59-73 Indo-European dispersal 37, 38f, 49-57, 5If, 99-100,102,109 Indo-Iranian 99 Indonesia 212,222,306n. 7 inflection, lack of in Creoles 81 inheritance 30,32,44-45,68,121,288n. 22 innovation 61-2,63,66-7, 74,80,83,87,111, 125-6,151,297n. 9 see also change Inquisition 220 interethnic war 252,257 interference 17, 79,110,125,134 interlanguages 101,155,157,159,190,192 internal vs external motivation for change 29-32, 58,110,160-1,182-3 interpreters 76,103,109,188,198,258,259 "invisible hand" 2,3-4, 59-73,92,116,119,134, 161,183,208 Ireland 56-7,237,241,244,309n. 7 Irish 14,56-7,98,215,231,241-2,299, 307n. 8 Irish English 45,237 isiNdebele 268 isolation 39-40,128,278-80,282 see also segregation Italian 14, 53,56,200,215,229,231,249,298n. 2 Jamaica 35,43,48,54,140,143,145,195,223 Jamaican 42,44,86,175 Jamaican language 300n. 18 Jamaican Creole 106,108,124,140-1,162,170, 174-5, 307n. 10 Japanese 111, 181,189,236,246 jargon 41,190,229,275, 307n. 9 Java 194 Jim Crow laws 291n. 11,312n. 2 "joints" 279 Kenya 238-9,252 Khoekhoe 128 Khoisan 31,228,238-9,270 Khoisan populations 228,238-9,252,266,267 Kikongo 145,153,156,288n. 20 Kikongo-Kituba 34, 56,152,153,155,166,258 "killer" language concept 208,223,294n. 11 English 207,225-52,254,271-2 Kimanyanga 152,153,155,258

Kinshasa 216,263,269,31In. 18 Kinubi 258 Kituba 152,153,155 see also Kikongo-Kituba Kiyansi 166 kludges 173 koines 21,31,33,34,45,52,84,100-1,148,184-6, 190,221,274-5 koineization 5,21,45 koineize 195-6,221 Korean 189,251 Kwa languages 12,122,136,140,142,150 Kwa-Zulu Natal 196 La Francophonie 225,240,242-3,264 lack of pride 208,211,212,231-2,257 Lamarckian nature of linguistic species 27,65-6, 68 see also organisms, idiolects as Ian9ados 305n. 3 language language as practice 32 at risk 55 language bioprogram 119,121,146 language bioprogram hypothesis 13,18,95, 104-5,136,138 language birth 1,205-24 language choice 4,254,261 language convergence 96,109,296n. 20 language custom 183 language death 1, 209, 210,217-18,234, 286n. 7, 306n. 2 language economics 254 language endangerment 7,32,207,210,211,224, 225-6, 231-2,250,253, 263-4,272, 290n. 6, 294n. 11 language faculty 90,94,109 language loss 196,201,206-7,215, 223-4, 232, 234, 249, 252, 309n. 4 language missionaries 255 language practice 1,4,6,241-2 language purity 30,46,102 language shift 33,41,45-6,47,69,193-9,201. 211,214-15,219,248,251,263 language stratification 259 language vitality 238-9,258-65,267,276 language war 28,223,226 languages of culture 94 world language 15,237,239-40,251,310n. 13 language bioprogram hypothesis (LBH) 13,81, 83-4,95,104-5,121,136,144,146-7,151 language rights 224,226,243,245,246-50

Subject Index langues minorees 210,212, 306n. 4 langue minoritaire 306n. 4 Latin 5, 20, 37,46-9, 56-7,99,124,165, 209, 235 Classical Latin 46,48, 52,209,212-13,218,235, 237-8, 243, 307n. 8 creole-like evolution into Romance languages 30-1,45-9,102,127,167,212, 218-20,234-8 dead or not? 46-7,209,213 Vulgar Latin 5, 31,40,46-7,48-9, 52, 57, 72, 127-8,165,166-7,179,193,209,211-12,219, 223, 235, 237, 246, 307n. 8 Latin America 211,215,230,232,262 Brazil, lack of Creoles 37-40,186-7 layering 168,169 learning, imperfect 35,104-5,183 least effort, principle of 123,144,146-7,151,155 legionaries 47,219,251 leveling 140 Leeward (Islands) 306 lexical learning hypothesis 151,297n. 14 lexifier 1, 5,20,33,36,40,41-5,47,48, 52,84-5, 102,104-7,109-10,117,119,121-5,127,130, 133-8,141,144,146-9,150,152,155,157, 159,160-6,168,169-70,172-6,179,189. 200, 274,276,288n. 18,291nn. 10,14,299nn. 14, 15, 300n. 16, 302n. 4, 303nn. 8,4, 305n. 11 life-cycle 41 Lingala 34,56,152,153,216,258,263,31 In. 18 lingua francas 55, 56,171-2,186,212,217, 222-3,226,227,233-4,236-7, 242-5, 258-65,270 Lingua Franca 46-7,54-7,134,135,173,187,198, 199,201,209, 212, 221-2, 226, 233, 236-7, 240, 243,244,255-6, 258-64, 267-8,269-72, 306n.7 Lingua Geral 190,229 linguascape 258 linguicide 223,226 linguistic aberrations 302n. 5 linguistic advocacy 243 linguistic area 29,145, 308n. 11 "linguistic capital'Vlinguistic market 67,224,254 see also market value linguistic diversity 22,246,258 linguistic drift 72 linguistic hybrid 122 linguistic market 254 London West Indian Patwa 293n. 22 Louisiana 44,211,265,306n. 1

Louisiana Creole 5,103 Louisiana French 2,103,127 Maa 238 Maasai 238,252 macaronic mixing 200 Macau 193 Madeira 291n. 12 maintenance, language 242,245-50,258-65 see also revitalization major languages 31 In. 13 Makhuwa 197 Malay 194, 266, 306n. 7 Malaysia 212,222 Mandarin 147,194,251 Mande 136,141,143 marginalization 264 marginalized language 306n. 4 marginalized languages (langues minorees) 210, 212,306n.4 markedness 31,129-31,144,149-56,158-9, 297n. 11 markedness value 157 scale of markedness 230 market value 67,201,211,224,242,243,253-71 Maroon 292n. 18 Martha s Vineyard Island 281, 312n. 7 Martiniquais 44 mass behavior 64-5 materiaux de construction 147,178 Mauritius 48,191,196,199, 269, 292nn. 15,18 McDonaldization 4, 207, 210-11, 227, 240, 254-5 media 260-1,267-8,282 Melaka 194 Melanesian expanded pidgins 85,130,137,138, 144,152,155,165, 296n. 5 Melanesian islands 135 Melanesian pidgins 130,134,137,144,152,155, 176,188,296n.5 mesolect application to non-Creole communities 106-7 mesolectal 43 metapopulation 25,292n. 14 migration 23-4, 70,156-7,256-7,279-80 Minnesota 70 minoration linguistique 264 minority language 206,212,263,306n. 4 miscegenation 39,43-4,187,215 mission civilisatrice 98,217

349

350

Subject Index missionary missionaries 219,240,255 mixed idiolect 122 mixed language 29,126,165,166 Mobilian 55,77,190,229 modernity 222,233,256,264 mondialisation 207,226-7,309n. 1 monolingual/monolectal speakers 135 rnonolingualism (societal) 231,244,250-1 Moors 212,218,220,236 moribund 53,69,283,306n. 3,307n. 8, 310n. 10 morphological erosion 146 Mulatto 34 multiethnicity 248 multilectal 22,94 multilingual 95 multilingualism individual multilingualism 180,201,241, 307n. 6, 31 In. 14 societal multilingualism 180-201,211,214,230, 242, 244, 247,248, 253-4, 258, 272 spatial multilingualism 201 multinational corporation 213,226-7,240, 251, 254, 255 museum culture 241 mutual accommodation 15,32,90,92,126 mutual intelligibility 15,27,107

nonnative accent 87, 300n. 17 nonnative speakers 172,192-3,194,199,251 nonstandard varieties 46-7,48,84,122,146, 155-6,163-4,186,213,237 normalization 82,91,153,296n. 8 Norman French 31,45,124,127-8,167,206,220 normative pressure 153 norms 3,61,78,89,116,134,157,159,183 North America see also Canada: Quebec French AAVE (African Amercian Vernacular English) 26, 39,41-2,43, 53,106-7,111-12, 186,197,239, 274, 276-8, 290n. 32, 291n. 11 European American English varieties 25-6,70, 105-6,111-12,200,221,244,289n. 28,295n. 14 indigenous (Native American) languages 105-6, 190, 211-12, 215, 229-30, 232,244,256, 257,262 pidgins 55 race 93-112 North American Free Trade Agreement 214 North Carolina 39,95,276,283,290n. 32 Northern Cities Vowel Shift 23,26, 54,67,95, 111-12 Norwegian 209,243,262 NP Accessibility Hierarchy 154 numeral classifying 129-30 Nyonya 194

Namibia 238 Native Americans 105-6,190,211-12,215, 229-30, 232,244, 256, 257,262 Native American English Pidgin 55 nativization 189-90 see also indigenization "natural selection" 116 see also "invisible hand" naturalistic language acquisition 3, 56-7, 89, 104-5,110,117,120,122,186,199,235 Ndjuka 145 Netherlands Antilles 39,186,187 network 25 communication network 62,93,115-16,117, 126,182 New England 44 New Netherland 192,306n. 7 New World 6,15,20,31,36-7,49,96,101,103, 138,141,143,266 Newfoundland 96 Nicaraguan Sign Language (NSL) 78-9,80,88 Nigeria 56, 76,107,148,229, 258 Nigerian Pidgin English 76,258 Niger-Kordofanian 128

obsolescence 205,209,245, 306n. 3 Occitan 206 Oceanic languages 144 Ocracoke brogue 281,290n.32 Ocracoke Island 281,283,312n. 7 official language 54, 57,199,209,233, 239, 255, 264-5, 31 In. 14 Old English 5,14, 31, 70, 72,124,165,166-7,221, 236, 340nn. 5, 6 Old French 5,48,128,166 Old Norse 31,127,128,167 Old World 15,229-30,257 one-drop rule 111 ontogeny ontogenetic 14,18,27,68-9 ontogeny-recapitulates-phylogeny argument 83-92 optimality theory 129 organism idiolects as 14-15,26-7,116 langauges as 13 organism as term meaning species 285n. 4

Subject Index ORIGIN VS INFLUENCE 121

origins of features 48 osmosis 148,178,184 osmotic 68,144,178,184,292n. 19 outlier 111 Ozark 53,278 pace-setter 64 Pacific 76, 77, 82,131,144,188-9,190,195,196, 229,251,296n.5 Palenquero 143 Papiamentu 39,42,48,302n. 8 parameter setting 83,91,130-1 purler de seconde generation 43 parochialism 33,201,306n. 8 partnership 225,228,242,264 pastoralist 207,217,228,238 patois 193,219,291n.8,307n.9 pattern emergence 2,59-73, 87 Penang 194 Pennsylvania Dutch Angolar 44,187 PERFECT 80, 84,123-4,144,153 personality, as a factor in language change 67,120 phylogeny 6,13, 74,83 phylogenetic 1, 89,90-2,298n. 4 pidgin English etymology 76,109,188,190,229, 258, 296n. 4, 305n. 3,305n. 5 pidginization index 138 pidgins see also expanded pidgins geographical locations of 35*, 102-3,188 grammar 79-80,91 lingua fr ancas 171-2 pidgin-to-creole evolutionary scenario 34,41, 74-92,102-3,138,164,171-2,190, 298n. 4 "primitive" languages 12-13,91,94-5 trade colonization 35f, 55, 74-8,102-3,171-2, 188-9,229 pied-piping 85,153-6,167,177 plantation settlement colonies 6,20-1,24,33-40, 51, 76-7, 82-3,100,102-3, 109,143,145, 171-2,185,186-93,197, 214, 289n. 27 polygrammaticalization 171 polyploidy 19,27,120 population genetics 11-28 population growth 43,44,148-9,186-7,188, 200-1 population movements 29-58 see also contact

population replacement 39,143,188,275 population structure 4, 6-7,24-6, 34, 39,44,46, 54,56,71,111 population subdivision 288n. 26 Portuguese 16,38,44,121,143,187-8,192,193,211 Portugese Creoles 187-8,193 potato potato farming 237 potato plantation 56,244, 309n. 7 pragmatic context, replacing grammar in pidgins 91 Predicate-Clefting 80 prelinguistic 85 preposition-stranding 84-5,110,153-6,177 preservation 213,215,242,245,247-8,31 In. 15 preservation, language 213,242,245-50 pressures adaptive pressures 6,297n. 17 communicative pressures 195,295n. 19 socioeconomic pressures 246,250 prestige 21,46-7,48,208,212,216,218,223, 233-4,237,239,242,256,257,263 pride 6 lack of 208,211,212,231-2,257 primary linguistic data (PLD) 75, 81-2, 85, 86-7, 89,107,182 primates, non-human 91,296n. 6,297n. 17 primitive 98 "primitive" languages 12-13,91,94-5,99,137, 295n. 15 Principe 48,143 Principense 44,187,291n. 12 principle of least 123,144,146-7,151,155 Proto-Indo-European 5, 54 protolanguage 74, 75, 79, 85, 88, 92,94 prototype 33,42, 75,151, 302n. 8 prototypic 136,269 prototypical Creole 42,147,152 pulverization 36,102,189 "pure" languages 34,46,102,103-8 Putonghua 236 Pygmy 31,128 Pygmy populations 228,238-9 Pyrrhic victory 17,20,51,207,235 Quebec 239,241,242,264 Quebec French 44,211-12,232,306n. 1 Quebecois 35,127,177 Queensland 193,296n. 5,302n. 7 Q-value 261

351

352

Subject Index race 30,334,93-112,231 race discrimination 39,171 race purity 30,109 race segregation 39,118,191 race-based segregation 103 racialism 93-112 radical radical Creole 36,136 ranking the variants (see also weighting) 30In. 8 reanalysis 171,173 recombination 118,120,147-8 feature recombination 3,103,299n. 15 gene recombination 3,18,27,120,127 reconstruction 2-3,18,85-9,173,182-4 recreation 32,87,102,121 rectilinear 46,49,70,162 "recycling" 178 refugeeism 234,257,272 refugees 223,234,257,272 relativity hypothesis 247-8 relativizer 124,305nn. 11,13 relexification hypothesis 40,105,110,137,139, 140,142,148,200,299n. 16 replication replicate 81,84,287n. 16 imperfect replication 3,36,86,182-3 perfect replication 117,125-6,182 Republic of Congo 212,252,258 reservations 170,230,244 Indian reservations 308n. 13 restructuring degree of restructuring 106 RESULTATIVE

123

resumptive pronoun 154-5 Reunion 47,48,191,292n. 15 revitalization 210,213,224,247,268,31 In. 15 rhoticity 115 rice 143,186,217,230 rights human rights 243 language rights 243,245,246-50 rights of languages 24,226,243,245,246-50 Roman colonization 30-1,218-20,234-8 Roman Empire 31,212,218,219,234-7,259, 308n. 16 Romance languages 5,14,20,31,45-9,52,108, 127-8,167,209,219,235,238 creole-like evolution 30-1,45-9,102,127,167, 212,218-20,234-8

Romanization 218,219,234,236 rules 3,117 see also communal norms; pattern emergence rules of engagement 119,224 Russian 251,262,293n.23 Sao Tome 48,75,143 Sao Tomense 44,187,291n. 12 salience 129,151,158-9 Sami 109,206 SanBasilio 143 San language 128 Sango 34,56,258 Sanskrit 99,212,237 Santo Domingo 38 Sapir-Whorf hypothesis 245, 307n. 6 Saramaccan 36,136,137,145,187-8,192 Scandinavia 49,70,215,262 schooling 56-7,196,199, 230,231,235,241, 245, 259,264-5, 277-8 scientific racism 97,101 see also social Darwinism Scottish Gaelic 306n. 3 seasoned slave 46-7, 52, 55,101,172,185,198,214 second-generation colonies 48,292n. 15 segregated 39,215 segregation divergence, factor in 39-40, 55-6,77,104,105, 118,164,276-7 and language vitality 55-6,111-12,197, 215, 229,256,278-80,282, 300n. 17 selection individual selection 82,120,127,173 group selection 66,120,127,173,183 see also competition and selection semantic transparency 19,150,151,153 semi-creoles 38,40-2,43-4 see also AAVE semi-creolization 42-3 semi-speakers 306n. 3 Senegal 56,187,212,256,258 serial verb construction 45,80,84,135,139,176 servants, indentured 21,33,52,84,96,100-1,109, 215,299n. 8 see also slaves settlement colonies 51-2,56-7,97-8,216-17,220, 228,229-34 see also exploitation colonies; plantation settlement colonies Sicily 100 sign languages 78-9,88,90 simplicity 146,150 Singapore 222,293n.24

Subject Index singulative 130 siSwati 268 SLA (second language acquisition) 35,104-5,125, 133,140,149-59,173,182-93 slave trade 6,101,142,143,187,189 slavery 6,193,196,222,275 slaves 52, 55, 82,96,100-1,104,109,137,140,143, 145,168,193,214-15 social advantage (speakers' selection aimed towards) 48, 55, 67, 77,125,194-5, 223, 224, 242,243,253-4,261-2 social clubs 279 social Darwinism 97, 111 see also social racism social ecologies 23,27,115,141 social isolation 43,266,278,293n. 23 socioeconomic structure 232,242,256-7,258,262 see also population structure Songhay 258 South Africa 54-5,195-6,265-9, 310n. 11 South African Indian English 196,266 South African Indians 196-7,266 South Carolina 43,100,276,278-9,282-3,290n. 32, 295n. 14 Southeast Asia 194,221 Southern Italy 100,231 Southern White vernacular 67 Spanglish 292n. 19 Spanish 38-9,240, 3lOn. 13 Spanish, as world language 240,251 see also Romance languages specialization 124,173-4,31 In. 16 speciation 1,5,14,16-17,24,29-30, 71-2,128, 205-6,207 species, languages as 11,14-15,18,21,23,25, 26-7, 63-73,119-20, 208-9 speech continuum 106 Sprachbund (language convergence) 13,96,109, 296n. 20 squatter camp 266,267,269,271 Sranan 135-43,163-4,167-8,170,187-4,192 Sri Lanka 193, 306n. 7 St.Kitts 43 St. Vincent 39 Stammbaum 12,14,30,109,207 see also cladograms statistical frequency of features 19,23,39,129, 143,151,157,158-9 see also critical mass stigmatization 21,119,185,197,213,231, 239, 241-2, 260,274, 275-6, 278-9,280

stigmatize 230-1,239,277-8 stigmatizing 281 Straits of Malacca 194 stratification, language 259,263 structural erosion 58 structural/typological resemblance 104-5,106, 129-30,139,144,151,152 sub-Saharan Africa 31,57,99,130,197-8,228, 258-65 substrate/substratum hypothesis 133-59 substrate influences 14,16-17,42,47-8,86, 100-1,104-5,110,118-19,121,133-59, 174-6,195,206,230,275 Sudan 258 sugarcane 33,37-8,186 superstate influences 20,135-6,137-8,144,206, 288n. 18 superstratum 20,47,288n. 18 superstrate 06 superstrate hypothesis 37 Suriname 5,41,43,101,164,171-2,291n. 10, 292n. 15 Surinamese Creoles 36,135,187-8,192 Surinamese Creoles 135-6,187-8,192 see also Saramaccan Sranan Sutherland Gaelic 279 Swahili 56,177,181, 212, 216, 244, 255,258,263, 265, 303n. 10 symbolic value 242 syntactococentrism 296n. 7 systematicness 78 TaiBoi 229 Taiwan 236 Tanzania 216,238,263,265 Target Shift 140 teams, speakers not behaving as 92,181,183 TENSE-ASPECT 84,148,152,153,156,291n. 12, 303n. 10 Thai 12 Third World 57,149,206,225,232,242-3,248, 251,272,280 tobacco 143,186,276 Tobago 43 TokPisin 146,160,175 tonal languages 42,150,302n. 8 Torres Straits 193 TownBemba 56,212,244,255,258

353

354 Subject Index township 266,267,269,271 trade pidgins 35i, 55,74-8,102-3,171-2,188-9,229 trade colonization 76-8,103,228-9,270-1 trade factory trade factories 76 transcommunal 197 transfer 133-6,149-59 "transmission", language see also "acquisition" breaks in 36, 85-6,104,109,161,164,190 concept of 17-18, 86-7,110-11,182-4 naturalistic 57, 89,149 see also naturalistic language acquisition triglossia 233 Trinidad 43,145,195,199,215,231,266 Trinidadian Creole 145 Tshivenda 268 Tswana 239 Tupi 77,229 Tupinamba 229 Turkish 40,236 Tutsi 252 typological diversity 150,208 typological kinship 106,139,143,151 typological/structural resemblance 104-5,106, 129-30,139,144,151,152 unbounded phenomena 64 unconscious vs conscious speakers' choices 116,126 unidirectional 169 uniformitarian 45, 57,95,109,149,162 uniformitarianism 1,12 unilinear 43,70,162,169,178 United Kingdom 224,239,240, 244, 264 United Nations 241-2 United States see North America Universal Grammar (UG) 83-4,147,155,303n. 4 universals 131,166,174,178 Upper Guinea 143 urban urban culture 212 urban immersion 256 urban vernacular 56-7,188-9,194-5,198,212, 216,222-3,233-4,239,242,255,258-9,260, 271,289n.27 urban vernaculars 194-5,212,216,222-3, 233-4,244-5, 255,258-9, 260

urbanization 218,232,262,267,269 USA see North America utilitarian explanation 293n. 24 Valley Girl Talk 53-4 Vandals 303n. 5 variation 12,23,44,66, 83,90-1,92,97 inter-individual variation 89-90, 111, 117,140, 296n. 2 variational evolution 124-7 vernacular 1,3,5,6,33,45,52-3, 55, 82,101,125, 172,184,185,192-4,198,214,222,233, 242-3,255,259-61,271 vernacularization 56-7, 76,188 viral species, languages analogous to 11,18,21,23, 25,26-7,69,208,285n. 1 Virgin Islands 306n. 7 Virginia 186,276,299n. 8,304n. 7 vitality, language 205-84 wars, language 223,226 weighting (see also ranking) 115-16,155-6,176-8 West Africa 12,109,122,141,142,146,161,176, 188,305n.3 West African Pidgin English 305n. 3 West Atlantic 150 West Germanic 42,52,72 West Indies 100 Westernization 207 whaling 189,296n. 5,302n. 7 white man's burden 98 Wolof 56,212,244,256,258,269 world languages 45,239-41,251,264 world order 55, 215-17, 230, 232, 247-9, 259, 267, 270,289n.27,310n.8 writing system 213,260-1,282 xenolectal 21, 34,40,47,55,82,118-19,123,125, 183,184-6,199,289n. 28,290n. 7,292n. 19 Xhosa 302n. 10 Xung 302n. 10 Yaaku 238 Yeyi 239 Yiddish 53,208 Yoruba 145,195,215,231 Zambia 56,212,255,258