Creole and dialect continua: standard acquisition processes in Belize and China

CREOLE AND DIALECT CONTINUA

CREOLE LANGUAGE LIBRARY (CLL) A companion series to the "JOURNAL OF PIDGIN & CREOLE LANGUAGES"

Editors Pieter Muysken (Amsterdam) John Victor Singler (New York)

Editorial Advisory Board Mervyn Alleyne (Kingston, Jamaica) Norbert Boretzky (Bochum) Lawrence Carrington (Trinidad) Chris Corne (Auckland) Glenn Gilbert (Carbondale, Illinois) John Holm (New York)

George Huttar (Dallas) Salikoko Mufwene (Chicago) Peter Mühlhäusler (Adelaide) Pieter Seuren (Nijmegen) Norval Smith (Amsterdam)

Volume 18

Geneviève Escure Creole and Dialect Continua Standard acquisition processes in Belize and China (PRC)

CREOLE AND DIALECT CONTINUA STANDARD ACQUISITION PROCESSES IN BELIZE AND CHINA (PRC)

GENEVIEVE ESCURE University of Minnesota

JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences — Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data Escure, Geneviève. Creole and dialect continua : standard acquisition processes in Belize and China (PRC) / Geneviève Escure. p. cm. - (Creole language library, ISSN 0920-9026 ; v. 18) Includes bibliographical references (p. ) and index. 1. Creole dialects. 2. Dialectology. 3. Language acquisition. 4. Sociolinguistics. 5. Gram­ mar, Comparative and general-Syntax. 6. Pragmatics. 7. English language-Social aspectsBelize. 8. Chinese language-Social aspects. I. Title. II. Series. PM7831.E74 1997 417'.22-dc21 97-4278 ISBN 90 272 5240 8 (Eur.) / 1-55619-173-1 (US) (alk. paper) CIP © Copyright 1997 - John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. • P.O.Box 75577 • 1070 AN Amsterdam • The Netherlands John Benjamins North America • P.O.Box 27519 • Philadelphia PA 19118-0519 • USA

Acknowledgments This book is first a tribute to the people of Belize, and more particularly the Creoles of Placencia and the Garifuna of Seine Bight, who shared with me their joie de vivre, and their love of language. There is no space here to thank all the men and women who opened their doors and inner lives to make me feel wel­ come. I especially want to acknowledge the warm friendship and support pro­ vided in Placencia by the Lopez family (Miss Cordelia, Clarence, Therese and Errol), Miss Tila (Athelo) Cabrai, Herbert, Miss Doris, Miss Lucille, Miss Sonia, Miss Tensy, Miss Pearl, Miss Lizzie, Blink, Dudu, Bobo, Captain Morgan, Peter, Karim, Philip, and all the pikni (many of them are now adults) especially Verna, Karen, Dennis, Dean, Edward, Chrissie,and Grayson. Special thanks to Errol who assisted me with the fieldwork in Placencia. This is also a tribute to the Seine Bight Garifuna. It was thanks to Mr. Flores, who has now passed away, that I discovered Seine Bight and, later, Placencia. We became friends during the long bus ride from Punta-Gorda, and he invited me to his distant home across the lagoon from Mango Creek. Had he not extended such warm hospitality, I would probably have bypassed the Placencia peninsula, then totally isolated from the main road from Punta-Gorda to Belmopan and Belize-City. My old friend Roman Zuniga also passed away. He was my Garifuna teacher, and he shared with me his humor and his pride in his people's, traditions and language. Little Mark, who at fourteen spoke Garifuna, Creole, Maya, and Kekchi was my most reliable guide in the bush of Toledo District. I cannot possibly mention the names of all the individuals who befriended me, showed me how to dive for conch and lobster, gut a fish, bake cassava bread, or invited me for fish and fufu. They were Creole, Garifuna, Maya, Kekchi, or Mestizo, in Orange Walk, Dangriga, Hopkins, Georgetown, Punta-Gorda, Barranco, San Antonio, Sartineha, Benque Viejo, Caye Caulker, and Belize-City. They instilled in me a profound love of Belize and its diverse people. In China too, I met warm and friendly people in the rural areas as well as

vi

CREOLE AND DIALECT CONTINUA

in the crowded cities. I gratefully acknowledge the cooperation and open-mindedness of the officials and professors at Huazhong Central University of Science and Technology in Wuhan, especially Cheng En Hong, Lei Guo Pu, and Mu Li Ya, who invited me to teach a stimulating group of teachers for an entire acad­ emic year, and at subsequent times, and gave me the opportunity to travel freely and get to know China better than if I had been a mere tourist. I owe a lot to my students, many of whom became my friends. They taught me a great deal about the sociolinguistic situation in China. Particular thanks to Chen Kairong, Cheng Kola, Deng Shan, Guan Xiangyang, Liu Lichun, Peng Yuanchu, Tan Xiangyun, Wan Youzhong, Wang Qunying, Zou Hong, and final­ ly to Zhang Xianping who gave me the opportunity to give lectures in his dis­ tant Hubei hometown of Xianfan. I fondly remember the enthusiastic welcome I received at Xianfan University. I also benefitted from a summer teaching at Tsinghua University in Beijing. Thanks to Lu Ci and Wu Guohua for their hos­ pitality, and mostly to Chun-Jo Liu for her initiative in organizing the summer program. I gained a deep appreciation of the people and linguistic diversity of China mostly through chance encounters. Strangers went out of their way to help me when I got desperately lost in Chengdu—unable to get the right bus because of my difficulties with tonal differences; they bought tickets for me in the crowd­ ed train stations of Wuhan, Shanghai or Kunming; and they offered me tea on dusty Tibetan roadsides. Those strangers were kind to me; they saw me as a per­ son, not as an alien (the distrusted waiguoren), and they contributed to human­ izing the anonymous crowds of the Middle Kingdom. During long train rides across the People's Republic, I pored over my Chinese notebooks, trying to figure out a language survival kit that would enable me to achieve basic communication in diverse linguistic areas. A novice learner relating new knowledge to old knowledge, and looking for shortcuts, it occurred to me that Chinese syntax is very much like Creole syntax. Thus ger­ minated the idea of a comparison of Chinese and Creole structures. However, bringing the project to completion was as lengthy, hazy, and rocky as a train ride in a smoky "hard seat" compartment. None of this work would have been possible without the steady support of the University of Minnesota Graduate School, that provided leaves, research assistants, and limited but steady financial support. I also acknowledge occa­ sional contributions from the English Department. I am particularly grateful to J. Lawrence (Larry) Mitchell, for his encouragements to pursue my interest in nonstandard varieties, as well as to Sandy MacLeish for his positive attitude and moral support. All my thanks also to the various students and research assistants who discussed issues with me, and helped me transcribe data: Gail Ostrow, Dennis Dougherty, Chao Shoushing, Mao Luming, Ding Ersu, Guan

vii

Xiangyang, Angela Karstadt, and most of all Anna Fellegy, who provided invaluable scholarly comments and editing help in the final stages of the mansucript. Ï hold a deep appreciation of Pieter Muysken's encouragements and patience. Chris Corne has my special thanks because he gave me the confi­ dence to push onwards by providing helpful critical comments on the first part of the manuscript. I also acknowledge the help of Chun-Jo Liu in reading and commenting on parts of the Chinese sections. In the last stages, the detailed contributions of the reviewers, Stephen Matthews and Virginia Yip were extremely helpful in sharpening issues in the book and relating my findings to current developments in the field of Chinese sociolinguistics. Finally, many friends and colleagues helped in indirect ways, in particular as contributors to the meetings of the Society for Caribbean Linguistics, the Society for Pidgin and Creole Languages, and the Institut d'Etudes Créoles.

Contents Acknowledgments

v

I.

Introduction: Developmental continua 1. General issues in linguistic change 2. Aspects of acquisitional studies 3. The linguistic analysis of nonstandard dialects 4. Creoles and noncreoles

1 1 3 10 18

II.

The Belizean speech community and the use of English 1. Introduction to the history of Belize 2. Ethnolinguistic composition and census data 3. The language situation 4. Focus on the Stann Creek district 5. Conclusion

25 25 28 36 43 55

III. Creole acroiects as innovations 1. Radical creoles, postcreoles, and decreolization 2. Acroiects and standard dialects: Social aspects 3. Acroiects and standards: Linguistic aspects 4. Conclusion

57 57 65 73 86

IV. The interaction of syntax and pragmatics in acroiects: Topic marking 89 1. Syntax and discourse features 89 2. Types of topic strategies 94 3. Distribution of topic strategies 103 4. Comparison with American English 117 5. Pragmatic aspects of syntax 120 6. Conclusion 122 V.

Sociolinguistic perspectives on Chinese 1. Chinese 2. Aspects of the linguistic history of China 3. Main subgroups of "Chinese" 4. The reform of Mandarin: Guanhua, Guoyu, Putonghua

125 125 130 135 139

CREOLE AND DIALECT CONTINUA 5. Varieties of Mandarin outside Beijing: Wuhan 6. Wu varieties and Suzhou Wu 7. Conclusions

ix

144 146 148

VI. Topic mechanisms in Chinese: An overview 1. Introduction 2. Topic fronting 3. Topic repetition 4. Topic presentation 5. Summary of topic strategies

151 151 156 163 166 182

VII. Literary and colloquial Putonghua 1. The Putonghua corpus: Methodology 2. Literature and the Cultural Revolution: Two New Friends 3. Literature after the Cultural Revolution: The Matchmaker 4. Beijing colloquial Putonghua 5. Overview of topic strategies in Beijing Putonghua

185 185 188 202 209 219

VIII.Topic strategies in varieties of Putonghua as second dialects 1. Introduction 2. Putonghua spoken in Wuhan 3. Comparison of Beijing and Wuhan Putonghua 4. Wu Chinese 5. Comparison of native and non-native varieties of Chinese 6. Conclusions: Toward a universal view of topic processes

223 223 224 239 243 258 260

IX. Conclusions: Pragmatic universals in second dialect acquisition 1. Patterns of topic marking: Summary 2. Topic marking in the Belizean continuum 3. Patterns of topic marking in Chinese 4. Dialect versus standard: Sociolinguistic universals 5. Conclusions 6. Postscript: Directions for education

263 263 265 271 273 285 286

Bibliography

289

Index

303

Chapter 1 Introduction: Developmental Continua Abstract The issue of dialect acquisition is compared to other types of linguistic developmental continua, including second language acquisition. The impact of social attitudes on language development is examined in various communities, as such attitudes may affect the acquisition of standard varieties by speakers of nonstandard varieties. Methods and models for the analysis of language in spontaneous discourse are dis­ cussed, and the linguistic units selected are illustrated. A combination of sociolinguistic methodology and functional grammar appears to be the best suited to this study. Finally, I present a justification of the proposed comparative study of a creole context (Belize) and a noncreole context (China). 1. General issues in linguistic change Language is an essential element of practically every human activity, yet its flexible and adaptive nature in social communication is largely ignored, and even denied, by its users. Despite the advances made in linguistics over the last fifty years, most speech communities cling to general normative attitudes toward language, believing that it is (or at least should be) static, monolithic, homogeneous, and primarily preserved, perhaps further refined, by scholarly experts, academicians, writers, and educators. While linguistic change is rec­ ognized, it is deplored as contributing to the deterioration of "pure" linguistic norms. Pop grammarians contribute to this prevailing view, claiming to correct the process of disintegration by denouncing the solecisms or barbarisms com­ monly attributed to the poorly educated and the younger generations. In fact, the popular view that blames change on specific social factors and age groups effectively signals a general awareness that linguistic change is highly depen­ dent on such factors. Correlating social and linguistic factors for explanations of language

2

CREOLE AND DIALECT CONTINUA

change is a relatively recent target in linguistics. The search for factors that motivate linguistic change has been applied to many areas of language study. Language change is most evident in such areas as historical linguistics, native language acquisition, second language acquisition, pidginization, and creolization which are all characterized by developmental continua—dynamic levels or stages which are not separated from each other by clear boundaries. In contrast to the early structuralist view that synchronic and diachronic processes were not comparable, there has been increasing awareness of the potential structural sim­ ilarities existing between different types of developmental continua. Such lin­ guists as Meillet (1921) and Martinet (1955) already emphasized the regularity of change, commonly referred to by structural linguists as the Uniformitarian Principle. According to that view, the same mechanisms which operated in the past can be observed in contemporary variation,' mostly because integrated rule systems are considered to be stable and to change or borrow only in the direc­ tion of "linguistic drift" or native tendencies. Sociolinguistic methodologies seem to reflect and endorse this approach in their use of apparent time (gener­ ational studies) and real time (longitudinal studies) to document linguistic change (Labov 1972a; 1972b). Recently, attempts have been made (Andersen 1983; Muysken & Smith 1986) to compare second language acquisition (SLA) to pidginization, creolization (Valdman 1983), and decreolization (Rickford 1983; Schumann 1978; Stauble 1978; Andersen 1983); first language acquisition, as well, has been compared to creolization/pidginization (Bickerton 1981). Nevertheless, there is still no clear consensus as to the nature of the structural similarities linking those developmental continua and the reasons motivating the use of operating principles in language development. According to Bickerton (1981:238) . . .no real connection exists between SLA and creolization: they differ in almost every particular. SLA is done alone, creolization is done in groups; SLA has a tar­ get, creolization hasn't; SLA is done mainly by adults, creolization mainly by kids; SLA gives you a second language, creolization gives you a first There is, however, general agreement that the adoption of formal princi­ ples does not operate in a vacuum but that it is linked to practical aspects of the learning situation, such as available linguistic input, psychosocial motivation, and historical factors. For some linguists, "the history of a language is a func­ tion of the history of its speakers" (Thomason & Kaufman 1991:4). Proponents of the imperfect second language learning hypothesis as applied to pidgins and creoles say that they display features of "interlanguage systems," such as invari­ ant verb forms, a lack of determiners, the use of demonstratives as determiners, the invariable placement of the negator in preverbal position, the use of adverbs to express modality, a fixed single word order, a lack of inversion in questions,

INTRODUCTION: DEVELOPMENTAL CONTINUA

3

and a reduced or absent nominal marking (den Besten, Muysken, & Smith 1995:97-98). In the search for a better understanding of the principles instrumental in language development, I will focus on two types of linguistic situations which can be characterized as subtypes of second dialect acquisition. Specifically, I will examine the acquisition of standard varieties of a language by speakers of the nonstandard varieties of that same language, with special attention to 1) a creole continuum (Caribbean) and 2) a non-creole situation involving extensive dialect variation (Chinese). The advantages of comparing similar phenomena in obviously distinct contexts are multifold and can be summarized as presenting the following possibilities: 1. 2.

3.

4.

the formulation of cross-linguistic generalizations in studies of acquisition, particularly as they apply to oral discourse; an inquiry into general social attitudes toward marginalized groups, and how this behavior affects the development and use of language varieties produced by those groups; the observation of the interaction of three primary language components— semantics, syntax, and pragmatics—in spontaneous oral communication; and an evaluation of the putative uniqueness of creole languages.

2. Aspects of acquisitional studies Although a substantial amount of linguistic research has been devoted to code switching phenomena between languages, and to second language acquisition (SLA) primarily by adults, relatively little attention has been given to the mech­ anisms underlying second dialect acquisition (SDA); yet, this type of acquisi­ tion is undoubtedly more widespread for the simple reason that dialect variation is universal. Whereas not all individuals find themselves in social situations which require the acquisition of a second language, there is no single human being whose repertoire is limited to only one language variety, style, or dialect. Furthermore, unlike second languages, second dialects are typically acquired earlier in life and continue to be acquired throughout adult life.1 Second dialect (usually standard) acquisition is often perceived to be an ambiguous, undefined linguistic process. In West Indian contexts, such learning situations have been characterized as those in which the standard to be learned is neither a native lan­ guage nor a foreign language (Craig 1971:376; Stewart 1964). SLA research has primarily focused on the analysis of acquisitional steps, particularly on errors occurring in the course of that process. Various mecha­ nisms have been hypothesized to account for "imperfect learning" and the mis-

4

CREOLE AND DIALECT CONTINUA

interpretation of linguistic input, which results in the learner's unstable interlanguage continuum; the mechanisms include language transfer due to native language interference, overgeneralization, simplification, and the operation of universal strategies put into motion when linguistic input is deficient (McLaughlin 1987). Errors are usually traced to conflicts between the phono­ logical or morphosyntactic systems of the two codes in contact (the native ver­ sus the target model). The learnability of certain features is accounted for by markedness relations, a theory asserting that a language is easier to learn if it contains more unmarked features than the learner's native language. Jakobson (1940) first introduced this mode of interpretation which points to correlations between the development of phonological features in first language acquisition, aphasic loss, and the patterning of phonological inventories across languages (see also Ellis 1994). Recent attempts to identify types of contact-induced linguistic change distinguish between two main mechanisms of interference, obviously relat­ ed to acquisitional patterns: "interference that results from imperfect group learning during a process of language shift, and interference that results from borrowing, by native speakers of a language, of features of some other language with whose speak­ ers they are in contact" (Thomason & Kaufman 1991:212). A more focused group of studies have investigated non-native varieties of English (the New Englishes) examining the relationship between the acquisition of such varieties and claims made by theories of language acquisition and change, such as error analysis, mother tongue interference, and markedness theory (Williams 1987). For example, Platt (1991:376) considers English in Singapore as "a kind of fossilized interlanguage [which] became a lingua franca in the English-medium schools among students whose home language might be one of the Chinese dialects, an Indian language or Malay." He shows that many features of Colloquial Singapore English (articles, copula use, tense and number marking) suggest Chinese influence (Chinese constitutes 76% of the population of Singapore), especially from Hokkien, the dominant Chinese dialect. However, a recent investigation of reported language use and identity in Singapore docu­ ments the effects of the 1987 compulsory bilingual education policy, which made English the medium of education and required the study of a second "eth­ nic" language (Mandarin, Malay, or Tamil). At the same time, the Singaporean government discouraged the use of Chinese "dialects," such as Hokkien, enforcing a "Speak Mandarin" policy. Results indicate that English is increas­ ingly used and that, among the majority Chinese group, there is a generational shift in reported identity from Chinese (especially specific Chinese group iden­ tity such as Hokkien, Cantonese, or Hakka) to Singaporean (Hvitfeldt & Poedjosoedarmo 1995). One may wonder to what extent these attitudes are affecting Singlish, and more specifically, whether Chinese substratal compo-

INTRODUCTION: DEVELOPMENTAL CONTINUA

5

nents are yielding to other features. Another New English acquired as a second language, Kenyan English is displaying mother tongue interference which reflects the learners' ethnic group. The Kikuyu show evidence of a liquid (1/r) phonological merger in their use of Kenyan English, whereas the Luo use affricates but no alveolar fricatives in their version of English. Both features echo native African features of Kikuyu or Luo (Schmied 1991:429). A similar distinctiveness linked to native language is found in Liberian English, a New English which happens to be integrated into a recognized creole continuum and similar to situations existing in other parts of Africa and in the Caribbean. Singler (1991:558), in his investigation of plur­ al marking in Liberian English, indicates that plural marking in Kru Pidgin English, whose speakers have Kru as a first language, is neither identical to plural marking in Liberian Interior English, whose speakers have Mande as a first language, nor to plural marking in Settler English, spoken by the descen­ dants of Black American immigrants to Liberia in the nineteenth century. The traces left by the native or ancestral language in the newly developed, typically contact-induced, variety are commonly referred to as substratai influ­ ences, which reflect the lower status of the speakers of the new language. On the other hand, any influence of the linguistic component traceable to the social­ ly dominant group is called superstrata!. Theories of genesis will be briefly dis­ cussed in chapter two, and a comprehensive presentation of existing approaches appears in Alleyne (1980), Arends et al. (1995:9-10; 99-109), Holm (1988), and Muysken & Smith (1986). Finally, pedagogically-driven research has focused on the identification of strata, or stages in the West Indian learning situation (and more particularly in Jamaica and Trinidad), and the implications for English-teaching methods (Bryan 1996; Christie 1983; Craig 1966; 1980; 1996; Pollard 1996; Roberts 1983; 1988; Shields-Brodber 1989). The important issue of the practical applications of lin­ guistic research to social change will be discussed in chapter nine. 2.1 Universalist explanations of acquisition One of the major explanatory approaches to issues of acquisition involves uni­ versalist explanations of the human language learning capacity. Chomsky's Universal Grammar (UG) provides innateness as an explanation for cross-lin­ guistic similarities. This theory postulates that all humans are biologically equipped with a language faculty that permits the generation of general lin­ guistic principles. The language faculty is available at birth and becomes acti­ vated when input is insufficient ¿o provide a full model of the target grammar, which is typically the case in first language acquisition. Grammar development consists of the setting of parameters for the specific language being learned. For

6

CREOLE AND DIALECT CONTINUA

UG-proponents, the study of a child's language development is a sufficient basis for the research and testing of language universals, and external factors, such as communicative needs and social interaction requirements, are consid­ ered irrelevant. The "Bioprogram Hypothesis," a version of UG introduced by Bickerton (1981), incorporates special reference to the emergence of language universals in pidgin and creole languages, a case of first language acquisition. The latter develop rapidly in contact situations in which speakers evolve a new system without the benefit of an already existing model, thanks to the resources of their innate linguistic ability. Bickerton's genetic view allows for some social input, since demographic, historical, and sociopolitical elements are necessarily involved in the genesis of pidgins, but he fails to include systematic reference to daily interactive needs. Assuming that Chomsky and Bickerton are right about the existence of a language faculty in the brain, there may be other types of universals dependent on external factors, in addition to the commonalities which depend on the inter­ nal properties of the language faculty (Butterworth, Comrie, & Dahl 1984). Internal and external explanations can be distinguished as follows: Cognitive and psychological explanations involve formal operations that the human mind can vs. cannot accommodate. . .while pragmatic or sociolinguistic explanations involve (formal?) operations that a human society or individual within a society can vs. cannot accommodate. . . (Hyman 1984:68). Muysken & Veenstra (1995:121-134) summarize universalist models applied to pidgins and creoles, identifying two types of universals: procedural universals (referring to universal properties of processes such as L2 learning, grammaticalization) and constitutive universals which designate the properties resulting from those processes (e.g.,TMA systems or word order). These prin­ ciples and their applications in acrolects will be further discussed in subsequent chapters. In view of the previous attempts at understanding language acquisition and change in various contexts, the question is open as to whether linguistic com­ monalities, internal or external, also emerge in the acquisition of second lan­ guages, such as pidgins and creoles, and second dialects, such as acrolects or standard varieties acquired by native speakers of nonstandard varieties. Clearly, the issue must be differentiated from first language acquisition, as presented by Chomsky and Bickerton, since learners of a second dialect or language already possess a native linguistic system which interacts with the target system (either causing errors, transfers, interferences, markedness patterns, or substratai and superstratal influences, according to existing theories). If there are contactinduced linguistic universals, how can they be differentiated from language-

INTRODUCTION: DEVELOPMENTAL CONTINUA

7

specific influences? In which language components are they more likely to be identified? These questions and their applications in acrolects will be further discussed in subsequent chapters. 2.2 The acquisition of second dialects The interpretations outlined above could conceivably be applied to other types of acquisitional situations of the "dialectal" type, especially in cases of second dialect learning—cases when there is no distinct mother tongue present in the competence of the standard learner. The general lack of interest in second dialect acquisition may be due to the difficulty involved in identifying second dialect situations as linguis­ tically distinct from first dialect contexts. In spite of the paucity of research on dialect acquisition, or perhaps because of it, there is a popular consensus that stan­ dard dialect acquisition is less traumatic than SLA. This attitude is obviously reflect­ ed in the educational establishment: a standard variety is never taught in a formal, organized manner and, in contrast to foreign language learning, is not supported by any language manual or the "bidialectal" equivalent of bilingual programs. The learner is assumed to know the standard before going to school. We can infer from such attitudes that there is a deeply entrenched conviction that the acquisition of dialects involves less distance between the native and the target varieties (D1 and D2) than is the case between L1 and L2. Of course, the nonstandard speaker may have previously acquired a passive or partial competence of the standard model, depending on the degree of exposure and motivation involved in a specific situa­ tion. Public opinion as well as educational systems assume that a speaker of Geordie, a variety of Newcastle English (Graham 1980), will find it easier to learn RP than to learn French, or that a speaker of American Black English should learn standard English faster than Spanglish, and Spanglish faster than Spanish. On the other hand, there is apparently no clear opinion concerning the relative learnability of nonstandard dialects by standard speakers. Would it be easier for a speaker of standard English to learn Black English2 or Geordie than Spanish? It is unusual, however, for speakers of prestigious dialects to strive to learn stigmatized forms, since there are neither the economic nor educational motivations to do so. Although the possibility is not excluded, SDA is typically a one-way social process. There are no empirical studies of dialect versus language acquisition that prove or disprove the validity of the above intuitive statements. Although some nonstandard varieties such as Black English have become the object of schol­ arly study (Labov 1972a), the actual process of the acquisition of standard American English by speakers of Black English still remains to be explored. The spontaneous acquisition of a standard code as second variety—whether we call it dialect or language—is highly constrained by social factors, subjective attitudes, and the psychosocial context of learning. It is, therefore, essential to

8

CREOLE AND DIALECT CONTINUA

distinguish between guided learning, the artificial classroom acquisition of a second language, and unguided learning which includes the acquisition of a standard dialect or language in a real life context. Unguided learning situations create the ideal circumstances for the study of the interaction between syntax and pragmatics, as will be discussed below. 2.3 Social attitudes and language/dialect acquisition An important component of language development lies in the existence of social attitudes toward languages and their speakers; yet, SLA studies general­ ly do not look at the potential significance of the relative social status of the tar­ get language and of the learner's native language. They also do not take into account the related but distinct issue of the social status of language learners and of native speakers of the linguistic target. For example, there is obviously a different social dynamics involved when comparing the acquisition of English by a Mexican migrant worker in Minnesota, a Mexican migrant worker in California, a Chinese graduate student in an American university, and a Chinese resident of Singapore. There is more of a group support system for Mexicans, or Chícanos, in California than in Minnesota, which has predominantly German and Scandinavian ethnicity and only a seasonal influx of Chícanos. Whether or not a large Mexican population in California, and the universal presence of Spanish or Spanglish in the streets and on the airwaves, facilitates or impedes the acquisition of English by Mexicans remains to be studied. In the other case, a Chinese graduate student on an American campus has a focused, academic motivation to learn English, and the process of SLA acquisition does not inter­ fere with his or her sense of Chinese identity, unless of course there are personal motives for rejecting it to embrace an American identity. For the Chinese in Singapore, the local English variety is a lingua franca which now symbolizes their identity as Singaporeans, as seen above. Since Mandarin Chinese is offi­ cially supported by the Singapore government through the "Speak Mandarin" policy, there is good motivation to preserve a Chinese ethnic identity. Social factors are essential to examine because they determine motivation, the availability of learning tools, and other factors instrumental in the develop­ mental process of language acquisition. As pointed out by LePage & TabouretKeller (1985:247), "neither 'race' nor 'ethnic group' nor 'language' turns out to be a clearly-definable external object," and linguistic choices are "projections" of identity. The learning of non-native varieties, whether they are better defin­ able as dialects or languages, is at least partially triggered by psychosocial and economic motives. Language is perceived as a marketable asset, a tool for upward mobility. An individual expects to reap benefits, such as status, "class," distinction, recognition and related economic bonuses from the acquisition of a

INTRÓDUCTION: DEVELOPMENTAL CONTINUA

9

new linguistic code. Thus, subjective attitudes toward languages, accents, dialects, and other perceived linguistic entities are directly derived from social factors, and the speakers' histories (Thomason & Kaufman 1199:4). 2.4 Linguistic bias Although the popular (not the linguistic) view defines a dialect as deviant from the norm, the difference between a language and a dialect may be impossible to determine unambiguously, either linguistically or socially. Bias plays an essential role in our perceptions of language varieties and in their acquisition. Linguistic bias reflects the social stratification present in every community, creating a cir­ cular pattern of cause and effect: language is a social mirror and, thus, perpetu­ ates social and ethnic bias because linguistic behavior shapes attitudes and opin­ ions. The lack of power of a stigmatized group is compounded by the negative values associated with the group's linguistic forms (typically described as incorrect, inadequate, inappropriate, uneducated, illogical, and politically dan­ gerous). A nonstandard variety is often the subject of denial by its own speak­ ers and by the society harboring it (Ferguson 1959). If at all recognized, it is assumed to be an immature linguistic habit which will be shed in the course of normal intellectual development. If such "bad" linguistic habits persist, users find it harder to establish their social credibility; they may even be denied access to the powerful world of standard speakers. Such attitudes often apply to foreign accents as well and, therefore, to the SLA process. Native speakers often associate—perhaps unconsciously—unflattering intellectual connotations to non-native renderings of their language. A related attitude is the feeling of embarrassment all language learners experience in the first stages of conscious adult acquisition. Clearly, we are strongly motivated to assess an individual's value as a social human being on the basis of that person's verbal ability and conformity to conventional local linguistic standards in addition to other super­ ficial factors, such as physical appearance and behavior. A prime example of language-based bias is reflected in the recent cam­ paign to identify English as the sole official language of the United States. The English Only movement has as its goal to deny official status to Spanish, as well as to other immigrant languages, a position which may partly be derived from the common misconception that language unity promotes political unity and economic development.3 The threatening power of linguistic and ethnic diver­ sity has long been reflected in the stigma attached to Black English varieties in the United States, as well as to creoles in the Caribbean and elsewhere. As reflected in the case of Spanish in the United States, varieties which are commonly referred to as "languages," and that have official and prestigious sta­ tus in a given country, may hold low status elsewhere. Castilian Spanish, for

10

CREOLE AND DIALECT CONTINUA

example, is the official language in Spain and constitutes the prestige form in most of Spain except in Catalonia, the Northeastern province, where it has, for political and historical reasons, a much lower status than the local language Catalan. The use of Catalan was made illegal during General Franco's twentyyear dictatorship. In contrast, Catalan was mostly tolerated in Catalogne, the southern province of France adjoining Spain, and it has fallen out of use there despite the fact that it is actively promoted by cultural organizations. Thus, lan­ guage differences sometimes disappear in cases of voluntary acculturation but are reinforced in cases of oppression. In Latin America, Spanish has a much higher official status than Indian languages (e.g., Quechua, Aymara) whose speakers are at the bottom of the social scale. In the United States, any variety of Latin American or Caribbean Spanish has low overt social status (which does not exclude high covert prestige) due to widespread negative social attitudes toward minorities, migrant workers, and recent immigrants. The circular pattern inherent in linguistic bias is particularly damaging to speakers of nonstandard or unofficial varieties who need to find creative ways to break the pattern of behavior and judgment. For example, the revalorization of Black English and African-American values through the medium of rap music illustrates the attempt to break the pattern of bias through a linguistic medium. Rap clearly has gone well beyond its original linguistic medium since it has been actively adopted in France with musical lyrics in standard Parisian French by Arab and African "rappeur" groups, as well as in Hong Kong as Canto-rap—rap with Cantonese lyrics by analogy with Canto-pop.4 Studies using the matched guise technique have documented the claim that social judgments are based on linguistic behavior, especially on the vague concept of accent (Giles, Bourhis & Taylor 1977). In England, Standard English (RP) is linked to traditional norms and the "public school" education reserved for the elite. A corpus of listeners found that RP evokes impressions of success, intelligence, and elegance but a certain lack of warmth, especially when used by women. Northern English varieties are associated with warmer but less educated and less successful individuals, which confirms that an official norm derives its prestige from the power of the speech community using it (Elyan et al. 1978). Such normative attitudes pro­ moting the supremacy of standard dialects or official languages are universally shared, and elementary and secondary education teachers are often encouraged to convey this view to their students. The two case studies to be analyzed in this book will illuminate such attitudes. 3. The linguistic analysis of nonstandard dialects Paradoxically, some of the most striking and productive developments in the

INTRODUCTION: DEVELOPMENTAL CONTINUA

11

field of formal linguistics have been concomitant with a denial of the existence and importance of nonstandard varieties, which indeed runs parallel to public ideology. Transformational Generative Grammar (both in its original form and in its later version as Government and Binding), because of the priority it assigns to the search for an abstract linguistic competence and the identification of the universal linguistic capacity shared by all human beings, has intentionally ignored all references to linguistic variability and to the extralinguistic fac­ tors—social, cultural, psychological and political—that determine linguistic variability. For formal linguistics, the database consists of the linguists' intu­ itions. Apparently, those intuitions have not included nonstandard verbal struc­ tures; thus, omissions of large segments of speakers' repertoires are likely to have contributed to the rejection of nonstandard dialects as a valid field of study and a valuable component of education, and may even have reinforced the non­ standard bias and the low status assigned to their speakers. Schiffrin (1987b:392) claims that one of the major differences between formal linguistics and sociolinguistics is "a difference in data," adding that "the selection of a data type has a profound influence on the range of phenomena which a model aims to represent and a theory aims to explain." 3.1 Labovian sociolinguistics Until twenty years ago, there had been no large scale study of the dialects of groups with low social status. Labov pioneered the field with his seminal stud­ ies of the social stratification of English in New York City (Labov 1966) and of Black English (Labov 1972a). Although the study of nonstandard dialects has since greatly expanded in the direction of increased observational adequacy, there is still a glaring absence of data documenting the degree of deviance exist­ ing between standard dialects as acquired by nonstandard speakers and the ver­ nacular (native) forms of those standard dialects. Labovian sociolinguistics attempts to deal with speech variability per se, and its relevance to commu­ nicative competence rather than with the Chomskyan internalized competence of the ideal speaker/hearer. For the first time, questions are raised about the validity of our judgments of standard and nonstandard speech, and full scale scientific descriptions of nonstandard speech varieties are conducted. As issues of descriptive accuracy and observational adequacy are brought to the forefront, it becomes necessary to develop sound methodologies for the collection of lin­ guistic data. What are the observations that will reliably reveal the mechanisms opera­ tional in language development? How can we ascertain the difference between linguistic facts and the abstracted interpretation of those facts? Indeed, what theory will provide an explanation of the significant linguistic facts? The intu-

12

CREOLE AND DIALECT CONTINUA

itions of generativists and the mailed questionnaires of dialect geographers obviously tapped limited, and manifestly selective, segments of the linguistic options open to speakers and are now no longer considered appropriate to sat­ isfy the sociolinguists' interests in establishing the range and extent of linguis­ tic variability. Matched guise techniques and laboratory experiments are com­ monly used to isolate linguistic features or identify subjective judgments in con­ trolled situations and produce interesting results, as in the case of the evaluation of RP outlined above; however, they have the disadvantage of creating artificial situations which can only remotely match genuine language use and, thus, pri­ marily define formal situations or tap conscious aspects of linguistic behavior. One of Labov's methodological principles—the Principle of Formality— states appropriately that "any systematic observation defines a formal context in which more than the minimal attention is paid to speech" (Labov 1972c). Labov advocates methods whose goals are to capture an individual's full per­ forming repertoire, from formal to casual, through elicitation of word lists and text reading (for formal speech) as well as through observations of spontaneous conversations (for careful to casual speech). The latter can only be successful if the fieldworker is a member of the speech community under investigation or has become fully accepted as a participant in the social activities of that com­ munity. Participation, therefore, helps solve the "Observer's Paradox," as one observes "how people speak when they are not being observed" (Labov 1972c). It has been claimed that sociolinguistics is merely a methodology and has no theoretical import. This is clearly inaccurate since all socially-oriented stud­ ies of language variation derive their premises from the basic theoretical view­ point that the social context determines human behavior, language being just one aspect of human behavior. Variation studies emphasize the necessity of quantitative analysis as a research procedure with the hypothesis that variabili­ ty in performance is predictable in terms of the correlated social and linguistic features of each natural interaction. Such approaches are traceable to a Marxist view of social class based on conflict and power. The issue of nonstandard dialects stands within this perspective because the social evaluation of language is based on power, or the perception of power and conflict, and on the human desire to change the power structure to one's advantage. Guy (1988:41) claims that Western social theories present an alternative definition of class based on social unity and status as opposed to the Marxist notions of power and conflict. He says that sociolinguists such as Labov have primarily been functioning from this "soft" perspective, with resulting gradient linguistic stratification, rather than the sharp stratification which is expected to evolve from violent conflict; however, sharp stratification in linguistic feature distribution or subjective atti­ tudes toward language types has been found to occur in issues of minority lin-

INTRODUCTION: DEVELOPMENTAL CONTINUA

13

guistic behavior—whether in Black English, creoles, or in politically-motivated language choice (such as the case of Catalan mentioned above). Consequently, it seems pointless to separate or choose between unity-sta­ tus and conflict-power dichotomies. Both types of social behavior obviously cooccur and contribute to the fashioning and development of linguistic behavior. The approach adopted here assumes that a social dynamic is essential in non­ standard dialect learning but does not presume to predict the type of trigger involved in the acquisition process. It is, however, hypothesized that nonstan­ dard native speakers will only effectively learn the standard if they have strong incentives to do so—if they envision a practical positive outcome to the acqui­ sition of the linguistic code. If they assess the achievement of power as unreal­ istic, it can be predicted that they will not learn much of the standard. The popular definition of the term dialect—a nonstandard variety which is somewhat "deviant" from the local norm—is a useful heuristic sociolinguistic concept because it is based on social attitudes regarding language varieties and their speakers. The notion will naturally be essential in this investigation of the process of acquisition of standard varieties by speakers of dialects or nonstan­ dard varieties. 3.2 Language in oral discourse and pragmatics The data base underlying this investigation of second dialect acquisition pri­ marily consists of spontaneous conversations. It is the most common type of ordinary communication, however, which poses the most serious problems of analysis. The principles underlying oral discourse lie within the overlapping ranges of pragmatics and sociolinguistics, Yet, there is currently no compre­ hensive theoretical framework which adequately predicts the organization of discourse, "no theory of paragraphs and its parts which is nearly as elaborate as a theory of sentences" (Linde 1981:85). The field of pragmatics encompasses a broad range of perspectives on lan­ guage in context. Green (1989:2) places it at the intersection of a number of fields within and outside of cognitive science: not only linguistics, cognitive psychology, cultural anthropology, and philosophy (logic, semantics, action theory), but also sociology (interpersonal dynamics and social convention) and rhetoric contribute to its domain. The fuzzy boundaries that pragmatics shares in particular with sociolinguistics and semantics, as well as with other disciplines, make pragmatics hard to define clearly, although vague definitions are not lacking. It is defined broadly as "the study of understanding intentional human action," requiring reference to the central notions of belief, intention, plan and act (Green 1989:3), and narrowly

14

CREOLE AND DIALECT CONTINUA

as "the interpretation of indexical expressions" (Green 1989:2). Levinson (1983) provides an interesting evaluation of various definitions of pragmatics, from its philosophical foundation to a more empirical data-driven approach, implying the growing importance of observational adequacy. Some of those defi­ nitions include "meaning minus truth conditions," "the study of the relation between language context that are basic to an account of language understanding," "the study of the ability of language users to pair sentences with the contexts in which they are appropriate." Levinson concludes that "conceptual analyses using introspective data [are] replaced by careful inductive work based on observation" (Levinson 1983:285). The functional approach endorsed by Dik (1980) and Givón (1979a; 1990), among others, also regards pragmatics as the primary framework within which "syntactic and semantic principles are explained in terms of the pragmatic pur­ poses and requirements of verbal interaction" (Dik 1980:2). The functional view of natural language claims to consider language primarily as an instrument of social interaction but, paradoxically, does not discuss real life contexts and pri­ marily aims to produce principles on the basis of isolated sentences. The advan­ tage of this approach is its search for "typological adequacy"; as Givón (1990:vii) notes, "surface diversity of cross-language typological facts masks behind it a great measure of commonality of human languages. Part of that com­ monality is due to semantic and pragmatic universals." Other linguistic research on language universals recognizes the combination of both formal and function­ al explanations for the commonalities found in human languages (Comrie 1984). 3.3 Theoretical framework for oral discourse The question remains of what theoretical model would most adequately account for the type of speech data to be considered here: namely, spontaneous dis­ course in the context of standard acquisition by native speakers of nonstandard varieties. The concept of communicative competence and, furthermore, polylectal competence, must be central to this putative model, since I will have to account for the competence of speakers who shift easily between different language varieties depending on the social context. The principles and methods of quantitative sociolinguistics, as briefly outlined above, are well-suited to the detailed analysis of specific linguistic variables. One major aspect of sociolinguistic analysis is empirical, as it involves close attention to social context and to the collection of reliable speech data, an essential element of any attempt at explanations of linguistic facts. In addition, an emphasis on typological generalizations will be essential in the projected comparison of two apparently unrelated linguistic contexts (Belizean and Chinese). Functional grammar provides that angle, with its spe-

INTRODUCTION: DEVELOPMENTAL CONTINUA

15

cial emphasis on universal pragmatic aspects of human communication, espe­ cially in areas such as topic and focus. Since I elected to exclusively investigate topic marking in spontaneous verbal interaction, the functional approach pro­ vides a valuable framework within which to place the Belizean and Chinese facts. The processes involved in topic presentation and topic organization pre­ sent particularly interesting insights into the ambiguous overlap of syntactic and pragmatic processes. Thus, without committing myself fully to either model, I find it convenient and eminently feasible to combine the sociolinguistic framework with the func­ tional-typological perspective. Indeed, I find the two perspectives to be comple­ mentary, in the sense that what is lacking in one is supplied by the other. Whereas sociolinguistics is generally restricted to individual linguistic situations, func­ tional grammar provides options for cross-linguistic generalizations; while func­ tional grammar has little to say about specific contexts and speakers, sociolin­ guistics has a well-developed methodology and theory to identify and control the extralinguistic variables which have significant impact on language choice. Finally, the overlap of semantics, syntax, and pragmatics is an important aspect of the study of discourse, and the fuzziness of boundaries is often invoked in functional approaches. Although an account of the absence of neat­ ly separated components presents a serious challenge, the issue cannot be skirt­ ed. The notion of overlap is particularly relevant to the study of marginalized varieties. Since nonstandard varieties are usually highly stigmatized, there is no formal or canonical description of their structure, and they are not subjected to a norm. They are, therefore, relatively free to evolve. It is the intent of this investigation to observe whether and how spontaneous contexts favor the clos­ er association of semantic, syntactic, and pragmatic processes, as noted in cre­ ­le situations. A hypothesis to be considered here is that unguided learners, as well as learners deprived of a stable model, are forced to rely on general prag­ matic or functional principles rather than on the formal syntactic rules of the tar­ get language because of the continued inaccessibility of the standard grammar. In other words, speakers may pragmatically rearrange the target syntax. The hypothesis will be advanced that such pragmatic reorganization may derive from universally shared principles, regardless of the specific linguistic situa­ tion. The issue of the putative overlap of pragmatic and syntactic devices will be investigated here more particularly in the acquisition of Belizean acrolects by native speakers of Creole, and of Chinese standard dialects by native speak­ ers of nonstandard Chinese varieties. 3.4 Units of analysis An essential prerequisite in the linguistic analysis of discourse involves delim-

16

CREOLE AND DIALECT CONTINUA

iting the domain of the study. The Chomskyan notion that the domain of lin­ guistic analysis can and should be the sentence unit constitutes the standard basis for most analyses—but it is easy to demonstrate that investigations limit­ ed to isolated sentences fail to provide essential information about the broader psychosocial context that generates the parameters within which a sentence occurs. Furthermore, it is often difficult to identify sentence boundaries and, thus, to break discourse into significant sentential units. This structural uncer­ tainty is represented in the following example, in which a Belizean speaker dis­ cusses "modernity":5 (1) This thing of modernity sometime, I am skeptical of it myself; I think when you talk of the economics of it you want to bring in some sophisticated system, make them become dependent on it, and, you know, look pan de land, look pan de people, and you just see them as factors of production rather than seeing them for what they really are, what the land can produce and how it will—de people who till it, rather than going into this sophisti­ cated system, say small is beautiful (Co, 42, 1980, Placencia) The definition of linguistic units in discourse, however, does require some reference to its boundaries and to its internal organization. Wald (1983:104) looks at narratives as "discourse units that are extended, potentially syntactical­ ly elaborate, commonly spoken, and easy to elicit." Such descriptions capture the complexity of discourse but do not contribute much empirical help to the analyst intent on defining descriptively adequate measurements. A more pro­ ductive attempt to delimit the "discourse origin" and "the end point" of a nar­ rative leads Wald to refer to a variety of cues, for example: "a coda such as that's it....or a falling intonation and slowing of the speech tempo finally laps­ ing into silence. . .[and] the following audience (or addressee) reaction" (Wald 1983:108). Labov and Fanshel (1977), in their study of the language of therapy, iden­ tify "units of talk" which function as speech actions, and their goal is to formu­ late the set of principles organizing those units of talk, either as concatenation rules sequéncing information, or mapping procedures which determine the pro­ duction and interpretation of verbal elements. According to Linde (1981:113), "discourse units have an internal structure that is as regular and accessible as the study of the structure of sentences." She identifies certain internal organizational principles which provide coherence in discourse, such as temporal ordering, relations of dominance and inclusion, and a wide set of social and cultural assumptions and presupposition. Linde defines major linguistic units in terms of genres, such as stories, jokes, recipes, narra­ tives of activities, and the like, all functioning as social units as well. One

INTRODUCTION: DEVELOPMENTAL CONTINUA

17

advantage of this approach is that it makes sense to speakers, who are far more aware of undertaking to tell a story than they are of undertaking to produce a sentence. A disadvantage of the "genre as unit" interpretation is that it may be too broad. A story is likely to be subdivided into various components, inter­ rupted by interlocutors' comments, then continued and led to completion; thus, a topic can be maintained over alternating turns of conversation. It may be practically more effective to define linguistic units shorter than the genre, while leaving open the potential of expanding beyond conventional idealized sentential boundaries. Sample (2) shown below illustrates the problem of discourse division. It is excerpted from a conversation taped in Placencia, between Co and his friend Ro. Co and Ro used to be coworkers at the income tax department and the customs office in Belize-City. They recall a common friend Xo. The topic of Xo is introduced by Ro's basilectal question [We he do now?], but Co immediately shifts to an acrolect (more appropriate when dis­ cussing aspects of their official functions as civil servants in Belize-City) and sometimes to a mesolect (turn 5), although Ro occasionally shifts back to a basilect (turn 4). 6 Beside the issue of lectal shifting, a point relevant to the def­ inition of discourse units is represented in the fact that the topic is interrupted on turn 6 by Ro's digression into recollections of office relationships, to which Co responds (turn 7); Ro then returns to the topic of Xo (turn 8). Thus, the prob­ lem is how to break down the following chunk of discourse. Should there be eight discourse units (each turn constituting one discourse unit)? Could it be three, the first discourse unit including turns 1-5 (discussion of Xo), the second (6-7) about office relationships, then the third resuming the topic of Xo (turn 8)? Or could the entire paragraph be considered as one single discourse unit: (2) Ro: We he do now? (what does he do now?) 1. Co: I think he wanted to get into merchandising and maybe in bond stores, duty free shops. 2. Ro: I used to like work with he you know, they gave he lee [little] bad name and so but. 3. Co: I found him alright, boy, I got on pretty well with him, ahm, we were from way back, you and I know him even before being a civil ser­ vant, and ahm, then I worked with him the first time at income tax, then I worked with him at customs, but, I'm, telling you, I think he was a. . .he had ideas, right. 4. Ro: Dat da we I like with a [that's what I like about him]; and he got good, to me, right, we got good staff relationship, right. 5. Co: He used to fine [He used to be fine] 6. Ro: I know one big place like customs funny for you, because the staff too big, right, but to me, like, I use to like ya [there], they make the staff

18

CREOLE AND DIALECT CONTINUA there feel at home, right, for instance da you lee birthday today and things like that. 7. Co: Yah, and I think it was good, right, kinda, you work together you must also fun together, you socialize, you find it makes for better working relationship. 8. Ro: Dat's right, I never find no fault with the man, man. If you m ga wan man de but da Xo de [If you had to pick a man, it would be Xo]; yah, he da something else, he da dead scatterbrains, man. (Placencia 1980)

For the purpose of this investigation, I will opt to take the entire paragraph as a discourse unit of analysis. My definition of "topic units" is centered around the topic: all structures related to the topic make up the topic unit, as will be further illustrated in chapters four and six. In many languages, including English, the basic syntactic structure of the sentence involves an SVO order in which topics do not have an especially high priority; yet, topic mechanisms occur more extensively in contextualized casu­ al and nonstandard varieties related to English, which then appear to assign topic structures to the pragmatic component. The incidence of topic mecha­ nisms in creole varieties, especially in acrolects, is of special interest here and is related to the claim that the pragmatic component takes over in spontaneous contexts. On the other hand, some languages clearly assign syntactic status to what is pragmatically constrained elsewhere, a phenomenon akin to grammaticalization. This may be the case in Chinese which displays consistent topiccomment discourse structures. The concept of topic chain (Shi 1989; Tsao 1979), to be also discussed in chapter six, may provide an appropriate framework for the study of creole topic mechanisms because of its flexibility in allowing a definition of the topic unit which combines syntactic and pragmatic elements. According to the theory of topic chains, the topic extends its domain to a sequence of several propositions. The topic chain, or topic unit is thus a dis­ course level phenomenon that raises the issue of the interaction of syntax and pragmatics, as illustrated in the Belizean examples above. And it thus appears that cross-linguistic observations provide a productive perspective on our understanding of topic mechanisms.

4. Creoles and noncreoles Parallel to the growing interest in Black English evidenced in the last thirty years, and perhaps even more extensive, has been the development of the field of creolistics over the last twenty years. This creole scholarship has endeavored to rehabilitate stigmatized creole languages in their own communities, and else-

INTRODUCTION: DEVELOPMENTAL CONTINUA

19

where, and to delineate a new field where social and linguistic factors are inter­ locked in language development. As indicated above, however, the attention devoted to the development of pidgin and creole-related continua has been somewhat limited. Several important books providing overviews of creoles, or case studies, have appeared since 1980 (Arends et al. 1995; Alleyne 1980; Bickerton 1975; 1981; Chaudenson 1992; Gilbert 1987; Holm 1988; LePage & Tabouret-Keller 1985; Muysken & Smith 1986; Rickford 1987; Romaine 1988; Thomason & Kaufman 1991). One of the current challenges in the field of creole studies involves pro­ viding satisfactory accounts and explanations of the linguistic variability inher­ ent in creole situations. There are apparently conflicting interpretations of vari­ ability, namely, the "continuum" and the "co-existent systems" perspectives (to be further discussed below). The varieties spanning a creole continuum are often characterized as unidimensional and displaying implicational patterns (Bickerton 1975; Rickford 1987). However, analyses leading to these interpre­ tations are usually based on phonological or morphological variables, rarely on observations of linguistic features beyond the sentence level. A different view holds that only a multidimensional model can accommodate creole situations. In particular, LePage & Tabouret-Keller (1985:180-185) claim that the linguis­ tic strategies used in Cayo District, a Western district of Belize with predomi­ nantly Spanish and Creole ethnicity, are necessarily multidimensional because each linguistic choice is an "act of identity" that implies solidarity with or dis­ tance from others. This perspective is in keeping with the speech accommoda­ tion theory, which involves the principles of "convergence" and "divergence" in human psychosocial interaction (Giles, Bourhis, & Taylor 1977). The multidimensional interpretation of linguistic variability need not be restricted to multilingual, multiethnic communities, as will be discussed later. The issue of decreolization is often linked to the continuum phenomenon. Decreolization has been interpreted as a development of the continuum in the direction of the standard model, leading from basilect to mesolect to acrolect, with concomitant loss of the "earlier" forms. The process has often been assumed to be a unidirectional substitution of the standard for the creole.7 Having embraced the decreolization view at early stages of my research (Escure 1981), I have come to the conclusion that the reality of creole continua, even in a relatively homogeneous speech community like Placencia, Belize, (see chap­ ter two) is more complex than can be accounted for by a unidimensional model. The learning of acrolects does not necessarily imply the loss of native basilectal vernaculars. On the contrary, the acquisition of acrolects involves an exten­ sion of the repertoire available to speakers rather than a systematic sliding of the continuum toward a more standard set of linguistic segments (Escure 1981;

20

CREOLE AND DIALECT CONTINUA

1982; Rickford 1987:31). There is a reasonable possibility that the process of change through extension and remodeling may be typical of any acquisitional situation, implying the addition of a second code to a first variety. This is like­ ly to apply to SLA as well. I firmly believe that the acquisition of a second lan­ guage does not systematically and irreversibly move forward but is subject to backsliding and the preservation of socially meaningful intermediate forms, not unlike the acquisition of varieties along a creole continuum.8 If this is true, SDA may not be drastically different from SLA, and creole situations may be struc­ turally similar to other language variability situations, although they may still differ in other respects, perhaps historical or political. For Muysken & Smith (1995:4-5), "creole languages are not in the slightest qualitatively distinguish­ able from other spoken languages." A comparison of second dialect acquisition in creole and noncreole con­ texts will help clarify the definition of the concept of "creole." In particular, the following problems will be addressed: are creole languages and their continua structurally different from others? Is the pattern of standard acquisition differ­ ent for a native creole speaker as opposed to a native speaker of a noncreole, nonstandard dialect? Those questions will be examined in the very limited con­ text of the discourse variables selected for analysis. There is a strong indication that, even though the term creole is widely used in various types of contact-induced situations, it is not clear that a creole is typologically different from other language types which develop through some type of sociolinguistic contact (Corne 1995a; Escure 1993a). It will be hypothesized here that second dialect acquisition is somewhat analogous, though occurring under apparently less traumatic conditions, to the situations which permitted the genesis of pidgins and creoles. Even though a model exists (the standard is usu­ ally codified in written form and is also subject to informal standardization), it is never systematically related to the learner's native system; thus, the learner is forced into the same decision making situations—what to select from a variety of options. Furthermore, the same social pressures exist: both the incipient pid­ gin speaker and the nonstandard learner are in positions to be affected by vary­ ing degrees of social oppression and made to feel that they cannot meaningfully participate in social life unless they control, among other things, the "proper" code—which nobody is attempting to teach them properly or at all! In the case of nonstandard learners, language universals may have already been actively used in the natural development of their uncodified vernacular. 4.1 Two case studies It is hoped that a detailed analysis of selected aspects of two ethnically differ­ ent types of second dialect acquisition will contribute to our understanding of

INTRODUCTION: DEVELOPMENTAL CONTINUA

21

cross-linguistic acquisitional mechanisms and of linguistic bias. First, the case of Belize, Central America: in that former British colony, the official language is English, directly inherited from the colonial past of what was previously identified as "British Honduras." English is the medium claimed to be used in all governmental and educational functions, but the reality is far from meeting official expectations. The Creoles (Afro-Europeans) overtly rec­ ognize the prestige of English, but their vernacular language is an Englishbased creole which has in-group prestige. Other groups manifest equally ambiguous attitudes and behaviors toward the standard acrolect. The acquisi­ tion of Belizean acrolects occurs in the context of a multilectal creole situation, which is likely to have always been variable, and definitely still is. It is a case of relatively recent language formation through pidginization and creolization. Second, the case of the acquisition of Standard Mandarin by speakers of other varieties of Chinese having local, but not official, status occurs in the more traditional context (noncreole) of dialect acquisition, so far undocumented. The People's Republic of China (PRC) has been actively promoting—through mass media and educational channels—the increased use of the new standard of Mandarin Chinese called Putonghua (the "Common Language"). Beijing natives claim that they cannot understand the provincial dialects spoken all over China, but most of the recent leading political figures of China are not natives of the Beijing province and sport remarkable accents: Mao Tse Tung was born in Changsha, Hunan, in the southern part of PRC, where a distinctive lan­ guage—Xiang—is spoken; Deng-Xiao-Ping is a native of Chengdu, Sichuan, whose Southwestern Mandarin variety is also very distinctive. Obviously, Belizean English-based creoles and Chinese dialects are genet­ ically and historically unrelated, and the countries in which those varieties are spoken are maximally different in size and ethnic composition. This apparent lack of connection is one reason why I elected to analyze the linguistic situa­ tions in those two geographically distant sites. Comparing the two types of acquisition should highlight universal aspects of acquisition, as well as the issue of whether there is any substance to the notion that creoles develop in any dis­ tinctive way (Escure 1993a; 1994). In addition, it is recognized that creoles constitute the linguistic product of relatively recent colonization and enforced transplantation and subjugation of one group to another region and culture, whereas a general assumption of homogeneity is typically associated with Chinese cultures and languages—a manifestly erroneous assumption. Both situations offer similarities in terms of the general social background underlying their respective developments. Chinese populations have been forever subjected to the constant oppression of various emperors and overlords invading each others' territories, and there has

22

CREOLE AND DIALECT CONTINUA

been a plethora of successive waves of colonization from the north to the south of China. The most recent invasion was the Manchu takeover which resulted in the domination of one nation by an outsider group. Linguistically, the dominant Manchu are said to have adopted the language of the subjugated nation and, more specifically, the Mandarin variety spoken in Northern China (Beijing vari­ ety). The Manchu invasion occurred at about the same time as the main slave trade from Africa, which changed the ethnolinguistic map of the New World. It is obvious that, historically, similar types of linguistic contact and ethnic mixing phenomena underlie China and Belize. 4.2 Outline of chapters It is expected that close observations of the standard versions produced by native speakers of nonstandard dialects will provide interesting insights into the universal processes underlying the development of language in society. This volume investigates the development of second dialects in the context of the two distinct sociolinguistic situations outlined above: The first part of this volume, chapters two through four, examines linguis­ tic variation in the context of English-based creoles and the English acrolectal varieties acquired by native speakers of creole basilects, as spoken in Belize (previously British Honduras) where I have conducted fieldwork from 1978 to 1994. Chapter two provides the demographic, geographical, social, and histori­ cal background of the speech community of Belize, highlighting some of the conditions under which the creole continuum is currently developing English acrolects and focusing more particularly on the Stann Creek District which is the locus of this research. Chapter three is concerned with the acrolectal seg­ ment, the product of the acquisition of the standard as second (non-native) dialect. The intent is to identify potential differences between the acrolect and the standard variety, the apparent target in acrolectal development. Chapter four specifically investigates the use of topic strategies in Belizean acrolects, com­ paring syntactic and pragmatic aspects of standard acquisition to basilectal cre­ ­le patterns and casual American English varieties. The second part of this volume, chapters five through eight, investigates similar problems in Chinese dialects spoken in the People's Republic of China, including the native vernacular Mandarin/Putonghua spoken in Beijing, the non-native Putonghua acquired by speakers of the Wuhan varieties (Wuhanhua) in Wuhan (Hubei), and the narrative variety of Wu used for traditional tales in Suzhou (Jiangsu) which is also acquired as a second dialect. Speech data were collected since 1985 (Escure 1987), and vernacular data spanning the period from 1963-1985 are derived from contemporary written texts. As a counterpart to chapter two in the Belizean section, chapter five provides parallel background

INTRODUCTION: DEVELOPMENTAL CONTINUA

23

information on the Chinese dialect situation with special emphasis on Beijing, Wuhan, and Suzhou. Chapter six outlines crucial aspects of Chinese syntax in discourse. Chapter seven looks into the use of various types of topic strategies in several varieties representing Standard Chinese/Putonghua. Chapter eight evaluates similar mechanisms in the context of two acquired dialects and leads to a comparison of first and second dialects in some representative Chinese var­ ieties. Finally, chapter nine establishes a comparison of the patterns observed in Belizean and Chinese standard acquisition and, more particularly, discusses the interaction of syntax and pragmatics, and the issue of the universality of dialect acquisition. From the above remarks, it follows that, should any structural sim­ ilarities be observed in the linguistic patterns adopted by learners of second dialects in Belize and in China, language-specific substrate and superstate mechanisms could be ruled out. Such potential findings would warrant careful scrutiny, with the ultimate goal of getting closer to an understanding of univer­ sal linguistic processes. Endnotes 1. Except in relatively restricted cases of simultaneous bilingual acquisition which does not involve diglossia-some sort of social discrepancy between the two languages. In fact, few, if any, such situations exist. 2. For example, Black English in the United States has become extremely popular among some white teenagers; however, I know of no full-scale study documenting the extent of its acquisi­ tion by groups other than African-Americans. 3. It has been demonstrated that there is no correlation between multilingualism and low eco­ nomic development or political instability, as discussed in Fasold (1984:5-9). 4. Thanks to Stephen Matthews and Virginia Yip for the information about Canto-rap. 5. In this excerpt, any representation of the phonological idiosyncracies of Belizean acrolects has been omitted since the point of this example is to discuss the challenge of defining discourse units. The only lexical unit that may be different from English is pan 'on' [look pan the land, pan the people]. 6. Linguistic features characteristic of basilects, mesolects, and acrolects will be discussed in later chapters. 7. A parallel interpretation of the acquisition of second languages assumes a unilateral shift from basilang to mesolang to acrolang (Stauble 1980:47; Schumann & Stauble 1983:263). 8. Although I don't know of any study testing in detail the hypothesis of the nonuniformity of L2 learning, all adult speakers of a second language experience the unreliability of L2 perfor­ mance which varies in terms of multiple factors in the context of interaction.

Chapter 2 The Belizean Speech Community and the Use of English Abstract This chapter provides general demographic and socioeconomic information on the Belizean community, with special focus on Belize's ethnic and linguistic pluralism. The language situation is briefly placed within the specific context of the Englishbased creole continuum. The locus of the research is Stann Creek district and, in par­ ticular, the Creole village of Placencia and the neighboring Garifuma community of Seine Bight. The linguistic range, from basilect to acrolect, is illustrated in sample texts, with a brief outline of idiosyncratic features defining each lect. 1. Introduction to the history of Belize Aldous Huxley (1934:32) said: "British Honduras is not on the way from any­ where to anywhere else, has no strategic value, is all but inhabited; if the world has any ends, British Honduras would be one of them." With due respect to Huxley, nothing could be further from the truth! Far from being the dead end implied by Huxley's quote, Belize is a vibrant land of diversity and paradox, with city slums and unspoiled beaches, lush jungles and deforested land turned into sugar plantations only a few miles apart from each other. In multicultural Belize, ethnic diversity and miscegenation are taken for granted, yet Belizean society has inherited the color-coding system of colonization. Literacy and edu­ cation are high priorities but educated Belizeans are offered few opportunities at home. Thus, many of them have emigrated to the United States which has a higher Belizean population than Belize itself. Thc young average age of the population confirms the continuing scope of emigration patterns. According to the 1991 population census, 75% of the population is under twenty-one, about the same as in 1980. Belize, which originated in a lively brew of cultures, lan­ guages, and dialects is still developing in multiple dynamic directions. Steeped in a violent past of invasions, clashes, migrations, exploitation, piracy and slav-

26

CREOLE AND DIALECT CONTINUA

ery, Belizean society offers fascinating insights into patterns of human adapta­ tion in the face of adversity and into the creativity of linguistic processes in complex communicative situations. The area was the scene of conflicts which almost eliminated Maya civilization, of clashes between Spanish soldiers and British buccaneers, of pirate landings, slave logging camps, Indian caste wars, Black Carib deportation, and finally Salvadoran refugees' immigration. In order to understand the range and overlap of the linguistic codes used in Belize, their social functions and iconic values, and eventually the nature of the second dialects spoken in the context of the traumatic history of Belize, it is essential to delve briefly into some historical aspects of the country. The name Belize was first applied to the main waterway and later to the principal city which became the capital of the independent territory of British Honduras. The origin of the lexical item Belize is somewhat unclear. It may be derived from the name of the assumed founder of the original settlement which eventually became Belize—a Scottish pirate named Peter Wallace, or Wallis, who was possibly a lieutenant of Sir Walter Raleigh.1 As reported in the 1827 Honduras Almanack,and as Bancroft (1883) and Asturias (1941) claim, the first settlement of British Honduras was founded either in 1617 or in 1638 by some eighty shipwrecked sailors led by Wallace/Wallis. The area settled would then have been named "Wallis" or "Ballis" by the Spaniards, due to some interfer­ ence of the Spanish bilabial fricative.2 The former British colony of British Honduras, self-governed since 1973, became independent on September 21, 1981, after a long period of negotiations with Guatemala which had long-standing claims to what is still referred to on Guatemalan maps as the Belice province. This 13,000 square kilometer territo­ ry (260 kms at its longest and 100 kms at its widest) is wedged between Spanish-speaking Central America (Mexico, Guatemala, and Honduras) and the Caribbean Sea. It is bounded on the north by Mexico's Yucatan state of Quintana Roo and on the south and west by Guatemala. Belize has the lowest population density of any Central American country, with an overall population of 189,392 according to the most recent (1991) pop­ ulation census. More than a quarter of the population is clustered in the capital Belize-City (population 43,710). Because of limited economic opportunities, Belize has been traditionally a land of emigration, and although it is difficult to have accurate figures, it can be assumed that there are at least as many Belizeans living in the United States as there are in Belize. The 1991 census finds that, overall, 3,000 Belizeans emigrated since 1980—over 85% of them to the United States—with a steady flow over the years (although no such data are available to document emigration before 1980). This figure is almost certainly an underrepresentation of the facts of emigration for two primary reasons stated in the

THE BELIZEAN SPEECH COMMUNITY

27

population census report: first, entire households who emigrated are not repre­ sented, and second, Belizeans residing in Belize may have been reluctant to report family members living abroad who have not yet regularized their resi­ dence status in their land of emigration. Considering those factors, I believe that a reasonable estimate of emigration since 1980 may be closer to 5,000 and, over thirty years may have reached as much as 150,000, and thus may have matched the current population officially residing in Belize. Such high emigration patterns appear to be confirmed by the relative youth of the Belizean population: 65% is under age twenty-four whereas the most pro­ ductive segment of the population (ages twenty-five to fifty-four) amounts to only 28%. Individuals over fifty-four constitute only 8% of the population. The 1980 census showed a similar trend in age distribution for the ten previous years. This overall pattern suggests that most of the breadwinners live abroad. Emigration is obviously an important factor likely to influence the lan­ guage situation, either because this young population is likely to have a power­ ful impact on language choice or because connections with United States resi­ dents may affect language development in Belize. Considering the pervading emigration, the population could be expected to show a slow increase or even a decrease. In fact, between 1980 and 1991, the population increased by 44,000, an increase of almost 30% and almost twice as much in real numbers as in the previous intercensal period (1970-1980).3 In comparison, the intercensal growth was less between 1970 and 1980 than in the previous ten years as shown in table 2.1: Table 2.1: Population growth in Belize (Population Census, 1991) Year

Population

1960

90,565

1970

119,645

1980

145,353

1991

189,392

Growth during decade 29,000 25,000 44,000

Why this apparent growth? In spite of regular emigration, increased childbirth is not the explanation. It appears that the traditional, or at least thirty-yearold, emigration channels to the United States have been offset by a strong incoming immigration trend. Over 25,000 people living in Belize in the last ten years reported being born abroad, which neatly accounts for the doubled inter­ censal growth noted above. Most of those immigrants are from Guatemala and El Salvador (both neighboring countries which suffered high political instabil-

CREOLE AND DIALECT CONTINUA

28

ity) and, to a lesser extent, Mexico and Honduras. Those four countries, all Spanish-speaking, constitute 85% of the native lands of recent immigrants. This fact is an important factor related to language use. Spanish, one of the languages of Belize, has now gained majority status over English-based varieties. The eth­ nic composition of the country is also affected. 2. Ethnolinguistic composition and census data The diverse Belizean population results from repeated waves of immigrants usual­ ly forced upon the land. Settling at different times but adding extraordinary diversi­ ty to the indigenous Amerindian population have been a range of peoples: shipwrecked British seamen, French and British pirates, African maroons and slaves, Miskito Indians brought from the Miskito Coast (now Nicaragua), Black Caribs (Garifuna) deported from St. Vincent, then Honduras, Spanish and Indian refugees from Mexico, Mennonites fleeing religious persecution, and more recently Salvadorans and Guatemalans fleeing political oppression. The contemporary pop­ ulation includes four major groups: Amerindians, Creoles, Mestizos, and Black Caribs (Garifuna), and other smaller groups. Escure (1983a:32) attempted to inter­ pret the 1980 census data (Population Census 1980); the distribution of ethnic groups and the languages spoken is represented in table 2.2: Table 2.2: Ethnic groups, population, and languages of Belize in 1980 ETHNIC GROUP

Creoles Mestizos Garifuna Maya Kekchi Whites East Indians Chinese Other: Lebanese Mennonites Salvadorans

%

SECOND LG.

THIRD LG.

39.7 33.1 7.6 6.8 2.7 4.2 2.1 0.1

Creole Spanish Garifuna Maya Mopan Kekchi English English Creole Cantonese

English Creole Creole English Creole** Creole English Creole

Spanish* English English Creole**

3.6

Creole Low German Spanish

English Creole**

POP.

FIRST LG.

Spanish* English

* in some areas only **rarely

Although each Belizean is quite clear about his or her ethnic and cultural identity, the official classification of ethnic groups offers particular challenges

THE BELIZEAN SPEECH COMMUNITY

29

in Belize. The 1980 census used nine categories, including the labels Negro/Black and Mixed, which drew most votes with 76% and 13% respective­ ly. There was no reference, however, to the locally used terms of identification: Creole, Black Carib or Garifuna, and Mestizo.4 This classification overlooked the peculiar ethnic configuration of Belize and local attitudes toward identity. At least three of the major groups are mixed (Creoles, Black Caribs, and Mestizos), and the first two can be considered "black." Specifically, the Black Caribs, who are now officially referred to as Garifuna (see below for a more extensive discussion of this term and other alternatives) are indeed mixed (Afro-Indians), although they usu­ ally have dark skin. Creoles are also by definition mixed (Afro-European), but they range over a broad spectrum of coloring and morphological features. Yet, Creoles and Garifuna are culturally and linguistically very different groups. As to the Mestizos, they are also mixed (Hispanic/Indian), although many members of the group identify themselves as "Spanish," and there was no label indicating a Spanish identity in the 1980 census. Furthermore, no distinction was established between the Maya and the Kekchi, two distinct groups of Amerindians who speak different, though related, varieties of Mayan. The 1991 census is a significant improvement over the previous ethnic census because it used labels popularly used in Belize, twelve altogether as optional responses to the question: "To what ethnic, racial, or national group do you think you belong?" The overwhelming majority fits into the Mestizo (43.6%) and the Creole (29.8%) groups. The foreign-born (14% of the population) account for the shift in the majority groups. This figure, however, does not clearly distinguish between recent arrivals and traditional Belizean Mestizos. It might have been help­ ful to include national origin such as Guatemalan, or Salvadoran, or the use of the label Spanish/Spaniard/Hispanic besides Mestizo, which would have shed light on the ethnic shift apparently ongoing in Belize. Table 2.3: Ethnic Groups in Belize in 1991 ETHNIC GROUP

Creole Indian Garifuna Maya Mopan Kekchi Maya Other Maya German/Dutch Mennonite Mestizo Chinese Syrian/Lebanese White Other

POPULATION (%)

29.8 3.5 6.6 3.7 4.3 3.1 3.1 43.6 0.4 0.1 0.8 1.0

CREOLE AND DIALECT CONTINUA

30

The most serious inadequacy of the 1991 census lies in its investigation of languages spoken and the discussion of the results. The census dealt with the language situation through two simple questions (Do you speak English?/Do you speak Spanish?) and three answer options in each case (very well/not so well/not at all). One might argue that subjective responses to fluency are generally unreli­ able because speakers tend to either overrepresent or underrepresent their lan­ guage skills, depending on the local or national status of those languages. This is a serious issue in Belize where English has the status of "official language" and serves as the language of education. However, the most serious problem may be that no attention is given to the use of the creole language. Furthermore, there is no attempt to assess the language acquisition and shifting strategies of the population. Indeed, the glaring omission of any reference to Creole confirms the continued presence of bias toward native vernaculars. Such an omission can only produce an inaccurate interpretation of the census data. Census figures indicate that 54.3% of the population believe that they speak English "very well," 22.5% "not so well," and 23.2% "not at all," sug­ gesting: "that just over half of our population speak the official language very well points to the need for some serious considerations, especially in view of the fact that fluency in an established language is an indispensable tool for the acquisition of any other form of knowledge" (Population Census 1991:8). On the other hand, the stated frequency of good Spanish fluency corre­ sponds to the percentage of Belizeans who, in 1991, identify themselves as Mestizo (43.6%), as shown in table 2.4: Table 2.4: Stated language fluency (1991 Population Census)

Speak English Speak Spanish

Very well 54% 43.8%

Not so well 22.5% 11.1%

Not at all 23.2% 45.1%

This table reflects a clear and complementary assessment of Spanish flu­ ency as "very well" and "not at all." The growing impact of Spanish from var­ ious proveniences suggests that other languages, such as the Mayan varieties spoken in the western and southern districts, may be affected by the develop­ ment of Spanish.5 In contrast to clear assessments of Spanish fluency, table 2.4 reflects the uncertainty related to what constitutes English fluency. The three major groups (Creole, East-Indian, Garifuna), who natively speak some form of English (assuming the creole is some form of English) constitute about 68% of the overall population, yet only 54% estimate that they speak English very well.

THE BELIZEAN SPEECH COMMUNITY

31

The uncertain piace and status of Creole vis-à-vis English is obviously part of the problem, an issue not addressed by the census. The issue of bilingualism or multilingualism is also not addressed. This official dismissal of Creole points to the necessity of including in census reports findings about the extent of sec­ ond language acquisition. Finally, it would have been interesting, if not crucial, to investigate which portions of the population estimate that they are fluent in Garifuna and in Mayan/Kekchi. Such figures may have helped assess the vital­ ity of the languages spoken by these two Belizean minority groups. The differential in language use and language fluency in English and Spanish is quite important and is obviously linked to the differential distribu­ tion of ethnic groups across the six districts of Belize, as represented in table 2.5. Creoles predominate in the Belize district (around the urban area of BelizeCity)6 and are also widespread in Cayo (west, near Guatemala) and Stann Creek (southern coastal area). Mestizos constitute the majority in Corozal and Orange Walk (both north near Mexico) and Cayo. The Garifuna only have a significant presence in Stann Creek. Finally, the southernmost Toledo district is clearly Maya territory, at least in rural areas, since the main town of Punta Gorda is in majority inhabited by Garifuna. Table 2.5: Ethnic groups in the six districts of Belize (1991) (arranged from northern to southern geographical locations; figures indicate percentages)

Districts COROZAL ORANGE WALK BELIZE CAYO STANN CREEK TOLEDO

Creole

Garifuna

Mestizo

Maya

7.6 7.4 67.9 23.0 25.1 5.7

1.3 1.2 5.3 1.7 36.2 10.0

74.1 71.7 18.7 58.0 23.7 11.9

5.0 9.1 1.2 8.7 8.0 62.8

A brief overview of Belizean ethnicity is outlined below, group by group, and in terms of their chronological presence or arrival on the scene of the territory which now constitutes Belize. 2.1 Amerindians The most ancient ethnic group in Belize is the Amerindian group. Maya civiliza­ tion flourished primarily in the Yucatan peninsula, extending into Belize and neighboring areas of Honduras and Guatemala, but abruptly declined in the ninth century A.D., possibly because of a violent popular revolt against the religious rulers. There is, however, archaeological evidence of continued occupation of the

32

CREOLE AND DIALECT CONTINUA

Maya sites (presumably by farmers) well into the fourteenth century and proba­ bly much later, up to the time when the British began settling in Belize in the early seventeenth century. Ruins of cities and pyramids are scattered and half-buried under the tropical jungles, a silent testimony to the existence of a lost civilization forgotten even by its descendants. Some of the sites that have been excavated or surveyed include Altun Ha, San Jose, La Milpa, Mun Diego and others in northern Belize; Xunantunich, Actun Balam, Caracol, Cahal Pech and Tzimin Kax in the West (Cayo District near the Guatemalan border); and Lubantum and Pusilha in the southern Toledo District. The Mayas in Belize now consist of three groups, each speaking a differ­ ent language variety and living in separate areas. The Mopan and Kekchi Mayas are the descendants of the original inhabitants of Belize who reentered the area from Guatemala during the end of the nineteenth century; they live pri­ marily in the Toledo district, in southern and western Belize near the Guatemalan border. On the other hand, Yucatecan Mayas came from Mexico after the 1847-1853 Indian Caste War, and they now live in the northern Corozal and Orange Walk districts, close to the Mexican border. Most of the Mopan and Kekchi-speaking Mayas still live in isolated villages and, until recently, have had little contact with other groups. There is an increasing admixture of the Mopan and Kekchi, but Mopan speakers have a somewhat higher status locally, perhaps due to the fact that the largest village of San Antonio (pop. about 200) is primarily Mopan, with some assimilated Kekchi. The southern Mayas consider themselves to be different from the northern Yucatec Mayas who primarily speak Spanish and have been largely assimilated to Mestizos. Yucatec Mayan is still spoken in some rural areas, though Spanish is now becoming the first language as it is associated with progress and social mobility for members of this group. Northern Mayas and Mestizos, thus, close­ ly identify with Mexico, and Spanish is the primary marker of ethnicity for them. Creole, however, competes with Spanish as an indicator of youth soli­ darity and interethnic friendship. Even though Spanish-speaking people demographically dominate the northern Belizean community, the size of the Creole group has increased in the city of Corozal, and the limited Creole/Mestizo inter­ action allows for some Creole expansion in the community. In the western Cayo District, the Mestizo and Creole populations mostly balance out; thus, Spanish and Creole have roughly equal status in the commu­ nity at large. Interaction between the two groups is frequent, and intermarriage is more common than in the northern part of the country. Consequently, chil­ dren in Cayo have a choice of identities, whether Spanish or Creole, as attest­ ed in Le Page and Tabouret-Keller (1985).

THE BELIZEAN SPEECH COMMUNITY

33

2.2 Creoles (Europeans, Africans, Miskitos) The Spanish colonization of Central America in the sixteenth century was con­ centrated primarily in the Gulf of Campeche and barely affected the area of the Bay of Honduras; however, French, Dutch, and British buccaneers had occa­ sional contacts with the Belizean coastline. According to sparse records pre­ sented by Asturias (1941), Bancroft (1883), and Burdon (1935), British settlers had brought Jamaican lumbermen as early as 1655 for the purpose of logging the swamps along the Belize River. As buccaneers increasingly converted to the lumber business after privateering was outlawed, a larger labor force was need­ ed and African slaves were imported to the area from other parts of the Caribbean. In a 1724 report, a Spanish missionary referred to "300 English [liv­ ing in Belize], besides Mosquito Indians and negro slaves, these latter having been introduced but a short time before from Jamaica and Bermuda" (Bancroft 1883, vol.7:626). The Mosquito (now called Miskito) Indians mentioned in the above statement came from the British settlement located on the Pacific Coast area of present-day Nicaragua. The Miskitos had become the allies of the British against the Spaniards and, in exchange, had acquired certain privileges, for example the right to crown their kings in the Cathedral of Belize. British lumbermen moved freely between the two settlements until 1787, when the British had to evacuate the Miskito Coast. At that time, some 2,000 British, their African slaves, and Miskito allies moved to the Bay of Honduras to join the "Old Baymen" as the British Bay settlers called themselves (Escure 1983a:30). Miskito Indians merged with the rest of the population (they have also become Afro-Indians in Nicaragua), although many Belizeans still refer to them by the term Waika (generally as a derogatory term). On September 10, 1798, the Spaniards were defeated by the Old Baymen at the Battle of St. George's Cay, which is commemorated as Belize's National Day. Yet, it was not until 1862 that British Honduras became a colony; it became a Crown colony in 1871. Slavery was abolished in 1838, but freedmen were unable to acquire land because land ownership was monopolized by a very small num­ ber of absentee corporate landlords. The forests had been practically exhausted in northern Belize, and the country's economy could not be adequately converted to agriculture. Today, agriculture-related occupations involve only 5% of the popu­ lation now, with 42.4% stating that they have never worked. As indicated in tables 2.2, 2.3, and 2.5, the Creole population, which was a long-time majority group, has now slipped to second place after the Spanish-speaking Mestizo group, great­ ly expanded by recent Central American immigration.

34

CREOLE AND DIALECT CONTINUA

2.3 Garifuna (Black Caribs) The ethnolinguistic composition of the Belizean population was further modi­ fied in the first part of the nineteenth century. The first significant event was the arrival, starting in 1803, of the "Black Caribs," a group of Afro-Indians whose native name is Garinagu, though they are now officially called Garifuna (the Arawakan term for their language). However, the anglicized terms Carib for the language, and Carib or Black Carib for the people, are still used freely and without negative connotations by the Garifuna-speaking Garinagu.7 The Garifuna group arose from the miscegenation of Arawak and Island Carib Indians with some African maroons who had been shipwrecked on the island of St. Vincent in 1635. After a failed uprising, supportėd by the French, the Garifuna were deported by the British in 1797 to the island of Roatan off the coast of Honduras. Most of them then moved to the coastal areas of Honduras (near La Ceiba and Tela), south to Nicaragua (Bluefields area), or north to Guatemala (Livingston) and to the Toledo and Stann Creek districts of Belize. The largest Garifuna town in Belize is Dangriga, the population of which was 6,435 in 1991 out of 18,085 for the entire Stann Creek district. Dangriga (previously called Stann Creek, like the district) was founded in 1823 by a group who had sided with the losing royalist cause against the republican gov­ ernment of Honduras. In 1939, other Garifuna fled the Honduran village of San Juan, again for political reasons, and founded the village of Hopkins, just south of Dangriga in Belize. There are now six Garifuna communities in Belize: Dangriga, Punta Gorda, Barranco, Seine Bight,8 Hopkins, and Georgetown— the only inland village, founded by Hopkins inhabitants after the destructive hurricane Hattie (Hopkins now has a sparse population). African phenotypical characteristics predominate in the Garifuna, who are physically indistinguish­ able from the darkest Creoles. Their African heritage is mostly represented in the folklore, whereas the Amerindian component has been preserved primarily in their native language, Garifuna, which is mostly Arawakan in structure (Taylor 1977), and will be briefly outlined at the end of this chapter. 2.4 Mestizos Another significant event leading to greater ethnolinguistic diversity was the Indian Caste War (1847-1853) in Mexico which led to a heavy migration of Yucatecan Mayas and Mestizos (Spanish/Indians) from Yucatan into the north­ ern (Corozal and Orange Walk) and western (Cayo) districts of Belize. Mexican Mestizos have continuously trickled in, many settling as fishermen on some of the islands off the northern coast of Belize in Ambergris Caye and Caye Caulker, for example.

THE BELIZEAN SPEECH COMMUNITY

35

The most recent wave of Mestizo/Hispanic immigrants consists of Salvadoran and Guatemalan refugees who started moving to Belize in the 1980s in order to escape their country's civil wars and political persecution. They have been greeted by the Belizean government as a welcome substitute for a consis­ tently depleted Belizean population due to massive economic emigration to the United States. They especially fulfill a dire need for agricultural workers in the banana and citrus plantations of the Stann Creek District, where they have settled in hastily built shantytowns. If not yet into a torrent, the Pania9 trickle has at least turned into a regular­ ly flowing stream, which is currently changing the ethnic and linguistic composi­ tion of Belize, turning at least half of the country into a Spanish-speaking nation. The other half—part of which I will be investigating—is still predominantly the domain of English and Creole. 2.5 Other groups Other ethnic groups arrived at different times to fill the need for indentured labor. The Chinese came from the southern Guangdong province of China in the 1860s to find labor; now, most own small businesses in Belize-City (dry goods stores or restaurants). They have preserved close ties with their ancestral coun­ try, especially with Hong Kong and the Guangdong province of the Peoples' Republic of China and, thus, speak Cantonese even now. Many still bring their wives from China through the age-old tradition of prearranged marriage.10 East Indians came shortly thereafter, in the 1880s, though not by choice. They were Sepoy soldiers, deported to Belize by the British after an aborted revolt against British rule in 1857. The British showed consistency in sending unwanted rebels twice to the backwaters of Belize (first, the Black Caribs, then the Sepoy), an attitude toward the forlorn colony obviously reflected in Huxley's quote which introduced this chapter. The East Indians mostly eked out a meagre living on the southern sugar plantations of the Toledo District which had been offered as land grants by the Belizean government to a small number of Americans planters fleeing the United States after the Civil War. The East Indians are now largely assimilated to the Creoles, both linguistically and eth­ nically, and are concentrated near the city of Punta Gorda. American plantations generally have been abandoned. Some merchants of Middle-Eastern descent, mostly Lebanese (usually referred to as "Syrians") have settled in Belize, as elsewhere in the Caribbean, at various times during the nineteenth and twentieth centuries and have started small businesses which compete with the Chinese. A more recent wave of immigrants arrived in the 1950s, as German Mennonites from Canada and the United States established farming communities in the western part of Belize.

36

CREOLE AND DIALECT CONTINUA

They rarely participate in the life of the country, restricting their interactions to market activities and to small businesses such as carpentry; many move on to other communities (in Central or South America) after a few years. As is the case elsewhere, they have preserved their Low German dialect. 3. The language situation Belize is certainly a multiethnic nation and can also be identified, overall, as a multilingual nation; yet, individual Belizeans are at most bilingual and usually monolingual, with superficial, if any, knowledge of some other language, as indicated by the census surveys reported above. Language use and the reasons for selection and code-switching vary from one region to the next and depend on the ethnic composition, the degree of interaction with other groups, the amount of education, and the relative status of each individual. As is usually the case in multiethnic societies, ethnocentricity and ethnic stereotypes exist and reflect the relative socioeconomic status of co-existing groups. Language choice is an accurate indicator of those attitudes. Since Creole is not officially recognized as distinct from English and is still overtly assigned a low value, one may wonder why it has so vigorously survived and is gaining prestige and a wider speakership. This vitality may be partly due to a "double standard" that has always prevailed in Creole societies. Even though the vernacular may be overtly denigrated by its very speakers and avoided for certain outgroup con­ tacts (see 4.2.3 below), it remains covertly—and sometimes overtly—presti­ gious and highly valued as a marker of ethnic identity and increasingly as a marker of Belizean national identity. Such a situation entails an almost schizo­ phrenic mechanism that requires subtle code-shifting based on a precise assess­ ment of the social situation (Hellinger 1972). 3.1 Language policy and education The official language of Belize is English. English may be unanimously recog­ nized as the language that one must acquire to participate in official government activities, but it is not commonly used in its external (American or British) stan­ dard form. This fact is largely ignored or unidentified by both language users and language planners. Although there is a general, informal agreement as far as the creole end of the continuum is concerned, what constitutes the "correct" local norm called English—the acrolect—is realized differently by different people, and they are not necessarily aware of such variations. The official (and unoffi­ cial) position is that any "non-English" variety is brokop—broken English— which often includes the creole varieties as well as the mesolectal area of the

THE BELIZEAN SPEECH COMMUNITY

37

continuum. Such is the persistent legacy of colonialism. Thus, in spite of the fact that it is gaining speakers and prestige, Belizean Creole is still officially ignored. A typical list of the languages spoken in Belize, as seen on a sign posted at the entrance of Punta Gorda, includes: English, Spanish, Garifuna, Maya, and Kekchi, with the glaring omission of Creole. This is disturbing in view of the facts presented above, and in particular when considering that English is not the Belizeans' first language but a second language or dialect for practically all. The paradoxical coexistence of two varieties, one carrying overt prestige and the other covert prestige, is not an unusual phenomenon. Indeed, this situation occurs wherever a "dialect" competes with a "standard," an issue to be exten­ sively discussed in the following chapters. Official directives to instruct in the overtly prestigious model are naturally passed on to educators; yet, the overlap of the model and the vernacular predictably accounts for the linguistic complexity encountered by children in the classroom. Teachers are expected to teach English, and are believed to be teaching English, but they are often unsure of the exact nature of English as distinct from Creole; more commonly, they are unaware that there is any difference at all. Furthermore, they frequently control a limited segment of the acrolectal end of the continuum which, ironically, is the only segment that they are instructed to teach. Because of the scarcity of teachers, young teachers and teachers' aides are put to work imme­ diately after primary school, with little or no teacher training. They admit that they frequently switch to Creole in the classroom for more fluent explanations. Similar problems are involved in teaching basic subjects to non-Creole Belizean children. Because teachers are often recruited from among the Garinagu, the Maya-speaking children of a remote village are likely to be intro­ duced to reading and writing in English by a native speaker of Garifuna who does not know a word of Maya and who has modeled his or her English on the fluctuating varieties she or he has heard in school, on radio reports which pre­ sent BBC news and local news in Belizean standard, and on Billy Graham's homilies which reach the most distant villages. The resulting effect can only lead to the maintenance, or even the expansion, of the bewildering array of options which make up the creole continuum. For example, Juan, a sixteenyear-old Maya teacher from the Mopan village of San Antonio who had just started teaching in the Garinagu village of Seine Bight told me that he could not speak either Garifuna or Creole, although he could understand the latter, but felt comfortable with English: "Mos'ly, I fin' it difficult wid der language because dey only speak Carib" (3.2: 3). Dennis, a forty-five-year-old Garifuna teacher in the same village, made the following comments, which illustrate, first, the linguistic confusion that must result from the contradictory messages received by the children, and sec-

38

CREOLE AND DIALECT CONTINUA

ond, the socio-psychological factors that determine language choice in the com­ munity at large: You start speaking English to dem [the children], then you give dem de definition in Creole which is much easier to dem because dats what dey hear around dem. Outside you find dat de majority of people, der parents speak Creole, dey speak Carib and Creole. When I first started teaching, while I spoke Creole intermittent, and finally I found out dat I spoke too much Creole, and den I put a stop to it because I had learned English. So I practice my English more. I practice it more even when I speak to my friends. So you would find out when I would stand in front of a class and I speak to dem, I try to speak as fluent as I can. When dey [my friends] speak Carib to me, I speak Carib. When dey speak Creole, I speak English to dem, dats de way I treat dem, right? (3.1: 3). But, later in the conversation, he says: Let's say I go among some friends and, like, we going to have a fine time. Okay, de boys start talking and all everybody just discussing Creole. So instead of trying to make myself feel, or try to show off myself dat I am a better man or I am different dan all o' dem, I associate myself wid dem. So I speak Creole, understan'? So nobody can just say he talk funny, he must be from a different land. Ya understan'? I play when you are in Rome you do like the Romans do. (3.1: 10). Of course, the cost of training teachers to achieve proficiency in English and to acquire a sound methodology for teaching English as a second language or dialect is no small problem, and Belize is still a long way from solving it. It seems, however, that a first significant step would be to recognize the separateness of English and Creole, their distinct linguistic and social values, and in par­ ticular to recognize the intrinsic value of Creole as a linguistic system in its own right. In the recent past, government policies have acknowledged the existence and cultural heritage of the Maya, Garinagu, and Mestizo communities, thus encouraging the use of Maya, Garifuna, and Spanish in school and in local pub­ lic events. The traditional November 19 celebration of the first Carib settlement in Belize has become another Belizean national holiday earning full coverage in the local mass media. In 1985, Radio Belize broadcast an average of three to four hours in Spanish out of a total seventeen hours; however, earlier attempts at using Creole on the mass media (short programs, storytelling on Radio

THEBELIZEAN SPEECH COMMUNITY

39

Belize, and brief articles in local press) have been discontinued. 3.2 English and the Creole continuum In spite of the official designation of English as the language of choice, English is not anyone's native language in Belize. Since the goal of this study is to exam­ ine how English is acquired by speakers of related dialects, in particular by native speakers of the English based creole, an essential prerequisite is to define the patterns which characterize Belizean Creole. Only then will it be possible to interpret potential differences between the target model and the version achieved by its learners. 3.2.1 Putative origins of Creole The variety which is overall the most widely spoken in Belize is, undoubtedly, Belizean Creole, the English-based creole which is the vernacular of the Creole (Afro-European) group. Like other Caribbean creoles, Belizean Creole proba­ bly developed from contacts between Africans and Europeans who emigrated to the American continent. It also ftmctions as the lingua franca for most Belizeans of various ethnic origins. The sociohistorical outline presented in pre­ vious sections suggests that the African slaves were imported via the Miskito Coast, Jamaica, or other Caribbean islands such as Barbados, which served as a stepping-stone for British colonization (Alleyne 1980:23). Once in Belize, those slaves either created ab ovo the language which was to become Belizean Creole, or they may have brought with them some form of English-based pid­ gin or creole. It is also possible that the pidgin developed through earlier con­ tacts between Africans and Europeans along the West African coast (Hancock 1980 assumes a Proto-Creole originating in the Upper Guinea Coast). However, it is practically impossible to ascertain with any degree of accuracy the under­ lying components of Belizean Creole, both because of the complex population movements which occurred in the area and because of the absence of historical linguistic and demographic documentation. Such paucity of data is largely due to the traditionally low status assigned to creoles as the vernacular languages of a long oppressed group. It is one of the tantalizing mysteries presented by cre­ ­le languages that, even though they are very recent languages, so little is known of their precise origins, and thus, practically any interpretation is specu­ lative.11 In this respect, Belizean Creole is no different from other creoles; in fact, its origins may be more difficult to trace because so many ethnic groups were involved in its development. Although the focus of this study is on the cur­ rent status of some selected segments of creole continua rather than on their ori­ gin, it is useful to briefly review current theories of creole genesis, as they illu-

40

CREOLE AND DIALECT CONTINUA

minate possibilities in developmental processes (a comprehensive critical overview appears in Arenas et al. 1995:87-134). Proponents of the monogenesis of pidgin and creole formation hold that all creoles have a common origin. The earlier interpretation assumed predomi­ nantly European (superstrate) linguistic influences, dating back to fifteenth- or sixteenth-century European arrivals. According to this perspective, the ancestor of creoles may have been either a nautical lingua franca or a Portuguese-based version of the Medieval Mediterranean trade language (Sabir), later relexified. The nautical element may have been acquired by slaves in Africa or during their sea voyage between Africa and America, or it may even have been borrowed through contacts with white colonists after their arrival in the Caribbean (den Besten, Muysken & Smith 1995:92-93). Some nautical lexical items are evident in most Belizean lects. Terms now obsolete in Standard English are in common use; in particular, frequent function words are likely to have originated in Scottish or other Northern English dialects, such as yonder 'over there,' pan
THE BELIZEAN SPEECH COMMUNITY

41

contact, as predicted by the bioprogram hypothesis (Bickerton 1981). As briefly outlined in chapter one (2.1), the bioprogram is hypothesized to be an innate set of features which emerge in dramatic contact situations where the learner has no adequate exposure or time to learn a new language and is therefore forced to use his own (biolinguistic) resources. However, most lin­ guists now seem to adopt mitigated positions which involve recognizing that both superstratal, especially dialectal, and substratai influences contributed to the formation of creoles (Mufwene 1986; Rickford 1987 inter alia). This polygenetic interpretation of pidgin and creole genesis posits a diverse origin, assuming that each of the languages involved in colonizing situations deter­ mined the development of related pidgins and creoles. For example, on the basis of the study of several morphological and syntactic features, Corne (1995a) argues for the transfer of substratai Melanesian conceptualizations in the build­ ing of Tayo, a French-lexicon language spoken in St. Louis, New Caledonia, which developed in the last 150 years. Yet, this substrate influence results in an entirely new code: Tayo is "in no way a modification of French, nor a relexification of Melanasian language but. . .it is a new creation which is typologically Melanesian" (Corne 1995a:20), or at least "has been deeply influenced by Melanesian conceptualizations" (Corne, personal communication). This "cre­ ative" interpretation presents a valid explanation of developmental aspects of pidgins or creoles, one that will prove helpful in explaining the emergence of second dialects, such as acrolects in Belize. 3.2.2 Lectal range Defining what constitutes "the creole," even in the specific context of Belize, is a difficult challenge. The term creole is an abstract construct which has no clearly identifiable boundaries. Creoles are typically embedded within complex language continua. For example, the Belizean continuum consists of overlapping Englishbased varieties ranging over a wide spectrum whose extreme poles can be identi­ fied as 1) Creole and 2) English.13 As is the case in other parts of the Caribbean, the labels "creole" (or "pidgin") and "English" do not accurately refer to the var­ ieties involved. They merely indicate some abstract contrast which is part of the competence of the Caribbean speaker. Different individuals may produce quite different linguistic forms that they would refer to as "creole" and "English," depending on 1) sociogeographical factors (exposure to, respectively the local variety and the outside norm), 2) educational factors (the higher the education level reached, the more the likelihood that the English model is known), and 3) psychological factors (language attitudes determine choices among the varieties available). All three types of factors are intimately interrelated. The creole includes various basilects spoken as vernaculars (first dialects)

42

CREOLE AND DIALECT CONTINUA

and is often unintelligible to speakers of a standard English (American or British). In Belize, it is sometimes referred to as "raw" or "broad" creole, espe­ cially by the inhabitants of the main city, Belize-City, when referring, somewhat deprecatingly, to the language spoken da bush—in rural areas. These basilectal varieties are overtly characterized by a limited but distinctive morphology, especially involving special preverbal markers of TMA systems, with an emphasis on aspectual categories. The phonological inventory is also signifi­ cantly different from that of standard English. Furthermore, the syntax presents interesting idiosyncracies in a more covert way, which spread throughout the entire spectrum to the uppermost acrolects, as will be discussed below. The acrolectal end of the spectrum, commonly called "English," also differs in impor­ tant ways from models external to Belize, such as British or American English var­ ieties. Hence, the term acrolect, briefly defined in chapter one as a set of local stan­ dards distinctly different from external models, is fully justified to represent such varieties; although apparently modeled on English, the acrolect displays several non-English features, such as topic structures and serialized verbal forms. A more comprehensive discussion of acrolects is provided in chapter three. Between the basilectal varieties, natively spoken in various degrees by urban and rural Creoles, and the acrolectal end of the spectrum, officially pre­ sented as the norm or linguistic target but interpreted in widely different ways, there is a vague intermediate area called the mesolect, with linguistic and social characteristics of its own. Linguistically, it is identified by an absence of basilectal and acrolectal markers, or it may show redundancy in the use of both types of markers, although these occurrences may be erratic (Escure 1982; 1983b; 1984). Mesolects, which constitute "deviations" from the basilects and are typi­ cally marked by morphological substitutions, are not necessarily indicative of the Creole speaker's step toward the acquisition of the correct English or an unsuccessful attempt at mastering the prestigious standard, although it may be either. In most cases, the mesolect has a special and distinctive value which denotes a rejection of the social values associated with either the basilectal or the acrolectal ends of the continuum. In Belizean social contacts, the mesolect constitutes an effective, subtle instrument of communication in non-peer con­ tacts, whether in in-group or out-group contexts. For example, whereas the "raw" creole is appropriate during informal and spontaneous in-group contacts, and the acrolect reserved for professional situations or very formal speech events, the use of a variety stripped of both creole and standard features is appropriate in semiformal speech events, such as interaction with respected members of the Creole community or with members of another ethnic group, such as the Garifuna (see 4.2.2 below). Thus, the use of the mesolect typically

43

THE BELIZEAN SPEECH COMMUNITY

marks a certain social distance, reflecting either deference, especially with elders, or lack of familiarity, diffidence, even disapproval or dislike, especially with members of a different ethnic group. 4. Focus on the Stann Creek district Since this investigation of Belizean Creole is restricted primarily to the commu­ nities of Placencia and Seine Bight in the Stann Creek district, a southern section of the country, it is relevant to quote the census figures surveying ethnic group (table 2.6) and language distribution (table 2.7) in the Stann Creek district: Table 2.6: Ethnic distribution in Stann Creek district (Pop: 17,477) Creole 25%

Garifuna 36.2%

Mestizo 23.7%

Maya Mopan 6.8%

Kekchi 1.1%

Indian 3.8%

Table 2.7: Stated language fluency in Stann Creek district

Speak English Speak Spanish

Very well 56.2% 26.0%

Not so well 23.5% 9.3%

Not at all 20.4% 64.7%

Of the six Belizean districts, the Stann Creek district has the most evenly distributed percentages of the major ethnic groups. It also has the highest per­ centage in the country of Garifuna people (36.2%) and the second highest of Creoles (25.1%).14 As far as the language situation is concerned, the stated flu­ ency in English is the same in the Stann Creek district as throughout the coun­ try (table 2.4), despite the different distribution of ethnic groups in all six dis­ tricts (table 2.5). The combined Creole, Garifuna, and East-Indian population in the Stann Creek district, amounting to almost 70% of the population of the Stann Creek pop­ ulation, roughly corresponds to the number of individuals claiming excellent or lesser fluency in English (around 80%) and concomitantly lacking fluency in Spanish (65%). The "not so well" category is likely to reflect distinct individual definitions of what constitutes English. Garifuna individuals typically claim that they speak English and not Creole, whereas Creoles are more concerned with the fine distinctions between the different codes in use, as discussed above. On the other hand, the small number of Mestizos in Stann Creek (23.7%), relative to the northern districts, corresponds to the number of people claiming excellent Spanish fluency (with the addition of a few Mayas). Because of the design of the language

44

CREOLE AND DIALECT CONTINUA

questions, it is, however, impossible to determine the respective roles of English and Creole in daily communication and whether Creole or Garifuna people are most likely to consider themselves fluent in English. 4.1 Placencia and Seine Bight According to the 1991 census, rural Stann Creek has an official population of 11,650, showing an increase of about 4,000 during the last ten years. This pop­ ulation increase, however, has not visibly affected the two small Stann Creek rural/coastal communities of Placencia and Seine Bight. The overall increase of the Belizean population is due to immigrants from neighboring Spanish-speak­ ing Central American countries, who usually do not settle on Stann Creek coastal areas. The overall population in Placencia (about 400) and Seine Bight (about 500) has remained remarkably stable, which does not exclude population movement; some individuals emigrate and others return home, thus maintain­ ing the overall balance. Placencia and Seine Bight are located on sandy, coconut-lined beaches about five miles apart on the Caribbean coast, on a long, narrow strip of land running parallel to the mainland and separated from it by a lagoon. Placencia is built at the very southern tip of the peninsula, around a graceful sheltered bay (said to have been a pirates' refuge) punctuated and protected by Placencia Caye, a small island (called a caye/key, from Spanish cayo). This geographical location caused the two communities to be very isolated, and reachable only by a six-hour trip by bus on a dirt road from Belize-City to Mango Creek, followed by a canoe trip across the lagoon. Since 1993, however, Placencia, and Seine Bight to a lesser extent, have begun giving in to the lucrative mirage of tourism: an airstrip near Placencia has been built and a new dirt road now provides direct (two-hour) access from Dangriga along the peninsula, enabling trucks to bring building materials and food supplies. This easier access is changing the appear­ ance and economy of the area, especially in Placencia. The airy wooden hous­ es on stilts, which used to provide safety in times of floods and hurricanes, are now modified to accommodate backpackers. Cubicles for tourists are built on the ground floors of several houses, thus putting an end to a traditional favorite leisure activity of the village, namely, lying in hammocks under the breezy, shady space underneath the house and indulging in gossip with friends and rel­ atives. Increasingly, men abandon their traditionally prosperous fishing activi­ ties to launch into what is perceived as the more lucrative occupations of tourist guides or hotel workers. Such changes are on-going, and there is no obvious impact on the language situation yet, as attested to by this study, but the situa­ tion bears watching. These areas were selected in the late 1970s because of their isolation and the particular dynamics linking Placencia and Seine Bight.

THE BELÍZEAN SPEECH COMMUNITY

45

Placencia has a homogeneous Creole population, which used to function smoothly around a self-contained prosperous economy based on fishing. Men would go away to the cayes along the coral barrier reef for one or two weeks at a time, catching fish or skin diving with special hooks for lobster and returning to the village with a full load, which was then processed and exported through the local fishing cooperative. Women would work at the cooperative and take care of the households and the organization of village life during the men's absence. Seine Bight is ethnically and economically different from Placencia. Its population, one of the original Black Carib settlements from Honduras, is almost exclusively Garifuna, with the exception of two or three Maya families who recently moved there from an inland village. Although fishing and raising some limited crops (in particular cassava) have long been subsistence activities in Seine Bight, they have never been done on a sufficiently organized scale to sustain the local population, as in Placencia. Garifuna men and women either work on neighboring banana or citrus plantations, try their luck in other parts of the country, training as teachers or police officers, or emigrate to the United States, sending paychecks to the older women who typically raise several gen­ erations of children. Another means of support for Seine Bight inhabitants, especially in the 1980s, involved working for the Creoles in Placencia. To do so, Seine Bight men and women would walk four to five miles to and from Placencia each day. Their work included a variety of menial tasks (cutting bush, cleaning fish, washing clothes) which clearly placed the Garifuna in a lower social position vis-a-vis the Creoles of Placencia. Such interactive patterns between Creoles and Garifuna are particularly interesting as they document the fimction of social factors in the development of language patterns (Escure 1982; 1983b; 1984). In the context of Placencia, where Creoles were always secure in their identity and relative comfort, the local Creole language is strongly associated with friendship, solidarity, and folk culture which are more highly valued than the external values of education and refinement asso­ ciated with English. This general sense of security is reflected in the graceful hos­ pitality which I associate with Placencia, and probably explains why the acquain­ tances and friends I made in Placencia over the years were from the very begin­ ning readily cooperative and supportive, though often amused, at my efforts to learn about and tape-record their speech. Since they already had an innate sense of pride in their local history, they were delighted to note an outsider's interest in the language which they always valued in reaction to official efforts to put down vernacular forms as "brokop." They assign strong values to their "mixed" genet­ ics, which they attribute to the hospitality extended by Africans over the years to French and English pirates, then to German, Spanish, and Portuguese expatriates

46

CREOLE AND DIALECT CONTINUA

among others, often through the conduit of their connections with relatives in Roatan, one of the Bay islands of Honduras, and especially the village of West End in that island. Their "hybrid" awareness and identity is displayed in the var­ ious shades of skin, eye, and hair color is evident throughout the village, and not uncommonly in members of the very same family.15 The respondents, whose speech samples I have used in previous publica­ tions and in this work, range across all age groups and types of status. There is no socio-economic discrepancy in Placencia of the type found in urban Western societies (upper, middle, and lower/working classes). There are, however, local status differences linked to history. For example, at least four families hold a certain "status" and most of the land in the village because they were the founders of the community. Yet, a person's family affiliation does not translate visibly into privileged economic or social treatment. Other factors defining local status and respectability include age, leadership in village administration, and personal behavior. Placencia is a tolerant society which accepts personal weaknesses, as long as they do not interfere with the public order. For example, regular drinking is tolerated, but regular unruly behavior linked to intoxication is frowned upon and despised. Similarly, sexual escapades are common and a favorite topic of local gossip. Early pregnancies are not uncommon, and young girls' children are well taken care of by the immediate or extended family; how­ ever, a girl who leaves her children behind with her mother too often while attending the local drinking holes will acquire a bad reputation and may even­ tually be forced to leave. Although the village is nominally Anglican, the little circular church by the sea is rarely attended by more than a couple of older women. The Anglican Church even discontinued the appointment of the resident priest who lived in Placencia in the 1980s. There are, however, a few dedicated "Seventh Day Adventists": one of them, T, now in her 80's, is also one of the most interesting characters in the village, who holds a special status in Placencia. Throughout her life, she was the village's sole healer and midwife, providing the only medical care available, delivering most of the current inhabitants of the village, and administering cures to various kinds of physical and psychological illnesses through every possible means, whether antibiotics, herbs, obeah, or prayers. She is highly respected and loved and was awarded a special decoration by the Queen during the monarch's visit to Belize in March 1995. Her most endearing quality is her special gift for telling stories, with a general predilection for mesolectal forms, which seems to reflect her unique mediating role in the village. Seine Bight is still perceived by the Creoles as an immigrant grouping, without community roots and without status. The Creole attitude toward the Garifuna is typically scornful and supercilious; this is not lost on the Garifuna,

THE BELIZEAN SPEECH COMMUNITY

47

who hold in return a certain resentment toward the Creoles. The fact that the Garifuna are often destitute and without stable economic activity, and in part dependent on the Creoles for their survival, has contributed to this stereotype. The situation has somewhat improved for the Garifuna, some of whom now work as employees in the new small hotels. Today, the Garifuna take the bus instead of walking to Placencia, and they no longer wash clothes for Placencia women, pri­ marily because washing machines have been recently introduced. The last episode in this mutual dependence is linked to the development of tourism. Whereas traditional Carib drums (essentially an African tradition) were previous­ ly scorned by the Creoles and restricted to the village of Seine Bight as a local form of entertainment, Placencia's new entrepreneurs have discovered that tourists like Carib drums and punta (the Garifuna dance); therefore, they hire Seine Bight drummers and dancers to come and perform in Placencia. Thus, Garifuna life has been, and still is, separate but intermingled with Creole life. The distinctiveness of their language has always been an advantage that the Garifuna have had over the Creoles. They can have verbal exchanges in the presence of the Creoles, often at the latter's expense. On the other hand, they are able to communicate in Creole, which they informally learn through their regular interactions and probably pass on early to their children. Thus, children become fluent in Creole prior to primary education and certainly before they are exposed to English. Creole is established early in life as a functional variety with high local prestige value, at least for younger generations, and perhaps also for others. Most older Garifuna people in Seine Bight appear to embrace the official disparagement of the creole language, perhaps because it is a convenient excuse to assert the superiority of their own "real" language; they overtly claim their respect for English as opposed to Creole, which often results in hypercorrected forms of English, laughed at, in turn, by Placencia Creoles who feel they know better. This clearly shows that the language situation serves as an accurate mir­ ror of social and economic conflicts. Yet, there are also those who, like Dermis the teacher, are sensitive to the appropriateness of linguistic adjustments: ". . .1 associate myself wid dem; so I speak Creole, understan'? so nobody can just say he talk funny. . . . " The acceptance and increased use of Creole is reinforced by the fact that the Garifuna have traditionally held a social status inferior to that of the Creoles. Such attitudes are still evident in many subtle ways in group interaction despite the recent trend toward the rehabilitation of Garinagu culture and the Garifuna language. Garifuna holds in-group prestige among its members (usually older) who are intent on preserving traditional cultural values (songs, festivities, wakes, etc.) and who use it for daily interaction in the villages. Yet, Creole can

48

CREOLE AND DIALECT CONTINUA

be heard more than occasionally among Gariftma children, and some young parents encourage them to speak Creole and English rather than Garifuna. Increasingly, Garifuna teenagers seem to favor the creole over their ancestral language, and they may no longer have this unique linguistic advantage. Furthermore, there appears to be a recent, and not altogether positive, link developing between Creole and Garifuna teenagers with the arrival of crack cocaine in the area, the concomitant novel occurrence of crime in a previously peaceful area, and the construction of a drug rehab center north of Seine Bight (which local inhabitants suspect to be in fact a drug distribution center). In 1995, Placencia and Seine Bight have lost their innocence. 4. 2 Samples of Belizean texts The linguistic flexibility available to the Placencia and Seine Bight population in spontaneous situations is illustrated in the following speech samples which represent each of the three general segments of the continuum. Speech data were collected in the context of natural conversations during fieldwork done between 1979 and 1994. The samples provided below are derived from conver­ sations collected in the 1980s. A longitudinal comparison will be presented in the following chapters. 4.2.1 Text 1: Basilect In the following conversation, a Creole woman (S, 43) talks about the profes­ sional future of her son (R, 24) to a visiting friend (E, 26). It is an informal inter­ action, since the participants are long-time acquaintances (R and E are child­ hood friends): 1. S: 2. 3. 4. E: 5. S: 6. E: 7. 8. S:

R. wan tek wan korespondens cos, i me wan tek it befor i kum out i me ds plan fu tek it dat da ti taim de term don, di kos don. a tink i se i wuda me wan kos ina akountin. dat expensive. i kuda me get wan jab da Big Krik, bot dat agen da wan problem. dat da onli paat taim, do. Grif me de tel mi bout da ting, tu. i soun laik da onli wen de banana sizan de aan. Yes da onli den, bot dat de aan wan gud wail, tu.

Translation 1.S: 2.

R. will take a correspondence course, he will have completed it before school is over. He was planning to take it so that by the time the term is over the course will be over too.

THE BELIZEAN SPEECH COMMUNITY 3. 4. E: 5. S: 6. E: 7. 8. S:

49

I think he said he would have done a course in accounting. That's expensive. He could have got a job at Big Creek, but that again is a problem. That's only part time, though. Grif told me about that, too. It sounds like that's only for the banana season. Yes, it's only then, but that lasts a good while, too.

Linguistic characteristics of the basilect represented in text 1 include: 1.

2. 3. 4.

5.

the use of the creole past or anterior morpheme me, which can also com­ bine with the continuous aspect marker de for continuous past, as in i de plan (line 2), with the future marker wan (z me wan tek, line 1), or with a modal for irrealis/past conditional (z kuda me get, line 5); the absence of inflectional morphology, e.g., the nonmarking of tense/ per­ son/number as in i kum out (line 1) 1 6 ; the absence of copula {dat expensive, line 4; da wan problem, line 5); the use of de as locative or existential verb {when de banana sizan de aan, line 8 (literally, 'when the banana season is on' e.g., 'when the bananas are ripe'); and the extensive use of focusing expressions with da/dat, which is used as a deictic pronoun, as in da onli den and bot dat de aan (both on line 8).

4.2.2 Text 2: Mesolect in Creole/Carib interaction In this conversation, a Creole fisherman (R) from the village of Placencia talks with a Garifuna man (A) from the neighboring village of Seine Bight and a Creole/Lebanese man (L) from Belize-City: 1. R: 2. 3. 4. 5. 6. L: 7. A: 8. 9. 10. R: 11. 12. 13. 14.

a se, mai Bobo, we da fran, wen i kom op, i kech some gud, en wen i kom op, en i kom outa di wata ina de bout, di blodi stingre i bring out—gaan, sok it op pa mi breda. di bon went rait in, en i kudn get it. yu no samtaim dey juk yu, i gat it gud. di paizan grupa dey no kom in so kloz, do. yu si, wat ar yu keshin en di lain aal, if yu straik dem wid a haapun put di haapun aal iz, laik, stan op—jos laik bon laik it iz nat di wan dat kal flain fish? di mantare, doz ting ar ogli. Dey hav a saata haan, en dey hav tu red, red blood, red aiz. Bot doz tingz ar ogli. Mi breda se waz daivin out, i de wid iz hukstik, rait, i lukin doun aroun di rak, trai huk op a krefish, rait, en wen i luk op, dis big ting waz rait, rait yu ar fra mi.

CREOLE AND DIALECT CONTINUA

50 Translation 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

I say, my Bobo, in front, when he came up, he had caught something, and when he came up, when he came out of the water into the boat, the bloody stingray that he brought out was attached to my brother. the bone [sting] had gone right in [his foot], and he couldn't get it out. you know, sometimes dey stab you, it was pretty bad. The poison grouper, they don't come so close, though. you see, what you catch with the line and all, if you strike them with a harpoon the way to throw the harpoon is to stand up—just like a spear Isn't it the one they call flying fish? The manta rays, those things are ugly. They have a sort of horn, and they have two red, bloody red, red eyes. Those things are ugly. My brother said that he was diving out there, he had his hookstick, he was looking down around the reef, trying to hook up a lobster, and when he looked up, this big thing was right there, [about as close as] you are from me].

Mesolects are intermediate varieties, linguistically as well as socially (see 2.3). They often constitute a compromise between an overly formal speech and an inappropriately informal or friendly mode, and the linguistic structure of mesolects reflects the delicate nature of such situations. Mesolects can, thus, be divided into three general subtypes: 1.

2.

3.

First, they may include primarily neutral variants, stripped of both acrolectal and basilectal features. This would imply that the mesolect speaker knows of the existence of such features and controls those other marked varieties but has determined not to make full use of them. On the other hand, mesolects sometimes include a combination of basilec­ tal and acrolectal features, which may be interpreted as a certain hesitancy in the choice of the most appropriate code, as is clearly reflected in text 2. There is another possible combination of acrolectal and basilectal variants which leads to redundancy or hypercorrection rather than to alternation as in the second subtype. This variety of mesolect is more likely to indicate insufficient control of both ends of the continuum, or at least of the acrolectal end, and is more typically found in native speakers of languages other than Creole, as is the case with the Garinagu man in text 2.

Text 2 illustrates such sociolinguistic ambiguities. There is a social dis­ tance between A, a fífty-fíve-year-old Garinagu man who works menial jobs in

THE BELIZEAN SPEECH COMMUNITY

51

the Creole village of Placencia, and the two young Creole men, both in their mid-twenties, who had been schoolmates in Belize-City. L, who belongs to the middle-class, is visiting his friend R, a fisherman's son in Placencia. On the one hand, age demands respect for the older man, but his ethnic membership, low paying job, and lack of formal education assign him a lower status. On the other hand, L brings with him the relative status of the "city-dweller." The linguistic choices evident in all three participants reflect the ambiguity of the situation and the conflicting power struggle underlying variants of the mesolect. Thus, R, the first Creole speaker, initially produces a primarily unmarked mesolect (lines 1-5) of the first type, with zero marking for past tense (with the exception of went in line 4) and zero copula but also an absence of basilectal morphemes. The second part of his conversation (lines 10-14) belongs to the second mesolectal type, with the coexistence of English copular forms (lines 10, 11,13, 14), the basilectal copula/locative de (i de wid iz hukstik, line 12), and of zero copula combined with progressive -ing for the aspectual form i lukin also in }ine 12. This apparent inconsistency may indicate an attempt at linguistic con­ vergence, in response to A's language (lines 7-9) which includes three struc­ turally hypercorrect copular forms, such as the preposing of the auxiliary in a cleft construction in line 7, which, furthermore, uses a progressive English form to refer to a habitual action (wat aryu keshin 'what you catch'). Similarly, the absence of copular contraction on line 9 (it iz nat di wan. . .) indicates a hypercorrect level of formality which often occurs in Garinagu interpretation of English. Garifuna phonological interference is also reflected in the form keshin showing the substitution of the palatal fricative for the affricate, since A has no /č/ in his native system. Throughout this sample, and in spite of the various mesolectal interpreta­ tions just discussed, the creole syntax is preserved through focusing devices such as left dislocation (e.g., mai Bobo. . i kech some gud. . ., line 1; di paizan grupa dey no kum in. . ., line 6) and clefting (line 7) as well as adverbial fronting (lines 1, 2, 5). 4.2.3 Text 3: Acrolect The acrolect is usually considered to represent the Creole speaker's best approximation of the standard model, in this case English. Since chapter three is primarily dedicated to acrolects, only a short sample is provided here. It is an excerpt from an Anglican service in the Creole village of Placencia, and the speaker is P, the Creole "priest" (as Anglican ministers are called in Belize): P:

1. The Jews were expecting someone to come to redeem them. 2. But because Jesus did not fulfill their requirements,

52

CREOLE AND DIALECT CONTINUA 3. in other words, they wanted a military rula 4. like how dey had known it in de days of Salomon and David. 5. Yes, dey were expecting, how do you call it, 6. let's call it someone den who would willing to take up arms. 7. In oda wods, de reason why I believe dat Judas betrayed Jesus, 8. Jesus did not, you see, Judas was trying to push Jesus in da corna, right?

A religious service, and in particular a sermon with biblical references, is a situation that requires the most formal style available. However, this acrolectal sample is different from English and reveals occasional Creole interference, especially when the preacher extrapolates from biblical quotations. For exam­ ple, note the absence of the auxiliary in who would willing (line 6), the hypercorrect use of progressive aspectual forms such as were expecting (lines 1 & 5) and was trying (line 8), as well as the frequent phonological substitution of the dental stop for the interdental. 4.3 Garifuna A brief overview of Garifuna is included because its speakers constitute an important minority in the area under investigation, interact frequently with the Creoles, and necessarily learn or teach segments of the creole continuum. There are additional reasons why some aspects of Garifuna must be highlighted. For example, differences perceived in the Garifuna people's English-based varieties (from basilects to acrolects), as compared to the Creoles' versions could possi­ bly be traced back to structural elements of Garifuna; on the other hand, if acrolectal forms produced by the Garifuna are similar to those spoken by native Creoles, they might provide support for a universal interpretation of the pat­ terning of second dialect/language acquisition. Taylor (1951; 1955; 1956a; 1956b; 1958; 1977) and Stochl (1975) provide extensive descriptions of Garifuna, refering to it as Central American Island Carib. It is an Arawakan language closely related to Arawak (Lokono) which, due to the historical events mentioned earlier, has borrowed lexical forms from Carib (Karina), Spanish, French, and English, in that order of temporal sequence and influence. As indicated above, the key event was the deportation of those AfroIndians from the Lesser Antilles, especially St. Vincent and Dominica to the Bay Islands of Honduras (there are still Carib villages on the north coast of the island of Roatan). Some general characteristics of Garifuna that are likely to surface in its speakers' usage of English or Creole include the following. 4.3.1 Phonological features of Garifuna As far as the phonological inventory is concerned, Garifuna lacks the voiceless

THE BELIZEAN SPEECH COMMUNITY

53

palatal affricate, which leads to frequent interference between /š/ and /č/ throughout the continuum, thus producing [f i č] for fish or [d i š] for ditch. This is perhaps the most noticeable feature identifying a Garifuna native to other members of the Belizean community, even if she or he is a fluent speaker of Creole or English. Garifuna is also characterized by the absence of the velar nasal (but has a palatal nasal as well as the bilabial and the alveolar), and of the lax vowels. The latter features are represented in the Garifuna people's produc­ tion of the continuum but do not differentiate them from other ethnic groups, since speakers of Creole, Mayan, and Spanish also lack lax vowels, Spanish lacks the velar nasal, and Creole only uses it occasionally in acrolects. Indeed, the West Indian standard does not contrast lax and tense high and mid vowels (Alleyne 1980; Escure 1978). 4.3.2 Morphological features of Garifuna The morphology of Garifuna includes a wide range of affixes: prefixes and suf­ fixes for personal pronouns, attributive and privative prefixes, plural suffixes for animate beings only, some gender morphemes (see below), and aspect/tense markers that combine similarly with verbs and adjectives. Thus, Garifuna treats verbs and adjectives as a single grammatical class, as illustrated for the adjec­ tive mani 'quiet' and the verb ariha 'see' and in sentences (1) and (2):

Progressive Future

mani 'quiet' maninau 'she is being quiet' manibaun 'she will be quiet'

ariha 'see' narihinau 'I am seeing her' tarihubaun 'she will see her'

1) Bidiba arabaun haruga? Mafuna. Will you go to the bush tomorrow? Maybe. 2) Nichugubai baun ladega uaraguatibu. I'll give it to you because you are honest.

Like Garifuna, but unlike English, Creole uses adjectives like verbs: there is no copula before adjectives (i sick), including contexts with future and past markers: 3) i wan sick [s/he+ future + sick] 's/he will be sick' 4) i me sick [s/he+ past anterior+ sick] 's/he had been sick'

54

CREOLE AND DIALECT CONTINUA

In this respect, Garifuna and Creole function similarly, and Garifuna speakers can be expected to have no problem learning those Creole patterns. However, there is a difference involving the progressive aspectual marker which can mod­ ify adjectives in Garifuna but not in Creole. Possibly related to Garifuna inter­ ference is the Garifuna (more so than Creole) use of the creole continuative marker de before adjectives as well as before verbs (i de quiet versus native Creole i quiet) and the increasing incidence of de before Stative verbs, as in, for example, young Creoles' and Caribs' usage a de tink for 'I think.' 4.3.3 Carib women's language It has been said that Garifuna is one of few languages illustrating structural dif­ ferences between men's and women's speech patterns: women's speech would be related to the ancestral language Arawak whereas men's speech would evi­ dence Karina elements, the language of the warlike Amerindian group who invaded St. Vincent and exterminated the Arawakan men but not the women. Those differences, however, are now minimal and fast disappearing from mod­ ern Garifuna, according to my own observations and inquiries made in the Seine Bight and Hopkins communities. If differences are maintained, they occur in the first and second singular personal pronouns, and in a very limited number of nouns, as illustrated in table 2.8: Table 2.8: Some male/female differences in Garifuna MALE SPEECH

FEMALE SPEECH

au amuru yagu yudi yaugu yagane wegeri irahu wuri iraha gunubu wayunu gunara

nuguya buguya nunine nui nilegen nugune eyeri irahu hinaru iraha huya here winoga

GLOSS

I you my drink my meat my pet my boat boy girl rain land crab tomorrow

The linguistic forms traditionally used by male speakers are receding and are now only used by the older generation, though there are two exceptions: the male speech pronouns (of Karina origin) au and amuru for T and 'you' seem on the contrary to become unmarked, as they are now being increasingly used by young people of either sex, whereas the female Arawakan forms nuguya and

THE BELIZEAN SPEECH COMMUNITY

55

buguya, also meaning 'I' and 'you,' are predominantly used by most other members of the Garifuna community. This seems to be a case of innovation ini­ tiated by the younger generation, and it will be interesting to note if this trend continues. Taylor (1977) also claims that there is an unusual type of grammati­ cal gender marking which depends upon speaker sex: abstract nouns would be masculine in women's speech, and feminine in men's speech. However, Stochl (1975) does not mention this feature, and I have found no evidence of this dis­ tinction in Seine Bight Garifuna. As far as grammatical gender marking is con­ cerned, it is still preserved for some concrete items. For example, the feminine gender is assigned to articles of clothing, buildings, instruments, plants as well as to females, whereas the masculine gender is attributed to body parts, tempo­ ral and seasonal referents (such as 'sun,' 'moon,' 'month') as well as to males. The latter feature obviously accounts for the high incidence of 'he' and 'she' as used for subjects and objects in Garinagu usage of the continuum, in contrast to Creoles who often use 'i' regardless of sex, a reflex of the lack of gender mark­ ing in the basilectal vernacular. 5. Conclusion Given the remarkable multilingual complexity represented in Belize, it is obvi­ ous that a lingua franca is needed in interethnic communication. Although English is officially claimed to fill this role, the reality of language use in Belize presents quite a different picture, as indicated earlier. The closest to a Belizean lingua franca turns out to be the English-based creole used as native language by the Creole group which constitutes the majority of the population in the country. Because of its official status, however, English must also be spoken, and when it is, the resulting code is either an acrolect or a mesolect, which are the normal outcomes of second dialect acquisition. The following chapter will investigate the processes underlying acrolectal speech behavior. Endnotes 1. Come (personal communication) suggests this may be an example of folk etymology. 2. The French word balise "beacon" has also been suggested as the etymology of Belize, per­ haps due to the legendary visits of French pirates. Another possibility is that Belize was derived from Mayan and later hispanicized or anglicized. 3. By comparison, the U. S. population increased by about 10% between 1980 and 1990, (from about 220 to 250 million according to Time 1/30/95:54). 4. Other categories were East-Indian, Chinese, Amerindian, Portuguese, Syrian/Lebanese, White, and Other Races. 5. It would be particularly important to study the current status and usage of Maya and Kekchi

56

CREOLE AND DIALECT CONTINUA

in Belize; this study, however, is restricted to English-based varieties. 6. The capital city of Belize is Belmopan, which is restricted to a small administrative center. 7. The name change was made official after a campaign in the 1970s, the goal of which was to restore pride in Black Carib culture and, generally, to inform Belize's ethnic groups about their respective traditions in reaction to the destructive effects of British colonization. The term Garinagu is the native Black Carib word referring to Carib ethnicity, and Garifuna refers to the language itself. Both terms are now used in government publications. Because the term Garifuna is most frequently used to refer to both the language and the people, it is also the term commonly used in this text. 8. Seine Bight is the village that I have investigated because of its social and economic relations with the neighboring Creole village of Placencia (see Escure 1979; 1982; 1983a; 1983b) 9. Pania is the term locally used to refer to the Latino/Hispanic population. 10. I was personally able to verify this fact during a 1986 trip in Hong Kong. As I was going through customs before boarding the train for the People's Republic of China, I noticed a young Chinese woman standing just ahead of me and bearing a Belizean passport. She explained that she had recently married in absentia a Belizean Chinese and was on her way to visit relatives in Guangzhou (Canton) before emigrating to Belize to join her unknown husband. 11. Corne (p. c.) believes that this paucity of data is due to a general lack of academic interest in exploring the relevant archives and says that exciting new information regarding Cayennais and Louisianais has been found through archival research. He also points out that the origins of several languages are known, in particular Hiri (Police) Motu, Bislama, Tayo, Mauritian, Isle de France Creole, and Reunionnais. 12. For example, haal
Chapter 3 Creole Acrolects as Innovations Abstract This chapter investigates the nature of the acrolectal set of Belizean English varieties and its place in relation to the overall creole continuum. A dynamic view of lectal development is offered as a substitute for the notions of "radical creole" and "postcreole." Acrolects are defined as innovations and extensions of speakers' repertoires. Both the social functions and linguistic characteristics of acrolects are examined in some detail and clearly differentiated from their standard counterparts. 1. Radical creoles, postcreoles, and decreolization As indicated in the previous chapters, the very notion of acrolect presupposes that the creole continuum cannot be defined strictly in terms of the creolized output of a pidginized variety. I intend to show that regardless of the historical accidents which have determined creole genesis, the resulting linguistic situa­ tion is just as complex and multifaceted as any other situation. One important aspect of this complexity is that the creole continuum does not eventually dis­ solve into a standard but develops its own innovative standard, the "acrolect" or set of "acrolects." To demonstrate the independent development of the acrolec­ tal segment, some important issues must be briefly reviewed in order to high­ light the current status of knowledge in the field of creole studies and, perhaps, add innovations to the picture. Two hypotheses predicting crucial synchronic and diachronic aspects of creole situations will be reviewed and evaluated. The first is the radical creole hypothesis which claims that certain "conservative" creoles remain close to their genesis and, thus, reveal substratai influences; the second is the postcreole hypothesis which, to the contrary, states that certain "innovative" creoles reflecting superstratal influences have evolved rapidly to the point of extinction—creoles that have disappeared after yielding to stan­ dardized varieties.

58

CREOLE AND DIALECT CONTINUA

1.1 The radical creole hypothesis One of the widely held assumptions about pidgin and creole situations has been that creoles constitute entities that can be clearly separated from their superstrates or lexifier languages. True or radical creoles are often claimed to be identifiable primarily in rural areas and in older speakers. For example, Bailey (1966:2-3) who concentrated on "the speech of older people and members of the lower levels of the social scale" in her introduction to Jamaican Creole syntax, says: "There is a hard core [of Creole speakers]—the unschooled, ranging from pre-school children to the elderly, with the concentration at either end of the scale—living in isolated villages removed from the centers of culture." This approach has been embraced by Alleyne (1971; 1980), Byrne (1987), Winford (1993), and others who consider Saramaccan, a Surinamese creole with two European lexifier languages (English and Portuguese) to be one of the most "conservative" creoles and, thus, of particular interest to creolistics: "We can ċonstruct a scale starting with Saramaccan and ending with Barbadian English, and passing by varying degrees of retention or elimination of African speech habits" (Alleyne 1971:181). The scale shows at one end some African languages existing virtually intact, and a number of other languages which. . .reveal a high preponderance of structural forms, and less significantly, of vocabulary that can be rather conclusively derived from West African languages. These are, in descending order, Saramaccan and Sranan. . .Krio and Gullah. In the next position on the scale would be located the speech of rural inhabitants of Jamaica, Antigua and Guyana; toward the end of the scale, the speech of the urban working class of these same areas, followed by the speech of the urban Black working class of the United States, and finally the speech of the Black middle class everywhere (Alleyne 1980:18). The view that rural speakers, such as maroons who escaped to the bush, are more conservative because they are sheltered and isolated from the progressive influence of urban centers appears to have been inherited from the field of dialectology which made similar assumptions. In fact, the term creole has come to refer exclusively to basilects, the set of "conservative" vernacular varieties. They were, and still are, the primary object of study in creolistics because they are claimed to best reveal the process of creolization. Those creoles which are thought to be particularly representative of an early stage of creolization have been labeled "radical creoles." For some, a radical creole is assumed to include a higher inci­ dence of substratum features (see Alleyne quoted above), whereas for others a "deep creole," such as Saramaccan, is closest to the core grammar and best

CREOLE ACROLECTS AS INNOVATIONS

59

reflects "the essential structures and categories of the creolization process, and by implication, pertinent aspects of our faculté de langage " (Byrne 1987:15), a position in keeping with Bickerton's bioprogram (Bickerton 1981; 1984). The consensus of what constitutes a radical creole is rapidly evolving as more information emerges about potential source languages. The hypothesis of the African substrate is endorsed by McWhorter (1992) who believes in mas­ sive Kwa/Nigerian influences in the development of Saramaccan serial verbs through gradual creolization as a result of universal principles. On the other hand, Bickerton (1994), a proponent of radical, or rapid, creolization disagrees, arguing that the Bantu population exceeded the Kwa population and, therefore, the substrate hypothesis is not viable (Bantu languages are said to have little evidence of serial verb constructions). Bickerton finds serial verbs in Seselwa (the Seychelles Isle de France dialect with French lexical base), which has developed in the social context of a predominantly Bantu population: If universal principles could bring about serialization in the Indian Ocean, where there was no significant Kwa substrate, then they could well have brought it about in the Caribbean. The substratist explanation for Caribbean serialization becomes superfluous (Bickerton 1989:176). However, Corne et al. (1995) suggests that "the so-called 'serial verbs' in Seselwa are in fact the result of clause reduction in asyndetic coordination." This debate illustrates how data (demographic and linguistic) determine inter­ pretations: McWhorter and Bickerton rely on different stages and types of Suriname demographics, the former now assuming that Saramaccan was shaped on the West African coast before the establishment of maroon commu­ nities, and the latter relying on recently published records of slave shipments (Postma 1990). Corne et al. (1995:1) rely on early Isle de France Creole texts to show that "serial verbs appear to be largely unchanged over time. . .in free variation with semantically asyndetic coordinated clauses, and the result of reduction of such clauses by the deletion of recoverable elements." It would appear that anybody's favorite theory can be at least partly sup­ ported by some segment of the historical and linguistic facts. There is no deny­ ing the fact, however, that serial verbs (a major aspect of the argument that Saramaccan is a radical creole) are widespread in many world languages, and it is possible to claim that serial constructions develop naturally in the acquisition of contact languages. This claim in itself can lead to different and, indeed, opposite interpretations. Bickerton and Byrne believe that serialization, at least in Saramaccan, generated spontaneously and did not originate through influ­ ence from other languages. Corne's interpretation is that there is vernacularization, a "transfer of 'substratai' conceptualizations in the building of a creole lan-

60

CREOLE AND DIALECT CONTINUA

guage" (Corne et al. 1995:23). This approach is appealing because it can explain linguistic similarities which may be evocative of substratai structures (Manessy's (1989) "creole cryptotypes"). Considering that several native African languages were usually represented among speakers creating the creole, the valid scenario presented by Corne et al. (1995:23) is that "only what is in common, a sort of highest common factor, was transferable to the emergent cre­ ­le, giving only a pale reflection of the syntactic strategies of any given African language" (see also Corne 1989). Tayo, a French-lexicon creole language spoken in New Caledonia is the prime example of the rapid (fifty years) transfer of substratai (Melanesian) con­ ceptualizations in the building of a creole language with some French syntactic influence (Corne 1994; 1995b). This scenario is further supported by the fact that constructions akin to serial verb structures occur widely even in the ver­ nacular, nonstandard, or colloquial forms of many languages in which serial­ ization is not formally recognized, as shown below from a small set of English and French sentences. In these sentences, it may be difficult to. differentiate parataxis from serialization, or it may be the case that these phenomena coa­ lesce in the pragmatic context of syntax: (1) English: Come get it; Lemme do it.1 (2) French: Ils sont partis se cacher. (They ran away to hide.) Elle va se faire attraper. (She is going to be chided.) Je me suis fait couper les cheveux. (I had my hair cut.) Il est allé courir acheter du pain. (He ran out to get bread.) Va te faire voir chez les Grecs. (Go to hell.) (3) Belizean acrolect: Dey make a ting call 'em casarina. Another approach to the definition of radical creoles involves surveying a large number of features that may yield deeper insights into possible influences on creolization. For example, although Haitian and Saramaccan were considered to be "on a par as radical creoles," Muysken (1994:305) claims that "Haitian stands much further away from potential African source languages than Saramaccan." This opinion is based on observations that Saramaccan includes African (Fongbe) substrate features that Haitian does not have: pre-nasalized stops, co-articulated stops, incorporated articles, paragoge, or agentive forma­ tions inter alia. On the other hand, features which may also be traced to African substrate influences, such as serial verbs, predicate clefts, postposed determin­ ers, and postposed adverbial deictics, are shared by Saramaccan and Haitian. A related hypothesis holds that the separation is so sharp between the creole (radi­ cal, deep) and its superstate that there is no intermediate set of varieties or mesolects. The hypothesis of distinctive lects appears to be an unwarranted

CREOLE ACROLECTS AS INNOVATIONS

61

extrapolation from the data, which does not stand careful scrutiny and obscures the constantly changing nature of the continuum. The same static perspective is reflected in the common reference to "creole" or "basilectal" speakers versus "standard" or "acrolectal" speakers and the implication that they are mutually exclusive. Bickerton (1975:24) summarizes that view in the following manner: We will begin by assuming that speakers who use features such as the tenseaspect markers bin, a, don, doz etc.—features which are not found in any acrolec­ tal speaker—may be basilectal speakers, and that the features mentioned may be basilectal markers. If our assumption is correct, we will find that those speakers who use such features most frequently will use acrolectal features seldom, if at all; and that, conversely, those speakers who use acrolectal features with high fre­ quency will use basilectal features infrequently, if at all. Bailey (1966:1), however, while identifying distinct codes in Jamaica and restricting the tenn creole to the extreme nonstandard end of the spectrum, also demonstrates her awareness of the ambiguous linguistic situation: A given speaker is likely to shift back and forth from Creole to English or some­ thing closely approximating English within a single utterance, without ever being conscious of this shift. Most observers of language in Jamaica have encountered extreme difficulty in distinguishing between the various layers of the language structure, and indeed the lines of demarcation are very hard to draw. Clearly, Bailey states that Jamaicans can be at the same time basilectal and acrolectal speakers. In spite of her insights, her book focuses exclusively on basilects and outlines a static grammar of Jamaican Creole, which paved the way for the circularity and confusion that has contributed to limitations in the development of the field of creole studies; for the most part, artificial labels have too often been imposed on a continuum that cannot be clearly separated into dis­ tinctive codes. DeCamp (1971:350) identified the weaknesses of this approach, saying that "there is no sharp cleavage between creoles and the standard." 1.2 The postcreole hypothesis Another assumption related to a unilateral view of creole situations is that many sites where pidgins and creoles had developed, under the impetus of social con­ flicts such as slavery in the New World, have now reached "the final stages of the creole life cycle," or "postcreole" stage (DeCamp 1971:349). DeCamp was the first to coin the phrase postcreole continuum to identify the range of lin­ guistic variants he had observed in Jamaica. He defined postcreole communi­ ties as "communities in which is in the process of merging with a standard" and

62

CREOLE AND DIALECT CONTINUA

refers to them as "a complex but related set of switching activities, all triggered by the presence of one stylistic feature [+oratorical]" (DeCamp 1971:349, 353). Those valuable insights into the concept of variability have been restricted to imply a linear developmental process of change. By "postcreoleness," later interpreted as "decreolization," the traditional view has assumed that the origi­ nal creole is in the process of disappearing and is being replaced by the stan­ dard language holding prestige in the area, the process being facilitated by increased educational opportunities and social integration. The alleged linear progression in the direction of the overtly valued model would lead to the death of the original creole—hence the claim that a radical creole could only exist if kept separate from external sociolinguistic contacts. Decreolization has, thus, come to mean the loss of the creole, as clearly stated in the following: The result of decreolization is to create a continuum of intermediate varieties between creole and superstrate. If the process is sufficiently long and intense, the continuum may be progressively eroded at its creole end. The result may be a syn­ chronic state in which the most conservative variety recoverable is already con­ siderably different from (and considerably closer to the superstrate than) the orig­ inal creole (Bickerton 1981:47). The Trinidadian and the Bajan/Barbadian situations are often cited as prime examples of decreolization. 1.3 Problems with radical creoles and postcreoles The radical creole and postcreole hypotheses appear to be mutually exclusiveor to apply to two mutually exclusive types of communication—because they correspond to two types of assumed creole histories: the former occurring in isolated, usually rural, communities and the latter in integrated, usually urban, communities. Although it is expected that creole societies have been exposed to differing degrees of external influences, it is unlikely that they would be so strikingly contrasted in tenus of their linguistic development. First of all, a prac­ tically static community preserving a conservative creole would be highly dys­ functional, and it would be in direct contradiction with what is known of lan­ guage history. Secondly, it is unrealistic to assume that there ever was a stable creole, considering the highly unstable conditions which gave rise to pidgins and creoles. A more satisfying scenario would posit that the initial stage of creolization operated in terms of a developmental continuum from the very start. The notion of an original creole is a vague concept at best because there is no evidence of earlier stages beyond sparse written fragments. In fact, it is possi­ ble, and very likely, that there was a wide-ranging continuum from the begin-

CREOLE ACROLECTS AS INNOVATIONS

63

ning of African-European contacts. Alleyne first argued in favor of the position that creoles were only a segment of a continuum of variation, showing evidence of "considerable variation from their inception rather than early and rapid crys­ tallization (Alleyne 1971:170). These [intermediate] forms existed from the very inception of bilingual con­ tact they were generated by that contact situation in the same way as the other socalled "pidgin" or "creole" forms. . . .they derive from an underlying protodialect, a proto-mtermediate Afro-American. Hence, they were generated in the contact situation at a very early stage when some segments of the African popu­ lation began to be assimilated by rigorous acculturative processes (Alleyne 1980:182-3). Regardless of the original departure point of the creole continuum, which is a matter of speculation in the almost total absence of hard evidence,2 the position of mesolects within the continuum has been variously interpreted. The standard view has been to refer to a mesolect as an intermediate or imperfect approximation of the model variety reflecting the overall progression toward the acquisition of the target language. In this sense, mesolects constitute a tem­ porary stage, not unlike the interlanguage stage typically associated with sec­ ond language acquisition. The adoption in SLA research of the equivalent terms basilang, mesolang, and acrolang obviously signals the perceived linear simi­ larities between the two linguistic situations (Stauble 1980:47; Schumann & Stauble 1983:263). This one-way upward shift is interpreted as implying a lin­ guistic restructuring of the creole system, with the replacement of basilectal morphemes or structures by their standard reflexes. 1.4 A dynamic view of lectal development A differing view, and the one to which I now subscribe, holds that mesolects and acrolects are not temporary stages in the acquisition process but constitute separate, overlapping varieties which are assigned distinctive psychosocial functions in the overall system, display their own linguistic characteristics, and usually co-occur in a single speaker's competence. Rickford (1983:304; 1987) notes that "[i]n the decreohzation process, individuals retain the competence in the basilect (or any other lects) that they had at a previous stage, and add to this their newly acquired competence in a higher lect." I also noted (Escure 1981; 1982:247-252; 260) the extraordinary flexibil­ ity and code-shifting ability evident in two Belizean villages. I observed that Creole speakers of Placencia moved away from the basilectal values of five diagnostic morphosyntactic features when they found themselves in relatively informal situations, especially in intergroup contexts involving contacts with

64

CREOLE AND DIALECT CONTINUA

Black Caribs. For example, the same speakers showed evidence of past tense marking 63% of the time in intergroup situations, whereas they produced only traces of past marking in intragroup basilectal contacts. All features examined functioned in similar fashion (marking of third-person singular present mor­ pheme, use of do in negative forms, copular usage, and plural morpheme). To describe this adaptive process, I used the term synchronic decreoliza­ tion, which refers to the observed intra-individual variation, with the intent to signal that no definite loss of the creole basilect is implied by the absence of a basilectal feature. In the creole situation, learning the standard does not entail a monolithic progression toward the target language (with the replacement of old, linguistic forms by new forms) but, rather, the development of a broader reper­ toire, allowing for the co-existence of native and acquired forms in a speaker's competence. Creole speakers have been found to extend their repertoires, at least in English-based creole continua, rather than keep them constant through the sub­ stitution of mesolects for basilects and acrolects for mesolects. In other words, mesolects do not constitute just a temporary interlanguage stage. This is not to say that all speakers extend their repertoires to the same degree. Only repeated observations of the same speakers in different contexts and with different participants, and an extensive knowledge of the community, can lead to a reasonably accurate assessment of individual and community com­ petence. Previous claims that decreolization produced postcreole continua were based on isolated and, thus, limited lect observations, which did not permit a full comprehension of the extent of code-switching available to individuals; therefore, a thorough investigation of all linguistic varieties is the first crucial step in any study of creole (and any other) situations. 1.5 Decreolization as acquisition Instead of defining decreolization as the gradual loss of the creole basilect lead­ ing to a postcreole stage, the term decreolization will be used to refer in its broadest sense to the acquisition of non-creole forms, the learning of linguistic forms which are more closely related to the superstrate. This process, which involves the acquisition of new codes, results in the possibility of extensive code-switching triggered in diglossic situations. This definition allows for the preservation of a (possibly separate) basilectal creole system, although not nec­ essarily in its original state: (1) Decreolization: the extension of the basilects to include linguistic forms derived from the superstate, and resulting in mesolect or acrolect acquisition. According to the interpretation proposed, decreolization would not be unlike second language acquisition, in which it is generally never claimed that

CREOLE ACROLECTS AS INNOVATIONS

65

L2 replaces L1 unless the L2 learners are displaced, isolated, and never again exposed to L1. It has been said, however, that one important area where creole situations differ from SLA is that SLA implies the replacement of earlier forms by new ones; acrolect acquisition does not and, instead, involves "extensions of one's linguistic repertoire from an earlier stage" (Rickford 1983:304). Selinker (1972), however, mentions "backsliding" phenomena in SLA. There is ample evidence that SLA, if studied in a natural context and not in the lab, is no more irreversible than acrolect acquisition, or for that matter, any second dialect or language acquisition, except in cases of early L2 acquisition leading to com­ plete bilingualism where there is no clearly dominant language. When assessing the amount of attention paid to the creole continuum, basilects (or whatever variety is assumed to be basilectal) turn out to have been the primary object of study. It is only recently that questions have been raised about mesolects (Rickford 1987), especially in relation to their role as dynam­ ic elements of the continuum instead of errors of interpretation. As far as acrolects are concerned, however, they have been practically ignored. Consequently, the next section will attempt to identify social and linguistic characteristics that may justify the existence of the specific term acrolect. 2. Acrolects and standard dialects: Social aspects The fact that acrolects have so far raised little interest in creole research is prob­ ably due to the underlying assumption (though never actually discussed) that acrolects are identical to the standard variety of the lexifier language, and that they are the successful outcome of the acquisition process which is the educa­ tional target in Creole societies. Yet, the very use of the term acrolect as opposed to merely standard variety would seem to indicate that there is, indeed, a per­ ceived difference between the higher lect in the continuum and the standard. 2.1 Acrolect and standard defined The assumed equivalence of an "acrolect" and its corresponding "standard" has been largely untested so far. Most observations indicate, to the contrary, that acrolects are not identical to the target language but that native speakers of basilectal creoles produce their own version of the target language. This does not imply that speakers are "unsuccessful" in their attempt at learning the stan­ dard. Their goals may simply be different and it seems inappropriate to assume that acrolectal patterns are approximations of the target language. It may be more relevant to refer to acrolects as innovations. If the notion of individual or community goals and choices in language acquisition is taken into account as a

66

CREOLE AND DIALECT CONTINUA

significant aspect of acrolectal development, then the crucial question is why acrolectal speakers would choose, albeit unconsciously, to produce varieties dis­ tinct from the official standard. Psychosocial aspects of the differentiation are par­ ticularly relevant here and will be investigated more specifically in the context of Belizean society. As a starting point, it may be helpful to refer to definitions of the terms acrolect and standard before further discussing the situation. The term standard is typically defined in composition manuals or gram­ mars, and such definitions are likely to influence, more particularly, teachers as well as official language policies. The normative view derives "correctness" from the written and literary "educated" medium and assumes that spoken var­ ieties should match the written variety; therefore, a grammarian's definition of the standard assimilates spoken and written varieties. Linguistic definitions are careful to separate written and oral varieties and refer to social class as instru­ mental in the arbitrary determination of the official model. Both types of "stan­ dard" definition are offered below. (2) Standard: a) written and spoken language of educated people (Leggett, Mead & Kramer 1988:268); b) a dominant or prestige dialect (Fromkin & Rodman 1983:251). On the other hand, the term acrolect has only been used in creolistics but always as an equivalent of the standard target language, although one may won­ der why a different term would be needed under those circumstances. Bickerton (1975:24) uses the term "acrolect to refer to educated Guyanese English (a variety which differs from other standard varieties of the language only in a few phonological details and a handful of lexical items)." My observations indicate that there are ample reasons to avoid equating "standard" and "acrolect." Acrolects may be defined according to two distinct sets of criteria, either in terms of pragmatic use (and more specifically in relation to the speaker's overall repertoire) or in strictly linguistic terms. First, an acrolect may be assigned a pragmatic, discourse-oriented definition in terms of its speakers' intents and abilities. An acrolect would, thus, be the highest variety an individ­ ual is able or willing to produce in the relevant formal contexts. This definition implies that an assessment of the whole range of a speaker's competence has been made, and this in turn requires extensive knowledge of the individuals under investigation. It might be more revealing to investigate a few individuals exhaustively in as wide a variety of situations as possible than to acquire single lects from a large number of individuals. This method is likely to provide insights into linguistic processes that would not be possible from a more restricted set of samples. Secondly, an acrolect may also be defined in terms of its strictly linguistic

CREOLE ACROLECTS AS INNOVATIONS

67

characteristics and by comparison to basilects and mesolects within the same continuum. For example, basilects are relatively well differentiated from the other varieties, in terms of the incidence of certain morphemes, such as the use of basilectal de (for copular forms) or ms (for past) versus mesolectal null forms or inflectional past marking on verbs; however, the boundaries, if any, separating mesolects from acrolects are extremely difficult to establish (Escure 1982; 1983a; 1983b). Consequently, the following definition of acrolect is pro­ posed, which characterizes the resulting effect of the decreolization process as defined in (1) above, but allows for variability in its actualization. (3) Acrolect: A non-native version of the standard language, which is acquired through spontaneous or guided learning, functions in formal contexts, and extends its speaker's repertoire without necessarily leading to the loss of the speaker's vernacular. The acrolect is not necessarily a dominant or prestigious dialect, although it is usually associated with education. The definition of the term acrolect implies that there is an available standard that provides a linguistic model instrumental in the formation of acrolects, even in the absence of any guided learning. In Belize, the official English standard is also the superstrate, the lan­ guage which has yielded the lexical base of the creolizing varieties. All Belizean citizens know that the official language of Belize is English and that they are "supposed" to speak English, regardless of their ethnic identity as Creole, Garifuna, Mestizo, Maya, or Kekchi (see chapter two). 2.2 The label "English" in Belize It is questionable, however, whether in Belize Standard English is either "dom­ inant" or "prestigious," as stated in the definition given above. The general label "English" is extremely confusing to Belizeans, primarily because English is an external standard, never physically present, although it is officially proclaimed to be the language of Belize. English is not spoken in Belize by any particular group, excepting a handful of immigrants, including a few American families who emigrated to Belize after the Civil War; there are few British subjects liv­ ing now in Belize. Due to the recent status of Belize as a British colony (the colony of British Honduras), and because it is still a member of the Commonwealth, it is tempting to assign the official English variety to British English or RP, but the influence of RP seems to have been slight. There were isolated recent contacts with some varieties of British English, when in the late 1970s Britain maintained a defensive force in Belize to prevent a potential Guatemalan invasion. Most British soldiers stationed in Belize happened to be

68

CREOLE AND DIALECT CONTINUA

speakers 'of northern or Scottish English or were Gurkha soldiers from Nepal and, thus, speakers of a Nepalese L2 version of English. Contacts with those British soldiers were definitely not conducted in RP and, in fact, provided access to other nonstandard varieties of English. Interestingly, this situation may have reenacted the early contacts of the Belize settlement between Africans and nonstandard speakers of European English. The only direct access to native renderings of a standard English variety appears to be through radio programs: BBC news and American evangelical broadcasts. Standard English, then, has practically no spoken presence within the country of Belize, though the written medium is, indeed, standard English. 2.3 British and American values in Belize In spite of the absence of direct access to standard English, there are general­ ized official and unofficial pressures to learn it, but not all individuals respond to the pressure, or those who do, do it in highly variable ways. It is well under­ stood that mainstream values or the prestige of a variety external to the com­ munity do not constitute the only, or even the major motivation, for change through dialect acquisition. First, RP is associated with a remote culture, one which does not have realistic value within the Belizean context. Although most Belizeans do not seem to hold strong negative feelings toward the language of the early colonizers, and even appear to preserve some allegiance to England (many Belizeans still keep in their homes portraits of Queen Elizabeth and of other members of the royal family), English/British values are simply irrelevant to Belizean life, and are as quaint and exotic amidst Central American mangrove as the royal hats, Big Ben, or London smog. In contrast, English/American val­ ues are much more meaningful, in part because of the geographical proximity and the economic reality of job opportunities in the United States. Belizeans nowadays are more likely to have had contacts with American English since many have spent some time working in the United States or have relatives there. Many Belizean children are supported by the money orders sent home from Houston or New York to hard-working grandmothers who com­ monly raise three or four generations of Belizean babies. Those generations of Belizeans working in the United States return home sometimes, usually for Christmas or when they retire, and they bring back gifts, including their ren­ dering of American English. However, American English does not necessarily carry with it the prestige associated with the economic benefits assumed to result from living in the United States (although the reality often falls short of what homecoming emigrants are willing to admit). American accent affecta­ tions are jokingly disparaged as "taakin amerikaan." In reality, the Belizean emigrants' contacts with American Standard varieties are determined by the

CREOLE ACROLECTS AS INNOVATIONS

69

restricted opportunities available to aliens (often illegal) who also happen to be mostly black. Belizeans have more contacts with speakers of Black American English than with speakers of American standard varieties. Thus, American influences are likely to penetrate creole via the Black English medium, rein­ forcing nonstandard forms of English such as multiple negation. The real world in Belize is one in which the creole, the mesolect, or some other language is spoken, and the vernacular variety has a high prestige at home, not necessarily of the covert type. Although they are aware of the official condemnation of local varieties, most Belizeans generally feel comfortable with their native basilect and take it for granted. 2.4 Range of linguistic choices and attitudes There is a wide range of attitudes regulating the appropriateness of linguistic codes. For example, Kim L. a twenty-five year old urban Creole man proclaims his preference for what he identifies as an intermediate variety—"de creole and English mix togeda"—which he refers to as "broken English": (4)

. . .I like the broken English, I don't like speaking creole, sound too bad.. .you see, you get people who really speak de creole really really bad, you know. Sometime taakin an you don really unerstan what dey sayin, right?. . . I can speak creole myself, you know, but I don like it. . . When I first staated, staated out creole right? After a while I, now I try get out a it for a while. So keep out a it up to now. [10.1: 2]

Kim's "broken English" may in fact be easily interpreted as an acrolect. Having spent some time in the United States, Kim is probably aware that the variety he favors is not standard English, and he therefore identifies it as a mesolect. His characterization of his own variety highlights the ambiguous, adaptable nature of the Belizean continuum. Also extremely revealing in sever­ al respects is his friend Rick L.'s insinuation that Kim switched away from the creole solely to appeal to "hoity-toity girls": (5) . . .he use to mess with da kinda girl, dey kinda haity-taity girl, right? You wan taak English, you raas (expletive) exactly we happen to a Carib. [10.1:4] (He used to date, you know, classy girls. If you speak English you'll sound just like the Garifuna people.) The first part of Rick's comment suggests some important social reasons underlying the acquisition of noncreole forms: one of them is impressing edu-

70

CREOLE AND DIALECT CONTINUA

cated young women. To younger Creole men, in particular, the use of English is a commodity and is portrayed in sample (5) as part of a con act, with the ulte­ rior purpose to obtain a date with a rich girl. The implication is that the language of education has prestige with wealthy girls whose parents can afford to send them to a private school run by nuns, and that there is an obvious equation between money and English. For that reason, English has no inherently signif­ icant value for the average Creole individual. The prestige of English in Belize is strictly reserved for a small privileged segment of Belizean society, and thus, it is no more than an occasional commodity, deprived of local identity and like­ ly to be perceived as an illusion, a parody, and a trap for those who fall for the externally imposed standard. 2.4.1 Language attitudes and gender The claim that women especially value the standard language has been made before in relation to various linguistic communities. In the Gullah-speaking community of South Carolina, women were found to learn English more fre­ quently than men because the working opportunities for women (as teachers and nurses) require interaction with a standard-speaking environment on the mainland; men get better jobs working in construction or other manual labor on the island where the English-based creole (Gullah) is widely spoken and holds local prestige for them (Nichols 1983). Thus, women, but not men, are moti­ vated to learn the standard variety. In this situation, the acquisition of the stan­ dard variety is sex-based because, according to Nichols, men and women are offered different economic opportunities. However, in my opinion, women's choice of standard English varieties does not mean that women abandon their native vernacular as they learn English; in fact, the issue of creole loss is not overtly mentioned in the context of the South Carolina Sea Islands, and it could be implied from the absence of such a discussion that Gullah is lost by women, and that "decreolization," in the traditional sense, is implemented by them. Furthermore, it is assumed throughout the study, though never actually stated or discussed, that those women who learn standard American English actually speak in a manner identical to that of native speakers who are not members of the Gullah community. This, of course, is highly unlikely. On the other hand, Edwards' study of a segment of Detroit's urban Black English community found the opposite: women use more vernacular (Black English) variants than men. For four of the five linguistic variables used in this study (four phonolog­ ical variables and the copula), females chose more of the BE variants than males. This is attributed to the fact that the women living in that neighborhood are less socially mobile than men (Edwards 1992:104-5). In the Belizean rural context, middle-aged women who are most actively

CREOLE ACROLECTS AS INNOVATIONS

71

involved in village activities were found to be more likely than men and other women to use a wide range of varieties, from basilectal to acrolectal (Escure 1991b). Those older women do not give special preference to the standard vari­ ety but show flexibility and discrimination in their linguistic selections, which reflects their sensitivity to the needs of the community. There is also some evidence that women in the same Belizean village are directly involved in linguistic diffusion and that they accurately reflect the ten­ dencies of ongoing linguistic change. For example, in a study of copular varia­ tion, it appeared that when they use basilects, women preserve more than men the conservative features that are also strong vernacular variants, namely, the pre-verbal de morpheme and the pre-adjectival zero-copula. On the other hand, women also accentuate innovations when they use acrolects, in particular the loss of the locative verb ds and the pre-nominal zero-copula, which is increas­ ingly replaced by the English be copula (Escure 1993c: 128-9). This study con­ firms the variability existing in individual repertoires, as discussed above, as well as the multidirectionality of the linguistic development in creole situations. It shows that individuals (especially women) can elect to be more conservative or more innovative depending on their assessment of the situation. In the Detroit black urban community, as well as in the rural context of Belize, age, sex, and participation in community activities, or isolation from external groups, are found to strongly stimulate allegiance to the local basilec­ tal vernacular. The Sea Islands Gullah women who are now involved in main­ land activities show growing usage of English. The choice of prestigious (offi­ cial) or stigmatized (local) varieties may be more related to social variables such as class and status than to gender per se. Such patterns are not unique to postcolonial or immigrant communities. Milroy (1980) has shown that the con­ cept of "social network" plays an important role among Belfast working class groups in terms of preserving working class varieties of Irish English as opposed to the adoption of Standard English. 2.4.2 Language attitudes and ethnicity Interethnic conflicts are also revealed in the latter part of Rick's statement (5), which can be less literally translated as "You try to speak English, and you end up sounding like a fool, just like the Caribs." The last reference points to con­ flicts between the Garifuna (Black Caribs) and the Creoles, who live in neigh­ boring villages in southern Belize (see chapter two). They are traditionally involved in minor bickerings in which language plays a major role. The Garifuna overtly scorn the creole and apply themselves to the acquisition of "good English," although they mostly use the creole in daily activities. On the other hand, the Creoles snicker at the Garifuna's second-language versions of

72

CREOLE AND DIALECT CONTINUA

both creole and English. The Garifuna people, however, keep a step ahead of the Creoles in the linguistic arena, since they speak a native language that the Creoles do not understand. The Gariftina people's mixed reactions to the creole are often shared by other ethnic groups. Some individuals claim that they do not understand the creole, and others say that they speak it when they have to, such as Martin D., thirty-eight, the Gariftina teacher quoted in chapter two who speaks creole with his friends "so nobody can just say he taaks funny he must be from a different land" (3.1: 10). He mentions that his reason for not liking to use creole (although he does) is that it has no rules, a belief unfortunately shared by many Belizeans: (6) Now the reason why I would say dat because de Creole—de Carib you can follow the trend, right, because looking at it grammatically dere are rules for use of words and so, right?—but I think in Creole, I just cannot under­ stand, you know. Unless if dey can, but so far I haven't seen, I couldn't write a book in Creole dat's a language book, you know. (3.1: 9) The same view is shared by Peter ., a man of mixed Creole/Hispanic descent with a college education, who works as a civil servant in a ministry in Belmopan, the official capital of Belize. Peter states a strong preference for his "Spanish" roots rather than his Creole origin. He says: (7) Creole is notin to be proud of really, you cannot be, I mean, I grew up in de ting. I be frank wid you, dere is notin to be proud of in creole, I tell you de trut. I am from my moder's side, I'm African descent, right; in my fader's side, I'm Spanish descent. Now, if I was to pick de two evil, I pre­ fer de Spanish, not because of color but because of de moral fiber of de people. (9.2: 6) This quotation suggests that cultural attitudes may be extremely important in determining allegiance to any given variety. The ethnophobic self-depreca­ tion evident above reflects the lasting undermining effects of colonization, per­ haps its most devastating legacy, which still shapes some individuals' attitudes toward their peers and their languages. Although the group of Creoles is, by def­ inition, mixed with combined African and European ancestry, Belizeans often equate "Creole" with "African," as in the above quotation. They also classify as "Spanish" or "Panya" any Central American Hispanic group (mostly Mexican, Guatemalan, or Honduran) which may have interacted with Belizeans after emancipation; indeed, there is a Spanish/Mestizo majority population in the

CREOLE ACROLECTS AS INNOVATIONS

73

northern and western parts of the country, near the Mexican and Guatemalan borders, as discussed in chapter two. Such statements addressing the putative "no rule" status of creoles capture an awareness of the changing nature of creoles. Colonization has naturally con­ tributed to the pervading stigma attached to the languages spoken by slaves and their descendants in the Caribbean. Yet, in spite of the overt stigma attached to the basilect and its speakers' apparent denials of its value and existence, creoles are still alive and well and, in fact, gaining popularity with certain groups. Similar attitudes are observed universally as languages and dialects derive their status (high or low) from their speakers' degrees of status; however, the reverse is not necessarily true. A speaker does not automatically gain social status through the use of a high status dialect. It is also crucial to take into account the often subconscious perceptions of individuals who accurately assess their social position and their chances for upward mobility, balancing the potential socioe­ conomic gain involved in the acquisition of the standard against the learning effort required and the potential loss of identity incurred by estrangement from the native vernacular, albeit temporarily. It would be naive and presumptuous to assume that all members of a given society subscribe to official linguistic policy and agree to embrace mainstream values and adopt the standard variety selected by educators, academicians, or planning authorities. 2.4.3 Acrolects as psychosocial entities The reasons why individuals choose to learn or not learn a variety are complex, as briefly previewed in chapter one, and are documented in an increasing num­ ber of case studies exploring the complexity of the psychosocial mechanisms involved. Acrolects in creole societies, particularly in Belize, are different from the recommended official standard for two major psychosocial reasons. First, acrolects develop independently because the officially prescribed standard is not present in the country; there is simply no linguistic exposure to a consistent British or American standard of English, either in daily activities or in school contexts, as indicated in chapter two. Secondly, Belizeans are not strongly moti­ vated to learn that remote standard, precisely because there are no cultural or psychosocial links to standard-speaking groups. Since acrolects do not develop according to a specific model, they are highly variable, reflecting the multidimensionality of human experience. 3. Acrolects and standards: Linguistic aspects The high variability of acrolects is a phenomenon triggered or facilitated by the co-existence and mixing of lects. Individuals may have widely differing reper-

CREOLE AND DIALECT CONTINUA

74

toires, due to social and psychological factors (actual degree of exposure to the standard, degree of education, type of model presented by teachers, group or individual identity; personal ambitions, inter alia). Indeed, one person's acrolect may be another's mesolect. Lects cannot be neatly separated from their superstrate and from each other, but they are perceived as being different from the codified superstrate, and they are almost certainly intended to sound differ­ ent, as indicated in statements 4, 5, and 6. Although most Belizeans widely use the word English, this term clearly denotes to them a local standard variety which differs both from American and British varieties. It now remains to iden­ tify how the distinctions between Belizean acrolects and English superstates can be linguistically characterized, if at all. It is essential to verify whether the superstrate/acrolect difference is real—actually represented in linguistic terms, beyond psychosocial factors. In the following section, some of the most promi­ nent acrolectal features will be briefly examined without speculating as to the underlying reasons for the putative acrolect/superstrate distinctions. The two texts selected below illustrate some acrolects commonly found in Belize. Both are excerpts from conversations I had with the speakers at an early stage of my fieldwork.3 In the creolistic tradition prevalent in the early 1980s, my first reaction had been to discard the texts as noncreole, too English, and, therefore, uninteresting because they appeared to reflect the creole speakers' "highest" available varieties, produced for my own benefit and as a courtesy to a stranger ignorant of the local creole. I have since come to realize that the "English" versions produced by those speakers of creole offer invaluable insights into the interactions of the creole and the standard varieties and help us understand how acrolects develop. I later found that "foreigner talk" as proffered to a stranger was not truly different from the acrolects that later occurred in for­ mal situations in intragroup contexts, such as village council meetings or Sunday church sermons and in out-group contexts with other Belizean ethnic groups. 3.1 Acrolectal texts 3.1.1 Text 8: Local food (Peter, 40; Creole, Belize-City) 1. 2. 3. 4. 5. 6. 7. 8. 9.

I will agree one of the things about a fisherman is that they don't have a very good diet, you know, is one of the things we don't bother about that in Belize, you know, we eat anyting we like. But what happened, we used to have all things that sort of compensate for vegetables, you know, like callalou. Yeah, some kind of cabbage, it look like cabbage, but it doesn't grow like cabbage; I didn't eat it for years, but I saw it, and I think it's in Sartineha a couple months ago, dey grow it up dere, in Sartineha village, in the naat (North),

CREOLE ACROLECTS AS INNOVATIONS 10. 11. 12. 13. 14. 13. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

yeah, its a green vegetable, right. The Caribs, dey have different kind of cassaba, you know, dere have two different type; you have de type, de Caribs, dey use plenty, which they use it for all different type of things; dey use it for -um-dey make a thing call em "casarina," which is a -it's a dried and process; the cassaba were strain, it comes into a kinda powder like -um-coarse powder; you heat up de water, right, and you put de casarina in dat water and you stir it and it comes like a powder, somting like dat. You know, it's a syrup, it's strain' to be very rich, you know; I tink de Americans dey call it "Farina," I tink. (Indians) dey call it plantations -milpas- These are farms where they cut down the land and they burn it in a dry, see. Den, well, actually dis (e)rode de soil a bit, you know. And after workin dat for tree, for a couple years, dey find dat production taper off badly, you know, bekaas dey destroy de soil dat way. They would come dig up around those old stumps and things and plant their corn diet, and then they plant things like that, rice. Yeah, after a couple of years, de production drops off, dey change again. Dey leave dat an dey go-whole village change, yeah. Well nowadays, government is helping them and they've improving the land. They using a little mechanisations for better results, you know. They use certain additives like, urn, fertilizers, and so on, to the soil. It takes time for dem to really accept, you know, um, de need for change, to different approach, you know. But dey are basically corn, rice, an plantains; an pigs an chickens; yeah ground food, peppers, naturally; bikaa dey make pepper soup; beans; got beans to go wid de corn tortillas.

3.1.2 Text 9: Dugu (a Black Carib variant of voodoo) (Mark, 16; Garifuna, Punta-Gorda) 1. 2. 3. 4. 5. 6.

Like in voodoo they have this certain woman that comes in the middle of the hall and starts to hypnotize everybody. If you aren't on guard, they'll come up and slap you right down off your seat. Some of them don't touch the ground at all. You know, just dance, knocking about like they're dying, like when people have fits. The lady hypnotizes them, they no

75

16 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

CREOLE AND DIALECT CONTINUA know it. . Then afterwards they have some of the men that, like, suddenly, they'd take off their trousers and shirt and they would go up to the beam, with lone -ah- like these shorts, and suddenly they'd come down and tell you all what's wrong with you and your parents. . .They have some woman, like, after the celebration is finished, you go along with which or whatever you call the lady—the bouye—and she takes you to certain part of the burial ground and you have to pray along with her to the dead person you want to pray to. And then, suddenly, like, you take along a person along with you to dig a hole, and put a pig's head in there and chickens.

3.2 Linguistic strategies in acrolects Acrolects present regular idiosyncratic features, and those phonological or grammatical features are the creative results of the late acquisition of standard patterns in contact with creole patterns. I have identified three common strate­ gies which appear to be particularly involved in the development of acrolectal features, and can be related to procedural universals, referring to universal prop­ erties of processes in language development (Muysken & Veenstra 1995:120). They include frequency variability, structural hypercorrection and relexification leading to grammaticalization. 1) The frequency variability strategy is commonly associated with the incorporation of a relatively new standard variant into the acquired acrolectal system, typically leading to a variable combination of vernacular and standard variants for the same feature, for example, [d] and [ð] as reflexes of the voiced interdental, in contrast to the categorical use of the vernacular variant [d] in basilects, and the categorical use of the standard variant [b] in the external standard system. 2) Structural hypercorrection is a phenomenon which involves the overgeneralization of a standard rule by introducing new forms. For example, the acrolect may extend pluralization via standard rules to abstract or generic items which are not habitually marked for plural in standard English, as represented in They using a little mechanizations (8:33). 3) Finally, relexification, which often coincides with some type of grammat­ icalization, is perhaps the most fascinating and creative strategy found in acrolects because it is often concealed under a conventional morphological or lexical item borrowed from the superstate. This phenomenon generally involves the substitu­ tion of a standard lexical item for a creole item but with the additional dimension of a functional shift that meets the syntactic demands of underlying patterns of the

CREOLE ACROLECTS AS INNOVATIONS

77

native basilectai system, or perhaps universal constraints. For example, the use of were in "the cassaba were strain, it comes into a kinda powder" (8: 14-17) may be the result of the relexification leading to grammaticalization of the creole anteri­ or past preverbal morpheme mε, which in the sentence quoted above corresponds exactly to the anteriority of cassava straining: "after the cassava has been strained, (as a result) it becomes powder" (Escure forthcoming). The three major acrolectal strategies will be further illustrated below, 3.3 Linguistic features Although the focus of this study is primarily on the discourse component, I will first review some prominent features of the other components—phonological, morphological and syntactic—in order to provide an overview of acrolectal pat­ terns not found in the lexifier language. 3.3.1 Phonological features One of the most striking phonological characteristics of acrolects is that they contain a statistical combination of standard and creole variants for any given variable, illustrating particularly well the first linguistic strategy identified above as frequency variability. Interdental fricatives constitute a prime example of a diagnostic feature because interdentals are practically absent from creole basilects where they are represented by stop reflexes. They variably occur in fairly high proportions in acrolects, though never as much as in the standard.4 It can be safely assumed that interdentals do not exist in the creole phonemic sys­ tem but, instead, are acquired and added to the speaker's competence in the process of learning a standard. Given the ethnic differences in Belize, specifi­ cally at the research site of Placencia and its neighboring Garifuna community, it is particularly interesting to examine how Creole and Garifuna speakers respectively acquire the standard interdentals. The creole basilect and the Garifuna language both lack interdentals, which means that the linguistic base is unified in this respect for both groups of speakers. If they show differential behavior in the acquisition of interdentals, then extralinguistic factors must have played a crucial role in the acquisition. It is possible to gain insights into the potential Creole/Garifuna differences by comparing text 8 produced by Peter, a Creole, and text 9 spoken by a Garifuna man. Although the transcripts are relatively short, the generally high incidence of English interdentals or their reflexes are so widely distributed in normal discourse that even a brief sample includes a sufficient amount of those sounds (seventy-six potential interdental sites in text 8). To illustrate the vari­ able distribution of phonological features, texts 8 and 9 have been scanned for

CREOLE AND DIALECT CONTINUA

78

the incidence of two variants of the phonological variables /ð/ and / θ / , respec­ tively, [d] and [ð] and [t] and [θ], as represented in table 3.1. Table 3.1: Interdental variation in acrolects (Texts 8 and 9)

PETER MARK.

N [61] [30]

[ð] 40.9% (25) 96.6% (29)

[d] 59% (36) 3.3% (1)

— —

N [15] —

[θ] 60 %(9) —

[t] 40 %(6)

In acrolectal text 8, Peter uses almost equivalent amounts of the creole variant [d] (59%) and the superstrate voiced variant [ð] (40.9%). The latter constitutes a very high incidence, considering that Peter's basilect consistently lacks interdentals. He also has an even higher frequency of the voiceless inter­ dental (60%), although the latter phoneme occurs in much smaller proportions than its voiced counterpart, as is the case in all English varieties. In contrast, in text 9, Mark uses primarily the formal variant in this short excerpt (no voiceless interdental context occurs). The Creole and the Garifuna respondents exhibit highly differentiated linguistic behavior, which is reflected in the Creoles' gen­ eral opinion that the Garifuna people hypercorrect their speech and the Garifuna's own statements that they value English more than the creole ver­ nacular (see chapter two). This would confirm that linguistic behavior matches ethnic identity and, more specifically, the different attitudes toward English held by the two ethnic groups. The choice of the variants [ð]/[d] and [θ]/[t] is somewhat difficult to predict., unless one takes into account the psychosocial values they hold for speakers. In addition, the choice of variants is determined by the immediate linguistic environment. Word-internal phoneme position seems to be a relevant factor in these acrolects: interdentals, especially the voiceless, rarely occur in medial or final word positions, except in naat (North) or in quantifiers such as anyting, somting. In fact, the first use of the creole variant occurs in anyting (1.3) after seven consecutive incidences of interdentals, all in word-initial position, including the marked combination [ðəθiηz] the things. However, an interesting pattern of variability emerges in the use of the interdental/stop variants of text 8 when the context is more closely examined. Even in this limited excerpt with a total num­ ber of seventy-six putativc interdentals, interdentals and stops are clustered dif­ ferentially in separate paragraphs with a varying interdental frequency which differs from the overall figures shown in table 3.1:

CREOLE ACROLECTS AS INNOVATIONS

79

Table 3.2: Interdental frequency in three segments (Text 8) Segment 1 Segment 2 Segment 3 Overall

(1.1-10) (1.11-21) (1.22-40) (1.1-40)

73.3% (11/15) 23.8.% (5/21) 45.0% (18/40) 44.7% (34/76)

The shifts in the three sections seem to follow naturally from the level of attention paid to speech and the speaker's degree of involvement and familiar­ ity with his topic. In the first part of the text, he is more formal and self-con­ scious, basically giving a fairly formal overview of nutrition problems in Belize; this is accompanied by a high incidence of interdentals (73.3%). The second segment is more relaxed because he positively presents a familiar topic ("Carib foods"); this corresponds to a predominantly creole pronunciation and a very low incidence of English variants (23.8%). Finally, the third segment reverts to a degree of formality, as he moves on to Mayan methods of land cultivation and concludes with a critique of those methods (45%). Obviously, the text aver­ age frequency of 44.7% obscures the text-internal variability which is only matched in the third segment. Such variation illustrates the extreme sensitivity of phonological features to context and mood. Clearly, this limited example demon­ strates that acrolects are very different from the standard, at least in respect to interdentals and their variants. This is not tantamount to saying that standards are stable, but it is obvious that acrolects are much less stable, showing that the acquisition of interdentals is not a simple, rigid, and unilateral process, but reflects psychological and social variations in discourse processing. An acquired phonological feature is a finely-tuned instrument that helps signal the speaker's intention and evaluation of the topic and/or the audience, and is an important aspect of interpersonal and intergroup mechanisms which must hold an important place in the speakers' communicative competence, as demon­ strated in Labov (1972) for Black English, for example. 3.3.2 Morphological features Most diagnostic morphological features found in English-based varieties are also distributed in quantitatively variable proportions according to the frequen­ cy variability strategy. The most common features include postnominal or postverbal markers of tense, aspect, and case in Standard English: plural, thirdperson singular present, past, and possessive. The tense-aspect system in native varieties is represented differently, as well as the marking of number (typical­ ly postnominal markers, as in da bali dem 'those friends') and of possession (typically simple juxtaposition or the use of pre-possessor morphology as in Jan fren/fi Jan fren "John's friend").

80

CREOLE AND DIALECT CONTINUA

A quick overview of texts 8 and 9 indicates that the addition of Englishderived morphology occurs in fairly high proportions in plural marking and in equivalent amounts for both speakers (87.9% and 83.3%). Other features (e.g., the third-person singular present morpheme and the past morpheme) occur less frequently in these small texts but indicate the same trend, at least for the Creole speaker (Peter): Table 3.3: Morphological acrolectal features PLURAL N [z] 0 Peter 33 87.9% (29) 12.1% Mark 6 83.3% (5) 16.7%

N 8 4

3S/PRES [z] 0 62.5%(5) 37.5% 100% 0

N 4 0

PAST [d] 75% (3) -

0 25% —

Morphological features occur less frequently than phonological items; thus, a short text is less revealing of morphological frequency patterns of vari­ ation, and it is not possible to say whether the ethnic difference is represented in morphological variation. Plural marking, however, indicates the same textinternal variability (for text 8) that was observed for interdental variation, indi­ cating again that the second segment is more informal and more likely to include traces of the native basilectal system. The third segment actually includes hypercorrected plurals, or overpluralization, which reflects structural hypercorrection, the use of the second linguistic strategy, with a 112% rate of plural marking since nineteen nouns are pluralized in the acrolect while only seventeen would be in Standard English: Table 3.4: Plural marking in three segments (Text 8) Segment 1 Segment 2 Segment 3 Overall

(1.1-10) (1.11-21) (1.22-40) (1.1-40)

100% (6/6) 50% (4/8) 112% (19/17) 87.8% (29/33)

A similar variability in number marking occurs more broadly in the com­ bination of plural and singular coreferents in nouns, pronouns, and verbs, as in "one of the things about a, fisherman is that they don't have a very good diet" (8: 1) and "dey make a thing call 'em casarina, which is -it's -a dried and process"; the cassaba were strain, it comes into a kinda powder" (8: 14-17). These excerpts illustrate the combined effects of the creole system, represent­ ed here in zero-number marking, number-neutral value of creole 'em function­ ing as an alternative for a relative pronoun, and the superimposition of the

CREOLE ACROLECTS AS INNOVATIONS

81

English system (an attempt to conjugate the verbal is/were, and the marking of third person singular present). Other features display a similar frequency vari­ ability of the creole and standard variant of a single linguistic feature: for exam­ ple, the marking and non-marking of third person singular as in "it look like cabbage, but it doesn 't grow like cabbage" (8: 6-7). The frequency variability strategy can be extended to include standard and creole variants of different variables, a phenomenon which involves feature mixing: one standard feature is marked and another is not, as in the following sentence in which the verb is marked in standard fashion and the negation is nonstandard: "the lady hypnotizes you, they no know it" (9: 6,7). Similarly, there is frequent variation in the use of the V-ing English progressive. In text 8, lines 33 and 34, the same verb form occurs twice in succession, first with a cop­ ula-less progressive and then with the unmarked form: dey using a little mech­ anisations. . . dey use certain additives. . .The same combination with the verb say occurs in (8): (8) Sometimes talking, and you don't really understand what they saying, right. You no understand what they say.[K, 10.2:1] Structural hypercorrection, the second acrolectal strategy identified above, is represented in the peculiar use of past marking, with some confusion of pre­ sent perfect and preterit, as in '7 didn't eat it for years, but I saw it, and I think it's in S. a couple months ago" (8: 7, 8); also in other examples not included in the two sample texts: (9) We have shown we not responded to the opportunities favorable. (Peter T9; 1:20) (10) We were being taught to speak English from grammar school days. (Errol, T.7; 3:1.3). Finally, the third acrolectal strategy, relexification, is commonly used to grammaticalize nonstandard interpretations of standard structures. It was men­ tioned above that the basilectal past/anterior preverbal morpheme mε may be replaced by the preterite form were/was. Other items, in particular certain adverbials, may function as substitutes for the basilectal marker. This is the case for the adverb just in we just have dat ship illustrated in (11). The cruise ship mentioned had, in fact, discontinued its visit to the village a couple of years before this conversation, thus it is clear that the use of just (and was) must have

82

CREOLE AND DIALECT CONTINUA

a function distinct from the notion of immediacy which is the usual English gloss of just (11) We just have dat ship comin in here, dat was good because it only spend couple hours; de touris' it just stay couple hours. (That ship had just come in; it's a good thing that it only spent a couple of hours here; the tourists just stayed a couple hours.) [S:11:23; 1989] Additional evidence of the new function of had/was is also clear in the fol­ lowing excerpt from a conversation with the village midwife. She consistently uses had/was to refer to the setting of the story and to events anterior to the actu­ al topic (the death of the second baby of a pair of twins she had to deliver dur­ ing a storm). (12) I had one case, it was twins. Well, one come da Wednesday, and we had bad weda, Placencia was unda wata, no plane can land at di airstrip, no boat come, de road block off, de bridge dey flood. Well data baby neva come. When di wata brok Friday morning, and we had to catch di plane and gone up. Dat baby born on di plane just when di plane pitch into Belize. But di ambulance come wid, it had oxygen, you know, come a help di baby, and di baby died on de way to hospital. [Ti: 12:22-23] Note in (12) the use of neva, which is the negative equivalent of the past/ante­ rior morpheme me but is easily transferred into acrolectal speech with an equiv­ alent value. Furthermore, brok (in "di wata brok") is no indication of past tense marking; creole basilects have long reanalyzed certain standard past forms {broke, left, got, gone, born) as regular tenseless creole verbs, as seen in the fol­ lowing sentences: (13) Di dance had to brok up. (The dance was interrupted.) (14) I had to lef dat one. (I had to leave that one behind.) (15) Everybody use to got baby easy. (Everybody used to deliver their babies without problems.) (16) Dat mos' got cancer. (It must be cancer [that she has].) The various examples provided above clearly indicate that a borrowed

CREOLE ACROLECTS AS INNOVATIONS

83

form may be interpreted differently at separate stages of the continuum: some irregular preterites are interpreted as basic verb forms, and some preverbal aux­ iliaries (was, were, had, did) are grammaticalized as anterior preverbal mor­ phemes, as are other preverbal adverbials such as just. Similar adaptation occur in many other areas of the language. Adverbial forms or other categories can be assigned aspectual functions that they do not have in the lexifier language: steady as iterative/habitual (17); suddenly as a combination of inchoative/focus marker (as discussed above; see text 9:10, 16); will/would and contracted forms such as "they'll start cursing in (18) and usta are used to mark habitual/durative aspect (8: 4, 13; 9: 3, 8, 9, 10); start/begin are also often used as what appears to be an iterative rather than an inchoative aspect marker (18): (17) When a mε li bway da Belize wi de sell crab stedi ting. (When I was a lit­ tle boy in Belize-City, we used to sell crab regularly.) [B. 20:15] (18) They (Mayas) drink that (rum) down and they begin cursing and jumping around like fighting with people. Like they'llstart saying the word "smishna," shi-mi_ah-na. . .your mother is a. . .and they 'll start cursing people like that. The relexification/grammaticalization strategy is a pervasive phenomenon which provides flexible options in acrolectal development and, in effect, blurs the boundaries between morphology, syntax, semantics, and pragmatics. The above discussion of grammaticalization could as easily fit into the following summary of acrolectal syntax, since the syntax of acrolectal discourse is depen­ dent on the types of relexified morphemes just illustrated. 3.3.3 Syntactic features It is obvious from the acrolectal texts presented in this and previous chapters that acrolectal syntax is different from standard syntax in numerous ways, some more covert than others. This is most clearly represented in high frequencies of salience structures, the presence of resumptive pronouns in relative clauses, ser­ ial verb-type forms, and special aspectual forms. Those patterns are more fre­ quent and diversified in creole acrolects than in their superstrates. In syntax the most common strategy is frequency variability, a combination of standard and creole features: 1) The negative construction can be represented both by the use of the doauxiliary and the negative morpheme according to standard patterns, or by reg­ ular preverbal no or multiple negation widespread in creole syntax. Only the

84

CREOLE AND DIALECT CONTINUA

standard negation is used (four times) in Peter's text, whereas Mark uses two standard forms and one creole negative constructions. 2) Copular absence or presence in acrolects is another common combina­ tion found in variable frequencies. Standard copula forms are rarely found in a consistent fashion in acrolects, but the combination copula/zero-copula is typi­ cal of acrolects, as in "government is helping them. . . .they using a little mech­ anisations" (8: 32,33). The absence of copula usually entails the absence of the passive voice in acrolects, which, in this respect, matches the creole absence of passive even though an agentive preposition may occur in the acrolect: (19) The school I used to go to used to teach by nuns. (Kim T. 10; 1:1) (20) West Indies, yeah, well, we influenced a lot; we influence by West Indian custom, West Indian music, everyting. (Errol, T.10.6: 3) In addition, copula usage is overextended to fulfill focusing functions which are not common in standard English, especially in presentative positions or cleft/pseudo-cleft constructions, with the frequent use of is/it's before topics, as in: "is one of the things we don't bother about that in Belize" (8:2); "I think it's in Sartineha a couple months ago" (8: 8). 3) Relativization shows interesting patterns in acrolects. Although the reg­ ular standard relative pronouns "which, that, zero" are commonly used in acrolects instead of creole we, they are often accompanied by resumptive pro­ nouns such as: (21) Is one of the things we don't bother about that in Belize (8:2-3) (22) They have two different type, you have the type, the Caribs, they use plen­ ty which they use it for all different type of things. (8:11-13). Resumptive pronouns in relatives are highly related to the focusing/repetition phenomena represented in topic structures (Escure 1988). In (22), the NP the type is fronted and presented by a special topic structure (you have), relativized, and pronominalized in the relative clause. In (23) the NP the woman is fronted and similarly presented, relativized after its preposition, and associated with a paraphrase: (23) They have some woman, like, after the celebration is finished, you go along with which or whatever you call the lady -the bouye-. (9:11-13)

CREOLE ACROLECTS AS INNOVATIONS

85

4) Serialized verbals are often disguised in acrolects as semantically asyn­ detic coordinated structures similar to those found in Tayo (Corne et al. 1995), as discussed above. Such combined fragments are certainly part of what pro­ duces idiosyncratic acrolectal structures, more paratactic than syntactic, as illustrated in the following examples: (24) dey use it for-um dey make a thing call em casarina. (8: 14) (25) they cut down the land and they burn it in a dry. (8: 23) (26) they would come dig up around those old stumps. (8: 27) In addition, (27) presents serials in a more meaningful sequence: (27) Sometimes they'll want to kill you, chop you up with a machete. . .This man he was saying that the guy came up to his house that night drunk, raped his wife and chopped 'im up all over 'is head, cut 'im over there. Like last month a guy got drunk around Silk Grass there and he went into this lady-he-used-to-live-with- house, and got a look at her, looked at her and looked at the kids, and just went about mincing them up with his machete. [12:34] The verbals qualify as serials because they do not amount to separate actions but fit together to depict one overall event, some verbals fulfilling the roles of prepositions, providing place, manner, or time information; thus, the sequence kill you chop you up is considered to be a serial structure because it includes the main action verb followed by a secondary verb indicating the man­ ner of killing. Similarly, the sequence [chopped 'im up over 'is head cut 4m over there] is basically made up of verbals which seem in English to be almost paraphrases of each other; however, there is a subtle, but essential, distinction within a creole context. The first verb chop is more specific as it indicates a rapid, repeated manner of cutting, whereas the verb cut means more generally "attack with a sharp object." Therefore, the creole serial concept of comple­ mentary verbals is represented in the sentence shown above, even though it seems merely repetitive in English. Most of the syntactic features outlined here tend to overlap with discourse features. This is the case for the extension of the "copula" to topicalization and the use of resumptive pronouns in relative clauses which is related to repetitive patterns represented in creole discourse. Serialization is also related to the flow of pragmatic organization and cohesion represented in creole discourse. Those

86

CREOLE AND DIALECT CONTINUA

factors will be more closely examined in the following chapter. 3.3.4 Effects of basilectal influence in acrolects This chapter has provided a summary of a few phonological, morphological, and syntactic features felt to be representative of acrolects. The selected features include interdental/stop variation, negation, copular variation, relat vization, verb serialization, and the marking of number, tense, and aspect. Observations of those linguistic phenomena strongly suggest that basilectal substratal pat­ terns are incorporated into acrolects through various innovative strategies. Those strategies include the combination of creole and standard variants result­ ing in variable frequencies which reflect the introduction of new variants, struc­ tural hypercorrection, and relexification. Acrolectal differences from the corresponding "target" language can thus be explained in terms of first language/vernacular input resulting in a new code. In other words, there is evidence of substratal/native vernacular influence on an acquired variety, a strategy no doubt similar to that underlying pidgin genesis as well as second language acquisition. Basilectal/vernacular linguistic behav­ ior appears to be reinterpreted through superstate or "target" equivalents, although it cannot be viewed merely as a calquing process. It is much more like­ ly that the acquisition target is to produce a new, creative code that captures contexts and speakers socially different from the standard speakers, contexts, and speakers. It is a set of intended differences representing different identities. The relexification of basilectal morphemes into standard morphemes and the substitution of basilectal phonemes for standard phonemes represent vernacular influence in interdialectal situations, especially when the same speakers control different dialects. 4. Conclusion This chapter has established that acrolects cannot be assimilated to standards. The few features (derived from various linguistic components) that were examined provide evidence of distinctive patterns not found in the standard. Clearly, the acrolectal continuum subset constitutes another language category that differs as much from the external standard as it does from the native basilectal subset (cre­ ­les) and the mesolectal subset (semi-formal). Acrolects correspond to the most formal subset of the creole continuum which combines features of the native and acquired language. The evidence presented indicates that the acquisition of acrolects is not a unidirectional development toward an official target but rather a complex, variable, and socially significant process showing the combined effects of basilectal and superstratal influences, as well as the possible emergence of uni-

CREOLE ACROLECTS AS INNOVATIONS

87

versal features, an issue to be discussed in the following chapter. In sum, it is claimed here that when two different but regularly interacting systems, as typically represented in creole continua, conflict in one specific lin­ guistic area, diverse creative strategies are implemented to produce an ade­ quately formal style while at the same time meeting the structural needs of the native system. It may well be that the process of piginization and the process of acrolect acquisition, as well as other more traditional cases of second language acquisition, function in similar ways. This view would also lend support to the claim that early stages of pidginization and creolization are not cases of simpli­ fication but, to the contrary, include extensive variable systems in contact, sim­ ilar to those we can now observe in real life second dialect/second language or acrolect acquisition situations. Endnotes 1. It is generally said that a serial verb construction is characterized by a single intonational contour. This is the case in all the examples produced here. 2. Currently, there are valuable attempts at documenting the early development of creole through research of eighteenth- and nineteenth-century creole sources, such as Lalla & D'Costa (1991) for Jamaican Creole and Rickford & Handler (1994) for Barbadian. However, written documents can only represent restricted aspects of the varieties used at the time, since written representations of the creole presuppose the author's personal interpretation of an oral variety of which he is usually not a native speaker. 3. Both control repertoires which include basilects, mesolects, and acrolects (see Escure 1982 for samples of those variations and a classification of the contexts and speech acts which trig­ ger such linguistic variations). 4. This feature is also characteristically variable in Black English and in all second dialects of English; generally, interdentals are universally marked phonemes that are relatively rare in world languages.

Chapter 4 The Interaction of Syntax and Pragmatics in Acrolects: Topic Marking Abstract Topic marking strategies are examined in the context of spontaneous Belizean acrolectal discourse. Three major strategies are identified (fronting, repetition, and presenta­ tion) in a representative sample. Synchronic and diachronic comparisons are estab­ lished between basilects and acrolects, with the intent to assess crosslectal variability. Belizean patterns are then compared to an equivalent standard (not creole-related) con­ trol sample (American English), in order to test the putative distinctiveness of dis­ course features in creole acrolects as second dialects. The remarkable stability observed in pragmatic structure suggests that paratactic strategies are consistently priv­ ileged in Belizean varieties and closely linked to topicalization strategies. 1. Syntax and discourse features The previous chapter briefly identified idiosyncratic aspects of Belizean acrolects. Acrolects display a unique set of phonological and morphological features, which distinguish them from their basilectal and lexifier counterparts in corresponding phonological and morphological areas. Acrolectal syntax also appears to be characterized by a high degree of flex­ ibility, reflecting the three major types of strategies outlined in chapter three: frequency variability combining basilectal and standard forms in varying degrees, hypercorrection of standard forms, and relexification of basilectal ele­ ments. However, with the possible exception of some limited verb-bound syn­ tactic rules, such as negation, the scope of syntactic rules is likely to expand beyond simple sentential units, requiring reference to a discourse context, espe­ cially to a natural discourse context, as previewed in chapter one. There are obvious advantages to be derived from the study of discourse

90

CREOLE AND DIALECT CONTINUA

produced in natural social situations. The first advantage lies in the presence of an interpretable context, as opposed to working with sentences artificially elicit­ ed through interviews or lab protocols. Spontaneous communication is undoubt­ edly the correct setting to evaluate how second dialects, such as acrolects, are acquired. Secondly, this type of context is particularly conducive to an under­ standing of the articulation of syntax and pragmatics. Furthermore, the investi­ gation of natural discourse is most likely to offer privileged insights into the dynamics of linguistic change which led to the development of pidgins and cre­ ­les, and the concomitant acquisition of acrolects, since all of these are acquired in spontaneous, unmonitored contexts. The analysis of the syntax of discourse, however, poses serious challenges, both theoretical and methodological, as outlined in chapter one. Although the definition of structural units of analysis selected here may be controversial, and the discourse model adopted perhaps not entirely satisfactory, the approach fol­ lowed here will nevertheless provide crucial insights into rarely investigated areas, and in particular acrolectal performance as representative of second dialect acquisition, and the interaction of syntax and pragmatics in acrolects. The area of discourse selected for analysis involves the typology and distribu­ tion of topic structures. 1.1 The theoretical interpretation of topics In theoretical analyses of topic structures in English or other Indo-European lan­ guages, one major concern has been the interpretation of the anaphoric relations which hold between a topic and some other sentence constituent. In the early transformational generative framework (Chomsky 1973; 1977; Ross 1967), top­ ics are fronted by movement rules, and their pronominal copies may be left in the pre-extraction position in the underlying base sentence. Although I will use in this discussion the terms topicalization, left dislocation, and clefting which were first developed as part of the movement analysis, I assume here no such interpretation. These terms simply constitute convenient ways to identify sepa­ rate surface structural mechanisms dealing with topic presentation, some accompanied by anaphoric pronouns and others not. Similarly, the term fronting does not necessarily imply that the topic was elsewhere at another level; it mere­ ly refers to initial placement. In fact, it is possible to assume a basic topic-first structure, instead of a subject-first sequence (for SVO languages). In the later generative framework, and in particular in government and binding (GB) theory (Chomsky 1977; 1981), the process of topicalization is accounted for by two base rules:

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

91

a) S" → TOP S' b) S' → COMP S' An OSV structure such as, "My mother, I like" is accounted for by mov­ ing an underlying WH-phrase in S' into COMP position, then deleting it, leav­ ing an "empty element." An OSV Pro sentence (left dislocation) such as "My mother, I like her" is explained by assuming that the base contains the pronom­ inal anaphora co-indexed with TOP, rather than a WH-phrase and, thus, no movement applies. However, the GB framework does not readily account for all topic structures, except perhaps in some restricted styles of English. The scope of topic marking is an issue which conflicts with the GB inter­ pretation. The traditional sentence-bound interpretation of topics does not ade­ quately account for the scope of topics in natural discourse contexts such as the ones providing the data in this analysis. This is because, as shown in chapter one, spontaneous conversations can hardly be broken into simple sentences, a fact which led to my choice of the "topic unit" as the unit of analysis. The scope of a topic may extend to several verb units, although in some cases it can natural­ ly be restricted to one simple verb-phrase. Another problem with the GB framework is its specification that co-indexing must exist between the topic and some other element (empty or lexicalized) in the comment. In fact, there are structures in which the comment or predicate contains elements which are not anaphorically bound to the topic, even though some semantic relation may hold between them, as shown in the following sentence: (1) Citrus dey had a lot of competition « with Argentina, but now de bananas we can't produce enough « bananas to put in de market for Europe.1 (BL=Pom/94.3) In (1), the topic citrus is not linked to any overt constituent anaphor, although it is clear that it is topicalized out of a prepositional phrase [for citrus] with con­ comitant loss of the preposition, and originating as a complement of competition. In addition to the topicalization of citrus, this sentence also illustrates left disloca­ tion in bananas, with a full copy in the extraction position. In their fronted posi­ tion, citrus and bananas are similar to presentative structures with adverbial or prepositional highlighting morphology of the type as far as X is concerned, or as to X in English. A likely standard English version appears in (la): (la) As far as citrus is concerned, they had a lot of competition from Argentina, and as to bananas, we can't produce enough to put on the market.

92

CREOLE AND DIALECT CONTINUA

Chomsky's GB theory cannot account for the wide use of zero anaphora in creole varieties, as well as in some types of casual English discourse. The prag­ matic inferences involved in assigning the antecedents for anaphora will also be discussed in the following chapters dealing with Chinese (chapters six, seven, and eight), with reference to a pragmatic theory of anaphora (Huang 1994). 1.2 Topic There is no consensus on the issue of the basic definition of topic, except for the fact that topics have multiple facets. A special issue of Linguistics (1992:30) illustrates some of the perspectives on topic-hood, with special attention to the grammatical, pragmatic, psycholinguistic and crosslinguistic nature of topics in utterances and in discourse. For example, Givón refers to psycholinguistic notions of mental processing derived from artificial intelligence research. "Referents ('topics') serve as file labels in the episodic memory for stored text" (Givón 1992:5). Cadiot refers to topics in French in terms of form-function rela­ tionships, hypothesizing that "the syntactic devices associated with topic-hood match pragmatic functions and strategies" (Cadiot 1992:57). Schiffrin discuss­ es the duality of the notion of topic functions at the communicative level, as well as at the linguistic level (Schiffrin 1992:165). Yule and Mathis define speech topics in terms of "'staging', which organizes the various elements of the speaker's contribution into background and foreground positions, and. . .'constructed dialogue', which captures the use of direct speech forms (though not verbatim quotes) within reports of previous conversations" (Yule and Mathis 1992:199). De Beaugrande examines "'topic' in respect to the econ­ omy of discourse, that is, the active system of values, focuses, and priorities among textual elements, and to the agenda of discourse, that is, the ongoing docket of actions, needs, motives, and goals" (de Beaugrande 1992:243). Not surprisingly, considering the diversity of such definitions or descrip­ tions, it is stated in the introduction of this series of articles that "Definitions of 'topic' and 'topic-hood' are obviously controversial" (Dittmar 1992:21). Such complexity even leads some authors to a radical disposal of the notion of topic. Schlobinski and Schütze-Coburn (1992:89) show that "problems arise when topic is equated with 'given' or 'known' information, with 'point of departure', with 'what the sentence is about', with 'communicative dynamism." Since a definition of topic is unobtainable, they consequently propose to abandon that cumbersome notion, substituting for it another potentially cumbersome mecha­ nism, as they propose "the independent treatment of all linguistically relevant categories (at all levels) that are usually factored into the topic equation in one place or another." Notwithstanding disagreements on topics, the following discussion will

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

93

rely heavily on realistic data, and especially chunks of spontaneous conversa­ tions to provide an empirical analysis of topic units in Belizean and Chinese varieties, and perhaps eventually reach an interpretation of its place in human language processing. 1.3 Topic and focus It is often presumed that topic and focus are distinct concepts and there has been various attempts at distinguishing between those two notions. For example, topic is often defined as "the subject about which something is said" or as "given information" (Crystal 1987:421), whereas the focus is "an element in a sentence to which the speaker wishes to draw special attention" (Crystal 1987:432). This presumed contrast may, in fact, lead to confusion since there is no reason why given information may not be that to which a speaker wants to draw attention. Some linguists believe that there is a clear formal distinction between focusing and topicalization, in addition to a "comma intonation" which would occur with topicalization but not with focusing: In topicalization, the topic NP is not coindexed with any null element in an Aposition, but is coindexed with some overt pronominal (or quasi-pronominal) form: "Mary, I saw her yesterday". . .In contrast, focused NPs are always coin­ dexed with a null element in an A-position, never with any overt element: "It was Mary that I saw" (Bickerton 1993:192). Bickerton thinks that focusing (here, clefting) is the dominant syntactic device to achieve constituent emphasis, in contrast to topicalization mecha­ nisms, which according to him occur in few creoles. However, spontaneous dis­ course contexts do not validate a clear distinction between topic and focus. Whether reference is made to presupposition, scope of assertion, assertive ver­ sus counter-assertive focus, or old/given versus new information, there seems to be no basis for treating topic and focus as separate constituents or as involving different processes. Focus appears to be a mark of salience given indiscrimi­ nately to old or new topics, and the focusing process can be applied to any kind of constituent, and anywhere in the sentence or proposition, as will be demon­ strated in various types of Belizean and Chinese discourse. The general associ­ ation of the topic with the notion of "aboutness" in the sentence or in larger units seems to be satisfactory because of its very generality (Keenan & Schieffelin 1976; Wald 1983:104) . The notion of "aboutness" is also related to that of "firstness" or frontness (initial, pre-clausal position), although it will be shown that focusing can apply almost anywhere in the sentence or utterance. In the fol­ lowing discussion, the term topic marking will be broadly used to refer to focus­ ing mechanisms of various types.

94

CREOLE AND DIALECT CONTINUA

2. Types of topic strategies This analysis will refer to three major types of topic structures commonly found in the world languages, although not necessarily with the same degree of fre­ quency. The aim of this study is to determine which of those common topic strategies regularly occur in acrolects, and whether the relative incidence of such strategies separates acrolects from other lects and from their lexifier languages. The strategies are: 1) fronting of the topic, that is, various types of focused structures, whether nominal or verbal, appear in sentence—or clause-initial position, regardless of logical position, or place of extraction; 2) copying or repetition of the topic within the discourse unit; and 3) the use of specific presentative elements external to the actual sentence as topic markers. Those three categories of topic structures are presented in more detail, then ana­ lyzed in acrolectal texts. Such topicalization strategies are assumed to be much more frequent and diversified throughout the continuum than in the lexifier lan­ guage (English), which is not topic prominent (Escure 1988). In order to verify this assumption, a brief control sample of standard English varieties external to Belizean acrolects, American English, will be the subject of analysis for com­ parative purposes. 2.1 Topic fronting Sentence or clause-initial placement of the topic element takes place in struc­ tures traditionally referred to as topicalization with simple fronting [citrus in (1), everything in (2)]. It may appear as left dislocation, which involves the nominal or pronominal copying of an initial topic. The topic may be a subject [Carihs in (3)] or a fronted direct object [the second topic of (1) bananas, or de unclean in (4)]. It appears that any constituent, nominal (2, 3, 4), verbal (5), or adverbial (6), may function as topic and may be fronted. (2) Everyting we have to pay for « (Escure 1988:163) (3) The Carihs, « they have different type of cassava (Escure 1988:170) (4) De Bible tell you dat de God made all different animals unclean and de unclean « dey shouldn't eat, right? 'The Bible tells you that God created various unclean animals, and the unclean they shouldn't be eaten, right' (T=Adv/94)

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

95

(5) Lock on me, she mε<
96

CREOLE AND DIALECT CONTINUA

(7) Like this side, allot so much money for <
INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

97

The full or partial repetition of certain constituents is widely observed in acrolects and other lects, and appears to fulfill a necessary function which is not represent­ ed to such an extent in other English varieties. The resumptive pronoun appear­ ing in "which they use it " (9) is also related to the repetition process, as are the anaphora which occur as part of the multiple presentation strategies typical of the pragmatics of acrolectal topic structures. Partial or full constituent repetition, pro­ ceeding from sentence to sentence, results in a peculiar overlapping of syntactic structures, which seems to be rooted in pragmatic processes and occurs in many non-creole-related, nonstandard varieties of English. This "stacked" syntax requires a special analysis, which will be discussed in section 3.1 presenting the procedures for analyzing the distribution of topic structures in acrolects. 2.3 Topic presentation 2.3.1 Cleft, pseudo-cleft, and existential structures Presentative structures constitute another common focusing strategy in acrolects as well as in other segments of the continuum. They include standard clefts (with it's/that's), as in (12), and pseudo-clefting structures (13), often combined as in (12); and also frequent structures of the existential type, such as it have, they have, have, dey got, i got, which are basically equivalent to the English existen­ tial structures there is/are. Existential have strategies appear in (9) "they have two different type," (13) "we used to have all things," and in the following: Fn3 (12) That's what holding up people from buy, say too expensive. 'That's what's preventing people from buying (land): they say it is too expensive' (H=14.4) (13) What happened, we used to have all things that sort of compensate for vegetables 'We had various foodstuffs to compensate for the lack of vegetables' (P=1.4). (14) De ones dat is smarter, dey get ahead while de other one dey hangin back, have no one dat take care of dem at all. 'The smarter ones [children] get ahead whereas the others trail back because there is no one (teacher) to take care of them' (T=Mi 1.94.7) Note that the semantic possessive value of have often overlaps with the more likely existential values of there is/are, which is likely to have contributed to

98

CREOLE AND DIALECT CONTINUA

the semantic shift or lexification of there is to have/got. (15) den we have dis ting about eating 'Then there are these things [restrictions] about food' [in Seventh Day Adventist beliefs] (T=Adv/94) 2.3.2 Presentative particles Beside widespread existential morphemes, a wide variety of presentative ele­ ments, or topic particles, appear in acrolectal topic structures, basically fulfill­ ing highlighting functions similar to those of clefting, pseudo-clefting, existen­ tial, and fronting mechanisms. For example, the phrases dis ting (15); one of the things (16); dat's one ting (17); or de next problem (18), lead naturally to left dislocated structures: (16) One of the things about a fisherman is that they don't have a very good diet 'Fishermen don't have very balanced diets' (P= 1.1,2) (17) Dat's one ting wid him, he can remember everytin de tel him 'He can remember everything you're telling him' (T= Mi 1/94.5) (18) De next problem, I was worried about my children 'I was worried about my children' (=Mi 1/94.2) As is apparent in the sermon excerpt in (19), the use of highlighting phras­ es helps achieve the prominence of crucial concepts. In the following, the pas­ tor is trying to persuade his audience that the village needs to achieve a stronger sense of togetherness in order to redress failing "family values." To get his point across, he is using a metaphor that involves comparing the community to a group of scientists conducting a lab experiment, and he uses such rhetorical devices as a large quantity of repetition, introductory elements, tags, and others. The preacher is especially fond of using the phrase the point (it occurs ten times in this text), which is a particularly didactic device used to focus atten­ tion, even when the point of the peroration is not clear.4 (19) The point is, the point is that, urn, if for instance, let's say a group of sci­ entists going to get togeda, an all o' dem -wait wait wait- all o' dem, all o' dem have some idea to dat particular problem, right, so let's say dey

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

99

have some idea; but somebody will lead them on. Now, dey do deir own experiment in deir lab, dey may come up wid de same findings, but dey have done it different ways, and dey come togeda in a group, an dey, wid all deir shortcomings, all deir mistakes, right, dat dey have made. [E= Not scientific!] (Interruption) Not scientific- don't make a mistake? Oh broda, come on! Well, de ting is, how do you see de community? I just want to finish what I want to say. So de point is dat dese scientists come togeda wid all deir experiments, right, wid all deir shortcomings and so on, an' dey've tried it ova, an' start ova, an so on right? Den, if it was an experiment, I mean, it's like what talkin about last night. You've got, you make mistakes an' so on, you now start to make, you start, an' so on, an' you come togeda, but somebody have got to generate the talk (R= 12.1.8-9) Not uncommonly, as represented in several sentences, there is an accumulation of presentative discourse markers; thus, (20) displays no less than five presentative morphemes followed by two fronted elements, one dislocated (us. . .we don 'wan ') and the other further repeated (Jesus we don 'wan'to accept Jesus). (20) But yet, you know, you see, like us, you see, Jesus, we don' wan' to accept Jesus, or many people don' wan' to accept Jesus in deir lives because he makes requirements of us, you see. (R=12.1). 2.4 Semantic change in presentative adverbs One of the characteristics of acrolects is that the seemingly standard-like usage of certain lexical items conceals distinctive functional or semantic processes which are not derived from the standard. An example of this acrolectal phe­ nomenon is represented in a class of adverbs or conjunctions which appear with very specialized presentative functions that they normally do not have in stan­ dard English. It is clear in Belizean acrolectal (and basilectal) contexts that cer­ tain English adverbs are endowed with supplementary semantic qualifications which contribute to the emphatic marking of the topic. The adverbials suddenly, still, because, only, again are illustrated below in sentences where their basic standard English meanings are modified. In (21), several adverbials introduce the topic, including the adverb suddenly, which appears to denote more than its habitual semantic value implying an element of surprise. There is some indication that change of state is

100

CREOLE AND DIALECT CONTINUA

involved: the speaker talks about traditional obeah ceremonies and a succession of expected rituals. Suddenly introduces a sequence of events, and appears to function more like a relexified morpheme to match a basilectal topic marker (such as da or now), indicating a change of state or a new ritual. (21) And then suddenly, like, you take along a person along with you to dig a hole. 'What happens next is that you get someone to dig a hole' (M/Ca) Similarly, in Belizean still seems to have acquired a contrastive value, which differs from the predominantly continuative value found in the lexifier.4 In (22), T, a twenty-fíve-year-old woman argues for the Seventh Day Adventist belief that Saturday should be the rightful day of rest instead of Sunday (because Sunday appears first on calendars): (22) So how would you rest on first day of the week when God the creator he still rested on the last day? 'How could you rest on the first day of the week while God the creator on the contrary rested on the last day?' (T= Adv/94) (23) We still speak creole, a like my creole language only because G say dat we take it and mess it up. 'We speak creole [in spite of the fact that tourism is developing], I like my creole language even though G says that we mess up (English).' (Bl=Ma2/94) It must be noted, however, that still is sometimes semantically ambiguous. In (23), a fisherman's response to a question concerning the survival of creole in the face of increasing tourism, still may have either the standard meaning of [we continue to speak creole] or [we speak creole in spite of the influx of tourism]. The latter and stronger interpretation is the most likely, given the prior context of the conversation which dealt with the growing presence of tourists in the village, due to the relatively recent construction of a few small hotels. This sentence includes another interesting combination of presentative adverbials; only because has obviously no limiting or causative connotation, as it would have in the standard lexifier.5 The standard interpretation [I like creole because G says it's bad] is absurd in this context, and exactly the opposite of the creole meaning. In a light bantering context, BL means to say, unambiguously

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

101

in this case, that he likes his creole in spite of the mild mockery of the local language indulged in by G—an American tourist—and BL implies that G's opinion is irrelevant and uninformed. Other adverbials such as already, although, now are similarly not restrict­ ed to their standard English semantic value; they clearly have an emphatic func­ tion. For example, the following excerpts from a text more extensively quoted in (29) below show that already (produced with emphatic stress) means more than reference to a prior state, and that although has basically the same value as already, but certainly not the standard implication of presenting a new fact: . . .dey already put him in tomb before de Sabbat began. . .aldough dey wanted to put him in dat tomb before de sun come down. Now, for example in (1) and (19), is not used with any temporal reference, but it clearly has a strong topic marking, or topic change function, which is also now widespread in standard casual English: (1) [now de bananas we can't produce enough] (19) [now dey do their own experiment] Finally, (24) shows that again, another presentative adverbial, must be inter­ preted as indicating a change of state (from cool to warm) instead of its stan­ dard value of repetition. TI, an eighty-year-old woman, repeatedly exhorts a lit­ tle girl not to drink what is left in a glass sitting on the kitchen table because 'it is no longer cold': (24) Dat not cool again, it's hot now, it isn't cool again, dat isn't cool again, it hot, it isn't nice again, don't do it, don't This (drink) is no longer cold, it's warm, so don't drink it' (Th/ MA2/94) Although this study focuses on acrolects, similar values are found in the creole vernacular, as shown in (25), a basilectal example produced by the same speak­ er BL, though in a different context and at a different lectal level.7 Again, the ambiguity of still may have the standard value of [they continue to want to go], or function as a marker of contrast, not unlike its meaning in (22). I favor the latter interpretation, because the contrastive meaning fits best within the con­ text of the utterance, as it did in (23): (25) T. call me tell me dey me got two people we me wan 8 come in, a tell em

CREOLE AND DIALECT CONTINUA

102

a mε got dis trip line up for today anyway, but a don't know if dey still going 'T. called me to tell me that they had two people who would have come (on the present trip to an island), but I told him that I had this trip lined up for today anyway, but I don't know if they will go for sure' (BL=Ma2/94) The idiosyncratic behavior of certain adverbials indicates that there are signifi­ cant semantic differences between acrolects (and basilects) and the standard lexifier. The adverbs discussed in the Belizean acrolects, and others, indicate a change of state, or some contrast in action or feeling, both notions being akin to topic marking. Those adverbial elements often, but not always, carry emphatic or dramatic connotations, which are regularly underscored by verbal stress. The standard and acrolectal/basilectal glosses of the English items discussed above are summarized in the following table. They appear to represent roughly two general semantic categories: a change of state value and a contrastive value. Both are clearly related to an important pragmatic topic shift in discourse: Table 4.1: Semantic

variation in lexifier and Belizean

English item

Standard gloss

suddenly (adv.) again (adv.) already now still (adv.) only (conj.) because (conj.) although

with effect of surprise repetition accomplishment time without change with limited scope causation restriction

items

Belizean gloss (all emphatic) change of change of change of change of contrast contrast contrast contrast

state state state state

The varying glosses shown in table 4.1 are merely illustrative of a complex process which reflects the fluidity of the creole continuum and, perhaps, of every language situation. The approximate semantic/pragmatic values provid­ ed as the Belizean glosses do not imply fixed Belizean semantic values. They indicate possible, though frequent, speakers' choices in specific discourse con­ texts. Thus, it cannot be inferred from the adverbial variations discussed above that all such adverbs merge into change of state or contrasting ele­ ments, all with strictly topic marking functions. These individual items still preserve somewhat distinct, though often blurred and even ambiguous seman­ tic notions. When observing such semantic variations between a creole continuum

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

103

(whether at acrolectal or basilectal levels) and the lexifier language, it is appro­ priate to investigate dialectal or historical lexical correlates. Certain English dialects have loaned specialized and often conservative semantic values to cre­ ­les, due to the well-known emigration of Scottish and Irish elements to the Caribbean, either as sailors or as plantation workers. In the case of the presentative adverbs discussed above, there is little evidence that the Belizean seman­ tic values can be dialectal remnants of earlier forms of English. The nonstan­ dard "change of state" semantic value of again seems to be the most likely traced to Irish dialects (Cassidy & LePage 1980:5; Holm 1982:2), as in [I don't love him again]. In addition, still has a seventeenth-century meaning of 'even then', 'notwithstanding' (ODEE 196:869), as discussed in section 2.4. above. I find it more likely, however, that the distinct semantic values found in Belizean lects result from the operation of a grammaticalization process. A sub­ stratai requirement determining the shape of overt presentative structures may have led to the special semantic values assigned to some adverbial elements. This shift may have been facilitated by the occasional contextual ambiguity illustrated in some of the examples shown above. Similar mechanisms may have operated in the semantic shift of have from possession to existential ('we have' sometimes means 'there are', as in "we have books here"), in the devel­ opment of da as focus particle from a combination of copula and deictic pronoun (Escure 1983c; 1993b), or in the ongoing change of me as anterior aspect marker to emphatic element (Escure forthcoming). This evidence indicates that the acrolect has great potential for modifying the standard lexicon in innovative ways that match the pragmatic requirements of specific interactive situations. The fluidity of the grammatical categories is particular­ ly striking because they may adopt different values, as required by the context. 3. Distribution of topic strategies In the following section, acrolectal samples are analyzed to evaluate the distri­ bution of the three types of topic strategies illustrated in section 2. As argued in chapter one, I consider it essential to observe language in its natural context, not elicited in an artificial interview but recorded in the course of spontaneous inter­ action. I have included in my corpus of natural discourse excerpts from con­ versations recorded in Belize at two different time periods: in the 1980s (between 1980-1985) and in 1994. When dealing with natural discourse, there are two challenging tasks; to define the concept of topic and to assess the extent of its scope, both of which can only be dealt with if a meaningful discourse unit is available.

104

CREOLE AND DIALECT CONTINUA

3.1 Definition of discourse units If the syntax of discourse is best characterized in a natural context, then it is cru­ cial to identify the structural unit within which topic structures can be identi­ fied. The traditional domain of syntactic analysis, the sentential unit, fails to provide a satisfactory, realistic tool in the study of the syntax of discourse, as shown in chapter one (section 3.4). The major reason underlying this short­ coming is the difficulty involved in defining clear sentence boundaries in nat­ ural discourse. In creole and acrolectal discourse (which are likely to be repre­ sentative of all natural speech situations), communication is not conveyed in a neat sequencing of clearcut SVO structures. In fact, messages are encoded in the form of what may be described as stacked units of meaning, involving a great deal of overlapping as shown in examples (26-28). Consequently, it is an extremely difficult and often arbitrary task to determine what constitutes a rea­ sonable discourse unit. The fieldworker's overall knowledge and understanding of the context, her intuitive perceptions of the speakers' intents, and her aware­ ness of community values constitute somewhat subjective but essential vari­ ables required in any assessment of discourse units. It is extremely unreliable to analyze speech samples recorded in absentia or to do so in communities with which the investigator is not intimately familiar. In my attempt to define a viable semantico-structural unit as a basis for analysis, I use all of the above subjective types of information, as well as more objective guidelines. The premise is that the topic is the central concept of any type of communication. Therefore, I define a topic unit (TU) as a portion of dis­ course that starts with the introduction of a topic (old or new) and ends when the topic is changed, or is replaced by an expansion of that topic into another topic or a subtopic. At a basic level, a TU differs from a sentence in that it may include more than one verb unit. While the criteria on which TU identification is based remain tentative, and even sometimes arbitrary, they are nevertheless preferable to a blatantly erroneous breakdown into neat sentential units, con­ sidering that stacked syntactic units constitute the bulk of spontaneous oral communication. Once methodological guidelines are established, the analysis of discourse topic strategies can proceed in a relatively systematic manner. The three strategies defined above, namely topic fronting, its repetition, and its structural presentation through various morphological means, contribute to the construction of the TU, and each of the three strategies may require one or more additional verbs, because verbal topics can be fronted, repeated, and used as part of a presentative phrase. The following basilectal (26, 27) and acrolectal (28,29) examples illustrate the structural TU that is pragmatically the smallest identifiable domain of a meaningful unit. It appears that acrolects pre­ serve the same communicative mechanisms that occur in creole basilects:

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

105

(26) Dey bite me, mosquito bite you, dey bite you good, bite hard. Lot of peo­ ple me out de ms ds de da shallow, shallow part, my pa de fisherman, he go down, when he go dive, he hook up pan his back ina string, an' i go dive ina wata, i go ina wata. 'I got bitten by mosquitoes, when we were there with other people at the shal­ low part [of the caye], and my father, the fisherman, went diving there with his hookstick tied to his back with a string [to catch lobster]' (Ed, 8, 1994). (27) You have to tell him fu check wid me when dey come ya, so dat dey no bring people when I have people here because I tell dem people da nobody else wan de da camp, right. I wan mek him know dat when I gan dat dey can't do dat. 'Tell him [an associate] to check with me and not to bring people [tourists] to the caye when there are some already camping there, I want him to know that when I'm there they can't do it [bring people to the caye]' (B1, 40, 1994). (28) You saw dat canal dat dug out da back, dat goes into de land? maybe pass­ ing through de lagoon have a big canal, dey have a canal from de lagoon an dey go right into de land. 'There is a canal in the back from the lagoon to the land' (HA, 70, 1994). (29) Jesus, dey took one man body and it was a Friday, what dey call Good Friday, I remember dey already put him in a tomb before de Sabbat began, we believe dat de sun is going down and Sabbat has begun dat he has pass away den, and Jesus in de tomb before de Sabbat come creeping up on dem in dere, so it's Good Friday, dey put him in dere, so it's Good Friday, aldough dey wanted to put him in dat tomb before de sun come down. 'They lay down Jesus' body in the tomb on Good Friday before sundown, before the beginning of the sabbath' (TH, 25, 1994). It is clear that each of the topic units shown above can hardly be broken down further. In each case, the message includes several related and overlap­ ping components. For example, in (26) the topic diving with (Ed's) father, includes reference to 1) the background atmosphere (mosquitoes), 2) the actu­ al setting (shallow part of the cayc), and 3) the father's diving equipment. In (28) reference to the canal going from the lagoon in the back of the village to the front is repeated three times. In (29), TH repeats four times that Jesus was entombed before the Sabbath (repeated three times) and before sundown (repeated once) on Good Friday (three copies). Topic units, such as examples 26-28, constitute meaningful discourse

106

CREOLE AND DIALECT CONTINUA

components that define the domain of the analysis, and the relative frequency of the three types of topic strategies identified in section 2 will be assessed in relation to that unit. The three TUs (26-28) display a high incidence of topic strategies: (26), has eight cases of repetition and two instances of fronting (including one type of verb fronting [dey bite me, mosquito bite you, dey bite you good, bite hard]); (27) has four topic structures, (two repetitions and one presentative element da ); and (28) has five topic strategies (four repetitions and one presentative). Although the examples presented here are typical and repre­ sentative of Belizean acrolects, it is a fact that not all TUs include topic strate­ gies, and this is why a relatively large sample is required to assess as accurate­ ly as possible the actual incidence of such mechanisms for each speaker and for the group or lect under consideration. 3.2 Analysis of topic strategies in acrolects Fifteen samples of spontaneous acrolectal speech are tabulated below in order to assess the incidence of fronting, repetition, and presentative morphemes. The overall pattern is presented in table 4.2 and details of individual samples are presented in table 4.3. Of those fifteen samples, four were collected around 1980-1984, and the remaining eleven were collected in 1994. In order to assess the respective importance of each topic strategy, its overall number (TS) is divided by the number of TUs, giving an overall approximation of the speak­ er's—and ultimately the community's—use of topic marking mechanisms. The overall comparison of total relative use of the three strategies repre­ sented in table 4.2 reveals an important discrepancy. Fronting remains highly restricted with an average incidence of only 9.8%, whereas presentation and repetition are both widely used. The presentative strategy, involving the use of particles or highlighting structures, is clearly the most highly valued with an overall average incidence of 53.6% for all texts and an intermediate 36.7% for the repetition strategy: Table 4.2: Topic strategies in 15 Belizean acrolects (percentages represented in parentheses) (N= total number of topic strategies)

TOTAL

TS

Fronting

Repetition

Presentation

717

70(9.8)

263(36.7)

384(53.6)

Table 4.2 shows a detailed tabulation of the fifteen texts which make up the cor­ pus. Of those fifteen samples, four were collected between 1980-1984, and the

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

107

remaining eleven were collected in 1994. There is no significant difference between the lects and speakers investigated. All samples analyzed display large quantities of topic enhancing strategies (repetition and presentation), whereas fronting is the only strategy involving minimal or nonexistent incidence. As indicated in the previous discussion of topic marking mechanisms, sev­ eral of the topic strategies may be combined. Several presentative morphemes combine with fronting, a topic may be simultaneously repeated and fronted, or it may be highlighted by a presentative element and also duplicated in the same topic unit. Thus, there are on average two topic strategies per topic unit (717 TS for 369 TUs). Furthermore, considering the preference for presentation, it not surprising to find that there is an overall average of more than one presentative element per topic unit (384 presentan ves for 369 TUs): Table 4.3: Topic strategies in acrolects (percentages represented in parentheses) TS/TU

Fronting

Repetition

Presentation

Prt/TU

#l=HABw/S 9 #2= SOw/H #3=HAB #4=REV #5=DOR

34/18 15/18 45/24 217/36 103/26

6(17.6) 6 (40.0) 9 (20.0) 6 (2.7) 16(15.5)

6 (17.6) 3 (20.0) 15 (33.3) 55 (25.3) 43(41.7)

22 (64.7) 6 (40.0) 21 (46.6) 156(71.8) 44 (42.7)

1.2 0.3 0.9 4.3 1.0

(1994) #6=TH #7=TH/Ad #8=BL #9=CO #10=TIL #11=COw/T #12=THw/C #13=COw/L #14=Lw/CO #15=ERw/CO

24/44 51/58 30/20 74/35 64/43 15/12 6/8 9/7 16/11 14/9

0 1 (1.9) 3 (10.0) 9(12.1) 4(6.2) 1 (0.6) 4 (66.6) 0 3 (10.7) 2 (14.2)

10(41.6) 21 (41.1) 7 (23.3) 44 (59.4) 30 (46.8) 11 (73.3) 1 (16.6) 4 (44.4) 9 (56.3) 4 (20.5)

14(58.3) 29 (56.8) 20 (66.6) 21 (28.3) 30 (46.8) 3 (20.0) 1 (16.6) 5 (55.5) 4 (25.0) 8(57.1)

0.3 0.5 1.0 0.6 0.7 0.3 0.1 0.7 0.4 0.9

(1980-1984)

TOTAL 717/369 70(9.8) 263(36.7) 384(53.6) 1.04 TS: topic structures; TU: topic units; Prt/TU: number of topic particles per topic unit; /w indi­ cates that the speaker is conversing with another, and each lect is analyzed separately

There are, of course, internal variations related to the topics treated and to the number of participants. For example, a conversation between two women (a seventy-year-old woman [13] and her granddaughter [12]) shows a discrep­ ancy, possibly related to the fact that the older woman leads the conversation,

108

CREOLE AND DIALECT CONTINUA

producing twice as many topic units, and also a higher percentage of topic strategies (1.2 versus less than 1 per unit for T, the younger woman). But T, when speaking to me uninterrupted about three distinct topics [6, 7, 12] is even more restrained in terms of her topic strategies, using in each case less than one topic particle per topic unit. Her discourse style is atypical of the discourse strategies found in other speakers. She primarily communicates in terms of sim­ ple subject/verb/object sequences in (30), with only one presentative phrase that's one ting wid him; though there is frequent subject repetition—he/him occurs eleven times in the sentence, which also exhibits one case of left dislo­ cation (Omar, he bright): (30) Omar he bright, he is very bright, he started talking about two years, they took him out of preschool two years, now they take them at three, he went to school at two, everyday he come home, he can count and rhyme, and he sing, that's one ting wid him, he can remember everyting dey tell him. 'Omar is very bright, he began talking at about two, and started school at two (now they take them at three); whenever he comes home from school, he counts, and sings new rhymes, he remembers everything' [TH 6/94] TH's three samples clearly lower the overall average of topic marking. At the other end of the spectrum, the sermon (text 4) shows the highest incidence of topic marking, an average of six topic strategies in each discourse unit, as illustrated in some of the examples presented above (19). In view of the range of individual speech styles available to individuals, I am led to conclude that an average of two topic strategies and one topic particle per topic unit is represen­ tative of the overall style of communication operational throughout the spec­ trum of lectal styles. CO, the older woman mentioned above, is a habitual "repeater." She clear­ ly highlights her topic through repetition, producing repetitive devices 59.4%, 73.3%, 44.4% and 64% of the time, respectively, in the four separate samples she produced (9, 11, 13, and 21). The highest user of repetition is the eight-yearold boy (19), whose sample is also illustrated in (26). He produces as much as 85% of repetition (thirty-four times for forty topic units). This may provide some insights into the process of creole acquisition, although more children's speech data would be required to confirm this hypothesis. 3.2.1 Lectal homogeneity For comparison, six basilectal samples (table 4.4) are adduced, showing a homogeneous distribution of topic strategies across lects, regardless of sex or

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

109

age. There is a consistent average of two topic strategies per topic unit (189 TS for 91 TUs), as in the acrolects displayed in tables 4.2 and 4.3. The basilectal corpus, however, appears to use in relatively equal amounts the repetitive and the presentative strategies, 46% and 45.5% respectively. This differentiates basilects from acrolects which, as shown in table 4.2, favor presentation: Table 4.4: Topic strategies in Belizean basilects (percentages represented in parentheses) TS/TU

Fronting

Repetition

Presentation

Prt/TU

(1980-1984) #16=ERw/C #17=Cw/E

46/17 45/16

7 (15.2) 3 (6.6)

11 (23.9) 12 (26.0)

28 (60.8) 30 (65.2)

1.6 1.9

(1994) #18=BL #19=ED #20=ERw/CO #21=COw/E

15/12 40/21 18/12 25/13

0 3 (7.5) 1 (5.5) 2 (8.0)

5 (33.3) 34 (85.0) 9 (50.0) 16(64.0)

10(66.6) 3 (7.5) 8 (44.4) 7 (20.0)

0.8 0.1 0.7 0.5

TOTAL 189/91 16(8.5) 87(46) 86(45.5) 0.9 TS: topic structures; TU: topic units; Prt/TU: number of topic particles per topic unit.

The two corpora (acrolectal and basilectai) are shown in table 4.5. Although acrolects use a higher incidence of presentative strategies overall, both sets of lects use a similar average incidence of particles per topic unit (about one per TU): Table 4.5: Comparison of basilects and acrolects TS: topic structures; TU: topic units; Prt/TU: number of topic particles per topic unit. TS

Fronting

Repetition

Presentation

Prt/TU 1.0 0.9 1.0

All Acrolect All Basilect

717 189

9.6% 8.5%

36.7% 46.0%

53.6% 45.5%

OVERALL

906

86 (9.5%)

350 (38.6%)

470(51.9%)

3.2.2 Individual style variability As indicated previously, some respondents have produced several samples. The ideal situation would be one in which the fieldworker can follow her respon­ dents throughout their daily activities, capture on tape their complete repertoires and, thus, fully characterize their speech performance and the degree of vari­ ability they display. In order to come closer to this ideal goal, an essential

110

CREOLE AND DIALECT CONTINUA

requirement is for the fieldworker to become a trusted member of the commu­ nity and be fully aware of the social and cultural attitudes which bind the mem­ bers of that community together. Only when these conditions are met can spon­ taneous recordings in the private and the public domains be completed. Though practical problems always interfere, I have achieved both goals to a large extent through my long time contacts with the Placencia community. Although this study focuses on the acrolectal segment of the Belizean con­ tinuum, the following set of samples illustrates the range of styles produced by four separate speakers, in terms of their use of topic strategies which, as seen above, tend to display crosslectal homogeneity: Table 4.6: Individual variation in four speakers TS: topic structures; TU: topic units; Prt/TU: number of topic particles per topic unit. (percentages represented in parentheses) Sample-Speaker

TS

TU

Fronting

Repetition

Presentation

TH#6 (acrolect) TH#7 (acrolect) TH#12 (acrolect) TH Overall

24 51 6 81

44 58 8 110

0 1.9% 66.6% 6.2%

41.6% 49.1% 6.6% 39.5%

58.3% 56.8% 16.6% 54.3%

CO#9 (acrolect) CO#ll (acrolect) CO#13 (acrolect) CO#21(basilect) CO Overall

74 15 9 25 123

35 12 7 13 61

12.1% 0.6% 0.0% 8.0% 9.8%

59.4% 73.2% 44.4% 64.0% 61.0%

28.3% 20.0% 55.5% 20.0% 29.3%

ER#15(acrolect) ER#16(basilect) ER#20(basilect) ER Overall

14 46 18 78

9 17 12 38

14.2% 15.2% 5.5% 12.8%

20.5% 23.9% 50.0% 30.8%

57.1% 60.8% 44.4% 56.4%

BL#8 (acrolect) BL#18(basilect) BL Overall

30 15 45

20 12 32

10.0% 0.0% 6.7%

23.3% 33.3% 26.7%

66.6% 66.6% 67.7%

As appears clearly in table 4.6, speakers display a certain amount of variation in their individual use of topic structures per topic unit, even within one single lectal group. Yet, they do not evidence a broader variability when comparing one set of lects to another. For example, CO's basilect (21) and acrolect (13) display similar proportions of topic structures, with a higher incidence of repe­ tition in both cases (59.4% and 64%). In fact, there is more differentiation between her various acrolects. She uses more presentation in acrolect 13

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

111

(55.5%) but more repetition in acrolect 11 (73.2%). Her basilects and acrolects are differentiated in terms of phonological and morphosyntactic features, but it is clear that those differences do not affect the treatment of topic marking. Thus, the individual variability of topic strategies is related to the specific conversa­ tional situation, not solely to the lectal level. In spite of context-related changes, average individual speech behavior matches overall group frequencies, all favoring presentation over repetition. This suggests a remarkable homogeneity in the handling of topics in discourse. 3.2.3 Acrolect distinctiveness In spite of the crosslectal homogeneity observed in topic structures, it must be reiterated that acrolects never actually get confused with basilects or with the English lexifier for that matter (see chapter three). Some features commonly found in first dialects (basilects) still occur, though less frequently, in the cor­ responding second dialects (acrolects), while others do not occur at all. Native creole features, such as zero-copula before adjectives or multiple negation, occur with reduced frequencies in the formal contexts defining the acquired acrolectal level, while others characterized by specifically creole lexical items (such as aspectual or locative ds, or anterior me) are excluded from acrolects, although they may become apparent as relexified forms. The samples produced by TH are good examples of such acrolectal char­ acteristics. Her texts, as discussed above, display among the lowest topic struc­ ture frequencies of all acrolects analyzed here; yet, her speech includes a large number of idiosyncratic features which cannot be ascribed to a corresponding non-Belizean standard English and, therefore, demonstrate that she is produc­ ing distinctively Belizean varieties. As briefly illustrated in (14), she exhibits: 1) frequent absence of subject-verb agreement ('de ones dat is smarter; de oder one dey hanging back'); 2) zero-auxiliary/copula ('dey hanging back'), which occurs 52% of the time in Text 4 (13/25), and freely combines with copula use as in (30) 'Omar he bright, he is very bright'. 3) a combination of past marking and non-past marking ('dey wanted to put him in dat tomb before de sun come down'). Absence of marking in simple past contexts occurs 29.4% of the time in 4 (10/34). 4) hypercorrection in past marking or other forms ('if we didn't have a beginning we wouldn't have had a end'), in which the past anterior conditional is a strange occurrence in the context of the situation, since 'the end' [of the human race in an Adventist context, which predicts upcoming final judgment days] occurs after 'the beginning', and therefore has not yet happened.

CREOLE AND DIALECT CONTINUA

112

Thus, acrolects do present distinctive features in numerous phonological and morpho-syntactic aspects. Yet, syntactic/discourse features seem to be char­ acterized by the remarkable consistency which obtains across lects and vari­ eties. In view of the homogeneous patterns of topic marking, which contrast with the crosslectal differences observed by phonological and morphological features, it seems appropriate to explore the longitudinal dimension before assessing the significance of the results presented above. 3.3 The longitudinal dimension The repeated observations that I have been able to conduct in Belize over a fifteenyear span provide insights into the reality of linguistic change in the creole contin­ uum over a limited time span. The longitudinal dimension is represented in table 4.7, showing a comparison of data collected in 1980-1984 and in 1994: Table 4.7: Longitudinal study (1984-1994 texts) TS: topic structures; TU: topic units; Prt/TU: number of topic particles per topic unit. TS

TU

Fronting

Repetition

Presentation

Prt/TU

ACRO/84 ACRO/94

[414] [303]

122 247

10.3% 8.9%

29.4% 46.5%

60.1% 44.5%

2.0 0.5

BASI/84 BASI/94

[91] [98]

33 ■58

10.9% 6.1%

25.2% 65.3%

63.7% 28.5%

1.8 0.5

A glance at the longitudinal data displayed in table 4.7 indicates that acrolects and basilects follow a similar trend toward a lesser amount of presen­ tation concomitant with an increased use of repetition as a focusing strategy. This trend is also reflected in the diminished frequency of topic marking particles. When comparing 1984 and 1994 acrolects, there is an overall reduc­ tion from (2) particles per topic unit to less than one (exactly 0.5 overall). In basilects, the incidence of particles reduces as well from 1.8 to 0.5 per topic unit. The frequency of the topic fronting remains less than 10% in every case, and even seems to have been reduced since the 1980s. However, individual characteristics of the samples must be carefully scru­ tinized before concluding that the samples collected in 1994 are less topic-ori­ ented, or at least are changing topic strategies. In fact, the average of the five acrolects recorded in 1980-1984 is primarily increased by one single long sam­ ple, the sermon produced by the Anglican priest (text 4). This sample displays an extremely high incidence of six topic strategies per unit (217 for thirty-six TUs) and 4.3 particles per topic unit. A sermon is a highly rhetorical genre and

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

113

an emotional context in the Caribbean. Religious arguments involving a certain degree of exaltation may cause speakers to monitor their speech closely because the priest is addressing a church audience. This conjunction appears to be particularly favorable to the emergence of topic strategies, and especially the use of presentative elements, as illustrated in (19). If that sample is removed from the earlier texts, the relative use of the three topic strategies is evened out as compared to the data shown in table 4.5, decreasing the presentation to 47.2% (and by the same token increasing the frequency of fronting to 18.8%). The average topic particle incidence also comes down to 1.1, much closer to the 1994 acrolectal average.10 If table 4.7 is revised to exclude REV (4) from the earlier texts, the results are much less dramatic: Table 4.8: Longitudinal study (1984-1994 texts) TS: topic structures; TU: topic units; Prt/TU: number of topic particles per topic unit. TS

TU

Fronting

Repetition

Presentation

Prt/TU

ACRO/84 ACRO/94

[197] [303]

86 247

18.8% 8.9%

34.0% 46.5%

47.2% 44.5%

1.1 0.5

BASI/84 BASI/94

[91] [98]

33 58

10.9% 6.1%

25.2% 65.3%

63.7% 28.5%

1.8 0.5

Regardless of the reasons for the peculiar topic strategies associated with one individual, this discussion of some individual differences demonstrates the necessity of assessing individual performance, as well as the difficulty of making conclusive judgments of performance and change. On the basis of the evidence represented in the Belizean corpus, two observations emerge: Table 4.8 suggests no drastic changes in topic marking mechanisms over the ten-year span observed, but there is a shift in the choice of strategies that is appar­ ent even in the revised table. Acrolects have decreased fronting and increased rep­ etition so that repetition and presentative elements are almost equally used. The second finding is that second dialects (acrolects) function very much like native dialects in terms of topic marking mechanisms. Basilects are highly favorable to the repetitive strategy, a feature which may have influenced acrolects. This may be due to the vernacularization of the linguistic models pre­ sented informally within the school system, and related to the sheer absence of external models of the standard (see chapter two). There is also a possible sociodemographic explanation for this apparent homogeneity, and the apparent transfer of basilectal strategies to acrolects. It is a fact that in Belize, as elsewhere in the Caribbean, the younger generation has

114

CREOLE AND DIALECT CONTINUA

been almost exclusively raised by grandparents because of the massive emigra­ tion of working adults to the rich countries of the North (as discussed in chapter two). This entailed closer contacts between telescoped generations, and grand­ parents' speech patterns may have had a more direct influence on children, and thus encouraged a preservation of conservative linguistic characteristics.11 3.4 Homogeneity in discourse and second dialect acquisition This analysis of topic marking in Belizean acrolects provides useful insights into the process of second dialect acquisition because acrolects are acquired as second codes by native speakers of Belizean Creole. Since the issue of dialect acquisition has been widely overlooked in the field of linguistics, the only point of reference is second language acquisition (and it is conceivable that there might be no principled difference between the two acquisition types). However, few points of reference for this investigation of a discourse feature are offered by SLA studies, which deal almost exclusively with the acquisition of phonemes and morphemes, phonological rules, morphological patterns, and restricted syntactic rules, such as relative clause formation. 3.4.1 Relatives and parataxis in acrolects Gass (1980) investigates the acquisition of English relative clauses by adult speakers of nine typologically diverse languages, with the intent of testing the universal accessibility hierarchy of grammatical relations as applied to relativization (Keenan & Comrie 1977). According to that hierarchy, the subject is the easiest position to relativize, followed by direct object, indirect object, and so on, with the object of complement position being lowest on the scale. Preliminary observations seem to indicate that the accessibility hierarchy applies to relativization in creole continua as well. Although I have not quanti­ tatively measured their distribution, it is clear that, when they occur, relative clauses follow subject and direct object nouns, whether in basilects (with the use of the invariable relativizer we) or in acrolects, as shown in the following: (31) De life we you live wid alcohol, da notin good 'The life that you live (using alcohol), it's no good' (CO 1994) (32) You know what I talk 'You know what I am talking about]' (CO 1994) (33) What you watching? -Real time stories about people we got problems. 'What are you watching? Real stories about people in trouble' (the televi-

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

115

sion show "Police Stories") [CO 1994] (34) De younger lady dat was one of his patients told me about dis 'The younger woman who was one of his patients told me about this' (TIL 1994) In extensive discourse, relativization has a rare incidence, whether in basilects or in acrolects. The following Belizean sentences and their English counterparts illustrate what I consider to be an avoidance of relativization: (35) My husband was her grandmother uncle, I deliver all her children, I deliv­ er all the grandmother children 'My husband was the uncle of her grandmother whose children I deliv­ ered' [TIL 94] (36) I told him 'Boy, I can't make anoda trip on dat road' we neva had dis road yet, oda road 'I told him: "Boy, I can't make another trip on that road, which was the other road, not the new one" TIL 94] (37) De older one stick up 'is wife at the bank, she was working at the bank.. . .a tink i me ds go ina de bank, she use to work in a bank, right, dat guy's wife, i stick up his own wife Deir father had big banana farm, use to do trucking, i use to build barges 'The older one [of two brothers] held up his wife at the bank where she was working; I think he had gone to the bank where that guy's wife used to work, and he held up his own wife. . . .Their father who had a big banana plantation used to own trucks and build barges)' [MAR/94] It is clear that acrolectal discourse is heavily slanted toward paratactic structures which avoid embedding, especially relativization, and all kinds of complementation. This phenomenon appears to be quite common in widely dif­ ferent types of contact languages. For example, in Swedish American English, the English acquired by Swedish adult immigrants to the United States, has relativizers in subject and object positions, but the habitually invariant relative particle may be deleted in subject position, unlike either standard English or Swedish (Karstadt 1996). In addition, some of the data collected by Karstadt in Minneapolis and Lindsborg also suggest that paratactic strategies are involved

116

CREOLE AND DIALECT CONTINUA

as shown in the following sentences in which topics are developed through var­ ious juxtaposed clauses, presentative devices, and repeated tokens: (38) One is Sout American, eight inches long, one dey call kubabee, dat two inches, a little over two inches long, and dis is da only ruby tro and da only one som [Swedish relativizer] kom here. 'One [of the varieties of hummingbirds] is South American, which is eight inch­ es long, one they call the 'Cuba bee' [?], which is two inches—a little over two inches long; and this is the only ruby-throated [hummingbird] and it is the only one that comes here to Minnesota' (Karstadt 1996) (39) His name was Jonsson, too, so I wen der and, and stayed der for da sommer, da next sommer denn. And, and denn he vas one guy der lived der all da time, he was one guy dat lived der, married to one of his daughter der. And him and I was out in da woods cuttin' timber, uh, went together, and he says, and he told me. 'why don't you write to Olof?' (Karstadt 1996) This pattern is necessarily combined with extensive repetition and presen­ tation, and other methods of referentiahty closely related to the phenomenon of topicalization discussed in this chapter. It seems that topic marking strategies constitute pragmatic compensatory devices for absent syntactic subordination mechanisms. Of course, this may work the other way around: namely, topic strategies may constitute inherent priorities in discourse and, thus, obviate the need for strict syntactic structures such as relative clauses. 3.4.2 SLA and second dialect acquisition in discourse Some SLA studies have started to investigate structures beyond the sentence patterns elicited in completion tasks, that is, in the area of conversational skills in adult learners. Some attention has been given to illocutionary speech acts, such as greetings, closings, apologies, refusals, or requests for clarification (Scarcella 1983). A recent review of the current status of SLA studies states that any work involving pragmatics has focused on illocutionary acts, or on sociopragmatic or pragmalinguistic errors—learners' socially inappropriate behavior when using L2 (Ellis 1994:23; 159-190). Investigating a learner's failure to respond to a native speaker's compliment, for example, is very different from investigating the syntactic structure of discourse in L2 communication (regard­ less of whether the L2 learner thanks the complimenting L1 speaker, or not).

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

117

Schmidt (1983:163) provides a brief but interesting reference to the discourse component as part of his overview of a Japanese learner of English living in Hawaii. He finds his subject's "development of topics particularly impressive," with extensive topic presentation and repetition, an observation which appears to confirm my investigations of acrolect acquisition. Unfortunately, only one sentence is provided to illustrate this pattern, and it is not clear whether L2 acquisition is found in Schmidt's study to display any distinctive discourse structure. Observations of dialect acquisition indicate that discourse features stand out because they are shared equally by the native and the acquired lects. Although acrolects and basilects are differentiated in the areas of phonology and morphology and some limited verb-bound syntactic structures, such as negation, they are not in the case of discourse-related syntactic phenomena. Thus, the major finding related to the issue of second dialect acquisition, and supported by the Belizean data, is that there is no observable difference between first and second dialect discourse structure, at least as far as topic strategies are concerned. All lects use equivalent frequencies of the three topic marking types identified as topic fronting, topic repetition, and topic presenta­ tion. All lects also show a similar preference for repetition and especially pre­ sentation in topic highlighting. The processing of topical information which requires the organization or conjoining of meaningful elements appears to be effected in a unified manner throughout the lectal continuum, a finding that also applies to the longitudinal dimension represented in the brief real time study comparing 1980-1984 with 1994 speech data. A possible explanation of crosslectal homogeneity in discourse is that there is transfer, or the extension of basilectal strategies to acrolects. In conclu­ sion, it can be hypothesized that topic strategies constitute an integral part of the structure of basilectal discourse, and that in the process of second dialect acquisition, acrolects acquire some of the basic structures of the native creole vernacular (Escure 1988). One possible way to verify the validity of this possi­ bility is to compare lectal performance in creole discourse to corresponding aspects of the lexifier language. 4. Comparison with American English The lectal range of Belizean varieties shows great pragmatic homogeneity of topic strategies, but one may wonder about the reasons underlying such consis­ tency, considering that other features, phonological and syntactic, vary exten­ sively crosslectally. A comparison of Belizean English varieties with an exter­ nal standard—conversational American English—permits an assessment of the

118

CREOLE AND DIALECT CONTINUA

putative basilectal transfer hypothesis. The choice of American English is strict­ ly arbitrary (and dictated by convenience), and it is as valid in the context of this comparison as any other variety of standard English, since the point is to provide an external control for Belizean acrolects. This control allows a preliminary eval­ uation of the reasons why basilects and acrolects are similar in the domain of prag­ matic constructions. The following possibilities are envisaged: if American English differs from acrolectal patterns, then the substratai transfer hypothesis is strengthened; on the other hand, if the external English variety exhibits similar pat­ terns, then the substratai transfer explanation would be weakened, though not nec­ essarily eliminated. But in this case, universal explanations may be involved. At the beginning of the chapter, it was hypothesized that a creole continuum would not be found to differ in major ways from casual forms of the lexifier lan­ guage. This theory was based on a previous study of crossreferential strategies, which concluded that various lects in Belizean showed no significant difference from casual American English, although formal varieties (of American English) were markedly distinctive in terms of anaphoric marking (Escure 1994). 4.1 Casual American sample The comparative data summarized in table 4.9 are derived from a conversation between three lawyers, one female and two males. They casually discuss legal problems, e.g., professional ethics, particularly confidentiality when a client is known to have committed a crime such as child abuse or murder; the practice of law by non-lawyers; the obscurity of legal jargon; and status differences between lawyers (public defenders versus corporate lawyers). This context was selected because it represents a natural situation which is intermediate: neither totally relaxed or light-hearted (the topic is serious and linked to ethical issues for young lawyers) nor formal (the conversation takes place between close friends). In spite of the professional overtones of the context, the incidence of topic struc­ tures is quite high, as can be seen from the results presented in table 4.9: Table 4.9: Topic structures in casual American speech TS: topic structures; TU: topic units; Prt/TU: number of topic particles per topic unit. (Topic structures per topic unit = 4.6; topic particles per TU = 3.4) (Percentages represented in parentheses) Speakers

TS

TU

Fronting

Repetition

Presentation

Prt/TU

K(f) J(m) D(m)

198 37 17

31 16 7

8(4) 5(13.5) 0

44(22.2) 8(21.6) 2(11.7)

146(73.7) 24(64.8) 15(88.2)

4.7 1.5 2.1

119

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS OVERALL

152

54

13(5)

54(21.4)

185(76.4)

3.4

An example of the heavy use of the presentative strategy through topic marking phrases, typical of speaker K, is provided in (40). (40) K=So you are, urn, I mean, one thing I should say is sometimes when you get into a case it takes on a life of its own and you lose perspective, I mean it's easy to do, where you, from the outside, if it wasn't your case, you could look at it and see the faults better, where you 're into it and you believe in it, and you kind of get this one, it's hard to remove yourself from it, to remove the emotional investment that you make in a case, so a lot of times I think lawyers just get invested that way in their case, like these people who try to get murderers off death row, and you think why in the world do they want to do that? J=

Good point, why do they?

K= They probably look, well, they look at the system, no, I don't know enough about that, but say these lawyers who get these drunk drivers off when they know they're guilty and they just lookfor the procedural flaws because they just believe that if the system has faults, they look at the sys­ tem and they say if the system that has the flaws, then we should win, they don't look at it and say: this person should be behind bars so he doesn't kill somebody, or she doesn't kill somebody, that's the way they look at it, they always look at the system. [Leg. Conv 1992:7) Two topic units were identified in this segment of discourse; the second topic is introduced at the end of the first topic unit as a natural development of the first. This is also reinforced by J's question which is a repetition of K's warning of topic redirection: TU #1 [lawyers get so involved in their cases that they lose any (moral) perspective] TU#2 [they lose perspective because their strategy is to look for flaws in the (legal) system] Yet, those two simple topics are wrapped inside multiple layers of presentation and repetition (the various strategies are italicized and the topics appear in boldface) as well as paraphrasing, which basically constitutes the remainder of the discourse, but has not been incorporated among topic strategies. In this example, as well as in the overall results of the conversation, casu-

120

CREOLE AND DIALECT CONTINUA

al American English produces the same type of stacked syntax with concomi­ tant multiple topic strategies that was observed in creole-related varieties. Note how the second topic {why lawyers lose perspective) is introduced gradually through expansion of the topic from verb to verb+object complement to verb+manner complement with six repeated versions of the topic: [they proba­ bly look> they just look at the system> they look for procedural flaws ]. The step by step progression in this sentence is not unlike the sentence produced by the Placencia boy: (26) dey bite me, mosquito bite you, dey bite you good, bite hard. Thus, when comparing Belizean and American varieties, both display equally efficient information processing mechanisms which are highly moni­ tored by the participants. The long conversation produced by K., the female lawyer, is as rich in such topic phenomena as the preacher's sermon, with an equally high incidence of 6.3 topic structures per topic unit. She also favors the presentative strategy, averaging 4.7 particles per topic unit. On the other hand, the other two participants in the lawyers' conversation show a more moderate incidence of 2.3 and 2.4 topic strategies, and 1.5 to 2.1 topic particles, which more closely matches the Belizean average. It is also interesting to observe that presentation with various morphemes is more clearly favored in casual American English (as represented in the small control sample used here) than in creole lects in which there is a less marked preference for presentation over repetition. The similar strategies that are operational in first and second dialects can hardly be ascribed to basilectal influence if the same strategies are found to be equally common in the casual forms of the external lexifier. In conclusion, observations of Belizean varieties strongly support the claim that more than basilectal influence is at work in acrolect performance. 5. Pragmatic aspects of syntax The remarkable distinction between formal and informal speech has been strik­ ingly overlooked in previous theoretical approaches to discourse analysis (see chapter one). According to Givón (1979: 228), "the linguist has to face the pro­ found and disturbingly scalar difference between the formal and informal reg­ isters of adult language. The first, the formal register, has been the traditional stock in trade of the linguist, and almost the sole data input." The seemingly chaotic ungrammatical nature of unplanned informal speech leads to parallels with the situation found in early child language.

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

121

The survey of topic marking in casual American varieties and in Belizean leads to conclusions similar to those which emerged from my previous analy­ sis of crossreferential strategies (Escure 1993a; 1994). In that investigation, I show that acrolects generally function like basilects and mesolects: all use equivalent frequencies of pronominal referents and nominal repetition (rather than paraphrasing). Furthermore, in the case of that feature, the Creole continu­ um (including acrolects) is found to function like spontaneous varieties of American English. The investigation of another discourse feature, topic mark­ ing strategies, confirms the results of the previous crossreference study. In both features, which primarily involve the pragmatic dimensions of discourse, there is no evidence that acrolects differ from casual forms of the standard external lexifier. The formal dimension of the lexifier, in this case standard American English, also represented in written English, turns out to be as distinct from casual English as it is from the entire lectal range of Belizean first and second dialects. The main difference is that formal American English uses a high fre­ quency of paraphrasing in order to avoid the redundant crossreferential strate­ gy which is characteristic of creole varieties, acrolects, and others. This crosslectal homogeneity is not tantamount to saying that there is no difference between creole acrolects and casual forms of American English. The significance of this finding is that, although acrolects are found to differ from their standard English counterparts in multiple ways, in phonological, morpho­ logical and some syntactic aspects, the discourse component turns out to dis­ play communicative mechanisms that use the same syntactic strategies as spon­ taneous versions of the standard variety. This is the area where syntax and prag­ matics are blurred, where they are not clearly distinguishable from each other. Although topics correspond to basic grammatical categories including noun phrases, verb phrases, simple verb units, or whole sentences, the repre­ sentation of topics in discourse is subjected to a complex set of syntactic strate­ gies that greatly modify the overall system of language, whether in the creole context or in the context of noncreole casual communication, even when deal­ ing with serious topics such as professional legal issues. In spite of the fact that American English is said to be of the SVO syntactic type, and poor in topic strategies (as compared to French or Chinese, for example), there is an astound­ ing number of syntactic structures that become intertwined with the basic lexi­ cally-filled categorics of the SVO type. The implication of such findings is that, at least in the specific domain of discourse, basic strategies underlying infor­ mation processing are identical in all language types, regardless of their origin through recent creolization or otherwise. As demonstrated, a simple SVO sen­ tence may emerge in discourse as a full paragraphs—defined as a topic unit— which may include as much as twenty-one verbal units (many repeated) as in 

122

CREOLE AND DIALECT CONTINUA

the lawyer's second TU of (38). It is obvious that more than basilectal influence is at work in acrolectal development, since the same phenomena occur in casual forms of the lexifier. An alternative explanation for the uniformity of topic marking across dialects is that universal aspects may be at work in determining the selection of certain features. It is perhaps the case that many of the syntactic (or phonological) devices represented in acrolects are universal features that emerge most easily in uncodified dialects (creole basilects, as well as nonstandard varieties) than in their codified counterparts. They then spread to creole mesolects and acrolects in the process of English acquisition, thus defining a local standard of English which is distinct from the model standard. Note that mesolects and acrolects are uncodified as well because they are a priori assumed to be "English." Finally, since I am dealing here with direct, oral communication, it is also clear that communication strategies favor all of the processes observed in acrolects. The pragmatic component may account for the similarities observed between acrolects and nonstandard varieties of English, and crosslinguistic gen­ eralizations may evolve from the recurrent communicative habits that manifest themselves in discourse patterns (Hopper and Thompson 1980). The influence of communicative strategies leads to the selection of the most unmarked, uni­ versally preferred strategy. In all such cases, a speaker is faced with conflicting systems, and the choices to be made are determined by universal ranking phe­ nomena. Givón (1979:223) defines two extreme poles: the pragmatic mode and the syntactic mode. He explains several types of development in terms of a promo­ tion from loose paratactic (pragmatic) discourse structures to tight grammaticalized syntactic structures; however, the Belizean data seem to indicate that the acquisition of acrolects does not involve a major development from "loose" pragmatic into "tight" syntactic structures. The juxtaposition of Belizean data with American informal speech also indicates that, if there is such a develop­ ment (from pragmatic to syntactic), it is not unilateral. It is plausible to assume that adults do not forget their first acquired communication skill, the pragmatic mode, but merely add to it via the gradual rise of the syntactic mode. This inter­ pretation seems to explain the discourse strategies found to pervade all varieties of Belizean and American English investigated so far. 6. Conclusion Of the three types of topic strategies found in Belizean acrolects, two clearly dominate this type of discourse: repetition of topics, in 37% of the cases, and presentation of topics through the often innovative usage of various particles or

INTERACTION OF SYNTAX AND PRAGMATICS IN ACROLECTS

123

adverbials, in 54% of the cases (table 4.2). These two strategies often co-occur in the same discourse unit. Another finding is that most speakers are consistent in their use of over­ lapping discourse units (table 4.3), a phenomenon which displays parataxis, or juxtaposition of structural units, as the preeminent organizational principle in discourse, rather than syntactic embedding. This is shown to be particularly evi­ dent in the avoidance of relativization. In this respect, acrolects are very simi­ lar to basilects (table 4.5). Differences in the relative frequencies of topic strate­ gies within speakers' repertoires may be due to other factors than lectal level, as shown in table 4.6. This is also illustrated in the discussion of one individual who differs from group norms in his use of presentative strategies (section 3.3). Longitudinal comparisons of 1984 and 1994 Belizean speech data do not display any major structural change. This suggests that the topic strategies iden­ tified here constitute a stable aspect of the parataxis of discourse in Belize. Thus acrolects, which, in regard to phonological and morphological features, are dis­ tinctive from other Belizean lects, and from external English standards, clearly merge with basilects and mesolects at the pragmatic level of overall discourse structure. The hypothesis that substratai parataxis is responsible for the shape of acrolectal discourse is, however, weakened by observations that casual dis­ course in American English displays topic marking patterns of the type preva­ lent throughout the creole continuum. This finding leads directly to the possi­ bility that universal communication strategies are largely responsible for the presence of paratactic topic mechanisms. Observed regularities in discourse patterns eventually reinforce the notion, presented in previous chapters, that the term creole merely captures a relative­ ly new stage of contact-induced language. Language vernacularization, a process actively operational at any stage of natural language processing, leads to a universal emergence of pragmatic strategies in discourse beyond the sen­ tence unit. In order to test the hypothesis that all casually acquired forms prior­ itize pragmatically effective communicative strategies, the second part of this volume will adduce data on second dialect acquisition in another language group: Chinese. Endnotes 1. The logical extraction site is represented by <<.Topics are italicized. 2. This sentence carries a basilectal feature (anterior mε which appears here with a simple past value) within a generally acrolectal context. 3. In all examples, topics are italicized, and structures highlighting topics appear in bold char­ acters (i.e., repeated elements, introductory phrases or tags following topics). 4. It is also perhaps a desperate move to keep the floor, in the face of some disruptive inter-

124

CREOLE A N D DIALECT CONTINUA

ruptions illustrated within the text of (19). 5. Note, however, that still may also have a contrastive value in Standard English as a sentenceinitial adverbial, as in [Still, I don't think I'll do it]. 6. An example of a standard counterpart of this conjunction combination would be "I mean, just because I was opposed to the theory of the case wasn't enough for me to get out of doing the case." This sentence is an excerpt from a casual conversation between lawyers which is used as as a lexifier comparison to the Belizean varieties (see section 4.1). 7. The basilectal level of (25) is reflected in the use of the anterior past marker mε, as well as of the relative marker we, and the serial type verb sequence call me tell me. 8. The combination me wan of anterior + future preverbal markers, corresponds to a past con­ ditional. 9. Habat and Sonia each produced coincidentally the same number of topic units in the same conversation. 10. 1984 data excluding REV: 249-156 = 93 particles for 122-36 = 86 T/U. New average =1.1 11. This notion was suggested by Ian Hancock (p.c. 3/1996).

Chapter 5 Sociolinguistic Perspectives on Chinese Abstract This chapter presents some major historical, geographical, and social aspects of the sit­ uation in China, and an overview of the genetic classification of Chinese language var­ ieties in as much as they relate to the linguistic analysis presented in the following chapters. More specifically, the scene is set for the three varieties to be studied: the Beijing standard (Putonghua), a local standard spoken in Central China (Wuhan Putonghua), and a special version of Suzhou Wu, spoken south of Shanghai. 1. Chinese "In an Olympic event for World's Greatest Language, English might scrape the silver, but Chinese would be unchallenged for the gold" (Sampson 1989: 229). It is not clear that languages can be hierarchically ranked in terms of qualitative criteria, but if "great" is defined quantitatively, then Chinese is certainly the winner on several counts. It can boast the highest number of speakers, more than twice as many as English, even considering the growing number of speak­ ers of English-based varieties around the world. Chinese also has the longest recorded history of any known language, going back more than 3,300 years. Considering the length and complexity of China's history, the traditional concept of Chinese as a unified language appears to be, at the least, paradoxi­ cal and hardly believable. It is more logical to assume that the notion of Chinese as the universal language of China is a purely abstract, or sociopsychological, construct; perhaps that notion is merely based on the fact that all speakers of Chinese share the same written system of ideograms. In a search for the mean­ ing and social function of language variability, the case of Chinese poses fasci­ nating challenges. For example, the following questions necessarily arise in this

126

CREOLE AND DIALECT CONTINUA

context: Does a language evolve in a consistent manner over the course of a 5,000 year human history? Can the development and reconstruction of a lan­ guage be based on written materials, which, by definition, reflect the language of the literate elite? To what extent is a language shaped by its social and polit­ ical history? Can sociolinguistic methodology shed light on the current state of Chinese and, indirectly, on its development? Do crosslinguistic comparisons between an old language, such as Chinese, and young languages, such as cre­ ­les, yield useful information on linguistic mechanisms in language contact? All of these questions will be addressed and applied to very specific case studies documenting in particular aspects of discourse in Chinese. 1.1 Brief notes on Chinese history To understand the concept of Chinese, it is crucial to review some prominent aspects of the long histoiy of China. Yet, it is clear that this overview cannot possibly provide a comprehensive outline of Chinese history, which can be doc­ umented as far back as 2200 b.c. The selection of certain aspects and the exclu­ sion of others present only a partial view of an extremely complex situation, but my intent in doing so is to provide a cursory, but meaningful, background from the perspective that all languages are the product of combined sociohistorical and linguistic influences. China (the word for China in Chinese is Zhōng Guó, the "Middle Kingdom") has a long, checkered history of ethnic diversity, violent conflicts, repeated invasions, and population movements spanning north to south, east to west, and west to east. Their descendants are now called the Hàn people, and live in the highly populated eastern half of the country, an area geographically equivalent in size to the regions of the United States spreading east from the Mississippi to the Atlantic Coast. The Hàn people are often contrasted with the Man (Barbarians), a term which refers to the minorities living around the fringes of the Chinese territory. There are now about fifty-five major ethnic groups on mainland China, but the Han people make up about 94% of the total population. Although little is known about the early history of China, say before 2,500 years ago, it is assumed that most of the south of what is now the People's Republic of China, and even as far north as the mouth of the Yangtze River and the Wu area, was inhabited by populations etlinically related to the Thai or Burmese. Repeated waves of colonization from north to south led to the exten­ sive mixing of peoples and languages, and what is generally described now as the sinicization of mainland China.

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

127

1.2 The warring states: political and demographic background The cradle of Chinese civilization is believed to be in North Central China, in the basins of the Wei, Luo, and middle Yellow rivers, corresponding to the Shaanxi and Henan provinces. The first attested dynasty is the Xia (2200-1700 b.c.), followed by the Shang Dynasty (1700-1066 b.c.) that left written records on bronze or shell materials. Throughout the history of China, periods of extreme confusion alternated with relatively brief eras of peaceful artistic and technical development. Thus, after the capital was moved to Luoyang (Henan), the Eastern Zhou Dynasty (770-256 b.c.) encouraged a flourishing of philosophy and poetry, with such writers as Confucius (551-479 b.c.) and Lao Zi. The Qin Dynasty (221-206 b.c.) unified China and founded the Empire, organizing the country into a net­ work of counties under centralized control, and many colonies developed in South China. In the later part of the Han Dynasty (206 b.c.-220 a.d.), Buddhism was brought from India. The following centuries were marked by competing rulers (such as the three kingdoms of Wei, Shu, and Wu in the third century a.d.) and multiple raids from northern and western nomadic bands between the fourth and the sixth centuries. A dozen southern and northern dynasties succeeded one anoth­ er or competed until the relative unity of the Sui, Tang, and Song dynasties (from the sixth to the thirteenth centuries). The Song Dynasty was particularly marked by the development of the arts and sciences, including the invention of the magnetic compass, gunpowder, and movable type. The thirteenth century (after 1279) was marked by the invasion of the Mongol hordes led by Gengis Khan, who controlled all of China and founded the Yuan Dynasty that ruled China until 1368, moving the capital to the north­ ern site of what is now Beijing (Cambaluc or Khan Buluc) for the Mongols. At that time, Marco Polo and, generally, European merchants and missionaries reached China, and contacts with the west expanded rapidly under the Ming Dynasty that overthrew the Mongols in 1368. The Ming were, in turn, con­ quered by invaders from the northwestern territories of Manchuria, and the Manchus reigned for four centuries, establishing the last imperial dynasty, the Qing (1644-1911). At that time, China extended as far as Manchuria, Mongolia, Tibet, Taiwan, and Turkestan. Western trade deeply penetrated China in the nineteenth century, and China was weakened by European and Japanese exploitation and wars in the nineteenth and twentieth centuries. Russia, England, France, and Japan appro­ priated territories and leases, which amounted to a virtual colonization of China. Some of the most destructive events included the Opium War (18391842), after which Hong Kong was ceded to England; the Taiping Rebellion, a

128

CREOLE AND DIALECT CONTINUA

failed attempt at overthrowing the Manchus (1850-1864); the Sino-Japanese war (1895); the ill-fated Boxer Rebellion (1900)1, that resulted in Western retal­ iation and the further weakening of the Manchu rulers, who were forced to pay heavy taxes to the foreign powers; finally, the Japanese invasion of Manchuria (1931) and ensuing Sino-Japanese war continued until 1945. Manchu rule was abolished in 1912 by a revolution that began in 191l as an army mutiny in Wuchang, Hubei (part of Wuhan). The Chinese Republic was then proclaimed and first ruled by Sun Yat Sen (who later consolidated his Nationalist party—the Guomindang—in Guangzhou), while warlords were tak­ ing over the Beijing regime. A long period of internal conflicts followed. Throughout foreign attacks, civil strife developed between the Nationalists and the new Communists inspired by the 1917 Russian revolution. The Guomindang (Nationalists) recruited primarily among the landowners and gath­ ered under the military leadership of Chiang Kai-shek, who first had its capital in Nanjing (down river on the Yangtze) in 1928, later moving it to Chongqing, Sichuan (upriver on the Yangtze). On the other hand, the Communists led by Mao Ze-dong recruited their membership primarily among the impoverished peasant population, who engaged in the famous Long March (1934-1935) to northwest China (Yanan) where Mao kept his headquarters until his victory in 1949. The civil war that erupted after the Japanese left China in 1945 resulted, in spite of the support provided by the United States, in the defeat of the Nationalists who fled to Taiwan, still a point of contention as events, such as Chinese threats at the time of the 1996 Taiwanese elections, have demonstrated. The People's Republic of China (PRC) was proclaimed on October 10, 1949, and Mao Ze-dong launched a Stalinism-inspired series of economic and political reforms which were meant to restructure all economic and education­ al areas of China according to the egalitarian tenets of orthodox marxism. Between 1949 and 1976, Mao's China was shaken by a succession of political movements, including the Land Reform, the War to resist United States aggres­ sion, the Aid to Korea, the Great Leap Forward, the Four Clean-ups, and the Cultural Revolution (1966-1976). The proletarian revolution and restructuring intended by Mao was inter­ preted as a substitution process that required erasing and eliminating feudal tra­ ditions and imperialist influence and to replace them by new progressive ideas, such as the "four news" (new ideas, new culture, new customs, new habits). Public opinion was bent by slogans, such as "eliminate the pernicious influence of the revisionist line," or "eliminate counter-revolutionaries." This philosophy was applied by the zealous Red Guards as a total rejection and destruction of the imperial past and the ineffectual system of the twentieth century. During the ten years of the Cultural Revolution, chaos and violence tore China apart in a brutal class struggle, which involved criticism, imprisonment, dispossession,

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

129

and rehabilitation of all remaining intellectuals and landowners through politi­ cal indoctrination and field work in country communes. Classical literature, art, and philosophy were also branded as poisonous weeds to be forcibly extermi­ nated, which led to the extensive destruction of ancient temples and other works of art. For example, some of the fifteenth-century Beijing walls, and two of the gates of the Imperial City, were dismantled in 1958 to permit intensive urban development. The subway (still uncompleted in 1996) opened in 1969 along the lines of the old walls, a fitting symbol of modernization defined as eradication of the structures of the past. Mao's death in 1976 was followed by more unrest, and it was not until 1977 that stability was restored with Deng Xiao-Ping (who himself had been criticized and imprisoned during the Cultural Revolution). 1.3, Naming as a reflection of history The brief synopsis of population movements presented above clearly suggests extensive social interaction that is bound to have influenced the linguistic devel­ opment of Chinese varieties. Chinese names are particularly interesting in that they function as mirrors of the society (Lu 1989). According to the traditional Chinese naming practice, the surname, or family name, precedes the given name, or personal name. The overall number of surnames is extremely limited, and indeed has been estimated to a totality of between 6,000 and 8,000 for all of Chinese history (Lu 1989:265). In contrast, given names are virtually unlimited, and usually reflect contemporary political, social, and cultural conditions. The choice of a child's given name may be cru­ cial because it reflects the parents' attitudes, and it may even determine societal acceptance in the child's future life. The personal name may be made up of one or two ideographic characters, but there is an apparent return to the traditional use of single character personal names. The importance of names was particularly felt at the time of the Cultural Revolution. External appearances were a matter of survival, when failure to dis­ play "politically correct" behavior could cost you your life. Green army uniforms became the fashion—and also the only available garments—because Mao wore such army uniforms as an indication that the People's Liberation Army was in control. In a parallel pattern, personal names given at the time reflected allegiance to the Red Army and the Proletarian Revolution. A survey of 714 individuals compared the names of children born during the Cultural Revolution (1966-1976) and after the fall of the Gang of Four in 1976 (Lu 1989). The per­ sonality cult of Mao Ze-dong is reflected in the prevalence of names containing the word dong: for example Jing-dong 'love and respect Mao'; Wei-dong 'safe­ guard Mao'; Xiang-dong 'be loyal to Mao' (Lu 1989:267). However, dong no longer appears after 1970, indicating that people's interest in the movement had

130

CREOLE AND DIALECT CONTINUA

abated long before the Cultural Revolution ended. The other two most frequent naming compounds found around 1966 include j un 'army', for the People's Liberation Army, and wen 'culture' for the 'Cultural Revolution', producing such names as: Wei-jun 'the great Army'; Hong-jun 'the Red Army'; Ai-jun 'love the Army', or Wen-ge 'the Cultural Revolution'. During those years, boys and girls could be given such names, in contrast with traditional distinctions. Girls' names were typically associated with flowers, grace, and purity, as in Bin-bin 'refined and courteous', or Yu-hua 'Jade-like flower'. On the other hand, boys' names were evocative of strength and activity, with common reference to mountains, flowing streams and tall trees, as in Hai-bo 'waves in the sea', or based on items such as shan 'moun­ tain'; min 'agile'; or song 'pine tree'. During the Cultural Revolution, people were either forced to take new names, or they spontaneously renamed them­ selves. However, more recent trends indicate a return to traditional sex-distinc­ tive personal names. Thus, it is clear that even in the very restricted area of nam­ ing, external societal events have tremendous influence. How could it be dif­ ferent in other areas? 2. Aspects of the linguistic history of China Although the written ideographic system of Chinese, and its orthographic ver­ sion of pIny In,2 allows for cross-variety wider communication, mutual oral intelligibility between speakers of distinct Chinese "dialects" is not generally possible. This calls into question the notion of dialects as subgroups of a lan­ guage, as well as the common assumption that Chinese constitutes one single language. The term dialect has often been used to refer to language varieties which can be significantly differentiated. Although varieties within the seven major subgroups of Chinese shown in table 5.2 are to some extent mutually intelligible, for example Chaozhou and Amoy/Taiwanese within the southern Min group (Matthews and Yip p.c, 1996), most Chinese so-called dialects are not. Thus, within the Yue group, a rural dialect is likely to be unintelligible to a speaker of standard Hong Kong or Guangzhou Cantonese. Communication problems are also illustrated in the specific case of three Linxian (Henan) speakers interacting with Mandarin speakers, although it is apparent that the Linxian natives attempt to approximate the phonological and tonal patterns of the official variety (Liu 1979). The Linxian area is located in the northernmost part of Henan province, and thus closer to Shaanxi and Hebei provinces (to the north and west of Henan), which in turn include distinct varieties of Northern Mandarin. Such complex linguistic situations are by no means unusual. Communication is similarly difficult between speakers of different rural vari-

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

131

eties spoken around Wuhan, Hubei. Consequently, the issue of the putative rela­ tionships between different varieties of Chinese is crucial, especially in terms of those varieties that will be analyzed in the following chapters. In view of the long and complex history of China, it can be hypothesized that multiple instances of language mixing (whether or not such mixing actually involved pidginization and creolization) led to the extensive contemporary spectrum of Chinese varieties. An appropriate characterization of this diversity is represent­ ed in the Chinese saying: "speech changes every ten miles" (Wang 1989:197). 2.1 Genetic classification of the languages of China Notwithstanding problems of uncertainty involved in defining genetic linguis­ tic relationships (see section 2.4), table 5.1 presents the six distinct language families that are historically attested in mainland Southeast and North Asia (Blust 1994; Norman 1988; Ruhlen 1991). They are all at least partly repre­ sented in mainland China. In addition, there is an Indo-European language, Tajik, spoken in the northwestern part of PRC (west Xinjiang). This family is not represented on table 5.1.

Table 5.1: Language families in Southeast Asia [N=number of languages spoken in each family; Chinese sites are shown in italics] FAMILY 1. Austroasiatic 2. Sino-Tibetan 3. Tai-Kadai

N 155 258 57

4. Hmong-Mien

4

5. Austronesian

12

6. Altaic

63

SITES OF LANGUAGES SPOKEN NE India to Vietnam; Yunnan (SW China) to Malay Peninsula E.Thailand/Burma (Myanmar) to E. Himalayas (Tibet); PRC S.China {Guizhou, SW Hunan, Guangxi, Guangdong, border Yunnan and Vietnam); Thailand, to N. Malay Peninsula small pockets in S. China (Hunan, Guangxi, Yunnan), and northern SE Asia (Vietnam, Laos) Chamic (Hainan Island, S.China; Vietnam; Kampuchea) Malay (Malaysia); Moken (Myanmar and Thailand); Taiwan N. China; Mongolia; Korea; Japan; SW Asia as far as Turkey.

The consensus is that most of the languages spoken in China belong to the SinoTibetan family which extends all the way from Tibet (and Nepal) throughout China. In addition, some of the languages spoken in the south and southwest of China may be related to Tai-Kadai, Miao-Yoa, or even Burmese languages. There has been disagreement, however, as to the relationship of Tai and Burmese languages to Chinese. The consensus now seems to be that Tai lan­ guages, but not Burmese, are more likely to be related to the Austronesian group of languages than to the Sino-Tibetan group.3 Since the Austronesian homeland

132

CREOLE AND DIALECT CONTINUA

is thought to have been in coastal Southeast China (perhaps in the Fujian region opposite the island of Taiwan), and inhabited by populations ethnically related to the Thai and Burmese people, it is difficult to discount the putative influences of extensive early contacts between Tai-speaking people, and others coming from the north, on the development of mainland Chinese. The genetic classification of Chinese languages is based on such features as lexical roots, syllabic structure, consonantal or vowel correspondences, or tonal systems. For example, at least two facts separate Southern from Northern Chinese in a way that suggests the pervasive influence of a Tai substratum. First, the lan­ guages of South China (the Yue group, including Cantonese) have more tones than the Northern Chinese varieties, and secondly, those southern languages have a modified followed by modifier sequence, more often than the opposite order (mod­ ifier followed by modified), which occurs in Northern Chinese. In those two respects, Southern Chinese varieties function like Tai languages. But there are many exceptions: "all Cliinese languages have head-final NPs with modifier before the modified elements" (Matthews & Yip 1994;p.c. 1996). Matisoff (1994) argues against the view that Chinese and Austronesian are genetically related and maintains that, in spite of the absence of clear regulari­ ties between Chinese and Tibeto-Burman, there are a number of phonologically parallel cognates between Tibeto-Burman and Old Chinese. However, some narrow interpretations of typological differences may have given rise to "exag­ gerated claims of actual divergence" (Blust 1994:73). According to Matthews/Yip (p.c. 1996), the consensus is now that the "Tai-Kadai family and Sino-Tibetan may be part of a larger Austric Phylum along with Austronesian, Miao-Yao, and perhaps also Austro-Asiatic." 2.2 Aboriginal languages in South China Aboriginal languages, the Hmong-Mien (or Miao-Yao) group, which includes several divergent varieties, have long been spoken in the south and southwest of China, and they are assigned to the Austric phylum, not to Sino-Tibetan. Their speakers still live in isolated ethnic communities but have been subjected to a long history of conflicts with the Han people, and have been involved in an extensive diaspora across Southeast Asia (a large group of them, the Hmong or Miao, who had emigrated to Laos, Kampuchea, and Vietnam, now live in France and the United States). Interestingly, a similar number of tones is found in the surviving aboriginal languages (Miao or Yao), as well as in the Tai vari­ eties. The controversy over the affiliation of the Tai and the Miao-Yao groups to Chinese is directly related to the difficulty of explaining apparent similarities. For example, are the roots and tone systems shared by Tai and Cliinese the residue of a common origin or the traces of early loans, and if the latter is true, what was the

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

133

direction of the borrowing? As suggested by Matthews and Yip (p.c 1996), bidi­ rectional influences must have been involved from Tai and Hmong/Mien to Chinese and vice versa, through cultural prestige and bilingualism. 2.3 Altaic Finally, another putative source of influence is that of the Altaic phylum, which is represented in the north and northwest of China. One major subgroup is the Turkic family, of which several languages are spoken in the northwest provinces of Xinjiang and Qinghai, including Tatar, Kazakh, Kirghiz, Uzbek, and Uighur. The Mongolian group is especially represented in the language Manchu of the Tungus subgroup of the Altaic family. Scholars talk of the relatively simple structure of Manchu, which may be due to a long intimate exposure to Chinese, and of large borrowing of Sinitic vocabulary (Ramsey 1987:221). The Manchus invaded Northern China in 1644 and stayed in power until 1911. Although Manchu was the official language of the Qing dynasty, it is assumed that it ceased being spoken in the eighteenth century because the Manchu invaders fully embraced Chinese culture while taking over the politi­ cal governance of China. For instance, literary scholars note very little change in syntax between the Pre-Manchu Chinese and the Chinese spoken at the time of Manchu dominance, as represented in major fiction writings, such as Xiyu Ji Journey to the West, written in 1592, and works written in the eighteenth cen­ tury.4 However, literary evidence does not exclude the possibility of stylistic differentiation. The Manchus may have adopted the Mandarin style in written form, which would not obviate the hypothesis that the spoken vernaculars may have been affected by Manchu (as by Tai much earlier). However, it is difficult to embrace the view that Manchu merely vanished and, furthermore, that it dis­ appeared without leaving even substratai traces in Chinese. This view is conso­ nant with what is known of the development of world languages and implies that Chinese varieties could be the result of merging or borrowing phenomena. Indeed, it has been suggested that the Chinese bă construction, which is part of a SOV construction (S bă O V), could be the result of influence from the SOV order prevalent in Altaic languages (Norman 1988:11). This interpretation gives a plausible explanation to the presently ambiguous word-order pattern of Chinese, variably SVO and SOV. Altaic influence could have initiated the addi­ tion of SOV order in Chinese. 2.4 Problems in genetic classification The Manchus' alleged acceptance of Mandarin as they ruled China in the sev­ enteenth century would appear to involve a situation analogous to that of the

134

CREOLE AND DIALECT CONTINUA

Norman dominance of England in the eleventh century. Although it is often assumed that French eventually yielded to English after two hundred years of French-speaking occupation, it is nevertheless obvious that English has exten­ sively borrowed—lexically, phonologically and syntactically—from French, and that the fairly simple structure of English as it is now (as compared to German, for example) may be the result of some degree of creolization, or at least language mixing, due to intimate language contacts with French.5 Similar observations apply to Manchu: Manchu not only has a very large number of loan words, but it also has a struc­ ture significantly affected by its contact with Chinese. Its relationship to Evenki, the prototypical Northern Tungus language, can be likened to the position that English occupies within the Germanic family, since English structure and vocab­ ulary have also been altered through intense exposure to another culture, namely that of the Norman French (Ramsey 1987:213). Whether or not Middle English or Manchu was a creole6 is not at issue here, and it is controversial whether creoles can be identified through uniquely distinguishable linguistic characteristics, as discussed in previous chapters. The hypothesis that creoles develop through "abnormal transmission" (Thomason & Kaufman 1988:211) is open to debate considering that social disruption seems to be the rule rather than exception in the history of mankind. Yet, in consider­ ation of the extremely long, rich, and turbulent past of China, it may appear that the development of Belizean Creole and Chinese may not be so different in principle, and in nature, after all. Unambiguous and appropriate support for a creole-linked interpretation is not readily identifiable, however, although it can be inferred from indirect cir­ cumstances and, in particular, the absence of available norms or institutions (such as schools) capable or interested in developing such norms. Chinese var­ ieties evolved in a natural, unguided environment because throughout the long history of China the powerful literate elite focused exclusively on the written model. The unified linguistic standards established over 2,000 years ago, at the time of the founding of the Qin Empire (221 b.c.), only applied to the written classical language without any link to the vernaculars (Forrest 1973:214). The establishment of linguistic standards is often assigned to a major piece of schol­ arship, the Qièyùn Rhyming Dictionary (a.d. 601), which "represented a stan­ dard of correct speech common to the educated classes of the north and south China in the sixth century. In other words, it was a cultivated 'Mandarin'" (Pulleyblank 1984:2). It follows from the above observations that spoken vernaculars and their speakers have been largely ignored over the centuries until the violent political upheavals of the twentieth century. Consequently, traditional elitist attitudes,

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

13 5

concomitant with a sharp separation from the uneducated masses, have permit­ ted language to evolve without the interference of a standard model. Independently of the putative creole origin (an issue that may be similarly raised in connection to most languages), there is the ambiguous issue of genet­ ic affiliation. Even though English is still classified as a Germanic language, there might be arguments in favor of linking English to the Romance family as well. A similar debate arises when defining the genetic classification of creoles as newly evolved languages. Whether they belong to the Indo-European family (languages from which their lexical base originates, at least in the case of Caribbean creoles), or to a West African family because of some aspects of their syntax and morphology, is an issue which is not settled. Generally, the very issue of genetic classification is one that may appear suspect when considering that the basis for classification is typically derived from written forms or restricted to formal styles of a language. More specifi­ cally, the major source underlying all historical reconstructions of Chinese is the language presented by the Qièyùn Rhyming Dictionary written in 601 a.d., which was intended as a guide to the recitation of literary texts (Karlgren 1915; Pulleyblank 1984). This observation reflects the widespread—and perhaps unavoidable—fact that the notion of genetic relationships is linked to an assumption of stability and homogeneity, and consequently this approach over­ looks the spectrum of varieties which undoubtedly exists in any language at any time (Norman 1988). 3. Main subgroups of "Chinese" In the light of the preceding discussion, which only brushes the surface of the incredibly complex sociolinguistic history of China, it is interesting to note that the Chinese people generally believe that they speak variants of a single language, and that they are an ethnically homogeneous Han people. The term Hănyŭ (Chinese) suggests the egalitarian ideal endorsed by the new Communist regime of Beijing and reflects the traditional concept of language as a sociopolitical construct. A con­ sequence of this Chinese linguistic identification is that the southern and northwestern "minorities," including the Tibetans, were effectively ignored or excluded from participation in Han activities until recent times. Most of the classifications of major Chinese subgroups conform to the dominant thesis of uniformity. The family tree method has been traditionally adopted to explain the relationships between Chinese languages. More recent­ ly, the hypothesis of lexical diffusion has been proposed (Wang 1969), with off­ shoots such as the diffusion overlapping method (Hsieh 1973). This approach is based on the concept of lexical gradualness in sound change: the more phono-

CREOLE AND DIALECT CONTINUA

136

logical forms two or more dialects share, the more likely they are to be closely related. It is said that: apart from some minor divergences, such as indirect object before direct object in the Wu dialects and Cantonese for which Mandarin (like English) has the oppo­ site order, and slight differences in the order of the negative in potential comple­ ments in some of the southern dialects, one can say that there is practically one universal Chinese grammar (Chao 1968:13). 3.1 Chinese

subgroups

Some of the traditional divisions of Chinese subgroups are presented in table 5.2:

Table 5.2: Major subgroups of Chinese CLUSTERS

LOCALE

1. MANDARIN

(all of North and Southwest) Northern: Beijing Northwestern: Taiyuan Southwestern: Chengdu, Wuhan Lower Yangzi: Nanjing (coastal area around Shanghai, Zhejiang) I: Suzhou II: Wenzhou Nanchang

2. WU

3. GAN (Jiangxi) 4. XIANG (Hunan) Old New 5. HAKKA

PERCENT OF POPULATION

Shuangfang Changsha (widely scattered from Sichuan to Taiwan) Meixian 6. YUE (Guangdong, Guangxi ) Yue-hei: Zhongshan Qin-lian: Lianzhou Gao-lei: Gaozhou Si-yi: Taishan Guei-nan: Yulin 7. MIN (Fujian, coastal areas in South) Northern: Fuzhou Southern: Chaozhou

71.5%

8.5%

2.4% 4.8%

3.7% 5.0%

4.1%

The term Mandarin is still widely used to refer to the majority standard language spoken in China by over seventy percent of the population, and now also actively being taught in Hong Kong in preparation for the Chinese retrieval

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

137

of the territory leased by the British since 1942. Some type of Chinese (SinoTibetan) is said to be spoken by an estimated 927 million speakers, 680 million of them speaking Mandarin (Ruhlen 1991). The "Chinese" continuum is some­ what arbitrarily divided into seven language categories, and displays language distribution in terms of approximate percentages of the overall Chinese popu­ lation, as as shown in Table 5.2, which is based on Li & Thompson (1981:3), Ramsey (1987:87), and Ruhlen (1991:143). 3.2 Some phonological differences in Chinese Chinese varieties are traditionally differentiated from each other in terms of their phonological and suprasegmental inventories. It is assumed that a split occurred toward the end of the Tang era, around 900 A.D. At that time, Middle Chinese voiced stops were lost in northern varieties. Since it is generally assumed that there was linguistic diffusion from the north to the south, varieties spoken in the south are claimed to be more conservative, preserving sounds lost in the more innovative northern varieties. It is true that some southern varieties have preserved a voiced/voiceless stop contrast; only Wu and Old Xiang, and some varieties of Gan still have it (Lü 1980), as shown in table 5.3. However, Yue has lost voiced stops, although it has preserved syllable-final voiceless stops, which have disappeared from Mandarin. The number of tones is another feature which differentiates dialects. Tone is considered to be related to the voicing of initial consonants because of a his­ torical change which effected a tonal split conditioned by the degree of voicing of syllable initials. According to this split, tones derived from syllables with Middle Chinese voiced initial consonants (called yang) are generally lower than those derived from syllables with Middle Chinese voiceless initial consonants (calledyin) (Pulleyblank 1984; 1994). There are exceptions to this claim, but it is overall validated by a computer-based investigation of tone systems in 737 Chinese dialect locations (Cheng 1973). Cheng shows that northern dialects have fewer tones and higher pitch values than southern dialects, and that high tones predominate in all dialects. However, there is extensive variation even within what is identified as a single dialect area. For example, Cheng (1973:96) finds that Northern Mandarin exhibits three tones in seventeen sites, four tones in 291 sites, five tones in thirty-eight sites, and even six tones in one site. Similarly Yue (Cantoncsc) is found to vary between six and ten tones, a varia­ tion which may be mostly due to the analysis itself (a nine-tone system being the most common type). In view of the extensive variability suggested by the findings of Cheng and others, table 5.3 represents a simplified outline of gen­ eral voicing and tonal distinctions that are traditionally used to separate north­ ern and southern dialects. The extensive difference occurring within each sub-

13 8

CREOLE AND DIALECT CONTINUA

group make it impossible for many speakers of different "dialects" to commu­ nicate verbally.

Table 5.3: Some differences between dialects Initial voiced stops MANDARIN WU YUE GAN XIANG -Old -New HAKKA MIN

Initial voiceless stops

No Yes No No Yes No No Yes [b,g]

No Only [q] Yes Yes/No Yes No Yes Yes

N of Tones 3, 4, 5, or 6 5, 7, or 8 6,7, 8, 9, 10 6 4, 5 or 6 4, 5 or 6 6 7

3.3 Evidence of early sociolinguistic factors Many facts of Chinese history suggest the operation of extensive sociolinguis­ tic factors. Karlgren (1915) claims that during the Tang Dynasty, a standard lan­ guage based on the speech of the capital spread as a koine over the whole coun­ try. There is a variety of the Min dialects, called Tang Min which is said to have been introduced by refugee scholars from the north on the downfall of the same Tang Dynasty, resulting in two pronunciations for most of the vocabulary of that variety (Ramsey 1991:218). Another case of migration inevitably leading to linguistic change involves the history of the Hakka people who now live in scattered communities all over Southern China, as well as in other parts of the world, notably in Singapore and in the United States (Hashimoto 1973). The area around Meixian in the moun­ tainous northeastern corner of the Guangdong province, and in adjoining Jiangxi and Fujian coastal areas, is considered to be their homeland. However, the ancestors of the Hakka migrated from the northern plains in a series of waves over the centuries. The name Hakka, that they adopted as their own, means 'guest, stranger' in Yue (Cantonese). Even now, the Hakka are perceived as non-Han people (Man), like the Miao or the Yao, and they are still treated as outsiders. The Hakka dialects have some Northern features; for example, like Mandarin, they have lost their Middle Chinese (seventh century) voiced stops (but so has Cantonese). However, Hakka is more like Yue in other respects, hav­ ing preserved final voiceless stops like Yue (see table 5.3 below). According to Ramsey (1991:111), "Hakka dialects are not much more like Mandarin than Cantonese is." Some merging has obviously occurred in the case of Hakka as

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

139

well; we can assume that this is the rule rather than the exception, and that China was a huge melting pot, whose ingredients still surface separately from each other. No matter what classification is adopted, it is highly likely that mul­ tiple, probably conflicting influences were at work (and still are) in the devel­ opment of Chinese, almost certainly leading to some leveling, as is illustrated in the parallel evolution of the tonal systems of Middle Chinese," and cognate Sino-Tibetan languages, Common Tai, Vietnamese and Miao-Yao: The common tonal development which all these languages share is a striking areal phenomenon, no doubt arising from a high degree of bilingualism between Chinese and the ancestral forms of the present-day Vietnamese, Tai and Miao-Yao languages in what was then South China (Pulleyblank 1994:87) In conclusion, an attitudinal similarity can be found between the arrogant sep­ aration of the Chinese mandarins from the masses, and the dominance exerted in the New World colonies by European plantation owners and overseers over their African slaves, who were denied any access to a formal education. A similar socioeconomic setting is one which is a prerequisite for extensive variability. In the following discussion, I will discuss some specific aspects of Mandarin, and of Wu, since they constitute the data base of the following chapters. 4. The reform of Mandarin: Guanhua, Guoyu, Putonghua Mandarin is said to be. spoken in most of the northern part of China. The Chinese word for Mandarin is Gūanhùa, which means literally 'officials' lan­ guage'—the upper-class language spoken by the literate and governing elite. This language was standardized as an informal "official" language since the fif­ teenth century, mostly for administrative purposes, and has been assumed to be the standard Chinese until the founding of the Republic in 1912. The term Mandarin is still used outside China (in translation) to refer to Chinese, in spite of the fact that it has long been discarded, both in name and in nature, from Chinese usage. The brief outline of Chinese history presented above is suffi­ cient to demonstrate that the present Chinese population and language is the product of heterogeneous groups and languages, whose only common ground was that, until the twentieth century, they were uniformly oppressed and kept uninformed of national and international politics, as well as excluded from the mandarins' educated lingua franca. The masses were suddenly brought to the forefront at the beginning of this century and became abruptly active in Chinese politics, a fact which triggered important linguistic consequences. After the overthrow of the Manchu Dynasty in 1911, the new government attempted to integrate an innovative language policy into the construction of a

140

CREOLE AND DIALECT CONTINUA

new, independent nation free from oppressive Chinese traditions as well as from foreign domination. Mandarin was no longer a valid standard because it was associated with Qing imperialism. It took twenty years of conflicting debates over language policy to publish, in 1932, the Vocabulary of National Pronunciation for Everyday Use. This dictionary established that the new "National Language" (Guóyǔ) would be based on the dialect spoken in Beijing. Even then, disagreements persisted about the validity of imposing a single stan­ dard at the exclusion of other varieties; it seemed to go against the intended pro­ motion of democratic principles in the new China. In the meantime, the new policy about teaching Guoyu in every school was not successful: "In those parts of China where the vernacular diverges most conspicuously from it, the Guoyu shows no signs as yet of any likelihood of supplanting the local speech for any purpose" (Forrest 1973:260-61; first pub­ lished in 1948). It was not until 1955 that the official language policy was fur­ ther clarified as part of the Communist goals of reunification, literacy, and mod­ ernization. The new standard is called the Common Language (Pǔtōnghuà) and is defined in a 1956 official document Directions with Respect to the Promotion of the Common Language: The foundation for the unification of the Chinese [Han] language is already in existence. It is the Common Language [Putonghua], which has as its standard pronunciation the Beijing pronunciation, as its basic dialect the Northern Dialect, and as its grammatical model the exemplary literary works written in the modern colloquial. The principal method of achieving the complete unification of the Chinese language is to promote the use of the Common Language in the cultural and educational systems and in all phases of the daily life of the people (Ramsey 1987:14). It is also stated that Putonghua should include elements from the speech of the workers and the peasants. In this respect, Putonghua thus differs from Guoyu, and appears to be more in harmony with the goal of the new Communist Republic which officially aims at recognizing the rights of the common people. However, this lofty goal of creating a "common language" for all Chinese peo­ ple is far from being realized. Standard Putonghua, presented as a model on television for example, is a foreign language to most Chinese, except to the new literate elite. This also applies in Beijing, where many vernacular varieties cooccur in the urban districts and in the surrounding areas. Putonghua generally holds prestige but especially so with the educated class. In view of the above discussion, and considering the negative historical connotations associated with the term Mandarin, and the fact that Mandarin no longer exists officially in China, I will strictly use the term Putonghua (henceforth abbreviated as PH) to refer to the official language of the People's Republic.

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

141

The most recent in a series of traumatic events affecting China has been that of the "Cultural Revolution" (1965-1976), which ended with the fall of the Gang of Four. This period has had a strong impact on the current sociolinguistic situation for several reasons. First, it created a climate which encouraged the Red Guard movement, leading to the persecution of "revisionist" intellectuals and, in the better cases, to their forced exile to the countryside. The Red Guards, who were mostly urban youths, were also sent to rural areas to organize com­ munes. The result of these rapid population movements has been that individu­ als from different social groups and regions closely interacted for several years in unusual ways. Never before did Chinese citizens, now all tóngzhì ('comrades') travel so much from one end of the country to the other. Through this melting pót process, they became acquainted with language varieties that they would never have encountered otherwise. Another effect of the Cultural Revolution is found in the peculiar develop­ ment of literature in the 1960s. Liberty of expression was severely restricted; thus, writers confined themselves to a narrow range of "politically correct" themes glorifying hard work and dedication to productivity Peasants who had been the foundation of Mao's revolution, were particularly glorified as saviors of the nation, and it is logical to assume that rural vernaculars may have pene­ trated Putonghua. Consequently, part of my analysis of contemporary Putonghua varieties will include some excerpts of the typical vernacular-orient­ ed literature written in the early years of the Cultural Revolution and immedi­ ately after that period. My hypothesis is that those pieces of literature, comple­ mented by casual and semi-formal samples of Beijing Putonghua, will provide adequate insights into contemporary linguistic patterns (chapter seven).

4.1 Beijing The reform of Mandarin originated from Beijing, the seat of the strongest cen­ tral government in China's history. The key national party and government newspapers are printed there, and the national state and political congresses take place there. Beijing is located at the northwestern edge of the great North China Plain, and its population (almost nine million) is second in China only to Shanghai (over eleven million). Nearly half of the population resides outside the urban center. Beijing has been China's capital for the last 900 years. The city first gained importance and prosperity after the connection of the Great Wall during the Qing Dynasty (third century b.c.), which was meant to stop invasions from northern tribes. It expanded during the Han and Tang dynasties, with the construction of a waterway from the Yellow River to the Beijing area. The tenth century was a period of turmoil, when, in spite of the Great Wall, the northern tribes controlled China. During the next century, Beijing became the southern

142

CREOLE AND DIALECT CONTINUA

capita! of the Khitan (Liao) Dynasty, whose name is the origin of the word Cathay, commonly used to refer to China in medieval Europe. The Mongols captured the town in 1122, and expanded it. In the course of the usual warring encounters, the city was destroyed, then rebuilt by Kublai Khan in 1261. The Yuan Dynasty was a prosperous time for Beijing, as attest­ ed by Marco Polo who arrived in Beijing in 1275. When the Ming Dynasty took over in 1368, the Mongols left the city in ruins, but the Imperial seat was moved to Beijing in 1421, new city walls were built, with a fourteen-mile battlement around the city and four gates, and the Imperial Palace was built at that time. The city suffered looting and burning again as the Manchu conquerors (the Qing Dynasty) took over. Those non-Han conquerors undertook to rebuild the monuments to Chinese Imperial power. As the Manchu allowed foreigners to settle in Beijing, the latter ultimately contributed to the deterioration of their regime. The modern revolutionary movements can be traced back to the inva­ sion of Beijing by British and French troops in 1860. For fifty years Russians, British, French, Americans, and Japanese interfered with the imperial regime, eventually leading to the erosion of the imperial system (see also the history outline in section 1.2). The May 4, 1919, movement signaled the emergence of modern national­ ism, and opened the door to the establishment of the Communist Party of China in 1921. The student movement found widespread support from intellectuals, shopkeepers, peasants and workers, but ways to implement nationalist senti­ ments led to political rivalry and confusion. In 1928, the ruling Guomindang changed the name of Beijing ('north capital') to Beiping ('north peace'), when they moved to Nanjing ('south capital'). In 1937, Japanese troops occupied the city until the end of World War II, and in January, 1949, Beijing welcomed the Communist forces. Mao Ze-dong stood atop the Gate of Heavenly Peace, fac­ ing Tiananmen Square, and declared the founding of the People's Republic of China. Mao's huge picture still dominates Tiananmen Square, a tribute to the continuing strength of Chinese-style communism which could not be shaken by the June, 1989, "Democracy" demonstration on Tiananmen Square.

42 Phonological outline of Beijing Putonghua The standard Putonghua pronunciation is defined as based on the speech sounds of the Beijing dialect, and it is mandated that this standard be taught in all the nation schools, with the use of a romanized version of the pīnyīn phonetic alphabet. This system was developed with the help of Soviet linguists, and was officially recognized in 1958. Putonghua has a relatively small number of syllable types (about 400), compared to those of southern varieties such as Cantonese and to the several

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

143

thousand different combinations which occur in English. This limitation in syl­ labic types is due to the fact that consonants primarily occur as syllable initials, the phonotactics of syllable finals allowing only two nasals (dental and velar) and the retroflex liquid beside vowel and diphthongal finals. Table 5.4 illus­ trates the rich consonant system of Beijing Putonghua, and provides the pinyin representation beside a more standard, though simplified phonetic alphabet (Ramsey 1987):

Table 5A: Consonant initials in Putonghua (adapted from Ramsey 1987) Labials Phonetic [p p ' m f Pinyin {b p m f

Dentals

Retroflex

t t' n 1 ts ts' s chr ch'r shr r zh ch sh r dtn1zcs

Palatal

Velar

chy ch'y sy jqx

gk kk'

Glottal h] h}

There are twenty-one syllable-initial consonants, including stops, frica­ tives, affricates, two nasals, and a liquid, representing eight points of articula­ tion (bilabial, labiodental, dental, dental, retroflex, palatal, velar and glottal). The velar nasal only occurs in syllable-final position. Putonghua has two major distinguishing characteristics that set it apart from other Chinese varieties, as well as from other world languages. First, it has no voiced obstruent but has, instead, a non-aspirated/aspirated contrast in obstruents in three sets of stops (bilabial, dental, and velar), and three sets of affricates (dental, retroflex and palatal). As shown in table 5.4, unaspirated phonemes are represented in pinyin by the letters {b d g z j zh} (for voiceless unaspirated stops and obstruents), whereas their aspirated counterparts are written {p t k c q ch}. Second, there is a retroflex series that is unique to the Beijing varieties, but only used in the most prestigious forms of the Beijing standard. In retroflexion, the tip of the tongue is raised behind the alveolar ridge. The four retroflex consonants include one palatal fricative shr, two affricates ch V and chr (one aspirated and one non-aspirated), and one liquid r. Karlgren (1915) describes retroflexion ("articulation supradentale") as a kind of assimilation to a following velar vocalic element which affected Northern Mandarin /k/ around the end of the Tang period (900 a.d.), applying first before back vowels, and later generalized to affect palatals before front vowels. Some scholars believe that the process of retroflexion may have occurred in successive stages (Forrest 1973:194; Pulleyblank 1984). Others assume a quasi-simultaneous shift. The "Pre-Mandarin Great Consonant Shift" would have changed velars to palatals before /i, y/, palatals to retroflexes before other vowels, and retroflexes to dentals before schwa, affecting all

144

CREOLE AND DIALECT CONTINUA

northern dialects collectively called Mandarin (Chen 1976:165-166). However, retroflexes do not exist in any variety outside the most elegant Beijing pronun­ ciation, and most Chinese learning Putonghua as a second dialect do not use retroflex consonants and, thus, do not distinguish retroflexes from the dental affricates or fricative. The pī nyīn transcription does not accurately reflect important vowel dif­ ferences, however, which are primarily predictable on the basis of a preceding consonant. For example, the symbol {i} has at least three distinct phonetic val­ ues, depending on whether it follows a dental initial — in this case, it is a sibi­ lant version of a non-high vowel, appropriately described as the "buzzing of a bee" (Ramsey 1987:45)—; a retroflex —in this case, it is a retroflex version of a non-high vowel; or occurs elsewhere —in this case, it is a simple high vowel. The four tones of Standard Putonghua include: the first tone, high and level, represented as ( - ); the second tone, high and rising ( ' ); the third tone, low and rising ( ˘) the fourth tone, high and falling ( '). The purpose of this brief summary is to permit a reading of the texts included in the following chapters. 5. Varieties of Mandarin outside Beijing: Wuhan Since Putonghua is now centered around Beijing Chinese, this term does not apply to related vernacular varieties spoken outside Beijing. For lack of a bet­ ter term, and since the goal of this study is not an exhaustive survey of the lan­ guages of China, I will use the term Mandarin as used traditionally to refer to those varieties which are spoken outside the Beijing urban community: north (Liaoning, Jilin, Helongjiang), northwest (Hebei, Shanxi, Shaanxi, Ningxia, Gansu), and south/southwest of Beijing (Shandong, Jiangsu, Anhui, Henan, Hubei, Guizhou, and as far west as Sichuan near Tibet, and Yunnan near the Vietnam and Myanmar borders). The so-called Mandarin group includes four major subgroups of dialects (or languages), as shown on table 5.1. As discussed above, the common term assigned by linguists to this group of languages implies a certain homogeneity which is more likely to be related to the sociopolitical context than to linguistic reality, since most of those varieties are not mutually intelligible. For example, certain varieties spoken in Hubei, are not understood by Beijing natives, although Hubei natives usually understand Putonghua, especially since it is mandated in the school systems nationwide. In most places, there is a name for the local language or languages.

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

145

5.7 Geographical and demographic outline of the Wuhan area Wuhan (population five million) is the capital of Hubei Province, and its major urban center. It is located in Central China, at the confluence of the Han and the Yangze rivers, roughly 1,000 miles from each of China's four largest cities: Beijing to the north, Guangzhou to the south, Shanghai to the east, and Chongqing to the west. The name Wuhan is a blend of the names of the three neighboring metropolitan centers of Wuchang, Hankou, and Hanyang. The three cities were connected in 1957 by the Changjiang bridge, the first and only to cross the Yangtze River between Chongqing to the west and Nanjing to the east. Wuchang, located east of the Yangtze, is the oldest city, going as far back as 200 b.c.; and was an important administrative center in the fourteenth centu­ ry; Hankou was a fishing village in the third century b.c., whereas Hanyang developed in the seventh century. 5.2 Wuhan language varieties Varieties spoken in Wuhan, called Wuhan-hua 'the language of Wuhan', are examined in chapter eight; therefore, only a few relevant facts will be present­ ed. According to table 5.2, Wuhan Chinese is part of the northwestern Mandarin area of the Loess Plateau along the Yangtze River. Like Beijing varieties, and contrary to southern Chinese, Wuhanhua does not include voiced stops in its phonological inventory, but it does have the two series of voiceless aspirated and unaspirated stops. Although it has been claimed that tone systems are most­ ly identical in all Mandarin varieties (including four tones), pitch patterns wide­ ly vary, often leading to significant consonantal and vowel shifts. In reality, lin­ guistic heterogeneity prevails in Wuhan as well as in other parts of the country. The Wuhanhua spoken by educated natives (which is not Putonghua) is very different from varieties spoken by local peasants, who sell their produce at the free markets, or by shopkeepers around town. When traveling around Hubei, I have often been told that Wuhan natives do not understand many of the varieties spoken in the surrounding rural areas. There is an important phonological feature which differentiates Wuhan varieties from Beijing Putonghua. Wuhan does not have the series of (four) retroflex consonants, two affricates, one fricative, and one liquid [chr, ch'r, shr, r] that chacterizes Putonghua, as stated above, and absence of retroflexion also characterizes most other varieties of Chinese. An important consequence of the lack of retroflexes is observable in its effect on contiguous vowels. Thus Wuhanhua (WH) contrasts many lexical items which are homophonic with retroflexes in Putonghua (PH). The following table illustrates some of the sig­ nificant differences found between WH and PH. In WH, there is a two-way

CREOLE AND DIALECT CONTINUA

146

complementary contrast in consonant+vowel sequences: dental affricates pre­ cede mid-back vowels and palatal affricates precede high front rounded vowels. PH only has a single combination: retroflexes precede high back vowels. In addition, there are extensive tone differences, resulting in PH homophones which do not exist in WH (Escure 1987): Table 5.5 Wuhan reflexes of Putonghua retroflexes WUHAN [cou] [] ['ou] [c'y] [sou] [sy]

PUTONGHUA [chru] [chru] [ch'ru] [ch'ru] [shru] [shru]

PINYIN zhu zhu chu chu shu shu

GLOSS help live begin go out uncle book

The acquisition of retroflexes and concomitant homophony is only one of the many problems which interfere with the learning of Putonghua by native speakers of Wuhanhua.7 In reality, even that educated minority who has gained access to the acquisition of spoken Putonghua eventually produces a different variety, which is somewhat intermediate between Putonghua and Wuhanhua, a variety which car­ ries local identity and is thus eminently comparable to an acrolect.

5.3 Wuhan data base The data on which I base my analysis of the Putonghua acquired as a second dialect in Wuhan were collected during the academic year of 1985-1986 which I spent in Wuhan. Speech data were collected among a group of young teach­ ers, ranging in age from twenty-three to forty, who were my students. Most of those individuals had spent several years in rural communes or had variously been involved in the violent confrontations of the Cultural Revolution. They were certainly aware of the broad range of linguistic variation. 6. Wu varieties and Suzhou Wu 6.1 History of Jiangsu province The Wu varieties are spoken by some eighty million people in a relatively small but densely populated area around Shanghai, comprising Zhejiang and part of

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

147

Jiangsu province in the Yangtze Estuary. This is the most densely populated area in China, whose major urban center is Shanghai, with a population of eleven million. Suzhou, which was settled over 3,000 years ago, is one of the oldest and most beautiful cities, and is still called "the Venice of the East" because of its network of canals and ancient stone bridges. The ancient capital of the king­ dom of Wu (the "Barbarian" state) that flourished by the end of the seventh century was located near the modern city of Suzhou. The people of the king­ dom of Wu were regarded as "barbarians," navigators, and skilled in the art of tattooing. Suzhou was particularly renowned in China for its silk industry, as early as the Tang Dynasty (618-906 a.d.), and it has repeatedly been the prize target of China's invaders. It became a Japanese concession at the end of the nineteenth century; and was again occupied by the Japanese during World War II and then it was held by the Guomingdang until 1947. For a long time, Shanghai was merely Suzhou's port (Forrest 1973:240). The development of Shanghai only began at the middle of the nineteenth century because of foreign concessions and capital, and then expanded exponentially to become the flourishing but overcrowded city that it is now—China's most popu­ lous city. Because of its expansion and extensive immigration from southern China, the language of Shanghai developed independently from that of Suzhou and has gained great prestige. On the other hand, the Suzhou dialect that was con­ sidered an upper-class dialect is now considered old-fashioned and quaint. 6.2 Wu As represented on table 5.2, Wu is classified as a separate language group, even though it has strong similarities to the neighboring varieties of Xiang and also to Hakka. It is also related to Mandarin, of which it is usually considered to rep­ resent a more conservative stage, because of the surviving presence in Wu of the voiced stop series in word-initial position, which have long dropped from Mandarin. However, not all studies agree on this point; Forrest (1973:240) claims that voiced stops occur only in medial position in Wu. The classification of Wu in a separate category appears somewhat arbi­ trary and unsubstantiated, especially when looking at vernacular varieties. The similarities or differences observed between Wu and any type of Mandarin range over a linguistic continuum which is also characteristic of all Mandarin varieties per se. For example, languages spoken in Hubei or Sichuan contain just about the same amount of distinctiveness or resemblance as is observed when comparing Suzhou Wu and Beijing Putonghua. The Wu group is said to include five varieties related to major urban areas and their surrounding com­ munities, namely, those of Wenzhou, Ningpo, Taizhou, Shanghai, and Suzhou.

148

CREOLE AND DIALECT CONTINUA

As was observed for Wuhan varieties, Wu varieties do not have any retroflex consonants, with similar problems involving differential phonetic representa­ tions, as well as distinct tone systems and concomitant homophony. This phenomenon was observed in recent dialect surveys conducted in China; although the purpose of such studies is clearly prescriptive, as seen in the following excerpt from a collective survey: Shanghai people do not distinguish 'zhaodao' (find), from 'zaodao' ('arrive', 'shiren' (poet) from 'siren' (private). The reason is that they pronounce zh, ch, sh, as z, c, s. In Shanghai dialect, there are no such sounds as zh, ch, sh. Any word beginning with zh, ch, sh and z, c, s in Putonghua is pronounced as z, c, s by Shanghai speakers. However, zh/z, ch/c and sh/s are different, and the differences are important in distinguishing word meanings. We cannot, however, replace 'zhaodao' (find) with 'zaodao' (arrive early), nor 'shiren' (poet) with 'siren' (pri­ vate) (Fudan Dialect Survey 1958:27). The underlying message of such analyses is that the Shanghai dialect is inade­ quate at conveying important semantic distinctions ("we cannot replace. . ."), a prescriptive method which conforms to planning strategies endorsed by most nations. Of course, these grammatical normative recommendations have little impact on the actual evolution of the language. In fact, Shanghai people and all Wu speakers are very proud of their linguistic idiosyncrasies and their separate identity, which is represented primarily in those features. My approach to.Wu will be very focused. First, I will exclusively study a specific variety of Wu spoken in the Suzhou area, namely the very stylized Wu used in traditional story-telling called pingtan. It is an important cultural compo­ nent of many Chinese regions, which has its roots in a long tradition akin to Chinese opera. As discussed in chapter eight, this variety of Wu is both formal and vernacular, with a great deal of style-shifting due to the humorous presentation of princely, noble characters being subjected to lowly human emotions. It represents a type of second dialect acquisition because it must be narrated in certain traditional ways and is broad­ cast as a regular radio program. 7. Conclusions The extreme heterogeneity of Chinese ensures that any approach will be chal­ lenging and intimidating. I have attempted to address some important issues from a sociolinguistic and comparative perspective by strictly focusing on one single aspect of discourse structure, namely the mechanisms underlying topic marking processes, thus providing a counterpart to the previous analysis of the same feature in Belizean Creole.

SOCIOLINGUISTIC PERSPECTIVES ON CHINESE

149

Endnotes 1. The Boxers were members of a secret society founded to foment popular rebellion against the Manchu's imperial reign, but they limited their immediate objective to the expulsion of foreigners and, thus, formed an uneasy alliance with the imperial government. 2. The pīnyīn (romanization) orthographic representation is the official spelling system since 1958 commonly used for Mandarin. It includes the symbols 'b d g' to refer to the three unaspirated voiceless stops, whereas 'p t k' correspond to the aspirated voiceless series. 3.Tai is classified among the Austro-Tai subgroup of the Austric phylum, whereas Burmese is assigned to the Tibeto-Burman branch of the Sino-Tibetan family (Ruhlen 1991). 4.Thanks to Chun-Jo Liu (p.cl995.) for this reference. 5. However, Thomason and Kaufman (1991:123-1237; 263-269) deny that there was any sig­ nificant superstratum interference from French in the development of English, and even deny that English is a mixed language. 6.Gorlach (1986) argues against the claim that Middle English was a creole, mostly on the grounds that criteria for defining a "creole" are contradictory and undefined, view also shared by Thomason and Kaufman (1991:211) and apparently based on the assumption that creoles develop through "abnormal transmission." 7. The emphasis in education is naturally on written aspects of Putonghua, and indeed, it takes years to learn a significant number of Chinese ideograms.

Chapter 6 Topic Mechanisms in Chinese: An Overview Abstract As part of my evaluation of the relative effects of first dialect interference versus lan­ guage universals on the process of second dialect or language production, I will now take a close look at a group of Chinese linguistic varieties which are genetically unre­ lated to the creole continua so far analyzed. This chapter provides a background refer­ ence to the analysis of second dialect acquisition in Chinese by first presenting an overview of the linguistic mechanisms which contribute to the marking of topics in Standard Chinese.

1. Introduction Chinese is generally considered to be a particularly topic-prominent language that uses a variety of devices to highlight the topic in a sentence. This chapter will investigate the notion of "topic," theoretical interpretations incorporating that notion, and the various strategies available in Chinese to mark topics. Such an outline is a prerequisite to the detailed investigation of topic mechanisms in Standard Chinese (chapter seven) and to the analysis of the same process in sec­ ond dialects of Chinese (chapter eight).

1. 1 Theoretical interpretation of topics in Chinese Traditional descriptions of Chinese refer to the frequent incidence of "word inversion," or "transposition," from Archaic Chinese times onwards, "for the sake of giving prominence to the predicative part" (Forrest 1973:72). According to Chao (1968:69), "the grammatical meaning of subject and predicate in a Chinese sentence is topic and comment, rather than actor and action." Li & Thompson (1981:15) say that "one of the most striking features of Mandarin

152

CREOLE AND DIALECT CONTINUA

sentence structure, and one that sets Mandarin apart from many other lan­ guages,1 is that in addition to the grammatical relations of 'subject' and 'direct object', the description of Mandarin must also include the element 'topic'." They argue that the Mandarin sentence must be analyzed in terms of its topiccomment relation rather than as a subject-predicate combination, as is usually the case for English (Li & Thompson 1981:19). Xu & Langendoen (1985) also discuss the grammatical relations that hold between topic and comment; they conclude that current syntactic theories, such as Chomsky's GB (Government and Binding) framework, do not readily account for all Chinese topic structures. They claim that because Chinese has no lexical complementizers without semantic content (like English that), there is no reason to postulate a base rale including a COMP constituent, but they pro­ pose instead to include a TOP (topic) category in the base, a system that will "leave open the possibility that material may move directly into the TOP con­ stituents from the comment clause" (Xu & Langendoen 1985:3). According to this approach, the only condition holding for topic structures in Chinese is that the comment be a well-formed clause which "relates" to the topic; thus, the topic comment relation is analogous to the relation of a pronoun to its antecedent but with no co-indexing required. They do not make any commit­ ment as to the possibility of a movement analysis, but say that "at least some topic structures in Chinese are generated without movement, namely those whose source structures would be ungrammatical no matter where the topics originate within the comment clauses" (1985:27). The issue of co-indexing is another reason why existing syntactic theoret­ ical frameworks such as GB—or rather UG (Universal Grammar), as GB's expanded version is now commonly referred to— do not appropriately describe Chinese (or creoles for that matter, as seen previously). UG specifies that coindexing must exist between the topic and some other element (empty or lexicalized) in the comment. However, there are large quantities of Chinese struc­ tures in which the comment contains elements which are not anaphorically bound to the topic, even though there is always some kind of semantic relation between them,2 as shown in (1): (l)Jiùshì yī shuō ba, j i āzhǎng dōu bǎ nǐmén dāng háizi kàn. [TI — speak- TE, parents AP -TI you take child see] 'As far as talking is concerned, parents treat you as children' or 'Parents talk to you as if you were kids' (Beijing TV 85: MSS.39:20)3 It is clear that in (1) the verbal topic shuō 'speak' is not directly bound to

TOPIC MECHANISMS IN CHINESE

153

any item in the comment [parents treat you like children], although the broader context of discourse (an evaluation of a science program for middle-school stu­ dents) provides an unambiguous interpretation of the relevant aspect of parentchildren communication. The issue of co-indexing has been extensively debated by syntacticians interested in Chinese, although primarily restricted to intrasentential structures, which in fact removes from the discussion crucial pragmatic information. For example, C. T. J. Huang (1984) attempts to fit zero anaphora in the GB para­ metric theory of UG, treating the null anaphor as an abstract trace left by a fronted empty topic; the zero anaphor is also taken to be an empty pronoun pro, in accordance with the GB system. Chomsky represents non-overt nominal expressions, also called empty categories (ECs) in terms of two types of abstract features for NPs, anaphor and pronominal, and those empty categories are abstract entities assumed to be syntactically present, in spite of the fact that they are phonetically non-existent. However, such formal explanations have not been readily accepted as valid. The issue of the co-indexation of the abstract pro is not easily solved because zero anaphora are not related to a clear control domain (as required by UG), and can receive various interpretations, especial­ ly in a spontaneous discourse context (Y. Huang 1994; 258). Furthermore, the empty topic/empty anaphor interpretation fails to explain a language that gives such prominence to topics, as well as the freedom to use those topics in various positions, through fronting, repetition, and presentation, as will be fully discussed below. Assuming the validity of the standard defini­ tion of a topic as the most prominent item in the sentence or discourse unit, how can empty entities underlie the most salient lexical constituents, as well as their referential elements? Attempts have been made at developing pragmatic theo­ ries of Chinese, following the general belief that: There seems to exist a class of languages (such as Chinese, Japanese and Korean) where pragmatics appears to play a central role which in familiar European lan­ guages (such as English, French and German) is alleged to be played by gram­ mar. In these 'pragmatic' languages, many of the constraints on the alleged gram­ matical processes such as intrasentential anaphora are, in fact, primarily due to the principles of language use rather than rules of grammatical structure (Y. Huang 1994:259-60). For example, unification-based Lexical Functional Grammar (LFG) is used by C. R. Huang (1990) to account for some idiomatic discontinuous struc­ tures, which cannot be interpreted as separate units by a GB analysis. Indeed, complex sequences can be fronted, as will be shown in the following examples. LFG also accounts for anaphora and topicalization phenomena, according to C. R. Huang (1990:294), who states that "topic phrases need not be identified with

154

CREOLE AND DIALECT CONTINUA

grammatical functions, they need only contain representations of a single TOPIC function." On the other hand, Y. Huang argues in favor of a "neo-Gricean pragmatic theory of anaphora," which he primarily applies to intrasentential mechanisms, thus primarily addressing the issue of reflexivization, although he claims that discourse anaphora could also be accounted for by the same approach. He pro­ vides an analysis of anaphoric production and resolution, the encoding and decoding of anaphoric relations, invoking pragmatic principles related to appro­ priateness, efficiency, effectiveness, reduction and economy, in terms of "speaker's maxims" (the Q-principle—do not say less than is required; the Iprinciple —do not say more than is required; and the M-principle—do not use a prolix, obscure, or marked expression without reason). This is combined with a constraint, the "Disjoint Reference Presumption" (DRP) which states that "the arguments of a predicate are intended to be disjoint, unless marked otherwise" (Farmer & Harnish. 1987). There is, however, a striking vagueness about such principles, especially because they are dependent on "world knowledge," a loose concept by defini­ tion, as represented in the following discussion of the DRP principle: "It could also be equally strongly argued that the DRP is based on world knowledge, given that the fact that one entity tends to act upon another could be due large­ ly to the way the world stereotypically is" (Y. Huang 1994:130). Spontaneous discourse appears to be frequently redundant, verbose, lengthy, fragmented, repetitive, disjointed, but all such properties may actually contribute to the effectiveness or emotional value of a narrative. Therefore, communication may be achieved through inference rather than reference. Functional analyses widely differ in their formal paraphernalia, but they all agree—at least as far as analyses of Chinese are concerned—in rejecting strict­ ly structural accounts; they all recognize the role played by external factors in processing information, beyond the purely formal concerns of generativists. Regardless of what may constitute the best theoretical interpretation of topics in discourse, it seems likely that the relation between topic-comment in Chinese is established via semantic and pragmatic inferential mechanisms, rather than by strict syntactic referential systems. This brief summary of some accounts of the thorny issue of Chinese top­ ics makes it clear that neither strictly syntactic analyses, nor general discourse approaches are capable of accounting fully for the tightly linked topic-comment constructions, which are common to all Chinese varieties. It makes sense, how­ ever, to embrace the compactness of topics and their related comments, a com­ plex constituent previously referred to as the topic chain. The notion of "topic chain" was developed by Tsao (1979) and later adopted by Shi (1989) as well.

TOPIC MECHANISMS IN CHINESE

155

According to Tsao, the topic extends its domain to a sequence of several sen­ tences. The topic chain is equivalent to a juxtaposition of topic-comment constructions, and it is thus considered to be a discourse-level phenomenon that has no place in syntax. However, Shi attempts to bridge the gap between dis­ course and syntax by arguing that the "topic chain is actually a basic unit in Chinese syntax" (Shi 1989:223). Since this analysis is primarily concerned with meeting the requirement of observational adequacy, which seems to constitute a first step required to achieve explanatory adequacy, the interpretation adopted here will be that the basic struc­ ture of Chinese must be of the topic chain sequential type (a concatenation of top­ ics and their attached comments), rather than of the subject-predicate type, with no assumption of any movement phenomena. This position is strongly supported by the high frequency of topic-first sequences, and even more so when repeated observations of natural speech in various dialectal contexts are adduced. It is obvi­ ously not restricted to Chinese, since the same unit of analysis, the topic unit was found to provide the most satisfactory analysis in the study of Belizean Creole dis­ course (chapters three to four). Whether this chain-like unit of analysis fits within syntax, or discourse will be discussed below.

1. 2 Types of topic strategies The categories outlined in the previous chapter for the description of creole acrolects are equally productive and meaningful in the analysis of Chinese struc­ tures. It is the case that practically any constituent can function as topic, including pronouns and nouns, whether they function as subjects, direct or indirect objects, verbs, adjectives, or adverbs. The topic-first position is a regular feature of Chinese syntax, and topic repetition is widely used. In addition, there is a broad variety of particles, or presentative elements, which can occur in the sentence, either to emphasize the topic, to mark the juncture separating topic from comment, or to underscore the termination of the comment. Topic salience, which will be referred to by the general term topicalization, can be achieved through various mechanisms which can be divided into three major strategies: 1) an especially prominent placement of the topic in sentence initial position, often at the immediate beginning of the sentence or near the front, that is, a topic may be fronted in relation to its logical position, though actually fol­ lowing other constituents; 2) the repetition of the topic which may occur in various forms, nominal or pronominal; and 3) the use of additional morphemes as presentative particles which accompa­ ny or mark the presence of topics.

156

CREOLE AND DIALECT CONTINUA

These three types of strategies are often combined in the same sentence; thus, a topic may be fronted or repeated, and accompanied by a topic particle. The pragmatic result is that the major constituents of the Chinese sentence are always identifiable. This is an important notion, because it is clear that the canonical Indo-European definition of a sentence as consisting of noun phrase/verb phrase sequences does not fit the structure of spontaneous Chinese discourse, no more than it was found to apply to creole discourse. The general strategies for topic marking are briefly illustrated below, and further analyzed and discussed in the following chapters, in the context of three corpora of writ­ ten and spoken data.4

2. Topic fronting A topic (in italics below) may be placed in a prominent position if it occurs first in the utterance (topic unit), as is (2) and (4). A wide variety of constituents appear to be subjected to this initial placement, affecting, in addition to simple nominais such as (3) and (5), verbal predicates with no overt subject, such as (1) and (4), and complex "possessive" constructions, such as (2) and (4). In fact, this construction applies, with a general function of modification, beyond a sim­ ple possessive phrase and to a wide range of prenominal qualifying phrases, including the equivalent of English compound structures, adjectival, adverbial, and relative constructions; see for example (11) below as a representation of a complex two-level fronting of a topic with its preposed relative clause includ­ ing another level of fronting. The particle de (glossed as G) signals a link between the preposed modifier and its head. A topic is usually introduced by a presentative element, as in (1) and (3). It can also be fronted and placed after a sentence-initial subject, as in (5) to be discussed below. Four distinct subtypes can be identified in this category of topic strategy, which assigns a prominent position to topics in discourse: 1) simple fronting with gapping, or co-indexed empty element (the traditional topicalization movement); 2) fronting with lexically-filled pronominal anaphora (the traditional left dis­ location movement); and 3) fronting without anaphoric binding in the comment.

2.1 Topicalization: Fronting with gapping When fronted, the topic, which can be any constituent or group of constituents, is placed in sentence-initial position (2), or its close vicinity, as in (3) where the NP object topic is fronted, but following an introductory adverbial (fǎnzhèng)

TOPIC MECHANISMS IN CHINESE

157

which also has a presentative function: (2)

Shī

shē

ng-de

guānxì,

[Teacher student-G

yě

rapport, AP

x ī wà ng nǐmén z à i =

hope

you

at]

=j i érnù zhōng gě i y ǐ yīding-de zhōngshì, shìba [program middle give certain-G emphasis, —EP] '1 hope that you will give in your program a certain emphasis to teacher-stu­ dent contacts' (Beijing: MSS 4: 1-2, 1985) (3)

Fǎnzhèng, nǎ jí yě bù quē. [really (TI) any part AP not miss] '(I wouldn't) want to miss any part (of it)' (Beijing TV B.1.3, 1985)

Pronominal subjects and objects which are not topics may be omitted, resulting, for example, in surface sequences of the (Adv) OV type, shown in (4) and (7), in which the subject ŏ (I) is missing. Another superficial combina­ tion may be SV, as in (4), which is overtly missing some crucial elements. The surface topic kán-de dong dehuà is, in fact, a serial verb sequence, and really a sentence without overt subject or object, yet punctuated by a special topic ender (dehuà): (4)

à-de d ǒng dehuà, hěn hǎo xiào-de[see-G understand—TE, AP good laugh-G] 'If (we could) understand (it), (it would be) funny (Beijing 1985)

Beyond the general pattern of topic fronting, a great flexibility is possible in surface sequences, as already represented. When the topic is an object, it is often fronted toward the beginning of the sentence, as in (2) and (3), yielding an OV or Adv OV structure. In topicalized structures with overt subject personal pronoun, that pronoun remains in initial position, yielding a surface sentence of the SOV type, such as (5) and (6). Thus, both the SOV and OV word orders constitute top­ icalized sentences. An adverb may simultaneously be fronted, producing the S (Adv) OV structure in (6). This double fronting may also be punctuated by a topicending particle (e.g., me), to be further discussed below: (5)

Wǒmén

[We

zhè

zhī

this (TI) CL

diànyǐng

film

kàn bù dǒng.

see not understand]

We saw this film but did not understand it. (Beijing 1985)

15 8 (6)

CREOLE AND DIALECT CONTINUA Wŏmén xiànzài [We now

jiùshì gēqǔ liúxíng AP songs popular

gēqu me tīng le. songs TE hear ASP]

'All we hear now is popular songs' (Beijing TV 85: MSS.44: 20) A topic can also be made prominent if it is fronted within an embedded sentence, as in (7) where the topic z h ège d i à n s h s ì j i é m ù ' television program' is fronted within the sentence embedded into the verb hope: (7) Ye [AP

xīwàng [wǒmén zhège diànshsì j i émù a, nénggòu a= hope [we this TV program TE, can TE]

=bù dān nénggòu miàn xiàng zhōngxuésheng,= [not only n face to middle school students,] =érqiě [but also

nénggòue miàn xiàng can face to

shìjìe the world]

'(I) also hope that we can direct this program not only to middle school stu­ dents but to the world as well' (Beijing, MSS, Off A:7, 1985)

2.1.1 Multiple topics The issue of topicalization is, thus, somewhat complicated by the fact that sev­ eral topics may co-occur in the Chinese sentence or discourse unit. In (6), it is possible to consider both the adverb xiànzài, and the duplicated NP gēqǔ 1 iúxíng geqǔ as co-topics. A topic which is not placed in initial position may still be fronted within a topic phrase and can be identified in other ways, for example, by the presence of a preceding topic-introducer (TI), such as the object marker ba in (9). Sentences (8) and (9) are paraphrases, both produced within a few minutes by a seventeen-year-old boy. Whereas in (8) the wordorder is SVO and the subject is also the topic, (9) has a SOV word order with two topics. The first, a verbal subject in initial position, and the second topic, a preverbal object introduced by the transitive marker ba.5 This object is clearly fronted because the comment onset is demarcated by the particle dōu (see below for a discussion of pre-comment particles): (8)

Nàxiē yīngxóng-de [Those (TI) hero-G

bàogào fēicháng gǎndòng rén. reports AP move people]

Those reports of heroic stories are extremely moving' (Beijing TV 85: 32)

TOPIC MECHANISMS IN CHINESE

159

(9)Tīng-le bàogào yǐhòu, bǎ wŏmén dōu fēicháng gǎndòng. [Listen-ASP reports after, TI us AP AP move] Those reports move us deeply/we are all deeply moved by those reports' (Beijing TV 85. MSS:32:3-6) In (10) a conditional clause is the subject of the topic unit and contains a dual fronting of the adverbial bú kuàizhediǎn 'not quickly' and the bene­ ficiary of the verb guīnǔ 'daughter', introduced by the verbal preposition gei. However, the direct object gènuxù 'husband' occurs after the verb xuan 'choose', which is complemented by the sequence Jià-chū-qù—literally composed of a serial sequence of three verbal units marry + enter + go, and refers to the marrying to be accomplished (as the result of the mother's selec­ tion of a husband). Clearly the daughter is the major topic to be fronted, but the adverbial topic quickly is also an important topic and also fronted: (10) Rúguŏ bú kuàizhediǎn gĕi guīnǔ xuǎn gènǔxù jiàchūqù= [if not quickly TI daughter choose a husband marry] =zhēng kĕnéng dānwù-le ta yí bèizǐ ya! [really might delay-ASP her AP all life EP!] 'If she does not find a husband for her daughter very soon, her whole life may be jeopardized'.6 (2NF, 1963). Another instance of the complex structure of Chinese utterances is repre­ sented in the several layers of topic fronting in (11). This sentence contains an embedded clause which follows the order OV, and the fronted object is modi­ fied by a relative clause that also contains a fronted object. Thus, the first level of object-fronting involves a discontinuous topic gēn zhègè. . .zhào shū xián 'with this. . .Zhao', the fronted object of lā jìnhū xiàn yīnqín 'establish contact and flatter Zhao Shu Xian', and is introduced by two elements (the preposition gēn and the deictic zhègè). This fronted object is discontin­ uous because it includes a relative clause modifying Zhao (the mother).7 The embedded relative preceding its "head" contains another level of fronting which applies to the object of the verb of the relative clause (dang 'decide'); that fronted object is introduced by a prepositional introduccr:duì nǔér hūnshì '(about) the daughter's marriage business' (topic 2). Again, the English translation can hardly represent this complex fronting-embedding:

160

CREOLE AND DIALECT CONTINUA

(11) .. .Xiǎo huōr bìng méi sǐxīn j i è j ī h u ì j i ù gen zhègè= . . .young man AP not give up take chance AP 77 this =duì nǔér hūnshì néng dāng bǎifēn zh ī jiǔshí jiāde= [TI daughter marriage can decide 100 out of 90 family-G] =Zhào Shū Xián lā jìNhū x iàn yīnqín. Zhao Shu Xian pull close, show like-please] [In spite of an earlier refusal] the young man did not stop trying to estab­ lish contact with, and flatter that famous Zhao Shu Xian (topic 1) because she had almost full authority in the family to make decisions about her daughter' s marriage (topic 2) (MM, 1984).

2.1.2 Topic position and given/new status Descriptions of Chinese topics, usually based on western views of topic versus focus, as discussed in chapter four, often state that preverbal (object) con­ stituents refer to given or definite information that came up earlier in the con­ versation, or to knowledge assumed by the speaker to be shared by the listener, whereas postverbal ones are understood as new information, or related to indefiniteness (Li & Thompson 1975; Ramsey 1987:71; Ross 1983:228). However, the conversational data presented here indicate that both types of information (old or new) may appear in any position, and that in fact any fronted element (especially an object) carries a special, emphatic value. For example, in (6) a new indefinite topic gēqǔ 1 i ú x í n g gēqǔ 'songs, popular songs' is preposed before its verb: (6)

Wŏmén xiànzài jiùshì gēqǔ liúxíng gēqǔ me t īng le. [We now AP songs popular songs TE hear ASP] 'All we hear now is popular songs' (Beijing TV 85: MSS.44: 20)

On the other hand, specific, previously mentioned topics are often marked by deictics such as zhè (5), zhèg (7), regardless of their position in relation to the verb. However, there are also frequent cataphoric deictics that present new information, such as nàge in (12) that obviously introduces new information about the woman named Huang Yunxiang. On the basis of recently collected data, it would appear that the fronted object topic (whether it occurs initially or after a subject) is not differentiated from non-fronted objects by the new/given status of the information but rather by some focusing function. The following claim has been made about Mandarin:

TOPIC MECHANISMS IN CHINESE

161

Mandarin is in the process of shifting its favored word-order from SVO to SOV; . . .that progress toward the conclusion of this word-order shift has been impeded by the rise of the use of word order to signal defíniteness; but. . .the evidence points to the relentlessness of the process of shifting to an SOV word order;.. .the ability of position relative to the verb to signal defíniteness is thereby weakened (Li & Thompson 1975:189). If the hypothesized change to SOV order is confirmed, then perhaps focusing mechanisms may contribute to such a change. A similar word-order alternation in the bǎ and bei constructions is illustrated in sections 4.1.5 and 4.1.6 below. 2.2 Left dislocation: Fronting with pronominal anaphora In addition to topic fronting, left dislocation involves the presence of a coindexed pronominal element. This phenomenon is usually identified by Chinese grammarians as wàiwèi ǔ, or "outside element. . .normally placed at the beginning of the sentence though it may appear at the end," while there is some anaphoric reference "inside" the sentence (Liu 1981:99). Although most Chinese speakers will accept the validity of left dislocated structures when presented with them, these structures seldom occur in spontaneous speech, and only rarely occur in literature, as in (12) where tā is the pronoun copy of the left dislocated subject mínbīng 'soldier' (which is preceded by a relativized clause): (12)Nàge shēnduàn miáotiáo, li ǎ nr hěn xiùqì-de nǔ= [that (TI) body slim, face AP pretty-G female] -mínbing, tā j iào Huáng Yúnxiāng [militiaman she is called Huang Yunxiang] 'That woman warrior with a slim body and pretty face is called Huang Y.' Li Ruqing: Haidao Nu Minbing 'Militiawoman on the island' (from Liu 1981:99) The extensive speech data recorded in Beijing television interviews (see chapter seven) include only one example of topicalization with pronominal anaphora. Although it may appear counter-intuitive to say that left dislocation is rare in any form of Chinese (Matthews and Yip p.c. 1996), it is a fact that this structure barely occurs in all my samples (oral and literary). This may indeed be an accident of my corpus, but I have no choice but to accept it as a fact. The scarcity of left dislocation observed in spontaneous colloquial and native Putonghua, as well as in non-native Putonghua varieties, suggests the testable

162

CREOLE AND DIALECT CONTINUA

hypothesis that such a topic strategy may be a feature restricted to formal dis­ course, and perhaps characteristic of more traditional literary texts, particularly in view of the fact that the examples provided by Liu (1981) all came from ear­ lier literary works. In order to test this hypothesis, a survey of written texts was conducted. Three volumes were examined, and in two collections of short sto­ ries published between 1919 and 1949, only one example of left dislocation was found; in the third collection, which included prize-winning stories written in 1978 (Anonymous 1980), only four instances occurred in the thirty-seven sto­ ries sampled (647 pages), two of which are shown below: (13) Dài Yù,tā kànguò [Dai Yu, she read

"qīng chūn zhīgē", nà shì yí f ù= "young spring song" TI TI one]

=duôme chouè-de zu ǐ liǎn [such ugly G mouth face] 'Dai Yu, she read 'The Song of Youth', what an ugly soul she has' Lu Xinghua: Shānghén 'Scar of the Soul' (Anon. 1980:246) (14)Zhèxiē pò zìhuà, guà tā gànshá? [These old paintings, hang them for what?] 'Why hang these old paintings?' Zhou Libo: x iāngj iāng YÍ Yè 'A Night in Xiangjiang' (Anon. 1980:134) Note that all four sentences occur in a highly emotional oral context (anger, astonishment). For example, (13) is produced with intense hatred by Dai Yu's daughter criticizing her parents during the Cultural Revolution for having read a "corrupted" piece of literature. It is also significant that in another poten­ tial position for left dislocation, it is not the personal pronoun tā which is used as proform, but the deictic ná, which is a different type of anaphoric reference, much more frequent than left dislocation. The more preferred type of topicalized sentence is definitely one which does not include personal pronouns as anaphora. Therefore, left dislocation will not be further discussed. 2.3 Fronting without anaphoric binding In contrast to left dislocation, it is not uncommon to encounter Chinese sen­ tences in which a fronted topic is not anaphorically bound to semantically-related elements, such as (1)—repeated below—and (15):

TOPIC MECHANISMS IN CHINESE

163

(l)jiùshì yī shuō ba, j iazhǎng dōu b ǎ n ǐmén dang háizi kàn. [TI — speak TE, parents AP TI you take child see] 'As far as talking is concerned, parents treat you as children' (Beijing TV 85: MSS.39:20) (15)zhème dà-de ǔ, ye gāi zàinǎlǐ bì yī b ì a, [That (TI) big-G rain,  should at there hide AP hide EP] 'You should have found shelter in such a heavy rain' (Wang Wenshi 1963:1.10: 129) In (1), the implied meaning is that parents talk to their children in an authoritative fashion; however, the verb topic shuō 'talking' is not anaphorically bound to any item in the comment, whereas the comment in (15) 'finding shelter' is related semantically but not grammatically to the topic ǔ 'rain'. 3. Topic repetition: Nominal anaphora with or without fronting A topicalized element may be repeated later in the sentence in a nominal or pronominal form other than the simple personal pronoun. This strategy, tradition­ ally referred to as synecdoche (involving part-whole relations) could be interpret­ ed as a variant of left dislocation much more frequent than the construction with a simple pronominal copy. The duplicate copy is an appositional phrase, which may be an exact or a partial copy of the topic or a paraphrase. This copy may be a reduced indefinite or quantifier, such as ŏude 'some', ige 'one', or yīqie 'everything', and dōu, dàdōu 'all' (when the topic is a plural noun). Such items function as comment introducers, referred to as adverbial particles (AP), clearly separating topics from comments, in which the comment represents a subset of the topic, rather than being strictly coreferential with it: (16)Cǎocōng zhōng bèi j iǎobù j ī n g q ǐ - d e yīxiē xiǎo dòngwù= [grass in by footsteps stirred-G some small animals] =yŏude cuàn xià tīyan, yŏude zhènyì fēi xiàng yúntiān [some ran down dam, some flutter fly toward sky] 'The little creatures in the grass that were stirred by the footsteps either ran down the dam or flew towards the sky'. Yang Mo: qing chun zhi G 'Song of Youth' (Liu 1982:99)

164

CREOLE AND DIALECT CONTINUA

(17) Wŏ-de liǎnggè gēgē, y īge shì [My-G two-CL brothers, one -CL TI

xuéshēng, student,

yīgè shì gōngrén. one-CL TI worker] 'One of my brothers is a student, the other a worker' (Beijing 1988) (18) Yuǎnchù-de shù, jìnchù-de cǎo, nà shīlùlu-de yīfú,= [Far-G tree, near-G grass, TI wet-G clothes,] =nashuāng j ǐ n b ì - d e y ǎnjīng.. .yīqie dōu xiàng zhěnggè= [TI pair closed-G veyes. . . everything AP like whole] =cǎodì yíyàng,wù méngméng-de. [grass same, wet foggy-G] The trees far away, the grass nearby, the wet clothes, the closed eyes, everything looks like the grass, wet and foggy.' Wang Yuanjian, qigen huochai

'Seven Matches'(Li 1981:99)

(19)zuótiān chīfàn-de rén dōu bìng-le [Yesterday eat-G people AP sick ASP] 'All the people who ate yesterday's meal got sick' (Beijing 1987) The anaphor of a topic can also be a full nominal. Various patterns occur in cases of repetition. When an object is not fronted, as in (20), it is still made prominent or identified as a topic by the presence of the attached referent diǎn shìr 'those things'. In (21) and (22), the fronted topic contains several nomi­ nal or pronominal paraphrases of the topic diànshì jù (the TV show), and all copies remain outside the comment, whereas a similar sentence (23) has nomi­ nal copies within the comment but still in a preverbal position: (20) Ai, t ā fǎnyìng lǎo Běijīng, nèi diǎn shìr. [Yeah, it mirrors old Beijing, this(TI) type things] Yeah, it depicts the old Beijing, things like that. (Beijing TV 85:C:5.9) (21) S ì shì tóng táng ne, zhè biǎoyǎn jiù bú shì. [Four generations same house TE, this (TI) show AP not TL "Four Generations in the Same House," this show is not (like other TV shows) (Beijing TV 85: A.l.1)

165

TOPIC MECHANISMS IN CHINESE

(22)Diàmshì j ù a , nǐ xǐang zánmen pái-de, Sì shì tóng táng a, nèi piānzi pái-de, zhēn nǎo. 'This TV show, you know, the one that we made, "Four Generations in the Same House," that film we made, is very good' (see text 1 (8-9) for literal translation) (23)

Sì shì tóng táng fǎnzhèng, wŏ chóngxTn yī jí yī= [Four generations same house TE, I again one part one] =mù kàm dào m ò l i ǎ o . [act watch till end] "Four Generations in the Same House," I have watched every part (of that show) twice. (Beijing TV 85: B.l.1-2)

On the other hand, a verbal constituent may be the main topic without being fronted. This is the case in (24) which displays several copies of the ver­ bal adjectives and the adverbials modifying those adjectives, all contained in the comment. This clearly signals that the main topic is the verbal constituent: (24)

Néixiē rénwù dōu tèbié [Those-(TI) characters AP AP =j i ù zhēnshí, [AP true

k ěxìn, tèbié realistic, AP

zhēnshí,= true,

lǎo Běijīng nèi yàngzi, suoyï tèbié = old Beijing this appearance, so AP]

=zhēnsh1í, kěxìn. [true, realistic] 'Those characters are realistic and true to life, they accurately represent the appearance of Old Beijing, so they are very true to life and realistic'. (Beijing TV 85: A.l.5-6) The previous discussion has illustrated the variety of devices used in Chinese to identify verbal or nominal topics in discourse. The first strategy examined was fronting, including topic fronting without anaphoric binding to the rest of the utterance; the second mechanism was repetition, with pronomi­ nal or nominal anaphoric elements. Finally, a third, and perhaps the most impor­ tant strategy, will be discussed below: it involves the additional use of special particles to mark topics.

166

CREOLE AND DIALECT CONTINUA

4. Topic presentation The presence of intervening elements between topics and comments has been identified in previous studies of Mandarin, but to a limited extent for two relat­ ed reasons: 1) little attention has been devoted to the syntax of any variety beside the most formal and officially sanctioned "Mandarin," and 2) most recent analyses have deferred to established normative descriptions of Chinese (also a Western tradition), thus overlooking important facts of language in dis­ course.8 For example, Chao (1968:67) mentions that ". . .[the] subject and predi­ cate [are] separated from each other by a pause, a potential pause, or one of the four pause particles: a (ia), ne, me and ba '." Li & Thompson (1981:86-87) fol­ low in Chao's footsteps, referring to the same four morphemes as "pause parti­ cles" or "topic markers," although they add that "a cursory survey of Mandarin speech suggests that they are not commonly used at all." There is no further information provided on how and where that survey was done, but recent fieldwork in Chinese dialects, including casual contemporary Beijing Putonghua, indicates that, on the contrary, such particles are quite frequent. In fact, many more particles occur even more frequently in similar positions, as widely rep­ resented in other sentences quoted in this chapter. Chinese discourse is, indeed, heavily interspersed with such particles whose major function appears to be to separate topics from their comments. The topic/comment basic structure assumed here for Chinese derives strong syntactic support from observations of particle distribution because particles seem to mark off those syntactic boundaries. Those "topic markers", all related to topic presentation, can be divided into four categories, depending on their position: 1 ) pre-topic particles (including copular morphemes), to be referred to as topic introducers (TI), 2) post-topic particles occurring before a pause (the standard "pause parti­ cles"), referred to as topic enders (TE), 3) pre-comment particles, which occur after the pause, usually adverbs, referred to as adverbial particles (AP) 4) post-comment particles, occurring at the very end of the utterance, referred to as end particles (EP). The four types of topic particles may co-occur in various combinations in the same meaningful unit, as shown in (1)—repeated below—which includes four topic particles altogether: three pre-topic (J iùshì, yī, bǎ), one posttopic (ba), and one pre-comment (dōu).9

TOPIC MECHANISMS IN CHINESE

167

(l)jiùshì yī shuō ba, jiāzhǎng dōu bǎ nǐmén dāng háizi kàn. [TI TI speak-TE, parents all AP -TI you take child see] 'As far as talking is concerned, parents treat you as children' (Beijing TV 85: MSS.39:20) The following utterance (25) is about students' opinions. It contains sever­ al topics combined with particles: the first topic is shèhuì 'society', punctuat­ ed both by the pre-topic preposition (TI) duì, and by the post-topic ba parti­ cle (TE). A second topic shēnghuó 'life' is immediately adduced with the same introducer; the second part of the sentence includes a subtopic j iāzhǎng 'parents' (part of 'society', followed by the post-topic particle ba, while the com­ ment has a fronted object women 'we/us' introduced by the verbal preposition ba: (25) Wŏmén duì shèhuì ba, duì shenghuo, dōu yŏu = AP have] [We TI society TE, TI life, = z ì j ǐ - d e kànfǎ kěshì jiāzhang ba, lǎo bǎ wŏmén = parents TE, AP TI us ] [self-G view but =dang h á i z i , [take child] 'We have our own view about society and life, but parents treat us as kids' (Beijing TV 85: MSS. 40:15-16) Each of the four categories is fully discussed below. 4.1 Pre-topic particles: Topic introducers (TI) A topic placed in a prominent position, whether or not it is fronted or repeated, is usually introduced by a particle. Those TIs consist of four distinct categories of grammatical constituents which are often combined in any given utterance. They are discussed in the following order: 1) "copular" particle shì, 2) deictics, 3) adverbials, and 4) verbal prepositions. 4.1.1 The particle shì Clefting and pseudo-clefting are strategies—often including a copular form— that give additional emphasis to the topic, regardless of other types of topic pre­ sentation. Thus, English, which is generally not regarded as a topic-prominent language, nevertheless uses the presentative phrase it is to perform this function

CREOLE AND DIALECT CONTINUA

168

in casual, as well as occasionally in formal contexts; French has c'est; and Belizean creole has a or da, which may in fact not be a copula, but has presentative value, as discussed in chapter nine. In Chinese, the moipheme shì is sometimes considered to be a copula (though only before nominals), but it usually carries the special emphasis associ­ ated with clefting when used before a given constituent anywhere in the sentence: "shì in cleft structures singles out the focused constituent" (Teng 1979:101); "the function of shì seems to be simply that of marking the focused constituent" (Hengeveld 1990:303). Yet, it is often assumed that there is another optional cop­ ular shì that occurs in equational structures. However, Hengeveld (1990:314) questions "why Mandarin should use copulas at all." In fact, it may be the case that shì always marks special prominence. This is essentially confirmed by the obser­ vation that shì is not required in a Chinese sentence where a copula would be expected in English, as illustrated in the following sentences which show that the phenomenon is not restricted to adjectival predicates, as in (4; 19; 24).10 It is also found before noun phrases as in (25) above: (4)

à-de dŏng dehuà, hen hǎo xiào-de, [see-G understand TE, AP good laugh-G] 'If (we) could understand (it), (it) would be funny (Beijing 1985)

(19) Zuόtiān chîfàn-de rén dōu bìng-le [Yesterday eat-G people AP sick ASP] 'All the people who ate yesterday's meal got sick' (Beijing 1987) For additional emphasis, a copula-less predicate may be simply repeated with additional adverbial forms as in (24), repeated below: (24)

Nèixiē rénwù dōu tèbié [Those (TI) characters AP AP

kexìn, tèbié zhēnshí,= realistic, AP true,

=jiù zhēnshí, l ǎo Běijīng nèi y àmgzi, suŏyǐ tèbié = [AP true old Beijing this appearance, so AP =zhēnshí, kěxìn. [true, realistic] Those characters are realistic and true to life, they accurately represent the appearance of Old Beijing, so they are very true to life and realistic'. (Beijing TV 85: A. 1.5-6)

169

TOPIC MECHANISMS IN CHINESE (26)

Jiù nèige qíngkuāng, jiù zhème ge shìqing [AP that (TI)situation, AP this CL thing} 'That was the situation, that's the way it was ' (Beijing Saga A:5)

Another relevant fact confirming that sh i is not a true copula is its use before regular verbs, such as bǐj iǎo 'compare' in (27). Shì clearly has a presenta­ tive function here, as well as in (28), where it appears (after an emphatic adver­ bial) at the beginning of a subjectless sentence. (28) also includes some hesita­ tion marked by the use of three types of emphatic markers, shì, the intensifi­ er tǐng and the deictic n é i g e . Shì may also occur before any other gram­ matical category, such as the indefinite zěnmē 'how' in the second part of (29): (27) Fǎnzhèng shì [Really (TI)TI

b ǐ j i ǎ o xiànzài, b j i a o guòqù, shìba. compare now, compare past, EP]

'Really, what it does is compare the present and the past, isn't it?' (Beijing 85: B.3:5) (28) Fǎnzhèng

shì

[Really (TI)77

tǐng

AP

-nèige

— dòng

zhī

xīn-de

— that (TI) — move one heart-G]

'(It) truly moves me deeply' (Beijing Saga B: 3) Even when shì occurs in an apparently copular position, in the first part of (29), it seems likely that it has a presentative function, much beyond a simple equative statement. In fact, sh i introduces a metaphor of the topicalized con­ stituent in (29) lǎoshl shì 'as far as teachers are concerned': (29)Lǎoshī shì xīnqín-de yuánd īng, wǒmén t î n g t i n g tāmen = [Teachers TI industrious-G gardeners. We listen-listen them =shi z ěnmē" xiǎng-de ba [TI how think-G EP] 'What teachers are, they are industrious gardeners. Let's hear what (it is) they think' (Beijing TV 1985: MSS.l: 1-2) In other cases, as illustrated in (30), the postposed copula (yǒudē shì) highlights a copy of the topic. The sentence really means 'as far as party members controlling their wives, there are many who do it':

170

CREOLE AND DIALECT CONTINUA

(30) Búguò lā zìj ǐ lǎopό hòutuǐ-de dǎngyuán, yě= [but pull own wife leg-G party member, AP] =yǒudē shì ne. [plenty TI EP] 'But many party members control their wives' (Wang Wenshi 1963: 124) The same morpheme is used for tag questions as shì bù shì , shì ba, shì ne 'isn't it?' (2; 27; 30), or for short assertions (shì de 'indeed, right'). The shì particle can be combined with other particles, as in (31), which displays both a fronted object with bǎ, and a postposed constituent topicalized by shì: (31) Bú yào ba j iàoyù kànchéng j i ù shì [Not want TI education consider AP 77

xuéxiào-de school-G

shir, matter]

'We don't want to consider education to be related to school alone'. (Beijing TV 85: MSS) The emphatic interpretation of shì is supported by historical evidence: shì was first used in Classical Chinese (eleventh century b.c.—third century b.c.) between nominais as an anaphoric demonstrative pronoun (Li & Thompson 1977:420; Pulleyblank 1995:85-89). It can be interpreted as a copy of a front­ ed object, as illustrated in an example taken from the Analects of Confucius (fifth century b.c.): Fu yu guo, shi ren zhi suo yu ye. 'Riches and honor, these men desire' (in Ramsey 1987:72). Li & Thompson (1977:424) assume that "the topic-comment constructions without a copula became a subject predicate construction with the anaphoric demonstrative pronoun shi being reanalyzed as a copula." However, it is possible that the copular interpretation in modern Chinese grammars of the old Chinese demonstrative may be attributable to a Westernoriented perspective which assumes that a copula is an essential sentential con­ stituent (a similar problem was noted in the interpretation of the copular/deictic da in Belizean Creole; see chapter nine). Spontaneous data confirm the hypothesis that the particle sh i occurs in topicalized structures, and often in sentences involving a contrast, such as (17) or (27), whereas the ancient verbless pattern for nominal predicates remains widely used in unmarked equative sentences, such as (26). It is also worth noting that shì always occurs in neg­ ative sentences, and this may be because the negative tends to have a marked contrastive value.

TOPIC MECHANISMS IN CHINESE

171

4.1.2 Deictics Some of the TIs represented in the data examined below are place deictics (zhè, zhege, zhème, zhèyang, nà, nàge, nèi, nèixiē), which often occur several times in a single topic-utterance, and more so in some speakers than in others. Such TI deictics are not necessarily semantically related to the items that follow them. For example, in (32) the initial deictic zhè introduces the time adverbial shǒuxiān 'first', and could be roughly translated as 'this is what I will be talking about first': (32)zhè shouxiān ne, cóng nu xuéshēng, wo zuòwéi yīgè= [77 first TE, from (TI) female student, I as (TI) one-CL] =nǔ tóngxué ha, wo juéde cóng nǔ tóngxué lai jiǎng ba= [female student TE, I feel from TI female student come say TE] =shǒuxiān yīngāi jiùsh ì t è b i è [first (TI) must AP AP natural

dàdafāngfang bié= not]

=n iǔn i un iēni e [affected] Tn the first place, female students, as a woman myself, I feel that female stu­ dents should be especially natural and unaffected' (Beijing TV 85: MSS 36) In (33) wǒ nà wèi, literally T that man (my husband/that husband of mine)' shows the deictic ná placed after the first person pronoun (here functioning as possessive adjective modifying wèi) and the noun topic: (33) Wǒ nà wèi, dàoshì chángcháng j iàoyù wǒ ne! [I that (TI) man, in the end (AP) often (AP) educate me EP] 'That husband of mine, he educates me indeed' (Wang Wenshi 1963:124; see also Text 2: 6a) Other examples of deictics introducing or accompanying topics are wide­ ly interspersed throughout the samples presented in this chapter. These deictics are clearly reminiscent of the broad use in creoles of sentence-initial da, dat, dis, and other demonstratives in various lects of the continuum. 4.1.3 Advertíais Some other pre-topic elements, e.g., j i ùshì (1), xiànzài

(see both in 34),11

172

CREOLE AND DIALECT CONTINUA

fan z hēn g (3; 27; 28), are more easily interpreted in English as adverbials because they fulfill pragmatic roles similar to those held by discourse markers such as so, well, anyway, like, you know. They may be characterized as mitigators, hedges, fillers, or connectors between utterances, and they constitute pragmatic particles which have little semantic value in spite of their dictionary gloss. That those semantic values are lost in natural discourse clearly appears in the following. In (34), xiànzài 'now' provides no time reference, and in this case is equivalent to the semantically empty use of now in English casual discourse; j i ùshì 'still, so' (34) has a strictly emphatic value, with usually no actual reference to sequencing. Similar observations can be made for other adverbials, such as shǒuxiān 'first' (32) which is not really referring to the first of a sequence.12 (34) Xiànzài, g ēn j iāzhǎng, dōu shuō bú shàng shenme-le,= [Now (TI) with (TI) parents, AP talk not result something-ASP] =jiùsh ì y ī shuō ba, j iāzhǎng, dōu [so (TI) AP speak TE, parents, AP

ba nǐmén dāngháizi kàn, TI you take child see]

'Now there is almost nothing you can talk about with parents, that is, when you talk, your parents treat you as children' (MSS.1985, 39:20). Chapter four documents similar pragmatic changes in Belizean varieties. 4.1.4 Verbal prepositions Some of the most intriguing TIs are those traditionally referred to in Chinese grammars as "prepositions," mostly because they superficially correspond to Western accounts of prepositions in Indo-European languages. This brief overview is restricted to the items most frequently occurring in my corpus, including duì 'to' (directional); gěi 'to' (benefactive); zài 'at' (location); cong 'from' (provenance); gēn, tóng 'with'; tōngguo, zuòwéi 'concern­ ing'; bǎ (an object marker); and bèi (an agent marker). The functions of the members of this category are more than strictly preposi­ tional; indeed, their grammatical status is at best ambiguous, and much closer to the notion of "serial verb" represented in African languages and many creoles. Historically, they were verbal elements, and they still preserve some of their ver­ bal functions as shown for duì in (35; 36). In those sentences duî does not com­ bine with a noun. It is clearly the second (in 35), or third (36) element of a verbal string modifying the topic. In those serial strings, it is associated with the postver­ bal aspectual (completive) marker le, a verb property. This verbal function is "frozen" in the common colloquial usage of duì 1 e 'right! that's correct'—literally [target reached].

TOPIC MECHANISMS IN CHINESE

173

(35) Nǐ zhègè qiao niángr j ì n g tóng wo zuoqï= [you TI pretty woman how with (TI) me use] =duì [target

lái-le come-ASP]

'But how can such a pretty woman as you dare compete with me!' (Zhang: 2NF; 8) (36) Yō, zhè guǐ tiān [oh, TI devil weather

zhēn tóng iao niáng zuòqǐ duì-le AP TI old mother contradict reach ASP]

'Hell, this lousy weather goes against mother's (predictions)' (Zhang: 2NF; 48) Other examples represent the highlighting function of verbal prepositions such as tōngguò 'regarding' (37) and cóng or zuòwéi 'from the perspective' (32)— repeated from above—-which in effect produce structures akin to pseudo-clefts: (37) Tōôngguò zhègè j i émù, yě yǐn j ì n zánmēn shèhuì = [77 this (TI) program, AP bring enter our society] =gège fāngmiàn guānxīn zánmēn qīngshàonián,= [each aspect care our young ] =háizǐ-de chéngzhang, [child-G growth] ' What this program does is examine all aspects of our society to foster the development of our children' (Beijing TV 85:MSS. 6.4-5) (32)

Zhè shǒuxiān ne, cóng nu xuéshēng, v/o zuòwéi yīgè nǔ tóngxué ha, wo juéde cóng nǔ tóngxué lái j i ǎ n g ba, shǒuxiān yīngāi j i ù s h ì t è b i è dàdafāngfang bié niǔniuniēnie 'In the first place, what female students are concerned about, as a woman myself, I feel that what female students are concerned about is that they should be especially natural and unaffected'.

Although many verbal prepositions can occur to introduce a subject topic, or a non-fronted object, some particles such as gěi, dui (sometimes), and bǎ. (always) occur before a fronted recipient or object; however, when the object immediately follows the verb, that verb is unmarked for direction or intention, and the meaning is usually that the object is not aware of, or is not affected by, the action. Compare, for example, [duì tā. shuō], [gěi tā shuō], or [bǎ

CREOLE AND DIALECT CONTINUA

174

tā shuō] = [to her speak] 'talk to her' (direction), as opposed to [shuō t ā ] = [speak her] 'talk of her'. This contrast supports the strong topic-marking role of verbal prepositions. For example, in (38), Yang, the commune director, is the person Zhang (the speaker) had to talk to in order to get permission to sleep in her friend's room (see chapter seven for a full discussion of the text). In (39), gei functions as a strong imperative, whereas a more neutral command would be dō huíqù ba "! 'all of you go back': (38) Wǒ g ěi Yáng k ēzhǎng shuō-le, zánliǎ zhùzài yìqǐ [I give/with Yang director speak-ASP, we two stay together] 'I talked to Director Yang, and we can stay together' (Zhang: 2NF; 36) ( 3 9 ) D Ō U gěi

wo huí

qù

[all give me return go] 'I order all of you to go back' (Zhang: 2NF; 49) A serial sequence (with verbal preposition) often results in causative-like structures. For example, the only way to say "show me" in Chinese is literally 'give me see' gěi wǒ kán, a close equivalent to English "let me see".13 In all cases mentioned above, it is clear that the so-called prepositions, or serial verbs, under examination are basic syntactic constituents and that they are not case markers. Their function is clearly pragmatic; they mark topics, which justifies my classification of such items as TIs. Two of the special elements (bǎ and bèi) have been particularly singled out in previous accounts of Chinese and associated with distinct constructions, respectively, the object construction (bǎ ), and the passive construction (bèi). I will argue that they really fulfill the same function of topic markers as duì, gěi, tóng, and others discussed above. 4.1.5  Like other verbal prepositions, bǎ originated as a verb 'to hold' in Middle Chinese, when it alternated with jiang 14 . It then appeared as a patient-marker, and by the turn of the century it became grammaticalized into a preposition-like marker of high transitivity (Sun 1995:160). The bǎ construction has been var­ iously called a disposal construction (Wang 1954; Chao 1968), an executive construction (Hashimoto 1971), an accusative or transitive construction (Teng 1977; Tsao 1979; Li & Thompson 1981: Sun 1995).  now exclusively introduces the object of a verb, though not necessar­ ily that of the main verb, and the sequence bǎ + object is preverbal, often pro-

TOPIC MECHANISMS IN CHINESE

175

ducing a causative-type construction (Tsao 1987). It is, thus, by definition a prominently fronted position (Tsao calls it a "secondary topic"), but a fronted object is not necessarily introduced by bǎ. It has been claimed that a fronted position defines a specific element or a situation previously referred to, and that a bare nominal after b, cannot have a non-specific or definite reading (Chao 1968; Li & Thompson 1981), whereas a postverbal object position would indi­ cate an unspecific or habitual action. However, this distinction is not apparent in my corpus, as represented in the following sentences. According to previous studies, perhaps best summarized in Sun (1995) and Zou (1993), the occurrence of bǎ must be related to distributional properties and constraints such as specificity effects, and dynamic change of state verbs. Temporal boundedness is said to be marked by the presence of the perfective aspect le (considered to be required whenever the bǎ construction occurs), or of a directional particle (Zou 1993:718). However, the sentences excerpted from literary discourse shown below, and others extracted from spontaneous discourse (in chapters seven and eight) do not support the claim that ba is sys­ tematically linked to specific, definite or individuated semantic properties, or to the emporai boundedness of an event, or even co-occurrence with the aspectual mark­ er le. Although the above-mentioned specifications may certainly hold in formal varieties of chinese, or isolated sentences, the samples examined here suggest that the alternation between strings with preverbal bǎ+object and strings with bare postverbal objects is much more loosely used in contemporary colloquial situa­ tions. In this case, they appear to be merely related to topic-marking strategies, as also previously discussed in the case of simply fronted elements in section 2.1.2. yīfú Both possible sequences are represented in (40): V+O in kǎogän-le '(she) dried (her) clothes'; and bǎ+O+V in bǎ. . .yī fú. . .dié nǎo' (she) folded (the) clothes (that L.Y. had given her)'. (40) Shū Lán kǎogān-le yīfú, huàn zài shēn shàng,= [Shu Lan dry-ASP clothes, change body on,] =bǎ Là Yuè gěi tāde y i f ú xiǎoxīn dié hǎo [TI La Yue give her-Gclothes carefully fold. . . .] 'Shu Lan changed back into her dried clothes, and carefully folded the clothes that La Yue had given her ( 2NF; Narrator: 67). In the V+O sequence (kǎogän-le yīfú) the object clothes is not unspecified in the context of the story. The fact that Shu Lan's clothes got drenched in the rain, and had to be removed to dry, was mentioned in the pre­ ceding utterance. The second preverbal object, her friend La Yue's clothes, was

CREOLE AND DIALECT CONTINUA

176

also mentioned before (La Yue had previously lent dry clothes to Shu Lan). So both objects are specific and "old," but the second clothes is specially empha­ sized (in relation to her friend's hospitality), which justifies the use of the ǎ+O+V construction in the second part of the sentence. Furthermore, it is interesting to note that the aspect marker le is associated only to the first verb (with postverbal nominal). This may suggest that the completive aspect need only be marked once in a topic unit (a fact also noted in Belizean Creole). There is a similar object placement variation in (41-42), but in both cases, the main topic is qí 'the red flag', the trophy to be won by worker-peasants, a very specific topic presented from the first line of the story. In (41) the ǎ+object qí construction reflects the speaker's surprise when she finds the flag in an unexpected location, thus, representing a typical topicalization. On the other hand, the V+O construction in (42) is appropriate, since the focus is already on defeat and win: (41) Zhāng j i ě , nǐ zěnme bǎ qí fàng zài zhèlǐ [Zhang sister, you why 77 flag put in(TI) here

ya? EP?]

'Sister Zhang, that flag, why did you put it here? (2NF; Wu: 79) (42) Wǒmén-de kǒuhuò shì:mǎtà nánèr shè. . .shōuhuí = [our-G motto TI: defeat Naner village. . .win back] =clà hóng qí [big red flag] 'Our motto is: defeat Naner village, win back the great red flag'(2NF:75) The emphasis related to the ǎ-construction is also clear in the following sentence. The speaker expresses her impatience at not being able to persuade her friend; she can talk to her but to no avail: (43)ǒ zhīdaò yīshì yě ba nǐ shuō bù zhuǒn. [I know AP AP TI you talk not change] 'I know that I can't persuade you at all' (2NF; Zhang:61) The examples given in this section are drawn from a short story to be dis­ cussed in chapter seven. However, the same topic-marking interpretation of bǎ i s confirmed in casual sentences, as represented in bǎ nǐmén (1); ba women (9) and ba j iàoyù (31), in which at least two of the fronted objects refer to gener­ al ideas about education, or habitual parental attitudes, thus contradicting the tra-

TOPIC MECHANISMS IN CHINESE

177

ditional notion that object fronting denotes new, nonhabitual references. This fits well with Tsao's characterization that "the function of bǎ is to mark the following NP as a special topic, and the meaning of the construction is to make clear the tran­ sitivity relation between the primary topic and the bǎ topic, and to bring into focus the result, as expressed by the verb and its complement" (Tsao 1987:1).

4.1.6 Béi Béi, like bà, is used only to provide a special focus on a noun before the verb, as is indicated in the meaning of the related lexical item bèizi 'blanket, quilt', a reminder of an earlier verb meaning 'to cover', which suggests additional effect on the item following. In contrast to ba (introducing an object), bèi introduces an agent, which is why it has been traditionally associated with the passive construction. However, the use of bèi does not require a specific aux­ iliary.15 Bèi can, therefore, be merely defined as an agent-marking device, when this agent is not placed in sentence-initial position. In this sense, it is truly equivalent to a cleft construction, as was observed with other particles: (44)Tā dàodǐ bèi wo jiàoyù-guò-lái la. [he after all TI I educate-ASP-come EP] 'After all, /have educated him (her husband)' (Zhang: 2NF;14) [It is I who educated him/thanks to me, he got educated] (45)Jǔzhǐ wénjíng-de Wú Shū Làn béé' Zhāng Là Yuè'= [behavior gentle-G Wu Shu Lan TI 'Zhang La Yue'] =zhègè tā céng shuōqǐguo duōshǎo c ì-de míngzǐ,= [this (TI) she once mention many times-G name, =bèi yǎnqián kàndào-de zhègè zhēnshí-de this (TI) real-G [TI eye front see-G

nǔrén,= woman,

=yǐjí tā nà chìluǒluo-de duì rén-de tàidù= [AP her that (TI) warm-G TI people-G attitude] =suǒ gǎnrǎn [is moved] 'The gentle Wu Shu Lan is moved by the frequently mentioned name of 'Zhang La Yue', by the real woman in person and her warm attitude' [It is the name of Zhang, it is the real woman, it is her warmth that moved Wu] (Narrator: 2NF; 7)

178

CREOLE AND DIALECT CONTINUA

In all of the above cases, bè i can be omitted from the context. The use of the particle indicates heightened emotions, anger, or excitement, as clearly rep­ resented in the previous sentences, thus, creating a pragmatic phrase not unlike clefting in English. However, it is difficult to render in English the multilayered topicalization involving Zhang La Yue, which is represented in the bèi -fronted phrases in (45). In conclusion, it is clear that ba and bèi like other verbal prepositions are not stable case markers of Chinese morphosyntax. They are truly pragmatic particles marking topic prominence, with dramatic effects to denote a emotional or sometimes formal context (as will be discussed further in chapters seven and eight). 4.2 Post topic particles: Topic enders (TE) Topic enders have the function of delimiting the topic and are usually followed by a pause in oral discourse, thus clearly separating topics from comments. They include primarily n e, a, as well as ba, na, ya, me, although n e and a occur most frequently. Several related topics in an utterance may be each fol­ lowed by a TE, e.g., ne in (46), which punctuates first an adverbial topic ( J i n t i ā n ne 'today'), then a full sentential topic (wǒmén 1áidào nǐ jiā ne 'we come to your home'). This outlines a strategy meant to set up the back­ ground of the main purpose of the interaction ('to interview parents about their children's education'): (46)

Jīntiān ne, wǒmén láidào n ǐ ' j i ā ne, xiǎng tīngting= [Today TE, we come your home TE, like listen =jiāzhǎng tóngzhine duì wǒmén diànshì t á i jǔbàn= [parent comrade TE TI our TV station run ] =zhōng xuésh ēng jiémù-de j iànyì hé yīxiē kànfǎ. [middle school student program-G suggestions and some opinions] 'Today, we have come to your home to hear parents' suggestions and opinions about our TV station's program on junior school students'. (Beijing TV 85: MSS)

In (6) the particle me accentuates a fronted topic further introduced by two adverbs; thus the topic is highlighted in four ways: fronting and three presentative elements:

TOPIC MECHANISMS IN CHINESE (6)

179

Wǒmén xiànzài j i ù s h ì gēqǔ liúxíng gequ me t ī n g - l e . [We now AP songs popular songsTE hear ASP] 'All we hear now is popular songs' (Beijing TV 85: MSS.44: 20)

See also shuō ba in (1); sì shì tóng táng ne in (21); shèhuì ba in (25); shǒuxiān ne, nǔ tóngxué ha and nǔ tóngxué lái jiǎng ba in (31); zhèģe d i à n s h s ì jiému a in (7), inter alia.

4.3 Pre-comment particles: Adverbial particles (AP) Adverbial morphemes functioning as comment introducers have rarely ever been identified as topic-related in the extant literature on Chinese syntax, except in a few cases. Jiù and cài have been singled out as focusing adverbs equivalent to 'even' and 'only' in English (Biq 1988:78). However, those adver­ bials (and others) may be fulfilling more complex roles in the organization of syntactic units. They constitute an essential element of discourse strategies found to be operational across the various dialects of Chinese, including Mandarin. More specifically, the juncture between topic and comment is almost always filled with some constituent which has the overt form of an adverb and may, in addition, have a semantic value related to the original adverb. However, more often than not, this adverbial is semantically empty and seems to hold a position which has the strictly syntactic function of delimiting topic and com­ ment. This is clearly shown by the high frequency of occurrence of those items, especially in vernacular forms, and is also supported by native speakers' sub­ jective reactions. For example, one of the most common adverbials preceding adjectives is hen (typically translated as 'very'), as in (5), in which the predi­ cate 'be funny' is not especially emphasized: (4)

àn-de [see-G

dǒng dehuà, hên n ǎo xiào-de, understand TE, very (AP) good laugh-G]

'If (we) could understand (it), (it) would be funny (Beijing 1985) Yet, all my respondents agree that a predicative adjective not preceded by any­ thing in the comment would sound, at best, awkward and is most likely ungrammatical. It seems that hěn no longer has (assuming it ever did) the function of an intensifier:

180

CREOLE AND DIALECT CONTINUA

*Zhège hǎo 'This is good' (awkward, unacceptable, or emphatic)16 Zhège h ěn hao 'This is good' (unmarked, neutral) Other adverbials (see table 6.1) must be used instead when intensification is required: Zhège fēi chati g hǎo. 'This is very good'(marked). This is, also illustrated in natural discourse, as shown in sentences presented above, such as (9; 25) and others): (9)

Tīng-le bàogào yǐhòu, bǎ wǒmén dōu fēicháng 'These reports, we are deeply moved by them'

gǎndòng,

(25)

Neixii rénwù dōu tèbié kěxîn, tèbié zhēnshí 'Those characters are very realistic and very true to life'

In the Putonghua conversations examined in chapter seven, predicate adjectives are always accompanied by some kind of modifier. In addition to the adjective intensifier category, other manner adverbials occur in the same posi­ tion, especially when the comment includes non-adjectival verbal units. Those adverbials are constituents of the comment for two reasons: they occur after the post-topic pause (if there is such a pause), and they sometimes provide seman­ tic value modifying the verb. For example, yě and j i ù/j iùshì are usually glossed according to standard dictionary entries as 'also; still; only', but this meaning is not accurately represented in most cases, since some of those parti­ cles also function as topic introducers, as in (3; 32) below, and also (1, 2, 6, 15, 37), and others: (3)

Fǎnzhèng, nǎ j í yě bù quē, [really (TI) any part still (AP) not miss] (I wouldn't) miss any part (of it)' (Beijing TV 85: B.1.3)

(32) Shǒuxiān yīngāi jiùshi tèbiè dàdafāngfang bié= [first (TI) must still (AP) very (AP) natural not] =n iǔni uniē i e [affected]

TOPIC MECHANISMS IN CHINESE

181

Another frequent pre-comment particle is d ōu 'all' which most commonly occurs after a plural noun phrase, as in (9); see also (1, 25, etc.): (9) T ī n g - l e bàogào yǐhou, ba women döu fëichang gândong. [Listen-ASP reports after, TI us all (AP) very (AP) move] 'Those reports move us deeply/we are all deeply moved by those reports' (Beijing TV 85. MSS:32:3-6) Those particles are reduced to the syntactic/pragmatic function of indicat­ ing a sequence of events in the comment or marking a comment juncture. Even when those adverbials preserve a shade of their semantic value, as for yïj Ing (in 47), the pragmatic function is still crucial: (47)Qunidn l d i Meiguó-de ren, yïjlng huí guó [last year come USA-G people, already (AP) return-ASP

le ASP]

'The people who went to the USA last year have already returned' (Beijing 1988) In addition to their potential adverbial semantic value, those items have taken up the important syntactic function of processing discourse information, and that is accomplished by separating the topic from its comment. This is con­ firmed by the fact that an AP is not necessary when another unit marks the junc­ ture, for example if the comment includes the negative particle b u (7) or bi e (32), or a modal, such as nénggóu (7). Some of the most common AP mor­ phemes identified in the various Putonghua samples examined in chapter seven are shown in table 6.1 : Table 6.1: Comment adverbial particles (AP) Before Adjective: hen, fëichang, ting, tèbié, lâo, zhën, gëng, zùi, bïj iào, bu, tài Before Verb:

jiù, jiùshï, hai, háishi, yë, cái, yïjîng, zhūn, döu, quän

4.4 Post-comment particles: End particles (EP) The end particles are generally the same as those occurring in post-topic posi-

182

CREOLE AND DIALECT CONTINUA

tion (TE), but they include primarily ba and a, and less commonly h a, ne, na, ya, la, shiba, me, ma.15 They function as pragmatic markers because they only occasionally punctuate sentence structure and typically finish the topic unit, usually to convey a variety of emotional messages (Han 1995). Some of the most common particles and their possible meanings are shown below, although as illustrated in the examples provided, it is often difficult to find appropriate translations: ba a ne 1a ya (2)

request; polite suggestion; imperative in (29); tag question, e.g., shiba in (2; 27). excitement, astonishment as in (15) disagreement; contradictory statement in (30; 33) satisfaction, triumph (44) distress, surprise (10; 41) Shî shëng-de g u ö n x ï , ye xTwctng nïmén z à i j i é m ù zhöng g e i y ï y ï d i n g - d e zhôngshî, shiba,

'Concerning the student-teacher rapport, I hope that you will give it a certain emphasis your program, won't you?' (15)

Zhème dà-de yu, ye g a i

zài

nâlï

bi

yî

bî

cu

'Such a heavy rain, you should have found some shelter, really! (29)

Women t ï n g t i n g tarnen shî 'Lets hear what they think! '

(33)

Wo nà v è i , d à o s h i

zënmë x i a n g - d e

ba!

chángchdng j i à o y ù wo ne!

'On the contrary, my husband often educates me indeed'.

5. Summary of topic strategies Of the three major topic-marking strategies discussed and illustrated in this chapter, the presentative category is the most proline, and this impression will be tested empirically in the following chapters. There is an extensive diversity of particles in Chinese, especially topic introducers (Tis), including deictics, adverbials, and an abundance of multifunctional verbal prepositions which ful­ fill important topic-marking functions. In addition, there is a large group of markers introducing comments (APs), which primarily include adverbs or quantifiers, some preserving their original

TOPIC MECHANISMS IN CHINESE

183

semantic value but many others serving purely syntactic, or pragmatic func­ tions, not unlike the Belizean elements discussed in chapter four. A more restricted group of morphemes serves to signal the end of a topic (TEs). Finally, there is a group of morphemes which punctuate the end of an utterance (EPs), often adding emotional content to the context. The ultimate goal of these mul­ tiple Chinese particles is to facilitate information processing by clearly break­ ing down the utterance into its units (topic versus comment), and mostly to identify prominent elements—a function also facilitated by other devices such as fronting and repetition.

Endnotes 1. Li & Thompson (1981:15) further state that ". . .topic-prominent sentence structure is a sig­ nificant typological feature of Mandarin in terms of which it can be compared to other lan­ guages, such as English." 2. This absence of anaphoric binding was also found in Belizean Creole (see chapter four). 3. The various particles included in this sentence are fully discussed below. In all following examples, some of the most common Chinese morphemes are identified as follows: ASP= aspect marker (e.g., 1 e, guo, marking anteriority or completed action; or zhe, marking pro­ gressive); G=possessive marker, also functioning as relative marker, i.e. de following a modi­ fying constituent, as in (2); CL=classifier preceding a noun, as in (3). In addition, all topicmarking particles which independently have other grammatical or semantic functions (e.g., adverbials) are abbreviated as follows: TI=topic introducer; TE=topic ender; AP=adverbial par­ ticle; EP=end particle. The use and functions of those particles are fully discussed below. Those elements are translated whenever possible; if they keep a semantic value they are enclosed in parentheses. When no appropriate translation exists, it is marked —, and = indicates that a sen­ tence is continued on the next line. 4. Full methodological details of the data collection are presented in chapter seven. Examples in this chapter are mostly drawn from a series of interviews made in Beijing in 1985 and from texts published in 1984. 5. See section 4.1.5 in this chapter for a full discussion of verbal prepositions, including b a 6. An English translation cannot truly reflect the prominence of both the adverbial and the beneficiary unless it is broken down into several sentences, as attempted here. 7. Literally, the relative clause shows the sequence '. . .with this [decides daughter's marriagewho] Zhao'. Relative clauses in Chinese are preposed and marked by the postclausal relativizer de. 8. See the preceding chapters for a discussion of similar normative attitudes in relation to Belizean Creole. 9. The two particles ba and ba are clearly distinguished in Chinese, both by different tones and different ideograms. 10. Verbal adjectives—adjectives functioning as full-fledged verbal units—are common in many languages, including créoles, as discussed above. 11. Note that part b) of sentence (34) also appears as (1) in parts of this chapter. 12. Similar morphemes are used as APs (pre-comment particles), as discussed below.

184

CREOLE AND DIALECT CONTINUA

13. Thus, it would appear that causatives and inchoatives in English are possible serial strings, as in "help me do it; make him do it; get him to do it." 14. J i an g has now disappeared from Mandarin Chinese, but it is still used in many conserv­ ative varieties, including rural varieties, and Suzhou Wu, as will be discussed in chapters seven and eight. 15. However, the morpheme sua, as in suo gan ran (45), can be viewed as a possible passive auxiliary, although it rarely occurs. 16. Matthews and Yip point out z hè hå  is fine in a comparative or contrastive context, this indeed confirms the unmarked value of h e n, the old intensifier, versus the marked emphatic value of zero to mark contrast. 17. Another particle ma, occurring at the end of the sentence, functions as a question marker. It is not included among the post-comment particle category.

Chapter 7 Literary and Colloquial Putonghua Abstract Two major varieties of Standard Beijing Chinese (Putonghua) are investigated in writ­ ten and spoken contexts that mirror social aspects of language change in the recent decades of the history of the People's Republic of China. Topic marking is used as the diagnostic feature according to the procedures outlined in chapter six. 1. The Putonghua corpus: Methodology I will now outline the contexts in which the topic marking examples presented in chapter six were collected, the rationale for selecting such data, and the methodological procedures implemented to obtain them. In consideration of the importance of methodological issues in the assessment of theoretical analysis, I have included as broad a data base as possible to document natural speech. In order to consider potential differences between various dialects, it is essential to first establish a point of reference, determining relevant aspects of the standard variety of Chinese which provides nonstandard speakers with a model for sec­ ond dialect acquisition. The linguistic feature under investigation, namely, the morpho-syntactic mechanisms involved in topic structures—as represented in the three categories broadly outlined in chapter six—will be examined in detail in different colloquial varieties of Chinese generally considered to be standard, since they were either published or were produced on the national Chinese tele­ vision network. The Standard (Mandarin) Chinese, now called Putonghua, "Common Language" (henceforth abbreviated as "PH") is based on a variety of Beijing Chinese. These PH varieties will provide the linguistic background against which other varieties can be compared. The other dialects to be investi­ gated in the following chapter are non-native Putonghua as spoken in Hubei, and Suzhou Wu, all based on recently recorded speech samples (see chapter eight).

186

CREOLE AND DIALECT CONTINUA

LI Problems of analysis: Topic units In establishing quantitative measurements of topic-marking strategies and par­ ticles, it became increasingly obvious that the sentence does not constitute a pragmatically valid point of reference, as already demonstrated in relation to creóle discourse (chapter four). The processing of information in Chinese syn­ tax is best understood if analyzed in terms of topic/comment combinations; that is, the topic determines one syntactic unit which includes one or more com­ ments relating to that topic, but not necessarily containing anaphoric reference to that topic. This unit is defined as topic chain by Tsao (1979) and Shi (1984), as discussed in chapter six, and I will refer to it as "topic unit" or "topic utter­ ance" (henceforth TU). For example, there are seven TUs in text 1 (a sample from "Two New Friends," shown in section 2.2 below, and analyzed in 2.2.1), with twenty-four comments attached to those seven topics. In one specific TU, the topic nan ren 'men' has five comments (text 1: lines 3a-3e): 1) men should not impede their wives' careers (3a); 2) men used to say that women interfere with their own careers (3b); 3) now men face a reversal of the situation (3c); 4) some men interfere with their wives' careers (3d); and 5) does Wu's husband belong to the latter type of man? (3e) Yet discrepancies in the actual amount of speech may be overlooked, since the topic-unit goes beyond the sentential unit. Thus one single TU may be made up of a short response, and yet include several presentative particles, as shown below:1 (1) ni ye h en qiào a [youAP AP pretty EP] 'you are very pretty' [1 comment; 3 particles] (Wu; 2NF-8) (2) ha i me i you [AP not have] 'not yet' [topic implied; 1 comment; 1 particle] (Wu; 2NF-12) (In response to: Have you applied to party membership?) On the contrary, one TU may include several comments, often constructed on the basis of parallel structures, repeated twice or more, as in (3-4), and a longer TU does not necessarily include more topic particles:

LITERARY AND COLLOQUIAL PUTONGHUA

187

(3) tóuhuí sheng, èrhui shóu, jintiem j i à n - l e miàn, [first stranger, second friends, today see-ASP face, j i ùshï qîn j ie mei l a . AP become true sisters EP] 'Although we have just met, we have already become friends' [topic implied; 4 comments; 2 particles] (Zhang; 2NF:6) (4) zhàngfu vàng-zhe tä, yáo tóu, zhôu méi, tàn qî [husband look-ASP her, shake head, contract brow, exhale sigh] 'Her husband is looking intently at her, shaking his head, with a frown and sigh' [topic 'husband'; 4 comments; 0 particle] (Narrator: 2NF; 22) Decisions as to TU range are often difficult and even arbitrary. For example, how does one treat the author's omniscient interference in a dialogue? Such problems, and the solutions adopted here are illustrated in the first lines of text 1. Regardless of the size difference between topic utterances, it appears that their number correlates with assertiveness: the more TUs, the more confident a speaker is (to be discussed below in section 2.4). 7. 2 Chinese varieties examined Two major varieties of Putonghua illustrated in four distinct samples are ana­ lyzed in this chapter. They can be considered to be representative of the present or recent state of spoken PH since they cover a wide repertoire of styles rang­ ing from written styles reporting vernacular speech to actually produced infor­ mal conversations. In addition, they provide a longitudinal dimension docu­ menting Chinese in the 1960s and in the 1980s. The first corpus represents a specific type of written PH and, thus, a relatively literary style spanning the last thirty years, whereas the second corpus investigates the spoken vernacular dimension of contemporary PH, as spoken from 1985 to 1988. 1.2.1 Literary Putonghua This corpus is derived primarily from two sets of short stories selected to rep­ resent the recent sociocultural history of the People's Republic of China in the last thirty years (1963 and 1984), from the period immediately preceding the onset of the Cultural Revolution to the subsequent open door policy. The first sample is a short story by Wang Wenshi: xin j iéshï de huobàn Two New Friends (Chinese Writers'Association 1963), which is typical of the themes and style of the 1960s with its emphasis on proletarian reconstruction efforts. The

188

CREOLE AND DIALECT CONTINUA

second sample is a story extracted from Hao Ran's selected works: gun i an g dale yào chuj i à The Matchmaker [literally: Big Girl Needs a Husband] (1984: 197-312). This story is also set in the Chinese countryside, and it mirrors the language and social scene of China twenty years later. This type of corpus is particularly interesting to investigate, given the ide­ ological motivation underlying Putonghua "the common language, the language of the people." My goal in analyzing literary language is to determine how writ­ ers over the last thirty years interpreted basic educated PH, and at the same time represented the vernacular speech of the workers and peasants who have become the official heroes in China's recent history, and thus the major charac­ ters in literature (see sections 2 & 3 below). 1.2.2 Colloquial Beijing Putonghua Spoken varieties were recorded in the context of television interviews of the Beijing population covering two distinct popular topics: a popular TV minis­ eries, and an educational program. These interviews were made in the street and in a school in Fall 1985, and directly recorded in Wuhan and Beijing during the 1985-86 academic year that I spent in the People's Republic of China. The cor­ pus yielded a wide range of vernacular PH and is representative of the Beijing model provided to the nation through the medium of television (see section 4). 2. Literature and the cultural revolution: Two New Friends 2.1 Sociocultural components At the pre-Cultural Revolution date of publication of this story (1963), peasants had gained high status and were considered to be the saviors of the nation. Productivity, competitiveness, and political activism are highly valued, thus, we can expect the main characters to be highly vocal and expressive of their lofty goals, and the major medium for such messages should naturally be proletarian speech, as much as it can be represented in the written form. It can, therefore, be assumed that the propaganda literature represented in Two New Friends con­ stitutes the ideal medium to investigate the stereotypical interpretation of peas­ ant speech. The "two new friends"—Zhang La Yuè and wu Shū La A?1—are two young and pretty model peasant women, clearly depicted to induce female interest in militant productiveness. Zhang has already achieved fame and acquired the nickname of chuångj iång, that is, "pathbreaker." However, Wu has just snatched from her the red flag assigned to the most productive commune. Thus the plot is really about Wu's metamorphosis from a submissive "perfect house-

LITERARY AND COLLOQUIAL PUTONGHUA

189

wife" (häo xífü) to a victorious commune leader, which, at that time, is one of the tenets of the new "Chinese-style" feminism. 2.2 Sample of written Putonghua: Text 1: Two New Friends The short representative sample shown below is extracted from Wang Wenshi's Two New Friends (Chinese Writers' Association 1963:124). This short excerpt has been selected to illustrate in detail the context of topicalization and method­ ological issues in analyzing such strategies. The complete text is then analyzed below. The following sample presents the first encounter of the two female characters, Zhang La Yuè (Z), already recognized for her commitment to hard work, and w u Shū Lan (W), who recently raised her social consciousness to move away from her housewife role to that of a competitive worker. The two women chat about their ambitions and their husbands.2 la(Z) "you! ni zenme bu néng ru dáng!"= [oh! you why not can join party!" ] 2.(Z) = Zhang La Yuè dèng-zhe jingqíde yânjîng,= [Zhang La Yue open-ASP surprise-G eye, ] lb.(Z) =kuài shinqïng ba a ! kuai shinqïng ba! ai n u " ["quick apply EP EP! quick apply EP! eh you. . ."] le (W) "yïj Ing shinqïng- le!" ["already(AP) apply-ASP!"] Id (Z) "nà j iu nao" [that(TI) AP good.]  (Z) ni nánrén  bu lä houtuï ba=. [your man must not pull leg EP] 3b(Z) =cóngqicm, tarnen diu shuö nurén lä nánrén houtuï= before, they AP say women pull men legs] 3c (Z)=xiànzài dào-guo lai-le=, [now reverse-ASP come-ASP] 3á(Z)=youxIe nánrén [some men 3e(Z) =nl nánrén shï [your man TI

laqï pull gè CL

nurén-de houtuï la= women-G legs - EP shdyàng rén ?" what type man?]

4(W)Shūlan dadào:"shî dangyuán." [Shulan replied: "TI party member.]

CREOLE AND DIALECT CONTINUA

190

5a(Z)"/?á gèng nao" Zhang La Yuè zhuangzhong-de shuo:= good! Zhang La Yue seriously-G says] [that(TI) AP 5b(Z)=búguo lä z i j ï laopó h o u t u ï - d e dangyuán ye youdë shi ["but pull own wife leg-G partymember AP plenty TI 5c (Z)---wo nàgè sïguï [I that(TI) devil Sdrø^keshi [but

jiùshî AP

ne= EP

z h è l ù huo=T this(TI) type]

ne, la dàodî bèi v6 j iàoyù-guo l a i là~= TE, he in the end(AP) TI I educate-ASP come EP

5e (Z) =duï z í j ï d e nánrén yao j í n g c h á n g j i à o y ù ne=, [TI own-G men must always(AP) educate EP] 5f (Z) =miandé tarnen ban shou ban j icio," [avoid they hinder hand hinder foot."] 6a(W)"wo na wei dàoshi chángchang j i à o y ù vo ne=" [My TI man AP often (AP) educate me (EP)!" 6b(W)=Shūlan vënshùn e r t a n s h u à i - d e shuö, [Shulan shyly(AP) but openly(AP)-G says]

Translation

1. (Z) "Oh, how come you haven't joined the Party yet"! Zhang Layue opened her eyes wide with surprise. "You will apply at once, won't you? Apply at once! Oh, you." 2. (W) "But I have already' 3. (Z) "That's good. Don't let your husband prevent you. In the old days, they were always complaining that women were obstacles, but now, it is the other way round. Some men discourage their wives. What kind of person is your husband?" 4. (W) Shulan answered, "He is a Party member." 5. (Z) "That's better"! Zhang La Yue said seriously. "But many party members now discourage their wives. My man used to be like that, but I have changed him around! Husbands should be taught not to hamper each and every one of our movements." 6. (W) "On the contrary, my husband often educates me"! Shulan said shyly, yet honestly.

LITERARY AND COLLOQUIAL PUTONGHUA

191

2.2.1 Topic strategies in text 1 The topic structures outlined in chapter six are briefly illustrated below in ref­ erence to text 1 : Topic presentation lb: kuài shënqïng ba a (two end particles) Id: na jiu nao 5 : n à  nao (one adverbial particle in each case 5d:keshi ne, tä dàodï bei wo j iàoyù-guo lai i a (one topic ender, one verbal preposition/topic introducer, one end particle) Topic fronting 5d: t d is extracted from its object position Topic repetition The repetition strategy ofshénqîng 'apply' [for party membership] three times (lc-d). There are also several copies of the hout uï (Teg pulling') image to refer to men (or women) impeding their spouses' careers. Table 7.1 displays the distribution of the three major topic strategies in this short excerpt, which includes the two female characters Zhang and Wu. The presentative strategy is clearly the most favored (63.2% of the time), whereas fronting and repetition are much less valued, respectively, 8.1%, and 28.5%. Table 7.1: Topic strategies in Text 1 (Two New Friends) TU=topic unit; TS^topic strategies; TS/TU= number of topic strategies per topic unit

Zhang Wu BOTH

TU 7 3 10

Fronting 4(9.7%) — 8.1%

Repetition 12(29.2%) 2(25%) 28.5%

Presentation 25(60.9%) 6(75%) 63.2%

TS 41 8 49

TS/TU 3.6 2.0 3.1

In (5b), the initial subject topic 1 à zïjï låopo houtuï-de dängyuan 'party members who impede their wives', a complex NP preceded by its rela­ tive clause with the clause-final relativizer de, is clearly separated by the par­ ticle ye from the comment youdë shi A?e'there are plenty (such men)', which itself includes a fronted NP punctuated by the clefting particle sh i (see chap­ ter six: 4.1.1), and finally the sentence is completed by the topic ender ne (see chapter six: 4.2). Both fronting and repetition (but especially fronting) may be accompanied by a presentative particle, or more than one, as shown in (5d) or (5e). Those two

CREOLE AND DIALECT CONTINUA

192

sentences are in fact paraphrases (with repetition of j iàoyù 'educate'), each stating the necessity to educate husbands. In both, the object topic "husband" is extracted from the comment and fronted. In (5d), there are two topics: The object t GL (he=the husband) is extracted from the comment and merely front­ ed, whereas the agent wo (1= Wu, the peasant leader) following the fronted object is introduced by the verbal preposition/TI bèi, thus giving special emphasis to that agent as well (chapter six: 4.1.6): (5d) tå clàodï bèi wo [he, AP TI I

j iàoyù-guo l a i l ä educate-ASP come EP]

'As far as he is concerned, I am the one who educated him'./He was educated by me'

In addition, in (5e) but not (5d), the verbal preposition du i introduces the topic "husbands": (5e)duï zîjïde own-G [TI

nánrén man

yào j ï n g c h a n g j iàoyù ne must always(AP) educate EP]

'Regarding (your) own men, (you) must always educate (them), don't you'

The variable use of object fronting and optional incidence of the verbal prepositions dui and bèi in synonymous utterances clearly support the argu­ ment presented in chapter six, namely that verbal prepositions/coverbs may have pragmatic presentative functions flexible enough to allow for changes in topic salience. This does not imply that all coverbs qualify as topic introducers, or if they do, that they are completely bleached of their original semantic value, such as goal or direction. For example, tóng (5), xiáng (7; 9), and duî (8) preserve their directional values while serving as topic introducers. It may be the case that those semantic roles predispose them to slide into a focusing func­ tion, especially in preverbal position. Since presentative strategies are predominant, it is important to identify carefully the overall distribution of those particles. Table 7.2 focuses on the pre­ sentation strategy in text 1, showing the distribution of particles in each of the two speakers, and the relative use of such particles per TU: 3.6 for Zhang and 2 for Wu. The four particle categories discussed in chapter six (topic introduc­ ers, topic enders, adverbial particles, and end particles) are all represented:

LITERARY AND COLLOQUIAL PUTONGHUA

193

Table 7.2: The presentation strategy in Text 1

TI:

Zhang

Wu

\)(shï)

2

1

2) deictics (nage, nà) 3) verb/prepositions (dui, bèi)

5 2

1

—

— —

4) adverbs TE: (ne) AP: (jiu, dou, gen g, ye) EP: (ba, ne, a, la) TOTAL:

1 8 7 25/7 (3.6)

3 1 6/3 (2.0)

BOTH SPEAKERS: 3 particles per TU (10 topic-units; 24 comments)

2.3 Complete analysis of Two New Friends Following the sample analysis of text one, Wang Wenshi's complete story was examined to assess the veracity of the preliminary findings. In addition to the two female characters' direct speech, the story includes the narrator's indirect language. Close examination of the eighty-seven page transcript confirms the results presented in tables 7.1 and 7.2. Table 7.3 shows that 69.7% of all the topic strategies in the complete text involve some presentative element (out of 857 topic tokens), with reduced proportions of fronting (12.6%) and repetition (17.7%). However, Zhang and Wu differ in the number of particles per TU that they produce (3.5 for Zhang vs. 1.9 for Wu), which is consistent with the pilot study of text one. There is significance in the fact that the narrator's language, which could be assumed to be "standard," is much like the reported speech of the two peasant women, at least in terms of the topic strategies studied here: Table 7.3: Complete topic strategies in Two New Friends TU=topic unit; TS=topic strategies; TS/TU= number of topic strategies per topic unit (average: 2.6 particles per topic unit) Presentation

TS

TS/TU

55

Fronting 26 (10.3%)

Repetition

Zhang

31 (12.3%)

195 (77.4%)

252

3.5

Wu

38

13 (12.4%)

19(18.1%)

73 (69.5%)

105

Narrator

133

69 (13.8%)

102 (20.4%)

329 (65.8%)

500

1.9 2.5

All

226

108(12.6%)

152(17.7%)

597(69.7%)

857

2.6

TU

One must keep in mind the political underpinnings of the literature written at the time of the Cultural Revolution: the language of the peasantry is then pro-

CREOLE AND DIALECT CONTINUA

194

claimed to be "standard" and prestigious, and one can assume that the narrator also endorses the rural identity of the hard-working commune worker. It is for this reason that I consider propaganda stories like Two New Friends extremely useful in terms of their documentation of the 1960s vernacular. Table 7.4 examines in more detail the favored presentative strategy in the complete story, and illustrates the distribution of particle patterns, including the four categories of topic introducers(TI), topic enders (TE), comment introduc­ ers (AP), and comment enders (EP): Table 7.4: The presentative strategy: Topic particles in Two New Friends (Percentages in parentheses)

Zhang Wu Narrator ALL

n/particles [195] [73] [329] [597]

TI 69 (35.4) 15 (20.5) 87 (26.4) 171 (28.6)

TE 3(1.5) 1 (1.4) 11(3.3) 15(2.5)

AP 83 (42.6) 38(52.1) 217(66.0) 338 (56.6)

EP 40(20.1) 19 (26.0) 14 (4.3) 73 (12.2)

The following patterns of particle usage emerge here: 1. Adverbial particles (pre-comment particles) are consistently favored over the three other types of particles. Table 7.4 shows the highest frequency in the narrator's lines (66%), relative to other particles. 2. Although TEs and Tis have a similar function of topic identification, the pre-topic marker is much preferred (28.6% overall), as opposed to the almost inexistent topic enders (2.5%), the lowest distribution among all par­ ticles for all three characters. 3. As to end particles (EPs), which always occur last in a topic utterance, they are primarily reserved for dialogues, denoting direct oral communication; they hardly ever appear in the narrator's text (4.3%), but they appear to have some important communicative and emotive significance in the direct speech of the two women. 2.4 Topic particles and psychosocial factors A complete analysis of the use of topic particles by the various characters clear­ ly underscores the psychosocial dimension involved in the function of dis­ course markers in literature. In order to capture the stylistic reflexes of Zhang and Wu's change of dynamics, as well as the narrator's response to it, the text was divided into three discrete segments for analysis, with special focus on the topic introducers, because they evidence variation between Zhang and Wu, as

LITERARY AND COLLOQUIAL PUTONGHUA

195

represented in table 7.5. As far as the narrator is concerned, he shows consistent participation in the three sections, but there are obvious differences in the dynamics between Wu and Zhang. The first section (thirty-one page transcript) presents the first meeting between Zhang and Wu, and provides the background of both characters. Wu has just snatched the red flag of "most productive commune" from Zhang. It is Zhang who initiates the encounter, nobly proffers friendship and admiration to her challenger, and does most of the talking. Zhang is a consistently assertive speaker: she describes herself as'Vo shîgè huopào xîngzi, y T diån jiu xiàng" T am impetuous, once fired, I will explode' (2NF: 5). In contrast, Wu is shy and her contribution is minimal, as reflected in the number of TU s she pro­ duces (11), versus twenty TUs for Zhang and forty-eight for the narrator (see table 7.5). The second section (next thirty-one pages) includes Zhang's com­ ments on the joys of "life away from home" that Wu is discovering during a meeting in a neighboring village. Again, Wu's direct participation is limited: it is mostly the narrator who expresses her reactions and feelings. The third sec­ tion (last twenty-six pages) focuses on Wu's growing self-confidence. Zhang and Wu return from the meeting and, caught in a sudden storm, stop at Zhang's house where Wu encounters Zhang's work team. Finally, Wu snatches the floor, as she has done with the red flag and decisively responds to the aggressive chal­ lenges of Zhang's team members. She, thus, truly emulates Zhang's behavior. The story ends as Wu, bent on not wasting any more time in spite of the rain­ storm, decides to make a late night return to her village, but not without mak­ ing a detour to visit Zhang's high-yield cotton fields. In this final part, it is clear that the narrator fondly effaces himself behind Wu who embodies the new superwoman: a team leader dedicated to land productivity. In this final section, Wu displays an active speaking style which matches her new behavior, increas­ ing her TUs to twenty, as well as the number of particles used, whereas Zhang and the narrator's numbers of TUs are reduced. The three major topic marking aspects of the discourse of Two New Friends appear to vary according to an individual's personality and degree of social involvement, and those variations are mirrored in the patterning of parti­ cle usage: 1) the number of topic-utterances, or discourse units; 2) the frequen­ cy of particles per topic-utterance; and 3) the relative use of the four types of particles. Table 7.5 outlines the distribution of the first two elements in each of the three sections of the text, and shows that there is a slight increase of Wu's ratio of particles to topic units in section 3 (up to 2.2 particles per TU), but Wu never quite matches Zhang's abundant use of particles (between three and four per TU). It is no surprise that speech—and here specifically topic strategies— would turn out to reflect personality and attitude. Wu is subdued and shy, in

CREOLE AND DIALECT CONTINUA

196

contrast to Zhang's fiery temperament. Although Wu's verbal contributions increase in proportion to her social participation and newly gained as serti veness, they never quite match Zhang's bold aggressiveness: Table 7.5: In-text variation ofparticle use in Two New Friends N=TU

n=particles

%n/N

Zhang-1 Zhang-2 Zhang-3 ZHANG (all text)

20 20 15 55

82 65 48 195

4.1 3.2 3.2 3.5

Wu-1 Wu-2 Wu-3 WU (all text)

11 7 20 38

21 9 43 73

1.9 1.3 2.2 1.9

Narrator-1 Narrator-2 Narrator-3 NARRATOR( all) ALL

48 47 38 133

123 130 76 329 597

2.6 2.8 2.0 2.5

226

2.6

There is naturally a political message underlying Wu's behavioral change: Her increasing self-assertiveness fits well with the psychosocial and political climate of the Cultural Revolution advocated in this story, and its message to women. The new "Chinese-style" feminism is wound around the values of pro­ ductivity and competitiveness as dedication to the development of the nationstate. Women are to be liberated and fulfilled through hard work and competi­ tion with men and with each other, and they should indeed compete at every level: in their work (farming), team-spirit, political awareness, and activism, as well as in beauty. It is clear, however, that this liberated role is intended strict­ ly for younger women. The old woman or mother-in law (iåo popo) takes care of the children and household chores and is depicted as full of admiration for the new shining woman who even surpasses men, and she is fully satisfied to stand in her shadow: (5) nàgè Wu Shu ban a, yÍding you sänge tóu, Huge bang=, have three-CL heads, six-CL arms, [TI-CL Wu Shu Lan TE, must =yïdîng bï wo gäo, bï wo zhuàng,= must than I tall, than I strong,

LITERARY AND COLLOQUIAL PUTONGHUA =rénjia others = vo j iu I AP

197

shuo nï zhangde bï wo xiù= say you look than me pretty bu xi n, x i â n g bu dào, nï zhègè qiao niángr=, not believe, think not arrive, you TI-CL pretty woman,

^ j i n g tóng vo zuô qï duî lai-le AP with(TI) me use challenge target come-ASP] (I have been thinking) 'that famous Wu Shu Lan must have three heads and six arms, she must be taller and stronger than me, people say she is prettier than me. I couldn't believe it. And yet here you are, a pretty woman challenging me' (Zhang; 2NF 7-8). (6)

j i ù q i a n ya, nanér zhî zai s ï f â n g , vu hú si hai = [before TE, men ambition TI everywhere, five lake four sea = j i ä o péngyou r ú j l n , zánmen nurén ye zhî zài s í fang la,= make friends; now, we women AP ambition TI everywhere EP =zánmén ye shi we AP TI

péngyou biàn tiânxià friends all over land]

'In the old days, men exercised their ambition everywhere and made friends in all corners of the land. Now we women can also be ambitious and it's our turn to have friends all over' (2NF: Zhang; 39). However, the narrator's voice constitutes most of the text and plays an important role in this type of didactic literature. The narrator is the Beijingbased teacher and moralist, yet a proletarian anti-intellectual, whose purpose in presenting the story is to incite good (^productive) behavior. His tone is there­ fore one of persuasion vis-a-vis the reader and highlights the function of certain discourse particles in narrative rhetoric, as compared to the direct speech pro­ duced by the two women. 2.4.1 Topic introducers and their rural variants in Two New Friends I will now take a closer look at a specific type of particles (Tis), as produced by the three characters. Although topic introducers are not as predominant as com­ ment introducers (APs), they present particularly interesting problems. Table 7.6 presents the distribution of the three subtypes of Tis defined in chapter six. A breakdown into the four TI categories (deictics, verbal prepositions, and the clefting particle shl) shows that verbal prepositions prevail (37.4% of the

198

CREOLE AND DIALECT CONTINUA

time) somewhat over deictics (31%), and shi (19.9%), with only 11.7% for adverbials functioning as topic introducers: Table 7.6: Topic introducers in Two New Friends (N=total number of Tis) N NARRATOR 87 ZHANG 69 WU 15 TOTAL 171

DEICTIC n % 20 23.0% 25 36.2% 53.3% 8 53 31.0%

V/PREP n % 51.7% 45 14 20.3% 5 33.3% 64 37.4%

ADVERB n % 11 12.6% 13% 9 0.0% 0 11.7% 20

SHI n 11 21 2 34

% 12.6% 30.4% 13.3% 19.9%

Verbal prepositions are not essential in most Chinese sentences, but are usually associated with a more formal, authoritative style, which is commensurate with the moralistic political teachings favored by the Cultural Revolution. It may at first appear paradoxical to find that this feature is commonly accompanied by the use of rural lexical items, but the paradox disappears when the sociopolitical context is taken into account. The writers of the propaganda typical of the Cultural Revolution apparently felt that they would be most per­ suasive if they mitigated their solemn paternalistic messages with the use of popular, and especially rural lexical items. The narrator's sociopolitical mes­ sage is sprinkled with the occasional usage of local rural features, which are typical of Northern Hebei rural speech or of a more conservative variety, and unusual in standard Beijing Putonghua.4 Most of those rural features appear in the use of certain variants for verbal prepositions, such as: x i ¿n g for du! (directional; focus), J i an g for ba (fronted object marker),5 and j iào for bèi (fronted agent marker). The distribution of those alternate verbal prepositions is represented in table 7.7: Table 7.7: Distribution of verbal prepositions in Two New Friends TI

NARRATOR

ZHANG

'to/reach'

18

3

12

2

ba

'take'

9

4

/jiâng bèi

1 7

—

'by/get'

—

—

duï

GLOSS

/xi an g

/j i à o

Wu

1

For each of those verbal prepositions, both variants co-exist in the same

LITERARY AND COLLOQUIAL PUTONGHUA

199

story, and the variant chosen typically signals some stylistic nuance: the rural variant is considered to be more casual, but no single variant is used exclusive­ ly by one character. For example, x i an g (casual) and dui (more formal) alter­ nate in the same constructions: in the complete text, x i an g occurs fourteen times and dui twenty-one times, all mostly produced by the narrator (only five occur in Zhang's reported speech: xiàng twice and dui three times). (7) ta you rèqingde xiàng z î j ï d e zhangfū shuö. [she again(AP) kindly-G to(TI) own-G husband say.] 'she again kindly told her husband.' (Zhang: 2NF; 58) (8)zhangfū dui ta shuô= [husband to(TI) her say.] 'Her husband tells her.' (Narrator: 2NF; 20) (9)td j i u dàshëng xiàng vū lï hândào. [she AP loudly to(TI) house inside shout.] 'she loudly shouts toward the house. (Narrator: 2NF; 50) (lO)zhäng dui haizïmën huïhui shou. [Zhang to(TI) children wave hand.] 'Zhang waves at the children' (Narrator: 2NF; 50) (11) Zhang xiàng ta j ï j i yan [Zhang to(TI) her wink eye] 'Zhang winks at her' (Narrator: 2NF; 57) (12) to" dui ta xiàoxiao, bu shuö shénme [she to(TI) him smile, not say something] 'she gives him a smile, but says nothing (Narrator; 2NF; 20) Other examples of northern verbal prepositions include J i an g for bà in the so-called object construction, and j iào for bei, an agent marker (or the socalled passive construction). Both bà and bèi are discussed in chapter six: 4.1.5 and 4.1.6): OBJECT-MARKER:

of ba):

ba/j iäng. (j i an g occurs only once for seventeen occurrences

200

CREOLE AND DIALECT CONTINUA

(13) ta yîbâ jiang [she at once(AP) TI

yixiù lu dào qi jianchù sleeve roll-arrive-reach shoulder]

'she immediately rolls her sleeve up to her shoulder' (Narrator: 2NF; 9) (14) kuài bà xiùzï là xià lái ba [quick TI sleeve pull down come EP] 'quick, roll down your sleeve' (Wu: 2NF; 10) AGENT-MARKER:

bei/j iào (j iào occurs once for eight occurrences of bei).

(15)zán nàmiàn hóng qi ne? j iào Wu Shū Lan diânzou la. [our TI red flag EP? TI Wu Shu Lan win ER 'Where is our red flag?- Wu Shu Lan has got it.' (Woman: 2NF; 29) (16) ta dàodï bèi wo j iàoyù-guo lái la. [he after all(AP) TI I educate-ASP come EP] 'After all, I have educated him (her husband)' (Text 1: 5d) As discussed in chapter six (4.1.4), none of those particles/verbal preposi­ tions is required to produce a grammatical structure, even when preceding a fronted element. They occur about 50% of the time when an element is front­ ed, and it is clear that they provide a special emphatic and emotional focus on the following element. The clefting value of those presentative elements is evi­ dent, as for example in (15) "j iào wù Shū Lan diânzou la" (It is Wu who won).6 In the early 1960s, high status was assigned to the proletariat, and to the peasant class (which was instrumental in the success of Mao Ze Dong's revo­ lution). This status is reflected in the presence of linguistic features which are typical of rural dialects. This is an interesting validation of language forms which have traditionally been stigmatized in Chinese literature. For the first time in the history of China, peasants are officially praised and consequently their language becomes a model. The recently established norm of the "Common Language" (Putonghua) is undergoing a dynamic change, thus demonstrating that linguistic change is bound to political and social change. 2.6 Summary of findings in Two New Friends In literary representations of rural speech, both narrative style and reported direct speech reflect the regular use of topic marking, and in particular presen-

LITERARY AND COLLOQUIAL PUTONGHUA

201

tative strategies: L APs (adverbial particles which introduce comments) seem to have become a regular component of Chinese syntax, as reflected in Two New Friends, with a high 56.6% overall incidence. Indeed, they appear to fulfill the function of syntactic comment boundary: they occur at the onset of practically every comment and are not especially associated with emphasis. They are thus likely to be unmarked (see also chapter six, section 4.3). 2. Other particles adjacent to topics, topic introducers (pre-topic) and topic enders (post-topic) clearly assume an important pragmatic function. They denote lively contexts and convey emotional overtones, and their usage moni­ tors the power relationships binding characters. Topic introducers obviously hold a very important position in the discourse of Two New Friends, and they occur in indirect as well as in direct quoted language. They are widespread in the narrator's presentation of the characters. Wu Shu Lan, the shy housewife, increases her use of topic markers as she discovers her potential as a victorious team-leader and develops her social participation and positive contributions, thereby gaining self-esteem, but she still has the lowest incidence of TIs (only 20.5% of all her topic particles serve as introducers). Among the three subcategories of TIs identified, deictic s and verbal prepositions are most widely used overall, but they seem to be associated to dif­ ferent stylistic contexts. Deictics are more colloquial (not unlike the use of top­ ical demonstrative adjectives in English as in "so this guy comes up to me and says"). Deictics are most used by Wu (53.3%), the unpolished peasant woman. In contrast, the verbal preposition has a definitely formal or elaborate value in most cases: it is used regularly by the narrator intent on making moralistic didactic statements (he alone produces 51.7% of all verbal prepositions occur­ ring in the text), and by other characters only when they are involved in rhetor­ ical persuasion. Finally, Zhang, the assertive, socially successful woman, uses a fairly important amount of the clefting particle shi (30.4%) which seems to be associated with direct speech and emotional involvement, as will be seen in some of the other texts. 3. In contrast to pre-topic particles, post-topic particles (TEs) are insignif­ icant in this story, amounting to no more than an average 2.5%). 4. EPs occur at the very end of the topic utterance and are associated with direct speech because they hardly occur in the narrative style (only 4.3%), but they have an important function in the two women's speech (20.5% and 26%). They create the necessary discourse connectives between the interlocutors Zhang and Wu, and they also convey emotional overtones, such as approval, request for approval, hesitation, authoritativeness, irony, and mischievousness inter alia.

202

CREOLE AND DIALECT CONTINUA

3. Literature after the Cultural Revolution: The Matchmaker The period immediately following the ten years of the "Cultural Revolution" is characterized by increasing materialism and cynicism. Hao Ran's The Matchmaker reflects the major themes of the 1980s in China: disenchantment with the values of hard work and self-sacrifice extolled earlier, the search for quick prosperity and material gratification and the use of any available strate­ gies to achieve those goals—a "good marriage" being one of those strategies represented in The Matchmaker. The story also places such social attitudes within the historical context of the times: it contains veiled references to the troubled years of the Cultural Revolution and the personal sufferings endured during that time. This story is eminently comparable to Two New Friends because both stories are set in the same Northern Hebei countryside outside Beijing. The profound anxiety grounded in an unstable recent past creates a deep sense of urgency to achieve fast gratifying results. Thus, Zhao Shu Xian feels that her daughter's happiness depends on being married as soon as possible to a man with marketable skills such as driving a truck (he most certainly should not be an intellectual!)-see TU (4) in text 2 below. On the other hand, the young man Zhang Lin who aspires to marry Zhao's daughter hopes to gain through marriage an instant promotion to the higher social status of the Li family (5-6) and will use any device, including flattery and gifts to meet his goal and gain the favor of his prospective mother-in-law, in spite of an earlier rejection. These goals of upward mobility through marriage (as the most expeditious method) are underscored by a cynical realization that hard work does not bring rewards. There is no point passing exams, says the mother, since her daughter will still be stuck with a menial job in the factory. The same sense of inadequa­ cy underlies Zhang Lin's mention of a dedicated model commune director who spent eight painful years to change his poor village into a productive orchard (5). He has been commanded in a Beijing paper and now the commune even boasts a color television, but the underlying question is: "Was it worth it"? In the long run, Zhang Lin uses every possible means—flattery and the promise of fresh fruit from the orchard—to gain Zhao Shu Xiang's favor. Thus, this second story provides an interesting counterpoint to the previous story Two New Friends. Lofty aspirations have vanished. Text 2 includes several excerpts illus­ trating Hao Ran's short story: 3.1 Text 2: The Matchmaker (excerpts) (la)shijiān bu deng rén, rèyuè cuî rén lao=, time not wait people, time make people old,

LITERARY AND COLLOQUIAL PUTONGHUA

203

(1b) =rúguǑ bú k u à i z h e d i ǎ n gei guinǔ xuan gè n j iachūqu= if(AP) not quickly(AP) TI daughter choose CL husband marry-go (lc)

=zhēng kĕnéng dānwù-le tā yí b è i z ï ya really(AP) might delay-ASP her AP all (her) life EP

(Time does not wait and makes people old. If she does not find very soon a hus­ band for her daughter, this might really put her whole life on hold!) ( 2 a ) z h ï s h ì m e i g è r é n yŎu m e i g è r é n - d e t ā n t u , = everyone have everyone-G like-desire TI (2b)=Zhang Lín t ā n t u pan gè gao zhîr= Zhang Lin like-desire move up CL high branch; (2c)=Lì jiā-de dìwèi gao, paizǐ Li family-G land-status high, sign

liang, bright,

(2d) you zhèyàng y í g è zhàngrén jiā duó g u ān g c a i . . .= have TI CL father-in-law family how bright-color. . . (3a)=Zhāng Lín Zhang Lin

yóuqí kànzhòng-le Li especially(AP) see-get-ASP

jiā-de zháiyuàn=, Li family-G reside-place,

(3b)=chéng q î n - d e shíhòu, bu yòng fùqîn cāoxTn=, become related-G time, not need father worry, ( 3 c ) = j i u you fangzï zhù. . .= AP have house live. . . (4a) =j ìnguǎn zāodào j ùj ué, xiǎo huar bìng méi s ï x î n = , though receive refusal, young man even(AP) not give up ( 4 b ) = j i è j ī h u ì jiù gēn zhègè = take chance AP with (TI) this(TI) (4c) =duì nué r hunshì néng dāng b a i f ēn zhī jiushí jiā-de^ to(TI) daughter marriage can decide 100 out of 90 family-G (4d)=Zhào Shŭ x i á n la jìnhū, x ià n y ī n qí n . Zhao Shu Xian pull close, show like-please. 2. 'Everyone has their own obsessions: Zhang Lin's obsession is to marry into a promi­ nent family. The Li family has a high social status and prestige. It would be brilliant to acquire such a father-in-law. 3. Zhang Lin especially covets the Li's residence. When he gets married, his own father would not have to worry about him, for they would have a house to live in. 4. Though he has been rejected once already, the young man has not given up yet. He

CREOLE AND DIALECT CONTINUA

204

is waiting for his chance to establish contact with Zhao Shu Xian (the mother) and flat­ ter her, because she can make 90% of the decisions concerning her daughter's mar­ riage.'[M; 10, ] (5a) lǎo duìzhǎng qīnzì guàshuài dàibïng,= old team-head himself head lead-army, (5b)=zài nàlǐ ānyíng TI there camp

zhāzhai zhengzheng bā nián=, set up camp fully(AP) eight year,

(5c) =y ì qiāo yì -de va dì vu chï,= CL shovel CL hoe-G dig earth five feet, (5d)=ba shàngbiān-de shítóu mai dào dïxià,= TI above-level-G stone bury-arrive earth below, (5e) =ba dïxià-de nao tǔ fan dào shàngmiàn=, TI earth below-G good soil turn-arrive above-level, (5f)=yìngshì ba luànshítān biànchéng-le hua guoyuan= actually(AP) TI scatter-stone-land change-ASP flower orchard. (5g)=rúj īn, tarnen nàge duî chéng-le quán gōngshè= now, they TI-CL team change-ASP whole(AP) commune (5h)=zuì fù-de duì, guāng qïchē jiù most rich-G brigade, even(AP) trucks (5i)=duïbù office (5j)=zài

you sānliàng= AP have three-CL,

ān-le y ì t a i 20 yïngcùn-de dianshìjî,= set up-ASP one-CL 20 British-inch-G tv set

bùu k à n - z h e z h è n g x î n j ï n - d e

AP not see-ASP

earn salary-G

y ǎn h ó n g - l e , =

eye red-ASP

(5k)=lián sïshí suì lǎo guānggùnr dōu qushàng xífù," even 40 year old single-stick all(AP) marry-get wife (The old supervisor himself directed (the commune). It took him eight years to reorganize the village: little by little they dug five feet below ground level, removed the soil from beneath the stone surface, and brought up that good soil to the top, eventually turning the land into a blooming orchard. Now, thanks to that brigade, the village has become one of the richest in the commune. It now owns three trucks, and the office is even equipped with a large television set. There is no longer any reason to envy [to look with red eyes] the bureaucrats. Even bachelors in their forties can find wives.)

LITERARY AND COLLOQUIAL PUTONGHUA

205

3.2 Topic Strategies in The Matchmaker The analysis of the complete story confirms the importance of topic pre­ sentation previously observed in Two New Friends: Table 7.8: Topic strategies in Text 2: The Matchmaker (Overall: 3.6 particles per TU) (Percentages in parentheses)

ZHAO (F) ZHANG (M) NARRATOR ALL

TU 7 12 29 48

Fronting 4(133) 14(21.5) 17(11.5) 35(14.4)

Repetition 5(16.6) 7 (10.7) 24(16.2) 36 (14.8)

Presentation 21 (70.0) 44 (67.6) 107(71.3) 172 (70.8)

N 30 65 148 243

Particles/TU 3.0 3.7 3.7 3.6

All three characters represented in table 7.8 (Zhao, the mother, Zhang Lin, the young man who aspires to marry Zhao's daughter, and the narrator) equally favor presentative strategies with frequencies hovering around 70%. They pro­ duce an overall amount of 3.6 particles per topic unit, which is somewhat high­ er than the average incidence of particles (2.6) represented in Two New Friends. 3.3 Topic Particles in The Matchmaker As in the previous section, I will look closely at the relative distribution of topic particles in the second story (table 7.9). Although presentative strategies are equally favored by all three characters in The Matchmaker and in Two New Friends, there is a difference in terms of the use of the different particles repre­ sented in the predominant presentative strategies: Table 7.9: The presentative strategy: Topic particles in The Matchmaker (Percentages in parentheses)

ZHAO (F) ZHANG (M) NARRATOR ALL

N [21] [44] [107] [172]

TI 12(57.1) 23 (52.2) 57 (53.3) 92 (53.5)

TE 0 0 0 0

AP 7 (33.3) 18 (40.9) 47 (43.9) 72(41.9)

EP 2 (9.5) 3 (6.8) 3 (2.8) 8(4.7)

Whereas APs (pre-comment particles) are consistently favored in the first story (Two New Friends), as seen in table 7.4 ( 56.6% of the time on average),

CREOLE AND DIALECT CONTINUA

206

they are no longer dominant in the second story, written in the 1980s (where they average 41.9%). TIs now take a prominent position, first of the four cate­ gories (53.5%) for all three characters, and for each individual taken separate­ ly. Topic enders are not used at all in this text, and they were scarce in the first story as well. As to end particles (post-comment), they occur minimally (4.7%), even less than in Two New Friends (12.2%). The relatively high frequency of TIs reflects a more common overall use of verbal prepositions (46.7%) in The Matchmaker, as shown in table 7.10, whereas deictics come in second place, with only 27.2%: Table 7.10: Topic introducers in The Matchmaker (Percentages in parentheses)

ZHAO (F) ZHANG (M) NARRATOR ALL

N [12] [23] [57] [92]

DEICTICS VERB/PREP 4(33.3) 5(41.7) 5(21.7) 12(52.2) 16(28.1) 26(45.6) 25(27.2) 43(46.7)

ADVERBS 0 0 9(15.8) 9(9.8)

SHI 3(25) 6(26) 6(10.5) 15(16.2)

Although only 41%) (thirty-five out of eighty-six objects in the story) are fronted, all those fronted objects are preceded by some verbal preposition, mostly duì, bǎ, and bei as found in "Two New Friends," (gei and gen also occasionally occur). Elsewhere, non-fronted objects generally refer to non­ essential elements of the sentence, as discussed above. All three participants are remarkably consistent in their particle choices in this story, more so in fact than in the previous story. There is indeed a complic­ ity between the narrator and the characters: he is them, understands their inner feelings and aspirations, and in fact holds most of the floor, as represented in the much higher number of topic units produced by the narrator (this was also the case in Two New Friends). 3.4 Stylistic level and dialectal features Like the previous text (Two New Friends), Hao Ran's The Matchmaker exhibits a large number of colloquial or strictly rural features, but contrary to the earli­ er story, which makes consistent use of the vernacular form of Putonghua, the 1980s story displays style shifting in a manner which is not reflected in topic strategies. The mother and the young man represent the "common people" and regularly speak the vernacular in colloquial forms. The narrator's voice is more varied, reflecting his role of mediator between the common people and a meta-

LITERARY AND COLLOQUIAL PUTONGHUA

207

level of social analysis and poetic description: he switches from casual to for­ mal or more traditionally literary expressions. Indeed, this style-shifting aptly reflects the ambiguity and instability of the post-Cultural Revolution time; the variable language mirrors the search for the appropriate balance between a fair recognition of the values of the rural classes and fledgling aspirations to a return to literary pursuits. It is interesting to note that none of the rural forms of verbal prepositions found in Two New Friends occur here. Colloquial or rural elements are repre­ sented solely in lexical items and structures. Style-shifting may occur within the same sentence as in (2b) tānt pan gè gao zhīr where the formal use of t antú 'to strongly desire' as a verb contrasts with the colloquial expression pan gè gāo zhīr 'move up to a high branch' (i.e., "marry above you"). Another example of colloquial and specifically rural (Hebei) expression occurs in text 2: lb-c: "ruguo bu kuài zhediǎn gei guinǔ xuǎn gè  j iàchüqù zhēng kěnéng dānwù le tā yí bèizï ya!" (if no husband is found soon for her daughter, her whole life may go down the drain!). The Matchmaker reflects a more recent literary interpretation of the ver­ nacular Putonghua than Two New Friends. There is some style-shifting between rural, colloquial and formal or literary modes in the narrator of The Matchmaker, which probably mirrors the author's uncertainty about the role of literature in the 1980s, and the degree of freedom of expression than can be expected in the depiction of Chinese daily life. This ambiguity no doubt reflects as well the insecurity of the masses, poised between a difficult past and a hope­ ful future, but at the same time wary ofthat future. However, this style-shifting is mostly evident in the choice of certain lexical items, not in the general mech­ anism of topic marking. In fact, the narrator's use of topic mechanisms match­ es almost exactly that of the two characters. 3.5 Comparison of The Matchmaker and Two New Friends The twenty years' time span separating the two stories is clearly reflected in their themes and general tone. Two New Friends has didactic, moralistic overtones, as the narrator assumes the role of stern political advisor and teacher, whereas the narrator in The Matchmaker acts as a sympathetic observer who depicts the inner hopes and justified foibles of a battered generation madly scrambling for a better life. In spite of those psychosocial differences, topic strategies are remarkably identical (table 7.11):

CREOLE AND DIALECT CONTINUA

208

Table 7.11: Longitudinal study of literary PH

1960s (2NF) 1980s (MAT) BOTH

TU 226 48 274

Fronting 12.6% 14.4% 13%

Repetition 17.7% 14.8% 17.1%

Presentation A11N 823 69.7% 243 70.8% 1100 70%

N/TU 2.6 3.7 2.7

There is no major linguistic change observable over twenty years of Chinese proletarian speech, as reported in the two stories investigated, at least as far as presentative strategies are concerned: they are clearly and consistently favored over other devices. However, some difference exists when comparing specifically the various presentative particles represented in table 7.12: Table 7.12: Comparison of Two New Friends and The Matchmaker

1960s (2NF) 1980s (MAT) BOTH

N 597 172 769

TI 28.6% (171) 53.5% (92) 34.2% (263)

TE 2.5% (15) 0.0% 2.0% (15)

AP 56.6% (338) 41.9% (72) 53.3% (410)

EP 12.2% (73) 4.7% (8) 10,5% (81)

Thus, to summarize the patterns observed in the two stories: 1. The previously observed preference for pre-comment particles in the 1960s story (56.6%) is also represented, but to a somewhat lesser extent in the 1980s story (41.9%). 2. Concomitant with the decrease of APs (from 56.6% to 41.9%) there is an increase of topic introducers (from 28.6% to 53.5%). 3. TEs are inexistent in both stories. 4. EPs show a minimal frequency as before. In both cases, verbal prepositions seem to be preferred as topic introduc­ ers (37.4% in the 1960s story, and 46.7% twenty years later). Does this trend toward an increase in topic introducers, and especially the category of verbal prepositions, represent a significant evolution of the patterning of topic strate­ gies? It is now necessary to investigate spoken Putonghua in order to compare the strategics observed above to real life contexts. Observations of contempo­ rary spontaneous Putonghua will help assess those findings in terms of their his­ torical and literary perspectives.

LITERARY AND COLLOQUIAL PUTONGHUA

209

4. Beijing colloquial Putonghua Two television interviews illustrating the Beijing vernacular Putonghua, both recorded in PRC in 1985, have been analyzed in detail in terms of their topicalization strategies; thus this dimension of oral spontaneous and contemporary discourse offers possibilities for a comparison with the literary Putonghua of the last thirty years represented in the previous sections. The corpora of speech data analyzed represent at least four hours of interviews, and illustrate stylistic vari­ ation, ranging from casual (the interviews collected in the streets of Beijing about a popular show) to semi-casual/semi-formal (the interviews about a sci­ ence television program). This range makes it possible to assess the status of contemporary Colloquial Putonghua, and in particular to estimate the relative importance of pragmatic versus syntactic factors in the use of particles across styles. 4.1 Informal Putonghua: sì shì tóng táng (Beijing Saga) The first series of four interviews was conducted in the streets of Beijing and broadcast in October, 1985, by the major Chinese television channel to investi­ gate popular reactions to a miniseries called sì shì tóng tang. A Beijing Saga (literally: Four Generations in the Same House), which depicted recent Chinese history through the life of four generations of Beijing natives. 4.2 Text 3: Sample of Beijing vernacular Putonghua In the following sample, (I) represents the interviewer, and (R) the respondent. Utterances are numbered and counted only for R, the man who explains what he likes about the popular television show, since the TV interviewer has a lim­ ited speech production: (I) NÍn nao NÍn shí lao Beijīng rén ma? [hello! You TI old Beijing person (Q)?] (R)

DUì.

[Yes] (I)- zài zhèr, nǐ zhù-le duóshao nián-le? [TI here, you live-ASP how many years-ASP?] (1)-R- w j i ù zài zhèi hútóngzi [I AP TI that (TI) street

zhù s ì s h í nián-le. live 40 years-ASP

-I- NÍn xïhuān kan guóchan diànshî j ù ma? [you like watch homemade TV show (Q)?]

CREOLE AND DIALECT CONTINUA

210

(2)-R-e- nà xïhuan na [oh- that(TI)like EP] -I-Neng bu néng tan yï tan nǐ x ï h u ā n - e n - = [ not  talk  talk you like -uh=nèi yï l è i - d e p i ä n z i which one kind-G movie

diànshï TV

jù? show?

(3a)-R-Di an shì j ù a, nǐ x i a n g zanmēn p ā i - d e = [TV show TE, you recall we make-G, (3b)-R-= n e i g e "Si Shì g áng" ,= [ TI "Four Generations Same House" , (3c)-R-=nèi p i ä n z i TI show

p ā i - d e zhin háo= made-G AP good]

(4)-R-=ai, tā fǎnyìng [Ah, it mirror

l a o B e i j i n g , nèi old Beijing, TI

dianshìr=. things]

(5)-R-=fǎnzheng wŎ kàn h a i s h ï j ī b e n hái néng kàndŏng=. [Really (TI) I watch AP basically(AP) AP can understand] (6a)-R=zhuyào shì da r i b e n s h í d à i [truly(TI) TI fight Japan period

guo l á i - d e , duì bú duï?= pass-come-G, true not true?]

(6b)-R=nï x i a n g mai húntún mián, nèi chǎng, j i ù s h ï [you recall buy dumpling flour, that(TI) scene, TI

nèi yàng=. that (TI) type]

(7)-R- =TOU t i a n vanshàng qù pái duì qù pái yì xiŭ,= [First day evening go stand in line go stand one night] (8a)-R=Dièr t i ā n j ǐ n g c h á l á i - l e dào z h è r ná b i ā n z i = , [Second day police come-ASP arrive here use whip] (8b)-R=yī hōng, zhun dǎpǎo, zhun mai bù zháo.= [AP drive, then(AP) beat away, then(AP) buy not touch] (8c)-R=hëi -hei - h e i , jiù n è i g e qíngkuàng jiù [Yeah TI that(TI) situation AP

zhème g è s h ì q í n g , that(TI) thing]

(I) Hello! Are you a Beijing native? (R) Yes. (I) How long have you lived here? (1)I have lived in this street for 40 years. (I) Do you like to watch home-made (PRC) TV shows? (2) Sure I do. (I) Can you tell me which kind of TV program you like?

LITERARY AND COLLOQUIAL PUTONGHUA

211

(3) There is that show, you know, that show made in China "Four Generations in the Same House" is a very good TV program. (4) Yeah, it depicts Old Beijing. (5) You see, I find it basically very easy to follow. (6) It is mostly about the time of the war with Japan, isn't it? For example, there is that scene about buying dumpling flour: (7) You get in line in the evening and wait the whole night. (8) The next day, the police come and drive you away with whips, you eventu­ ally can't buy anything. Ah, that's just the way it was. 4.3 Analysis of A Beijing Saga The sample represented in text 3 includes eight topic units with fifteen com­ ments. For example, topic 8 (dìèr t ian 'second day') is associated with six comments (lines 8a-c). The three topicalizing strategies are illustrated in this short excerpt. Fronting is represented, for example, in (1) and (2): (1)ŏ j i ù zài zhèi hutongzi 40 years] (2) e, nà xǐhuān

zhù s ì s h i nián-le ( I on that street lived

na. [Oh, that I like]

Topic repetition appears in (7) and (8): (7) qù pái duì qù pái [go stand in line, go stand] (8) zhun dapao, zhun mai bù zháo [then (they are ) beaten away, then buy nothing] Repetition is also widely represented with paraphrases and other types of anaphora: this is the case for topic 3, for example, (diànshì jù a 'television show'), which is accompanied by three paraphrases, two of which include rel­ ative clauses, all three introduced by topic particles. Table 7.13 displays the rel­ ative use of the three topic strategies in this short sample:

CREOLE AND DIALECT CONTINUA

212

Table 7.13: Topic strategies in Text 3: A Beijing Saga N 37

Fronting 2(5.4%)

Repetition 8(21.7%)

Presentation 27(73%)

TU 8

As previously observed in tables 7.1 and 7.8 for, respectively text 1 and text 2, there is an overwhelming preference for presentative strategies (73% of the time), with an average of over three particles per TU (3.3). The range of par­ ticles found in text 3 is represented in table 7.14, showing a breakdown very similar to that of table 7.2, its counterpart for text 1 : Table 7.14: Topic particles in Text 3: A Beijing Saga TI:

1)shi 2 2) deictics (nǐxiǎng, neǐge, nè i) 9 3) verbal/prepositions (zà i) 1 4) adverbs (fan zheng, J i ùsh i) 3 TE: (a) 2 AP: (jiù, hái, zhēn) 9 EP: (na) 1 TOTAL: 27/8 (3.4): 3 particles per TU (8 TU;15 comments) The topic strategies used in the complete sample of the four combined interviews, which make up the first part of the spoken Putonghua corpus, match almost exactly the short excerpt (8 TUs) represented in table 7.13, and turn out to be again very similar to the patterning observed in the literary representation of Putonghua in the 1960s and the 1980s; The presentative strategy prevails (63.8%) here too: Table 7.15: Topic strategies in the complete A Beijing Saga (set of 4 interviews) N 185

FRONTING 15(8.1%)

REPETITION 52(28.1%)

PRESENTATION 118(63.8%)

TU 37

Prt/TU 3.2

But a different pattern is apparent in some respects in the frequencies of the various particles (table 7.16). The trend toward an increase in topic intro­ ducers observed in the 1980s story (The Matchmaker) is also represented in the Beijing interviews in which TIs constitute as much as 45.8% of all particles. There is also a slight increase in frequencies of the two previously under-repre-

LITERARY AND COLLOQUIAL PUTONGHUA

213

sented categories of topic enders and end particles, such as shi ba in (17c), and others illustrated in text 3. But the most obvious difference lies in the low fre­ quency of APs (29.7%). as compared with 56.6% in text 1 (table 7.4). and 41.9% in text 2 (table 7.9), indicating a definite shift in the choice of topic strategies: Table 7.16: The presentative strategy in the complete A Beijing Saga(4 interviews) (Percentages in parentheses) N [118]

TI 54(45.8)

 18(15.3)

 35(29.7)

 11(9.3)

However, the increase of topic introducers conceals the variations in the internal distribution of those particles in literary PH and spoken Beijing PH. Although verbal prepositions occurred frequently in both short stories, this is not the case in Beijing PH in which they are sparsely used it is least used (only 5.6%o), whereas the favored topic introducers fall clearly into the deictic cate­ gory (55.6%): Table 7.17: Topic introducers in A Beijing Saga (Percentages in parentheses) N [54]

Deictics 30(55.6)

Verbal Prepositions 3(5.6)

Adverbs 13(24.1)

SHI 8(14.8)

In addition, the adverbial category functions somewhat frequently to intro­ duce topics as well (24.1%). For example, one man repeatedly uses the adver­ bial fanzheng 'really/you see/you know' to introduce each comment related to the topic (TV show Beijing Saga): (17a) fǎnzheng na j í ye bù qué [Really(TI) any part AP not miss] 'you know, I won't miss any part' b) fǎnzheng kàn le tǐng dong zhï xīn de [Really(TI) see-ASP AP move self heart-G 'you know, it's a moving show'

214

CREOLE AND DIALECT CONTINUA

) fǎnzheng shì bïjiâo xiànzai bïj iao guoqù, shì ba [Really(TI) TI compare now compare past EP] 'you know, it's a comparison of past and present, ok?' [Interview B] Does the observed shift in topic introducer choice represent an ongoing trend of spoken Beijing PH? The analysis of another corpus of spoken colloquial vari­ eties may help determine the answer to this question. 4.5 Semi-casual Putonghua: zhōngxuéshëng (Learning Science) The second corpus of Beijing vernacular consists of twenty-one interviews investigating the reactions of students, parents, teachers and school officials to a television science program for junior high school students entitled. Learning Science ghōngxuéshēng (literally: 'Middle School Student'). This series has the potential of exhibiting the patterning of linguistic variability in Putonghua as related to differences in social status, age and gender. Because of the topic treated, it is likely that none of the varieties produced in Learning Science, an educational program, will be as casual as those produced in the discussion of "Beijing Saga." The latter is a widely watched and highly popular show which triggers completely casual reactions. In contrast, the seriousness of the acade­ mic topic, namely, the evaluation of the educational value of television science programs carries a definite degree of formality. This is well represented in a general comparison of the two samples in terms of particle frequencies repre­ sented in the presentative strategy: Beijing Saga has an average of 3.2 particles per topic-utterance, whereas Learning Science has a much lower incidence (1.8 perTU). There are other discrepancies between the first and the second sets of inter­ views. The second set includes many more respondents (twenty-one) than the first set (only four), but the four interviews of the Beijing Saga are longer than the twenty-one samples produced by the Learning Science participants. Those 21 respondents also represent clear variations in age, and status, as stated by the interviewer who clearly announces: "now, we are going to see what students/ school officials/ parents/ teachers think of the science program." There is no such occupational definition of the speakers in the case of the four persons interviewed in Beijing Saga (basically only their sex is apparent, but they do not overtly appear to differ in their social status). The social/occupational variations announced in Learning Science are likely to be reflected in the distribution of the linguistic features under consideration. And indeed they are: in summary, the speech varieties produced by the officials (editors of the science magazine and school administrators) are more formal than those produced by students,

LITERARY AND COLLOQUIAL PUTONGHUA

215

whereas teachers and parents seem to hold an intermediate position. Table 7.18: Topic Strategies in Learning Science (21 respondents)

7 STUDENTS 2 PARENTS 4 TEACHERS 8 OFFICIALS ALL

N [216] [ 69] [ 68] [295] [648]

Fronting 24(11.1%) 5 (7.2%) 7 (10.3%) 22 (7.5%) 58 (9%)

Repetition Presentation 51 (23.6%) 141 (65.3%) 24 (34.8%) 40 (58.7%) 23 (33.8%) 38 (55.9%) 113(38.3%) 160 (54.2%) 211(32.6%) 379 (58.5%)

TU 72 17 20 102 211

Prt/TU 2.0 2.4 1.9 1.6 1.8

Table 7.18 displays an interesting gradation of topic strategies leading from students (most casual) to officials (least casual). Although all groups still lean toward the use of presentative strategies as observed in all other samples (an average of 58.5% here), an increasing formality is apparent, from students to education officials, concomitant with a slightly decreasing incidence of presentative elements form 65.3% (students) to 54.2% (officials). Students consti­ tute the most habitual users of that strategy, thus matching the speakers dis­ cussing the Beijing Saga show who choose presentative strategies 63.8% of the time. Parents, teachers and officials show a decline (in this order) of the pre­ sentative strategy, with corresponding increase in repetition, though not in fronting mechanisms. This relatively increasing use of the repetition strategy may be related to the obvious efforts made by interviewees to clearly express their views on the value of the science program for a television audience. For example, one of the young students (female) interviewed about the science program for middle-school stu­ dents produces the following statement, which includes extensive repetition (underlined) as well as presentation (particles appear in bold characters): 1- wŎ juéde zài xuéxiào zhōng you zhème yïgè [I think at(TI) school middle there is this(TI) one

wèntii= problem:]

2-—j iùshì nán nǔ tóngxúe jiāowang ne, hen bù zìrán= [AP male female student contacts ,  not natural] 3-=jiùshï nán nu shing shǖohùa ne, ye tǐng bièniu de= [AP male female students speak TE, AP AP awkward EP] 4-=wo j i ù xïwàng ne zài diànshï tài zhōng= [I AP wish TE, TI TV station on

216

CREOLE AND DIALECT CONTINUA

5=tíchū zhè yàng yigè v è n t i [raise this(TI) kind CL problem] 6-=zěnyàng bāngzhù women zhèngquède = [how help us correctly-G ] 7-=jiějué [solve

nan nú shëng z h î j iānde guānxì = male female students between-G relationships;

8—wo zìjǐ [ I self

gèrén r è n v é i , wŏ j u é d e nán shëng nu shëng jiāowâng ba= person think, I feel male student female student contacts EP]

9-=yīnggāi [ should

hěn z ì r á n d e j i ù s h ì s h û o , bù yào yŏu shënme j i è x i à n ^ AP natural-G; that is to say, not know have which boundary; ]

10=zhè shŏuxiān necóng nu xuésheng wo zuowéi y î g è nú tóngxué ha= first TE, TI female student -I as one-CL female student TE] [TI ll-=wo j u é d e cóng nu tóngxué i á i j i a n g ba= [I feel TI female student say speak EP] 12-=shouxiān yïnggāi j i ù s h ì t è b i e dàda fängfang, b i è n i u n i u n i ë n i e . [first must AP AP natural and poised, not affectedly bashful] Zhōngxuésheng

(St B: 35-36)

1. In my opinion, there is a problem in middle school: 2. namely, the interactions between boys and girls are not very natural 3. that is, boys and girls communicate in an awkward manner 4.I merely wish on television 5. to raise this specific question: 6. how can we help us to adequately 7. solve male/female relationships 8.I myself think, I feel that male-female contacts 9. should be natural, that is to say,(they) don't know the boundaries 10. So first, from a female student's (point of view), as a female student myself. 11.I feel, speaking as a female student, 12. that girls especially should act natural and poised, rather than affectedly bashful. When closely investigating the still predominant presentative strategy in "Learning Science," topic introducers lead (35.9%), but the categories of APs (especially) and TEs are not too far behind. The increase in topic enders is noticeable (26.9%) as compared to previous corpora, but note that the Beijing interviews showed more TEs (15%) than any of the literary PH texts. This trait may be relat­ ed to the communicative needs of a conversation. In contrast, end particles have very low frequencies as before (5.3%):

LITERARY AND COLLOQUIAL PUTONGHUA

217

Table 7.18: The presentative strategy; Topic particles in Learning Science (N) [379]

TI 35.9%

 26.6%

 32.2%

 5.3%

This distribution seems to indicate a trend toward a more balanced use of the various particles (except EPs). 4.6 Conclusions: Topic marking in spoken Putonghua In table 7.20, both sets of interviews—Beijing Saga and Learning Science—are placed side by side in order to highlight current trends in Beijing Putonghua. In this table teachers and parents are grouped together (because of the small num­ ber of individuals in each category, and also because they exhibit similar fre­ quencies; all three groups of Learning Science interviews are compared to the Beijing Saga interviews, ranging from most casual to most formal. No clear dif­ ferentiation emerges from this grouping, but there is in fact a generally uniform pattern which shows that TIs are most common, followed by APs and TEs, and finally EPs which are consistently minimal. Parents and teachers match most closely the first set of Beijing Saga street interviews in terms of their preference for TIs, which seems to play an important role in "adult" popular language (45.8% and 56.4%), but not so much in students' and officials' speech who use APs and TIs about equally in the 30% range, as shown in table 7.20 displaying topic particle frequencies across the whole Beijing Putonghua vernacular corpus: Table 7.20: Relative frequencies of particle-types in Putonghua

I. BEIJING SAGA II. LEARNING SCIENCE 1. Students/7 2. Parents+Teachers/6 3. Officials/8 ALL TOTAL 2 TEXTS:

(N) [118] 54

TI TE (45.8%) 18(15.3%)

AP EP 35(29.7%) 11(9.3%)

[141] 43 [ 78] 44 [160] 49 [379]136 [497] 190

(30.5%) 37 (26.2%) 49 (34.8%) 12 (8.5%) (56.4%) 16 (20.5%) 13(16.7%) 5(6.4%) (30.6%) 48 (30.7%) 60(37.5%) 3(1.9%) (35.9%) 101 (26.6%) 122 (32.2%) 20 (5.3%) (38.2%) 119(23.9%) 157(32.8%) 31(6.2%)

Table 7.21 displays specifically the distribution of topic introducers. The spoken Putonghua data are divided approximately in terms of stylistic level. The informal group includes all the Beijing Saga interviews, as well as part of the Learning science interviews shown on table 7.20, namely sets 1 and 2 which

218

CREOLE AND DIALECT CONTINUA

include students, parents, and teachers. The formal style is represented by the officials' interviews from Learning Science, (set 3 in table 7.20): Table 7.21: Distribution of topic introducers in colloquial PH DEICTIC

(zhè, z hè) n %

V/PREP

(dù i, ba) n %

ADVERB

shì

(fǎnz hèng) n %

n

%

INFORMAL

[Beijing Saga] [Learning sci: 1,2]

[54] [87]

30(55.6%) 32(36.8%)

3 26

(5.6%) (30.3%)

13 15

(24.1%) (17.2%)

8 14

(14.8%) (16.1%)

[49] 21(42.9%) [190] 83(43.7%)

8 37

(16.3%) (19.5%)

1 29

(2.0%) (15.3%)

19 41

(38.8%) (21.6%)

FORMAL

[Learning sei: 3] ALL

Deictics are the most highly favored overall (43.7%), and consistently so across the different age/occupational groups. They seem to be especially com­ mon in the most vernacular style (Beijing Saga) in which they amount to as much as 55.6% of all TIs, but they are also frequent in the officials' formal speech (43.7%). Verbal prepositions (only 19.5% overall) no longer hold in spoken Putonghua the prominent position they had in literary representations of Putonghua, especially in the 1960s story. However they reveal an interesting differentiation in their distribution, Beijing varieties showing little use of coverbs, whereas almost a third of all TIs used in Learning Science informal speech falls into this category. One must therefore conclude that the use of ver­ bal prepositions is not just a formal device as was hypothesized previously in the context of literary Putonghua, since Learning Science officials, in their oth­ erwise more' elevated' speech rarely use them (16.3%). One possible explana­ tion for the high occurrence of coverbs in student/parent/teacher interviews is that they hypercorrect, and that verbal prepositions are perceived as a rhetorical device. Adverbials are especially avoided by the group of officials (there is only one occurrence in this group), which confirms that the use of adverbials as TI is primarily colloquial, but the clefting particle sh ï is particularly present in the officials' speech (38.8%), more so than in other Beijing Putonghua samples. Although it is impossible to look at each of the numerous particles used in Chinese, there is some indication that each type may have a particular function in discourse. For example, as noted above, topic enders which have been tradi­ tionally identified as the sole topic particles (see chapter six), were found to be

LITERARY AND COLLOQUIAL PUTONGHUA

219

almost inexistent in most varieties. Yet, they account for almost 30% of the topic particles in the "science" interviews, and officials display the highest inci­ dence of the TE category; in particular, interesting patterns emerge from the dis­ tribution of the topic enders a and ne, which constitute about 75% of all topic enders. They seem to be socially diagnostic and to be linked to the speaker's occupational status. In the Learning Science corpus, a denotes high "cadre" sta­ tus, and is perceived by native speakers of Chinese as "assertive" or "snooty," as reported to me, whereas ne is not. This evaluation is in fact supported by both television samples, since 60% of the thirty-seven instances of a found in both interviews occur in the eight officials' speech, whereas the seventeen remaining speakers produced only fifteen, and prefer to use ne (thirty-four times), as shown in table 7.22: Table 7.22: Distribution of the topic enders a and ne

a (high status) ne (non-status) Both

Officials (8) 22 18 40

Others ( 17) 15 34 49

Total/All TEs 37 52 89/119(74.7%)

There is no doubt that the two sets of interviews analyzed above exhibit some linguistic variability conditioned by stylistic level, topic and respondent status. However, general topic marking strategies show a remarkable degree of agreement: there is consistent and widespread use of presentative elements of all types to mark off syntactic boundaries. Since an overall total of 248 utter­ ances were identified in these two texts, with a total number of 497 particles, this means that exactly two particles occur in each utterance on average. These findings clearly confirm that topic particles play an important role in the oral discourse processes of native Putonghua. The only observable differences involve the specific choices of particle type: some individuals prefer pre-comment particles and others favor topic introducers. Furthermore the specific grammatical categories of those subtypes seem to be open to some option: deictics are preferred in spoken Putonghua, and verbal prepositions in literary PH. 5. Overview of topic strategies in Beijing PH The primary goal of this chapter was to establish some empirical evaluation of the current state of Putonghua that could function as a point of reference for the study of second dialects of Chinese. The conclusions presented are based on a corpus of data representing 1) recent literary forms of Putonghua displaying the

CREOLE AND DIALECT CONTINUA

220

socio-political dynamics defining the standard, as represented in two short sto­ ries written in 1963 and 1984, and 2) spoken Putonghua as derived from two sets of television interviews recorded in 1985. Quantitative measurements support the claim that topic strategies, includ­ ing the extensive use of topic particles, constitute a major element of Chinese discourse which has not changed since the 1960s, and have become an essen­ tial feature of the syntax. In a language devoid of case marking and of most morphological markers, except aspect morphemes, the topic elements effec­ tively identify propositions, and are in fact essential to language processing. This seems to indicate that particles serve to grammaticalize topic-comment structure. The results presented in this chapter are summarized in table 7.23, which incorporates the longitudinal dimension as well as the stylistic variation: Table 7.23: Comparison of topic strategies in four sets

Lit PH Lit PH Sp Saga Sp L. Sc

Year (1963) (1984) (1985) (1985)

Fronting 12.6% 14.4% 8.1% 9.0%

Repetition 17.7% 14.8% 28.1% 32.6%

Presentation Part/TU 64.7% 2.6 3.7 70.8% 3.2 63.8% 58.5% 1.8

(Table 7.3) (Table 7.8) (Table 7.15) (Table 7.18)

Below is a summary of findings that will be relevant to the next chapter: 1) Topic presentation is consistently the preferred topic marking strategy throughout all corpora. There is, however, an increase in the repetitive strat­ egy in spoken varieties. The only clear difference across varieties appears to involve the specific choice of topic particles: Topic introducers are general­ ly favored (53.5% in the 1984 story vs.38.2% on average in spoken PH), although the 1965 story uses more APs). The other difference relates to the specific choice of topic introducers, indicating a preference in literary Putonghua for verbal prepositions (37.4% and 46.7%), and in spoken Putonghua for deictics (average of 43.7%, ranging from 55.6% to 36.8%). 2) There is no significant difference between the vernacular represented in the 1965 text, and the vernacular represented in the 1984 story, and the one actually spoken in the 1980s. 3) The peasant vernacular reflected in both short stories matches the Beijing dialect recorded for television interviews in the 1980s. This would imply that the official efforts to standardize Chinese as a "common language" have been successful. 4) A short sample text appears to be representative of a much larger text. This indicates that my limited samples are accurate representations of the current

LITERARY AND COLLOQUIAL PUTONGHUA

221

status of the accepted Putonghua norm. The lack of differentiation in terms of overall particle usage between Literary and Colloquial PH may be surprising in another situation, but not in the Chinese context. In fact, such relative homogeneity reflects quite interestingly the recent social history of the People's Republic of China and .closely mirrors the attempts made to eliminate previous social inequalities (De Francis 1950). Because language is a reflex of social differences, the emphasis on the ver­ nacular represented in the literature examined here reflects primarily the authors' interpretations of the common people's language. In fact, their inter­ pretations turn out to match quite accurately the actual language of the people interviewed in the streets of Beijing in 1985 about a popular TV show depict­ ing the daily lives of the common people. The official policy to legalize the "Common Language" thus seems to be fully justified and implemented, at least in Beijing, and in the literature written by writers raised and educated in Beijing. It remains to be seen how other speakers handle the situation, and this will be indeed the subject of the following chapter. However, it would be inaccurate to claim that there are no "class" differ­ ences represented in Putonghua. The interviews about the science program clearly shows that the new "cadre" class is also linguistically separated from the rest of the population. The difference is observed here at least in terms of parti­ cle usage. Given the above findings, and in particular the overwhelming impor­ tance of topic presentation as a pragmatic mechanism of Chinese syntax in nat­ ural discourse. I will.now proceed to examine other dialects of Chinese pro­ duced as second dialects. Endnotes 1. The various categories of particles are identified as TI (topic introducers), TE (topic enders), AP (adverbial particles), and EP (end particles), as discussed in the previous chapter. Topic par­ ticles are highlighted in boldface. 2. A literal translation appears within brackets, and a regular translation is provided below the text. Topics are italicized, all particles are identified as TI, TE, AP, or EP, and utterances are numbered for easy reference. 3. There are other types of "topic repetition," such as the four paraphrases, pronominal and nominal, of the initial topic 'nánrétf man/husband' (3a-e), a topic taken up again in 5c as sǐguï 'devil', and again repeated and paraphrased four times after that tā (5d), zïj ǐ-de nánrén (5e), tārnen (5f), and wo ná wei (6a)]. However, only the actual duplication of a topic is included in this analysis. 4. This evaluation is based on the judgements of Beijing educated natives who spent part of the Cultural Revolution in the Northern Hubei countryside.).

222

CREOLE AND DIALECT CONTINUA

5. j iāng was historically standard in Beijing Mandarin but was later replaced by bū, as dis­ cussed in chapter six. It is still found in conservative varieties, such as as Cantonese (Matthews &Yip 1995). 6. Jeung, the Cantonese cognate of J iāng, is said to be "primarily restricted to cases where motion takes place" (Matthews & Yip 1994:144).

Chapter 8 Topic Strategies in Varieties of Putonghua as Second Dialects Abstract This chapter analyzes topic strategies in a corpus representing varieties of Chinese acquired as second dialects, and compares them to the first dialect varieties of Beijing Putonghua examined in chapter seven. The varieties analyzed here include Putonghua as acquired by native speakers of Southwestern Mandarin (especially Wuhan-based) and a stylized variety of Suzhou Wu spoken as a second dialect and evidencing exten­ sive style shifting. 1. Introduction The previous chapters revealed regular topicalizing patterns in the vernacular represented in literature spanning the last thirty years, as well as in the collo­ quial Putonghua spoken in Beijing. Cross-stylistic homogeneity of topic struc­ tures in first dialects is characterized by the overwhelmingly preferred use of the presentative strategy over the other two types (fronting and repetition). Having established a reliable background reference, non-native forms of Putonghua and of Suzhou Wu will now be examined in order to ascertain to what extent they display similar topic marking mechanisms, and ultimately to assess whether the selected discourse feature will exhibit differences that may be linked to the effect of second dialect acquisition. On the hand, if no such dif­ ferences emerge, we would be led to consider the potential function of such dis­ course mechanisms as procedural uni versais. The corpus investigated in this chapter includes the following varieties: 1) Non-native informal Putonghua as produced by educated speakers in Wuhan, Hubei (Central China) during the 1985-1986 academic year, and in spring 1988. Conversations were recorded from central and southern China

224

CREOLE AND DIALECT CONTINUA

natives, most from Hubei, Hebei, and Hunan, and range from an informal stan­ dard Wuhan dialect (Wuhanhua, a variety of Southwestern Mandarin) to forms of Putonghua, produced as a second dialect. 2) Suzhou traditional oral tradition: The Suzhou story is a well-known type of oral literature, spoken in a variety of the Wu dialect (see chapter five). The stories are very popular and regularly told on the radio, and some segments of those tales were recorded in Suzhou in spring 1988. They illustrate a variety of Chinese which differs from modern colloquial Suzhou Wu and is acquired as a second dialect. These two distinct varieties were selected both for practical and strategic reasons. First, Wuhan varieties were directly available to me and observable because I was living in Wuhan, Hubei and in close contact with educated Wuhan natives who had to speak Putonghua in their daily professional activities. Furthermore, I personally experienced and struggled to overcome the discrep­ ancies existing between the "Mandarin" I had learned and the broadly-ranging versions I heard in Wuhan. As far as the choice of Suzhou Wu is concerned, it constitutes a second dialect as well, since it is a traditional variety of Wu used in the recitation of well-known vernacular stories. Secondly, it seemed like an appropriate strategy to investigate widely different second dialects of Chinese in order to compare them to creole acrolects, and thus come to conclusions in rela­ tion to the place of universals in pragmatic aspects of syntax. 2. Putonghua spoken in Wuhan As in most regions of the People's Republic of China, the five-million population of Wuhan, located in the Central China province of Hubei, do not speak Beijing Putonghua (PH) as a native variety. The Wuhan vernacular (Wuhanhua) is a vari­ ety of southwestern Mandarin which differs extensively from PH, and most strik­ ingly in its phonological and suprasegmental features (see also chapter five). However, PH is, at least officially, the language of education, officially mandated to be regularly used in schools and universities. The situation almost anywhere in China is one of typical diglossia, combining the use of the local variety at home and in casual contexts, with that of the national language for official business and educational activities; therefore, all educators are expected to be at least bi-dialectal, though they are more likely to be multidialectal. 2.1 The Wuhan corpus This corpus of speech data consists of spontaneous conversations between col­ lege instructors ranging in age from twenty-five to forty-five, my students in the

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

225

linguistics courses that I taught as part of the Advanced Teacher Training Class in Central China University of Science and Technology in Wuhan, Hubei. Students in the class came from various parts of central or southern China, including Guangdong province, where the very distinctive Cantonese (Yue) is spoken, and many among them could not understand their classmates' respective native dialects. It was thus natural for them to use Putonghua in informal interaction.1 The sample selected for analysis and illustrated in text 1 below represents the relatively casual use of Putonghua as lingua franca by three women who had all learned the official variety as a second language, but claimed, and were reported by colleagues, to be fluent in it. Those three women (PG, WG and WN), all in their early or mid-thirties, met in my apartment in a relaxed setting, and had a long con­ versation about the year of study they had spent in Wuhan, and the sacrifices that this situation imposed on their families. PG is from Hunan, and WG and WN from Wuhan; the conversation lasted about two hours. Their speech can be characterized as mostly informal, but not completely casual, in spite of the fact that they are very close friends. They are very much aware of the purpose of the recording, and of my interest in dialectal varia­ tions.2 Furthermore, since they primarily associate PH with the educated speech required in a professional context, it seems unlikely they can ever be complete­ ly informal in their usage of the standard, and it is perhaps the case that their stylistic range in this second dialect is limited as well. This type of context—a casual conversation between friends from different linguistic backgrounds who use PH as a lingua franca—permits to identify the upper extent of their sponta­ neous control of PH acquired as a second dialect. 2.2 Text 1: Wuhan non-native Putonghua In the following excerpt from the conversation between the three young teach­ ers, PG (age thirty-two) expresses her sadness at the thought of leaving her classmates to return home after a full year of study, and reflects with some degree of regret and guilt on the difficult times that she went through during her one-year separation from her young son. Non-native PH shows a wide variety of tonal idiosyncrasies, and no attempt is made to represent those complex differences because they do not affect the discourse feature under consideration. Speakers of second dialects usually have difficulty acquiring standard tonal patterns because native suprasegmental features often interfere in their version of Putonghua as a sec­ ond dialect. (1) aiya, wo yizhi zheme xiang, zenme gao-de, hen kuai ya! [ah!, I AP TI think, how come-G, AP fast EP!

226

CREOLE AND DIALECT CONTINUA

(2a) zhi you liangge xingqi, women jiu yao zou-le [AP have two weeks, we AP will leave-ASP (2b) yi xiang-dao women sange ren zai zheli gao-le yi nian na, [once think-arrive we three persons TI here arrive-ASP one year EP, (2c) yixiazi yao zou-le, zhende xinli haoxiang bu tai shufu [ AP will leave-ASP, AP at heart seems not AP comfortable 3a) jintian jiushi women chi-guo fan yihou, women huilai, [today AP we eat-ASP meal after, we return-come, (3b) jiu zou yi zou, wan yi wan, zheyang hao yixie. [AP walk-walk, play-play, this kind(AP) good a little. (3c) xianzai hai bu tai xiang jia, shuo ming-de, [now AP not AP miss family, say truth-G (4) wo jiushi zai wuyijie yiqian a, wo tebie xiang jia, tebie xiang xiaohai a [I AP TI May Day before TE, I AP miss family, AP miss child TE (5a) yige shi, zhuyao shi wo airen, ta gongzuo you mang a, [one TI, AP TI my husband, he work AP busy TE, (5b) you shi yao shang ke, yao dai haizi [AP TI must teach class, must look after child (5c) women jia, baba, mama, dou meiyou zai-le, jiushi, yao ta gao le, [our family, father, mother, all(AP) not be-ASP, TI must he do-ASP, (5d) xiaohai zheme xiao, cai si sui ban, shenme shiqing, quan yao ta zuo [child AP small, AP four years half, whatever thing, all(AP) must he do (5e) you yao dang mama, you yao dang baba [APmust act mother, AP must act father (6) suoyi wo a, youdian xiang ta, xiang xiaohai la. [TI I TE,AP miss him, miss child EP (7) you shihou xiaohai, lao bu zai jia la, xin zhong, jiu wen wo [sometimes child, AP not be home TE, letter in, AP ask me:

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

227

(8a) "mama shenme shihou hui-lai ya? mama zenme hai bu hui-lai? [mother what time return-come EP? mother how AP not return-come? (8b) mama yijing ba women dou wang-le" [mother AP TI us AP forget-ASP (9) aiya, wo na shihou tebie xiang hui jia [ah, I that(TI)time AP want return home (10a) houlai, j iushP wo yi fang-le jia yihou ma,jiuhui-qu-le, [later, TI I once have-ASP vacation after TE, AP return-go-ASP, (lOb)hui-qu yihou ne, wo jiu dao youeryuan qu jie ta le, [return-go after TE, I AP arrive kindergarten go meet him ASP, (10c) jie ta yihou le, hen duo xiaohai zai nali, wo jiu shuo: [meet him after ASP, AP many children TI there, I AP say (10d) mingming, mama hui-lai-le, mingming, guolai" [Mingming, mother return-come ASP, Mingming, come over" (1 la) ta genben jiu bu li wo, ta ba tou zheme mai-qi-lai, [he simply AP not acknowledge me, he TI head thus(TI) bury-bend-come (11b) "mingming, mama hui-lai-le," ta hai bu li wo [Mingming, mother return-come ASP, he AP not acknowledge me, (12a) houlai tamen ayi jiu shuo: "mingming, ni mama hui-lai le, [later their teacher AP say: "Mingming your mother return-come-ASP (12b) ni kuaidianjiao ni mama," ta dou meiyou jiao. [you quickly greet your mother,"he AP not greet (12c) aiya, wo dou yao ku qilai-le, zhende. [Ah, I AP want cry come-ASP, EP. (13a) houlai, wo jiu qu bata bao jilai, bao-zhe, [later, I AP go TI him carry-come, carry-ASP, (13b) women jiu hui jia-le, [we AP return home-ASP,

228

CREOLE AND DIALECT CONTINUA

(13c) jiushi3 wo na shihou dou hai meiyou jin wu, [TI I TI time AP AP not enter house, (13d) wo jiu xian qu kan ta- hui-dao jia yihou ne, [I AP first go see him-return-arrive home after TE, (13e) ta hai bu li wo. [he AP not acknowledge me (14a) houlai, yihou deng-dao dier tian ne, ta cai qinjinwo, [later, after wait-arrive second day TE, he AP close me, (14b)cai kaishi jiao wo zuo mama, [AP begin call me as mother; (14c) yihou, wo jiu xiang: aiya, zhen bu rongyi! [TI I AP think: Ah, AP not easy] (1) Well, I was thinking: how could time fly so fast? (2) Only two weeks, and we'll be gone; we have been here, all three of us, for one year and it's sad to think of leaving. (3) Today after lunch, we took a walk and chatted, and I felt a little better; so now, I don't feel homesick, to tell you the truth. (4) Before May Day, I very much missed my family, especially my child. (5) One major problem is that my husband has so much to do: he must teach, and look after the child, since our parents don't live with us, so he has to do all of that (alone); especially since the child is so small, only four and a half, so he has to do everything, be a mother as well as a father. (6) I miss my child very much. (7) Since I have been away from home, my child sometimes asks in the letters: (8) "When will mommy come back? How come mommy isn't home yet? Mommy has forgotten us." (9) Ah, I especially want to go home at such times. (10) Later, when I went home for the holidays, and ran straight to the kinder­ garten to meet him. There were lots of kids, and I called for him: "Mingming, mommy is back, come over"; (11) But he simply ignored me, hiding his face like this. I said again: "Mingming, mommy is back"; he still refused to greet me. (12) Then, the teacher told him: "Mingming, your mother has come back, say hello to her, quick"; he still refused: I almost cried, really. (13) Then, I picked him up and took him home—I had gone to him directly,

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

229

even before coming home. When we arrived home, he still ignored me. (14) Not until the next day did he start opening up and calling me 'mommy'. Well, I thought, this is not easy! (WUH-1: 25-33) 2.3 Topic strategies in the Wuhan corpus 2.3.1 General patterns The sample represented in text 1 exhibits familiar patterns of topic marking which are also largely confirmed by an analysis of the whole text: the predom­ inance of presentative strategies is the most salient in PG's speech, with a high 78.2% frequency, and a much lower use of fronting and repetition (table 8.1): Table 8.1: Topic strategies in speaker PG (text 1) [5.6 particles per T/UJ N 101

Fronting 6(5.4%)

Repetition 16(15.8%)

Presentation 79(78.2%)

PG produces a large number of topic particles, seventy-nine particles for fourteen topic units, thus an average of 5.6 particles per TU. This very high incidence of topic particles in PG's speech (restricted to text 1) appears to be closely linked to the emotional content of her communication, as she recounts the difficulties involved in spending a year away from her two-year-old child and from her husband. In another part of her conversation with the other women (they plan to visit a friend's home in Wuhan, and lengthily discuss complicated directions), the incidence of particles drops, and PG averages 2.4 particles per TU for the whole text, thereby matching the other women in this respect (table 8.3, see below) as well as in other features. The three women's overall prefer­ ence for presentatives is reflected in the general analysis of topic strategies shown in table 8.2, which is based on the complete conversation between the three women: Table 8.2: Topic strategies in Wuhan Putonghua Women PG, 35 WG, 32 WN, 30 ALL

N 156 117 127 400

Fronting 8(5.1%) 4 (3.4%) 7 (5.5%) 19 (4.8%)

Repetition 34(21.8%) 36 (30.8%) 34 (26.8%) 104 (26%)

Presentation 114(73.1%) 77 (65.8%) 86 (67.7%) 277 (69.2%)

Tables 8.2 and 8.3 testify to the three women's remarkably uniform speech

230

CREOLE AND DIALECT CONTINUA

behavior. Not only do they produce equivalent amounts of speech (as measured in terms of the number of topic units, as shown in table 8.3), but they also dis­ play the same overwhelming preferential choice for presentative strategies, ranging between 66% and 73%. 2.3.2 The fronting strategy In contrast to presentative strategies, the other two topic marking mechanisms have a very restricted incidence, especially fronting which amounts to an aver­ age of only 4.8%. The left dislocation type of fronting with resumptive pronoun, previously found to be rare in PH, occurs only twice out of nineteen cases of fronting, and a total of 115 topic utterances produced by the three Wuhan speak­ ers. Both instances are represented in text 1 (sample above, as produced by PG): ( 1 ) yige shi, zhuyao shi wo airen, ta gongzuo you mang a, [one TI AP TI my husband, he work AP busy TE, you shi yao shang ke, yao dai haizi [AP TI must teach class, must look after child] (Text 1: 5a-b) 'One problem is that my husband, his work keeps him very busy, he not only must teach classes, but he must also look after the child' (2) Wo haizi, ta bu ting wo ma-de. [My child, he not listen my mother-G] 'My child, he does not listen to my mother' Other types of object fronting are also relatively rare. Fronting occurs only nineteen times in the text out of 400 strategies (Table 8.2). Out of the twentyeight objects occurring in text 1, twenty-two are postverbal, and only six are preverbal, and in most cases (five out of six), those preverbal objects are further introduced by verbal prepositions. The following examples illustrate the Prep+O+V construction in (3) and (4) from text 1, including the bǎ construc­ tion in (4) discussed previously (chapters six and seven), and in (5) extracted from another part of the women's conversation. Note that (3) includes both a fronted complement zai zheli 'at this place', and a postverbal complement yi nian 'one year', both arguments of the same verb, which is not an uncommon occurrence for emphasis in discourse. This confirms that the given/new or def­ inite/indefinite status of an item is not directly related to the placement of a topic in relation to the verb. In (3), the preverbal (with preposition), and postverbal objects are both known and specific.

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA (3)

231

Women sange ren zai zheli gao-le yi nian na, [we three persons at(TI) here arrive-ASP one year EP] 'The three of us arrived here a year ago' (PG; Text 1:2b)

(4) mama yijing ba women dou wang-le [mama AP TI us AP forget-ASP] 'Mother has really forgotten us' (PG; Text 1: 8b) (5) ni jiu bie gen ta shuo-le [you AP not with her speak-A SP] 'Don't explain to her any more' (WG: 52) As indicated before, verbs can also be fronted, as seen in 6 (Text 1:10b-10c): (6) hui-qu yihou ne, wo jiu dao youeryuan qu jie ta le, [return-go after TE, I AP arrive kindergarten go meet him ASP, jie

ta

yihou le, hen duo

xiaohai zai nali, wo jiu shuo:

[meet him after ASP, AP many children TI there, I AP say In addition, the rare (at least in Wuhan) initial O+V construction, that is, not preceded by a preposition or adverbial, is illustrated in (7); here the object shenme shiqing 'everything' —an indefinite quantifier---occurs comment-ini­ tially as part of a long topic chain (the topic airen combined with at least seven comments. It is followed by a pronominal copy functioning as adverbial parti­ cle quan 'all', preceding both the modal yao 'must' and the pronominal agent: (7) shenme shiqing, quan yao ta zuo [everything, all(AP) must he do] 'He must do everything' (PG; Text 1: 5d) As indicated previously, object fronting may occur inside the sentence, comment-internally, following a subject and a particle, as it marks the begin­ ning of the comment. This particle is often an adverbial that has been at least partly bleached from its original semantic gloss, and has acquired a mostly salient value. In (5), both the adverbial jiu and the negative bie strictly mark the

232

CREOLE AND DIALECT CONTINUA

beginning of the comment; in (4) the adverbial yijing 'truly, already' functions as comment juncture preceding the fronted object with ba. In addition, 79% of those fronted items (15) are introduced by verbal prepositions.This structure is interpreted as a type of clefting, and the verbal prepositions (zai, gen, ba) are identified as members of the category of topic introducers. This matches close­ ly PG's sample (text 1) in which five out of six preverbal objects (that is, 83.3%) are introduced by verbal prepositions. 2.3.3 The repetition strategy Repetition has a moderately important function in the Wuhan conversations, averaging 26% (table 8.2). Text 1 has sixteen instances of topic copies. As illus­ trated in the following sentences, a verb is often repeated, with an intervening particle yi to indicate continuity, repetition, or intensity (8): (8) jiu

zou yi zou, wan yi wan, zheyang hao yixie.

[AP walk — walk, play — play, that(AP) good a little]. 'We walked, played, that was pretty good' (PG; Text 1: 3b) Repeated verbs may be associated with different items. Thus, in (9) bao is the first part of a serialized string, then it is combined with a completive aspect marker, both series being redundant. (9) houlai, wo jiu qu ba ta bao-jilai, bao-zhe, [later, I AP go TI him carry-come, carry-ASV] 'Later, I took him home' (PG; Text 1: 13a). There are also in text 1 several repetitions of the most important concepts, such as bu li wo [literally: not acknowledge me] '(he) did not acknowledge me', in which PG refers to her son's resentment toward her because of her long absence (text 1: 11a, l1b, 13 a), or the repetition of xiang in xiangjia, xiang xiaohai 'miss family, miss child' (text 1: 4; 6). 2.3.4 The presentation strategy As indicated in table 8.2, this is the most salient mechanism observed consis­ tently in Wuhan Putonghua, as well as throughout the varieties of Beijing Chinese examined previously. The three women whose speech is represented above consistently use more than two particles per TU:

233

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA Table 8.3: Frequency of topic particles in Wuhan Putonghua Name PG WG WN TOTAL:

T/U 47 31 37 115

Particles 114 77 86 277

# Particles per T/U 2.42 2.48 2.32 2.4

When all particles are classified according to the four categories identified before as topic introducers, topic enders, adverbial particles and end particles, the pre-comment particle category emerges as the most broadly used—63.3% in table 8.4, which represents PG's sample in text 1. This is confirmed by the comparative analysis of topic particle distribution in the complete text (table 8.5), in which the three women exhibit once again similar choices of particles: Table 8.4: Distribution of particles in speaker PG (text 1) N 79

TI 15(19%)

 9(11.4%)

 50(63.3%)

 5(6.3%)

Table 8.5: Relative frequencies of particle-types in non-native PH

PG WG WN TOTAL:

(N) (114) (77) (86) (277)

TI 28 (24.6%) 23 (29.9%) 24 (27.9%) 75(27.1%)

TE 15 (13.2%) 8 (10.4%) 18(20.9%) 41 (14.8%)

AP 61 (53.5%) 43 (55.8%) 42 (48.8%) 146 (52.7%)

EP 10 (8.8%) 3 (3.9%) 2 (2.3%) 15 (5.4%)

All favor APs (pre-comment adverbials mostly) over half of the time, although their overall incidence (52.7%) is somewhat lower for all three women than for PG alone in text 1. Topic introducers rank second with a substantial average of 27.1% (WG producing as much as 30%). TEs are third with a 14.6% average, and finally EPs show minimal incidence, as was the case elsewhere. The frequent pre-comment adverbial markers APs are mostly made up of a few very common items in Wuhan Putonghua. There is one item in particular jiu, and its less common variant jiushi, which is the single most frequent particle in the complete Wuhan corpus analyzed here, and accounts for as much as 45.2% of all APs in the Wuhan samples (see table 8.6). This is much higher than the incidence of this particle in Beijing Putonghua (19.7%) (see table 8.13 below). The large number of APs used by PG (text 1) includes twelve instances of jiu, two of jiushi,3 and in addition five cases of hai (with or without the nega-

CREOLE AND DIALECT CONTINUA

234

tive meiyou ), and seven of dou (meiyou). Taken together, those three items con­ stitute 52% of the fifty APs produced by PG in this sample. The other two speakers of Wuhan Putonghua exhibit an even higher incidence of the specific jiu, item, which accounts for as much as 67.4% of adverbial particles for WG. Table 8.6: Frequency of jiu/jiushi in Wuhan Putonghua (N=total number of APs)

PG WG WN ALL

N 61 43 42 146

Wuhan PH 19(31.1%) 29 (67.4%) 18(42.9%) 66 (45.2%)

This shows a clear preferential trend for a few specific adverbials func­ tioning as pre-comment markers, which contrasts with the diversification observed in native Beijing Putonghua. Although topic introducers are relative­ ly less widely distributed than pre-comment elements in Wuhan, they are more diversified as seen in table 8.7, which compiles the four major classes of TIs identified in previous chapters (deictics, verbal prepositions, adverbials, and the clefting shi particle which has for the first time a significant presence, con­ trasting with its lesser frequency in native Putonghua (only 19.8% on average). Table 8.7: Distribution of topic introducers in non-native Putonghua N PG WG WN ALL

[28] [23] [24] [75]

Deictics: zhè,zhège 5 (17.9%) 4(17.4%) 7 (30.4%) 16(21.3%)

Prep: dùi, bâ, g e i , cóng 4 (14.3%) 2 ( 8.7%) 0 6 ( 8%)

Adv: j i ù s h ì , fnzhèng 9(32.1%) 5(21.7%) 8 (33.3%) 22 (29.3%)

sh ì

10(35.7%) 12 (52.2%) 9 (38%) 31(41.3%)

The relatively important overall use of shi in this conversation (41.3%) requires some discussion, as it may be particularly indicative of the non-native status of the Putonghua spoken by the three Wuhan women. For example, (1) has three consecutive examples of shì which, as observed in the previous chapter, functions here again as an especially highlighting clefting structure. That it is not a simple copula is apparent in its preverbal occurrence (actually preceding a modal+verb combination in 'shi yao shang ke'in (1). In this sen­ tence, PG emphasizes the astonishing fact that in her absence her husband accomplished alone what mothers usually do.

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA (1)

235

yige shi, zhuyao shi wo airen, ta gongzuo you mang a, [one TI AP TI my husband, he work AP busy TE, you shi yao shangke, yao dai haizi [AP TI must teach class, must look after child] (Text 1: 5a-b) 'One problem is that my husband, his work keeps him very busy, he not only must teach classes, but he must also look after the child'

The first sequence yige shi functions as the first "attention-getter" 'there is one (problem)'; the second sequence shi wo airen, redundantly introduced by the adverbial particle zhuyao, clearly identifies the main topic airen (husband); and in the third sequence you shi yao shang ke, the particle shi —again preceded by another adverbial, you, serves to topicalize a verbal unit, emphasizing the husband's workload; it could be translated as a pseudo-cleft 'what he does, (he) must teach classes'. This sentence includes as many as eight topic marking strategies. Beside the rare use of a variant of left dislocation {wo airen, ta gongzuo), and the three instances of the clefting shi discussed above, there are two adverbials as topic introducers {zhuyao and you ), and two instances of repetition of distinct items (you and yao ), an excellent example of redundant salience which highlights the emotional content of the utterance. The absence of true equative copulas noted in Beijing Putonghua (chapters six and seven) is also apparent in Wuhan Putonghua, as shown in (9), display­ ing two cases of zero-copula, before noun, and before adjective: (9) wo shijian shao; wo rongyi-de hen [I time little; I easy-G very] 'It will take me less time; it's very easy for me' [WN: 54] These various examples provide support for the position that I have previ­ ously adopted in the discussion of Beijing PH (chapter seven), namely that the use of shì is distinctively associated with topic marking, thus apparently reflecting the earlier deictic/demonstrative status of that lexical item (Pulleyblank 1995:85-89). The relatively high frequency of the shì particle, and of adverbial particles, all contributing to topic intensification in discourse, are particularly characteristic of the non-native status of Wuhan Putonghua, and may suggest that Wuhan varieties are more conservative than Beijing PH.

CREOLE AND DIALECT CONTINUA

236

2.4 The non-native status of Wuhan Putonghua Interestingly, most of the segments identified by my Beijing consultants in the Wuhan Putonghua sample as "deviant" from native Putonghua involve either the use of the shi particle before verbs, or the use of certain adverbials as APs. Those features appear in (1): the clefting construction with shi three times, and the repetition of the adverbial intensifier y ou. The non-native character of this sentence is apparently represented in the hypercorrect use of the third instance of shi in you shiyao shang he, which—according to my Beijing respondents— would not occur in the PH version. This is the same situation in which sh ì can­ not be considered a true copula because of its preverbal position. In addition, the adverbial comment introducer (AP) yòu would not appear after gōngzuo (that is, inside the comment) in native Putonghua; rather, it would be expected in the last comment but before the modal yào, producing (yòu yào da i háizl). Furthermore, the use of zhuyao as topic introducer is claimed by several native speakers of Beijing Putonghua to be impossible. Thus a native PH counterpart of (1) would read as (PHI): (1) yige shi, zhuyao shi wo airen, ta gongzuo you mang a, (PHl)Yīge shì (--) shi wŏ àirén, ta gongzuò (-) mang a, [one TI,

AP

TI my husband, he work

(lb) you shi yao shang ke, (--) yao dai (PHlb)yòu(_) yào shàng kè, you yào dài

AP busy TE, haizi háizl

[AP TI must teach class, AP must look after child 'One major problem is that my husband has so much to do: he must teach, and look after the child'. Other examples of non-native usage of shì are shown below. They are too numerous to be interpreted as accidental slips of the tongue. They are, there­ fore, likely to fulfill a systematic substratai (or universal) function in the acquired standard produced in Wuhan: (10) wo shi jiang Putonghua [I TI speak Putonghua] 'Indeed I speak Putonghua' [PG: 19]

TOPIC STRATEGIES ÍN VARIETIES OF PUTONGHUA

237

(11) ta shi meiyou yijian-de [they TI not (have) objection--G] 'They (the school) haven't raised any objections' [PG: 27] (12) ta baba shi 'lazybones'. 4 , zhuyao shi you wei bing [his father TI lazy, AP TI has stomach illness] 'his father is lazy, basically, it's (because) he has stomach problems' [WG:36] Other types of non-native situations involve certain adverbials, such as j iù, which is considered by Beijing natives to be awkward in (13), perhaps because the pseudo-copula shì already functions as introducer of the topic "parents," and perhaps also because of the fronted object xiaohai without resumptive pronoun:5 (13) Wo xiaohai jiu shi fumuqin dai (PH13) wo xiaohai (—) shi fumuqin dai [my child AP TI parents

look after]

'My child, it's my parents who look after (him)' (WN:38) The same situation obtains in (14), where the second instance ofjiu (14b) —but not the first, (14a)—is perceived as misplaced, presumably because the focus is on the indirect question, introduced by the question marker shi bu shi, 'is it or isn't?': (14a) Wo shuo: 'mama xingqiwu jiu yao huilai'; suoyi, mei yi tian [I (14)jiu

say: 'mama Friday

AP will return; so,

everyday

wen shi bu shi xingqisi geng xingqiwu a, mama hui

[AP ask TI not TI Thursday or

Friday

bu hui.

TE, mama return not return]

'I told him: 'Mama will return on Friday'; so he asks everyday if it is, or not, Thursday or Friday yet, if mama has returned yet, or not' (WN:39) Many more examples of similar structural differences between native PH and PH as second dialect appear in this corpus, and those differences primarily involve the additional usage of particles, and of cleft constructions with shì, which seems to indicate that the use of shì is much more common in Wuhan

238

CREOLE AND DIALECT CONTINUA

Putonghua in preverbal position, than it is in native counterparts. This assessment of the non-native status of Wuhan PH confirms that this acquired variety can be defined as an acrolect. However, no attempt is made here to provide a comprehensive analysis of other linguistic features of this variety, or to include a comparison with native varieties spoken in Wuhan, which could help evaluate the relative effects of substratai influences on the acquired standard. Nevertheless, some general patterns can be identified, which point to first dialect interference. As indicated in chapter five, the distinctive suprasegmental patterns of native Wuhanhua are represented in the language of Wuhan speakers of PH, entailing numerous vowel changes (this is the reason why tones are omitted in the Wuhan texts). Tonal distinctiveness is sufficient per se to identify the non-native status of Wuhan PH. There are also extensive phonological differences, including the consistent absence of retroflex conso­ nants. It is likely that such suprasegmental and phonological features, as well as morphosyntactic and lexical elements would exhibit some of the procedural mechanisms identified in chapter three (section 3.3.4) for Belizean acrolects, including strategies such as frequency variability, structural hypercorrection, and relexification leading to possible grammaticalization. Structural hypercor­ rection is particularly well illustrated in the morphosyntactic/pragmatic "deviance" noted in (1), with the non-PH use of you, ski, or the additional use of adverbials, such as zhuyao. Matthews' (1996) study of convergence in Chinese dialect grammar also finds that double marking or hypercorrection may lead to a new structure with elements of both varieties in contact. When southern varieties (Cantonese, Chaozhou, and other southern Min dialects) and standard Mandarin are in con­ tact, the result is the coexistence of two alternative forms or ditaxia, a special case of diglossia applied to syntax. The lexical or distributional idiosyncrasies identified in Wuhan PH suggest that substratai influences are operational in areas other than that of discourse top­ ics. What is remarkable is that Wuhan PH reveals a more advanced degree of salience in spontaneous speech than was found in Beijing PH, and a more spe­ cialized use of particles, including the extended use of the pseudo-copula shi. 2.5 Wuhan Putonghua: Conclusions In summary, non-native Putonghua was represented in the speech of one native speaker of the Hunan dialect (PG) and two native speakers of the Wuhan dialect (WG and WN). In spite of their different geographical origins (which involves native dialect differences), those three speakers' varieties are very much alike in terms of their general use of topic strategies, and of topic particles. It is assumed that this relatively small corpus is representative of the speech of speakers of

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

239

Putonghua as a second dialect, who have acquired a higher education level requiring standard fluency; this claim is highly testable. 3. Comparison of Beijing and Wuhan Putonghua 3.1 Topic strategies in Beijing and Wuhan No major difference emerges from a comparison of the overall use of the three topic strategies in, respectively, native (Beijing Putonghua) corpus, as present­ ed in chapter seven, and the non-native (Wuhan Putonghua) sample Both cor­ pora agree in assigning high preference to the presentative strategy, namely 65.5% in native PH, and just slightly more (69.2%) in Wuhan PH, as shown in table 8.9, including literary and spoken PH: Table 8.9: Comparison of Beijing PH and Wuhan PH (Beijing PH includes four corpora-literary and spoken-see chapter seven) N Beijing PH (ALL) [1933] Wuhan PH [400]

Fronting 11.2% (216) 4.8% (19)

Repetition 23.3% (451) 26%.0(104)

Presentation 65.5% (1266) 69.2% ( 277)

The two general corpora examined here are somewhat disparate in size, because the Beijing corpus, including both literary and colloquial Putonghua, is much larger than the Wuhan corpus. Nevertheless, the stylistic level of the Beijing spoken varieties is quite adequately comparable to that of the Wuhan casual/semi-formal corpus. On the other hand, the speech assigned to literary characters reflect the author's personal interpretation of peasant speech6, at least as far as repetition is concerned, and is therefore less directly representa­ tive of real native dialects than the Beijing PH corpus. Therefore, the literary varieties — as reflected in texts from the 1960s and the 1980s—, and the spo­ ken Beijing speech data are separated in table 8.10 in order to test their respec­ tive similarities with the Wuhan corpus. It is clear that the strategies used in Wuhan Putonghua are comparable to both sets of data, especially in terms of the prominent presentative strategy:

CREOLE AND DIALECT CONTINUA

240

Table 8.10: Comparison of (2) Beijing P H and Wuhan P H

Literary PH Spoken PH Wuhan PH

N [1100] [833] [400]

Fronting 13.0% (143) 8.8% (73) 4.8% (19)

Repetition 10.7% (188) 31.6% (263) 26.0% (104)

Presentation 69.9% (769) 59.7% (497) 69.2% (277)

All varieties definitely favor presentation over other devices as the topic mark­ ing mechanism of choice. And the actual overall incidence of presentative par­ ticles also follow the same pattern: between 2 and 3 particles per TU both in Beijing and in Wuhan Putonghua, but with a somewhat lower incidence in the vernacular of television interviews representing Beijing PH (table 8.11): Table 8.11: Overall incidence of particles

Literary PH Spoken PH Wuhan PH

n=particles 769 497 277

n=T/U 274 248 115

n= part. per T/U 2.8 2. 2.4

The fact that there are also similarities between literary Putonghua and Wuhan Putonghua suggests that the author's representation of commune-pro­ duced PH may be quite accurate. Considering that all intellectuals, or privileged classes, including the Wuhan women, now living in the People's Republic of China, were exposed to peasant-influenced PH during compulsory months or years spent working rice-fields in the countryside, the leveling of a worker-ori­ ented common language may have been effectively accomplished. On the other hand, it might have been expected that the three Wuhan teachers would perform more like the speakers interviewed for television, especially the officials' and teachers' group; this is not the case. In fact, Beijing students have the highest incidence of presentative elements (65.3%, as shown in table 7.18 in chapter seven), very comparable indeed to the three Wuhan women's performance. It is perhaps the case that Wuhan Putonghua (as second dialect) standardizes prag­ matic patterns which are considered informal in Beijing, as is also represented in the two authors' rendering of peasant speech. 3.2 Comparison of topic particles The four categories of topic particles are represented in a general comparison of presentatives in the two corpora (table 8.12). As discussed in chapter seven, some differences in particle distribution were observed in the two samples of

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

241

Literary Putonghua, that is, the 1960s story (Two New Friends), and the 1980s rural story (The Matchmaker). Although overall strategies were identical, the 1960s story included more adverbial particles (56.6%; see table 7.4) than the second story written twenty years later (41.9%; see table 7.9). A lower inci­ dence of APs was counterbalanced by a higher proportion of TIs. Again, Wuhan speech is closer in its particle distribution (almost 53% of APs) to some types of Literary PH. In contrast, contemporary spoken Beijing PH has rela­ tively fewer APs (only 32.8% overall), but more TIs: Table 8.12: Presentative particles in PH and WH

Literary PH Spoken PH WH

(N) [769] [497] [277]

TI 34.2% 38.2% 27.1%

  2.5% 53.3% 23.9% 32.8% 14.8% 52.7%

 10.5% 6.2% 5.4%

(chapter 7: table 7.12) (chapter 7: table 7.18) (chapter. 8:table 8.5)

Thus no clear differentiation emerges from the comparison of topic parti­ cle distribution, but there are subtle indications that Wuhan Putonghua and the peasant vernacular represented in literature over 20 to 30 years present simi­ larities, especially in terms of a special predilection for pre-comment particles. Among those, specific particles are favored in Wuhan: the adverbial jiu stands out, as represented in table 8.13, showing striking distributional differ­ ences between Wuhan and Beijing Putonghua (based on the Beijing Saga1): Table 8.13: Frequency ofjiu/jiushi in Wuhan and Beijing and P H (N=total number of APs in each case) N 146

Wuhan PH 66(45.2%)

N 157

Beijing PH 31(19.7%)

Another difference emerges in the choice of topic introducers. It was found in chapter seven that whereas literary varieties favor verbal prepositions, spoken Putonghua prefers the use of deictics. In contrast, Wuhan varieties of Putonghua give special prominence to the topic marker shi, which is used pri­ marily as a clefting device, rather than as a copula, with an overall average of 41.3%), the next category being that of adverbials which amount to 29.3%>; deictic introducers follow closely (21.3%), but verbal prepositions which are highly favored in the short stories examined have a restricted distribution in Wuhan speech (table 8.). In conclusion, a comparison of Wuhan and Beijing PH varieties reveals more similarities than differences in topicalization strate-

242

CREOLE AND DIALECT CONTINUA

gies. Differences occur, but primarily in the detail of lexical idiosyncrasies, and category distribution. 1) Educated speakers of non-native PH use more topic particles overall than a similar group of Beijing speakers. In Wuhan, the three women used an average of more than two particles (2.4) in each utterance but Beijing speakers of a similar education level used less than two on average. 2) In terms of specific particle selection, Wuhan speakers prefer to use adverbial particles. The AP category is over-represented in the Wuhan version: 52.7% of all particles are pre-comment adverbials, as opposed to 30.6% for the officials, and 38.2% overall in native Putonghua. One single particle jiu is espe­ cially common in non-native PH. 3) As far as TIs are concerned, only a moderate amount of deictics (21.3%) is produced in Wuhan (compared to 44.1%, for native PH), but the use of the clefting particle shi prevails. EPs constitute the only group showing the infre­ quent distribution previously observed in native Putonghua. As far as topic mechanisms are concerned, non-native varieties match most closely in appearance the speech produced by the characters in the two vernacular stories. This is not overly surprising, considering certain facts of Chinese history. Those stories were generally commissioned by Maoist directives issued in the 1960s, and the intent was to glorify peasant working classes, and extol a work ethic that would boost production. It is thus natural to find commune leaders producing the official PH with non-native accents. On the other hand, the three Wuhan women grew up during the period of the Cultural Revolution, and were exten­ sively exposed to the overwhelming propaganda of the time. Furthermore, they all spent several years in the countryside for "re-education," and became further attuned to peasant language. They are very much aware of the importance of the popular standard and of the status it holds, even though they are not themselves officials or "cadres"—the highest professional level to be attained in China.8 But even the egalitarian ideology (that led to the ubiquitous use of tongzhi 'comrade') would seem to encourage a leveling of all language varieties as is indeed the intent behind the term Common Language. This could at least partially explain the stan­ dardization of informal topic marking strategies. A cultural explanation partly jus­ tifies the unified pattern of topic marking strategies found in all varieties of Beijing Chinese and Wuhan PH, offsetting the substratai effect of native dialects in acquired dialects. On the other hand, the possibility of universal pragmatic strate­ gies in presenting topics cannot be discounted, especially when considering that there is independent evidence (suprasegmental, phonological, lexical) of first dialect interference. Observations of another second dialect variety will help in this investigation of the universal nature of topic strategies.

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

243

4. Wu Chinese The second variety of non standard Chinese to be examined is a type of Suzhou Wu produced in traditional story telling. It was selected for several reasons. First, it is extremely popular in the province of Jiangsu, and the story-telling genre uses speech forms believed to be representative of earlier forms of Wu; second, it is a vernacular, informal speech form; third, it is not directly related, linguistically, to northeastern Mandarin (Beijing Putonghua), nor to southwestern Mandarin (Wuhan); and finally, it is an acquired dialect. 4.1 The Suzhou oral narrative The Suzhou story is a classical and traditional form of oral literature which is spoken in the Wu dialect used in the Suzhou area, Jiangsu province (south of Shanghai). It is an ancient tradition of story-telling called pingtan which was already well-established as a professional art at the time of the Yuan Dynasty, and was at its peak by the late Ming and early Qing periods in the sixteenth and the seventeenth centuries (Bender 1988:57). The tales can be performed with musical accompaniment, and are present­ ed in serial fashion over weeks or even months, as they recount the exploits of heroes, princes, scholars and beautiful ladies. Episodes are heard regularly on the Suzhou radio station, and are probably as popular, at least among older peo­ ple, as Chinese Opera (such as Beijing Opera and other local forms). The lan­ guage used in these stories is typically identified by Suzhou natives as a formal speech style mostly because it contains many literary words which do not occur in colloquial forms of contemporary Suzhou dialect. Not surprisingly, consider­ ing universal conservative attitudes toward linguistic change, the Suzhou story style is said to be "the real Suzhou dialect", still spoken only by older people, and preserved in the traditional stories. However, what makes those stories par­ ticularly intriguing is that they contain a great deal of style-shifting, at least in the segment studied here. In spite of some formal or literary features (including occasional sentences in Suzhou Putonghua), they are for the most part spoken in a vernacular style, and contain colloquial passages, which create a realistic and often facetious presentation of the characters, in the narration itself, as well as in short dialogues. The sample on which this analysis is based constitutes the twenty-second episode of the tale of Yang Nanwu and Xiao Baicai, told in about two hours by two narrators, one male and one female, and recorded from Suzhou radio in spring 1988. The two narrators' voices alternate, the male voice presenting the hero Yang Nanwu's actions, thoughts, emotions and words, but also referring to other male characters, whereas the female narrator deals with the events and

244

CREOLE AND DIALECT CONTINUA

emotions affecting the female character Xiao Baicai. The latter has a limited participation (as befits a shy heroine in traditional Chinese tales), thus most of the speech is produced by the male narrator (150 out of 188 topic-units), and there is no apparent difference related to sex. The story recounts the misfortunes of Yang Nanwu, scholar and former dignitary who was thrown in prison because of the accusations of Xiao Baicai, the woman who was his former lover. The specific chapter examined here pre­ sents the encounter in Yang Nanwu's cell of the two protagonists, which takes place the night before Yang Nanwu's scheduled execution. The unfortunate scholar attempts to persuade Xiao Baicai to admit that she framed him, which is especially important since he is well aware that high dignitaries are spying on their encounter through a hole in the wall. 4.2 Text 2: Suzhou Wu (narrative style) The following excerpt from chapter 22 of the tale of Yang Nanwu and Xiao Baicai depicts the observations and feelings of one of the high dignitaries, Prince Chen, who has been pushing the other (lesser) dignitary out of the way in order to get a better view of Yang Nanwu through a discrete opening in the wall of his cell. The story humorously dwells on the spying mandarin's physical discomfort in his cramped position, and on the lesser dignitary's resentment at being nudged away. An interesting discourse feature represented here as well as throughout the tale is the narrator's use of several omniscient voices, switching from observations of each dignitary's physical behavior, to revelations of their inner feelings, as well as to the main character Yang Nanwu's inner thoughts. This text also exhibits linguistic characteristics of Chinese oral discourse observed in the context of other samples, in particular revealing the existence of stacked predicates relating to one single topic. For example, the first topic utterance (1) consists of one topic—the prince's reactions after watching through the hole, and nine subsequent comments (la-li). Some of the most obvious characteristics of Suzhou Wu (henceforth SU) are suprasegmental and phonological. The tone system in SU is fundamentally different from the native PH system (see chapter five). However, since the focus of this study is on morphosyntactic features, tonal details are not included in the following transcrip­ tions. An approximate transliteration of the Suzhou phonetic representation has been attempted in pinyin, an orthographic representation of Chinese loosely based on Wu pronunciation, although normally associated to PH. Each sentence is accompanied by its PH version, when possible (when no direct translation is available, the approximation is parenthesized, as well as by a literal English translation. A regular English translation following the text. Many lexical and morphological items found in the Suzhou narrative do

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

245

not have exact counterparts in PH, or are assigned different semantic or stylis­ tic values, some of them to be discussed below: (la) Ge PH (zhègè

Yang Nanwu jinze Yang Nanvu j ī n t i a n

gongfu gōngfu

[This(TI) Yang Nanwu today (lb) wang ya PH (váng yé

skill

se jiù AP

gang he, zheyang hao) AP

good

sezei qiy fe sie, s h í z à i chî bu x i ǎ o )

[prince master AP (lc)genji PH(èrqiě)

eat

wang ya vàng yé

not digest neisi biyjia, nàixîn b i i j i a o )

[besides(TI) prince master patience poor (ld) nin me li zi sang yue, PH (rén m e ( z h a n ) - l e chang yuan) [body TE stand-ASP long (Ie) dongdong me [hole hole TE ye yào

9

leide e

P H (dòngdòng me ( f ē i c h a n g )

(If) ge PH (zhègè

time

(bù shūfū)

AP

awkward

me wei qi zi, me vān__zhe)

[this (TI) waist TE bend—ASP (lg) dei sen de, PH (tóu shun dao) [head hang down, (lh) sue

z

sue

de lei,

P H (sāan

shì

sūan

)

[stiff

TI

stiff

EP]

yibi te tài

feel AP

di, dî)

low

246

CREOLE AND DIALECT CONTINUA

(li) a sei fe si ze PH (yě (rěn) bù zhù le) [AP endure not stay EP] (2a) dei wei gulei dei Wang Xin mangmang: PH (tóu hui guòlái dùi wcng xīn wangwang) [Head turn back

TI Wang Xin look look]

(2b)nei jinze siang kue, PH(nï j ī n t ï a n xiǎng kan, [you today

want

nei kue nǐ kan

fa, gou gue ba, vǒ gān

nue niang nei. yuan rang nï

watch, you watch TE, I willing let

(3a) bilang Wang Xin a xiang fe lo: PH(yìpang) wáng xín (mèihuà kě shōu) [at side, Wang Xin AP make sound not say:] (3b) "yizei me niang gou, gou zen e a fe ye kue". wǒ, wǒ bù yào kàn) PH (xìanzài me rang ["now

TE let

me, I

AP AP not want watch]

(4a) name wang ya ciy sei yiyqie yiyqie qie PH (name vang yéqian suì yíqiao yíqiao qiào [TI

prince 1000

years limp

(4b) de

le

lang

(PH) dào.

lâo wèizhi shàng zuo

weizi

arrive old seat (5a) "Ai" yi PH (ai

la

tei

you zài tan

on

sou din. )

sit down]

qi qì

ze le)

["Ah" again ASP exhale breath EP. (5b) gerne Yang Nanwu nahang? PH(zhege Yang Nánwu zenyàng ) [TI

limp

Yang Nanwu what?]

limp

you]

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA (5c) siang siang ndo jinze fe ye tei PH ( nïmén j ī n t i ā n bù yào tàn [ TI

TI

you

today

247

qi. qi)

not must exhale breath]

(5d)"ηou Yang Nanwu li wo lei" PH (wo Yáng Nanwu zì

si yu beifa." yǒu banfa)

["I Yang Nanwu link-down-come self have method", (5e) yinci PH (yîncï

siang de geilang Yang Nanwu dang din zi ji, xiang dào zhèshàng Yang Nanwu dǎ dìng zhu j iàn)

[consequently(TI) think

AP that

(5f) jinze yiy fangmi P H ( j ī n t ï a n yì fangmiàn [today

qu quàn

Yang Nanwu make up own mind li qiy jiu ta (hē) j i u )

one direction-side persuade her drink wine

(5g) lin yiy fangmi PH (lìng yì fāngmiān

ye yaò

keisi dei quezi * ze, kāishï dōu quānzǐ le)

[other one direction-side must begin

go around

EP

(5h) yindang zenben keisang men li zenzin: "sue mei. . ." PH (yïndāng zhèngběn kāishï wen tā zhënqíng:"shuān mèi..." [should seriously begin ask her truth: "Shuan sister."] (Chapter 22:108-110) (1a) The skill Yang Nanwu displayed today was so great that (1b) the prince was overwhelmed (1c) Furthermore, the prince had little patience (1d) He had been standing for so long (le) bending in front of this /(1f) awkwardly low hole, (1g) his head inclined (1h) that he was getting very stiff, (1i) and he could not stand it any more (2a) Turning his head toward Wang Xin, he said: (2b) "You wanted to watch today, so watch! I am willing to let you do that" (3a) Wang Xin did not know what to say: (3b) "You finally let me watch now, but I don't really want to any more". (4a) The noble prince (of thousand years) limped(and limped and limped)

248

CREOLE AND DIALECT CONTINUA

(4b) back to his previous seat. (5a) [Prince sighing]: "Ah". [Narrator, talking about the prince]: There he is, sighing again! (5b) Now, what is Yang Nanwu up to? (5c) [Talking of Yang Nanwu's thoughts] You should not sigh today. (5d) [Yang Nanwu thinking: "I, Yang Nanwu know what to do next" (5e) His reflections had led him to decide this: (5f) today, he would first make her drink; (5g) and then, he would skillfully persuade her (5h) to tell the truth: "Shuan sister. . . ." 4.3 Topic strategies in Suzhou Wu SU exhibits the same three-way system of topic mechanisms that was found in other varieties: a limited amount of topic fronting, repetition, and the use of presentative elements, including a type of clefting. The distribution of the 40 topic marking strategies exhibited in text 2 (comprising a total of five topic units) is rep­ resented in table 8.14. Once again the presentative strategy is clearly dominant: Table 8.14: Topic strategies in Suzhou Wu (text 2) (N^total number of topic strategies) N [40]

Fronting 10% (4)

Repetition 13% (5)

Presentation 77.5% (31)

It was observed before that short samples (that is, the two literary Putonghua excerpts, as well as excerpts from Beijing interviews and Wuhan conversation) are representative of the complete texts. This is also the case in the story of Yang Nanwu, which includes 188 topic units altogether. Table 8.15 presents the results of the analysis of topic strategies in the story, showing sep­ arately the male narrator's and the female narrator's productions. As indicated before, there is a discrepancy between the amount of speech produced by the two narrators: the female character—Xiao Baicai—has a subdued role in this chapter, and only occasionally responds (mostly through the indirect musings of the female narrator) to the male character—Yan Nanwu—now in jail and await­ ing execution because he was betrayed by his love. It is clear, however, that the small female sample matches very closely the more substantial male sample:

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

249

Table 8.15: Topic strategies in the Wu story

Male narrator Female narrator Both

N [679] [154] [833]

Fronting 55(8.1%) 16 (10.4%) 71(8.5%)

Repetition 184(27.1%) 35 (22.7%) 219(26.3%)

Presentation 440(64.8%) 103 (66.9%) 543(65.2%)

In both, presentative strategies consistently rank within the 60% to 70% range, whereas topic repetition ranks moderately high (around 26%), and fronting has a minimal incidence (8.5%). Those two minor strategies are briefly illustrated below, then the dominant presentative mechanism is discussed. 4.3.1 Fronting The fronting of complements, either to the beginning of a topic unit, or internally, is rare in Suzhou Wu, as it is elsewhere, but when it happens, it often co-occurs with topic repetition, typically combining preverbal and postverbal complements, as illustrated in (15) in which objects are italicized: a postverbal structure occurs in (15a) following the verb mang 'forget', whereas the preverbal (fronting) strategy appears in (15b), before the same verb fronting similar types of objects (referring to previously felt emotions such as affection, anger, and others), thus indicating that colloquial Suzhou Wu, like other forms of Chinese, does not, at least in this case, take into account general/specific or known/new information. (15a) jinze be mang siy (PH)[jïnzhāo bù wàng qián (today

siyn, qíng ];

not forget previous affection)

(15b)rjousei ni lixiang-ne hen yu (PH) [wǒ soin n ián l i - d e hèn yuan (I three years — G

qi qì

siybu cei mangji te quánbù wàngjì dào]

hatred, complaint, anger, all

forget-arrive)

'Today (I) still remember my previous affection, and I have forgotten all the hatred, complaint, and anger of the last three years' (Ch 22:11) Finally, left dislocation is even scarcer in the Suzhou text than in native Putonghua, or in Wuhan varieties. Only two cases were found out of 71 cases of fronting (less than 3% for 188 topic units), as in (16):

CREOLE AND DIALECT CONTINUA

250

(16) Ge Sen cinwang li le kue ya (PH) [z.hè chén qīnwang, tā zài kàm a] (TI Chen Prince,

he ASP see TE)

'Prince Cheng, he watched.' (Ch.2: 36). In most cases, even with a long initial topic, there is no resumptive pronoun, as represented in (17): (17) name wang ya ciy

sei yiyqie yiyqie qie

(PH) (name vang yeqiän suì yíqiao yíqiao qiào

[TI de

prince 1000 years limp limp le weizi lang sou din.

(PH) dào läo wèizhi shàng

TI

old seat

on

zuò

limp

)

sit down]

'The noble prince (of thousand years) limped painfully back to his previous seat' (Text 2:4) 4.3.2 Repetition Onomatopoetic repetition or reduplication is a prominent feature of the SU oral narrative which is more widespread in this variety than in PH. A duplicated ele­ ment adds emphasis to the text, warmth, and often humor to the rapport between the narrators and their audience. Several examples occur in text 2, (e.g. mangmang 'looked intently' (2a); yiyqie yiyqie qie 'limped (very stiffly), as shown in (17) above. The following sentence—the complete version of (16) shown above—includes several repeated topics, especially in reference to the shed­ ding/dripping of tears {dada di and do ŋeili a di celei): (16a) Ge Sen cinwang, li le kue ya,  -de PH ( chén qînvàng, tā zài kàn a, kàn-dào [TI Chen prince 16b)

gabiy gébì

he ASP see TE, see-arrive next door two persons]

kuo de lei rjeili dada di, li de mei sue Hang, PH [kū-de yànlèi dî, tā dào man shàn Hang [cry-G AP tear

lianggadei liànggèrén)

drip drop, he AP AP kind

good,

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA 16c)

251

wei donghuo zin qi, longde e wangya PH(huì tónghùa j ìn qù, shï'dé zhè wángyé [ AP same

enter go, make-G TI

prince ]

16d) ge Hang do qeili a di celei, ŋe PH (zhè Hang dî yanlèi ye dî chū-lai, áo [TI

two

drop tears

fe le ze. bù liǎo le)

AP drop come out, suffer not end

EP]

'When Prince Chen saw those two persons shedding tears next door, and because he had a kind and generous heart, he could not contain his tears'. (SU 22: 36) Similarly, the four-way reduplication yeye hehe nienie lolo of yehe nielo 'be close friends' in (18) evokes a warm friendship which stands in ironical contrast to the lovers' anticipated demise due to one lover's betrayal of the other. Both expressions have no exact counterparts in Putonghua, as indicated in the cognate approximations, and furthermore, the English translation is unable to capture the humorous or rhetorical effects of the repetitive strategy. Note also the repetition of sei 'hand' in a serialized verbal structure: (18a) Menze PH(Mingtian

ni yeye hehe women yàohao

[To-morrow we (18b) sei cei PH shou chān

nienie lolo rènao

be close friends hot and jolly

sei de shou dào

yingei yīnj iān)

[ hand support hand arrive netherworld] 'Tomorrow we should be very close friends and hold hands to enter the nether world' (SU 22: 22) Topic repetition is also evident in Suzhou Wu classifiers (measure words) which in Chinese grammar are usually systematic markers of certain semantic properties of a noun (i.e., length, bulk, height). In the following text, they func­ tion as reduplicated intensifiers, for examples zangzangjiji, buebue,yangyang, to indicate excellence, a device not much used in PH. The semantic link is that the adjective applies in every dimension—flat, long, thick, etc.—a notion which has no direct counterpart in English:

CREOLE AND DIALECT CONTINUA

252

(19a) Be PH (Bié

ninga

di-de,

li la zonggui

rénjid

tí-dào,

t a ya z h ō g g ū i

[Other people

mention-arrive she TE always

se shūo

say

19b) si yangyang he, buebue he, jiji he, zangzang he, sese he, PH (ta (yàng) hǎo, (bîan) hǎo, (jǐ ) nao, (zhāng ) hdo, (shìshì)hao [she very good, 19c) jise

very good,

very good, very good, everything good,

mbe yiy yang fe he.

PH(shénmì

döu hao)

[simply not one kind not good] 'All are unanimous in saying that she is perfect (good in every possible way') (SU 22: 67) Finally, a recurrent expression in chapter 22 involves the presentative phrase se gang zong {cue jin/yangzi) 'such type of a miserable situation/atti­ tude', which is usually placed at the end of the utterance, thus often functioning like a variant of right dislocation, as in (19) or (20). It is also reminiscent of phrases such as nèi yang, or n èi diǎn shìr 'those kinds of things' used in  similar manner in Beijing PH (see chapter six: text 3: 4; 6b), and would have counterparts in English, such as whatever, this kind ofthing, stuff like that. (20a) Sei ni du lixiang hei-de Yang Nanwu, PH (san nián dūo lǐ hài-de Yang Nanvǔ, [Three years more — 20b) se

gang

PH j i ù

zhèyàng zhōng cāo

[AP this (TI)

zong

suffer-G Yang Nanwu,

kind

cue

jin,

yinci

jì ng, yîncï

mbe mikong ji

li.

méi

tā)

miserable state, so(AP) not

(liân)

j iàn

face

see him]

"In the past three years and more, she had caused so much suffering to Yang Nanwu that she was ashamed to face him" (SU 22: 11)

4.3.3 Presentative strategy: Topic particles Considering the privileged status of presentation, it is no surprise to find an average of almost three particles per utterance, the same frequency for both nar­ rators, despite discrepancies in sample size (table 8.16). Chapter 22 contains an

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

253

overall number of 457 particles for 188 topic-units, when combining the contributions of both the male and the female narrators: Table 8.16: Frequency of topic particles in Suzhou Wu TU 150 38 188

Male narrator Female narrator Total: CH. 22

Particles 440 103 543

Percentage 2.9 2.8 2.9

As attested by tables 8.17 (text 2) and 8.18 (complete chapter), the Suzhou narrative style has a predilection for adverbial particles, that amount to as much as 43.8%) of all particles in the whole story, but there is an almost equally high percentage of topic introducers (38.9%)), which is in part due to the common use of clefting with z/si (see below). Table 8.17: Topic particles in Suzhou Wu (text 2) (n=number of presentative particles) n [31]

TI 38.7% (12)

 16.1% (5)

 32.3% (10)

 12.9% (4)

Table 8.18: Topic particles in Wu

Male Female Both

n [440] [103] [543]

TI 169(38.4%) 42(40.8%) 211(38.9%)

 59(13.4%) 11(10.7%) 70(12.9%)

 188(42.7%) 50(48.5%) 238(43.8%)

 29(6.6%) 0(0.0%) 29(5.3%)

Although pre-comment particles (APs) play an important role in the topic unit, there is no evidence that any one single morpheme has a predominant dis­ tribution, as is the case for jiu in the Wuhan sample examined above. Suzhou Wu utilizes the j iù reflex se (see text 2: la) 43 times out of 238 APs (18.1%), thus much less than the 4 5 % found in Wuhan for the same adverbial (table 8.6). In contrast to the selective use of a few APs in Wuhan PH, the variety of adver­ bial particles is overwhelming in Suzhou: over forty different particles occur before verbs or adjectives. In addition to se, one finds a/ya/ye (=PH yě) 14.7%, wei (=PH hái)—7.9%, and da/de (=PH dào, but rare as AP in Putonghua 1 0 )—6.3%. In this sense, the variety of AP particles in SU is more like that found in native Putonghua than in Wuhan Putonghua. This particle diversity is represented in all four particle categories, as is

254

CREOLE AND DIALECT CONTINUA

briefly illustrated in table 8.19. The varieties represented in the SU oral narra­ tive include a combination of literary (or conservative) items, and others which are currently used in colloquial forms of Suzhou Wu. 1. Literary items: Some particles found in Suzhou pingtan tales are no longer in use in contemporary Suzhou varieties, such as  'consequently' (PH b iàn); the variants ya/ye 'still, also' (PH ye), now replaced in Modern Suzhou Wu by a, as illustrated in Text 2: 1 i (16d); genji 'besides' (PH érqi e) in text 2:1 c; or da/de; (PH dào) which is used in conservative Suzhou Wu to indicate surprise in predicate-ini­ tial position, a usage atypical in PH, shown in text 2: 5e. (5e) yinci siang de geilang Yang Nanwu dang din ziji consequently(TI) think to(AP) this YN make up own mind 'These thoughts led YN to decide this' The more regular usage of de is represented in text 2:lg, or in (16a): (lg) dei sen de head hang-arrive (serial verb sequence) '(His) head hung down' (16a) kue de gabiy lianggadei see-arrive next door two people '(The prince) saw next door two peop]e(16a) 2. Colloquial items: Other morphemes are commonly in use in vernacular varieties of Suzhou Wu. Most common are probably the deictics e, ge, gerne, gezi, name (la; If; 4a; 5b) either before a noun or an adverb; and adverbial mitigators which also function as topic introducers (i.e., siang reduplicated in 5c), which has no PH reflex. Prepositional/verbal morphemes include the object introducer nuo (PH bǎ. ) and dei ( PH dùi ). There is a wide range of a SU adver­ bial predicates which have no lexically related equivalents in PH. Some com­ mon pre-adjectival particles functioning as intensifiers include leide. . .e (text 2,le), z. . .de lei (text 2, lh); other predicate particles occur before verbs, such as a (see text 2: li; 3a), or se (text 2,1a), to mention just a few. Those Suzhou items do not occur in Beijing PH or Wuhan PH, and are generally difficult to translate, though they are pragmatically equivalent to English discourse mark­ ers such as as far as (X) is concerned/regarding/kinda/like or even the increas­ ing usage of the suffix -wise in American English (meteorological forecasts con-

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

255

stantly include temperaturewise; weatherwise). All such particles function as intensifiers: they are almost completely bleached of their original semantic value, although their emphatic value is probably due to their original meaning (i.e.sezei 'really'). Table 8.19: Particle reflexes in Putonghua (PH) and Suzhou Wu (SU) [C=colloquial; L=conservative or literary] TOPIC INTRODUCERS

SU:

genji(L)

e/ge(C) geme(C)

PH:

érqie

zhègè

SU: de i (C) PH: du i

he(C) hǎo

gezi(C)

name siang(C)

segang(C) nuo(C)

te(C)

zhèyàng

gēng

bǒ

si/z shî

TOPIC ENDERS

SU: ba(L) fa/va/ve(C) va(C) la(C) ya(C) PH: ba ba a a a

na(C) na

niang(C) nie(C) me(C) ne ne me

ADVERBIAL PARTICLES

SU: da/de(L) ya/ye(L) bi(L) PH: dào ye biàn SU: PH:

zen(zen)(C) zhēn

mei(C) yu(C) man you

leide e(C) a(C) ye

sezei(C) shizài (quèshi)

sei/seibu(C) se/sezi(C) döu/quān(bu) j i ù / j i ù s h i

te(C) tài vei(C) hái

END PARTICLES

SU: ne/ni(C) la/ya,a(C) PH: ne na/ya/a

niang(C) ne

ze(C) le

me(C). me

The relatively high incidence of Tis shown in table 8.18 (38.9%) is partly due to the common use of the two SU reflexes of shi (si and z), which function as a clefting structure. This morpheme is the most common topic introducer (40.8%), as represented in table 8.20 which displays the use of topic introduc­ ers by both narrators in the Suzhou story. The use of this particle and its variant appears twice in (20a), where both si and z occur on the same line. There is no apparent constraint on the use of either variant (the variant z occurs slightly more often than si):

CREOLE AND DIALECT CONTINUA

256

(20a) Ge jinze si nei fe he, dedi nei z hei ninzin, PH (zhègè jīnzhāo shí nǐ bù hǎo, dàodí nǐ shí hài rěnjīng [T]

today

AP you not good, AP

you AP betray-person-spirit]

(20b) heide Jou Yang Nanwu, se gang zong yangzi PH (hài-de wǎ Yáng Nanwu, jiù zhèyang zhōng yàngzi [frame-G me Yang Nanwu, AP this

kind of situation]

"Today, (it's) you (who) are to blame, after all, (it's) you (who) are the trai­ tor, (who) framed me Yang Nanwu (and put me in) this miserable situation" (SU 22: 43). Obviously, this clefting particle represented by the two variants si/z can be placed either before the topic si nei or after the topic to be given salience nei z, with the same value 'it's you', and this double strategy is reminiscent of the functioning of the equivalent cognate particle shi in Wuhan and Beijing, as well as its counterpart da in Belizean Creole (see chapter nine, section 4.2). Table 8.20: Topic introducers in Suzhou Wu

Male narrator Female narrator BOTH

N [169] [42] [211]

Deictics 56(33.1%) 15(35.7%) 71(33.6%)

Verb/Prep 12(7.1%) 4(9.5%) 16(7.6%)

Adverbs 32(18.9%) 6(14.3%) 38(18%)

Shi 69(40.8%) 17(40.5%) 86(40.8%)

As far as the other introducers are concerned, deictics are quite frequent as well (33.6% overall), whereas adverbial introducers are limited. Finally, verbal prepositional TIs, such as nuo, te, dei, have a relatively minor function in topic presentation, as compared to the other categories, contrary to PH. 4.4 Stylistic shifts A particularly interesting aspect of SU narrative lies in its unique mix of formal and informal forms, literary and colloquial items, involving lexical morphemes as well as particles, which provides insight into the stylistic constraints on lex­ ical choice. Lexically related items are associated with different stylistic levels (ranging from colloquial to literary) in SU and PH, as illustrated in table 8.21. Typically, some unmarked PH variants have formal counterparts in SU, and vice-versa, informal SU variants may have formal reflexes in PH:

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

257

Table 8.21: Stylistic variations in lexical items (SU=Suzhou Wu; PH=Putonghua) [< > indicates that the item ranges over formal and literary styles ] Informal mang kàm

Formal Literary kue <wàng, d ī ng>

Gloss 'see, watch'

mikong 1 idn

lian miàm (kŏng)

'face'

PH SU PH

minze; jinze minti; jinti 'tomorrow; today' m í n g t i ā n ; j í n t i ō n <mingzhāo;j īnzhāo>

SU PH

SU

SU PH

Nonstandard zang, za qiáo

qiy hē

ha yĭn

yong yòng

'drink'

Table 8.21 shows for example that kue 'see, watch' is very formal in SU (2b in text 2) and typically used by high dignitaries, whereas the corresponding cognate PH item kán 'see' is unmarked. SU also has a colloquial counterpart mang (reduplicated in 2a), which appears when the narrator jocularly describes some of the characters' frustrations. The PH reflex vàng has a formal or poetic connotation ('gaze'). Similarly SU mikong 'face' is stylistically equivalent to the colloquial use of PH 1 i an, whereas the etymologically related PH item miànkŏng is very formal. This crossover pattern is extremely common. There are, however, some SU items which do not have clear equivalents in PH. For example, some SU vernacular forms rarely used in formal style, such as  'awkward' (le), or liwolei 'to arrive (at a decision)' (5d) have no appropriate PH counterpart, according to my Chinese sources. There is ample use of colloquial metaphors idiosyncratic to Suzhou Wu, such as qiy fe sie (lb), literally 'eat not digest', meaning 'to be overwhelmed', or xiang fe lo (3a), literally 'make sound not say', meaning 'did not know what to say'. In addition to those stylistic discrepancies, there are interesting semantic shifts, which obviously indicate an early relationship between SU and PH. For example, whereas PH differentiates between chl 'eat' and he 'drink', SU uses qiy, a reflex of chī for both 'eat' and 'drink', and with identical tones ( l b ; 5f), and uses ha for 'drink' only in very formal contexts (table 8.22). On the other hand, Suzhou Wu differentiates distance in time and place, in contrast to PH (or English "far"), indicating that both varieties, though related, developed in dif­ ferent directions.

258

CREOLE AND DIALECT CONTINUA Table 8.22: Semantic differentiation in Suzhou Wu and Putonghua

Gloss SU PH

'stand';'stand up' li li zhàn lí

'eat';'drink' qiy qiy chī hē

'far'(time);'far'(place) yue yue jiu yǔan

4.5 Suzhou Wu: Conclusions This brief analysis of the dialect used in Suzhou oral narratives, as represented in one radio-broadcast chapter of the tale of Yang Nanwu and Xiao Baicai shows that the morphosyntactic correlates of topic prominence occur also in varieties not directly related to Northern varieties of Mandarin, though also acquired as second dialects. In view of the objective pursued here—namely, to identify the nature of the linguistic features attributable to second dialect acquisition or, as it has become increasingly obvious, to eliminate those features which cannot be attributed to second dialect acquisition—it is of significance to observe that the narrative variety of Suzhou Wu surveyed in this chapter displays high topic prominence. This is identified primarily through the structure of the tales and the incidence of topic strategies. In those tales, Suzhou Wu has a unique value in that it is simultaneously formal/literary and colloquial/popular. The language used in those radio-broadcast tales represents in a sense a highly stylized, and some­ what literary and conservative variety because it describes high dignitaries, princes and mandarins in the tradition of Chinese opera. Yet, on the other hand, it shows the human foibles of those privileged characters and is intended to serve as a popular form of entertainment for mass consumption. Consequently, there is constant shifting between formal and colloquial variants depending on the context and the characters portrayed. It is thus remarkable to find that, regardless of the stylistic level, topic strategies exhibit regular patterns, as observed above. This is clearly represented in the excerpt selected because of its vernacular, often jocular stylistic level, and in the distribution of topic strate­ gies in the complete chapter. The relevant conclusion is that topic prominence is a universal feature of all Chinese varieties, both literary and spoken, and, therefore, one that cuts across native and non-native varieties.

5. Comparison of native and non-native varieties of Chinese When comparing the Suzhou Wu narrative style, the acquired Putonghua of Wuhan, and varieties of native (Beijing) Putonghua, it is clear that all three var-

, TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

259

ieties highly favor topic strategies in a very similar manner, and more particu­ larly the use of topic particles which have been specifically investigated and quantified in the three corpora. Table 8.23 compares the overall use of topic mechanisms in all three types of varieties: Table 8.23: Overall topic strategies

Beijing PH Wuhan PH SuzhouWU

N [1933] [400] [833]

Fronting 11.2% 4.8% 8.5%

Repetition 23.3% 26% 26.3%

Presentation 65.5% (Table 9) 69.2% (Table 9) 65.2% (Table 8.15)

Although differences related to second dialect acquisition clearly exist at the suprasegmental, phonological, morphosyntactic, and lexical levels, there is a remarkable homogeneity of topic mechanisms observed throughout all vari­ eties of Chinese, across the diachronic, synchronic, and stylistic dimensions, as well as across small or large samples. The syntactic/pragmatic level, however appears to function uniformly. This supports the universal hypothesis that dis­ course requires special structures and morphemes to mark syntactic sequences of communicative units. Topic particles range between two and three per TU, and occur at specific locations in discourse. Table 8.24 presents the general dis­ tribution of the four categories of presentative morphemes investigated throughout this analysis of topicalization: Table 8.24: Particle-types in three varieties of Chinese

Beijing: Wuhan: Suzhou:

(N) [1266] [277] [543]

TI 35.8% 27.1.% 38.9%

 10.6% 14.8% 12.9%

 44.8% 52.7% 43.8%

 8.8% 5.4% 6.6%

There is also a striking homogeneity in particle use: all three varieties use predominantly the pre-comment particles (APs), with Wuhan Putonghua lead­ ing in this respect (52.7%). The TI category follows closely overall, though the gap is wider in Wuhan again, which complements its higher use of APs with a lesser frequency of TIs. The other categories of particles do not pattern differ­ entially in any significant way. However, the internal composition of each par­ ticle category may present certain differences, beside the actual lexical idiosyn­ crasies which are most obvious in Suzhou Wu, as discussed in the previous sec­ tion. The most obvious differences in this respect involve:

CREOLE AND DIALECT CONTINUA

260

1) the specialized use of a few specific adverbial particles in Wuhan PH (jiu and jiushi in particular) as pre-comment particles, whereas other varieties, both native Putonghua, and especially Suzhou Wu, display much more diversi­ fication in the use of particles, and 2) different priorities in the choice of topic introducers from the four cate­ gories identified throughout this study. It is clear from table 8.25 that whereas native Beijing Putonghua as a whole assigns equivalent importance in the 30% range to deictics (zhe, zhege, zheme, zheyang), and to verbal prepositions (ba, dui, gen, bei), both non-native vari­ eties of Putonghua and Suzhou Wu show a clear preference for the use of the clefting particle shi, which amounts to about 40% of all topic introducers in each case. All three varieties, however, widely use deictics, which is obviously a frequent feature of vernacular discourse. The consistent frequencies of presentative deictics (ge, gerne, gang in Suzhou for example) reflect the sponta­ neous level of most varieties investigated here, as well as the importance that Chinese assigns to the popular language, as reflected in the two short stories analyzed here: Table 8.25: Topic introducers in Chinese ("All native PH"=combined varieties from chapter seven)

All native PH Wuhan PH Suzhou WU

N [449] [75] [211]

Deictics 35.6% 21.3% 33.6%

V/Prep 31.8% 8.0% 7.6%

Adverbs 12.7% 29.3% 18.0%

shi 19.8% 41.3% 40.8%

However, the combined frequencies of Beijing Putonghua varieties obscures the fact (pointed out in chapter seven) that a striking difference between Literary Putonghua (reported 'peasant Putonghua') and direct record­ ed Spoken Putonghua is represented in the use of verbal prepositions, which is high in literary PH (40.7%), but less valued in spoken PH (19.4%), and even less so in other varieties: only eight percent in Wuhan PH and 7.6 percent in the Pingtan style of Suzhou Wu 6. Conclusions: Toward a universal view of topic processes I have thus far evaluated the topic patterns represented in Belizean Creole (chapters two to four), and in various types of Putonghua (chapters six to eight). The Chinese varieties analyzed in the latter chapters included literary interpre-

TOPIC STRATEGIES IN VARIETIES OF PUTONGHUA

261

tations of the vernacular in the context of a political system which assigned sta­ tus to lower class varieties; a longitudinal perspective on two types of literary Putonghua spoken over twenty to thirty years; spoken Beijing Putonghua in the 1980s (a native variety); spoken Wuhan Putonghua in the 1980s (an acquired variety); and the pingtan style of Suzhou Wu. In spite of some minor differences in particle choice, topic strategies are remarkably homogeneous across all dimensions investigated, whether in terms of diachronic change, stylistic variation, or acquisitional types. Although the observations conducted and discussed here are limited in their scope and time range, I believe that they are extensive enough to suggest that spoken varieties of native and non-native dialects use similar discourse strategies, and that such pragmatic consistency may be due to universal patterns in the use of topic mechanisms, a theory which is substantiated when Chinese patterns are com­ pared to Creole patterns (see chapter nine). The intensive use of presentative morphemes noted above applies in a remarkably consistent manner to all varieties examined, whether native (Beijing informal Putonghua) or acquired as second dialects (Wuhan Putonghua and Suzhou Wu), as well as in creole acrolects. The only potential for a differential occurs within the prominent presentative strategy. There are specific differences related to the choice of certain particles, such as the preferential use of a few par­ ticles, versus the diversified use of a broad set of elements fulfilling similar func­ tions in certain pivotal articulations of the topic utterance. However, it is not clear that those particle choices separate native and non-native varieties of Chinese. For example, although it is tempting to conclude from the observed frequency of the clefting particle shi in Wuhan and Suzhou that it is a characteristic of non-native patterns, it must be remembered that the group of eight officials commenting on the value of the science program for middle school students all consistently use that particle shi 38.8% of the time (table 7.21 in chapter seven), thus differing from students, parents and teachers. More data are needed to assess the putative effect of stylistic context on particle choice, but there is no doubt whatsoever that topic prominence is an essential part of discourse processes in all varieties of Chinese, and that this type of information presentation is primarily dependent on the intensive use of particles in a variety of positions at structural boundaries. Important boundaries are obviously located at the beginning of comments, or before topics anywhere in the utterance, whether at the very beginning of the utterance, or after a sub­ ject which is not the main topic, or even within comments when another topic is introduced. Particles clearly contribute to the grammaticalization of structur­ al units, and appear to fulfill a universal function essential to information pro­ cessing that is not dependent on the type of learning.

262

CREOLE AND DIALECT CONTINUA Endnotes

1. There are mutually unintelligible dialects or languages, even in the close proximity of Wuhan and the neighboring state of Hunan (the Shangsha area is well-known for a distinctive language Xiang) 2. They were also particularly aware of dialectal variation in Chinese, since they had been assigned various sociolinguistic research projects involving local variation. 3. Although jiushi is normally an AP, two instances in this text qualify as topic introducers (13; 10), because they provide the setting for the whole proposition, not just the topic. 4. This borrowing is not typical of Wuhan Chinese. It is due to the fact that the two women are English teachers. 5. My consultants did not identify shi as deviant here, perhaps because it does not immediate­ ly precede the verb dai. 6. As indicated in chapter seven, there is no observable difference between the two short sto­ ries analyzed. 7. Cantonese also makes extensive use of jauh, a cognate of jiu, as attested by Matthews and Yip (p.c. 1996) (see also Matthews & Yip, forthcoming). 8. However, PG, as party representative, probably holds a fairly high status relative to her classmates and is most likely to become a cadre. 9. According to my consultants,  and liwolei are Wu idiomatic expressions with no direct correlates in Putonghua (see section 4.4). 10. The same particle da/de functions like PH dao when it occurs as a verb ('arrive') or as a preverbal preposition ('to') typically in serial verb constructions (see example in "Literary items," section 4.3.3 of this chapter).

Chapter 9 Conclusions: Pragmatic Universals in Second Dialect Acquisition Abstract This final chapter draws conclusions from the comparison of the discourse features observed in the acquisition of standard versions of Belizean English and Chinese. Those genetically and areally unrelated distinct corpora will serve as a basis for a dis­ cussion of such issues as the putative distinctiveness of creole languages, the signifi­ cance of adult language acquisition in language change, the interaction of syntax and pragmatics, and universal aspects of second dialect acquisition in discourse contexts. 1. Patterns of topic marking: Summary It has been assumed that, unlike child language acquisition, adult acquisition, broadly defined as involving the acquisition of second dialects or formal styles, does not contribute much to language change: The ability to master a language like a native, which children possess to an extra­ ordinary degree, is almost completely lacking in the adult. . . .I conjecture that changes in later life are restricted to the addition of a few rules in the grammar and that the elimination of rules and hence a wholesale restructuring of his gram­ mar is beyond the capabilities of the average adult. (Halle 1964:344) Halle's view—originally applied to phonology—may appear to be sup­ ported by observations of discourse strategies which exhibit remarkably homo­ geneous mechanisms across first and second dialects. On the other hand, one might argue that adults, on the contrary, play a significant role in language development simply by ensuring the maintenance of topic mechanisms in acquired dialects (Hopper & Traugott 1993:212).

264

CREOLE AND DIALECT CONTINUA

The survey of topic marking phenomena, as presented in the preceding chapters, suggests a striking homogeneity of topic strategies among two distinct sets of varieties-—the acrolects of Belizean Creole and two corpora of Chinese dialects (Wuhan Putonghua and Suzhou Wu), all acquired as second dialects. In both cases, learners give a high priority to topic-comment structures, in which topics are clearly promoted and highlighted according to at least three system­ atic strategies: fronting, repetition and presentation. The focus was kept intentionally and unavoidably narrow because of the methodological approach adopted in the context of this analysis. This perspec­ tive is predicated on the principle that observational adequacy can only be appropriately met if the spectrum of acquired varieties is spanned as thorough­ ly as possible; this requires a corpus including speakers' extensive repertoires, in which social context and individual characteristics cannot be overlooked. Consequently, the choice was made to forfeit an analysis of a greater number of linguistic features that would have necessarily entailed a corpus limited to an unacceptably small subset of the linguistic repertoire. Extensive, if not exhaus­ tive, repertoire investigations demand time and effort, but they constitute pre­ requisites to the successful evaluation of a given variety, and more specifically in the case of second dialect acquisition. Restricting this study to one single linguistic feature can only provide a par­ tial understanding of the overall process of the acquisition of standard varieties by speakers of second, or third "dialects," or nonstandard varieties. Nevertheless, the feature selected is a complex linguistic variable which anchors the communicative event and has far-ranging effects over the whole sentence, proposition, or topic unit. Topic mechanisms relate to crucial aspects of communication and are clearly linked to distinctive grammatical structures such as serialization or relativization. Topics are presented effectively and persuasively through the use of various devices, such as special placement in the utterance, repeated copies, and a wide array of presentative structures and grammatical elements underscoring those top­ ics. Thus, in effect, topic marking mechanisms collapse syntactic and pragmatic processes in a challenging fashion. Close observations of versions of the standard produced by speakers of nonstandard dialects in Belize and in China, as compared with their native pro­ duction of nonstandard varieties, provide a broad perspective on discourse mechanisms. Patterns of topic marking in acquired second dialects turn out to be similar, quantitatively as well as qualitatively, to patterns of topic marking in first dialects, whether standard or nonstandard—at least for those examined here (Belizean basilects, American standard English, and Chinese/Putonghua). Ordinary discourse does not consist of independent structural units, but of a concatenation of meaningful sequences which are likely to be goal-oriented in

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

265

the sense that speakers and hearers intend to exchange messages. To this effect, they develop strategies that enable them to communicate productively. One such strategy is the general set of mechanisms presenting core information or topics. Although some language may assign topics a relatively stable syntactic position in the grammar (in the case of topic prominent languages), it is likely that all types of languages, whether the product of native acquisition or adult acquisition develop similar mechanisms. 2. Topic marking in the Belizean continuum As far as the creole situation is concerned, it was shown that acrolects can be differentiated both from other lects and from the standard lexifier in terms of a number of phonological and morphological features. However, such lectal dif­ ferentiation does not apply to discourse features. Topic strategies are found to be remarkably alike in all Belizean English-based varieties—ranging from basilects to acrolects, regardless of context, style, status, or gender. Another interesting, though still preliminary finding, provides insight into patterns of diachronic change in the creole continuum. Varieties produced twenty years apart do not evidence any major difference, either in terms of dis­ course topics. All speakers investigated use large amounts of topic enhancing strategies, with somewhat differential usage of the three identified mechanisms. The most favored strategy overall for highlighting topics is the one involv­ ing the use of presentative structures, occurring 53.6% of the time in acrolects and 45.5% in basilects (see tables 4.2 - 4.5). Thus, presentation is slightly more favored in acrolects, with a predominance of various introductory phrases, such as one of the things; dat s one ting wid him; de next problem; the point is. These strategies are innovative because their function is clearly one of grammaticalized focus, much beyond the literal semantic value of the lexical items includ­ ed in those strings. Several presentative elements can occur contiguously before or after a topic (including any combination of NP, V, VP or adverb), and the topic itself may be repeated, commonly producing a string, as shown in (1): (1)

# (PRESENTATIVE)11 + TOPIC (X) (PRESENTATIVE)11 + (TOPIC)11 (Y) #

This pattern is illustrated in (2) and discussed in chapter four, which includes as many as three separate presentative phrases even before the first occurrence of the verbal topic, a post-topic element (now), and three addition­ al introductory phrases preceding the second topic reference:

266

CREOLE AND DIALECT CONTINUA

(2) #You see, the point is, if you goin' to think PRES

PRES

PRES

+

now,

ok,

TOPIC + PRES, PRES

let's say, PRES,

for instance, each one of us thinhin ' differently.1 # PRES

TOPIC

Topic repetition closely follows presentation as a favored topic marking strategy, with an average 38.7% repetition overall (36.6% in acrolects and 46.0%) in basilects). In my corpus, basilects and acrolects slightly differ in their choice of strategies, since basilects show an equivalent distribution of repetition and presentation strategies, 46.0% and 45.5% respectively, whereas acrolects show a preference for presentation. However, it should be noted that repetition is closely linked to presentation, since the most common strategy is a version of left dislocation with a full nominal copy. Focus is assigned to a topic by the use of introductory phrases, as indicated above; the topic is then repeated fully or partially later in the sentence, and additional topic copies can also be pre­ ceded by other presentative elements. This pattern is illustrated in (2), with the repetition of the topic think. Finally, fronting, the least favored of the strategies represented in my corpus, never amounts to more than 8% or 10% of all topic strategies in a given lect. The crosslectal homogeneity of topic mechanisms in the Belizean varieties examined suggests that a second dialect offers extensive similarities with a first dialect in the context of nonstandard native speakers, and in particular, that the discourse/pragmatics level is the language component least likely to show first dialect substratai effects. This is generally confirmed by the longitudinal study of Belizean varieties over a ten-year interval, including data collected in 1984 and 1994, which reveals no significant difference in topic marking, except a slight decrease that would need to be confirmed by further study. The surpris­ ing finding that there is crosslectal homogeneity in Belize runs counter to offi­ cial educational policy and popular belief that acrolects are regular "good" stan­ dard English, and basilects are rural, uneducated, and "inferior" dialects. This discovery was made possible through extensive comparison across lects and speakers. Clearly, these findings have important implications involving educa­ tional policy and putative curriculum changes, as will be discussed below. 2.1 Topic consistency in English-based continua When comparing Belizean data to American English, it was found that casual American English presents strong evidence of general trends in topic marking similar to those found in Belize. Comparing Belizean to an external standard English not necessarily in direct contact, but serving as a possible formal exter-

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

267

nal model and educational tool, permits an assessment of the universal status of the pragmatic strategies implemented in topic mechanisms. Although American English was selected as a matter of convenience, any other standard variety (British, Canadian, Australian, etc.) would have achieved the same goal, name­ ly, to assess whether non-native standard dialects differ from native standard varieties. Surprisingly, and contrary to popular belief, a semi-casual, profes­ sional American English sample (see chapter four, and table 4.9) exhibits an even more polarized use of topic strategies, as summarized in table 9.1: Table 9.1: Topic strategies in Belizean and America

(semi-casual)

N

Fronting

Repetition

Presentation

Basilects Acrolects

[189] [717]

8.4% 9.7%

46.0% 36.6%

45.5% 53.5%

ALL BELIZEAN

[906

9.4%

38.6%

51.8%

SEMI-CASUAL AMERICAN

[252]

5.0%

21.4%

76.4%

Presentative structures rank high—76.4%—on the list of topicalizing devices used in casual American English (as compared with an average 52% for Belizean), whereas repetition averages only 21.4%, and fronting shows mini­ mal incidence. This sample is only a brief control because the issue here is not to achieve a full comparison of Creole English or English acrolects with stan­ dard varieties such as American English but simply to shed light on the putative distinctiveness of second dialect acquisition processes. The conclusion is that some styles of English do use large quantities of topic marking strategies, just like creole acrolects. In this case, casual, as well as semi-casual, American English would seem to definitely favor presentation over repetition, even more so than Belizean varieties. Considering the wide range of adverbials found to introduce topics or sepa­ rate topics from comments, an adequate description of topics versus nontopic ele­ ments must be accomplished in relational terms because human language is always relational and cannot exist out of natural contexts. Comrie (1989:64) argues that English does not grammaticalize topics because it would be inappropriate to intro­ duce a chapter with a cleft construction, e.g. in this chapter, it's relative clauses that we are going to talk about. However, one can argue that a formal written style is not a necessary and sufficient test of the structure of English. From a more realis­ tic viewpoint, spoken informal varieties are statistically much more widely repre­ sentative of human communicative activities than such formal tasks as introducing chapters, and therefore, it seems that an assessment of linguistic properties such as the grammaticalization of topics (to be discussed in section 4.2 below) should be more appropriately based on casual speech, or at least should refer to comparably

268

CREOLE AND DIALECT CONTINUA

relaxed spoken or written contexts. As indicated above and amply represented in the lawyers' conversations analyzed in chapter four, and summarized in table 9.1 above, there is ample evi­ dence that English varieties—at least casual varieties2—have multiple ways of assigning grammatical functions to lexical items in an effort to distinguish top­ ics from their comments, and that they do so with a high frequency of topicstrategies, especially topic particles per topic unit (over four). 2.2 Topic marking in formal and written styles Formal and written styles also use creative devices to present topics. To cite only one example of academic prose, a linguistic discussion displays plentiful usage of a variety of presentative (though semantically empty) verbal structures which function exactly like the traditional (informal) cleft construction. In each of the following sentences, the message is in essence: "this chapter/section is about. . ." or "it is X that we will talk about." (3) In this section, we will contrast two major methodological approaches to language universals that have been adopted in recent linguistic work (Comrie 1989:1). (4) In this section, we will establish some more practical reasons why the study of language universals must operate with data from a wide range of lan­ guages (Comrie 1989:3). (5) In the present chapter, we will look in considerably more detail at one par­ ticular aspect of this problem, namely the definition of subject cross-lin­ guistically (Comrie 1989:104). The above introductory sentences demonstrate unambiguously that devices such as paraphrases, verbal introducers (look at one particular aspect of this problem, consider, contrast) or adverbial introducers {namely and even in considerably more detail) truly function like the cleft (it is X. . .that Y. . .), which occurs primarily in spoken varieties. Even semi-formal argumentative styles do not escape the elaborate setup provided by presentative strategies. To illustrate this situation, I adduce a few excerpts from a conversation conducted in an American university context. The following sentences are drawn from an evaluation of junior faculty performance. Participants are all senior faculty members in the English department of the University of Minnesota. It is clear that academic contexts partake in the widespread reference to topic marking devices. The samples shown below include four introductory phrases, and an

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

269

instance of threefold repetition in (6); the recurrent use of the discourse mark­ er you know, which may function as topic introducer sentence internally, as well as as topic ender (both in (7) and also in (8 and 9); the use of deictics as in this thing .. .talking to people, and this taking too many people in (8), which is also related to a type of left dislocation combined with clefting in (8) this taking too many people, that's something; there is also a type of verbal topic presentation in (9) which is a variant of a pseudo-cleft what she has done, she has accept­ ed; there is even a type of end particle in (10) so, which occurs at the end of one person's sentence; finally, there is evidence of prepositions functioning as topic introducers with fronted topics—with A in (II)—thus, evidencing a behavior not unlike that of verbal prepositions/coverbs in Chinese, such as bǎ, duí, gĕi, or cόng, as illustrated in (12) and (13) below: (6)

In other words, maybe, I think, it's also worth saying that A is in an unusu­ al position. There is no- she has no faculty member, there really is no one else who has established networks in the arts [A, Minneapolis, 11/2/94]

(7)

So she talked about, you know, different categories of people, you know [B, Minneapolis, 11/2/94]

(8)

And so, what she has done, she has accepted too many of those invita­ tions. So, I think, this thing, you know, talking to people in different ways, this taking too many people on [committees] that's something we should talk about [A, Minneapolis, 11/2/94]

(9)

I find it, it s, you know, someone has to do something for her, you know. But I think, one of the things that we could do, you know what I mean [B, Minneapolis, 11/2/94].

(10) This is really rich, so [B, Minneapolis, 11/2/94] (11) I think with A, it's not because she can't do it [A, Minneapolis, 11/2/94] (12) Xiànzài, gēn jiāzhǎng, [Now (TI) with (TI) parents, ASP]

dōu shuō bú shàng shenme l e = all (AP) talk not result s o m e t h i n g -

=jiùshí yī shuō ba, j iāzhǎng, dōu bǎ nǐmén dāng= [so (TI)—{AY) speak (TE), parents, all (AP) (TI) you take ] =háizi [child

kán. see]

270

CREOLE AND DIALECT CONTINUA

'Now there is almost nothing you can talk about with parents, that is, when you talk, your parents treat you as children' (MSS.1985, 39:20). (13) duí zíjǐde nan ren yào j ī n g c h á n g j iàoyù ne 'Regarding (your) own men, (you) must always educate (them), don't you' These short excerpts suggest that contemporary American English, even in professional contexts, is as richly subjected to the ubiquity of topic particles as any of the Belizean or Chinese varieties examined. Deictics are also used as introducers, as well as prepositions (11). Particles introducing topics, such as so in (8) can also be used to close an utterance, as in (10). This may be a relative­ ly new development, but it is nevertheless noticeable in the United States, and it is definitely not restricted to the routinely lamented "sloppy" speech of teenagers. Structural units are generally linked into tighter amalgamations by the use of such phrases as in terms of, on the basis of, or given that, as also observed by Hopper & Traugott 1993:177). There is clearly a shift from the tra­ ditional meanings of adverbials or phrases (so, you know, I think, really, actu­ ally, etc.) to a grammatico-pragmatic function, which pervades all stylistic con­ texts. In brief, there is ample evidence that even semi-formal English in the con­ text of those educators whose mission is to instruct in the use of elegant English makes extensive use of topic presentation in particular, and repetition as well. 2.3 Topic marking by listeners There are other creative ways to mark off syntactic boundaries in discourse, and they are also represented in listeners' utterances to signal interest in the speak­ ers and punctuate semantic/pragmatic units. For example, minimal responses, such as mmhmm, yeah, uh huh, right appear at strategic places at phrase bound­ aries in predicates in 95% of all cases (Fellegy 1995), as shown in (14-16): (14) Oh urn g[V is [ N p velocity] [pp of the apparatus]] in the-I think it's just a ya ya beam of light sorta (15) s[The alleged victim may remove [NP any member] [pp of the committee] mmhmm [pp for good cause]] and an alternative. . . mmhmm

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

271

(16) s [she had this kind of mystical experience [pp with this woman]] ya s [coming [pp down]] and s [resting [pp by her]] mhmm mmhmm (Fellegy 1995:190) This pattern is found to be consistent across various groups of speakers in Minneapolis and not related to gender or style variables. The minimal response thus effectively functions as "a hinge between linguistic and communicative competence" (Fellegy 1995:196), which demonstrates that even paralinguistic elements can fulfill needed functions in the syntax of discourse. In conclusion of this brief summary, the interesting finding emerges that all varieties of English make active use of topic marking devices, either in the form of straightforward presentative structures or under the guise of adverbs, verbs, or more complex structures which in effect illustrate the ongoing grammaticalization of topic marking. In view of the findings presented in the previ­ ous chapters and summarized here, there is justification to doubt the validity of the general belief that "English does not have any grammaticalization of topic versus nontopic (comment) status" (Comrie 1989:64). In fact, Comrie and oth­ ers somehow mitigate such statements when admitting that there are "restricted kinds of topic or focus that can be grammaticalized," such as contrastive topic and focus, which require reference to a context, and in particular to pragmatic relations between noun phrase arguments and their predicates. 3. Patterns of topic marking in Chinese The examination of native and non-native varieties of contemporary Chinese demonstrates the importance of topic mechanisms in all forms of Chinese as well, not surprisingly in light of the general knowledge that Chinese is a topic prominent language, in contrast to other languages such as English. What is unexpected is the extent of the topic strategies used in Chinese and the equal importance assigned to topic strategies in native as well as non-native varieties, an homogeneity parallel to that found in Belizean and American texts. All var­ ieties give a high priority to presentation, but use fronting minimally. Yet, Chinese differs from all English-based samples in as much as left dislocation is practically absent from the Chinese corpus examined here, whether in the reported peasant speech of the 1960s and 1980s literature, in spontaneous Beijing and Wuhan conversations, and in pingtan Suzhou story-telling style. Each of these corpora includes no more than one or two examples of left dislo­ cation, in the context of an already restricted usage of the fronting strategy. Chapter six (section 2.2) also adduces evidence from literature that confirms the

CREOLE AND DIALECT CONTINUA

272

extreme rarity of resumptive pronouns with fronted elements. However, this is not tantamount to saying that left dislocation is ungrammatical in Chinese, as it may occur in varieties not examined here. For example, Matthews and Yip (forthcoming) say that Cantonese structures with topicalized sentence subjects are grammatical with or without the resumptive pronouns. The question is how frequently the resumptive pronoun is realized. Topic strategies are remarkably alike in native and acquired varieties of Chinese. There is an overwhelming frequency of the presentative strategy, which prevails over the other two strategies in the 65% to 69% range in all three corpora: Beijing (native) Putonghua, Wuhan Putonghua and Suzhou Wu (nonnative varieties) as shown in chapter eight (table 8.23). Table 9.2: Topic strategies in Chinese (table 8.23 in chapter eight)

Beijing PH Wuhan PH Suzhou WU

N [1933] [400] [833]

Fronting 11.2% 4.8% 8.5%

Repetition 23.3% 26% 26.3%

Presentation 65.5% 69.2% 65.2%

No major difference in this respect emerges either from the longitudinal comparison of the peasant vernacular represented in a sample of 1964 literature, a similar type of vernacular represented in 1983 literature and, in 1985-1986, semi-casual television interviews and spontaneous conversations. Topic presentation is so pervasive in Chinese that four categories of topic particles were identified to sort out the complex mechanics of topic patterning in Chinese: 1) pre-topic morphemes (topic introducers) include, besides regular adver­ bials and deictics found in other languages, a category of verbal prepositions which have no equivalent in English, as they require the fronting of the topic they modify; 2) post-topic particles (topic enders) which closely follow a topic, and also occasionally occur in English, such as the now illustrated in (1) above, or the ubiquitous you know; 3) pre-comment particles, dubbed "adverbial particles" for Chinese because this class is primarily composed of adverbials or intensifiers, which structurally separate the topic from a following verb-centered commentary on the topic; and 4) post-comment particles (end particles) which signal the finish of an utterance (in English they can appear as disclaimers). All of those topic parti­ cles were found to have similar parsing functions, namely, to separate topics

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

273

from comments and, in other words, to grammaticalize the structure of topicoriented utterances. As indicated in previous chapters, APs and TIs—the precomment and the pretopic particles—are most frequently represented, which indicates that the initial position, or the position at the beginning of any syntactic unit, is the most efficient in presenting information. This pattern applies particularly to the nonnative Putonghua spoken in Wuhan, where the AP category amounts to 52.7% of all topic particles. Furthermore the Wuhan variety is characterized by the use of very few recurring elements (such as jiù, and yě). Such extreme specializa­ tion suggests that some adverbs previously restricted in their structural func­ tions are now grammaticalized into complementizers or topic-comment links. A similar specialization is represented in the dominant choice of the socalled Chinese copula shî or its cognate si/z in Suzhou Wu as topic introduc­ er or focus marker, which constitutes 41.3% of all introducers in Wuhan, and 40.8%) in Suzhou (but only 19%> in Beijing Putonghua), as summarized in chap­ ter eight (table 8.25). Thus, although all varieties of Chinese show that topic strategies and more particularly topic particles are essential in discourse structure, non-native vari­ eties appear to exhibit a more generalized version of the systematic ongoing grammaticalization of topic marking through the selection of a few representa­ tive elements. 4. Dialect versus standard: Sociolinguistic universale There are obviously differences between Belizean acrolects and Chinese acquired dialects; for example, the Belizean varieties investigated here make ample use of left dislocation, whereas that type of structure rarely occurs in Chinese. Further research is needed to verify those putative differences, but they appear to be relatively minor compared to the similarities observed in the ways Belizean acrolects and Chinese second dialects in Wuhan and Suzhou assign prominence to topics. There is ample evidence that the generally presumed distinction between topic-prominent and non-topic-prominent languages is a matter of degree rather than a clear-cut distinction. All varieties of natural discourse, including the peasant speech reflected in vernacular Chinese story writing of the Cultural Revolution include some form of topic prominence. Presentation is consistent­ ly favored—and even more so in the short casual American sample than in the Chinese corpus. On the other hand, Belizean acrolects display more internal variability in terms of the use of presentative elements; in some contexts, there is an accumulation of particles, whereas in other situations, there is more

274

CREOLE AND DIALECT CONTINUA

emphasis on paratactic types of structure and repetition. In English-based situ­ ations, the use of particles may be more commonly triggered by a highly charged emotional context than in the Chinese context in which focus strategies constitute a regular part of the communicative mechanisms. This suggests that such mechanisms are not yet as deeply grounded in the syntax of English-based varieties as they are in Chinese, in which the systematic use of particles at syn­ tactic boundaries indicates a more advanced, or established, stage of the grammaticalization of topic morphology, and extensive bleaching of adverbials functioning as topic markers. The patterns shared by the Belizean and Chinese continua suggest the operation of sociolinguistic universals that develop in the context of dialect encounters, or putative conflicts between standard and nonstandard varieties. As indicated earlier, conflicts between "dialect" and "standard" appear to lead universally to the co-occurrence of overt and covert prestigious varieties. To that extent, the educational endorsement of an official standard, and concomi­ tant de facto exclusion of other dialects, do not seem to succeed in uprooting stigmatized varieties because such varieties preserve strong grassroot values. This may be a factor in achieving the retention of certain endangered minority languages,3 as well as the survival of ethnicity-linked features in groups long removed from their original native lands. For example, elderly Swedish-Americans in Minneapolis, Minnesota, and Lindsborg, Kansas, exhibit syntactic Swedish features in discourse, such as rel­ ative markers and other particles, which help maintain their Swedish-American mixed identity (Karstadt 1996; forthcoming). Based on oral history interviews, Karstadt found that the vernaculars produced by elderly Swedish-Americans consistently differ from those of socially comparable native speakers of Minnesota English in their predominant use of nonstandard variants, such as dat and a zero relativizer in subject position, as also discussed in chapter four, and illustrated in (17-18): (17) Ja, and, and, and den dey sold it to people[ø] came in dar to da farmers (18) So in 1947 I went to Bloomin'ton [ø] was all farm out dar by Minnesota wiver and bought me tree acres and made my home dar (Karstadt, forth­ coming) Although a substratai explanation can account for the invariable use of dat, which parallels its Swedish counterpart som, the absence of a relativizer in sub­ ject clauses cannot be traced to Swedish, or to standard English for that matter. Thus, this innovation may be typical of contact-derived varieties, and point to

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

275

universal explanations of acquisition. Research in New-Ulm, Minnesota, a German-American community also suggests the implementation of linguistic devices, such as deictics, that differentiate the speech of German-Americans, though long established in the United States, from that of other ethnic groups (Fellegy forthcoming). Fossilization has been used to refer to the persistence of external influ­ ences in a system, but the term fails to capture the dynamic, innovative, and— at least subconsciously—intentional use of old features to preserve a sociolinguistic identity distinct from the majority (usually dominant) group identity. Decreolization has been viewed from a similar perspective—as an innovative acquisition of strategies linking the past to the present rather than as the result of an approximation of a new target implying the intended abandonment of an earlier system or identity. It is clear that the creole in Belize, as well as local vernaculars in China remain strongly established in their respective communi­ ties, in spite of increased access to the educational system. Attachment to native vernaculars remains strong, in spite of the fact that most vernacular speakers are persuaded that their native nonstandard varieties have "no rules," "no gram­ mar," a belief that seems to be universally shared as well, in view of the wide­ spread purist views of language discussed in chapter one. It is possible to pro­ vide a sociolinguistic definition of the vernacular in terms of its emotional value, in contrast to the standard which may be defined in terms of its intellec­ tual value. The acquired second dialect or acrolect may have a more complex mixed identity in the sense that linguistic behavior matches ethnic identity. Contact-induced change may be characterized, at first, by the co-existence of two or more socially-conditioned variants, followed by hybridization producing new structures. For example, spoken Cantonese requires obligatory agent phrases in passives, but agentless passives (Mandarin-influenced) occur in for­ mal or literary registers (Matthews 1996:1275-1276). This phenomenon is a form of diglossia applied to syntax (named ditaxia by Matthews), which even­ tually may result in hybridization, with the possible outcomes of redundancy, double marking, or hypercorrection (Matthews 1996:1277-1278). Such linguis­ tic consequences of languge contact are widely represented in the study of topic structures, whether in Belizean acrolects or in Chinese acquired dialects, all showing a high incidence of repetition, extensive particle use, and the extra assignment of salience to various lexical items. 4.1 Causes of pattern consistency The crucial issue that must ultimately be raised is that of the formal cause(s) of the observed acquisitional consistency of topic marking. What is the relative importance of the substrate on the uniform development of second dialects

276

CREOLE AND DIALECT CONTINUA

(whether Belizean Creole or Wuhan Putonghua, for example)? What of univer­ sal patterns of dialect acquisition? What is the function of dialect leveling as interpreted by historical linguistics? Societal and educational attitudes and century-old biases are partly respon­ sible for the generalized acquisition of second dialects, and the confusion and conflict thus created is certainly causally involved in the pattern consistencies observed. From a more strictly linguistic perspective, it has been claimed in prior discussions of linguistic change that clashes between different systems produce universal structures, sometimes referred to as the operation of natural­ ness conditions, or the production of unmarked features. The two approaches are not necessarily equivalent, and the issue of what constitutes "natural," or "unmarked" features is far from being settled. For many linguists a "natural" structure is defined as "unmarked" or more frequently "distributed." Speakers may revert to Universal Grammar (Givón 1979c), or they may produce forms which reflect the reemergence of bioprogram features originating in their innate system (Bickerton 1981). There is general disagreement as to the relative effects of substratai, superstratal, and universal influences on contact-induced change, perhaps because widely different sets of data, and often severely restricted data, have constitut­ ed the data base underlying interpretations. It has been said that "the universal­ ist and substrate hypotheses complement one another" (Arends 1986; Mufwene 1986). On the other hand, some superstratal influence is typically taken for granted, since the bulk of the lexical component is derived from that superstrate which is assumed to be the target language. Yet little has been said about the potentially continuing effect of the superstrate, thus leading to an extension of the linguistic repertoire without necessarily implying decreolization in the tra­ ditional sense, as discussed in chapter three, or the total loss of an earlier sys­ tem. In this sense, the acrolect constitutes an alternative to the standard (exter­ nal target), and perhaps an end in itself. This is attested by "a matched-guise experiment conducted in Hong Kong by Kingsley Bolton and Helen Kwok in which they found that Honk Kong speakers saw as their target fluent English with a Hong Kong accent, rather than standard American or British English." (Matthews and Yip pc 10/96). In fact, the more qualitative 4 linguistic data are adduced, the more com­ plex the situation appears to be, and this is really not an unexpected finding, since it is likely that ongoing linguistic change happens no matter where and when, and that "creoles do not enter a situation of stasis once nativized" (Muysken & Smith 1986:9). Therefore, traditional explanatory notions used in historical linguistics to explain language development, such as "borrowing," "dialect interference," or "dialect leveling" turn out to be vague and rarely rel­ evant when dealing with extensive corpora of spontaneous speech in the full

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

277

context of discourse. Some linguists have attempted to differentiate more clearly between bor­ rowing and substratum interference: .. .in borrowing proper, many words will be borrowed before any structural inter­ ference at all occurs; but in substratum interference . . .structural interference comes first. . . . In fact, the distinction between borrowing and substratum inter­ ference is so important that.. .we will hereafter use the term "borrowing" to refer only to "the incorporation of foreign elements into the speakers' native lan­ guage," not to interference in general. (Thomason & Kaufman 1991:21) Attempts at differentiating borrowing and substratum interference, however, often fail to account for the phenomenon of grammaticalization which widely occurs in the development of the creole continuum, and especially the acquisi­ tion of acrolects, Chinese non-native standard varieties, and the development of second dialects. It is doubtful that borrowing may occur without any kind of structural change or some restructuring, including the assignment of "old" (per­ haps substratai or native) functions to a newly borrowed lexical item. A "for­ eign" term can apparently be borrowed and still be assigned, without delay, a function that may be a projection of the substratai system, frequently because the borrowed item has some phonological resemblance to a substratai item. I have identified three general patterns occurring in cases of language mix­ ing. They serve to identify multiple, complex, or ambiguous identities. Frequency variability quantitatively alternates old and new variants, yet in a nonerratic fashion. If marked, new variants—such as interdentals—occur first in highly visible strong positions, such as word-initial placement; past marking occurs first in auxiliaries, though often with a pragmatic topic marking function (see chapter four; also Escure, forthcoming). The second strategy, structural hypercorrection, is characterized by multifunctionality. It occurs when a value is marked twice or more frequently; or a new item acquires a second function while still preserving its old function. This strategy can combine with frequen­ cy distribution. The new function, or variant, occurs less frequently than the old function and may gradually overlap. The third strategy, grammaticalization, relates to structural hypercorrection as well, as new items acquire new func­ tions, and those functions become part of the system. I am thus moving toward a causal interpretation of topic marking consis­ tency in discourse, according to which substratai and superstratal elements are manipulated to convey old (substratai) functions through new (superstratal) items. The selection of such forms may reflect an essential universal of lan­ guage contact.

278

CREOLE AND DIALECT CONTINUA

4.2 Grammaticalization in topic marking strategies Meillet was apparently the first to identify, though very succinctly, a phenome­ non akin to grammaticalization. He referred to the grammatical system as the most stable part of language, though progressively undergoing change: "La part de la langue qui se maintient de génération en génération en se transformant progressivement est le système grammatical" (Meillet 1924:6). Martinet also mentions the mechanism leading to the development of grammatical forms out of earlier lexical items: Les nouveaux indicateurs de fonction (prépositions, conjonctions, locutions pré­ positives ou conjonctives) sont formés à partir d'éléments autonomes. Ceux-ci peuvent être des monèmes autonomes (l'anglais up dans he went up the hill, employé comme fonctionnel dans up the hill) ou des syngtames autonomes (fr. sans égard [pour]). (Martinet 1961:180). Grammaticalization, less commonly called grammaticization is thus a linguis­ tic process involving structural change in the system, perhaps the most dynam­ ic and fundamental vehicle operational in second dialect development. It has been variously defined as "the linguistic process, both through time and synchronically, of organization and coding" (Traugott & Heine 1991, vol. 1:1), as "the harnessing of pragmatics by a grammar" (Hyman 1984:73), and more extensively as: the process whereby lexical items and constructions come in certain linguistic contexts to serve grammatical functions, and once grammaticalized, continue to develop new grammatical functions. . .it refers not only to processes observable in language, but also to an approach to language study, one that highlights the interaction of use with structure, and the non-discreteness of many properties of language (Hopper & Traugott 1993:xv). Any definition of this dynamic process must account for the interaction of structure and use, or of syntax and pragmatics. Thus it is a crucial issue to deter­ mine whether grammaticalization is the result of discourse strategies, or is due to other influences. Although it is impossible to identify and separate all the fac­ tors, cognitive, distributional, social, substratai or superstratal, which may con­ tribute to language change in general, and to grammaticalization in particular, crosslinguistic observations are likely to shed light on the operation of gram­ maticalization in the context of topic processing. The overlapping of the prag­ matic properties of topic marking, and of the syntactic mechanisms contributing to the delineation of propositional structure has been amply demonstrated in all the corpora examined here. What is involved is the transformation of various lex-

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

279

ical items (deictics, adverbials, verbs or prepositions) into items that fulfill gram­ matical functions necessary to the communicative process. The items selected for those functions are usually frequently distributed, and probably selected because of their visibility. In some varieties, topic marking functions rely on a few selected items which recur with high frequency (da in Belize; j i u in Wuhan), whereas other varieties, such as Beijing Putonghua, assign those same functions to a wide range of lexical items. This grammaticalization process is well illustrated in the extremely wide­ spread use of da which can alternatively function as topic introducer (19) or as post-topic marker (20) in Belizean basilects, in parallel structures and with equivalent meanings: (19) Da

Habat#

mέ

di kapten

TOP Herbert PAST ZERO COPULA the captain

[It's/that's H. (who) was the captain] 'H. was the captain/The captain was H.' (Escure 1983c: 194) (20) Habat# Herbert

da

me —

di kapten

TOP PAST ZERO COPULA the captain

[Herbert, he/that (was) the captain] 'H. was the captain/The captain was H.' (Escure 1983c: 196) The occurrence of the morpheme da can be interpreted as the appropria­ tion of the superstratal deictic/demonstrative that which because of its frequen­ cy and iconicity became easily extended to fulfill pragmatic functions, and at the same time remained fixed in syntactic positions in the close proximity of topics. It is likely that the shifting from restricted morphological categories to broader syntactic-pragmatic values was facilitated by the influence of analo­ gous substratal mechanisms (Escure 1983c). An issue to be raised is the reason why certain items are systematically and selectively subjected to grammaticalization, whereas others are not. The logical response to this puzzle lies in the universal priorities found in communicative processes. Highlighting the topic appears to be one of these universal priorities. Thus, the morpheme da becomes so important in topic marking that it can also occur after the topic as well, in comment-initial position, thus functioning as a left-dislocated structure (20), as well as in topic initial position, thus function­ ing as a cleft (19). A consequence of this ubiquitous position of da in Belizean basilects is that this topic marker has often been mistaken for a copula because of its postnominal position.5 It is manifestly impossible to identify this da as a copula

280

CREOLE AND DIALECT CONTINUA

because the same highlighting value is associated to the element contiguous to da. Both sentences are cleft/dislocated types which provide salience to the fact that "Habat" is new information in regard to the identity of the captain of the boat.6 In addition, in both cases da is followed by typically preverbal morphol­ ogy (here the past or anterior marker me; elsewhere the progressive/iterative marker ds), which means that da cannot be a copular verbal element; thus, m s must precede a "zero-copula." The universal aspect of pseudo-copula/topic marker ambiguity is supported by data from Chinese. As indicated previously (see chapters seven and eight), the same situation obtains in Chinese under identical conditions: shí (si/z in Suzhou Wu) can occur both in topic-initial and in comment-initial positions, with the same topicalizing function in what can be considered to be a copula-free language, con­ trary to some assumptions that sh i is an occasional copula: (21) Fǎnzhèng shí tĭng =nèige —dòng zhī xīn-de [Really TI AP —TI —move one heart-G] '(It) truly moves me deeply' [Beijing Saga B: 3; see chapter six (28)] (22) yīge shi, Tzhǔyào, shi wŏ àirén, ta gōngzuò yòu máng a, you shi yào shàng kè, yào dài háizī (Wuhan:Text l: 5a-b) 'One problem is really about my husband, his work keeps him very busy, yet he not only must teach classes, but he must also look after the child' [chapter eight (1)] (23) G e j i n z e si nei fe he, dedi nei z hei ninzin, heide Yang Neiwu se gang zong yangzi (Suzhou Wu 22:43)

ηou

'Today, (it's) you (who) are to blame, after all, (it's) you (who) are the traitor, (who) framed me Yang Nanwu (and put me in) this miserable situation' [chap­ ter 8 (20)] The universal interpretation of the grammaticalization of deictic elements receives strong support from classical Chinese data, as represented in texts such as the Confucian Analects and Mencius (dating from 551 to 221 b.c.). In those early texts, the deictic-demonstrative shi is "usually anaphoric, with no par­ ticular implication of closeness or remoteness. . .and functions to recapitulate a phrase or a series of phrases which is the subject of a noun predicate out of which its later use as a copula develop" (Pulleyblank 1995:85). However, the interpretation of shí as "modern copula" does not appear to be substantiated

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

281

by the acquired data adduced in the corpus presented here. It is clear that shî still has a primarily salient value. Thus, the striking similarity of the cleft (pseu­ do-copula) in Chinese and Belizean creole-related varieties at all lectal levels, including acquired dialects, reinforces the universal hypothesis that some spe­ cific particles serve to focus a topic in pre- or post-topic position. If those par­ ticles do not exist, they will be created through grammaticalization. A second example illustrating the topic marking universal comes from the use of other items such as adverbials. For example, it was shown in chapter four that many adverbials, such as still, suddenly, again in acrolectal varieties have assumed topic marking or topic-comment delimiting functions. In the case of acrolect acquisition, this process is interpreted as grammaticalization of simple adverbials which function as syntactic boundaries and partly lose their original meaning. The bleaching of adverbials is even more advanced in all varieties of Chinese, to the extent that practically every topic chain requires its chain links. The particles which operate the smooth articulation of topics and comments as substitutes for the most common basilectal da. In Chinese also, numerous adverbials function as topic markers. This can be viewed as a case of substrate or first dialect interference through borrowing, and also involving restructur­ ing of the system in the sense that the grammaticalized adverbs are added in positions and in quantities not used in formal standard versions of the same sen­ tences. Finally, there is the case of the development of formal styles which use stilted verbal introducers in lieu of simple clefts or pseudo-clefts, as illustrated above in samples of chapter introductions (see 3-5 above). Thus, similar types of essential but subtle grammaticalization occur throughout the developmental continuum—whether in acrolect acquisition or in stages leading to the devel­ opment of the creole basilect, and in the stylistic continuum of standard lan­ guages as well. Tracing this phenomenon to the universality of topic presenta­ tion in natural discourse does not conflict with the substratai interpretation, but only confirms why certain features are selected over others. Those features which are common to substrate and superstrate are eventually preferred. These examples support the claim that reanalysis is a major factor in change (Hopper & Traugott 1993:48), and therefore that change may be accomplished by adult learners of second dialects. 4.3 The syntax of discourse The consistency of topic marking observed in the previous chapters yields insights into the relation between syntax and discourse. Perhaps the question to be posed is: Is there syntax in discourse? However, this question only arises if syntax is defined narrowly as a subset of formal properties defining an

282

CREOLE AND DIALECT CONTINUA

autonomous level of structural organization, as it is in traditional linguistic the­ ory. If so, the answer must necessarily be negative, considering the shifting val­ ues of grammatical elements, both in Belizean and in Chinese. An autonomous level would exclude the possibility of interaction with the pragmatic level, of concrete communication. Perhaps, a better question to ask would be: Is syntax independent from pragmatics? Based on the evidence provided in the previous chapters, the answer must also be negative. Clearly, syntax is highly dependent on discourse strategies—even centered around the communicative events of normal dis­ course communication. My data strongly support Givόn's definition of syntax as a "dependent functionally motivated entity" (Givón 1979b:208). The Belizean and Chinese corpora—all collected in natural discourse situ­ ations as documented above—illustrate what I called a stacked syntax, that is high incidence of paratactic structures with a low frequency of embedding and avoidance of subordination and complementation. This phenomenon has also been referred to as clause chaining by DeLancey (1991) who describes the ori­ gin of verb serialization in Tibetan through the grammaticalization of verbs. Avoiding complementizers, including relative pronouns, naturally leads to rep­ etition as shown in some of the examples of relative clauses, and others, given in chapter four. However, repetition is not a redundant, ponderous vehicle for communication but rather a necessary enlightening device—Ariana's thread running through the complex convolutions of natural discourse. The fact that there is a predominance of parataxis, as found in Belizean and Chinese discourse, does not mean that there is a lack of structural organization. Discourse is highly structured around topic-comment constituents: an initial phrase (topic) sets the stage for what is said about it (comment). Thus, although in formal contexts syntax avoids redundancy by the use of pro-forms or para­ phrases, creole and Chinese discourse—spanning the whole continuum from native varieties to acrolects/acquired dialects—shuns proforms to favor repeti­ tion and presentation as topic-marking devices. This results in a fluid type of structural organization which cannot be neatly broken into delineated NP-VP sequences, thus forcing an analysis of discourse units as spanning semantic units beyond sentence boundaries. But even formal contexts may use clitics which will follow the initial topic, as in "the judge, however, took a different view," implying a contrast between "the judge" and some other participant, and thus clearly marking "the judge" as the topic of this sentence (Hopper & Traugott 1993:134). Such strategies allow structural boundaries to separate effectively topics from their comments. Givón's discussion of the diachronic process of syntacticization con­ tributes some stimulating ideas which are highly relevant to the present discus­ sion. Second dialect acquisition can indeed be considered diachronic since it

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

283

involves the sequential addition of one system to another. Givón assumes that it is typical for "loose paratactic pragmatic discourse stmctures [to] develop— over time—into tight grammaticalized syntactic structures" (Givón 1979a:208). In fact, he posits two extreme poles of communication: the pragmatic mode and the syntactic mode. His discussion implies that the pragmatic mode is a stage prior to the syntactic mode, with the following associations: Pragmatic

Syntactic

Pidgins → Creoles Child language → Adult language Informal language → Formal language (Givón 1979a:223) However, such a characterization seems to imply that there is a unilateral development from a "loose" pragmatic organization to a "tight" syntax, and more importantly, that the syntactic stage is more evolved than the pragmatic stage. The situation is manifestly more complex than what seems to be implied by this claim of directionality toward syntacticization.7 It appears more likely that the situation is such that both "modes" coexist, interact, and are simultaneously available to speakers in different degrees, and subject to the effects of context. Indeed, formal versions of standard languages are learned later than informal varieties under normative conditions, and typi­ cally through the strict constraints dictated by educational systems. Yet, infor­ mal varieties are never abandoned, and the acquisition of formal styles merely implies an extension of the stylistic repertoire available to individuals. Similarly, the acquisition of non-native acrolects in Belize and in Wuhan, Hubei (PRC) expands the stylistic/lectal range available to native speakers of Belizean Creole and of nonstandard Wuhanhua. Observations of production data in SLA also support the claim that dis­ course functions and syntactic positions are simultaneously elaborated. Japanese and Chinese learners of English produce a high frequency of interlanguage structures representing topic prominence, especially periphrastic topic constructions withfor and pseudo-passives in which the fronted topic is an argu­ ment of the verb (Yip & Matthews 1995), as illustrated below: (24)

For first language acquisition, the acquiring process is unconscious...while for L2 learners, they also imitate and follow the pro­ nunciation of their target language (Yip &Matthews 1995:26)

CREOLE AND DIALECT CONTINUA

284

(25) These ways can classify two types (Yip &Matthews 1995:22) This structure is parallel to its Chinese version (26), in which these ways is the object of the verb: (26) Zhexie fangfa keyi fen These way

liang

zhong

can classify two types

'These methods can be classified into two types' (Yip & Matthews 1995:22) In conclusion, a close look at second dialects in natural discourse indicates that the syntactic structures of acquired dialects partake of the same pragmatic strategies. The organization of discourse in terms of clearly marked topic chains (or series of topic/comment sequences) is a natural development in non-native dialects, regardless of the degree of topic prominence of the corresponding native systems of second dialect learners. What is remarkable is that informal (spontaneous) varieties of a language also exhibit highly topicalized structures (as represented in the casual American sample presented in chapter four). The major finding emerging from this study is that discourse is a great equalizer, the common denominator to all languages and all varieties. Regardless of language type, and irrespective of whether the language is a cre­ ­le or not, natural discourse exhibits similar types of pragmatic structures in topic presentation, or pragmatic inferencing (Hopper & Traugott 1993:63). In other words, the role of speakers negotiating meaning in communicative situa­ tions cannot be discounted and may in fact be a strong motivator in language acquisition. This is a strong claim which is testable, and indeed, should be test­ ed in the context of more languages and more extensively documented speak­ ers' repertoires. The finding that pragmatic principles are universally pre-eminent in spon­ taneous narratives is not tantamount to saying that such discourse is incoherent, disorganized, and redundant (which is the popular belief). On the contrary, top­ ics and their related propositions are tightly coded with high frequencies of par­ ticles—or various lexical items—grammaticalized to function as topic particles, as well as as by minimal responses produced by listeners, which clearly sepa­ rate the major semantic units of discourse for easier processing. If a language does not have pre-existing presentative particles, other elements will become grammaticalized units, thus creating a true syntax of discourse ruled by prag­ matic principles. Finally, the issue of the social context of second dialect acquisition must

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

285

ultimately explain a striking feature of second dialects. Formal (written) forms of a standard language like American English (which can be viewed as second dialects acquired after the native vernacular) do exhibit only one type of topic marking (presentation through formal verbal structures), whereas the informal end of continua evolving from nonstandard dialects exhibit richly varied topic and multipropositional structures, with extensive repetition as well as presenta­ tion. One of the reasons for this style-linked differential may be that the acqui­ sition of formal dialects in a standard context is highly constrained and codified within the educational system. Repetition is a strong taboo, whereas elaborate introductory devices are prestigious. Such attitudes are primarily determined by social constraints acquired in educational and professional situations. In contrast, the acquisition of standard dialects by nonstandard speakers is not directly addressed within an educational system, since nonstandard varieties are rarely given overt recognition, whether in school or in any official context. Speakers of Belizean Creole are "assumed" to speak English, and speakers of Wuhanhua are "assumed" to speak Putonghua. It is generally "assumed" that the acquisition of the standard is not problematic. In China, the problem is com­ pounded by the fact that the written logographic system, which signals formal literacy, does not reflect extensive spoken differences between different forms of Chinese. Even the use of the pinyin written system in the early years of pri­ mary education does not resolve the issue of dialect discrepancies because the pinyin orthography is based on Beijing Putonghua. Thus, the acquisition of acrolects and other standard forms is totally spontaneous and obviously func­ tions like a vernacular. In this case, language acquisition (and change) is goaloriented, or even teleological, a perspective that fits with a ftmctional view of language. Language change is determined by strategies used by speakers/hear­ ers in producing an understandable flow of speech, and this requires a strong anchoring of topics. 5. Conclusions A thorough examination of unrelated vernacular languages demonstrates that the discourse component, and specifically topic marking, functions in a univer­ sally manner consistent with such psychosocial motivations in communication as "economy, efficiency, clarity, expressivity or routinization" (Hopper & Traugott 1993:67). Many different dimensions were investigated in this survey of second dialect acquisition, namely, wide geographical distribution (Central America, China, and to a limited extent, the United States); genetically unrelat­ ed languages (African/Indo-European for Belizean; Sino-Tibetan for Chinese); acquisitional diversity (native versus acquired varieties); longitudinal compar-

286

CREOLE AND DIALECT CONTINUA

isons (1960s and 1980s for Chinese; 1980s and 1990s for Belizean); and final­ ly stylistic discrepancies (from written to casual styles). Throughout those diverse dimensions, the varieties examined inescapably demonstrate that the syntax of discourse has universal procedural priorities, all varieties displaying the systematic marking of syntactic boundaries through a regular three-way sys­ tem of topic mechanisms. When there is no formal pre-existing topic morphol­ ogy, the grammaticalization process assigns required syntactic values to avail­ able items such as adverbs, deictics or pseudo-copulas, and even minimal responses. Furthermore, in languages like Chinese that are already programmed for topic orientation, the language develops supplementary items to increase the syntactic definition of topics and their predicated propositions. In this sense, the increasing use of lexical items, adverbs, coverbs, deictics, etc., may be viewed as the process by which lexical elements, partially or totally bleached of their original semantic value, come to mark grammatical relations. In dialect acquisition, it is clear that the incorporation of substratai features is not merely a calquing process but that it is, in part, a dynamic reinterpretation through superstratal influences, and also affected by contact universals that appear to operate most effectively at the pragmatic level. Universal elements may be primarily identifiable in terms of the common features which are those transferrable into the new code. This implies that a new system pragmatically evolves, in priority, a discourse-oriented syntax. Acrolects thus have great potential for modifying and remodeling an available standard target in innova­ tive ways that match pragmatic requirements. 6. Postscript: Directions for education—-from language to social change The ultimate purpose underlying any study of socially marginalized groups in developed or developing nations is to eventually achieve the demarginalization of those groups. One of the major stumbling blocks interfering with a balanced incorporation of standard learners into the world of native speakers of the stan­ dard resides in the uncertainties of those learners who are regularly faced with language conflicts. Linguistic insecurity is naturally a reflection of the lack of recognition of nonstandard dialects and related social fallout. This entails igno­ rance of the linguistic social and linguistic distinctiveness between standard and nonstandard varieties. This gaping absence has an immediate practical conse­ quence in the sense that learners have no clear starting point, or point of reference, to help them develop productively throughout the required standard acquisition process. The assumption that all Belizeans speak English, or that all Chinese speak Mandarin/Putonghua, is counterproductive because it is manifestly wrong. The artificial sense of security resulting from the fallacy of linguistic commonality

PRAGMATIC UNIVERSALS IN SECOND DIALECT ACQUISITION

287

results in confusion in the minds of teachers and students alike. I noted once the equation here-hair-hear figuring prominently on the blackboard of a Belizean classroom. This clearly represented the instructor's intention to teach distinctive orthographic representations of what he perceived as a homophonic set pronounced as [her]. On another continent, my Chinese graduate students had never realized that their acquired Putonghua omitted Beijing-type retroflexes (with concomitant vowel and tone changes) until they started sociolinguistic fieldwork in their hometown of Wuhan. Dangerous ped­ agogical traditions are illustrated in both cases, namely, the habitual priority given to writing and reading, and the general omission of any systematic refer­ ence to spoken forms, as well as to the linguistic variability linking basilects and acrolects. Education is potentially the most effective equalizer because it can reach children early enough in life. The issue of language acquisition and the associated sense of social and ethnic identity in relation to others is proba­ bly the most crucial, yet the most downplayed, element in educational systems. The main findings of these observations of Belizean and Chinese dialect continua include the general distinctiveness of first and second dialects, espe­ cially on the suprasegmental and phonological levels, the strong covert prestige assigned to nonstandard vernaculars, the special psychosocial value assigned by their speakers to non-native acquired standards, the universal spontaneous strategies used to process information, and the general overlap of syntax and pragmatics. All of these observations can easily be applied to develop progres­ sive, nonprescriptive, and more flexible systems of education. Some possible innovations might include: 1) providing equal educational time and respect to native and non-native varieties co-occurring in a given society. Thus varieties no longer need conflict, but can co-exist peacefully; 2) training teachers to incorporate locally-appropriate discussions of the social functions and linguistic aspects of the varieties in contact, and; 3) recognizing that spontaneous oral communication is the easiest medium through which to initiate the acquisition of second dialects (rather than through formal and written forms), since they are identical to primarily oral vernaculars, at least in the area of topic strategies. Such commonalities would provide a sta­ ble and positive base to the acquisition of the standard. The task of creating unity in plurality is not as paradoxal as it appears, and leveling or exclusion is not the solution. The Roman phrase e pluribus unum has been interpreted in Belize as Belize da all a wi. Perhaps not surprisingly, a stan­ dard English translation "We are Belize" fails to capture the salience of the Belizean structure, more appropriately reflected in its casual left dislocated ver­ sion: "Belize, that's all of us."

288

CREOLE AND DIALECT CONTINUA

Endnotes 1. This sentence is discussed as (11) in chapter four. 2. It is reminded here that the lawyers' conversation could be qualified as "semi-formal" rather than casual because of the professional context. 3. Thanks to Stephen Matthews and Virginia Yip for pointing out this important implication of "dialect" versus "standard" conflicts. 4. "More" is used here in a qualitative, not a quantitative sense, as stated above. It is felt that a greater range of the community's or the speaker's repertoire should serve to constitute the data base, as opposed to using more data of a single type. 5. An unconscious bias to compare any language to familiar Indo-European grammars cannot be discounted. 6. This is a case where a topic provides new information, as opposed to the traditional claim that the topic introduces old information, followed by new information in the comment. In both Belizean and Chinese data, it is common to find new information topicalized, as occurs in English in such cases as in "As far as X is concerned. . ." or "Now about X. . ." where X is a new topic. 7. Givón later states that "syntactic structure in time erodes via processes of morphologization and lexicalization" (Givón 1979b:208), thus leading apparently to cyclical uses of pragmatic and syntactic modes.

Bibliography Alleyne, Mervyn. 1971. Acculturation and the cultural matrix of creolization. Pidginization and creolization of languages ed.by D. Hymes, 169-186. Cambridge University Press. —. 1980. Comparative Afro-American. Ann Arbor, MI: Karoma Publishers. Andersen, Roger, ed. 1983. Pidginization and creolization as language acqui­ sition. Rowley, MA: Newbury House. Anonymous Editors. 1980. 1978 prize-winning short stories. Beijing: People's Literature Publishing House. Arends, Jacques, Pieter Muysken, & Norval Smith. 1995. Pidgins and creoles. Amsterdam: John Benjamins. Arends, Jacques, Silvia Kouwenberg, & Norval Smith, eds. 1995. Theories focusing on the non-European input. Pidgins and creoles: An introduction. ed. Jacques Arends, Pieter Muysken, & Norval Smith, 99-109. Amsterdam: John Benjamins. Arends, Jacques. 1986. Genesis and development of the equative copula in Sranan. Substrata versus universals in creole genesis ed. by P. Muysken & N. Smith, 103-128. Amsterdam: John Benjamins. Asturias, Francisco. 1941. Belice. Guatemala: Tipografía nacional de Guatemala. Bailey, Beryl Loftman. 1966. Jamaican Creole syntax. London: Cambridge University Press. Bancroft, H.H. 1883-87. History of Central America Vol 6-8. San Francisco: The History Company. Bender, Mark. 1988. The storyteller's aesthetics and 'Song Sigong' Journal of the Chinese Language Teachers Association, 23.2.55-67. Bickerton, Derek. 1975. Dynamics of a creole system. Cambridge University Press. —. 1981. The roots of language. Ann Arbor, MI: Karoma Press. —. 1984. The language bioprogram hypothesis. Behavioral and Brain Sciences 7.173-88. —. 1989. Seselwa serialization and its significance. Journal of Pidgin and Creole Languages 4.155-83.

290

CREOLE AND DIALECT CONTINUA

—. 1993. Subject focus and pronouns. Focus and grammatical relations in cre­ óle languages ed. by Francis Byrne & Donald Winford, 189-121. Amsterdam: John Benjamins. —. 1994. The origins of Saramacan syntax: A reply to John McWhorter's 'Substratai influence in Saramaccan serial verb constructions'. Journal of Pidgin and Creole Languages 9.1.65-78. Biq, Yung-O. 1988. From focus in proposition to focus in speech situation: Cai and jiu in Mandarin Chinese. Journal of Chinese Linguistics 16.1.73-107. Blust, Robert. 1994. The Austronesian settlement of Mainland southeast Asia. Papers from the Second Annual Meeting of the Southeast Asian Linguistics Society ed. by by K. L. Adams & T. J. Hudak, 25-83. Tempe: Arizona State University. Bryan, Beverley. 1996. How much English does the Jamaican primary school child hear? Perceptions of a group of primary school teachers. Paper pre­ sented at the Eleventh meeting of the Society for Caribbean Linguistics. St Maarten, N.A. Burdon, Sir John Alder. 1935. Archives of British Honduras. London: Sifton Praed. Butterworth, Brian, Bernard Comrie, & Osten Dahl, eds. 1984. Explanations for language universals. Amsterdam: Mouton. Byrne, Francis. 1987. Grammatical relations in a radical creole. Amsterdam: John Benjamins. Cadiot, Pierre. 1992. Matching syntax and pragmatics: A typology of topic and topiclike constructions in spoken French. Linguistics 30: 57-88. Cassidy, Frederic G. & R. B. LePage. 1980. Dictionary of Jamaican English. Second Edition. Cambridge University Press. Chao, Yuen Ren. 1968. A grammar of spoken Chinese. Berkeley: University of California Press. Chaudenson, Robert. 1992. Des îles, des hommes, des langues: Essai sur la créolisation linguistique et culturelle. Paris: L'Harmattan. Chen, Matthew. 1976. From Middle Chinese to Modern Peking. Journal of Chinese Linguistics 4.2/3.113-277. Cheng Chin Chuan. 1973. A quantitative study of Chinese tones. Journal of Chinese Linguistics 1.1.93-110. Chinese Writers'Association. 1963. Duan Pian Xiaoshuo, Vol. 3. Selected Short Stories. Beijing: Writers' Press. Chomsky, Noam. 1973. Conditions on transformations. A festschrift for Morris Halle ed. by Stephen R. Anderson & Paul Kiparsky, 232-286. New York: Holt, Rinehart & Winston. —.1977. On WH-Movement. Formal syntax ed. by P. Culicover et al., 71-132. New York: Academic Press.

BIBLIOGRAPHY

291

—. 1981. Lectures on government and binding. Dordrecht: Foris. Christie, Pauline. 1983. In search of the boundaries of Caribbean creoles. Studies in Caribbean language ed. by L. Carrington, 13-22. Trinidad: Society for Caribbean Linguistics. Comrie, Bernard. 1984. Form and function in explaining language universals. Explanations for language universals ed. by B. Butterworth, B. Comrie,, & O. Dahl, 87-104. Amsterdam: Mouton. —. 1989. Language universals and linguistic typology. Chicago: The University of Chicago Press. Corne, Chris. 1989. Un créole à base lexicale française en Nouvelle-Calédonie: le tayo ou le patois de Saint-Louis. Etudes créoles 12.2.29-42. —. 1994. Relativization and thematization in Tayo and the implications for cre­ ole genesis. Journal of Pidgin and Creole Languages 9.2.283-304. —. 1995a. A contact-induced and vernacularized language: How Melanesian is Tayo? From contact to creole and beyond ed. by Philip Baker, 121-148. London: University of Westminster. —. 1995b. Nana  nana, nana  napa: The paratactic and hypotactic relative clause of Reunion Creole. Journal of Pidgin and Creole Languages, 10.11.57-76. Corne, Chris, Deirdre Colemann, & Simon Curnow. 1995. Clause reduction in asyndetic coordination in Isle de France Creole: The 'serial verbs' prob­ lem. From contact to creole and beyond ed. by Philip Baker. London: University of Westminster. Craig, Dennis. 1966. Teaching English to Jamaican Creole speakers. Language Learning 16: 1-2 —. 1971. Education and Creole English. Pidginization and creolization of lan­ guages ed by D. Hymes, 370-391. Cambridge University Press. —. 1980. Models for educational policy in creole-speaking communities. Theoretical orientations in creole studies ed. by A. Valdman & A. Highfield. New-York: Academic Press. —. 1996. Putative absurdity in student writing. Paper presented at the Eleventh meeting of the Society for Caribbean Linguistics. St. Maarten, N.A. Crystal, David. 1987. The Cambridge encyclopedia of language. Cambridge University Press. de Beaugrande, Robert. 1992. Topicality and emotion in the economy of dis­ course. Linguistics 30: 243-266. de Francis, John. 1950. Nationalism and language reform in China. Princeton: Princeton University Press. DeCamp, David. 1971. Toward a generative analysis of post-creole speech community. Pidginization and creolization of languages ed. by D. Hymes, 340-376. Cambridge University Press.

292

CREOLE AND DIALECT CONTINUA

den Besten, Hans, Pieter Musken, & Norval Smith. 1995. Theories focusing on the European input. Pidgins and creoles ed. by J. Arends, P. Muysken, & N. Smith, 87-98. Amsterdam: John Benjamins. DeLancey, Scott. 1991. The origins of verb serialization in modern Tibetan. Studies in Language 15: 1-23. Dictionary of English Etymology. 1966. Oxford: Clarendon Press. Dik, Simon. 1980. Studies in functional grammar. London: Academic Press. Dittmar, Norbert. 1992. Introduction. Linguistics 30:1-3. Edwards, Walter. 1992. Sociolinguistic behavior in a Detroit inner-city black neighborhood. Language in Society 21.93-115. Ellis, Rod, ed. 1994. The study of second language acquisition. Oxford University Press. Elyan, O., P. Smith, H. Giles, & R. Bourhis. 1978. R. P.-accented female speech: The voice of perceived androgyny? Sociolinguistic patterns of British English ed. by P. Trudgill, 122-31. London: Edward Arnold. Escure, Geneviève. 1978. Vocalic change in the Belizean English-Creole con­ tinuum and markedness theory. Berkeley Linguistics Society IV, 158-167. Berkeley: University of California Press. —. 1979. Linguistic variation and ethnic interaction in Belize: Creole/Carib. Language and ethnic relations ed. by Howard Giles & Bernard SaintJacques, 101-116. Oxford: Pergamon Press. —. 1981. Decreolization in a creole continuum: Belize. Historicity and varia­ tion in creole studies ed. by A. Highfíeld & A. Valdman, 27-49. Ann Arbor: Karoma. —. 1982. Contrastive patterns of intragroup and intergroup interaction in the creole continuum of Belize. Language in Society 11.2.239-264. —. 1983a. Belizean Creole: Chapter Two. Central American English ed. by John Holm, 29-70. —. 1983b. The use of creole as interlanguage by the Black Caribs of Belize. Mid-America linguistics conference papers ed. by Frances Ingemann, 271282. Lawrence: University of Kansas Press. —. 1983c. The Belizean copula: A case of semantic shift. Studies in Caribbean language ed. L. Carrington, 190-202. Trinidad: Society for Caribbean Linguistics. —. 1984. The acquisition of creole by urban and rural Black Caribs in Belize. York papers in linguistics 11: Urban creoles ed. M. Sebba & L. Todd, 95106. York, England: University of York. —. 1987. The acquisition of Putonghua (Mandarin) by speakers of the Wuhan dialect. Variation in language: NWAV-XV at Stanford ed. by K. Dennings, S. Inkelas, F. McNair- Knox, & J. Rickford, 121-136. Stanford University: Department of Linguistics.

BIBLIOGRAPHY

293

—. 1988. Topic structures as language universals. Journal of Pidgin and Creole Languages: 3.2: 159-5176. —. 1991. Gender roles and linguistic variation in the Belizean Creole commu­ nity. English around the world: Sociolinguistic perspectives ed. by Jenny Cheshire, 595-608. Cambridge University Press. —. 1991. Serialization in creole oral discourse. Development and structures of creole languages ed. by F. Byrne & T. Huebner, 179-192. Amsterdam & Philadelphia: John Benjamins. —. 1993a. Anaphoric reference in creoles and noncreoles. Language Variation and Change 5.3.233-259. —. 1993b. Focus, topic particles and discourse markers in the Belizean Creole continuum. Focus and grammatical relations in creole languages ed. by F. Byrne & D. Winford, 233-247. Amsterdam & Philadelphia: John Benjamins. —. 1993c. Gender and linguistic change in the Belizean Creole community. Locating power Vol. I, ed. by K. Hall et al., 118-131. Berkeley Women and Language Group, University of California. —. 1994. Référence anaphorique dans l'espace créolophone. Etudes Créoles. Paris: Didier Editions 17.1.68-83. —. Forthcoming. Gender and tense/aspect in Belizean Creole. Contact languages ed. by F. Mugler & J. Tent. Canberra: Pacific Linguistics, The Australian National University. Farmer, A. K. & Hamish, R. M. 1987. Communicative reference with pronouns. The pragmatic perspective ed. by Verschueren & Bertuccelli-Papi, 547-65. Amsterdam: John Benjamins. Fasold, Ralph. 1984. The sociolinguistics of society. Oxford: Basil Blackwell. Fellegy, Anna. 1995. Patterns and functions of minimal responses. American Speech. 70.2: 186-199. —. Forthcoming. Locative Deictics as Discourse Markers. University of Minnesota PhD dissertation. Ferguson, Charles. 1959. Diglossia. Word 15: 325-340. Forrest, Robert A.D. 1973. The Chinese language. London: Faber & Faber. Fromkin, Victoria & Robert Rodman 1983. An introduction to language. Third edition. Holt, Rinehart and Winston. Fudan daxue fangyan Diaocha gongzuo zu. 1958. Jiangsu Shanghai Fangyan Diaocha. (Dialects survey group of Fudan University Jiangsu and Shanghai Dialect survey). Vol.1. Shanghai: New Knowledge Press. Gass, Susan. 1980. An investigation of syntactic transfer in adult second lan­ guage learners. Research in second language acquisition ed. by Robin Scarcella & Stephen Krashen, 132-141. Rowley: Newbury House Publishers.

294

CREOLE AND DIALECT CONTINUA

Gilbert, Glenn (ed.) 1987. Pidgin and creole languages: Essays in memory of John Reinecke. Honolulu: University of Hawaii Press. Giles, H., R.Y Bourhis, & D. Taylor. 1977. Towards a theory of language in eth­ nic group relations. Language, ethnicity and intergroup relations ed. by H. Giles, 307-348. London: Academic Press. Givón 1979a. On understanding grammar New York and London: Academic Press. —.1979b. Discourse and syntax, syntax and semantics 12. New York: Academic Press. —. 1979c. Prolegomena to any sane creology. Readings in creole studies ed. by Ian F. Hancock, 3-35. Ghent: Story-Scientia. —. 1990. Syntax: A functional-typological introduction. Vol II. Amsterdam: John Benjamins. —. 1992. The grammar of referential coherence as mental processing instruc­ tions. Linguistics 30: 5-56. Görlach, Manfred. 1986. Middle English—a creole? Across historical and geo­ graphical boundaries ed. by Kratovsky & S., 329-344. Graham, Frank. 1980. The new Geordie dictionary. Newcastle upon Tyne: J. & P. Bealls Ltd. Green, Georgia. 1989. Pragmatics and natural language understanding. Hillsdale, N.J.: Lawrence Erlbaum Associates. Guy, Gregory 1988. Language and social class. Linguistics: The Cambridge survey. Vol IV: Language: The sociocultural context ed. by F. Newmeyer, 37-63. Cambridge University Press. Halle, Morris. 1964. Phonology in generative grammar. The structure of lan­ guage: Readings in the philosophy of language ed. by Jerry A.Fodor & Jerrold J.Katz. Englewood Cliffs, NJ, Prentice-Hall. 334-52. Han, Yang Saxena. 1995. A pragmatic analysis of the ba particle in Mandarin Chinese. Journal of Chinese Linguistics 23.2.99-127. Hancock, Ian. 1980. Gullah and Barbadian: Origins and relationship. American Speech 55.17-35. Hao Ran. 1984. Hao Ran Xuan Ji Vol. 2. (Hao Ran's Selected Works). Tianjin: Baihua Press. Hashimoto, Anne Yue. 1971. Mandarin syntactic structures. Unicorn 5. Princeton University. Hashimoto, Mantaro J. 1973. The Hakka dialect: A linguistic study of its phonology, syntax and lexicon. Cambridge: Cambridge University Press. Hellinger, Mariis. 1972. Aspects of Belizean Creole. Folia Linguistica (Acta Societatis Linguisticae Europaae) 6.Y2.118-35. Hengeveld, Kees. 1990. A functional analysis of copula constructions in Mandarin Chinese. Studies in Language 14-2.291-323.

BIBLIOGRAPHY

295

Holm, John A. (with A. Shilling) 1982. Dictionary of Bahamian English. Cold Spring, N.Y.: Lexik House Publishers. Holm, John. 1988. Pidgins and creoles. Vol. I. Cambridge University Press. Hopper, Paul & Sandra Thompson. 1980. Transitivity in grammar and dis­ course. Language 56: 251-99. Hopper, Paul J. & Elizabeth C. Traugott. 1993. Grammaticalization. Cambridge Textbooks in Linguistics. Hsieh-Hsin-L 1973. A new method of dialect subgrouping. Journal of Chinese Linguistics 1.1.64-92. Huang Chu-Ren.1990. A unification-based LFG analysis of lexical discontinu­ ity. Linguistics 28.263-307. Huang, C.-T. James. 1984. On the distribution and reference of empty pro­ nouns. Linguistic Inquiry 15.531-74 —. 1987. Remarks on empty categories in Chinese. Linguistic Inquiry 18.321-37. Huang, Yan. 1994. The syntax and pragmatics of anaphora. A study with spe­ cial reference to Chinese. Cambridge: Cambridge University Press. Huxley, Aldous. 1934. Beyond the Mexique Bay. New York and London: Harper & Brothers Publishers. Hvitfeldt, Christina & Gloria Poedjosoedarmo. 1995. Language use as identity across generations in Singapore. Paper presented at the Fifth Meeting of the Southeast Asian Linguistics Society, May, 1995, Tucson, AZ. Hyman, Larry. 1984. Form and substance in language universals. Explanations for language universals ed. by B. Butterworth, B. Comrie, O. Dahl, 67-86. Amsterdam: Mouton. Jakobson, Roman. 1940. Kindersprache, Aphasie und allgemeine Lautgesetze. Selected writings I ed. by R. Jakobson. The Hague: Mouton. Karlgren, Bernhard. 1915. Etudes sur la phonologie chinoise. Stockholm: Leyde Imprimerie Orientale. Karstadt, Angela. 1996. Relative markers in Swedish-American English: Evidence for a contact language phenomenon? American Speech71.1:27-43. —. Forthcoming. Swedish American English: A longitudinal study of linguis­ tic variation and identity. University of Minnesota PhD dissertation. Keenan, Edward & Bambi Schieffelin. 1976. Topic as a discourse notion. Subject and discourse topic ed. by Charles Li. New York: Academic Press. Keenan, Edward & Bernard Comrie. 1977. Noun phrase accessibility and uni­ versal grammar. Linguistic Inquiry 8: 63-100. Labov, William. 1966. The social stratification of English in New York City. Washington, D.C.: Center for Applied Linguistics. —. 1972a. Language in the inner city. Philadelphia: University of Pennsylvania Press.

296

CREOLE AND DIALECT CONTINUA

—. 1972b. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. —. 1972c. Some principles of linguistic methodology. Language in Society 1.97120. — & W. Fanshel. 1977. Therapeutic discourse. New York: Academic Press. Lalla, Barbara & Jean D'Costa. 1991. Language in exile: Three hundred years of Jamaican Creole. Tuscaloosa: University of Alabama Press. Le Page, Robert & Andrée Tabouret-Keller. 1985. Acts of identity. London: Cambridge University Press. Lefebvre, Claire. 1986. Relexification in creole genesis revisited: The case of Haitian Creole. Substrata versus universals in creole genesis ed. by Pieter Muysken & Norval Smith, 279-300. Amsterdam: John Benjamins. Leggett, Mead, & Kramer. 1988. Handbook for writers. 10th edition. Prentice Hall. Levinson, Stephen. 1983. Pragmatics. Cambridge University Press. Li, Charles N. and Sandra Thompson. 1975. The semantic function of wordorder: A case study in Mandarin. Word order and word order change ed. by Li & Thompson, 47-45. Austin: University of Texas Press. —. 1977. A mechanism for the development of copula morphemes. Mechanisms of syntactic change ed. by Li & Thompson, 419-443. Austin: University of Texas Press. —. 1981. Mandarin Chinese: A functional reference grammar. University of California Press. Linde, Charlotte. 1981. The organization of discourse. Style and variables in English ed. by Shopen & Williams, 84-113. Cambridge: Winthrop Publishers. Liu, Lanying (ed.) 1981. Grammar and rhetoric. Nanning: Guangxi Publishing House. Liu, William W. 1979. Dialect features and communication problems in Linxian. Journal of Chinese Linguistics. 7: 29-43. Lu, Shuxiang. 1980. On language. Beijing: San Lian Press. Lu, Zhongti. (with Celia Millward). 1989. Chinese given names since the Cultural Revolution. NAMES 37.3: 265-280. Manessy, Gabriel. 1989. De quelques notions imprécises (bioprogramme, sémantaxe, endogénéité). Etudes Créoles 12.2: 87-111. Martinet, André. 1955. Economie des changements phonétiques. Berne: Francke. —.1961. Eléments de linguistique générale. Paris: Armand-Colin. Matisoff, James. 1994. Following the marrow: Two parallel Sino-Tibetan et mologies. Papers from the Second Annual Meeting of the Southeast Asian Linguistics Society ed. by K. L. Adams & T. J. Hudak, 213-234. Tempe: Arizona State University

BIBLIOGRAPHY

297

Matthews, Stephen. 1996. Ditaxia and hybridization in Chinese dialect gram­ mar. Proceedings of the Fourth Conference on Pan-Asiatic Linguistics, Bangkok 1274-1283. — & Virginia Yip. 1994. Cantonese: A contemporary grammar London: Routledge. —. Forthcoming. Verb-copying constructions in Cantonese. Proceedings of the Third International Conference on Chinese Linguistics. City University of Hong Kong. McLaughlin, John. 1987. Theories of second language learning. London: Edward Arnold. McWhorter, John. 1992. Substratai influence in Saramaccan serial verb constructions. Journal of Pidgin and Creole Languages 7.1-53. Meillet, Antoine. 1921. Linguistique historique et linguistique générale. Paris: Champion. —. 1924. Les langue du monde. Paris: Champion. Milroy, Lesley. 1980. Language and social networks. Baltimore: University Park Press. Mufwene, Salikoko. 1986.Universalist and substrate theories complement each other. Substrata versus universals in creole genesis ed. by Pieter Muysken & Norval Smith, 129-162. Amsterdam: John Benjamins. Muysken, Pieter & Norval Smith. 1986. Substrata versus universals in creole genesis. Amsterdam: John Benjamins. —. 1995. Study of pidgin and creole languages. Pidgins and creoles ed. by J. Arends, P. Muysken, & N. Smith, 3 -14. Amsterdam: John Benjamins. Muysken, Pieter. 1994. Saramaccan and Haitian: A comparison. Journal of Pidgin and Creole Languages (Column). 9.2: 305-314. — & Tonjes Veenstra 1995. Universalist approaches. Pidgins and creoles ed. by J. Arends, P. Muysken, & N. Smith, 121-134. Amsterdam: John Benjamins. Nichols, Patricia 1983. Linguistic options and choices for black women in the rural south. Language, gender and society ed. by B. Thorne, C. Kramarae, & N. Henley. Rowley: Newbury House Publishers. Norman, Jerry. 1988. Chinese. Cambridge Language Surveys. Cambridge: Cambridge University Press. Oxford Dictionary of English Etymology. 1985. ed. by C. T. Onions. Oxford: Oxford University Press. Platt, John. 1991. Social and linguistic constraints on variation in the use of two grammatical variables in Singapore English. English around the world: Sociolinguistic perspectives ed. by J. Cheshire, 376-387. Cambridge University Press.

298

CREOLE AND DIALECT CONTINUA

Pollard, Velma. 1996. A singular subject takes a singular verb and other causes (?) of hypercorrection in Jamaican speech. Paper presented at the Eleventh meeting of the Society for Caribbean Linguistics. St Maarten, N.A. Population Census of Belize. 1980. Belmopan. Statistical Office: Central Planning Unit. Population Census of Belize. 1991. Major findings: Belmopan. Central Statistical Office. Postma, Johannes M. 1990. The Dutch in the Atlantic slave trade, 1600-1815. New York: Cambridge University Press. Pulleyblank, E.G. 1984. Middle Chinese: A study in historical phonology. Vancouver. University of British Columbia Press. —. 1994. The old Chinese origin of type A and  syllables. Journal of Chinese Linguistics 22.1.73-99. —. 1995. Outline of classical Chinese grammar Vancouver: University of British Columbia Press. Ramsey, Robert. 1987. The languages of China. Princeton University Press. Rickford, John. 1983. What happens in decreolization. Pidginization and creolization as language acquisition ed. by R. Andersen, 298-319. Rowley, MA: Newbury House. —. 1987. Dimensions of a creole continuum: History, texts and linguistic analy­ sis of Guyanese Creole. Stanford: Stanford University Press. — & Jerome Handler. 1994. Textual evidence on the nature of early Barbadian speech: 1676-1835. Journal of Pidgin and Creole Languages 9.22: 1-55. Roberts, Peter. 1983. Linguistics and language teaching. Studies in Caribbean language ed. by L. Carrington, 230-244. Trinidad: Society for Caribbean Linguistics. —. 1988. Second language acquisition theory and second dialect acquisition. Paper presented at the Seventh meeting of the Society for Caribbean Linguistics, Nassau, Bahamas. Romaine, Suzanne. 1988. Pidgin and creole languages. London: Longman. Ross, Claudia. 1983. On the functions of Mandarin de. Journal of Chinese Linguistics 11.2: 217-246. Ross, John. 1967. Constraints on variables in syntax. MIT Dissertation. Ruhlen, Merritt. 1991. A guide to the worlds languages. Vol. 1. Stanford University Press. Sampson, Geoffrey. 1989. Review of J. Norman Chinese. Journal of Linguistics 25: 225-235. Scarcella, Robin. 1983. Developmental trends in the acquisition of conversa­ tional competence by adult second language learners. Sociolinguistics and language acquisition ed. by Nessa Wolfson & Elliott Judd, 175-183. Rowley: Newbury House Publishers, Inc.

BIBLIOGRAPHY

299

Schiffrin, Deborah. 1987a. Discourse markers. Cambridge University Press. —. 1987b. Toward an empirical base in pragmatics. (Review article). Language in Society 16.381-396. —. 1992. Conditionals as topics in discourse. Linguistics 30: 165-198. Schlobinski, Peter & Stephan Schütze-Coburn. 1992. On the topic of topic and topic continuity. Linguistics 30: 89-122. Schmidt, Richard. 1983. Interaction, acculturation, and the acquisition of com­ municative competence: A case study of an adult. Sociolinguistics and language acquisition ed. by Nessa Wolfson & Elliott Judd, 137-174. Rowley: Newbury House Publishers, Inc. Schmied, Josef. 1991. National and subnational features in Kenyan English, English around the world: Sociolinguistic perspectives ed. by J. Cheshire, 420-422. Cambridge University Press. Schumann, John. 1978. The pidginization process: A model for second lan­ guage acquisition. Rowley: Newbury House. — & Stauble, Ann-Marie. 1983. A discussion of second language acquisition and decreolization. Pidginization and creolization as language acquisi­ tion, ed. by R. Andersen, 260-274. Rowley, MA: Newbury House. Selinker, Larry 1972. Interlanguage. International Review of Applied Linguistics 10: 209-231. Shi, Dingxu. 1989. Topic chain as a syntactic category. Journal of Chinese Linguistics 17.2: 223-254. Shields-Brodber, Kathryn. 1989. Standard english in Jamaica: a use of compet­ ing models. English World-Wide 10 1: 41-53. Singler, John 1991. Social and linguistic constraints on plural marking in Liberian English. English around the world: Sociolinguistic perspectives ed. by J. Cheshire, 544-562. Cambridge University Press. Stauble, Ann-Marie. 1978. The process of decreolization: A model for second language development. Language Learning 28.29-54. Stauble, Ann-Marie. 1980. Acculturation and second language acquisition. Research in second language acquisition, ed. by R. Scarcella & S. Krashen, 43-50. Rowley, MA: Newbury House. Stewart, William. 1964. Foreign language teaching methods in quasi-foreign language situations. Nonstandard speech and the teaching of English ed. by W. Stewart. Washington, D.C. Center for Applied Linguistics. Stochl, John. 1975. A dictionary of Central American Carib (3 vols.). Revised by R. E. Hadel and R. Zuniga. Belize-City: BISRA. Sun, Chaofen. 1995. Transitivity: The ba construction and its history. Journal of Chinese Linguistics 23.1.159-195. Taylor, Douglas. 1951. The Black Carib of British Hondura. Viking Fund Publications in Anthropology, 17. New York: Wenner-Gren Foundation.

300

CREOLE AND DIALECT CONTINUA

— 1955. Phonemes of the Hopkins (British Honduras) dialect of Island-Carib (Island-Carib, I).International Journal of American Linguistics 21:233-41. —. 1956a. Island-Carib, II: word classes, affixes, verbs, nouns. International Journal of American Linguistics 22:1-44. —. 1956b. Island-Carib, III: locators, particles. International Journal of American Linguistics 22:138-50. —. 1958. Island-Carib, IV: Syntactic notes, texts. International Journal of American Linguistics 24: 36-60. —. 1977. Languages of the West Indies. Baltimore: The Johns Hopkins University Press. Teng, Shou-Hsin. 1977. A semantic study of transitivity in Chinese. Taipei Student Book Company. —. 1979. Remarks on cleft sentences in Chinese. Journal of Chinese Linguistics 7.1: 101-113. Thomason, Sarah Grey & Terrence Kaufman. 1991. Language contact, creolization and genetic linguistics. University of California Press. Todd, Loreto. 1974. Pidgins and creoles. London: Routledge and Kegan Paul. Traugott, Elizabeth Closs & Bernd Heine. 1991. Approaches to grammaticalization (Vol. I & II). Amsterdam: John Benjamins. Tsao, Feng-Fu. 1979. A functional study of topic in Chinese. Taipei, Taiwan; Student Book Company. —. 1987. A topic-comment approach to the ba construction. Journal of Chinese Linguistics 15.1: 1-53. Valdman, Albert. 1983. Creolization and second language acquisition. Pidginization and creolization as language acquisition ed. by R. Andersen, 212-234. Rowley, MA: Newbury House. Wald, Benji. 1983. Referents and topic in discourse units. Discourse perspectives on syntax ed. by Flora Klein-Andreu, 91-116. New York: Academic Press. Wang, Li. 1954. A Chinese modern grammar. Beijing: Zhonghua. Wang, William S-Y. 1969. Competing changes as cause of residue. Language 45.9-25. —. 1989. Language in China: A chapter in the history of linguistics. Journal of Chinese Linguistics 17.2. 183-221. Williams, Jessica. 1987. Non-native varieties of English: A special case of lan­ guage acquisition. English World Wide 8.1.161-198. Winford, Donald 1993. Predication in Caribbean English creoles. Amsterdam: John Benjamins. Xu Liejong and D.T. Langendoen. 1985. Topic structures in Chinese. Language 61.1.1-27.

BIBLIOGRAPHY

301

Yip, Virginia & Stephen Matthews. 1995.I-interlanguage typology: The case of topic prominence. The current state of interlanguage: Studies in honor of William E. Rutherford ed. by L. Eubank, L. Selinker, & M. Sharwood Smith, 17-30. Amsterdam: John Benjamins. Yule, George & Terrie Mathis. 1992. The role of staging and constructed dia­ logue in establishing speaker's topic. Linguistics 30: 199-216. Zou, Ken. 1993. The syntax of the Chinese ba construction. Linguistics 31.4.715-736.

Index acrolect 6, 19, 22, 36-7,42, 51-2, 55, 57-123, 263-6, 273-5, 281 distinctiveness 111-2 longitudinal dimension 112-4 phonological features 77-9 morphological features 79-80 syntactic features 83-5 adverb(ial); 81, 99-103, 166182, 206, 209, 218, 233-7, 241-2,252-4,257,265,270, 281 see topic particles semantic change (of) 99-103 anaphora anaphoric binding 91-2, 1526, 161-3 nominal 163-5 pronominal 161-5 zero anaphora 153, 186 Arawak(an) 52, 34 aspect 49, 51-2, 61,79, 83, 103 ba 166, 182 see topic ender bǎ-construction 133, 158, 161, 166, 172-8, 198-200 see verbal preposition basilect 19, 22, 41, 48-9, 58-71,

77, 86-89, 101-2, 109, 122, 181-2,266-7,279-80 longitudinal study 112-4 Belize census 27-31 Chinese 28-9, 35 Creolesee creole, basilect emigration 26-37 ethnicity 28-36 history 25-36 languages 28-55 longitudinal dimension 1124,266 people 28-55 see Creoles, Garifuna, Maya, Mennonite, Mestizo Stann Creek 43-55 bèi -construction 161, 172, 1778, 192, 198,200 see verbal preposition bioprogram 6, 41, 59 Black Carib see Garifuna Black English (US) 7, 9, 11, 18, 70,79 bleaching 192, 274, 281, 283, 286 Cantonese(Yue) 28, 35, 130, 132, 136-8,262,275

304

CREOLE AND DIALECT CONTINUA

Catalan 10 Chinese Cultural Revolution 128-30, 141, 162, 187-208 continuum 137 dialect differentiation 137-41, 145-8, 187-221,223-261 genetic classification 131-7 history 125-30, 141-2 language see Cantonese Mandarin, Putonghua, Wu, Wuhan longitudinal dimension 187, 208 naming 129-30 post Cultural Revolution 2028 topics 151-261 see topic code-shifting 63-4 cleft(ing); pseudo-cleft 51, 84, 90,97-8,231-2,267-8,280-1 co-indexing 91, 152-3 Common language see Putonghua coverb see verbal preposition copula; pseudo copula 51, 71-84, 166,234,280-1 see da; shǐ creole continuum 3, 19-20, 39-43, 61-4, 102-3,265-6 genesis 39-41, 135 people(Creoles) 28-34, 42-48 post-creole 61-4 radical 57-63 structure see basilect use 30-2, 35-8,41-4, 52-4 creolization 2, 59, 121, 131, 134 da 103, 170,279-81 decreohzation 2, 19, 57, 62-5, 70,

275-6 definition 64 deictic 103, 160, 171, 197-8,218, 235, 269-70, 279 cataphoric 160 diglossia 64, 224, 238, 275 empty category 153 English American 42, 68, 117-20, 266-71 Belize 30-1, 35, 38,41-4, 523, 67-9 English Only movement 9 German-American 275 Singapore English 4, 8 Swedish-American 115-6, 274 focus 49, 51,83-4, 93-4, 103, 160-1,265 fossilization 275 French 41,52, 60, 134 frequency variability 76-82, 89, 277 functional analysis 14-15, 153-4, 281-5 fronting see topic gapping 156-8 Garifuna Garinagu 34 history 26, 34 language 52-5 people 31, 34, 37-8,44-9, 52, 71-2 use of creole/English 38, 49, 50-4, 71-3 women's language 54-5 gender 54-5, 70-1, 193-7 Geordie 7

INDEX Government and Binding 11, 902, 152-3 grammaticalization 77, 83, 265, 271-3,277-81 Guanhua 139 guided learning 7 Gullah 70 Guoyu 139-40 Haitian 40, 60 Hakka 138 Island Carib see Garifuna Jamaican Creole 61 jiù '166, 181,233-4,237,241, 253 Kekchi 28 Kenyan English 5 language acquisition 2-10, 63-5, 90, 114-7, 145-8,263-4,275-85 attitudes 69-73, 129,139-41 276 bias 9-10, 30, 71-2 education 286-7 identity 8, 32, 70-5 policy 36-9, 266 universals 77, 274-5, 280-6 left dislocation 51, 90, 94, 161-2, 230,266,271-3 lexical diffusion 135-6 Liberian English 5 longitudinal study see creole, Chinese Manchu 22, 128, 133-4, 139 Mandarin 130-46 markedness 4, 276-7

305 matched guise 10, 12, 276 Maya 26-32, 37, 43 Melanesian 41 Mennonite 28-9, 35-6 mesolect 19, 36, 42, 49-51, 55, 63,67 Mestizo 28-32, 34-5 Miskito 28, 33 multifunctionality 277 nautical lingua franca 40 nonstandard variety 10-1,61, 122, 264-75 objects fronted 160-1, 192,230-1 non-fronted 160-1, 230 observational adequacy see sociolinguistic methodology oral discourse 13-8, 209-21, 223-61 parataxis 114-6, 122,282 pidginization 2, 39-41, 57, 131 pingtan 148,243-61,271 possessive 156 pragmatics 14-18,263-88 pragmatic mode 122 presentation see topic Putonghua 21, 140, 185-221 Beijing Colloquial 22, 125-7, 141-4, 161-2, 166,209-21, 272-3 Literay 162, 187-208 phonology 142-4 rural variants 197-200, 206-7 Wuhan 22, 125, 144-6 rap 10 relativization 83-4, 114-6, 156,

306

CIŒOLE AND DIALECT CONTINUA

264,274,281 relexifïcation 40, 76, 81, 83, 89 repetition see topic resumptive pronoun 84-5, 230, 272 reflexivization 154 Saramaccan 58-60 serialization 59-60, 85, 282 Seselwa 59 shí/si/z 167-70,197,201,206, 218,234-8,255-6,273,280-1 sociolinguistic methodology 11-2, 155,264 Spanish 26-37, 52 style minimal responses 270-1 shifting 109-11, 206-7, 194-7, 256-8 standard 10-3, 37, 61, 65-74, 79, 81, 123,264-75 definition 66 structural hypercorrection 76, 801,89,277 substrate 40-1, 58, 266, 275-81, 286 superstrate 5, 40, 83, 277-81 Suzhou Wu see Wu Tayo 41,60 topic chain 18, 154-5, 186-90,2024,209-11,215-6,225-9,2458, 264-5 definiteness 160-1,230 definition 92-4, 153 fronting 84, 94-6, 106-20, 155-63, 191-220,230-2,239, 248-50 given/new status 160-1, 230,

280 multiple 158-60 particles see topic particles presentation 84, 94-120, 155, 166-182, 191-220,232-42, 252-6, 265-6 repetition 96-7, 106-20, 155, 163-5, 191-220,232,239 theoretical interpretation 903, 151

unit 104-6, 155, 186-93,2448 topicalization 90, 93-5, 153-6, 161 topic particles AP(adverbial particles) 166, 179-81,193,201,213,217 EP(end particles) 166, 181-2, 193,201,213,217 TE(topic enders) 178-9, 193, 213-9 TI(topic introducers) 166-78, 192-3, 197-201,206,213, 217 unguided learning 8 Universal Grammar 5, 153, 276 universals see language univer­ sals verbal preposition 172-8, 192, 197-200,206,218 with fronted objects 206, 270, 280 see topic particles (TI) see bǎ , bèi women's language see Garifuna, gender word-order 136, 156-61, 230-1

INDEX Wu (Suzhou) area 146-7 language 136, 147-8,243-61, 272-3 Wuhan area 145 normative status 236-8 use of Putonghua 145-6, 22442, 272-3