History, Society And Variation: In Honor of Albert Valdman (Creole Language Library)

History, Society and Variation Creole Language Library (CLL) A companion series to the “Journal of Pidgin & Creole ...

Author: Kevin J. Rottet | J. Clancy Clements | Thomas A. Klingler | Deborah Piston-Hatlen

325 downloads 925 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

History, Society and Variation

Creole Language Library (CLL) A companion series to the “Journal of Pidgin & Creole Languages”

Editor John Victor Singler New York

Editorial Advisory Board Mervyn Alleyne

Salikoko Mufwene

Kingston, Jamaica

Chicago

Norbert Boretzky

Pieter Muysken

Bochum

Nijmegen

Lawrence Carrington

Peter Mühlhäusler

Trinidad

Adelaide

Glenn Gilbert

Pieter Seuren

Carbondale, Illinois

Nijmegen

George Huttar

Norval Smith

Dallas

Amsterdam

John Holm Coimbra

Volume 28 History, Society and Variation In honor of Albert Valdman Edited by J. Clancy Clements, Thomas A. Klingler, Deborah Piston-Hatlen and Kevin J. Rottet

History, Society and Variation In honor of Albert Valdman

Edited by

J. Clancy Clements University of New Mexico

Thomas A. Klingler Tulane University

Deborah Piston-Hatlen Kevin J. Rottet Indiana University

John Benjamins Publishing Company Amsterdam/Philadelphia

8

TM

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 History, society and variation : in honor of Albert Valdman / edited by J. Clancy Clements ... [et al.]. p. cm. (Creole Language Library, issn 0920–9026 ; v. 28) Includes bibliographical references and indexes. 1. Creole dialects. I. Valdman, Albert. II. Clements, J. Clancy. III. Series. PM7831.H57 2006 417’.22--dc22 isbn 90 272 5250 5 (Hb; alk. paper)

2006048366

© 2006 – 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 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

CONTENTS Introdution J. Clancy Clements, Thomas A. Klingler, Deborah Piston-Hatlen, and Kevin J. Rottet

1

SECTION ONE: HISTORY Louisiana Creole at the periphery Thomas A. Klingler and Nathalie Dajko

11

Using and interpreting historical texts to analyze the formation and development of creole languages Marie-Christine Hazaël-Massieux

29

Lexical aspects of French and Creole in Saint-Domingue at the end of the eighteenth century Pierre Rézeau

47

The lexicalization – grammaticalization continuum J. Clancy Clements

77

Creole transplantation: A source of solutions to resistant anomalies John McWhorter

103

SECTION TWO: SOCIETY Creoles, capitalism, and colonialism Derek Bickerton

137

A curiosity of Mauritian Creole: Numerical slang Robert Chaudenson

153

Theoretical and practical conditions for the emergence of a koine among French-lexified creole languages Jean Bernabé

163

French in Haiti: Contacts and conflicts between linguistic representations Corinne Etienne

179

vi

CONTENTS

SECTION THREE: VARIATION Albert Valdman on the development of creoles Salikoko S. Mufwene

203

Diatopic variation in Haitian Creole Annegret Bollée and Pamela Nembach

225

Interrogative pronouns in Louisiana Creole and the Multiple Genesis Hypothesis Kevin J. Rottet

235

Gender in French creoles: The story of a loser Ingrid Neumann-Holzschuh

251

Tense, mood, and aspect and the Deixis Ordering Principle Anand Syea

273

Index

297

INTRODUCTION J. Clancy Clements University of New Mexico Deborah Piston-Hatlen Indiana University

Thomas A. Klingler Tulane University Kevin J. Rottet Indiana University

This collection of papers honors our colleague Albert Valdman, the creolist. For over 40 years, Albert has dedicated immeasurable time and energy to the study of Haitian Creole, the French-based creoles, and to pidgin and creole studies in general. After starting out in chemistry at the University of Pennsylvania, Albert was steered into linguistics by the eminent French linguist and phonetician, Pierre Delattre. After completing his B.A. at the University of Pennsylvania, he continued his studies at Cornell University, where the wellknown creolist Robert A. Hall was his dissertation director. Albert, however, did not discover his interest in Haitian Creole through his connection with Hall, but rather through his relationship with Hilde Wieners, the woman who would later become his wife. She had just returned from making a documentary film in the Caribbean, including Haiti, and encouraged him to work on Haitian Creole. It was primarily a desire to address the pressing need for the development of pedagogical materials and reference resources that led Albert to creole studies. Albert was hired at Indiana University in 1960 after working two years with the Foreign Service Institute and one year as an assistant professor at Pennsylvania State University. At Indiana, Albert held a joint appointment in the Department of French and Italian and the Department of Linguistics. During his long tenure at this university, he has been admired for his dedication to his departments and to the development of French studies in the United States. At his retirement in 2004, Albert had served Indiana University with distinction for more than four decades. No other professor we have known surpasses him as a model of professional service. Albert has two main research areas: foreign language teaching/learning and pidgin and creole studies. The present volume celebrates his contributions in the second area. (A Festschrift has already appeared [Gass, Bardovi-Harlig, Sieloff, and Walz 2002] in recognition of his contributions to the first area.) His major contributions to Haitian Creole are the textbook Ann Pale Kreyòl, and the dictionaries Haitian Creole-French-English Dictionary (with S. Yoder et al.) and A Learner’s Dictionary of Haitian Creole (with R. Jean-Baptiste and C. Pooser). He has also written on other varieties of French in the Caribbean, publishing Saint-Lucian Creole Basic Course (with L. Carrington) and Dictionary of Louisiana Creole (with T. Klingler, M. Marshall, and K. Rottet). He

2

HISTORY, SOCIETY, AND VARIATION

is the author of numerous books and over 200 articles, as well as the editor or co-editor of many volumes. Currently, he is working on the final editing of a new Haitian Creole-English dictionary and is involved in a project to produce a bilingual Haitian Creole-French dictionary for use by school children in Haiti. He is also overseeing a major project bringing together the leading scholars on Louisiana French to produce the first scholarly Dictionary of Cajun French. Beyond Albert’s commitment to research, he has also actively participated in the national and international academic community as an officer in many of the major scholarly associations devoted to linguistics and as an organizer of more than fifteen workshops and conferences. Not only did he direct the teaching of first- and second-year French, creating a model instructional program for graduate teaching assistants, but he also created the first Ph.D. program in French linguistics in the nation and taught actively in the graduate programs of the Departments of French and Italian and Linguistics, directing over 40 dissertations. His mentoring of students and colleagues has brought success and distinction to many who readily acknowledge his central role in their professional lives. One of the greatest honors Albert has received is the title of Commandeur dans l’Ordre des Palmes Académiques, presented to him October 26, 2000, by the French government. This award, first established by Napoleon Bonaparte in 1808, honors supreme achievements in teaching, scholarship and research. Albert Valdman is passionate about the academic profession. He believes in its tradition of research, in the formation of new generations of scholars, and in pursuing the best ways of learning and teaching. Indiana University and the profession at large can look at his great career only with admiration and gratitude. 1. The volume Because of the very nature of the object of research, the study of pidgins and creoles involves the study of interfaces, whether they be intercomponential, sociolinguistic, sociohistorical, or historico-typological. The present volume focuses on three areas involved in the formation and/or the development of pidgins and creoles: history, society, and variation. Although the studies in each area have a particular focus, they of course include aspects of the other two, as well. 2. History This section focuses on historical perspectives on creole languages. The study by Thomas A. Klingler and Nathalie Dajko uses forms from marginal Louisiana French varieties to better understand their evolution. Specifically, the article reports on language documentation efforts conducted in three peripheral parishes where Louisiana Creole (LC) is part of the linguistic makeup, but where that language or other French-related varieties, though once

INTRODUCTION

3

widespread, are today restricted to very small numbers of speakers, most of them elderly, and can thus be expected to disappear when the current generation of speakers is gone. The parishes examined are Natchitoches, Plaquemines, and St. Tammany. The authors also consider Louisiana Creole data from a fourth peripheral area, Mon Louis Island, Alabama, investigated by Marshall (1991), where the language has now died out. Although the research is still in progress, the work carried out thus far allows them to sketch the linguistic situation of each of the research sites and to demonstrate how language data from such peripheral areas can help to gauge the authenticity of early LC texts and gain further insight into the historical development of the language in Louisiana. Marie-Christine Hazaël-Massieux’s contribution addresses the problems inherent in using historical texts for documenting the development of creoles. The author focuses on three problems in particular: changes in the spoken language that are recorded in writing only much later, the questionable authenticity of the texts, and the multiple sources of variation that they show (e.g., the absence of an orthographic norm, pronunciation differences based on region, and pronunciation differences based on style). She emphasizes the necessity of taking such problems into account when interpreting the early attestations of creole found in these texts. Using examples from the French Antilles, she shows, however, that as long as proper caution is exercised, such early texts provide invaluable information about the development of the French creoles and the historical relationships between different varieties, as well as about the spoken French of the seventeenth century. A diachronic perspective on the lexicon of Caribbean French Creole is offered in the article by Pierre Rézeau. He comments on an unpublished anonymous manuscript, written around 1800, that contains numerous observations on the language of the period in Saint-Domingue and the Antilles. A considerable number of these observations concern flora and fauna. Around 100 of the most interesting terms appear in this commentary. For each item, a brief characterization is given, according to a typology of lexical variation: terms derived from French but involving a semantic shift, terms shared with French but markedly different in frequency, and terms that were borrowed from regional dialects, from an African language, or from an indigenous language of the Americas. The original entries written by the anonymous author of the text are provided and complemented by additional observations on the history of each term. The commentary includes carefully researched information about each lexeme's earliest attestation, comparisons with cognate terms from other places in the colonial world, and etymological notes. These data provide a highly interesting historical snapshot of a specialized part of the lexicon of Haitian and Caribbean French. In his study on the lexicalization-grammaticalization continuum, J. Clancy Clements argues that there are two separate but interdependent, bidirectional

4


clines in the grammaticalization-lexicalization process: (i) one whose endpoints are the semantics/pragmatics of a lexical item on the one hand, and the function of a grammatical marker on the other, and (ii) another cline whose endpoints are the unbound form on the one end and the affix on the other. Based on a variety of examples, including numerous examples from pidgin and creole languages, the author proposes that developments along both these clines are chaotic in nature, involving the notions of regularity, iteration, and unpredictability. A chaotic process can be defined as a process that under iteration shows a high degree of regularity, though it is ultimately unpredictable. Applied to the grammaticalization-lexicalization clines, although the path of grammaticalization (lexical item => grammatical marker; unbound form => affix) has been shown to be highly regular, it is ultimately unpredictable and not always unidirectional. Instances of movement towards lexicalization and free forms are discussed. Clements proposes the distinction between major and minor pathways along the meaning and the form clines in order to capture the observations that bidirectionality is found on both clines but that on both one direction is more common than the other. John McWhorter holds that in many cases the crucial aspect of pidgin formation was not the importation of large quantities of laborers, but rather the importation of an already-formed pidgin. He notes that creole genesis theory tends to focus on polygenetic scenarios, wherein a creole developed in the place it is spoken today as the result of sociohistorical factors specific to that location. Yet, he says, a great deal of linguistic and historical evidence shows that most creoles were brought to their current locations from elsewhere, or that a creole's emergence was profoundly affected by the previous existence of another contact language already spoken on site. His article demonstrates that this phenomenon allows principled explanations of various long-standing anomalies to creole genesis theory, such as the mysterious absence of creole continua in French plantation colonies and the paucity of Spanish-based creoles, as well as newer questions such as why a creole emerged in Hawaii among children who were being taught in English in school. 3. Society In his contribution, Derek Bickerton argues that the linguistic effects of the market forces in the history of creole formation have been under-appreciated. Noting that all plantation economies required a period of rapid expansion in order to recoup the considerable initial capital investment they involved, he points out that this expansion entailed rapid importation of large quantities of labor as soon as infrastructure was available, which in turn entailed a high proportion of linguistically naïve workers in the labor force during and immediately after the expansion stage. He argues that the large imbalance between the naïve workers and ‘old hands’ (i.e., those slaves who might have had reasonable access to the superstrate) ensured that, regardless of whatever linguis-

INTRODUCTION

5

tic developments took place in the earliest years, a structureless jargon must have emerged in all but the less successful economies and served as input for subsequent creolization. The chapter by Robert Chaudenson discusses a microsystem of slang in Mauritian Creole involving the names of numbers that can replace nouns and, in some cases, even verbs: sa ène niméro set (lit., ‘he’s a number seven’) ‘he’s a thief,’ get so kenz (lit., ‘look at her fifteen’) ‘look at her breasts.’ It is shown that the microsystem is used primarily in the urban areas of Mauritius, though some of its elements have also spread to the general population. Chaudenson first compiled an inventory of this lexicon while conducting field research in Mauritius some thirty years ago, but the inclusion of many of its elements in the Dictionary of Mauritian Creole (Baker and Hookoomsing 1987) attests to its continued vitality. In this study, Chaudenson gives a compilation of the major elements of this lexicon (including phonological forms and meanings) and examines how it functions semiologically, linguistically, and sociolinguistically. He also compares it to a similar but more restricted use of numbers in Reunion Creole and proposes a tentative hypothesis about its origin. Jean Bernabé argues that mutual intelligibility among creoles is a function of the dialectal distance that separates them and of the speakers’ knowledge of the lexifier. At present, he notes, creole-speaking researchers, rather than the naïve speakers, are in a position to evaluate the relationship between learning and the linguistic diversity of creoles. Bernabé claims that mutual intelligibility is becoming increasingly possible among speakers of different French-based creoles because of schooling in French, which makes French the common language of creole speakers from various nations. The fact that French is the common denominator of these creoles and that they are structurally similar increases the likelihood of mutual intelligibility. He believes that a significant rise in the number of communicative exchanges between creole speakers from various lands will be necessary to make this possibility a reality. Moreover, the creation of a veritable “linguistic market” will, over time, allow a new intercreole dynamic to develop. In this light, Bernabé examines institutional changes being brought about by the increased use of creole languages in the school setting. Corinne Etienne examines what it means to define a variety of French. Defining such a variety involves not only showing its specificity compared to other French varieties but also reporting on its speakers’ linguistic representations (Robillard 1993). Linguistic representations, or speakers’ categorizations and judgments of linguistic uses, constitute an essential element in any lexicological study, because they guarantee that the users will recognize themselves in the uses observed and described. In her article, she investigates how French in Haiti is defined by its users and whether they perceive it as a specific variety of French using data from semi-structured interviews. Twenty French-Creole bilingual Haitians were asked to define good French, good Creole, whether

6


they recognized the existence of a Haitian variety of French, and how they would define it. Findings indicated that only half the participants recognized French in Haiti as a local variety of French. The definition of Haitian French appeared to be inextricably linked to Creole. Participants’ epilinguistic discourses revealed various conflicts between linguistic uses marked by FrenchCreole contact and participants’ French and Creole ideal norms. 4. Variation Salikoko Mufwene provides a retrospective view of Albert Valdman’s work, incorporating it into recent discussions on the development of the French-lexified creoles. Mufwene notes that key researchers fail to utilize properly Valdman’s work in the current debate. This oversight, he notes, is shocking because the position they advocate is closely related to, if not the same as, what Valdman (1977) submits, adducing supportive data from the determiner system of French creoles of both the American-Caribbean and the Indian Ocean regions. Focusing on Valdman (1977, 1983, 2002), Mufwene highlights how this work anticipates the current debate on creole development and must be read by anybody who wishes to contribute new ideas to our understanding of the subject matter of the development of French creoles in particular. He argues that Valdman counts among the first of today’s creolists to have articulated positions consistent with the socio-economic histories of the relevant population of speakers. As the point of departure for their study, Annegret Bollée and Pamela Nembach take Albert Valdman’s (1987) article, “Le cycle vital créole et la standardisation du créole haïtien” (‘the creole life cycle and the standardization of Haitian Creole’), in which he pointed to the importance of dialectal variation in Haiti and provides some examples showing that a “central form” occurring in the region of Port-au-Prince is opposed to “peripheral variants” in the north or the south (111). Valdman cites Orjala (1970) who set up the hypothesis that there are three discernible dialects in Haiti: northern, central and southern. Bolleé and Nembach use Fattier's (2000) comprehensive description of Haitian Creole to argue that it is possible to verify this hypothesis by trying to identify dialect boundaries, possibly bundles of isoglosses that separate the southern peninsula, or the northern peninsula, from the rest of the country. By comparing the results obtained from the maps of Fattier’s Atlas linguistique d'Haïti with the picture presented by the Atlas linguistique et ethnographique de la Réunion, the authors raise the question about whether the absence of clearly delimited dialects (as, e.g., in Italy or Germany) could be typical of territories in which creoles have emerged. Kevin J. Rottet’s chapter examines the survival and role of qui meaning “what?” in Louisiana Creole (LC), with a view to shedding light on the pronominal system of that variety. The complexity and incomplete documentation of LC interrogatives has been pointed out in the literature and is potentially

INTRODUCTION

7

problematic, as creole interrogative systems have been of interest in recent years in addressing a variety of issues regarding creolization. Rottet addresses two such issues in this article, of which the first concerns the use of a focus particle in creole interrogatives. He then shows how interrogative data from LC can be brought to bear on a recent scholarly debate concerning the number of geneses of Creole in Louisiana and the role of Haitian Creole in the development of LC. Ingrid Neumann-Holzschuh investigates the representation of gender in French-, Spanish-, and Portuguese-based creoles. It is generally accepted that Romance creoles do not have grammatical gender, although it is possible in all varieties to mark sex differences with animate nouns by lexical or morphological means. As a grammatical category, gender seems to be a weak category. For example, it is strongly reduced in nonstandard varieties of overseas French, and it has been given up almost entirely in creolization. Due to prolonged contact with the lexifier language, however, the gender distinction may be reintroduced into creoles, as has been the case, for instance in Guadeloupean Creole and Louisiana Creole. New light is shed on this largely familiar topic by taking recent findings from linguistic typology into account. Typological research into gender differentiation starts from the basic assumption that gender has to be analyzed in the broader framework of nominal classification. This perspective raises the question whether creoles possess specific strategies to classify nouns, a problem directly linked, for example, to marking principles of the number category. The final chapter of the volume, by Anand Syea, examines clause structure and tense, mood, and aspect (TMA) markers in creoles. Members of the auxiliary system follow a general ordering pattern in creole languages in which tense markers linearly precede aspect and mood markers, and mood markers in turn linearly precede aspect markers. It is interesting that creole languages display this pattern with very little variation and naturally, questions arise as to why we get this word order pattern and why there is so little variation across the creole languages. This study proposes that the ordering of TMA markers in Mauritian and other creoles is independently determined by principles of syntax. In particular, it will assume something akin to Cinque’s (1999) Universal Base Hypothesis, which claimed that all languages have the same fundamental hierarchical structure and each projection is associated with a semantic notion. Syea argues that the ordering observed falls out from syntactic considerations. The article also addresses how current syntactic ideas about clausal structure and functional projections can provide a plausible alternative to semantic and historical explanations of TMA ordering. Syea speculates on the interaction between input and output in creole situations and draws comparisons with first and second language acquisition. *********

8


The editors would like to acknowledge support from the Department of French and Italian, Department of Linguistics, College of Arts and Sciences, and Office of the Chancellor at Indiana University, Bloomington, and from the Indiana University Office of International Programs and Office of the Vice President for Research. Many thanks are also due to Helen Hathorn and Amanda Edmonds for their help with preparation and formatting of the final manuscript.

References Baker, Philip & Vinesh Hookoomsing. 1987. Diksyoner kreol morisyen / Dictionary of Mauritian Creole / Dictionnaire du créole mauricien. Paris: L’Harmattan. Cinque, Guglielmo. 1999. Adverbs and functional heads: A cross-linguistic perspective. Oxford: Oxford University Press. Fattier, Dominique. 2000. Contribution à l’étude de la genèse d’un créole: L’Atlas linguistique d’Haïti, cartes et commentaires. Paris: Presses Universitaires du Septentrion. Gass, Susan, Kathleen Bardovi-Harlig, Sally Sieloff Magnan, & Joel Walz (eds). 2002. Pedagogical norms for second and foreign language learning and teaching: Studies in honour of Albert Valdman. Amsterdam: John Benjamins. Marshall, Margaret M. 1991. “The Creole of Mon Louis Island, Alabama and the Louisiana connection”. Journal of Pidgin and Creole Languages 6. 7387. Orjala, Paul R. 1970. A dialect survey of Haitian Creole. Doctoral dissertation, Hartford Seminary Foundation. Robillard, Didier de. 1993. Contribution à un inventaire des particularités lexicales du français de l'île Maurice. Vanves, France: EDICEF. Valdman, Albert. 1977. “Créolisation sans pidgin: Le Système des déterminants du nom dans les parlers franco-créoles”. Langues en contact: Pidgins-Creoles—Languages in contact, Meisel (ed) 1977 (Tübingen: Gunter Narr). 105-136. ----------. 1983. “Creolization and second language acquisition”. Pidginization and creolization as language acquisition, Andersen (ed) 1983 (Rowley, MA: Newbury House). 212-234. ----------. 1987. “Le cycle vital créole et la standardisation du créole haïtien”. Etudes Créoles 10. 107-125. ----------. 2002. “Comment distinguer la créolisation du changement linguistique ordinaire? ” Etudes Créoles 25. 123-141.

SECTION ONE:

HISTORY

LOUISIANA CREOLE AT THE PERIPHERY ∗

Thomas A. Klingler & Nathalie Dajko Tulane University

ABSTRACT: This paper reports on efforts to document Louisiana Creole in three peripheral areas where the language is in immanent danger of disappearing. We show that, beyond helping to sketch a more complete picture of the current state of the language, such work can shed new light on its historical development. Focusing on three features—the form of the verb “to have” (gen or ganyen), the marking of gender on possessive determiners, and the presence or absence of verbs showing two forms that are in complementary distribution according to grammatical context—we find that these data, in combination with data reported by Marshall (1991), reveal basilectal forms that are typical of nineteenth-century Creole texts but are no longer found today in the central Creole-speaking zone around Bayou Teche. We conclude that the peripheral zones have preserved older forms of the language and thus provide a window onto earlier stages of its development. 1.

Introduction It is widely recognized that Louisiana’s vernacular varieties of French are in danger of extinction. While language revival efforts may have some chance of succeeding in places where these varieties are still spoken by a significant proportion of the population, their disappearance is all but assured in others, in particular those that lie at the geographic periphery of the state’s ‘French triangle.’ Most of these regions have not been studied extensively, if at all, by linguists. Yet the documentation and description of the varieties in these peripheral areas are important, first, because they can afford us a more complete understanding of the linguistic situation in francophone Louisiana, but also because they can provide comparative data that offer clues to the development

* This is a revised version of a paper that the authors presented at the meeting of the Society

for Pidgin and Creole Linguistics in January of 2004. The research on which this study is based was funded in part by grants from the American Council of Learned Societies, the Cane River National Heritage Area, the Friedrich Stoll Endowed Scholars Fund, and the Department of Anthropology at Tulane University. We gratefully acknowledge the generous support from each of these sources. We would also like to thank Clancy Clements and Kevin Rottet for their valuable comments on an earlier draft of the study. Any errors that remain are our responsibility.

12


of Louisiana Creole (LC) and regional French and to possible relationships between them. This study reports on language documentation efforts conducted in three peripheral parishes where LC is part of the linguistic make-up and where that language or other French-related varieties, though once widespread, are today restricted to very small numbers of speakers (most of them elderly) and can thus be expected to disappear when the current generation of speakers is gone. The parishes we examine are Natchitoches, Plaquemines, and St. Tammany. We also consider LC data from a fourth peripheral area, Mon Louis Island, Alabama, investigated by Marshall (1991), where the language has now died out (see Figure 1). Although our research is still in progress, the work accomplished thus far allows us to sketch the linguistic situation of each of our research sites and demonstrate how language data from such peripheral areas can help us to gauge the authenticity of early LC texts and gain further insight into the historical development of the language in Louisiana. 2. Description of the linguistic situation LC is often found in coexistence with other French-related varieties, and the areas where we conducted fieldwork are reflective of this: In two of the three areas, we encountered speakers of another variety of French in addition to speakers of LC. In Natchitoches Parish, Francophones were found in the Cane River Valley, with LC spoken in the area between Natchitoches and Cloutierville known as the Isle Brevelle, and Louisiana Regional French (LRF) spoken in the surrounding towns of Cloutierville and Gorum. 1 All the speakers of LC in this area identified themselves ethnically as Creole. Although five of the LRF speakers were what could otherwise be classified ethnically as Creole, speakers of LRF generally rejected that label in favor of the label ‘French,’ a term they preferred when referring to their language as well. We emphasize that our classification of speakers’ language varieties as LC or LRF is based purely on our linguistic analysis of their speech, not on their own labeling of themselves or their variety of French. 2 1

We use the term Louisiana Regional French to refer collectively to all non-Creole and nonstandard varieties of French in Louisiana. The term is meant to encompass Cajun French, which we define as the variety spoken by the descendants of the Acadians and by those groups that have been assimilated by them, commonly known as Cajuns. Because of its strong association with this ethnic group, however, we consider the term ‘Cajun French’ to be inappropriate to refer to the French spoken by groups who do not consider themselves Cajuns and who live in areas that did not receive significant numbers of Acadian settlers. For the French of such groups we prefer the broader, ethnically neutral label ‘Louisiana Regional French.’ 2 For linguistic descriptions of LC, see Lane (1935), Neumann (1985), Klingler (2003a), and Valdman and Klingler (1997). For descriptions of LRF (generally labeled Cajun French in the studies in question), see for example Rottet (2001), Guilbeau (1950), and Papen and Rottet (1997). For a discussion of the problem of interpreting language labels in Louisiana, see Klingler (2003b).

LOUISIANA FRENCH ON THE PERIPHERY

13

In Plaquemines Parish, LC was found to be in coexistence with two other French varieties. The dialect spoken in the southern end of the parish, near the town of Venice, bears many similarities to the LRF of nearby Lafourche and Terrebonne parishes, and indeed, there is evidence that it was imported from that area near the turn of the century. One hallmark feature that these varieties share is the frequent replacement of the voiced palatal fricative [¹] with a voiceless glottal fricative [h], so that, for example, j’aime is often pronounced [h°m]. We located two speakers of this variety, one of whom identified herself as Indian. To the north, on the west bank of the river in the towns of Buras and particularly Port Sulphur, and on the east bank near Pointe à la Hache, a slightly different variety of LRF was attested. It also bears many similarities to LRF, but nonetheless maintains a distinctive regional character that sets it apart from other dialects of LRF, as in the pronunciation of a uvular [ë] rather than the apical [r] that is typical of most of francophone Louisiana. This variety was spoken by people of mixed race and by a single Black respondent. All speakers of this variety referred to both themselves and their language as French, adamantly rejecting the term ‘Creole’ both as a language label and as an ethnic label. The only speaker of LC in this parish lived on the east bank, in the town of Davant. Given that the parish formerly boasted a number of large plantations, it is likely that LC was once more widely spoken there. In recent memory, however, it was spoken only on the east bank, in the area of Davant/Pointe à la Hache. The third research area, St. Tammany Parish, differs from the others in that only LC is attested. Its speakers are centered around the town of Bayou Lacombe and the Bayou Liberty area of Slidell, both on the north shore of Lake Pontchartrain. All St. Tammany speakers of LC identified themselves ethnically as Creole. The U.S. census records illustrate quite clearly the rapid advance of language shift in all three of our research areas: in the 2000 census, only 0.9% of the population in Natchitoches Parish, 1.3% of the population in St. Tammany Parish, and 3.5% of the population in Plaquemines Parish claimed to speak ‘French’ (a term that encompasses the labels ‘Patois’ or ‘Cajun’) or ‘French Creole’ (U.S. Census Bureau: Census 2000). The small number of speakers we were able to locate is reflective of these numbers: in Cane River we found 13 speakers, only two of whom spoke LC; in St. Tammany Parish, we interviewed 7 Francophones, all speakers of LC; and in Plaquemines Parish, we were able to interview 20 speakers, only one of whom, as previously noted, was a speaker of LC. Those interviewed were all over fifty, and fluent speakers were in their seventies and eighties. Of the three areas, however, the Cane River Valley appears to be in the most advanced stage of language shift. Speakers here no

14


Figure 1: Map of Louisiana longer belong to a real French language community. Many of the interview subjects were unaware of any others in the region who spoke the language, this despite the fact that several of them knew each other; when they met, they always spoke English. The youngest speakers in this region who demonstrated any fluency were also ten years older than their counterparts in the other regions. In St. Tammany Parish, the youngest interview subject was in her fifties, and in Plaquemines Parish several people in their forties demonstrated some familiarity with the language (LRF). In both Natchitoches and Plaquemines parishes, where more than one variety of French was attested, LC was in further decline than the local variety of LRF. Fewer speakers of LC could be located, and those found were generally


15

less fluent than speakers of LRF. Two of the four LC speakers in Cane River could best be characterized as semi-speakers, and the remaining two as ‘rememberers,’ since they were only able to produce isolated words and phrases of a language they had heard in their youth. The single speaker of LC in Plaquemines Parish seems to be the last representative of a language community that has otherwise died out or dispersed. According to Marshall (1991:74), non-Creole varieties of French died out in the Mobile, Alabama area in the early twentieth century. LC survived somewhat longer, however, and in the 1980s Marshall was able to interview the last two living speakers on Mon Louis Island. Aged 89 and 92 at the time and now deceased, these women formed part of an old and well-established community of Creoles of color. In none of the areas considered here is French being passed on to younger generations in the home. In Plaquemines Parish, though LRF seems to be particularly strong in and around Port Sulphur, which boasts a large number of speakers all of whom speak the language with each other on a daily basis, it still is a language reserved for the elderly. Although many fluent speakers are the primary caretakers of their grandchildren during the day, they do not use the language to communicate with the children. In St. Tammany Parish, a drive to pass the language on does exist—the Creoles sans Limites “Creoles without Limits” group holds classes to teach both children and adults Creole—but only a small number of people take part in the classes, and even the strongest speakers in the group can only be classified as semi-speakers. 3. Linguistic findings As we have noted, the areas under consideration here are peripheral to the main French-speaking region of Louisiana. In three of them, the shift to English and the concomitant abandonment of French and LC is well advanced, and in the fourth, Mon Louis Island, these processes are now complete. If we consider only LC, the peripheral status of these areas is accentuated, because the demographic center for LC today—that is, the place where the language has the largest number of speakers and enjoys the greatest vitality—is a small and relatively compact zone centered around Bayou Teche. In addition to being peripheral, each of the three areas is considerably isolated from the rest of francophone Louisiana. These characteristics, though they do not bode well for the future of LC in these parishes, potentially render the data gathered there significant for the understanding of the historical development of the language. Before proceeding further, it will be useful to examine more closely the sense in which these areas may be considered peripheral and isolated, as well as the reasons why this status makes data collected there especially interesting. It is important to emphasize first that, in characterizing these areas as peripheral with regard to the demographically ‘central’ zone in the Bayou Teche

16


region, we are not suggesting that LC had its origin in the Bayou Teche region and was exported from there to the outlying areas. Klingler (2003a) argues that LC in fact arose on the Mississippi River plantations surrounding New Orleans and spread outward from there, a scenario that would also make the LC of the Teche region peripheral in relation to the point of origin of the language. 3 Once exported to the Teche region, however, LC became solidly implanted and thrived to a greater degree than elsewhere. While the language has now disappeared from its place of origin in the New Orleans vicinity, it is in the Teche region that the most numerous and, on average, the youngest LC speakers are to be found. In addition, this region is located in the heart of Louisiana’s ‘French triangle,’ a group of twenty-odd parishes in the southern part of the state where French language and culture continue to play significant roles. LC speakers there have long lived in close contact with speakers of LRF, who are present in greater numbers. Many people speak both varieties, and even those who only speak one typically understand the other. 4 The linguistic situation of French-related varieties in the Teche area has thus remained fairly dynamic: LC has, at least until recently, continued to be used as a vernacular among substantial numbers of speakers, and contact between LC and LRF has persisted until the present day. In contrast, the situation has remained much less dynamic in LC-speaking areas that are remote from the Teche region and lie outside the French triangle. Here the shift to English is much more advanced and French-related varieties have long since ceased to be used in any capacity by all but a tiny handful of the population. As a result, intensive contact between speakers of LC and other varieties of French has not continued to the present day. This makes it less likely that the LC of these areas has fully participated in the changes that have affected the LC of the central zone, especially those brought about by its contact with LRF. It is for this reason that the LC spoken in these areas holds the possibility of providing a glimpse into earlier stages of the language. 5 These findings can then fruitfully be compared with historical representations of LC to gain further insight into the language’s development.

3

Speedy (1994, 1995) proposes that LC had two points of origin, an early one along the Mississippi and a later one in the Bayou Teche region. Klingler (2000) argues against this view, citing demographic conditions in the Teche region that were not propitious to creole development, as well as linguistic and demographic evidence suggesting that LC was exported to the Teche region from the Mississippi region. 4 LC and LRF are not necessarily mutually intelligible in the case of speakers living in areas where only one variety is spoken. Such speakers sometimes report having difficulty understanding speakers of the other variety. 5 Cf. the areal norms of ‘spatial linguistics’ (Bartoli 1945), according to which older linguistic forms are typically to be found in isolated and peripheral areas (see Chambers and Trudgill 1998:168 and G. Hazaël-Massieux 1990:96).


17

As it happens, LC is fairly well documented in writing from the mid nineteenth to the early twentieth century (see especially Neumann-Holzschuh 1987), so we do have data on earlier—though not on the earliest—stages of the language. 6 However, as is well known, the reliability of early representations of creole languages is often in question. The case of early LC texts is typical in this regard. First, most of them were written by whites who were describing a highly stigmatized language associated primarily with the black population. Second, though some authors were native speakers of LC (e.g., Mercier 1880), this cannot always be assumed to be the case and, therefore, we cannot be sure just how well the authors knew the language they were describing. Finally, many early texts were written as satire, often with the intent of ridiculing Creole speakers, which means that we may be dealing with a caricature of the language rather than with an accurate representation of it (see HazaëlMassieux, this volume, for an examination of the many difficulties that arise in interpreting historical creole texts). 7 It is here, then, that comparative linguistic data from isolated and peripheral areas can be especially useful: to the extent that forms differing from those of the central area are found both in linguistic isolates and in historical texts, the authenticity of these textual representations is confirmed and we are in a better position to draw conclusions about historical developments in the language. We have selected for study three features that are typical of the LC of nineteenth-century texts, but that are either rare in the LC of the central zone (as described by Broussard 1942; Lane 1935; Neumann 1985) or have disappeared from it altogether. The features are: (1) the use of the form ganyen vs. the form gen for the verb “to have”; (2) the presence or absence of gender distinctions on possessive determiners (mo, to, so vs. fem. ma, ta, sa “my, your, his /her/its”); and (3) the use of invariant verb forms vs. the use of a class of verbs having both a long and a short form that are in complementary distribution (e.g., mon manje “I eat/ate” vs. mon manj “I eat”/mo manje “I ate”). Each of these is considered in turn.

6 The first description of LC dates from 1807 and is very cursory (Robin 1807). There is then a hiatus of thirty-nine years before the next documented text in the language appears (Cigale et la fourmi 1846, reprinted in Neumann-Holzschuh 1987:24). For eighteenth-century attestations of what may have been precursors to LC, see Klingler (2003a:25-46). 7 Neumann-Holzschuh includes numerous examples of such satirical texts in her collection (1987:98-113, 162-186). All were written during Reconstruction and are meant to ridicule newly freed African Americans. As Neumann-Holzschuh notes, “written in a deformed Creole, these satirical texts do not represent authentic specimens of nineteenth-century Creole; their interest is thus above all political and sociohistorical [écrits dans un créole défiguré, ces textes satiriques ne représentent pas des spécimens authentiques du créole du XIXe siècle; leur intérêt est donc avant tout politique et socio-historique]” (1987:25).

18


3.1 The verb “to have” In nineteenth-century LC texts, the equivalent of the verb “to have” most commonly takes a form of the French verb gagner, which, in contemporary French, means “to win” or “to earn.” Common orthographies in these texts include gagnin, gagné, gaingin, and gaigné. The shorter form gen, typically spelled gain or gaien, is also found, but it is much rarer (Neumann-Holzschuh 1987:19), and only the longer form appears in the satirical texts published in Neumann-Holzschuh (1987). Today, in contrast, gen is found to the total exclusion of gagner in most areas where LC is spoken, most notably in the central zone. The longer form is still attested, however, in two of the peripheral areas that we studied. In Cane River, gonyen or ganyen was used frequently by the most fluent speaker we interviewed, though she used the shorter form, gen, as well. 8 In St. Tammany Parish, ganyen was recognized by a single LC speaker who recalled having heard it in her youth; 9 as in the other areas, however, the longer form has today been replaced by gen. 10 Our data from Plaquemines Parish and Marshall’s data from Mon Louis Island show only the latter form. The traces of the archaic form ganyen or gonyen in two peripheral areas confirm the evidence from nineteenth-century texts that it was once a common feature of LC and, further, provide an indication of the extent of the geographical distribution it once had. While this form is clearly an archaism, it cannot be considered to be more basilectal than its shorter counterpart gen, since neither is closer to or further from French than the other. In this way, its status differs from that of single-form verbs and the absence of gender marking on possessives, both of which are at once more archaic and more basilectal than the features that appear to have replaced, or to be replacing, them in the central zone. 3.2 Gender markings on possessive determiners Unlike French, which marks 1 sg., 2 sg., and 3 sg. possessive determiners for the gender of the object possessed (masc. mon, ton, son vs. fem. ma, ta, sa), the LC of nineteenth-century texts shows no gender markings on possessive determiners. Instead, as shown in Table 1, the forms mo, to, so are used for all nouns, regardless of their grammatical gender in French or of the sex of their referent (Neumann-Holzschuh 1987:9-10). However, in the modern-day LC of the central area, the feminine forms ma, ta, sa alternate with forms unmarked 8

The other three LC speakers interviewed in Cane River, one of whom could be classified as a semi-speaker and the other two as rememberers, used only gen. 9 Valdman et al. (1998:180). 10 The speaker in St. Tammany Parish who recognized ganyen was interviewed not by us but by James Etienne Viator in the mid 1990s (James Etienne Viator, personal communication). We wish to thank him for providing us with this and other data from his work in St. Tammany Parish that proved useful for this study.


19

for gender, as shown in Table 2. 11 Since gender marking on possessive adjectives (as well as on other elements of the Noun Phrase) is absent from nineteenth-century texts, Neumann believes that its presence in the LC of the central zone is a recent development brought about through the influence of Cajun (LRF in our terminology). As she puts it, “In Louisiana Creole ... this category [grammatical gender] is in the process of being reconstituted under pressure from Cajun, which, with a few exceptions, has preserved grammatical gender” (1985:58, our translation). 12 While this is a logical conclusion to draw based on the available evidence, we might question the reliability of the data from nineteenth-century texts. As previously noted, these texts were written almost exclusively by whites, and it is possible that, in an attempt to caricature LC and its speakers or at least make them appear more exotic to readers, the authors exaggerated the basilectal nature of the language; they may have overrepresented features that distinguished it from French and left aside others, like gender marking, that may in fact have been used variably at the time, but that would have detracted from the stereotypical depiction the authors were striving for. Nineteenth-century LC French mo fame (p. 51) ma femme (f.) mo mait et mo maîtresse (p. mon maître (m.) et ma 138) maîtresse (f.) mo chimise (p. 131) ma chemise (f.) to mouman (p. 127) ta maman (f.) to la djole (p. 81) ta gueule (f.) so famme (p. 61) sa femme (f.) so popa (p. 37) son papa (m.) so pied (p. 131) son pied (m.) Source for LC data: Neumann-Holzschuh (1987).

English “my wife” “my master and my mistress” “my shirt” “your mother” “your mouth” “his wife” “her father” “his foot”

Table 1: Absence of gender on nineteenth-century LC possessive determiners

LC of Teche region French ma mezon ma maison tou ta vi toute ta vie don sa kour dans sa cour Source for LC data: Neumann (1985:128-29).

English “my house” “all your life” “in his yard”

Table 2: Gender marking on possessives determiners in the modern-day LC of the Teche region 11 The examples in Table 2 all come from black speakers. Gender marking in the Noun Phrase is much more common in the LC of whites. See Neumann (1984) and Klingler (2003a:116117, 186-187). 12 “Dans le créole louisianais ... cette catégorie est en voie de reconstitution sous la pression du cajun, qui, lui, a conservé—à quelques exceptions près—le genre grammatical.”

20


Let us now look at data collected in the peripheral areas under consideration to see if they can contribute anything to our understanding of the history of gender marking on possessives. In a brief interview conducted in 1994, our lone LC speaker from Plaquemines Parish uttered just five tokens of a possessive determiner susceptible of showing gender distinctions, all of them involving the 1 sg. As can be seen in Table 3, in each case he followed the French pattern of gender marking, using the feminine form ma before the noun fanm “wife,” whose referent is female, and the form mo before the nouns popa and pèr “father,” whose referents are male, and before jarden “garden,” which is assigned masculine gender in French. The picture is somewhat different for the other areas. In the data we have transcribed from St. Tammany Parish, only two tokens of a possessive determiner marked for feminine gender were found out of a total of 38 instances where a feminine form would be used in French, or just over five percent. In Cane River, the two semi-speakers we interviewed used no feminine forms of the possessive determiner, out of 21 utterances where such a form would be required in French. Finally, Marshall’s data from Mon Louis Island include just two examples of nouns having feminine gender in French that are preceded by a possessive determiner susceptible of showing gender distinctions, but in both cases the non-marked form of the possessive was used. Examples from all areas are provided in Table 3. 13 It is clear that the data from St. Tammany Parish, Mon Louis Island, and, especially, Cane River more closely match those found in nineteenth-century texts than do Neumann’s data from the central zone, in that they show little or no variation in the direction of the mesolect (though two St. Tammany Parish speakers did each use a feminine form once). 14 Coming from peripheral areas where older features may be expected to occur, these data corroborate those of nineteenth-century texts and thus increase the likelihood that the picture these texts present of the LC of the period is, at least as far as possessive determiners are concerned, a fairly accurate one. This, in turn, lends support to Neumann’s claim that the frequent use of gender marking in Breaux Bridge Creole is a recent development attributable to contact with LRF. The data from Plaquemines Parish, which do not fit the same pattern, remain to be explained. We return to them in the final section of the study.

13

All of the examples from St. Tammany Parish and most of the examples from Cane River come from translation questionnaires presented orally to speakers. The examples from Plaquemines Parish are drawn from a recorded conversation. 14 It is also worth pointing out that, in addition to taking possessive determiners unmarked for feminine gender, the nouns ladjel (ST), labouch (CR and MLI), and lamezon (MLI) have agglutinated la-. This is yet another feature that is typical of nineteenth-century representations of LC but that is only variably present in the LC of the central zone. As such, it serves to further reinforce the basilectal character of the language in these peripheral areas.


21

LC French English Plaquemines Parish *ma fanm (uttered twice) ma femme (f.) “my wife” mo popa mon papa (m.) “my father” mo pèr mon père (m.) “my father” mo jarden mon jardin (m.) “my garden” St. Tammany Parish mo momon ma maman (f.) “my mother” mo sœr ma sœur (f.) “my sister” *ma sœr ma sœur (f.) “my sister” to ladjel ta gueule (f.) “your mouth” to maman ta maman (f.) “your mother” *ta moman ta maman (f.) “your mother” so dèrnye fiy sa dernière fille (f.) “his last daughter” so tant sa tante (f.) “his aunt” so sœr sa sœur (f.) “his sister” Cane River mon sèr ma sœur (f.) “my sister” mo ti-fiy ma (petite) fille (f.) “my daughter” mo fonm ma femme (f.) “my wife” mo labouch ma bouche (f.) “my mouth” “your mother” to mòman ta maman (f.) to sèr ta sœur (f.) “your sister” son sèr sa sœur (f.) “his sister” son ti-fiy sa (petite) fille (f.) “his daughter” Mon Louis Island ta bouche (f.) “your mouth” to labouch (p. 77)† so lamezon (p. 77) sa maison (f.) “his house” * Indicates a form marked for feminine gender. † Page numbers refer to Marshall (1991). Marshall transcribed her data in IPA; we have changed them to make them conform to the system used elsewhere in this article.

Table 3. Possessive determiners in four parishes 3.3 Verbs with long and short forms As in the case of possessive determiners, there is a sharp contrast between the LC verbal system that is portrayed in nineteenth-century texts and the system that Neumann (1985) and others (Broussard 1942; Lane 1935) have described more recently for the central LC-speaking area. Whereas the historical texts show the vast majority of verbs to have a single, invariant form that is used in all contexts—a system that corresponds to those of Haitian Creole and the creoles of the Lesser Antilles—twentieth-century descriptions of the LC along Bayou Teche show the existence of a class of verbs having two forms, one long and one short, that are in complementary distribution according to grammatical context: The short form is used in the habitual or universal present, in the imperative, and after the impersonal expression (i)fo “it is necessary to/that,” while the long form is used in the preterit, after TMA markers,

22 Context

Present


Nineteenth-century texts [F]o mo fait mo frère oua ké cé so fame ki mangé so pichetache. “I have to make my brother see that it’s his wife who eats his peanuts” (Fortier 1887/1960:134) Mo coupé di boi. “I cut wood (for a living)” (Mercier 1880:380)

Torti la galopé pli vite pacé stimbotte. “That tortoise runs faster than a steamboat” (Mercier 1880:382)

Past

After TMA

Li galopé dan champ pichetache é li mangé ein rang. “She ran into the peanut field and she ate a whole row” (1887/1960:134) Tan compair Chivreil rivé coté primié zalon, li hélé : ‘Halo ! compair Torti.’ “When Bre’r Deer reached the first milestone, he cried: ‘Hello! Bre’r Tortoise!’” (Mercier 1880:382) [H]ibou, fouine avé blette apé mangé tou poule. “The owls, the polecats and the weasels are eating all the chickens” (1887/1960:103, 139) Ma chanté “I will sing” (Mercier 1880:379)

Breaux Bridge To monj sa ek de graton. “You eat that with cracklins” (Neumann 1985:195) (Cf. monje).

Jenn moun aster dons pa lakordeon. “Young people today don’t dance to the accordion” (Neumann 1985:195) (Cf. donse). Chop-la frem a siz-er. “The shop (always) closes at six o’clock” (Neumann 1985:196) (Cf. freme).

Somdi swar ye monje, donse, koze! “Saturday night they ate, danced, talked!” (Neumann 1985:195) (Cf. monj, dons, koz). Chop-la freme a siz-er. “The shop closed at six o’clock” (Neumann 1985:196) (Cf. frem).

Li s’ape monje. “He will be eating” (Neumann 1985:104) (Cf. monj).

Chop-la ape freme. “The shop is closing” (Neumann 1985:196) (Cf. frem).

Ouzote a galopé “You (pl.) will run” (Mercier 1880:382)

Table 4: Single-form vs. long- and short-form verbs in nineteenth-century LC and in Breaux Bridge and after auxiliary verbs (Neumann 1985:188-199). 15 The differences between the two systems are illustrated in Table 4. As in the case of gender marking on possessives, a comparison with the data from historical texts leads Neumann to conclude that this system of long and short verbs postdates the system of invariant verbs that characterizes nine-

15 Perret (1933) also describes two-form verbs, but he does not specify a geographical location for the variety of LC he describes.


Parish

Long form Tou mo piti ye parle angle e ma fanm parle angle. “All of my children speak English and my wife speaks English.” Kan le vyeu moun mele ansanm ye pale kreyòl. “When the old people get [lit., ‘mix’] together, they speak Creole.”

PP

ST

CR

Pwonn la hach na de mòn parle la bon franse. “In Pointe à la Hache there are people who speak good French.” Mo lemè parle. “I like to speak.” Tou se joen moun la parle jus angle. “All these young people only speak English.” I parle franse. “They speak French.” Sa arive de fwa. “That happens sometimes.” Ye pale kreyol. “They speak Creole.” O li li lèmen so labyèr li me li pa lemen diven. “Oh he likes his beer, him, but he doesn’t like wine.” Nou lève bònœr tou le maten. “We get up early every morning.” Kite mwen trankil. “Leave me alone.” Fèrme to ladjel! “Shut your mouth!” I parle franse. “They speak French.”

Parle a mon! “Talk to me!”

Li lenm son wiski. “He likes his whiskey.”

Gade mo piti pou mwen. “Keep my child for me.” (p. 82)† Antre mo frè. “Come in, my brother.” (p. 82) MLI Rèste lamezon. “Stay home.” (p. 78) Ye lenmen sa. “They like that.” (p. 77) Vou lenmen en goble dolo? “Would you like a glass of water?” (p. 81) † Page numbers refer to Marshall (1991).

23

Short form Le jœn pal plu kreyòl. “The young people don’t speak Creole anymore.” Twa tu parl meyœr keu mwen. “You speak better than I do.”

No short forms attested.

Parl pa ak li. “Don’t talk to her.” Nouzòt lèv tou le jour. “We get up every day.” Mo lenm le joli bouke. “I like the pretty flowers.” Ga li la. “Look at him there.” (p. 83) Pri pou nou. “Pray for us.” (p. 82)

Table 5. Long and short verb forms used in present-tense and imperative contexts in Plaquemines Parish (PP), St. Tammany Parish (ST), Cane River (CR), and Mon Louis Island (MLI)

24


teenth-century representations of LC. 16 Here again, however, we might ask whether the nineteenth-century texts in fact portray an artificially homogeneous system that is skewed toward the basilect. And once again, we can look to our data from peripheral LC-speaking areas for help in evaluating the reliability of the historical texts. The examples from our LC speakers in Plaquemines Parish and Cane River, as well as Marshall’s examples from Mon Louis Island, present a mixed picture, with long and short verb forms alternating in contexts where, in the LC of the central zone, only the short form occurs today (Table 5). In St. Tammany Parish, however, there are no examples of short verb forms in any context; only invariant long forms occur, just as in nineteenth-century texts. All four of these peripheral areas, then, show verb systems that are closer to the nineteenth-century system of invariant verbs than is that of the modernday LC of the central zone, which uses long and short forms in complementary distribution to convey grammatical information. The system in St. Tammany Parish is identical to that of historical texts, while those of Plaquemines Parish, Cane River, and Mon Louis Island, which mix long and short forms in the same contexts (thus making it impossible for these forms to consistently express grammatical information), occupy an intermediate space between the system portrayed in nineteenth-century texts and that of the modern-day LC of the central zone. 17 As with possessive determiners, the data on verb forms from the three peripheral areas where we conducted fieldwork, as well as from a fourth isolate investigated by Marshall, corroborate those found in historical texts and thus 16 She suggests, however, that the two-form verb system is an older development than the introduction of gender into the Noun Phrase, since the former but not the latter is noted in the early to mid-twentieth-century descriptions of Lane (1935) and Broussard (1942) (Neumann 1985:197-198). 17 For Marshall, the existence of long and short verb forms in Mon Louis Island is evidence that this feature has long characterized LC (1991:80). While this may be true, the significance of some of the forms she notes is questionable. Travailler, for example, rarely has a long form in any context, even in historical texts, and in the case of pri/priye, the example in which the short form occurs, pri pou nou, is identical to a typical congregational response during Catholic mass, prie pour nous “pray for us,” and may in fact be a relic from this religious context rather than evidence that the short form pri was part of the LC of Mon Louis Island. (Indeed, in discussing this verb, Marshall notes the possible influence of priests [1991:80].) More to the point for our analysis, however, Marshall does not state, nor do her examples suggest, that the long and short forms convey grammatical information as they do today in the LC of the central zone. Indeed, alongside the long forms occurring in present-tense and imperative contexts that we cite here, Marshall notes the short form dòn “to give” (cf. done, which is also attested in Marshall’s data) after the preverbal future marker a, where in the central zone we would expect to find a long form: ma dòn twa en bon vole “I’ll give you a good spanking” (1991:82). This indicates that, while some short forms did exist in this variety, they were not part of a system like that of the central zone, where long and short forms are used in complementary distribution and convey grammatical information.


25

allow us to have greater confidence in the accuracy of the historical record concerning LC. If, as Neumann maintains and our research tends to confirm, the system of invariant verbs represents an older form of LC and the system of long and short forms in complementary distribution is a more recent development, then we can conclude that, with regard to this feature at least, the St. Tammany Parish variety of LC is more conservative than are those of Plaquemines Parish, Cane River, and Mon Louis Island. These areas show movement towards the system found in the central zone, but they have not evolved as far in the direction of French and remain closer to the earlier, more basilectal system. It is possible that they provide a window onto an intermediate stage of development that the LC of the Teche region passed through as it progressed from a basilectal system of invariant verbs to the two-form verbal system it displays today. 18 4. Discussion and conclusions While all of the areas we studied retain archaisms that are useful in advancing our understanding of the historical development of LC, they also show interesting differences from each other that can probably only be explained through a careful examination of the ecology of LC in each area over time (Mufwene 2001). Our research in this domain is only in its initial stages, but there is one element of the linguistic situation that is likely to have played a key role: the continual presence of another variety of French alongside LC. As we noted earlier, in Plaquemines Parish and Cane River, LRF is more widely spoken than LC. In fact, our only speaker of LC in Plaquemines Parish declared that he also knew ‘French,’ which he had learned from people just a few miles down the road from his home. This claim was confirmed by the fact that, while some of his utterances were clearly LC, others appeared to be complete code switches to LRF. 19 It could also explain his marking of gender on possessives and his use of short verb forms in some contexts. 20 In St. Tammany 18

As has been pointed out by Hull (1983:9), Neumann (1985:197), and Klingler (2003a:8889), there is evidence in Broussard (1942) that the LC of the central zone also once had invariant verbs. In contrast to the two-form verbal system illustrated in the rest of his text, most of the proverbs cited by Broussard show invariant long-form verbs in the universal present. Since proverbs often preserve archaic linguistic features, it seems likely that Broussard’s proverbs reflect the basilectal nature of the LC that was originally exported from the Mississippi Valley to the Teche region. 19 An example of an utterance that could be LRF rather than LC is I m a di ki se ize pou parle italyen “He told me it’s easy to speak Italian,” in which the main clause features the 3 sg. subject pronoun i rather than li, the preverbal indirect object clitic pronoun m rather than postverbal, nonclitic mwa, and the past tense of the verb “to tell” formed with the auxiliary a and the past participle of the verb, di. All of these features are characteristic of LRF (and of SF, for that matter) but quite distinct from LC, in which the equivalent would be Li di mwen ... 20 The two clear examples where he used short verb forms, given in Table 5, could also be interpreted as code switches to LRF, since neither utterance contains elements that are unique

26


Parish, in contrast, we encountered only speakers of LC and no speakers of LRF. We do not know how long LRF has been absent from the scene in St. Tammany, but it seems likely to us that this absence helps to explain why the LC of the area retains a more basilectal verbal system than the LC of the other two areas. There remain some apparent anomalies in the data for which we are unable to provide satisfactory explanations. It is not clear, for example, why the data from St. Tammany Parish show two instances of feminine gender marking on possessives but no instances of short verbs, while in Cane River we found some short verbs but no examples of gender marking. It is possible that further research will reveal answers to these and other lingering questions. What is most significant for our analysis here, however, is the generally more basilectal character, in comparison to the LC of the central zone, of the language documented in each of the four LC isolates we examined. This finding, along with the discovery of traces of the verb ganyen in two of the isolates, is largely in accordance with nineteenth-century representations of LC. This does not, of course, make it possible to conclude that these representations were wholly accurate and devoid of caricature or exaggeration, and it should be emphasized that the type of Creole they portray is considerably more homogeneous than what is to be found in the peripheral regions. It does, however, suggest that nineteenth-century representations of the features examined here were not as far from the mark as we might have feared, based on the many reservations that have been raised regarding the authenticity of early creole texts. 21 Our findings further lend support to a scenario for the development of LC according to which the language arose in a single location and was subseto LC and not also found in LRF. A codeswitching analysis seems especially appropriate for the second example, since the pronoun tu is a feature of LRF but not of LC. Clancy Clements (personal communication) has suggested to us that, given how infrequently the speaker uses either variety, it is possible that he does not retain a clear sense of the boundary between them and perceives himself to be speaking a single variety throughout the interview, in which case a codeswitching analysis would not be appropriate. It is true that there is little in the interview itself that would allow us to assess the degree to which the speaker perceives linguistic boundaries in his own speech. However, his awareness that two varieties exist (or formerly existed) in the area and his statement that he learned to speak ‘French’ as well as Creole show that he perceives such boundaries within the community and lead us to prefer to analyze his variable linguistic production in terms of codeswitching. 21 Clancy Clements (personal communication) suggests that the authors of these texts may have selected, from among the pool of features available in nineteenth-century LC, primarily those that were most salient “in terms of frequency or in terms of contrast with the colloquial French of the time.” This strikes us as a very plausible account of how these texts were written. The result would be a representation of LC that is accurate in the sense that it makes use of features that were genuinely used, but partial insofar as it excludes competing features that were also in use but that were less salient.


27

quently exported to outlying areas, where its later evolution followed divergent paths depending on the specific context. For the central area, Neumann’s analysis of recent movement away from the basilect through sustained contact with CF appears to be accurate. The more basilectal quality of the LC of the peripheral areas examined here can be attributed to less intensive and sustained language contact resulting from the early disappearance of other French-related varieties, as in St. Tammany Parish, or from a very advanced stage of language shift in which English long ago replaced other varieties in most domains of use.

References Bartoli, Matteo. 1945. Saggi di linguisitica spaziale. Turin: Vinzenzo Bona. Broussard, James Francis. 1942. Louisiana creole dialect. Baton Rouge: Louisiana State University Press. Chambers, J.K. & Peter Trudgill. 1998. Dialectology (2nd ed). Cambridge: Cambridge University Press. Cigale et la fourmi, La. 1846. Revue Louisianaise 1(22). 538. Reprinted in Textes anciens en créole louisianais, Neumann-Holzschuh 1987 (Hamburg: Helmut Buske). 24. Fortier, Alcée. 1960. Bits of Louisiana folk-lore. Publication of the Modern Language Association of America, vol. 3. New York: Kraus Reprint Corporation (Original work published 1887). 100-168. Guilbeau, John. 1950. “The French spoken in Lafourche Parish, Louisiana”. Unpublished doctoral dissertation, University of North Carolina, Chapel Hill. Hazaël-Massieux, Guy. 1990. “Le guyanais et les créoles atlantiques à base française”. Etudes Créoles 13. 95-110. Hazaël-Massieux, Marie-Christine. This volume. “Using and interpreting historical texts to analyze the formation and development of creole languages”. Hull, Alexander. 1983. “Créole louisianais et créole haïtien: ressemblances et différences”. Paper presented at the Quatrième Colloque International du Comité International des Etudes Créoles, Lafayette, Louisiana. Klingler, Thomas A. 2000. “Louisiana Creole: The multiple-geneses hypothesis revisited”. Journal of Pidgin and Creole Languages 15. 1-35. ----------. 2003a. If I could turn my tongue like that: The Creole language of Pointe Coupee Parish, Louisiana. Baton Rouge: Louisiana State University Press. ----------. 2003b. “Language labels and language use among Cajuns and Creoles in Louisiana”. University of Pennsylvania Working Papers in Linguistics 9, Papers from NWAV 31. 77-90.

28


Lane, George S. 1935. “Notes on Louisiana French II: The Negro-French dialect”. Language 11. 5-16. Marshall, Margaret M. 1991. “The Creole of Mon Louis Island, Alabama and the Louisiana connection”. Journal of Pidgin and Creole Languages 6. 7387. Mercier, Alfred. 1880. “Etude sur la langue créole en Louisiane”. Comptesrendus de l'Athénée Louisianais 5. 378-83. Mufwene, Salikoko S. 2001. The ecology of language evolution. Cambridge: Cambridge University Press. Neumann, Ingrid. 1984. “Le créole des Blancs en Louisiane”. Etudes Créoles 6. 63-78. ----------. 1985. Le créole de Breaux Bridge, Louisiane. Etude morphosyntaxique, textes, vocabulaire. Hamburg: Helmut Buske. Neumann-Holzschuh, Ingrid. 1987. Textes anciens en créole louisianais. Hamburg: Helmut Buske. Papen, Robert A. & Kevin J. Rottet. 1997. “A structural sketch of the Cajun French spoken in Lafourche and Terrebonne parishes”. French and Creole in Louisiana, Valdman (ed.) 1997 (London: Plenum). 71-108. Perret, Michael John. 1933. “A study of the syntax and morphology of the verb of the creole dialect of Louisiana”. Unpublished master’s thesis, Louisiana State University, Baton Rouge. Robin, C.C. 1807. Voyages dans l’intérieur de la Louisiane (3 vols). Paris: F. Buisson. Rottet, Kevin J. 2001. Language shift in the coastal marshes of Louisiana. Berlin: Peter Lang. Speedy, Karin. 1994. “Mississippi and Teche Creole: A demographic and linguistic case for separate genesis in Louisiana”. Unpublished master’s thesis, University of Auckland, New Zealand. ----------. 1995. “Mississippi and Tèche Creole: Two separate starting points for Creole in Louisiana”. From contact to creole and beyond, Baker (ed.) 1995 (London: University of Westminster Press). 97-111. U.S. Bureau of the Census. Census 2000, Summary File 3. Age by language spoken at home for the population 5 years and over 2004. U.S. Census Bureau [cited January 20 2004]. Available from http://factfinder.census.gov. Valdman, Albert & Thomas A. Klingler. 1997. “The structure of Louisiana Creole”. French and Creole in Louisiana. Valdman (ed.) 1997 (London: Plenum). 109-144. ----------, Thomas A. Klingler, Margaret M. Marshall, & Kevin J. Rottet. 1998. Dictionary of Louisiana Creole. Bloomington: Indiana University Press.

USING AND INTERPRETING HISTORICAL TEXTS TO ANALYZE THE FORMATION AND DEVELOPMENT OF CREOLE LANGUAGES

Marie-Christine Hazaël-Massieux Université de Provence ABSTRACT: Although they present significant problems to the interested researcher, creole texts from the last three centuries are important sources of information on the state of early creoles as well as of early popular French. In this study, we will concentrate on methodological principles that should be borne in mind when analyzing historical creole texts. After discussing the nature of the evolutionary evidence available in written sources, we will turn our attention to two important methodological guiding principles: the question of the authenticity of historical texts and the treatment of variation found in these texts. Provided that great care is taken in their interpretation, we argue that historical creole texts can be a source of invaluable insight into earlier periods in the evolution of creoles. 1. Introduction Creole texts from the eighteenth, nineteenth, and early twentieth centuries are a source of fascinating discoveries. More particularly, they are indispensable tools for attempting to reconstruct, from a comparative perspective, the evolution of creole languages from their origin to the present day. However, the interpretation of these texts is fraught with numerous difficulties, so that great caution must be exercised when using them. Our purpose here is to outline the most important of these difficulties and to consider some methodological guidelines that must be established in order to ensure the validity of any conclusions drawn from the study of historical creole texts. 1 We will examine, in turn, the following: • • •

the question of the evolution of creole languages; the question of the authenticity of historical texts; variation and its causes.

1 By ‘historical creole texts’ we mean texts in creole written in previous centuries that make it possible to identify differences in past and contemporary usage. They were produced by a variety of writers, including missionaries or catechism teachers, playwrights, poets, and transcribers of works drawn from oral tradition. Also included are quotations from chroniclers who report sentences in creole from court trials or various situations in which creole was used.

30


2. The evolution of creole languages: evidence in written sources Like all languages, creoles evolve. To be sure, they evolved considerably more quickly in the first two centuries of their existence than they do now, but they continue to evolve. This evolution over time has brought about significant modifications of their grammatical systems. Lexical elements have acquired new functions and become grammatical morphemes—a process called grammaticalization. Continued language contact between (at least) the L-language (the creole) and the H-language (French and/or English) has resulted in regularizations and reanalyses that give rise to systems that are largely new and that are quite different from the linguistic systems assumed to have given birth to creoles. 2 Particularly interesting from this point of view are the verb systems, pronominal systems, and determiner systems, for only in systems such as these can one observe the operation of rules governing processes of restructuring that are so familiar in linguistics. Many people are not aware that it is quite likely that Guadeloupean Creole (GC) and Martinican Creole (MC) clearly diverged from one another in terms of the definite determiner as recently as the twentieth century, and especially after 1920, and that the differences between these two varieties have continued to increase. It was only during the nineteenth century, or perhaps at the beginning of the twentieth, that MC and GC adopted different possessive systems: liv-moin “my book” in Martinique versus liv-an-moin in Guadeloupe. The attestations provided by historical texts, which we have carefully studied, are confirmed by observers’ remarks. Father Goux, describing the speech of Martinique (we will see that MC stabilized earlier than GC), noted in 1842: The latter [possessives in MC] are placed after the noun, for example, ich moin “my child,” maman toé “your mother,” bagaïe li “his object.” In Guadeloupe, à is placed between the noun and the possessive pronoun, as in petite à moin “my child,” etc. (Goux 1842:2) [Editors’ translation]

2

Let us recall at the outset that the linguistic systems that are relevant for the genesis of creoles are not, as far as the European language is concerned, the normative written varieties found in the literature of the seventeenth and eighteenth centuries, but rather oral varieties that were spoken by the colonists and that were often regional (e.g., those of western France). Fortunately we have rather good documentation for the historical varieties spoken in seventeenthand eighteenth-century France. This is not the case for the forms of African languages spoken by the slaves. Comparative and historical African linguistics is much less advanced than is French historical linguistics, and for Africa we have no written documents, whereas in France there are letters written by people with little education, texts in dialect, authors’ attempts to represent popular speech (e.g., Molière in Don Juan), poems which enable one to reconstruct the actual pronunciation, and so forth. All of these various elements help to reconstruct the popular French that we can suppose to be at the origins of creoles.

CREOLE TRANSPLANTATION

31

Similarly, in a reference to the writings of Paul Baudot 3 (see also HazaëlMassieux 2000), Nainsouta (1940:20) noted: Along with the preposition à, bitin “thing, object, possession” creates the essential difference between the speech of Guadeloupe and the speech of Martinique. Bitin an moin “my possession” in the former, bagaïe moin in the latter. There is nothing absolute about this, moreover. They say chan-me à bain “bathroom,” chimise à hon-me “man’s shirt,” etc. We say zecodin-ne “(small variety of) mango,” pain mi (maïs) “cornbread,” papa moin “my father,” farine manioc “cassava flour,” grain-ne zaboca “avocado seed,” etc. A careful reading of Baudot shows that in this matter, a hundred years ago, the difference was even less absolute than it is today. [Editors’ translation]

He said in an empirical and intuitive way the same thing that we were able to establish by analysis: Assembling a collection of texts grouped chronologically between the eighteenth and the twentieth centuries shows, then, from an initial examination of determiners (possessives and definite articles), that creoles were gradually putting into place the rules that would characterize them later on, that there are still sometimes important developments taking place, and that certain features stabilize rather late (e.g., at the beginning of the twentieth century, for possessives in Guadeloupe), with some creoles stabilizing later than others, in other words diverging later from their neighbors. Although we cannot explain the choices made in each place, it is obviously impossible to deny that the varieties have diverged, and it is just as obvious that these developments cannot be ignored when their stages of development are so clear. (HazaëlMassieux 2000:58) [Editors’ translation]

To be sure, writing is relatively slow to record new forms. When established habits of written representation exist—even before it is possible to speak of a ‘written norm’—people tend to ignore phonetic modifications that are perceptible in speech, and these are therefore slow to be recorded in writing. However, this recording does gradually take place, as we can see in historical texts if we trace the representation of the postposed definite determiner la/là. This determiner was always given in its full form until the end of the nineteenth century at least, 4 but it increasingly came to be written –a in Martinique in the

3

A Guadeloupean author (1801-1870) who published numerous works in Creole in which, in fact, structures with and without à occur side by side. 4 In 1842, Father Goux, who was a very keen observer in this matter, described the form and function of the determiner clearly; he used the term ‘demonstrative’ because of the very strong deictic value of –la, which is retained in the modern creole and which explains the impossibility of its use in many contexts where French has an article: “The demonstrative pronoun is expressed by the monosyllabic form là, placed after the noun, as in pied bois là chèce, ‘that tree is dry,’ or bananes là mis, ‘these bananas are ripe.’ Often this same monosyllable is placed after nouns without being a demonstrative, as in the sentence jardin vous là bel ‘your garden is beautiful’” [Editors’ translation]. He was clearly talking about là and his use of the grave accent identified it as the second part of the reinforced French demonstrative. At the time of

32


twentieth century, which corresponds to the now current pronunciation –a after a vowel (thus liv-la “the book” but zozyo-a “the bird”). The disappearance of this l in writing did not become common until well into the twentieth century. It is found in the writings of Gratiant, whose first poems were written around 1940 (see Gratiant & Cohen 1958, Gratiant 1976, and Gratiant et al. 1996). 5 This does not mean that the pronunciation /a/ was never attested before that date, but it does mean that writers had become aware of it by then, and that they could even use this feature to emphasize differences between GC and MC. Even in Haiti, where the determiner –la underwent the same evolution as in Martinique, and where it is usually believed that developments took place earlier than in Martinique, we continue to find the full form with l through the end of the nineteenth century. In his 1896 poem, “Choucoune,” in which there are conscious attempts to represent phonetic reality, Oswald Durand nonetheless systematically wrote jou-là “the day” and temps-là “the time” where today we would find jou-a and tan-an. But it must be said that at that time, according to Father Goux’s description, the simple marker of definiteness was primarily attested in a form close to that of French; in other words, it was still preposed, despite the fact that some examples give the impression that it had partially lost its autonomous status and that, even when the written form did not indicate this, the definite marker should be considered an agglutinated article as in the forms lari (< la rue) “street,” monpè (< mon père) “father,” divin (< du vin) “wine” that one finds today. In “Choucoune” we find forms like grand la peine “great sorrow,” and we get postposed –là only in cases of deictic reinforcement, which is why Father Goux interpreted it as a demonstrative. In “Choucoune,” at the very end of the nineteenth century, we can see that the possessive was already always taking the form seen in malheur moin “my misfortune” or pieds moins “my feet,” with no trace of a preposition, confirming the developments that we have described for other historical texts; and yet there is no sign at all of the dropping of the /l/ in -la, and it is therefore quite certainly a later morphophonological phenomenon. Although the novels of Justin Lhérisson, 6 written in the same period contain much more French than Creole, the Creole portions show forms such as ti fi la or ti fi là “the little girl,” rather than ti fi a. In the work of Georges Sylvain (1929/1971), one still finds, for example, loup-là “the wolf,” d’l’eau-là “the water,” ti mouton-là “the lamb.” It is therefore only with his writing, there can be no doubt that the loss of the [l] between two vowels had not yet begun. 5 Most of the works collected in Gratiant et al. (1996) were published in the 1950s (see in particular Gratiant 1958), but many of the poems were written even earlier. His first poem, “Joseph lévé!,” probably dates from 1935. 6 The first edition of La famille des Pitite-Caille was published in 1929 (Lhérisson 1929/1975), while the first edition of Zoune chez sa ninnaine appeared in 1906 (Lhérisson 1906/1975).


33

Gratiant (Martinique), who published his first poems beginning in 1936 (Gratiant & Cohen 1958), that we can observe any change in the representation of the definite determiner. Several example from the writings of Gratiant are given in (1). Note that in (1f), the -s is purely orthographic, and that there is no trace of nasalization in (1h), which is quite probably a still later phonetic development that we would date after 1950-1960. (1) a. dans la tè glissan-a “in the sliding/slippery dirt” b. negg-la “the black man,” c. beuff-la “the bull,” d. docteu-a “the doctor,” e. doudou-a “the dear one,” f. coutlas-a “the machete” g. bon matin-a “in the morning” h. l’horizon-a “the horizon” j. savan-n-la “the field” vs. savann-lan (MC) Linguists should keep this development of creole languages in mind in order not to accept too readily a single hypothesis for interpreting old data involving mixture—in which apparently no rules can be established—of forms that are now associated with different dialects. Can we legitimately talk about dialect mixture for these historical periods? Do we know whether the varieties that we recognize today existed in the same forms two centuries ago? In fact, historical creole texts are characterized by a significant amount of grammatical indeterminacy, 7 and in an eighteenth-century text like the Passion de NotreSeigneur selon Saint Jean en langage nègre, 8 for example, we find features that would today allow us to systematically identify a text as being GC or MC occurring within two lines of each other, and throughout the entire text, which rules out the possibility that there were multiple copyists, each representing his own speech: During the Last Supper 9: “jesi prend pain, cassé li, séparé ba ïo tous; di ïo... prend li, mangé, cila sé corps moé, vous tende?... li prend couyambouc la outi li té metté di 7

By this we mean that as long as a form does not have a strictly grammatical function, it can be replaced at any time by another term having approximately the same lexical value (i.e., a synonym). 8 Text published posthumously in the collection of works by Guy Hazaël-Massieux (1996). 9 Let us recall that, since the origins of the Church, this narrative from the Eucharistic liturgy has been considered the essential moment of the celebration (i.e., the Consecration); the Church has always sought to leave the words unaltered. This variation is therefore particularly striking at this point in the narration, and it indicates that, whereas for us this would be considered a case of mixing of GC and MC, when this text was written, the writers had the impression that they were writing the same thing, insensitive to something that had not yet become a characteristic grammatical feature.

34


vin, li séparé ba ïo tous chaquin pitit quian dans coui ïo, li dire ïo: boire ça, c'est sang a moé, tendez ? [...]” 10 At the time of Jesus’ death on the cross, when he is speaking to Mary and John:

“chere mama, avla pitit a vous; pitit, avla mama vous” 11 The possessives in question are in boldface to facilitate their comparison; the proximity of these forms to each other in the text makes this feature all the more striking. It is difficult to imagine that these differences are the work of two authors or two copyists, as we said earlier. It seems equally unlikely that this is a case of dialect mixture arising from the presumed inability of the author, who had traveled around the islands, to distinguish them clearly. A much likelier scenario, which the presence of this type of mixture in all of the historical texts would tend to confirm, is that in the creole of the eighteenth century, many features that have now become systematized were used interchangeably. Either form could be used in free variation, just as one can use this or that lexical form in modern French or English, but the precise rules specifying their distribution, which constitute grammar, did not yet exist. It created no difficulty to use either corps moé “my body” or corps a moé. 12 Nowadays, one of them is associated with Martinique and the other with Guadeloupe, and they are not interchangeable; the second is highly unlikely in Martinique, while the first is just as unlikely in Guadeloupe except when naming one’s parents (cf. the example, papa moin “my father,” quoted by Nainsouta (1940:20), which is perfectly common even today). Creole languages, like all languages, continue to evolve. Thanks to recorded oral corpora it is now possible to follow this evolution. Recently Valdman (personal communication, 2003) indicated that Haitian Creole lan, a form which, like the MC lan, not long ago occurred only after a nasal consonant, has come to take on the role of sole determiner in certain regions for certain speakers in Haiti. We have, in fact, noted this form in all contexts on Radio France Internationale during some interviews. However, the examination of written texts that record language evolution helps us to understand what may have happened in past centuries. Thus, in Martinique, the forms of the definite determiner are constrained by the context, and grammarians have already clearly laid out the rules specifying their distribution. There is a perfect

10

Translation: Jesus took bread, broke it and gave it to them; he told them, “Take it and eat: this is my body, do you understand?” He took the wine vessel and shared it with them, each taking a little bit in his own vessel, and he told them, “Drink that, it’s my blood, do you understand?” [Editors’ translation] 11 Translation: “Dear mother, behold your child; child, behold your mother” [Editors’ translation]. 12 Before making this statement, we verified all contexts of use; in particular we verified that there was no phonetic conditioning (cf. Hazaël-Massieux 2000).


35

complementary distribution among la/a/lan/an according to the following principles: -la after an oral consonant: -a after an oral vowel: -lan after a nasal consonant: -an after a nasal vowel:

liv-la “the book” zozyo-a “the bird” fanm-lan “the woman” pon-an “the bridge”

The reading of contemporary texts has drawn our attention to a developmental phenomenon that we have also noticed in speech. To the above distribution of forms, the following form should be added (at least for popular speech; it is uncertain whether this pronunciation is found among all speakers): -ya/-yan after é/è and the corresponding nasal vowel, spelled here en This is attested, for example, in the following excerpt from a poem by Joby Bernabé (1984): Kimoun ki nòz malpalé fanm Jòdi ta-a anlè tè-ya? “Who on earth today dares to malign women?” [Editors’ translation]

Or again: Pawòl mwen sé an ti lafimen Ka filtré anba kannari-w Lafimen yan ké ay fè chimen-i. “My words are a wisp of smoke escaping from under your pot. The smoke will make its way.” [Editors’ translation]

This form indicates that the evolution of the system is ongoing, and that the following should undoubtedly be added to the system presented above for MC: -ya after a front oral vowel other than /a/ 13 -yan after a front nasal vowel other than /å~/. Some of the observations in Bernabé’s thesis (1983) confirm ours, though he also noted the form lanbi-ya whereas there are no occurrences in our corpus of –ya after i. He noted: Where front vowels are concerned, we have the realization –a /a/ after a non-nasal vowel when the vowel in question is more open than /e/. When the vowel in question is /i/, we have the form –ya /ja/; after the nasal vowel /e~~/, we have –an or –yan. After 13

Jean Bernabé (1983) seems to have encountered the form ya after a.

36


the vowel /a/, we have either –a /a/ or –ya /ja/. Moreover, it seems that for several years now a fourth pair of forms has been emerging, functioning after the back vowels /u/ and /o/, namely, -wa, and after /õ/, the form -wan (Note however that the latter form is rare and limited). (Bernabé 1983:644-45) [Editors’ translation]

3. The question of authenticity Another question that arises when analyzing historical texts, and which is in no way a minor one, concerns their authenticity. Even if a linguist succeeds in reconstructing the system rather easily, is the corpus on which the analysis is based really representative of the Creole of the period in question? This question of textual authenticity actually involves several issues. It is worth remembering that, especially for the oldest texts, we are often dealing with anonymous materials; we do not know who wrote them or when they were written (the exact place, city, island, etc., can greatly alter the interpretation by virtue of dialectal variation that may have been very significant). 14 These texts may sometimes even be signed, without the signature necessarily corresponding to the real author, who then remains in the shadows. Consider, for example, the Proclamations révolutionnaires in Creole by an unknown author and signed simply Leclerc or Bonaparte, but which quite manifestly were written by an unknown person (at the Cape, we are told) who knew a little Creole, but whose real knowledge of the language cannot be determined more precisely (Denis 1935). The Proclamations in many places seem to us to be more heavily influenced by French than were older texts, which were doubtless written by speakers who, even if they were non-native, had a better or longer established command of Creole. Nevertheless, we can assume that the oldest anonymous texts were written by whites (e.g., chroniclers and missionaries), whose Creole was probably marked by the French that they spoke and wrote. Without anachronistically conferring on the eighteenth century the modern-day opposition between a ‘basilect’ and an ‘acrolect’ that supposedly corresponds to racial categories— the basilect being the language of the phenotypically darkest segment of the population while the acrolect is used by whites—we must acknowledge that the Creole that is handed down to us in writing through historical texts comes 14 The question of variation in creoles, to which we will come back later, is a central question, but one for which data from historical writings are difficult to interpret, partly because of the lack of reliable phonetic data, but also because, in the absence of a standardized spelling, the variants are numerous. However, even if variants are not always clearly decipherable behind what are sometimes whimsical spellings, the absence of written norms can also be of great help to the linguist who is trying to recover likely pronunciations. The writers of the seventeenth and eighteenth centuries, much more so than today, did not hesitate to write what they heard. For instance, in numerous texts we find the 1sg pronoun spelled moins or moé. Even if the s serves no purpose, the first transcription, which has the same spelling that we associate with a different word from this period (< Latin minus), nonetheless indicates the nasalization that one could not deduce from the spelling moé, which probably indicates another form.


37

from people who were sufficiently educated to know how to write (and who thus had learned to write in French). Some—for example, missionaries— learned Creole thanks only to their stay in the islands, and they were therefore not native speakers. Although in daily life they were doubtless able to communicate with their interlocutors who were native speakers, this does not imply that their Creole corresponded exactly to that of native speakers: One may— indeed, one must—suppose that it was a language of foreigners, doubtless marked by particular features and perhaps notably ‘closer’ to French (which was their native language and which constantly motivated their interpretation, if only unconsciously) than the Creole of native speakers would have been if we had reliable attestations of it. Moreover, as is clearly visible in some texts, especially in the nineteenth century when many were written not by ‘foreigners’ (i.e., whites from the Metropolis) but rather by Creole whites (i.e., whites born in the colony) or educated mulattoes, the written use of Creole was very often motivated by the desire to entertain others of the same social category. This was accomplished by means of references or allusions to a world that all of them doubtless loved, 15 but that they regarded with the amusement of those who are to some extent on the outside, in particular because they also had a perfect mastery of French—and, more pertinently, of written French. This aspect, which occasionally implies caricature, exaggeration of features considered ‘typical,’ and probably also the minimizing or exclusion of more ‘normal’ features, must be emphasized in further developing the notion of authenticity. 16 Of course, we should always remember that the written language is never a faithful rendition of the oral, that a literary work always reflects a certain view of reality and not an exact duplicate. In any case, we have nothing but these historical texts to work with: We must make do with them and be glad that we at least have this much. Besides, it is well known that the language of the transcriber plays a role even in a strict phonetic transcription. An English person and a French person transcribing the same corpus, for example, will not 15

As surprising as the choice of the verb ‘love’ may appear, we remind the reader of what Rémy Nainsouta wrote to justify his work, Le langage créole: “J’aime profondément le créole” (“I love Creole deeply”) (1940:8). 16 By way of comparison, we note the techniques that are employed in some contemporary literature in order to write entire works in the regional French of the Antilles, a French that bears a Creole imprint. While this French is certainly not a pure invention of authors like Chamoiseau or Confiant, it is nevertheless—in their literary and esthetic perspective, or even in the political aim that the authors sometimes articulate of wanting to separate Creole and French as completely as possible—rather far removed from what speakers produce in the way of creolisms or speech marked by interference in everyday life. Often the authors exaggerate, create new linguistic forms and structures, or employ certain processes to such an extent that the result becomes almost a caricature of the linguistic reality of the Antilles—even if it does have an undeniable literary value.

38


produce the same results. But we must remember that what these historical texts provide is an interpretation of the Creole language as the authors believed they heard it (or as they analyzed it), and that modern-day linguists must try to equip themselves with as much knowledge as possible of the language of the authors, their history, and their linguistic attitudes in order to correctly decode the texts—knowing full well that, here again, interpretation runs the risk of ‘constructing an object’ that is not the ‘authentic’ Creole spoken during the period in question. 17 A final element that comes into play here is what the authors might legitimately have imagined to be a requirement of the genre they practiced: When one adapts the gospel into Creole for slaves, one seizes the opportunity to remind them, in this context of catechetical training, that it is necessary at least to cover one’s genitals and not to live naked. 18 The fable is a second genre that creole writers manipulated following moralistic goals. For the Creole adaptation, however, it was no longer the moral that La Fontaine formulated for the seventeenth century, but rather one that the bourgeois authors of the Antilles found amusing and full of allusions to local realities for a nineteenth-century audience in the colonies. 19 When Clément, an actor and white Creole author, born in Le Cap, gave us his “Negro parody of the Devin de Village,” (Camier & Hazaël-Massieux 2003) he transformed Rousseau’s text in order to make it amusing and entertaining to an audience with very different concerns from those who flocked to Rousseau’s operas in France: Social oppositions have thus been transformed into racial oppositions, voodoo magic appears in the text, and so forth. Although linguists generally prefer to study real language, taking into account pronunciation (even if it cannot be reconstructed with certainty), and 17

Linguists are well aware that all linguistic analysis tends to ‘construct’ the language object at the same time that it describes it and that there is no solution when faced with this general tendency of the human mind, except, of course, to keep it in mind in order to neutralize it as much as possible. 18 This is precisely the case in the famous Passion selon Saint Jean, cited previously, when the soldiers divide among themselves the clothing of Jesus on the cross and the writer adds the following—which, of course, is not in the original text of St. John nor in any of the other accounts of Jesus’ death: ïo quitté li avec ïon pitit morceau tout déchiré, pour cacher quiou a li (“They left him with a small piece of tattered cloth, to hide his lower half”). 19 Thus in “The Wolf and the Lamb,” Marbot (1846), like most of those after him who adapted the fables of La Fontaine into Creole, chooses first to ask about “the father” rather than proceeding directly to “If it’s not you then it’s your brother.” In this way, the lamb declares that he is a bastard (an allusion to situations that are well known in the islands): Gros loup là réponne comça, “Si c’est pas to, c’est papa Yche maman to.” – “Mais moin pas Ni papa, moin c’est bata.” – “C’est fouè to pouloss, moufi.”

“The big wolf answered in the following way: ‘If it’s not you, it’s the father of your mother’s child.’ –“But I don’t have a father, I’m a bastard.” –‘It’s your brother, then, my child.’”


39

although they are always mistrustful of a ‘representation’ of language, these historical texts nevertheless contribute numerous and solid elements that should not be overlooked: —With regard to the lexicon, we find attestations of words, some of which have disappeared or are preserved only in very conservative creole varieties. It is interesting, for example, to find forms that come from African languages and that have fallen out of use over the course of the centuries, often replaced by terms of French origin. Examples include bichi “between us” (which does not exist in modern-day dictionaries, with the exception of the Dictionnaire du créole de MarieGalante by Père Barbotin [Barbotin 1995]) and wanga “fetish, spell,” now attested exclusively in Haiti. —With regard to the grammar, it is here that the question of the evolution of these languages comes most sharply into focus. Thus, as we have said, the notation of a preposition à or its absence before the pronoun to mark the possessive in Caribbean creole has made it possible, by comparing texts spanning more than two centuries, to establish: • •

the progressive separation of the two possible forms, which now constitute precise dialectal identifiers, and grammaticalization phenomena and the theoretical conclusions that may be drawn from them.

—With regard to pronunciation itself, much can be learned through properly conducted comparisons and through an examination of the orthographic choices made by authors. Examples abound that make it possible for us to know, for instance, that a final consonant was pronounced. When the most scrupulous authors used the word chat “cat,” for example, it was written neither as chat nor as chatte (which would be ambiguous in French 20), but as chatt or chat’, which clearly indicated a pronunciation that, moreover, remains the same today. It would, of course, be inaccurate to imply that it was the feminine form that passed into Creole; rather, the word was exported to the islands and became part of the Creole language at a time when final consonants were still pronounced (especially in western France). In addition to chat or rat (which in Creole refer to either the male or the female of 20 In French chat [ßa] is masculine and chatte [ßat] is feminine. The use of the orthography chatte to represent the word in Creole would correctly indicate the pronunciation of the final consonant but could also easily be interpreted by French readers as a feminine form, when in fact the Creole pronunciation is identical when referring to the male or the female of the species [Editors’ note].

40


the species), we can cite dwèt for devoir “must,” which comes from French doit (3 sg. of devoir) and, depending on the region, was in the seventeenth century pronounced [dwε], [dwεt], or, rarely, [dwa]. Other interesting examples are the verbs fini “finish,” couri “run” (i.e., the -ir verbs in French) for which, as with the -er verbs, the final -r was not then pronounced. A little song from the sixteenth century (it has been established that it existed in 1574, see Hazaël-Massieux 1993:75) reminds us of this: Compère Guilleri Te lairras-tu mouri? “Brother Guilleri Will you let yourself die?” [Editors’ translation]

Very early on, these verbs in -ir are spelled with -i in historical texts. There is too often a tendency to consider the disappearance of the -r as an Antillean phenomenon, when in fact the forms that arrived in the Antilles were very likely couri, fini. This would have occurred, then, before the -r was restored in the pronunciation in France, perhaps only at the end of the eighteenth century. Thus tini (< tenir) “have,” vini (< venir) “come,” senti (< sentir) “feel, smell,” and many others are attested in the Passion selon Saint Jean (late eighteenth century), and even more such forms occur, in nineteenth-century Creole texts, where usage became completely and definitively fixed. In France, by contrast, the conservative nature of French orthography (cf. -ir/-er) would later lead to a return to an [ir] pronunciation, which became generalized, or even to an [er] pronunciation in liaison contexts in ‘Comédie française’ speech: “parler et mourir.”

While it is important to emphasize the precautions that must be observed and the care that must be taken when evaluating the authenticity of these historical texts in Creole, it is nevertheless possible to see that they do reveal a certain concern with reproducing as exactly as possible the language as it was pronounced. To conclude this section, and returning to what we said earlier, we emphasize that a written language is never, strictly speaking, ‘authentic’: It is always an attempt to represent the oral, and in that representation, there may be not only a time-lag but also exaggeration at various levels. However, the discipline of classical philology was built on the written record in this way, and there is no reason for creole languages alone not to use the available material on the pretext that it has somehow been falsified from the start.


41

4. Variation and its causes Variation appears extreme in all the historical texts. There is, of course, variation from one text to another, from one author to another, but as we have seen, there is also variation within a single text dated and signed by a single author. This variation is not the result of the same causes, and it is here that great care is required when it comes to interpretation. Moreover, an interpretation is not always immediately accessible; the texts must first be deciphered before they can be interpreted. Indeed, for the oldest texts, we must work on manuscripts, and allowance must be made for eighteenth-century handwriting practices: ornate letters, lack of any real regularity in the notation of dots on the letter i or of accent marks (often non-existent), the relative absence of punctuation and capitalization, uncertainty concerning word divisions (which result from the differences between Creole and French that the writers sensed), and so forth. It is therefore sometimes extremely difficult to interpret, to ‘translate’ these often ambiguous graphic notations of a language that, moreover, we do not know well and that we have to discover gradually. There is no dearth of examples of cases where one needs to hesitate a long time before deciding that a given symbol should be read as a t rather than an s, or that a given form is doubtless an e rather than an i, for example. The form of a letter changes depending on its position in the word and the adjacent environment. Here again, it is the careful examination and repeated reading of the whole text that makes it possible to decide, because as we become accustomed to the representations of the writer, we can eventually exclude such and such a letter here or there because we perceive that in such an orthographic context it would take a certain form, or we can better recognize a word in one context than in the preceding one, where it was hardly interpretable, and so on. It is quite certain, moreover, that the manuscript ‘interpreter’s’ knowledge of Creole and hypotheses regarding the reconstruction of earlier stages plays a great role in making this deciphering possible. 21 This holds true only for the oldest texts, generally up to the beginning of the twentieth century. After that, the texts have often been published and are thus printed or, at any rate, the writer’s handwriting is less difficult to decipher. Some of the causes of variation can be enumerated: —The absence of an orthographic norm, which, even when two items are pronounced in the same way by the same person, leads a writer to represent the same word differently within the same text. —Different pronunciations due to a lack of fixity: When a language is in the process of evolution and stabilization, the same speaker does 21

More than once, as a Creole specialist, we have been contacted by historians who, despite their sometimes significant knowledge of manuscript reading, could not correctly interpret a word written in a language they did not know.

42


not always pronounce things the same way; speakers are not always aware of this, and in each case they write what they believe they hear. —Different pronunciations depending on the region: Pronunciation may differ from one city to another or from one village to another, even when they are only separated from each other by a short distance. That is precisely what motivated the creation of linguistic atlases in nineteenth-century France. It is perfectly conceivable that Creole, even on the same island, was not identical everywhere. —Different pronunciations due to a change of region: An author might have traveled and thus been exposed to various influences, which might have caused him to use different pronunciations during the course of his life, or in different contexts. It is undeniable that there are changes in the course of a single individual’s life: We have emphasized that the presence of possessive forms with -a in contrast to forms without -a appears to have increased in the works of Baudot (1935/1980) 22 in the nineteenth century, at the same time that GC was becoming increasingly distinct from MC. —Different pronunciations linked to style: It is not certain that pronunciation is always represented in the same way in those cases where characters of different social milieus are portrayed. In addition, the literary genre may have a certain impact: poetry has its rhythmic constraints, which, for example, can lead to the choice at a given moment of a form having one more or one fewer ‘foot,’ yielding possibly a possessive form a-moin here and a form moin there, as long as the choice makes no difference with regard to meaning, to grammar, or to the norm. —More generally, the immediate context may generate variable representations that must be taken into consideration when interpreting some forms. An example that comes to mind is the form itou “also,” which occurs in the form tou when it follows li: li tou “him too,” whereas we more normally find moin itou “me too.” We should thus not be surprised by the high degree of variation. It is rather the presence of stable forms that is surprising in light of all of the possible reasons for variation. It is true that stable forms exist and, as is typical in the history of languages, the proportion of stable forms in Creole texts has risen over the centuries. Forms that authors interpret clearly have become more or less stabilized in the twentieth century, but today we still find variation tied to

22

Paul Baudot, who died in 1870, began writing in 1830 and continued until his death. His works were often distributed in manuscript form. Only after his death was a decision made to publish them in a collection.


43

misunderstandings on the part of authors, who sometimes divide words incorrectly, misinterpret them, and thus misrepresent them. 23 Continuing in this vein, it is appropriate to emphasize how orthographic choices can reinforce (or not reinforce) variation and sometimes even bring about the effective separation of two varieties that become two languages. Today, when there is a tendency to prefer a phonetic notation and to maintain a certain regional specificity, MC speakers may claim that they are unable to read a GC text, and, even more frequently, a Guadeloupean may claim to be unable to read MC if it is transcribed phonetically. From a linguistic point of view, we may consider that we are dealing with a single language, Lesser Antillean Creole, with variants that constitute perhaps two dialects but that are perfectly mutually comprehensible in their oral form. When these differences are recorded and fixed by written usage, and thus emphasized (they are in some cases nothing more than simple features of pronunciation, for example, the palatalization of /k/ and /g/ in MC), the result is a representation of each variety that is illegible to those who are not fluent in it or who are expecting a strictly phonetic representation of their own variety. In the domain of creole languages, there is a tendency to want to read Creole without having learned to read it, a tendency, moreover, that is in line with the mistaken notion that there is no need to learn to read if a language is represented phonetically. 24 We should not be surprised, therefore, by the difficulties that Guadeloupeans have in reading MC, which makes many more distinctions and is therefore farther removed from the source language than is GC. It should be clear, then, that the variation that is present in all Creole texts is simply a fact that ought not to be ignored. There is no single, correct way to address the questions that it poses, especially since, in the case of historical texts, it is very difficult to evaluate variation with precision. What is important above all is to invite caution on the part of readers or anyone who might try to analyze historical texts, because by neglecting variation, one runs the risk of drawing hasty conclusions about the state of the language.

23

We point out in our Internet course (http://creoles.free.fr/Cours) that Pipi’s statement to the mayor of Fort-de-France in Chamoiseau’s (1986) Chronique des Sept Misères—“Ebyen misié limè, séti manmay la té fin, danne ! . . .” (“Well, Mr. Mayor, these little children were hungry, damn it!”)—is written incorrectly (or at least, the way it is written reveals a false interpretation of the linguistic facts): The correct rendering would be sé ti-mamay-la, that is, ‘plural marker + ti-mamay (“children”) + definite marker’ for “the children,” or, if one prefers not to separate the plural marker from the noun, sétimamay-la. However, one cannot write séti, then later, mamay, which creates a false division of a group that the author has misinterpreted from a grammatical point of view. 24 The reader should recall that a functional written representation of a language cannot be strictly phonetic; it must also include certain syntactic features (words joined or separated, coherent groups, etc.) or morphological ones.

44


5. Conclusion The goal of this brief study has been simply to discuss a few major methodological principles for the analysis of historical texts. In conclusion, we wish to emphasize that just as the knowledge we already have of the popular French of the eighteenth-century facilitates the task of those wishing to learn about early creoles, acquiring increasingly accurate knowledge of these creoles also informs our knowledge of earlier spoken and popular French. For there are but a limited number of hypotheses about the origins and the earlier stages of these languages that could plausibly lead to the latest stages, which are well known. To be sure, we do not always have the ideal historical perspective. It is also regrettable that some texts are neither dated nor precisely localized geographically. We do have, however, clues to help situate them more precisely and to make it possible to carry out the indispensable task of reconstruction. It is thus important above all to research as thoroughly as possible these early texts and to arm oneself with the philologist’s patience in order to give voice to these corpora, which have often remained hidden for centuries and are sometimes discovered only by chance in the archives.

References Barbotin, Maurice. 1995. Dictionnaire du créole de Marie-Galante. Kreolische Bibliothek, Band 15. Hamburg: Helmut Buske. Baudot, Paul. 1980. Œuvres créoles (2nd ed.; Maurice Louis Edmond Martin, Trans.). Basse-Terre: Association des amis de la Bibliothèque centrale de prêt de la Guadeloupe; Pointe-à-Pitre: Association des archivistes, bibliothécaires et documentalistes francophones de la Caraïbe, Section Guadeloupe (Original text published 1935). Bernabé, Jean. 1983. Fondal-natal. Grammaire basilectale approchée des créoles guadeloupéen et martiniquais, 3 vol. Paris: L’Harmattan. Bernabé, Joby. 1984. Dabò pou yonn. Unpublished manuscript. Camier, Bernard & Marie-Christine Hazaël-Massieux (eds). 2003. Jeannot et Thérèse de Clément. Un opéra-comique en créole à Saint-Domingue au milieu du XVIIIème [Edited manuscript]. Revue de la société haïtienne d’histoire et de géographie 215 (April-September 2003). 136-166. Chamoiseau, Patrick. (1986). Chronique des sept misères. Paris: Gallimard. Denis, Serge (ed.) 1935. Nos Antilles. Paris: Maison du Livre français. Durand, Oswald. 1970. “Choucoune.” Rires et pleurs (poésies). Nendeln, Liechtenstein: Kraus Reprints (Original work published 1896). Goux, M. 1842. Catéchisme en langage créole précédé d’un essai de grammaire sur l’idiome usité dans les colonies françaises. Paris: Impr. H. Vrayet de Surcy. Available at: http://creoles.free.fr/Cours/goux1.htm.


45

Gratiant, Gilbert. 1976. Fab’ compè zicaque: Poésies originales antillaises en créole avec leur traduction française en regard. Fort-de-France: Désormeaux. ---------- & Marcel Cohen. 1958. Fab Compè Zicaque (poèmes en créole). Fort-de-France: Editions des horizons caraïbes. ----------, Isabelle Gratiant, Renaud Gratiant, & Jean-Louis Joubert. 1996. Fables créoles et autres écrits. Paris: Stock. Hazaël-Massieux, Guy. 1996. Les créoles: problèmes de genèse et de description. Aix-en-Provence: Publications de l’Université de Provence. Hazaël-Massieux, Marie-Christine. 1993. Ecrire en créole: Oralité et écriture aux Antilles. Paris: L’Harmattan. ----------. 2000. “Des références textuelles pour l’étude de l’évolution grammaticale des créoles dans la zone américano-caraïbe et de leur utilité pour l’étude historique. La question du déterminant.” Etudes Créoles 23. 40-65. Lhérisson, Justin. 1975. Zoune chez sa ninnaine: Fan’m gain sept sauts pou li passé (2nd ed.). Port-au-Prince: Editions Fardin (Original work published 1906). ----------. 1975. La famille des Pitite-Caille: Les fortunes de chez nous (2nd ed.). Port-au-Prince, Haiti: Editions Fardin (Original work published 1929, Paris: Firmin Didot). Marbot, François Achilles. 1846. Les bambous: fables de La Fontaine, travesties en patois créole. Fort-Royal-Martinique: E. Ruelle & Ch. Arnaud imprimeurs, Librairie de Frédéric Thomas. Nainsouta, Rémy. 1940. Le langage créole. Guadeloupe, Basse-Terre: Imprimerie officielle. Sylvain, Georges. 1971. Cric? Crac! Fables de la Fontaine, racontées par un montagnard haïtien et transcrites en vers créoles. Nendeln, Leichtenstien: Kraus Reprints (Original work published 1929, Port-au-Prince).

LEXICAL ASPECTS OF FRENCH AND CREOLE IN SAINT-DOMINGUE AT THE END OF THE EIGHTEENTH CENTURY

Pierre Rézeau Centre National de la Recherche Scientifique

ABSTRACT: An unpublished, anonymous manuscript, written around 1800 and held in Béziers, contains numerous observations on the language of the period in Saint-Domingue and the Antilles. A considerable number of these observations concern flora and fauna. Around 100 of the most interesting terms have been selected from this document for commentary. A brief characterization of each is given according to a typology of lexical variation. The original entries written by the text’s anonymous author are then provided, complemented by additional observations on the history of each term. The commentary includes carefully researched information about each lexeme’s earliest attestation, comparisons with cognate terms from other places in the colonial world, and etymological notes. These data provide an interesting historical snapshot of a specialized part of the lexicon of Haitian French. 1. Introduction On the eve of the French Revolution, a Frenchman from Languedoc traveled to New England, to the West Indies, and along the northwest coast of Colombia, settling for a period of time on a plantation in Saint-Domingue. His stay on this island, at that time the most prosperous of the colonies, would make quite a strong impression on him, and upon his return to his native village of Saint-Hippolyte-du-Fort (Gard) around the year 1800, he composed a bilingual Languedocien-French dictionary which he punctuated with numerous lengthy digressions about the language, flora, fauna, and customs of the West Indies, and especially of Saint-Domingue. Unexpected in such a work, this body of observations is appreciable both in quantity and in quality, which makes it all the more regrettable that only half of the manuscript of this dictionary survives to the present day, most accessibly in a photocopy with a few introductory pages (Bazalgues 1974). Its anonymous author, who must have been a doctor or a druggist, was an excellent observer, an enlightened commentator on language, and a scholar.

48


1.1 Characterization of the linguistic observations In the majority of cases, the author considers the features he reports to be widespread, and he introduces them with formulae such as “In the West Indies/in Saint-Domingue, this is called…” At times he specifies that a given word is characteristic of creole languages (such as amacorner, calumet see cachimbeau) or of Creole peoples (such as coucouye and hasiers), thus taking the point of view of an observer from France, and since his notes bear mostly on the lexicon, he does not see Creole as a linguistic system distinct from French. The only cases where he makes such a distinction—for example, when he refers explicitly to “Creole language” (see the entry maman) or when he twice puts an observed fact in the mouth of the “Negroes”—are matters where the lexicon is less in focus than the morphological or syntactic environment. In some cases, the author uncovers facts in Saint-Domingue which had nothing particularly exotic about them and which he could just as easily have gleaned in France outside of his own region. This is undoubtedly the case with terms such as bille “marble,” feuillard “hoop (of a barrel, cask),” and merveille “fritter, doughnut.” In this article I will leave aside a number of terms which are wellestablished in the West Indies and for which available documentation does not make it possible to improve on the lexicographic information already given by the major reference works. 1.2 Main sources of lexical variants not shared with Standard French This article will look at terms from the manuscript that are not shared with Standard French. Some of these terms are of French origin, but differ in their use. Others are terms that come from regional varieties of French in France. Still others are borrowed from other languages, usually Amerindian or African languages or Spanish. 1.3 The handling of these lexemes in general lexicographic works It is hardly surprising that this lexical material is often ignored by modern lexicography, insofar as the best French dictionaries often have trouble describing lexical variation in French inside of France. Any attention they might pay to overseas French is therefore sporadic and haphazard at best. For example, although TLF and ROB 1 record acomas, they do not mention mapou, and though the same dictionaries give the West Indian sense of cerise, they leave aside that of abricot. This self-centered attitude gives a very restricted view of linguistic reality.

1

Dictionaries used most frequently in this article are: Trésor de la langue française (TLF), Le Grand Robert de la langue française, 2nd ed., (ROB), and Französisches Etymologisches Wörterbuch (FEW).


49

In this article, it will be useful to take inventory of the most interesting facts catalogued by the anonymous lexicographer from Saint-Hippolyte-duFort. The entries, after the headword and part of speech, do not usually include a definition since the contextual examples themselves provide a kind of definition. A short commentary will then seek to give the oldest West Indian attestation of the term under consideration, taking into account a number of publications dealing with Saint-Domingue, especially from the eighteenth and nineteenth centuries, and to indicate whether and in what reference dictionaries it is treated (by ‘reference dictionaries’ I mean TLF and ROB). 2. Items from the Saint-Hippolyte-du-Fort manuscript 2.1 Terms of French origin differing from Standard French 2.1.1 Meaning change by semantic restriction. allure n. f. “In the West Indies, the amble [a gait] is called allure, but allure [“gait”] in French is a collective word that does not designate any particular gait, for a horse has good and bad gaits when trotting as when ambling” (Bazalgues 1974:34a) and “the amble, which is called allure in this country” (Bazalgues 1974:215b). ♦ First (?) attestation. Attested also in

1809, “These horses ... trained for the allure or the amble” (Descourtilz 1809/1935:91). Perhaps from Spanish ambladura; compare train, found in this sense by Jourdain (1956:137). — Absent from the reference dictionaries; to be added to FEW 24, 420b, AMBULARE. 2 calebasse n. f. calebasse sauvage or calebasse marronne. “There is another small squash, which is produced by a creeping and climbing plant called wild calabash (calebasse sauvage or calebasse marronne). Negro women, who in the state of slavery can hardly have any affection for the fruit of their womb, use a mixture of this to destroy it, sometimes paying for it with their own lives in the process, which has happened several times on our plantation in the space of a year” (Bazalgues 1974:147a-b). ♦ Perhaps related to calebasse zhèbe noted

in Jourdain (1956:273). — Absent from the reference dictionaries and from FEW 19, 85a-b, QAR’A. entourage n.m. “In that country [the West Indies], another enclosure is made, called entourage [“fence”]. It is constructed of stakes driven into the ground and of poles called gaulettes in the region, placed across the stakes and attached to them with ties called lianes locally” (Bazalgues 1974:93b). ♦ By analogy with French entourage, “that which sur-

rounds something” (only since 1780, TLF), this use is attested at the end of the 2

The Französisches Etymologisches Wörterbuch is a monumental etymological dictionary of French including, in principle, all Galloromance dialects. When I say that a word is to be added to the FEW at such and such a place, I mean that the FEW entry treats very closely related senses or forms and therefore it is clear where the Bazalgues attestations would logically go. A word described as absent from the FEW, on the other hand, is a word or sense not treated at all in the FEW.

50


18th c. in Saint-Domingue: 1776, “make sure that the surrounding woods [l’entourage du bois] are tight” (Debien 1956:51) and “Wild pineapple [...] This plant is used to make entourages that the Negroes and the livestock never dare to cross” (Nicholson 1776:149). Its presence on Réunion since 1734 indicates that it was already common in France (see Chaudenson 1974:693). — Absent from the reference dictionaries; Telchid (1997); to be added to FEW 13/2, 54a, TORNARE. garni adj. f. “cased in wicker.” “Many persons in the north as well as in the south of France and in the West Indies say une bouteille garnie [“a garnished bottle”], which is heard for garnie de paille, de jong [sic] [“garnished with straw, with cane”], etc.” (Bazalgues 1974:397a). ♦ Absent from the reference dictionaries and from FEW 17, 531a,

*WARNJAN. habitant n. m. “The habitant is a small bird the size of our wren and whose song or rather cry is about the same. Its plumage is Saxony green [a glossy light green color; Editors’ note] and its crop is blood red” (Bazalgues 1974:51b). ♦ Meaning absent from the reference

dictionaries and FEW 4, 369a, HABITARE. habitation n. f. “agricultural estate.” “It is rare in the West Indies to find an habitation to rent: the landowners improve them themselves and the absentees give their proxy to someone to manage them. For this they grant to the legal representative ten percent of all the year’s revenues” (Bazalgues 1974:63a). ♦ Attested since 1654 in reference to the West Indies,

“habitations in very good condition, well planted, furnished with good cabins, & with all sorts of tools to cultivate the land” (Du Tertre 1654:21). — Ditchy (1932/1977); absent from TLF and ROB; Jourdain (1956:197); Chaudenson (1974:598); Telchid (1997); Valdman et al. (1998), abitasyon; FEW 4, 369b, HABITARE. maille n. f. “uno anêlo is yet a link or ring of a chain that in Saint-Domingue they call a maille” (Bazalgues 1974:41a). ♦ By restriction of use, from French maille “ring of

metal” (since the eleventh c., TLF). — To be added to FEW 6/1, 14b, MACULA. 2.1.2 Meaning change by extension. cabri n. m. “In Saint-Domingue, the nanny goat and the billy goat are called cabri. In a herd of goats in which nanny goats, billy goats and kids are mixed, one would say What a lot of cabris” (Bazalgues 1974:164b). ♦ Attested also in 1809, “the wild goat [le cabrit

marron]” (Descourtilz 1809/1935:66); by extension from French cabri “kid, baby goat,” attested since the fourteenth c. (TLF). — Ducœurjoly (1802:304) gives Creole cabrit as equivalent to French chèvre; this meaning is not given for the West Indies in the reference dictionaries; to be added to FEW 2, 296b, CAPRA; Chaudenson (1974:716); Faine (1974) et Bentolila (1976), kabrit;


51

Valdman (1981) gives Creole kabrit as equivalent to English goat; Telchid (1997). coque n. f. “Sugarcane is subject to several diseases caused by insects that are called pucerons, and coque. The former gnaw on its roots and the coque, which is a kind of white bug, attaches itself to the leaves, sucks out the substance and makes the plant wither; then one gets little sugar of a very bad quality. To destroy them, one must set fire to the cane” (Bazalgues 1974:184a-b). ♦ Attested since the eighteenth c., “In 1748 ... there appeared

swarms of a kind of butterfly which had been showing up since 1744 and which reproduced prodigiously by means of a kind of larva that in the Colony they called Coque” (Moreau 1797/1984). — Absent from the reference dictionaries; to be added to FEW 2, 824b, COCCUM, where this sense is lacking. galerie n. f. “In the West Indies, one calls galeries covered street sidewalks, eight feet wide, that each person who builds is required to construct in front of his house for the convenience of the public and so that people go by sheltered from the rain and the sun” (Bazalgues 1974:424a). ♦ Attested since the 18th c., “The houses [of Léogane] have galer-

ies or sidewalks from 5 to 8 feet wide, but they are neither straight nor even, and consequently they are very often useless to passers-by” (Moreau 1797/1984:1096). — Absent, in this use, from the reference dictionaries and from FEW 4, 31b, GALILAEA. mama(n) n. “In the West Indies where, out of a kind of contempt, one calls the Negro and Négresse slaves papa, mama or maman, one says [to a man who carries eggs and to a woman who carries a basket] papa aux œufs [“egg Papa”], maman au panier [“basket Mama”] and in the Creole language papa z-œu, mama panié” (Bazalgues 1974:388a). ♦ Extension of

French maman “mother,” also noted on the Ile de France in 1812 (Chaudenson 1974:796). — 1820 (Brasseur [Hérault] 1991:35); Jourdain (1956:115); use absent from the reference dictionaries and FEW 6/3, 134b, MAMMA. ménagère n. f. “Ill-favored by fortune in general, the people of color (gens de couleur) rarely marry among themselves, though they are free. The women of that class, ambitious and coquettish, loving to shine in beautiful clothing and jewels, prefer to attach themselves to white men, who hold all riches, and live publicly with them under the name of ménagères [“housekeepers”] and congratulate themselves for having children by them, persuaded that it will one day result in a benefit to them and to their children” (Bazalgues 1974:239a). ♦ Attested in

the eighteenth c., “Most of them [les Mulâtresses “the Mulatto women”] live with a white man, where, under the little-deserved title of ménagères, they have all the functions of a wife” (Moreau 1797/1984:106); “the ménagères, that is, when their age permits, the acknowledged concubines of most white bachelors” (Wimpffen 1797/1993:120); semantic extension of French ménagère “wife” (since the seventeenth c., FEW). — “The men, in order to follow

52


the fashion of the country, live ordinarily with a woman of the country, called a ménagère” (Ducœurjoly 1802:82); 1816 (Brasseur [Hérault] 1991:37); absent from the reference dictionaries; to be added to FEW 6/1, 189b, MANERE. papa n. m. See mama(n) above. ♦ Extension of French papa “father.” — 1823 (Brasseur [Hérault] 1991:39); 1878 (Rézeau & Rézeau [Massé] 1995); Jourdain (1956:115); Chaudenson (1974:824), pa; usage absent from the reference dictionaries (ROB “français d’Afrique [“French of Africa”]”) and from FEW 7, 588b, PAPPUS.

2.1.3 Meaning change by analogy. abricot n. m. “fruit of the Mammea americana L.” “In Saint-Domingue the term abricot refers to a fruit the size of a large ball, which shares with our apricot only its color. It has three stones a little bigger than a pigeon egg, which are useful for curing horses of farcy [a skin disease of horses (Editors’ note)]. The tree which bears this fruit is a lovely tree; its leaf resembles that of the bay-tree” [Bazalgues 1974:46b]. ♦ Attested since the eighteenth c.: before 1730,

“the fruit which they call mamei and which we call the St-Domingue apricot” (Le Pers c.1730:f.38v°); 1797, “the gigantic apricot” (Moreau 1797/1984:435). — “The apricot of the West Indies bears no resemblance to that of Europe....” [Creole: z’abricot] (Ducœurjoly 1802:285); Jourdain (1956:289), zabricot; cf. Faine (1974), “The Haitian apricot tree is a large tree found in a number of tropical lands in the Americas, where it is known as the mamey de Santo Domingo”; Valdman (1981); Telchid (1997); absent from the reference dictionaries and from FEW 8, 284a-b, PRAECOQUUM. bigarade n. f. “What is there [in the West Indies] called a bigarade must not be the same species as in Europe, for they are ordinarily sweeter than those which are called sweet oranges (oranges douces): they are small, red like oranges from Portugal, somewhat flattened on the two ends, the navel and stem, and their skin is thick and bumpy” [Bazalgues 1974:462a]. ♦

Also attested in 1824, “Bigarades, crudely called sour oranges (oranges sûres [sic]) in the West Indies and bitter oranges (oranges aigres) in Europe” (Tussac 1808-1827 (3):67-68). — This usage is absent from the reference dictionaries and from FEW 4, 65b, *GARR-. bombe n. f. “The kettle is called bombe in the West Indies, but this word is not French with this meaning” [Bazalgues 1974:234a]. ♦ Attested since 1766 in Quebec with this

meaning, in Poirier (1998), which indicates that “this usage is doubtless connected to bombe ‘spherical glass vase’ which was current in France in the eighteenth c. (see Trévoux 1752 and TLF).” — “In Creole, a kettle is called a bônme” (Faine 1974); to be added to FEW 1, 431a, BOMBUS.


53

cerise n. f. “fruit of the malpighie”. “What is called a ‘cherry’ (cerise) in Saint-Domingue is a fruit that perfectly resembles the Languedocian agrioto in shape, color and taste. The tree also greatly resembles our agrioutié, but the difference between the two fruits is large when you consider the inside, after eating the flesh of them: in one, you are left with a round stone in your hand, and in the other, three small triangular seeds, which puts it in the ranks of seedbearing fruit” (Bazalgues 1974:212b). ♦ Attested since 1654, “little black cherries

rather similar to the wild cherries (merises) of Europe. In the middle of the fruit, there are three rather soft pits” (Du Tertre 1654:247); 1776, “This berry appears to the eye to be totally like our French cherry: it is attached to a little stem, and contains three striated, winged pits” (Nicholson 1776:205). — “The fruit called (by analogy) cerise in Haiti is not the same as the one in France of that name” (Faine 1974); Valdman (1981), seriz; TLF and ROB, cerise des Antilles; absent from FEW 2, 600a CERASEUM. charpentier n. m. “The charpentier [“carpenter”] is a kind of woodpecker or green woodpecker, as big as ours, with a plumage speckled white and black; it has the same cries and the same habits, in other words it climbs trees to bore into them with its beak, etc.” (Bazalgues 1974:52a) and “The charpentier (bird of the West Indies) is a kind of woodpecker, whose plumage is mottled with black and white, primarily on its wings. It has the same habits and the same call as our brayos-roujos [pic rouge]” (Bazalgues 1974:152a). ♦ Attested from

Corneille 1694 to Larousse 1866-1876, “wood-boring bird of the island of Saint-Domingue,” in FEW 2, 399a, CARPENTARIUS; “our green woodpecker […] after the Spanish Carpintero, we call it Charpentier” (Le Pers c.1730:f.28v°). — Absent from the reference dictionaries. jardin n. m. “In the West Indies, there are no other enclosures [but the quickset hedge] for the fields of sugarcane, coffee, cotton, for the pastures, etc. They are made three or four feet wide, planted with young seedlings of the tree called campêche and of the lemon tree or sown with their own seeds. These hedgerows are trimmed at least once a year and kept breast high. When they are well maintained and straight, they give a pleasant air to the fields. Thus they [the fields] are called jardin [“garden”] in that country, for one says, ‘How many Negroes do you put in the garden? I put so many Negroes in the garden to say I have so many Negroes cultivating my sugar, my coffee, my cotton’, etc.” (Bazalgues 1974:93a). ♦ Use well-attested in

the seventeenth (Rochefort 1658:306) and eighteenth c. (Wimpffen 1797/ 1993:169), notably in nègre de jardin [“field Negro”] (used for agricultural labor), as in Nicolson (1776) (see under entourage). — Absent from the reference dictionaries and FEW 16, 19a, GARD; Telchid (1997), jardin-créole. merle n. m. “The merle [“blackbird”] resembles that of Europe: it has the same size and plumage. Its beak is black, it is good to eat and travels in flocks” (Bazalgues 1974:52a); see also piperi. ♦ Attested since the seventeenth c., “There are also several black-

birds, thrushes, ortolans, and hawfinches, almost all of them similar to ours of

54


the same name” (Rochefort 1658:153). — Meaning absent from the reference dictionaries and FEW 6/2, 36a, MERULA; cf. Jourdain (1956:30); Peleman (1976) and Valdman (1981), mèl. orme n. m. “The orme [“elm”] in the West Indies does not grow quite as high as that of Europe. Its wood is of a medium hardness, its leaf somewhat resembles that of the grape vine and completely that of the cotton bush, because at first glance clumps of ormes are often taken to be cotton bushes. The seed of the orme is black when ripe, round and the size of a sorbapple, but woody like that of the alder-tree, which it resembles. It serves as food for horses, mules, etc., whereas in Europe only the leaf of the elm is given to livestock” (Bazalgues 1974:125b). ♦ Attested at the end of the eighteenth c.: 1776, “Bois d’orme [“elm

wood”]. Synon. Orme [...]; its flowers [...] become a spherical fruit, at first green, then black, in clusters [...] that the horses are fond of” (Nicolson 1776: 184-185); 1797, “a row of Saint-Domingue ormes” (Moreau 1797/1984:1096). — Meaning absent from reference dictionaries and FEW 14, 5b, ULMUS. ortolan n. m. “The ortolan is a small type of turtle-dove that is often found in pairs in the big roads. It is a little larger than a sparrow. It is a little lump of fat and for this reason good to eat” (Bazalgues 1974:52a); see also pipari. ♦ Attested since 1654, “The same must be

said of the ortolans of Martinique, which are small turtle-doves, which are no bigger than meadowlarks” (Du Tertre 1654:316); before 1730, “what we incorrectly call ortolans are only a kind of turtle-dove smaller than the others” (Le Pers c.1730:f.28v°). — Usage absent from TLF (which defines the word as a “variety of small sparrow of Europe”) and from ROB (“...of southern Europe”); cf. Jourdain (1956:30); Bentolila (1976), Peleman (1976), and Valdman (1981), zòtolan; to be added to FEW 4, 488b, HORTULANUS. perdrix n. f. “What is called perdrix [“partridge”] in that country [Saint-Domingue] is rather a kind of turtle-dove” (Bazalgues 1974:51a). ♦ Attested since 1654, “There are in

Guadeloupe [...] three kinds of perdrix [...], which to my mind were never anything but turtle-doves” (Du Tertre 1654:315). — Meaning absent from the reference dictionaries and FEW 8, 226b, PERDRIX; cf. Jourdain (1956:30); Valdman (1981), pèdri. raquette n. f. “figuier de Barbarie [“prickly pear,” Opuntia antillana (Editors’ note)]” “The nopal or figuier d’Inde, called raquette in the West Indies and roquette or cardasse or opuntia by the Academy” (Bazalgues 1974:446b). ♦ Attested since 1647, (Breton, in Petitjean

1980:280); 1654 (Du Tertre 1654:163;) 1658 (Rochefort 1658:107); FEW, which attests it since 1704, also found it in Saintonge but the historical data are lacking to suggest that this meaning was spread to the French of the West Indies from this region. — “RAQUETTE, n. f., oily plant, covered with thorns, a kind of nopal [“prickly pear”], whose fruit, green on the outside and red on the


55

inside, though of good quality, gives a bloody color to those who have eaten it” (Ducœurjoly 1802:348) [same word in Creole]; TLF; ROB; FEW 19, 144a, RAHA; Jourdain (1956:286); Telchid (1997). rossignol n. m. “the only one that sings somewhat pleasantly is called rossignol [“nightingale”] by comparison with the one from here. The American in New England calls it mokeur bird [sic], that is, oiseau moqueur, claiming that it imitates the song of all the other birds [In a note: This is Bomare’s poliglotte]. It sings at night like our nightingale, varies its song a lot, and has a rather pretty throat. It is as common as ours, or more so: always around the houses, sometimes on their rooftops, it hardly ever leaves inhabited areas. We call this common [familier] even though the saying is rossignol sauvage [“wild nightingale”]. The comparison ends with the plumage: the latter is nothing like the nightingale of Europe, it has grey and white feathers, a long tail, it nearly always jumps and it is a little bigger than our nightingale” (Bazalgues 1974:51a-b). ♦ Attested since 1654, “The bird that the inhabitants call

Rossignol, is very rare in Guadeloupe. It is rather similar to the wren of Europe; but it is a little bigger. It is the only one of all the birds that I have seen in the Indies that has a beautiful song” (Du Tertre 1654:318); before 1730, “another bird that is called rossignol as much for the beauty of its song as for some physical resemblance it bears to that of Europe” (Le Pers c.1730:f.29); 1797, “There are yet here some other kinds of birds, one of which is honored by the inhabitants with the name rossignol, because of its rather sweet song; very different from that of the one whose name it assumes” (Wimpffen 1797/1993:162). — Absent from the reference dictionaries and from FEW 5, 471b, *LUSCINIOLUS; cf. Jourdain (1956:30); Peleman (1976), ròsignòl; Valdman (1981), woziyòl, resiyòl. têtard n. m. “The têtard [lit., “tadpole”] in Saint-Domingue is a river fish so called because of its big head. It has the shape of our chabots [genus Cottus (Editors’ note)], about the size of an ordinary carp. Its flesh is soft, though rather good” (Bazalgues 1974:166b). ♦ Attested in

French since 1560 (TLF, “vieilli [“archaic”]”; ROB, “regional”), but in reference to a fish of France, while the example that the TLF gives applies to Louisiana (Baudry des Lozières 1802). — Peleman (1976), téta; to be added to FEW 13/1, 278b, TESTA. 2.1.4 Meaning change by metonymy. malingre n. m. “The word malingre in the mouth of the Creole of the West Indies signifies ulcer” [Bazalgues 1974:461b-462a]. ♦ Attested since the end of the eighteenth c.:

1775, “malingres [that are] often difficult to heal” (Debien 1962:125); 1787, “some fevers, some malingres” (Debien 1959:61). The presence of the word also in Réunion indicates that it was in use in France, where, however, it has never been found in this meaning. Chaudenson proposes to see in it the avatar of Middle French malendre “wound, ulcer” (FEW 6/1, 81a, MALANDRIA),

56


which a popular etymology could have brought together with the old slang malingre “beggar who with false wounds seeks to stir compassion” (FEW 6/1, 124b-125a, MALUS). — “MALINGRE [...] in Creole is any type of wound” (Ducœurjoly 1802:331); “[French] Plaie [“wound”] [Creole] Malingre” (Ducœurjoly 1802:342); “[French] Ulcère [Creole] Malingres” (Ducœurjoly 1802:353); Jourdain (1956:68), melingue; Chaudenson (1974:799); meaning absent from the reference dictionaries and FEW, loc. ci0074. 2.1.5 Meaning change by metaphor. cousin n. m. “In the West Indies, cousin is the name given to various plants similar to the burdock [Arctium lappa, Editors’ note] having, like the latter, flowers gathered into a flaky head, but smaller, flatter or rounder, armed with small hooks that for this reason easily cling to the clothing of passersby” (Bazalgues 1974:252a). ♦ Attested since 1654, “We have a

plant on this as on all plantations on these Islands, which we call Cousin, because of its seed, which is bristly and which clings to the clothes and hair of passersby: it is not much bigger than the head of a big pin” (Du Tertre 1654:163); 1776, “Cousin […] a capsular fruit […], armed with little spikes by means of which it attaches to the clothes of passersby” (Nicolson 1776:226227). — Meaning absent from the reference dictionaries and from FEW 2, 1074a, CONSOBRINUS; Jourdain (1956:275). décoré past part./adj. “On the other hand, the Negro contributes greatly to his getting dealt these kinds of punishments by his natural dispositions to theft, and to running away, for it is rare to see a Negro ‘decorated’ (décoré), as we call it, only by the whim of his master if he has not deserved it through some major offense” (Bazalgues 1974:130b). ♦ The decoration

referred to here is the branding with a fleur de lys which, along with the cutting off of the ears, was used to punish a slave the first time he ran away. — This use is absent from the reference dictionaries and from FEW 3, 26b, DECORARE. fou n. m. “In the seas of the West Indies there is found a bird called fou [“crazy”] locally, as well as by the Academy. The latter says that it resembles the crow of Europe, but we can verify that it has neither the breast nor the color of the crow; the crow is a beautiful black, and the fou, far from it, is not even a beautiful blue, nor much larger than a blackbird, whereas the crow is almost as big as a chicken” (Bazalgues 1974:399b). ♦ The bird in question is

manifestly not the Sula, the only one that is in fact called fou in the dictionary of the Academy and in the reference dictionaries; since 1725, according to TLF; already in 1654, “certain birds called Fregates & Fous” (Du Tertre 1654:93); this may be a case in which the author is mistaken. musicien n. m. “The musicien is a singular bird, which no one has seen and which many people say they have heard in the most isolated woods. It is said that it sings several musical notes in a row with the greatest possible accuracy” (Bazalgues 1974:51a). ♦ Attested since


57

1797, “the bird called musicien because of its brilliant voice and the ease with which it modulates several musical notes with a precision that charms man” (Moreau 1797/1984:164). — Recorded in Bescherelle 1845 as “name of diverse species of bird” (FEW), this meaning is absent from the reference dictionaries; to be added to FEW 6/3, 264b, MUSICA. procureur n. m. “The avacado[...], having a big pit, or rather seed, similar to the horse chestnut, called procureur [“prosecutor”] in amusing style” (Bazalgues 1974:416a-b). ♦ First (?)

attestation in this metaphorical meaning. — “it [the avocado] contains a big pit that is derisively called procureur” (Ducœurjoly 1802:72); absent in this meaning from the reference dictionaries and FEW 9, 416a, PROCURATOR. Tropique (avoir passé le –) loc. verb. “Happy he who does not inspire this proverbial expression Il a passé le Tropique [“He has passed the Tropic”], that is, he is witty, cunning, frivolous, and even crude” (Bazalgues 1974:364a). ♦ Figurative locution absent from the

reference dictionaries and FEW 13/2, 323b, TROPICUS. But “passer les deux Tropicques [“to pass both Tropics”]” (Rabelais, Tiers Livre, chap. 51) perhaps has a double meaning and would then be a precursor of this expression. 2.1.6 Probably of onomatopoetic origin. piperi, pipiri n. m. “Birds in the West Indies are, in general, small in quantity and among this number few are good to eat: one eats only blackbirds, ortolans, pipiri or piperi, nightingales, parrots, etc.; the latter two are always lean” (Bazalgues 1974:51a). ♦ No earlier attesta-

tions have been found. — Jourdain (1956:31), pipiri “Myobius audax (Cyml.)”; cf. Telchid (1997), pipirite “bird of the early morning”; absent from the reference dictionaries. toc-toc n. m. “The toc-toc, a bird that is all feathers” (Bazalgues 1974:52a). ♦ Absent from

the reference dictionaries; cf. FEW 13/2, 14b, TOKK-, toque-bois “green woodpecker” (Marne). 2.1.7 Derivation. calebassite n. f. “Calebassite (diminutive of calebasse) is a little round fruit, about the size of a pomme d’api apple, produced by a climbing plant and suitable for making arbours, cradles, etc. Its soft, sourish, rather tasty flesh is covered by a hard, brittle shell. Its flower is the passion flower, so distinctive in its shape” (Bazalgues 1974:147b). ♦ Attested since the

eighteenth c.: 1797, “Jérémie also has some cacones. There they are called callebassites” (Moreau 1797/1984:1407); 1809, “two little calebacites, one containing powder, the other shot [for the hunt]” (Descourtilz 1809/1935:4243). — Absent from the reference dictionaries and from FEW 19, 85b, QAR’A.

58


gaulette n. f. See under entourage. ♦ Derived from gaule [“pole”], attested since the middle of the 15th c.; in the West Indies since 1654, “they sew little gaulettes onto this oakum” (Du Tertre 1654:440); 1722, “trellis made of little gaulettes” (Labat, in Chaudenson 1974:774): cf. gaulettes “small gaules [“poles”] onto which tobacco is attached, in the West Indies, to dry it” (Trévoux 1752-1771, FEW). — TLF with no date, with a quote from J. Green (1935); ROB, “rare”; Telchid (1997), gaule; to be added to the FEW 17, 495b, *WALU. marronnage n. m. See décoré. ♦ Attested since 1671 in Saint-Domingue, marronnage, (Quemada & Rézeau 1970-1998); 1797, “desertion, which here is called marronnage” (Wimpffen 1797/1993:188). — Jourdain (1956:125); TLF cites an example from H. Grégoire (1826); ROB. recourir v. tr. “They say in Saint-Domingue it is necessary to recourir [lit., “to run again”] these cane stalks, these cotton bushes, these coffee shrubs, etc., that is to say, replace the seedlings where the first ones failed to grow, but recourir in this meaning is not French. However, in that country, one has nothing better to say than to use adopted terms if one wishes to be understood” (Bazalgues 1974:172a). ♦ Attested since 1776, “if I had had little

enough experience in the cultivation of indigo to recourir in large indigo” (Debien 1956:58). — 1823, Brasseur [Hérault] (1991:45); absent from the reference dictionaries; to be added to FEW 2, 1570b, CURRERE (which only notes this meaning since 1910, in French-speaking Switzerland). 2.1.8 Use of Compounding. araignée n. f. “The crab spider (araignée crabe) of the West Indies, whose body is almost as wide as the palm of the hand, is of a reddish brown color entirely covered with hair; it usually stays in damp places, where there are dead leaves that have fallen from trees, in underbrush, etc. It spins no web. Its bite is said to be very poisonous; what is certain is that it must sting quite hard, since it has two very pointed fangs on top of its head. Some people believe it to be the Apulian tarantula, whose bite causes the victim to fall into a deep slumber from which he is cured only by shaking vigorously, but this is erroneous: the former spider is not at all like this” (Bazalgues 1974:370a). ♦ Attested since the end of the eighteenth c.: 1776, “the

crab spider itself, however monstrous and vigorous it might be, is as quickly overcome as attacked [by the mongoose]” (Nicolson 1776:364); 1797, “crab spiders or red-bottomed spiders [à cul rouge], sometimes cause painful bites” (Moreau 1797/1984:265); 1809, “a hideous crab spider” (Descourtilz 1809/1935:98). — Absent from ROB in this meaning, this lexical item is erroneously classified in TLF under the label “a kind of crustacean similar to the crab” (with a misinterpreted example from Morand 1941); it is absent from FEW 25, ARANEUS.


59

arbre-homme n. m. “The arbre-homme [“man-tree”], if it exists, is one of those plants which, by virtue of its attributes, must be regarded, along with the sensitive plant, as a sentient step in the progression from plant to animal. It is said that when a person approaches this tree, the bulk of it trembles. I might add that its sap is as red as blood” (Bazalgues 1974:120a). ♦

Absent from the reference dictionaries, including from FEW 25, 88a-b, ARBOR. banane n. f. banane musquée / banane-cochon “Three species of bananas are known in the West Indies: the figue-banane [“fig-banana”], the banane musquée [“musk banana”] and the banane-cochon [“pig banana”], thus called because it is the least good of all although the biggest and the most common” (Bazalgues 1974:123b). ♦ Also attested in 1808, “the

musk banana, smaller than the others, but having a more delicate taste” (Tussac 1808-1827(1):63); cf. bananier-cochon and bananier musqué (Nicolson 1776:159). — Absent from FEW 20, 86b, BANANA. bois de fer loc. nom. m. “The trees called bois de fer [“ironwood”], acajou [“mahogany”], and acomas [“false mastic”] are the best for construction” (Bazalgues 1974:46a) and “The bois de fer is the best wood of all for construction, for burning and for making good charcoal. If someone wishes to keep a fire burning a long time, he covers a log, preferably of bois de fer, with ashes. Since this wood is not given to rotting, it is commonly used for making wooden posts to go into the ground. Its name is an indication of its hardness: it is the hardest of all woods in that country and perhaps also in Europe. Most nails bend when you try to drive them into this wood, and if the nail is not of a good caliber and well-proportioned, there is no point in proceeding, unless it has been passed through wax” (Bazalgues 1974:118a-b). ♦ Attested

since 1654, “the tree [...] which our colonists call bois de fer, because of its great hardness” (Du Tertre 1654:229), the term often designates several varieties (Rochefort 1658:72); “Bois de fer [...] Two species are identified, the white wood and the red” (Nicolson 1776:174), but it has also sometimes been used as a generic term for acomas and mahogany (cf. Petitjean 1980:305); as the name of a West Indian species, it is absent from the reference dictionaries and the FEW 15/1, 206a, *BOSK-; Jourdain (1956:270); Chaudenson (1974:256). bois de lance loc. nom. m. “a kind of wood in Saint-Domingue, called bois de lance [“lancewood”)] We do not know the shape of this tree, we know only that its wood is of medium hardness, and that carpentry uses it for many things, but especially for making billiard cues” (Bazalgues 1974:128a). ♦ Attested at the end of the eighteenth c., “Bois de lance

[...] It is used for making chairs, ladders and other similar furniture” (Nicolson 1776:179-180); “In the south there is the bois de lance, the bois trompette [“trompetwood”]” (Moreau 1797/1984:1330). — Absent from the reference dictionaries and the FEW 15/1, 206a, *BOSK-. bois espagnol loc. nom. m. “The heart of the bois espagnol [“Spanish wood”] is also good for making wooden posts, and the bois de brin of this wood, when it is with its bark and its sap-

60


wood, is good for making stockade stakes, because it is easily cut” (Bazalgues 1974:118b). ♦ Attested at the end of the eighteenth c., “Bois espagnol” (Nicolson 1776:174). — Absent from the reference dictionaries and the FEW 15/1, 206a, *BOSK-. (bois) patate loc. nom. m. “There are no artificial meadows in the West Indies, but rather natural ones around farmers’ houses, which are called savanes, where they keep their horses and mules all year, so that the latter are at pasture year-round in the countryside. In towns and on trips they are no less so, being fed with freshly cut plants and especially with the shoots of the sweet potato vine, or bois patate as it is called in the country, and of which they are quite fond” (Bazalgues 1974:57b); “The bois-patate, technically, is little more than a vine that perfectly resembles grapevines both in its wood and in its leaf. It is climbing and it produces a kind of potato that is very good to eat, which has the taste of the chestnut. This fruit is the ordinary food of the Negroes and colored people of that country, and many whites set great store by it. Horses and all animals of whatever kind are very fond of the sweet potato, some of them of its wood, and all of them of its fruit” (Bazalgues 1974:119a); and “The fodder of horses in the country about which we are speaking, whether in cities or on trips, is what they locally call bois patate, sugar cane heads, stalks of millet, that are always given fresh and such as have just been cut from the plant, so that horses there are at pasture all year long” (Bazalgues 1974:216a). ♦ Attested since 1654, “those climbing stalks which the farmers

call bois de Patates” (Du Tertre 1654:185); 1776, “Patate [...] The stalks are given to horses, and are called bois patate; they are used as fodder; the roots are the everyday food of Negroes” (Nicolson 1776:287); 1797, “The fodder which serves as food for animals are the sweet potato, in other words the stalk and the leaves of that root; that of the millet when it is green, Guinea grass [l’herbe de Guinée]” (Wimpffen 1797/1993:138). — Term absent from the reference dictionaries and from FEW 15/1, 206a, *BOSK- and 20, 57b, BATATA. bois trompette loc. nom. m. “The bois-trompette [“trumpetwood”] is a soft wood, which doubtless draws its name from the fact that the trunk of the tree as well as its branches and even the stem of its leaves are hollow and thereby suitable for making trumpets” (Bazalgues 1974:119a). ♦ Attested in the eighteenth c.: before 1730, “a large tree which the

Indians named jasuma and which we call trompette because of some resemblance of its branches to that instrument” (Le Pers c.1730:f.14v°); 1776, “Bois trompette” (Nicolson 1776:190). — Absent from the dictionaries consulted as well as from FEW 15/1, 206a, *BOSK-; Jourdain (1956:271). bout de pétun/de tabac loc. nom. m. “The bout de tabac or de petun is of the size of a blackbird; it is entirely black like the latter, but it appears bigger when it is flying, because it has a lot of feathers. It has an unpleasant cry, and on its hooked beak it has a sharp bump. It is common, it sometimes travels in groups, but it is always thin, it is not good to eat” (Bazalgues 1974:51b-52a). ♦ Attested since 1654, “There is also in Guadeloupe a very large

number of little black birds which are quite similar to blackbirds, the farmers


61

call them bout de petun” (Du Tertre 1654:317). — Absent from the reference dictionaries; to be added to FEW 15/1, 218a, *BOTAN. case n. f. 1. grand-case “the master’s hut, called the grand-case” (Bazalgues 1974:255b). ♦ Attested since the eighteenth c.: 1767, “la grande case” (Debien 1962:35); 1797, “a grand’case which is falling into ruin” (Wimpffen 1797/1993:125). — Ducœurjoly (1802:331) gives grande caze as the Creole equivalent of maison principale. This lexeme and the following one are both absent from the reference dictionaries and from FEW 2, 451a CASA. 2. case à bagasse “The roofs of houses in the West Indies are of three kinds: slate, wooden shingles, and sugarcane stalks. The first two are ordinarily used in towns and the third, prohibited in towns, is much used in the countryside, for Negro huts, and for cases à bagasse” (Bazalgues 1974:255b). ♦ Attested

since the eighteenth c., “little gusts of wind that unroof and knock over Negro huts, and even cases à bagasse” (Moreau 1797/1984:219). — Ducœurjoly (1802), see bagasse “It [bagasse] is stored in great sheds that are called caze [sic, given in the singular] à bagasse”; Jourdain (1956:198). chou n. m. 1. chou (palmiste) “Atop the trunk [of the cabbage tree], a cluster of leaves about 8 or 9 feet long form a splendid panache, which is topped by an upright plume rising from its center which is called the ‘cabbage’ (chou), which is good to eat” (Bazalgues 1974:120b). ♦

The lexeme chou (de) palmiste [“cabbage tree”] is attested since 1654, “the marrow or brains, that the inhabitants call chou palmiste” (Du Tertre 1654:235; Rochefort 1658:62), but already called chou in 1603 (Champlain, in Petitjean 1980:282). — TLF and ROB; Jourdain (1956:92, 274); to be added to FEW 2, 536b, CAULIS. 2. “The seeds of this cluster, which come out of the center of the tree [the banana tree] and at the end of which is a heart-shaped cluster of purple leaves called the chou of the bunch, resemble in shape our mélongènes or eggplant, though not in color, for they are green until they have completely finished growing, and yellow when they are ripe” (Bazalgues 1974:123b). ♦ This usage is absent from the reference dictionaries. 3. chou caraïbe “The chou caraïbe in the West Indies is a plant whose long-stemmed leaf is not eaten, but at its roots it produces, like the potato, several tubers which are very good to eat. It is one of the foodstuffs of that country and the name of this cabbage reveals that it is originally from there” (Bazalgues 1974:189b). ♦ Attested since 1654, “Of two kinds of cabbage that are

called Kareïbes” (Du Tertre 1654:157). — TLF, see caraïbe (undated; quote from Hugo, 1826); absent from ROB; to be added to FEW 2, 526b, CAULIS; Jourdain (1956:92, 274); Telchid (1997). figue-banane n. f. See banane. ♦ To name a variety of banana, figue (d’Inde) is attested since 1605 in a text concerning Guinea (Chaudenson 1974:560); 1647 (Breton, in Petitjean 1980:295); 1654 (Du Tertre 1654:204), but figue des Indes already in 1579, cf. Chaudenson (1974:560), who sees it as a calque of Indo-Portuguese). — “the figue banane is shorter, it is only eaten ripe”

62


(Ducœurjoly 1802:297); TLF, with a citation from Bernardin de Saint-Pierre (1814); ROB, with no reference; Jourdain (1956:276) and Telchid (1997), figue; absent from FEW 3, 496a, FICUS. figuier d’Adam loc. nom. m. “The bananier [“banana tree”] or figuier d’Adam, which is another type of reed about 15 feet high, must be regarded as the most useful of all the trees of that country by its fruit alone. This fruit, called banane, and the sweet potato, are the principal food of the Negroes and people of color, which they sometimes prefer to wheat bread” (Bazalgues 1974:123a). ♦ Attested since 1658, “Some have found this fruit [banana] so

beautiful and so delicate, that they have imagined it to be that of the earthly Paradise, which God had forbidden Adam and Eve to eat. Thus they name it Figuier d’Adam [“Fig tree of Adam”], or Pommier de Paradis [“Apple tree of Paradise”]” (Rochefort 1658:92-93; cf. Petitjean 1980:296). — Absent from the reference dictionaries and from FEW 3, 496b, FICUS; Telchid (1997), “former name of the banane [sic for bananier “banana tree”].” mil (petit –) n. m. “In the West Indies they say a grain of petit mil [“small millet”] and pronounce petit mi” (Bazalgues 1974:445a). ♦ This lexical item, here probably in con-

trast to gros mil [“large millet”] ‘corn,’ is attested since the seventeenth c.: 1658, “de gros Mil called Mais [“corn”], de petit Mil & and several others” (Rochefort 1658:437); 1775, “Le petit mil is a subsistence food” (Debien 1962:106); 1776, “petit mil ordinaire [“ordinary small millet”]” (Nicolson 1776:274). — Ducœurjoly (1802:323), see habelin “de la farine de millet, ou petit mil [“millet, or small millet flour”]”; to be added to FEW 6/3, 83a-b, MILIUM. orange n. f. 1. orange douce “l’orange douce [“sweet orange”], highly prized in Europe” (Bazalgues 1974:416b) and “The oranges douces of the fields & the gardens of these islands are highly prized in Europe” (Bazalgues 1974:462a). ♦ Attested since 1515, in the

translation of a text concerning the East Indies (Arveiller 1963:370); 1654, “Those who are fond of oranges douces [...]” (Du Tertre 1654:239); 1658, “As for oranges, there are two kinds in the West Indies. [...] Some are sweet, & the others sour, both are extremely delicate” (Rochefort 1658:47). 2. orange sure “The bitter oranges are called oranges sures in the West Indies, which amounts to the same thing, because sur [“sour”] and aigre [“bitter”] are almost synonymous” (Bazalgues 1974:462a). ♦ See above in bigarade; Jourdain (1956:290). 3. orange sure douce “Many bitter oranges are found in the woods of these islands [the West Indies]. There is still an abundance of those called oranges sures douces, which are indeed very sweet even though they are wild” (Bazalgues 1974:462a). ♦ These lexemes are absent from the refer-

ence dictionaries and FEW 19, 138a-b, NARANG(A).


63

oseille de Guinée loc. nom. f. “sour hibiscus.” “In the West Indies, preserves are made from a species of sorrel, called oseille de Guinée” (Bazalgues 1974:23b). ♦ Attested in 1722 in

Larousse 1866-1876 (FEW). — absent from TLF; ROB only makes reference to Africa; Jourdain (1956:281); FEW 24, 106b, ACIDULA. piment de chien loc. nom. m. “Le piment originates in America. It is found in Saint-Domingue in the roads, in the woods and everywhere, but it is uncultivated, the fruit is not as big as an ordinary olive, somewhat longer and thinner at the end. In the country it is called piment de chien” (Bazalgues 1974:247b). ♦ Attested at the end of the eighteenth c., “the

Piment à chien, which is flattened” (Nicolson 1776:288). — Lexeme absent from the reference dictionaries and FEW 8, 446a, PIGMENTUM. pois verts loc. nom. m. pl. “In Paris, des haricots verts [“green beans”], and in the West Indies des pois verts” (Bazalgues 1974:385a). ♦ The term pois was long used in the creoles

and the varieties of French in America to designate various vegetables “of twenty species” (Moreau 1797/1984:435) and it still is, as in certain regional varieties in France (such as in the West), to designate kidney beans. — Absent from the reference dictionaries; to be added to FEW 8, 606a, PISUM, where this lexeme is missing in this meaning; Chaudenson (1974:841); Telchid (1997). quatre piquets n. m. “The main punishment of the Negroes in the West Indies is the quatres piquets [“four stakes”]. It consists of three or four stakes driven into the ground, to which the Negro’s limbs are attached in order to whip him on the buttocks” (Bazalgues 1974:205b). ♦

First (?) attestation of this lexeme. — Absent from the reference dictionaries and from FEW 8, 454a, *PIKKARE. 2.1.9 Frequency of use. These items in this section enjoyed more frequent use in Saint Domingue than in France. “In the West Indies, the Creoles call the donkey by hardly any other name than bourrique” (Bazalgues 1974:70a). “Gaspiller is the verb used in the West Indies in the sense of to lavish, to waste” (Bazalgues 1974:373b). “The headcovering of Creole women, especially women of color, is the mouchoir” (Bazalgues 1974:228b). “In the West Indies, tombereaux are simply small carts, on which a barrel of sugar is customarily placed; they are pulled by five mules which normally trot” (Bazalgues 1974:200a). 2.2 Terms from regional varieties of French in France bigaille n. f. “In Saint-Domingue, there is a very small and almost undetectable black fly, properly called bigaille in the country, which bites hard. However the name bigaille is also regarded there as a generic term for insects which fly and which are bothersome, for one often hears people say ‘There is a lot of bigaille in this area’, and the reference is as much to the mosquito, which they call a maringoin, as to the gnat properly called a bigaille” (Bazalgues

64


1974:109b). ♦ Attested since 1722 (Labat, in Chaudenson 1974:705); 1809, “At

night, the noise of the bigaille (as those flying torturers [maringouins] are called) alone troubled my sleep” (Descourtilz 1809/1935:97). TLF and ROB give it with examples from 1826 (drawn from the novel Bug-Jargal by V. Hugo, set in Saint-Domingue) and with the customary remarks, “familiar” and “old-fashioned [vieux]” respectively, whereas this is primarily a geographically marked sense; the word is also used as a patronymic in Haiti. Also to be rejected is the association of bigaille with the word *GALLIUS (FEW 4, 42a), found in these dictionaries; we assign it rather to the etymon *PIKKARE (ibid., 8, 458b; similar forms under WIBBA 17, 575b should be added here as well). — “BIGAILLE, n. f. a very small gnat, usually in mangrove swamps or other marshy ground, very bothersome because of its sting” [same word in Creole] (Ducœurjoly 1802:299); Peleman (1976), bigay “tiny fly”; Valdman (1981), bigay “gnat, mosquito”; FEW 22/2, 294, n. 21. botte (tomber en –) loc. verb. “In the West Indies they say that a barrel tombe en botte when it comes apart on its own, when it falls apart. […] They also speak of making it tomber en botte, to refer to taking it apart” (Bazalgues 1974:327a). ♦ Attested since 1782 in west-

ern France (Rézeau 1984); Ditchy (1932/1977), “to faint, to swoon.” — Absent from the reference dictionaries and from FEW 15/1, 229b, BOTE. burgau n. m. “The burgau of the Americas, a shellfish from which the most beautiful motherof-pearl is taken, is shaped somewhat like our cagaraoulo [“snail”] but it is much larger and harder. It is claimed that the former inhabitants of Saint-Domingue ate them and that they were so abstemious that a single one was enough food for them for an entire day. But that seems rather fantastic to us” (Bazalgues 1974:170a). ♦ Attested since 1563 in Saintonge,

“and an infinite number of burgaux of various kinds and sizes” (B. Palissy, 1563:211-212), this word is never used in modern France except in the local speech of coastal Vendée and Charente-Maritime; Du Tertre (1654:430). — BURGAU, small shellfish found on rocks that are washed by the sea. They are cooked in water with salt, they are eaten with a pin” [the same word is used in Creole] (Ducœurjoly 1802:299); TLF; ROB; FEW 21, 266a-b ‘bigorneau’; Baldinger (1988-1998 (1):n° 751); Telchid (1997), burgot. caler v. tr. “The Creole, in Saint-Domingue, says caler (for écailler “to remove an eggshell”), which is wrong” (Bazalgues 1974:286a). ♦ This form (cf. Creole zécale “eggshell”)

probably came to the West Indies from Normandy. — “[French] ÉCALER [creole] Caler” (Ducœurjoly 1802:312); to be added to FEW 17, 77a, *SKALA. crebiche n. f. “Abundant crayfish are found in the rivers, streams and canals in the West Indies. In his language, the Creole calls them crebiches” (Bazalgues 1974:344a). ♦ First ( ?)

attestation of this form, which is probably of Norman origin (cf. creviche


65

“shrimp” Littré; FEW), still attested in Louisiana Creole; it is absent from the reference dictionaries. — Jourdain (1956:132), cribiche; Valdman et al.(1998); to be added to FEW 2, 298b, CAPRA. élingué adj. and n. “a tall, thin person, having a slender figure. Some people say an élingué, a grand élingué, especially among the Creoles of Saint-Domingue” (Bazalgues 1974:358a) and “In the West Indies he [a slim and ungainly man] is called an élingué, a grand élingué, a term that is expressive, but not French” (Bazalgues 1974:446a). ♦ Borrowed from the western

dialects (in particular those of Normandy, Brittany, Saintonge), the word also passed into the French of Canada and Louisiana (see Rézeau 1997:325). — Absent from the reference dictionaries; FEW 17, 147b, *SLINGA. expenter v. tr. “Many people say expenter [“to terrify” (Editors’note)] in French and in the same meaning as in the language Languedoc, especially the Creoles of the Antilles; it is not French” (Bazalgues 1974:356b). ♦ Doubled by épouvanter, which won out in Stan-

dard French, this type is still widely attested in the nineteenth c. in numerous regions of France, especially in the form épanter (FEW); expanter is doubtless a hypercorrection. — To be added to FEW 3, 304a-b, *EXPAVENTARE. grage n. f. “In the West Indies it [the grater, a household utensil] is called a grage, from the word grager [“to grate”]” (Bazalgues 1974:449a-b). ♦ Attested in this use since 1658,

“One first scrapes it [the cassava root] with a knife [...] & then one grates (grage, in the phrase of the country) it with a grater or a flat grage, of iron or copper” (Rochefort 1658:443; Breton 1665, in Friederici 1960). As Friederici (1960) rightly observes, Breton only uses grage and grager in French; it is in fact a term of Norman origin that was recorded in dictionaries from Trévoux 1752 to Augé 1948-1949, see FEW. — Absent from the reference dictionaries; Friederici (1960:262-263); Jourdain (1956:84); Telchid (1997); FEW 23, 40b ‘piler.’ grager v. tr. “Grager [“to grate”] the cassava or the cassava root to make from it the bread or the flat cake called cassave” (Bazalgues 1974:449b). ♦ Attested since 1403 in France in

reference to flax (FEW) and 1658 in the West Indies in reference to cassava (Rochefort 1658:443); 1797, “un moulin à grager” [“a grating mill”] (Wimpffen 1797/1993:125); see the preceding entry. — Ducœurjoly (1802: 347) gives this verb as the Creole equivalent of French râper [“to grate”], “Gragé. It is hardly said but for cassava – gragé manioc pour fair cassave [“grate manioc to make cassava”]”. Absent from the reference dictionaries; Telchid (1997); FEW 23, 40b ‘piler.’ hasier n. m. pl. “In the West Indies, the Creole people call them [les halliers “thickets”] hasiers” (Bazalgues 1974:96a). ♦ Attested since the sixteenth c. in Normandy

66


(FEW). — Rézeau & Rézeau [Massé] (1995/1879), braziers; Jourdain (1956: 22), raziers; Telchid (1997), razié; FEW 16, 121a, *HAISI. jambette n. f. “A jambette is a little folding pocketknife. The English make very pretty ones that are highly prized. In the West Indies, these are the only kinds of knives that are ever carried in the pocket, where they are not a great encumbrance, being no larger than our penknives” (Bazalgues 1974:93a) and “a jambette, which is a small knife whose blade folds into the handle, and that is hardly larger than a folding penknife” (Bazalgues 1974:175b). ♦

Attested since 1623 (see Littré), this meaning appears to be geographically marked according to FEW (in particular Brittany, Normandy, Anjou, Saintonge). — FEW 2, 114b, CAMBA; TLF without usage notes but with an example from the Rennes region; ROB “regional”; Jourdain (1956:179), canif d’écolier [“schoolboy’s penknife”]; Telchid (1997). nabot n. m. “A nabot, in the West Indies, 3 is an iron ring whose weight varies in proportion to the age and the strength of the Negro on whose leg it is placed, to prevent him from going marron [“maroon”] (a Creole expression that has become French), which means that he has run away, that he is a fugitive, that he has withdrawn to the woods to live there in freedom” (Bazalgues 1974:130b). ♦ Attested in Nantes since 1773, “the same chain links

two galley slaves together wearing a nabot on each foot” (Debien 1962:66). A mistaken division of un abot, this term is to be classified under FEW 15/2, 43a, *BUTT (also to be placed there are nabot “false link connecting the two fragments of a chain,” since Augé 1907; Nantes (Blain) “lock of the shackles of a horse, placed on one of its legs above the fetlock joint,” mistakenly listed ibid., 7, 8b NANUS). — Cf. TLF and ROB, abot “regional.” rabattre, rebattre v. tr. “to repair (a barrel).” “In the West Indies they say rabattre ou rebattre les barriques de sucre, de café, etc. [“to repair the barrels of sugar, of coffee, etc.”] in order to say to adjust them, to mend them, to put them in good condition ; but in France one must say, in speaking of barrels, relier [“to hoop”] barrels, hogsheads, etc.” (Bazalgues 1974:12a). ♦

Probably a word from the southwestern quarter of France, attested since 1777 in the French of the Sables-d’Olonne “futaille de deux barriques rabattues [“a barrel of two casks put together”]” (Gérard, 2002:200); 1818, “Relier un tonneau, et non, le rabattre [“relier a barrel, and not, rabattre (a barrel)”]” J.B.C., 4 Les Périgordismes corrigés, Périgueux: J. Danède, 54). — Absent from the reference dictionaries; to be added to FEW 24, 21a ABBATUERE.

3 A first version began as follows: “A nabot in the West Indies is one of those irons that shames humanity, and which will horrify posterity among sensible men”; the author struck this passage through, indicating in the lower margin: “It seemed to me that this article was in great need of revision.” 4 The author of this unpublished manuscript is known only by the initials J.B.C.


67

sous-tirer v. tr. “However, not all Negroes withdraw into the woods when they are fugitives: those accustomed to a certain part of the region, the Creoles accustomed to a softer life than the African imports, the Negro women [négresses], the Negro children [négrillons], wander around the towns or withdraw to the inhabited quarters they have frequented, sometimes also lured or sous-tirés as they say in the country by their relatives, by their friends to take advantage of their labor or to save them from the harshness of the punishment that is reserved for them by their master. To lure them in this way is a crime in the eyes of the law and of probity” (Bazalgues 1974:130b-131a). ♦ First (?) attestation of this pejorative use, attested

in the nineteenth c. in western France: Nantes soutirer “attirer [“lure”],” Charente-Maritime “to lure to one’s home with a fraudulent intention” and Angevin sourtirer “to draw, to lure, to seduce” (FEW 6/1, 405b and 418a, MARTYRIUM); the term is attested in the same period in the Ile de France (today, Mauritius) in the same context (Chaudenson 1974:869). — Absent from the reference dictionaries; Chaudenson (1974:869); Telchid (1997), soutirer “to conceal, to encourage misdeeds” and soutireuse “[a woman who is a] go-between, accomplice.” tille n. “The aissette is a tool used by coopers, clog-makers, etc. In the West Indies, everyone calls it tille” (Bazalgues 1974:26a); “In the West Indies, this tool is called tille” (Bazalgues 1974:192b). ♦ Term of vocabulary of marine carpenters, which spread from

Normandy. — Absent from the TLF; ROB, “technical”; to be added to FEW 17, 324b, TELGJA. 2.3 Borrowings from other languages acomas n. m. “The trees called bois de fer [“ironwood”], acajou [“mahogany”], and acomas [“false mastic”, Sideroxylon fœtidissimum] are the best for construction” (Bazalgues 1974:46a) and “The acomas is highly regarded as a hardwood for construction. It is still used in the building of ships” (Bazalgues 1974:118b). ♦ Attested since 1640, achomat (Bouton,

in Petitjean 1980:305); 1654 (Du Tertre 1654:223); 1658 (Rochefort 1658:69); a borrowing from Carib. — “ACCOMA, n. m., a large tree suitable for building, a very hard wood” [Creole: accoma] (Ducœurjoly 1802:287); Jourdain (1956:268), acomat; absent from Valdman (1981), akoma is included in the revised edition in progress; TLF; ROB; the FEW has no entry ACOMAS. cachimbeau n. m. “There is a type of fern in the West Indies from which pipe stems are made and which the Creole calls calumet, which are fitted to a bowl of baked clay, made locally and called cachimbeau” (Bazalgues 1974:388b). ♦ Attested in French at the end of the

eighteenth c., “Calumet franc [...] Its root is fibrous, and it bears several tubes. [...] These tubes are used in the Islands for smoking, by fitting them [...] to the bowl of a pipe made of earth baked in the sun, called a cachimbo” (Nicolson 1776:200); the reader will note in this example, as in the preceding one, the use of calumet by metonymy to denote a plant). — Ditchy (1932/1977), cachem-

68


beau; absent from the reference dictionaries; to be added to FEW 20, 64b, KIXIMA. caïeu n. m. “A caïeu in Saint Domingue and all over the West Indies is a kind of sardine, but a little smaller than the European one” (Bazalgues 1974:163b). ♦ Attested before 1748,

“On the second day of June, several people were poisoned by a kind of small sardine, that they call cayeux in the Islands” (Pouppé-Desportes 1770(1):108); 1797, “the sardine or cayeux” (Moreau 1797/1984:1239). — Noted in Encyclopédie in 1751, the word is absent from the reference dictionaries; “CAYEU, n. m. type of small sardine that is nonetheless eaten” [same word in Creole, though differing by one letter, cayeux] (Ducœurjoly 1802:302); Telchid (1997), cailli, cailleux, caillu “a little fish that lives in cailles [“coral reefs”].” calalou n. m. “In Saint-Domingue, the word calalou denotes a type of dish resembling our bourbouyado, but it is lighter in color; several greens go into, and especially a vegetable that they call calalou or gombeau, produced by a plant also called calalou. It is a stew specific to that country and which often gives rise to a party although it is of little consequence in itself. They say Let’s go have a calalou in such and such a place and sometimes this dish is followed by several others; it is like when Europeans say Come over and have some of my salad tonight. Among the greens in a calalou, one finds salted pork, fish, crab, etc. mixed in” (Bazalgues 1974:141a-b). ♦ Attested since 1751 as a dish (Encyclopédie, in TLF); 1797,

“okra (gombeau), from which the calaloux that the Creoles boast of are made” (Wimpffen 1797/1993:115); “CALALOU, n. m. a dish composed of several kinds of plants that are prepared like spinach, but in a more liquid form, and into which bananas, crabs, herrings, etc. are frequently mixed” (Ducœurjoly 1802:300). — TLF and ROB, only as the name of a dish; Jourdain (1956:94); Telchid (1997). cocojus/cucujus/coucouyo n. m. “The cocojus or cucujus of the West Indies, named coucouyo by the Creole people and with which one can easily read at night by the light which it emits from the two posterior stumps on its head, is a true beetle and not a bird as some dictionaries say, perfectly resembling our capricorn beetle, except for the antennae which it lacks” (Bazalgues 1974:82a). ♦ Term attested since the sixteenth c. (FEW); before 1730,

“a glowing fly that the Indians of this Island named cucuio” (Le Pers c.1730:f.29v°); 1797, “the glowing fly, that is here named coucouïe, from the Spanish cucuïos or cuyeros” (Wimpffen 1797/1993:235), borrowed from Carib via Spanish. —“COUCOUYE, n. f. insect the size of a cantharis, which flies about at night, emitting from its eyes a light that one would assume to be phosphorus” [same word in Creole] (Ducœurjoly 1802:306); to be added to FEW 20, 65b-66a, KUKÚI, where the forms coucouye and coucouyo are lacking.


69

coui n. m. 1. “In the West Indies the term calebasse [“calabash”] is used for a sort of fruit that resembles our pumpkins, whose insides are not eaten; but when ripe and hollowed out, it serves as a piece of household equipment for Negroes and people of color, and for many Whites who are poor or distant from the cities and who are not in a position to furnish themselves with vessels of any other material, and sometimes out of thrift. These sorts of vessels are called couis” (Bazalgues 1974:146b). ♦ Attested in this sense since the beginning of

the seventeenth c.: 1614 couy, (FEW); Du Tertre (1654:251); Rochefort (1658:433). — “COUI, n. m. half of a gourd or coconut, used as a drinking glass or even as a plate by Negroes” [in Creole: couï] (Ducœurjoly 1802:306); Brasseur [Hérault] (1991:23), couic (1822); the word is absent from the reference dictionaries; FEW 20, 66b, CUY; Jourdain (1956:85); Telchid (1997). grigri n. m. “The grigri, which resembles a little our white oaks, and which grows very high and wide, is yet a hard wood. It is used for casks, for sugar mill drums” (Bazalgues 1974:118b119a); “Grigri in the West Indies is a big tree” (Bazalgues 1974:451b). ♦ Attested since

the 18th c.: 1757 (Encyclopédie); 1797, “gris-gris with a diameter of four to five feet and several other sorts of very large wood suitable for sugar mills” (Moreau 1797/1984:728). — Absent from the reference dictionaries and FEW, 20. manchette n. f. “Cutlass [...] that in the West Indies is called manchette. The Negroes use them to trim the quickset hedges, etc.” (Bazalgues 1974:254b). ♦ Borrowing from Span-

ish machete, attested since 1704 in the form machette (Trévoux, see TLF); to be noted here is the nasalized form of the first vowel, absent from the reference dictionaries, still in use in Haiti in the twentieth c. (Faine 1974). — “MANCHETTE, n. f. In the Colony, one calls manchette a saber with a plain wooden handle, which is used to trim hedges” [same word in Creole] (Ducœurjoly 1802:332). mantègue n. f. “Lard [...] that everyone in the West Indies calls mantègue, from the Spanish manteca” (Bazalgues 1974:444a). ♦ Attested in this meaning in 1722 (Labat, in

Chaudenson 1974:617); 1808, “pork fat, which is called mantègue in this country” (Tussac 1808-1827 (1):61); borrowing from the Spanish mantegua. — Ducœurjoly (1802:349) gives “mantègue or mantèque” as the Creole equivalent of French sain-doux [“lard”]; absent from the reference dictionaries. mapou n. m. “The tree called mapou is the only one that, at a certain time of year, sheds all of its leaves in a short time, as do the trees of Europe, to grow new ones” (Bazalgues 1974:45b); “The mapou is a big tree whose wood is tender and of little use. It has the peculiarity of losing all its leaves at once, in a certain season, as do most of the trees in Europe. This observation can only be made by a European, and in fact few must have made it, for we have never heard anyone speak of it. From the wood of the mapou are made bowls, bathtubs, and other imple-

70


ments of this nature” (Bazalgues 1974:123b-124a); and “The leaves of the trees of the West Indies, like those in all hot regions, never fall, or, if they fall, it can only be by turns: some are shed in order to reproduce others and in such a manner that one cannot notice the transmutation or the slightest bare spot in the tree. The mapou alone, a tall tree and of tender wood, changes its leaves at one time like the trees of Europe” (Bazalgues 1974:392b). ♦ Attested

since 1645 (G. Coppier, in Petitjean 1980:311); 1654 (Du Tertre 1654:349), the word is of unknown origin (Friederici 1960:390). — Absent from the reference dictionaries and the FEW; Jourdain (1956:282); Faine (1974). marron, -onne adj. and n. 1. [In reference to a slave] “The marron [“maroon”] Negro in the towns, in the neighboring farmsteads cannot remain hidden long: sooner or later, sometimes on the first or second day of his escape, he is caught, because all who surround him have an interest in capturing him, the white like the black, and principally the local police” (Bazalgues 1974:131b-132a). ♦ Attested since 1658 (Rochefort 1658:322; TLF); borrowed

from the Carib mar(r)on, which itself is derived from Spanish cimarrón by apheresis. — Ducœurjoly (1802:310) gives it as Creole “Marron. It is said only of runaway slaves” and equivalent of French fugitif (1802:319); Rézeau & Rézeau [Massé] (1995/1880); Jourdain (1956:125); Chaudenson (1974:616617); ROB. 2. [In reference to an animal] “The cats in the West Indies se rendent souvent marrons [“often run away”], because of the ease with which they find food year round in the countryside” (Bazalgues 1974:206b). ♦ Attested since 1640 (TLF); the collocation se

rendre marron is attested since 1667 (Du Tertre 1654 in reference to a slave). — Rézeau & Rézeau [Massé] (1995/1880); Jourdain (1956:28); Telchid (1997); ROB. 3. [In reference to a plant] calebasse marronne [“wild gourd”]. ♦ The adj. marron is attested in numerous collocations in the domain of West Indian flora in the eighteenth c. (see also ananas marron at entourage). zombi n. m. “In Saint-Domingue, a will-o’-the-wisp, the goblin, is called a zombi, the zombi. One also says, ‘In such and such a place, there are zombis’. ‘Beware of the zombi’, one says to children” (Bazalgues 1974:304a). ♦ The word seems to appear here in French dis-

course; this would be its earliest attestation (already in Creole in 1797, “a Zombi tale [in a note: Creole word meaning spirit, ghost]” (Moreau 1797/1984:70). — Ducœurjoly (1802:349) gives it as a Creole word, equivalent of French revenant [“ghost”]; Rézeau & Rézeau [Massé] (1995/1879); absent from Littré, the word appears in TLF and ROB (since 1832); Jourdain (1956:53, 247); Telchid (1997). 2.4 Unknown or uncertain origin amacorner (s’) v. pron. “‘To whose place did he go to hole up [se fourrer]?’ The West Indian Creole says ‘Who the devil did he hook himself up to [s’amacorner]?’” (Bazalgues 1974:54b). ♦ This word of unknown origin is related to macorne “rawhide strap

for livestock.” Attested in 1776, “I am obliged to postpone sending the horned


71

livestock until tomorrow because yesterday I forgot to make the necessary rawhide straps (macornes)” (Debien 1956:49); Ditchy (1932/1977) gives it with the etymological definition “rawhide strap used to join two head of cattle by the horns.” — Brasseur [Hérault] 1991:14 (1820); Ditchy (1932/1977), macorner “to bind two things together; to put in pairs; se macorner, to live together without benefit of marriage”; Daigle (1984), se macorner; absent from the reference dictionaries. kasta n. m. “There are some travelers who call it [the mangrove or the mangrove-tree] or say that it is called paretuvier or paletuvier, in the East Indies, and kasta in the West Indies” (Bazalgues 1974:119b) and “in the West Indies it [tanner’s bark] is taken from the tree called mangle [“mangrove”] or kasta. This tanner’s bark gives the leather a red [tint] that is much brighter than that of the live oak and a much lower quality” (Bazalgues 1974:337a). ♦ Ab-

sent from the reference dictionaries and the FEW, 20. macoute n. “Reed baskets or cabas for horses; they could also be called sacoches [“saddlebags”] … In the West Indies they are called macoutes” (Bazalgues 1974:330a). ♦ Attested

before 1730, “from it [palm tree bark] are made various baskets, or pouches, that are loaded onto horses and that are very convenient for trips, we call them macoutes” (Le Pers c.1730:f.17). — Absent from the reference dictionaries and FEW; in the mid twentieth c., the term enjoyed a sinister notoriety in the word (tonton) macoute (see ROB). mal-fini, manfini n. m. “The manfini called by the people mal-fini is a bird of prey that is not quite as big as our falcons” (Bazalgues 1974:52a). ♦ Attested since 1654, “The Manse-

fenil is a powerful bird of prey, which in its shape and in its plumage bears so much resemblance to the eagle, that only its small shape can distinguish it” (Du Tertre 1654:313-314); 1658, “The Mansfeny is also a kind of small eagle that lives also on prey” (Rochefort 1658:159); before 1730, “the mansfenis or malfenis are somewhat like the falcon and the eagle” (Le Pers c.1730:f.29); 1797, “le mal-fini” (Moreau 1797/1984:265). The word is a “possible corruption of a form mans phœnix that is found in the first written accounts of the West Indies, sometimes transcribed mensfenil Falco spraverius caribæarum (Gm.)” (Jourdain 1956:31, see malfini). — Absent from the reference dictionaries; Telchid (1997), “frigate bird (seabird).” tache n. “from the sort of pod or husk that covers its [the cabbage tree’s] seeds, which are black and the size of a sloe, the Negroes make their bed; these husks are commonly known as taches [lit., “spots, stains” (Editors’ note)]” (Bazalgues 1974:121a). ♦ Attested since the

seventeenth c.: 1658, “The inhabitants of the islands call this bark [of the cabbage tree] Tache, & they use it for the roofs of their kitchens, & of the other small storehouses of their farmsteads” (Rochefort 1658:62); 1797, “magnifi-

72


cent cabbage trees. They provide beds and sheets for the Negroes. Their tâches [sic] are as much as 7 feet long by 3 feet wide; they can be beaten, they can be washed” (Moreau 1797/1984:1298); 1809, “wooden houses with surrounding porches many of which had roofs made of cabbage tree taches” (Descourtilz 1809/1935:26). — Absent from the reference dictionaries. 3. Conclusion At the intersection of history and geography, these few notes are simply intended to testify to the richness of a document that deserves a place among the other works devoted to Saint-Domingue from the same time period, namely those of Moreau and Ducoeurjoly. On both sides of the Atlantic, a number of other unpublished works which could shed light on the genesis of the French varieties of the Americas (or of French-based creoles), still await readers. This is the case with the manuscripts held by the Centre des Archives d'Outre Mer (CAOM) in Aix-en-Provence, or the numerous documents held in the departmental archives in Nantes and Bordeaux.

References Arveiller, Raymond. 1963. Contribution à l’étude des termes de voyage en français (1505-1722). Paris: Éditions d’Artrey. Augé, Claude (ed). 1907. Le Larousse pour tous: Nouveau dictionnaire encyclopédique (2 vol.). Paris : Larousse. Augé, Claude (ed). 1948-1949. Nouveau Larousse universel: Dictionnaire encyclopédique en deux volumes. Paris: Librairie Larousse. Baldinger, Kurt. 1988-1998. Etymologien: Untersuchungen zu FEW 21-23, Tübingen: Max Niemeyer Verlag. Band 1, 1988 ; Band 2, 1998. Baudry des Lozières, Louis-Narcisse. 1802. Voyage à la Louisiane, et sur le continent de L'Amérique septentrionale, fait dans les années 1794 à 1798. Paris: Dentu. Bazalgues, Gaston. (ed). 1974. Dictionnaire languedocien-français [letters AJ], Mimeograph. Republication of anonymous manuscript (c. 1800), 472 pp., kept in Béziers (Centre international d’études occitans), accompanied by preface by Bazalgues. Montpellier. Bentolila, Alain et al. (eds). 1976. Ti diksyonnè kréyol-franse. Port-au-Prince (Haiti): Éditions Caraïbes. Bescherelle, [Louis-Nicolas]. 1845. Dictionnaire national, ou dictionnaire universel de la langue française. 2nd ed. Paris: Chez Simon. Brasseur, Patrice. 1991. “Les régionalismes du français dans la correspondance de Sévère Hérault, colon nantais en Guyane (1805-1825).” Études Créoles 14. 9-52.


73

Chaudenson, Robert. 1974. Le Lexique du parler créole de la Réunion (2 vol.). Paris: Librairie Honoré Champion. Corneille, Thomas. 1694. Dictionnaire des arts et des sciences. Paris : J.B. Coignard Daigle, Jules O. 1984. A dictionary of the Cajun language. Ann Arbor, MI: Edwards Brothers. Debien, Gabriel. 1956. Études antillaises; XVIIIe siècle. Paris: A. Colin. ----------. 1959. Un colon sur sa plantation. Université de Dakar, Senegal: Publications de la Section d’histoire, 1. ----------. 1962. Plantations et esclaves à Saint-Domingue. Université de Dakar, Senegal: Publications de la Section d’histoire, 3. Descourtilz, Michel Étienne. 1935. Voyages d’un naturaliste en Haïti, 17991803. (J. Boulenger, ed) Paris: Plon (Original work published 1809). Ditchy, Jay K. 1977. Les Acadiens louisianais et leur parler. Geneva: Slatkine Reprints (Original work published Paris: Droz, 1932). Ducœurjoly, S.-J. 1802. Manuel des habitans de Saint-Domingue (vol. 2), Paris: Lenoir. Du Tertre, Jean-Baptiste. 1654. Histoire generale des Isles de l’Amerique, Paris: Jacques Langlois; Emmanuel Langlois. Encyclopédie : Diderot, Denis & Jean le Rond d’Alembert. 1751-1765. Encyclopédie ou dictionnaire raisonné des sciences, des arts et des metiers, par une Société des gens de lettres 35 vols. Paris : Briasson, David, Le Breton, Durand, Geneva : Samual Faulche. Faine, Jules. 1974. Dictionnaire français-créole, Montréal: Léméac. FEW: Wartburg, Walther von. 1922-2003. Französisches Etymologisches Wörterbuch: Eine darstellung des galloromanischen sprachschatzes (25 vol.). Bonn: Klopp, 1928; Leipzig-Berlin: Teubner, 1934 & 1940; Basel: Helbing & Lichtenhahn, 1946-1952; Basel: Zbinden, 1955-2003. Friederici, Georg. 1960. Amerikanistisches Wörterbuch und Hilfwörterbuch für den Amerikanisten 2nd ed. Hambourg: Gram, De Gruyter & Co. Gérard, Alain. (ed). 2002. Les Sables au temps de la grande pêche. Manuscrits de Collinet (1739-1782). La Roche-sur-Yon: Centre vendéen de recherches historiques (Mémoire de Vendée). Green, Julien. 1935. Journal Tome II : Derniers beaux jours (1935-1939). Paris: Plon. Grégoire, Henri. 1826. De la noblesse de la peau, ou, Du préjugé des blancs contre la couleur des Africains et celle de leurs descendans noirs et sangmêlés. Paris: Baudouin frères. Hugo, Victor. 1826. Bug-Jargal. Paris: Hetzel. Jourdain, Élodie. 1956. Le Vocabulaire du parler créole de la Martinique. Paris: Klincksieck.

74


Larousse, Pierre. (ed). 1866-1876. Grand Dictionnaire universel du XIXe siècle. Paris: Administration du Grand Dictionnaire Universel. Le Pers, Jean-Baptiste. c. 1730. Mémoires pour l’histoire de l’isle SaintDomingue. Unpublished manuscript. Paris: Bibliothèque nationale, ms fr. 8990. Littré : Littré, Emile. 1863-1872. Dictionnaire de la langue française. Paris: Hachette. Morand, Paul. 1941. L’homme pressé. Paris: Gallimard. Moreau de Saint-Méry, Médéric-Louis-Élie. 1984. Description topographique, physique, civile, politique et historique de la partie française de l’Isle de saint-Domingue. B. Maurel & E. Taillemite (eds). Paris: Société française d’histoire d’outre-mer (Original work published 1797). Nicolson, Father O.P. 1776. Essai sur l’histoire naturelle de Saint-Domingue, Paris: Gobreau. Palissy, Bernard. 1563. Recepte veritable, Keith Cameron (ed), Genève: Droz, 1988 (Textes littéraires français; 359). Peleman, Lodewijk. 1976. Dictionnaire créole-français. Port-au-Prince (Haïti): Bon Nouvèl. Petitjean Roget, Jacques. 1980. La Société d’habitation à la Martinique (2 vol.). Lille: Atelier de reproduction des thèses de Lille III. Poirier, Claude. (ed). 1998. Dictionnaire historique du français québécois. Laval (Quebec): Les Presses de l’Université Laval. Pouppé-Desportes, Jean-Baptiste. 1770. Histoire des maladies de SaintDomingue (2 vol.). Paris: Chez Lejeay. Quemada, Bernard & Pierre Rézeau. (Series eds). 1970-1998. Datations et documents lexicographiques. 2nd series. 48 vol. Paris : Klincksieck, CNRS-INaLF. Available online at http://www.atilf.fr Rabelais, François. 1546. Le Tiers Livre. Paris : De l'imprimerie de Michel Fezandat. Rézeau, Dominique & Pierre Rézeau. (eds). 1995. De la Vendée aux Caraïbes. Le Journal d’Armand Massé (1878-1884), missionnaire apostolique. Critical edition. Paris: L’Harmattan. Rézeau, Pierre. 1984. Dictionnaire des régionalismes de l’Ouest entre Loire et Gironde. Les Sables-d’Olonne: Le Cercle d’or. ----------. 1997. “Toward a lexicography of French in Louisiana.” French and Creole in Louisiana, Valdman (ed) 1997, (London: Plenum). 315-332. ROB: Robert, P., Rey, A., & Morvan, D. (eds). 2001. Le Grand Robert de la langue française: Dictionnaire alphabétique et analogique de la langue française 2nd ed. (6 vol.). Paris: Dictionnaires Le Robert. Rochefort, César de. 1658. Histoire naturelle et morale des Iles Antilles de l’Amérique. Rotterdam: Arnould Leers.


75

Telchid, Sylviane. 1997. Dictionnaire du français régional des Antilles (Guadeloupe, Martinique). Paris: Bonneton. TLF: Imbs, Paul (vol. 1-7) & B. Quemada (vols. 8-16) (eds). 1971-1994. Trésor de la langue française. Dictionnaire de la langue du XIXe et du XXe siècle. Paris: Gallimard. Available online at http://www.atilf.fr Trévoux. 1704: Dictionnaire universel françois et latin, contenant la signification et la définition tant des mots de l’une & de l’autre langue ... le tout tiré des plus excellens auteurs ... & glossaires qui ont paru jusqu’ici en différentes langues. Abbaye de Trévoux, 1704 ; Paris, 1721 ; Paris, 1732 ; Paris, 1743 ; Paris, 1752 ; Paris, 1771. Tussac, Fr.-Richard de. 1808-1827. Flore des Antilles (4 vol.). Paris: Selfpublished. Valdman, Albert et al. 1981. Haitian Creole-English-French dictionary. Creole Institute, Bloomington: Indiana University. ----------, Thomas A. Klingler, Margaret M. Marshall, & Kevin J. Rottet. 1998. Dictionary of Louisiana Creole. Bloomington: Indiana University Press. Wimpffen, Alexandre-Stanislaus de. 1993. Haïti au XVIIIe siècle. Richesse et esclavage dans une colonie française. P. Pluchon (ed). Paris: Karthala (Original title: Voyage à Saint-Domingue pendant les années 1788, 1789, 1790 par le baron de Wimpffen. Original work published 1797).

THE LEXICALIZATION–GRAMMATICALIZATION CONTINUUM*

J. Clancy Clements University of New Mexico

ABSTRACT: Instead of viewing grammaticalization as a one-way development, this study suggests that it be understood as a lexicalizationgrammaticalization continuum, on which two separate but interdependent, bidirectional clines operate: (a) one cline whose endpoints are the semantics/pragmatics of a lexical item on the one hand, and the function of a grammatical marker on the other, and (b) another cline whose endpoints are the unbound form on the one end and the affix on the other. Based on a variety of data, including numerous examples from pidgin and creole languages, developments along both the meaning and the form clines are argued to be chaotic in nature, whereby a chaotic process is defined as a process that under iteration shows a high degree of regularity, though it is ultimately unpredictable. 1. Introduction By far the most frequently encountered path of grammaticalization starts with a lexical item (including some closed-class words such as demonstratives, interrogatives, and some pronouns) which over time becomes a grammatical marker of some sort through reanalysis and a host of other modifications. This process is shown in (1). 1 (1)

A lexical item

B...C >...>

D grammatical marker

The process is not discrete in the sense that a lexical item may retain its lexical status and, for instance, evolve simultaneously as an auxiliary, as in the case of English have, which is used as a full-fledged lexical item with the meaning ‘possess’ and as a perfect-tense auxiliary that optionally cliticizes. Thus, a more representative picture of grammaticalization might be (2), where some of the ways different stages could co-occur appear in (3). *

This is a substantially reworked, expanded, and updated version of Clements (1995). In this study, grammaticalization is considered only at the level of the word, not at the level of entire syntactic structures. For discussion of grammaticalization of structures, see Haspelmath (1998).

1


78 (2) A lexical item Æ

B lexical items used in specific contexts Æ

C

D

clitic syntax Æ

affix morphology

(adapted from Hopper & Thompson 1993:94-95) (3) a. A Æ {A / B / C / D} b. A Æ {A / B / C} Æ {B / C / D} c. A Æ {A / B} Æ {B / C} Æ {C / D} (adapted from Hopper & Thompson 1993:36)

ÆD ÆD ÆD

Particular cases of (2) would be those in which a relational noun (4) or a verb (5) becomes a case-marking or a TMA-marking affix. (4) relational noun Æ adposition Æ agglutinative case affix Æ fusional case affix (5) full verb Æ auxiliary Æ TMA clitic Æ TMA affix (adapted from Hopper and Thompson 1993:107-108) Of course, a grammaticalizing element may stop evolving formally at some point along the path towards affixation. In other words, a given lexical item may become a clitic and then show no evidence of further formal evolution. Typical examples of this are Wackernagel-type clitics, which tend not to become affixes, but remain clitics (cf. Nevis 1986a, 1986b and references therein). Again, English have is a good example of this because it functions as a full verb and also as an auxiliary. As an auxiliary, it is optionally realized as the clitic [v] or [z] (e.g. you've come and he’s come). Although it is dangerous to hazard predictions, it is safe to say that auxiliary have is not likely to become an affix in the near future. One important point regarding grammaticalization is that it happens along a continuum (or cline), on which one or more stages are typically represented at the same time, as illustrated in (3) above. That is, a given item may be represented at any point from stage A to stage D, or at various points simultaneously on the continuum at any given time. The representations in (2-5) imply that the grammaticalization process is unidirectional, whereby an element moves away from being a member of the lexical component toward membership in the morphosyntactic component of the grammar. The issue of directionality on the grammaticalization cline can be addressed on two levels with the help of the following questions: Regarding form, is it possible for an affix to become a clitic or a free morpheme? Concerning semantics, is it possible for a grammatical marker to become a discourse marker or even a lexical item such as a noun or a verb? We will attempt to provide answers to these questions below. In the commonly accepted definition of grammaticalization, the process whereby a lexical item comes to be used as a grammatical marker is considered


79

gradual. As Bruyn (1996) insightfully points out, it is useful to distinguish at least three types of grammaticalization: (a) ordinary grammaticalization, which is a gradual and largely or wholly a language-internal process; (b) instantaneous grammaticalization, which proceeds much more rapidly and which according to Bruyn is attributable to transmission discontinuity and communicative pressure in a situation in which a new creole language is forming (e.g., Sranan wan “a(n)” from the English numeral one); and (c) apparent grammaticalization, whereby a grammaticalized feature that appears in a new creole language is the result of a process of grammaticalization in another language (e.g., Sranan gi “give” used in the benefactive sense “for” is a grammaticalization from an African language that was transferred into Sranan) (Bruyn 1996:3942). The elements in (6a, b, c) are arguably cases of instantaneous grammaticalization, while (6d) would represent the initial phase of instantaneous grammaticalization (Clements 2003). The cases in (6e, f, g) are interesting because we know that Portuguese já “already” and logo “afterwards, immediately” were already found in pidgin Portuguese as quasi TMA markers (Clements, in press; Naro 1978). Thus, these could be cases of instantaneous or of apparent grammaticalization. As our discussion of the lexicalization–grammaticalization continuum develops, we will keep these distinctions in mind. (6) a. Tok Pisin pinis COMPLETIVE (< English finish) (Holm 2000:186) b. Korlai Creole Portuguese p(V)- PRONOMINAL OBJECT MARKER (< clitic pər “PRONOMINAL OBJECT” < Portuguese preposition para “for”) 2 (Clements, forthcoming) c. Haitian Creole French te PAST (PERFECT) (< French été “been”) (Muysken & Veenstra 1995:155) d. Chinese L2 Spanish ya incipient PAST marker (< Spanish ya “already”) (Clements 2003) e. Mahim Creole Portuguese ja PAST (< Portuguese já “already”) (Dalgado 1906) f. Batticaloa Creole Portuguese ja PAST (< Portuguese já “already”) (Smith 1978) g. KCP l\ FUTURE (< Portuguese logo “afterwards, immediately”) (Clements 1996:112)

2

The Korlai Creole Portuguese (KCP) object pronominal system is: par(m)i “me” pn “us” pr “you-SINGULAR-INFORMAL” pudzo “you- PLURAL” -----puse “you- SINGULAR-FORMAL” pel ‘him, her’ pelo “them” The KCP form pər is found only in the combination pər k°~ (< Portuguese para quem “for whom”) “whom.”

80


In the last 25 years, an extensive amount of work has been carried out in the area of grammaticalization (see the recent Heine, Claudi, & Hünnemeyer 1991; Hopper & Traugott 1993; Traugott & Heine 1991 and references therein). Due to the fruits of this research, the general mechanisms of the grammaticalization process have become fairly well known. Areas related to grammaticalization which are less well understood involve the question of what the developmental path of an element is, semantically as well as formally, once it has grammaticalized. In what follows, we will propose an extended model of grammaticalization, using the general chaotic–theoretical notions of iteration, regularity, and unpredictability. Drawing on data from a number of languages, including creole languages, we will argue that: (a) an affix can develop into a free morpheme and, (b) a grammatical marker can develop into a lexical item. That is, we argue that the processes involved in grammaticalization are ultimately bidirectional. We posit a lexicalization–grammaticalization continuum, on which there are major and minor pathways. The development from a lexical item toward a functional element is more frequently attested than that from a functional element to a lexical item. Nevertheless, the latter do exist and there is a certain systematicity to the development. The data we will discuss also suggest that it is useful to make the distinction between the development of form and meaning of a given element on the lexicalization– grammaticalization cline. Although the form and meaning of an element are often interdependent, they can have separate developmental paths. 2. Traugott’s grammaticalization model Traugott (1982) argues that reanalysis in grammaticalization involves more than ‘semantic bleaching’ whereby a lexical item is gradually bleached of its semantic content while simultaneously acquiring a grammatical function. Given that “the normal condition of a speech community is a heterogeneous one” (Labov 1982:17), semantic bleaching would arguably take place over a given period of time. However, Traugott argues that even before or at the same time this bleaching takes place, a lexical item will maintain or experience an increase in its informational content. To capture theoretically this increase in information content, Traugott proposes a model consisting of three functional– semantic components: the propositional, the textual and the expressive. 3 Traugott notes that the paths indicated by the arrows in Figure 1 are optional, 3

As a point of departure for her own model, Traugott takes a model developed by Halliday and Hasan (1976), who distinguish the ideational, cohesive (cf. chs. 1 and 8), and interpersonal components, which, however, are not isomorphic with her propositional, textual, and expressive components. Moreover, in Traugott’s original flow chart, there is only one option of further development after an element has become a dummy marker. However, in her explanations she mentions that there are actually two options: Either a dummy marker may disappear or it may be reanalyzed into a meaningful marker.

81


representing tendencies and not a strictly unidirectional path of development. The broken lines pointing to the dummy marker indicate that there may be “several intervening stages of syntacticization and morphologization during which the grammaticalized [item loses] some or all of its meanings” (Traugott 1982:258).

Lexical items

Propositional Component ––––––––––––––––––––––– Grammatical Proposition Marker a) less personal

Textual Component ––––––––––––––––––––––––– Grammatical Text Marker a) less personal

b) more personal

b) more personal

Expressive Component ––––––––––––––––––––––– Grammatical Expression Marker a) less personal

b) more personal

Dummy marker

indicates pathways

Reanalyzed marker or Ø

indicates possible intermediary stages on the pathway

Figure 1: Pathways of grammaticalization, adapted from Traugott (1982:257) Thus, the propositional component involves the relation between a linguistic expression and the extra-linguistic reality it denotes, be it deictic-referential or solely referential. The textual component has to do with the resources available for creating a cohesive discourse. These include the various connectives, like but and therefore, (elements also subject to referential verification, but ultimately understandable only in terms of pragmatic discourse functions). They include anaphoric and cataphoric pronouns (e.g. anaphoric she in a woman . . . she; cataphoric this in I meant to say this: despite everything, you are still my friend), topicalizers, relativizers, complementizers, and so forth. These share the property of being directly linked to the unfolding of the speech event itself. (Traugott 1982:248)

In other words, the textual component involves closed-class elements that relate parts of text to other parts of the same text in a given speech situation.

82


The expressive component [...] bears on the resources a language has for expressing personal attitudes to what is being talked about, to the text itself, and to others in the speech situation. These include elements which show not only cohesion but also attitudes toward, even evaluation of, the propositions that cohere [...]. (Traugott 1982:248)

In comparing (7) with (8), for example, Traugott finds that the former, with the conjunction and, is evaluatively neutral, whereas the latter contains the adversative while, adding to the utterance an evaluative aspect. Also included in the expressive component are markers of turn taking and honorifics. (7) Fred likes peanut butter and Will hates it. (8) While Fred likes peanut butter, Will hates it. Explicit in Figure 1 is that a given lexical item need not pass through all stages (i.e., all components) of the grammaticalization process, to become grammaticalized. That is to say, a lexical item can acquire a propositional value and then take on an expressive value without having acquired a value as a textual marker. Moreover, the further along a given lexical item evolves along the path of grammaticalization, the more attitudinal or personal values the item reflects. While the propositional component deals with truth conditions, referentiality, and deixis, all of which are generally grounded in what we take to be objective reality, items with textual value abstract away from the language-real world relation and focus on cohesion of the text itself. Still more personal/attitudinal are the values in the expressive component. Thus, while a given lexical item may gradually lose its original semantic content in the process of grammaticalization, it increases in informational content from another source. However, an element can bypass any or all of the three components, as in instantaneous grammaticalization found in pidginization and creolization. In a case of abrupt creolization, for example, certain lexical items may become grammatical markers instantaneously, as in the case of Portuguese para “for,” which became an object pronoun marker in KCP (cf. [6b] above) over a relatively short span of time. A number of additional examples illustrate the pathways shown in Figure 1. For instance, Old English (OE) stead originally was a noun meaning “place, town,” but already at that time we find stead being used as a deictic in expressions such as on the stead “on the spot, there” (Oxford English Dictionary II, 2, d). Although in present day English we still find vestiges of nominal stead such as in the compound homestead, it is used either as an adverb (9) or in a compound preposition (10). Adverbial stead and the compound preposition instead of are elements of the textual component in that they relate one sentence to another in a given speech situation.


(9)

83

We were going to stay home but we decided to go to the movies instead.

(10) Instead of staying home, let’s go to the movies. A well-known example involves English will, which in OE was a verb of volition “want,” but had also developed an adverbial function, as in to go will “to go astray.” In OE, there were already instances of will functioning as a future marker. Today will marks the future almost exclusively and is realized as a free morpheme or a clitic. The development from the verb will “want” to a future marker most likely bypassed the textual and expressive components as it gradually became a reanalyzed marker. A third example, involving Spanish pues (from Páez Urdaneta 1982), represents a different type of development. In Latin, post(ea) “after” was a place/time adverbial, which was also used, albeit less often, in the sense “because, since.” In the twelfth century, pues is found as a temporal conjunction, as in (11), but much more frequently as a causal conjunction, as in (12). (11) Pues esto an fablado, pienssanse de adobar. after this have-3p said, start-3p to-prepare(-to-go) “After they have said this, they start getting ready to go.” (12) quito Castiella pues el rey he en yra. leave-1s Castille because the king have-1s in anger “I’m leaving Castille because I have angered the king.” Subsequently, pues developed an emphatic use, which in the words of Páez Urdaneta (1982:335) was “used to emphasize sentence constituents which are judged by the speaker to be emotionally or strategically important for satisfying his communicational intentions or his hearer's expectations. [The emphatic] pues is always found preceding adverbials, nominal phrases, embedded sentences, etc.” Examples of this emphatic pues are shown in (13). (13) a. Pues

en el barco que yo viajé, pues era grande. EMPH in the ship that I traveled EMPH was big “(Pues) in the ship I traveled on, (pues) it was big.”

b. La vida año que pasa, pues año que sube. the life year that pass-3s EMPH year that rise-3s “Not a year goes by (pues) without the prices going up.”

84


c. Bueno, pues sí, español. Well EMPH yes Spanish “Well, (pues) yes, Spanish.” The use of pues in (11)-(12) corresponds to the textual component, involving cohesion in discourse, while its use in (13) corresponds to the expressive component, dealing more with personal evaluation of the speech events. An example from Portuguese involves the adjective mesmo/-a “same,” belonging to the propositional component. Already in Middle Portuguese (at least by the fifteenth century), mesmo (invariable) had developed the additional use as an emphatic marker, a type of degree adverb (14a), or simply a marker of emphasis (14b). Both examples can be classified as instances of an expressive grammatical marker within the expressive component. (14) a. O meu amigo come muito mesmo. the my friend eats a lot EMPH “My friend really eats a lot.” b. Devemos chegar hoje mesmo. must-1pl arrive today EMPH “We have to arrive TODAY.” In KCP, the reflex of adjectival mesmo/-a is mez, while that of emphatic mesmo is mε(m). We see that their respective phonological shape has changed. In the former, the sibilant is maintained and the second nasal disappears, whereas in the latter the sibilant disappears entirely and the second nasal is variably realized. Thus, the fact that a Portuguese lexical item with a given phonological shape (i.e., mesmo) develops differentially in KCP corroborates that it had already developed two distinct meanings by the fifteenth century, one in the propositional component, the other in the expressive component. We take the more reduced form of emphatic mε in KCP to be a function of phonetic erosion due to higher frequency of use. Another example involves a use of non-anaphoric se found in Mexican and other varieties of Spanish. As early as the sixteenth-eighteenth centuries, the Spanish reflexive clitic se “himself/herself/oneself/themselves/yourself/yourselves” was used as an emphatic marker with motion verbs which did not, and still do not, require the presence of se. The examples we cite are from a 1578 legal document (García Carrillo 1988:66,119,128) in (15) and a diary from the early eighteenth century (Myers & Powell 1999) in (16).4 In these examples, se is used in utterances charged with emotion. In (15a, b), se appears in quotes 4

Thanks to Amanda Powell for pointing these examples out to me.


85

either reporting the escape of a prisoner (15a) or relating that a prisoner has escaped (15b). (15) a. oyó dar bozes hazia la Cárçel Real de corte a vn hombre, que dezían que era alcaide, diziendo: ‘que se huye el preso!’... (p. 128, lines 854-856) “he heard the yells directed to the Royal Jail ... by a man who they said was the mayor, and he was yelling, ‘the prisoner se is getting away!’”

b. e Martín Vélez, que allí estaua, le dixo: ‘Guillén es, que se a salido de la cárçel y se viene a la yglesia...’ (p. 119, ll. 330-331) “and Martín Vélez, who was there, said to him: ‘It’s Guillén who just se got out of jail and SE is coming to the church...’”

(16)

estando un dia en orasion vide a nr sr crusificado cerca de mi y vide como se desclabo bajandose de la [crus] y se vino para mi la crus se quedo parada y su mgd llego donde io estaba y se entro dentro de mi y abraso a mi [alma]... (vol. II, Folios 175 (verso) 176) “One day during prayer I saw our Lord, crucified, near me and I watched him as he took out the nails, lowered himself off the cross and SE came to me. The cross stayed upright and He approached where I was and SE entered within me and embraced my soul.”

Maldonado (1988) refers to a possibly related phenomenon in contemporary Mexican Spanish involving the use of se to mark unexpected changes in a situation. He cites the examples in (17): in (17a), the event is intended, whereas in (17b) it is accidental and unexpected. (17) a. Juan (*se) cayó al agua con toda elegancia. “Juan fell into the water with utter elegance.” b. Juan se (*Ø) cayó al agua vestido. “Juan fell into the water dressed.” In these examples, the 3sg reflexive (i.e., anaphoric) pronoun se, which belongs to the textual component, functions as an expressive grammatical marker or as a pragmatic marker, which would make it part of the expressive component. Another non-anaphoric function of se is as an aspectual marker with certain verbs, shown in (18). In (18a, b), it marks the completive aspect, and in (18c) the inchoative aspect.


86

(18) a. comer “eat” vs. b. beber “drink” vs. c. dormir “sleep” vs.

comerse “eat up” beberse “drink up” dormirse “fall asleep”

The respective developments of stead, will, pues, mesmo/-a, and se on the grammaticalization cline shown in Figure 1 are illustrated in (19). In (19), the single lines with the arrows indicate that the change has been completed; that is, that the former stage is attested vestigially or not at all. The double lines without the arrows indicate that both functions are still attested. For example, although Spanish se is used as an emphatic marker, it is still used as a reflexive pronoun as well. (19) ITEM Æ stead (noun)

Prop. C Æ

Text. C. Æ preposition/ adverb

Expr. C. Æ

tense marker (optionally a clitic)

will (verb) pues

Gram. item

temporal adverb/ conjunction

mesmo/-a (adjective) se

pronoun

se

pronoun

emphatic marker emphatic marker emphatic marker aspectual marker

We have seen various examples of how the flow chart in Figure 1 applies to the grammaticalization of very different classes of items, including a noun, a verb, an adjective, an adverb, and a pronoun. We now turn the discussion to counterexamples to the flow chart. 3. Counterexamples No matter which component a given item evolves from or which component it develops into, the general evolution of the item is said to move from Propositional to Textual to Expressive with the end being that point at which an item is fully grammaticalized, a functional element. But, is there no further evolution for elements once they have become grammatical markers? If functional elements can undergo further evolution, in which direction(s) or along which path(s) can/does this evolution take place? 5 In what follows, I discuss 5

Asking somewhat different questions, Newmeyer (1998) comes up with a sizable list of counterexamples to the notion of unidirectionality. In one part of his analysis, Newmeyer concludes that grammaticalization does not constitute a distinct process but is rather the


87

evidence that suggests that a fully grammatical (i.e., functional) element can undergo a number of further developments, both with regard to its form as well as its function/meaning. The first three cases involve a grammatical marker that has become an emphatic marker, part of the Expressive Component. In one of these cases, however—that of KCP—the emphatic marker subsequently disappears. During the creolization process that led to the formation of KCP, evidence indicates that the Portuguese temporal adverb ja “already” was reanalyzed as a past marker (Clements 1996). Recently collected data spanning four generations of KCP speakers suggest that, until relatively recently, KCP marked the past redundantly, with the preposed marker ja, as well as suffixally with –o or –w, as shown in (20) (cf. Clements 1990). (20) a. Lorenz ja Lorenz PAST “Lorenz sang.”

kato. (kata “sing”) sang

bebew. (bebe “drink”) b. Lorenz ja Lorenz PAST drank “Lorenz drank.”

irgiw. (irgi “get up”) c. Lorenz ja Lorenz PAST got-up “Lorenz got up.”

Moreover, we know from texts collected by Dalgado (1906) that at least two sister creoles of KCP, Mahim Creole Portuguese and Talasri Creole Portuguese, also marked the past redundantly exactly as KCP apparently did. Due to the increasing pressure exerted on KCP by Marathi, the SOV adstrate language in the Korlai area, KCP has been taking on—more recently at an accelerated epiphenomenal overlapping of the independently occurring processes of semantic change, downgrading (e.g., changing of an element from a verb to an adposition to an affix), and phonetic reduction (1998:252-60). Further on (1998:275-78), however, he discusses the reasons why unidirectionality is ‘almost’ true. He argues that downgrading and phonetic erosion are the effects of least-effort in language production. Functional elements, he suggests, require less coding material and thus less production effort than lexical elements. While this is a promising line of research to pursue, one would also need to take language processing into account. For example, it is known that in first language (cf. Bates & Goodman 1999) and second language (cf. Zobl 1982) acquisition, lexical elements are acquired before functional elements. That is, from a language processing perspective, functional elements are harder to recognize, parse, and understand than lexical elements precisely because they are given ‘short shrift’. This tendency in language processing would complement the least-effort effects apparently present in language production and alluded to by Newmeyer.


88

pace—SOV-type traits such as postpositions, clause-final complementizers and predominantly SOV word order (cf. Clements 1990, 1991, 1992, 1996, 2001). Part of the same development has involved the loss of ja as the past marker, leaving the suffixal past marking which KCP shares in this case with Marathi. Currently, no inhabitant of Korlai marks the past with ja; rather, all speakers use ja as an affirmative adverb “yes,” but only with sentences with dynamic verbs, as the example in (21). 6 (21)

Question:

Answer:

Jato use? had-lunch you-SING-FORMAL “Have you had lunch?” Ja. yes “Yes.”

However, the speech of a Korlai-born man who resided all but his first 10 years outside of Korlai reveals an earlier stage of KCP in which ja, though it had ceased to be a past marker, functioned as an emphatic marker. After the informant in question had left Korlai, he lived with some relatives in Mumbai 6 In earlier writings, I have identified ja as a proverb. However, it seems more accurately characterized as an affirmative adverb “yes.” In KCP, ja can be used interchangeably with hi~ (< Ptg. sim “yes”) as an affirmative response to questions that contain dynamic situations (in Comrie’s [1976] sense), such as those in (i). (i) a. Q: Use jave Š~nt? A: ja/hi~ came yesterday you-FORMAL “Did you come yesterday?” b. Q: Gaḍi jaho? A: ja/hi~ bus went “Did the bus leave?” c. Q: Chu kaiw? A: ja/hi~ rain fell “Did it rain?” d. Q: Use \k\ sirwis hedzew? A: ja/hi~ that work did you-FORMAL “Did you do the work?” However, when affirming stative or imperfective dynamic situations, only hi~ is possible, as illustrated by the examples in (ii) and (iii). (ii) a. Q: Use ti sab \k\ mahiti? A: *ja/hi~ PAST know that news you-FORMAL “Did you know that news?” b. Q: Puse mang kere? A: *ja/hi~ OBJ-you-FORMAL mango want “Do you want mangos?” (iii) Q: Use \k\ sirwis ti hedzen? A: *ja/hi~ that work PAST doing you-FORMAL “Did you do the work?”

89


(formerly Bombay) and subsequently settled there. While living in Mumbai, the informant had contact to KCP through his KCP-speaking wife, but had no sustained contact with the villagers in Korlai. An examination of a long monologue of his about his life (approximately 3700 words) reveals the use of ja to mark emphasis. By dividing up his narrative in three parts (the description of his life up until a violent confrontation at his workplace between union members and the police, the description of the violent confrontation, and the remaining part of the narrative), we see that the use of ja fluctuates substantially, depending on what was being described. During the first portion of his narrative, in which he describes leaving Korlai, living with relatives in Mumbai, finding a job, starting a family, and moving up in the company where he worked, ja was found in 24% of sentences (21/87). By contrast, in the description of the violent scene, the presence of ja increases to 72% (13/18), and remains relatively high during the rest of the narrative (63% [49/77]). 7 The data suggest that ja is not used to mark the past, but rather is used to mark emphasis and that its use is variable. I interpret this case as one in which a functional element (a past marker) becomes a marker of emphasis (part of the expressive marker) and is contrary to Traugott’s flow chart. Interestingly, the expressive stage of ja was a precursor to its total loss from KCP as an emphatic marker. It seems that once ja was disconnected from its function as a past marker, it served to mark emphasis for only a short while before that function disappeared. In G°~, a Kwa language spoken in Togo and southern Bénin, the particle ná obligatorily marks habituality on verbs, but only optionally on prepositions, shown schematically in (22a), and with an illustrative example in (22b). (22) a. [VP V-ná [PP P(-ná) NP]] nya sugbŠÑ [PP ná(-ná) b. [VP tó-ná emit-HAB word many to-HAB “... usually talks to Ayï.” (Lewis 1990:2)

Ayï]] Ayï

In many cases in which ná is optional, G°~ speakers use the particle to mark emphasis, in which case the presence of ná triggers focus movement. This is yet another case of a grammatical marker, ná, acquiring the function as an expressive marker and encoding the speaker’s view of a given situation. The third case involves the pragmatic use of ne in Swiss French. Historically, all varieties of French share discontinuous negation: ne … pas. FonsecaGreber (2003) reports that the negation particle ne in her 117,000-word Swiss 7 The raw numbers given here correct a typo in Clements (1990:116), where the raw numbers for absence of ja in the third part are given incorrectly as 32/18 (37%). The corrected version is 28/77 (37%). In the present discussion, we are focusing on the presence of ja, which, as just noted for the third part, is 49/77 (63%).


90

French corpus is only found in 2.5% (43/1982) of the negations. Where it appears, it seems to mark either formality or emphasis. We cite two examples of the latter use. In the first, a speaker (Speaker 12 in Fonseca-Greber’s corpus) makes sure that her interlocutors understand that it was her father’s wish, and not hers, not to have an answering machine. To stress this, Speaker 12 repeats her utterance, inserting ne each time when focusing on the father’s resistance to answering machines. This is shown in (23). (23) S1: et ben les répondeurs ça sert à quelque chose..non.. (…) “well answering machines are there for a reason aren’t they..(…)” S12: mais nous on a même pas de répondeur..mais papa il n'en veut pas.. “but we don’t even have an answering machine..and dad does not want one..” S13: hein non “yeah” S12: nous on a même pas de répondeur..et papa il n'en veut pas.. “we don’t even have an answering machine..and dad does not want one..” S13: non (S1, S12, S13, V-B) “yeah”

In another interaction, shown in (24), two speakers (the married couple Speaker 7 and Speaker 8) co-construct an emphatic repetition, as S7 reaffirms S8’s line of argumentation. Both insert ne for emphasis, which is further emphasized by the use of the relatively rare negator point. (24)

S8: n’oubliez pas que je ne suis point riche.. “don’t forget that I’m not at all rich..” S7: c’est vrai que je ne suis point riche..voilà exactement.. (IV-A) “that’s right that I’m not at all rich… that’s it exactly…”

In these three counterexamples, we see that a grammatical marker acquires the value of encoding the speaker’s view of a given situation, and all three cases involve a redundant marker (tense, case, or negation) taking on an expressive function. In another type of counterexample, we find a grammatical marker acquiring full-fledged verb status, while maintaining its affixal nature. Langacker (1977:90) states that: [m]any languages, including various Uto-Aztecan languages, use existential locutions such as ‘His N[oun] is’ to express the meaning ‘He has a N[oun]’. In this exis-

91


tential locution, the noun N would naturally take the possessed suffix *-wa, and the verb might be expected to be zero, which can in fact be reconstructed for ‘be’ in Proto-Uto-Aztecan and was clearly used in existential sentences. Another, less common possessive locution found in various Uto-Aztecan languages [is] ‘He has his N’. [It seems] ... that the ‘have’ sense of *wa came about through a construction reformulation in which expressions of the type ‘His N is’ were reanalyzed as representing locutions of the type ‘He has his N’ (which later simplified to ‘He has N’).

Thus, in this case a grammatical element marking the possessed nominal was reanalyzed as a full-fledged verb of possession, although it remained an affix. This reanalysis is shown in (25). (25) Ni ‫׀‬ X

POSSRi ‫׀‬ Y–

N ‫׀‬ Z

POSSD ‫׀‬ -wa

BE Æ ‫׀‬ Ø

Ni ‫׀‬ X

POSSRi ‫׀‬ Y–

N ‫׀‬ Z

HAVE ‫׀‬ -wa

Important here is that a crucial semantic change took place without any corresponding formal change, suggesting that the semantic evolution of a grammatical item can occur independently of its evolution as free form, clitic, or affix. This relatedness between form and function clines, as well as their simultaneous independence from one another along the cline of lexicalizationgrammaticalization, is supported by other data which will now be discussed. 4. Separateness of form and function on the lexicalizationgrammaticalization cline To support the argument that form and function on the lexicalizationgrammaticalization cline are separate though interdependent, let us consider the following two examples. The first one involves the Northern Saame Enontekiö (NSE) dialect from the Finno-Permic family, as discussed by Nevis (1986a). In this dialect, the abessive morpheme taga developed from an affix to a clitic to a free word without changing its status as a functional element. The form taga comes from an affixal sequence *pta-k-ek/n consisting of caritive *pta + lative *-k + an extra lative suffix *-k or n (with epenthetic e). This is shown in (26). (26) *pta -k caritive suffix lative suffix

-ek/n lative suffix

Æ taga abessive postposition 8

In a discussion of the development of taga, Joseph and Janda (1988:200) argue that: 8

The terms caritive and abessive are synonyms, both referring to a case marker that expresses the absence of the referent of the marked noun (= English “without”). The term lative is used to refer to a case that expresses motion up to the location of, or as far as the referent of the noun it marks.

92


evidence that taga developed from an affix is the fact that in the vast majority of Saame dialects and in fact throughout Finno-Permic in general, the reflexes of *pta-k are suffixal. Moreover, the sequence is clearly composed of two (or three in the case of Northern Saame) morphemes, each reconstructible on its own, and the singlesegment elements *-k and *-n simply could not have been independent words in the protolanguage. Finally, in most of the Finno-Permic languages, *pta shows idiosyncrasies of combination—in particular, attachment as a case suffix to verbal stems— which are unpredictable and unexpected and thus unlikely to have occurred independently in each language.

Contrary to the Uto-Aztecan example above which undergoes semantic change but remains an affix, here an affixal grammatical relation marker becomes a postposition without any change in function. That is, the change affix Æ free morpheme occurred independently of the semantics of the marker. The second case reported in Nevis (1986b) is similar: Although Old Estonian interrogative es and emphatic ep (both second-position enclitics) have become free words in Modern Estonian, there has been no change in their semantic interpretation. Yet another case involves a ‘loosening’ process of a bound morpheme, deaffixation, present in the derivational suffix –ism (also mentioned in Newmeyer 1998). In various European languages one can speak of many kinds of isms, referring to a doctrine or theory. Note that even though there is a change in the formal status of the affix—it becomes a free morpheme—its semantic content remains virtually unchanged. That is, the semantics of the affix –ism would be roughly “theory, doctrine of X,” where X represents the semantic content of the noun or adjective it attaches to. As a noun, ism is defined in the Oxford English Dictionary as “a form of doctrine, theory, or practice having, or claiming to have, a distinctive character or relation: chiefly used disparagingly and sometimes with implied reference to schism.” In Traugott’s model, ism > ism does not evolve beyond the propositional component, although the semantics of the free morpheme take on a pejorative connotation. Apart from the examples of grammatical markers and derivational morphemes, one also finds examples of items belonging to closed lexical categories becoming members of the major lexical categories of nouns and verbs. Hopper and Traugott (1993:127) mention the examples of uppers (N) < upper (Adj) and downer (N) < down (Adv). To these, various others can be added, such as ins and outs and ups and downs from the adverbs in, out, up, and down respectively. The adverb out is also attested as a verb, as in to out someone “to expose someone.” 9 9

As an example, the sentence “I’m sorry I outed you” appears in the film “Sweet Home Alabama” (2002, Touchstone Pictures) uttered by the character Melanie to childhood-friend character Bobby Ray in apology for revealing in a group of his friends that he was gay.

93


(27) GRAM. ITEM

KCP past marker ja G°~ habitual marker –ná Swiss French negative markers ne Uto-Aztecan possessed marker – wa

Expressive C.

>

Textual. C. > in, out (Adv) up, down upper (Adj) down (Adj ) out (Adv)

Propositional C. ins and outs (N) ups and downs (N) uppers (N) downers (N) to out (V)

emphatic marker emphatic marker emphatic marker verb “have”

The chart showing the development of these and other items discussed in this section is given in (27). The other examples, those in (28), involve a change in the form. However, in each case the semantic changes are slight and do not involve a change in component. (28) GRAM. ITEM

Expressive C. >

Textual. C.

>

Propositional C. English -ism “doctrine, theory, or practice of X” > ism “doctrine, theory, or practice”

NSE postposition taga “without” (< affix *pta-k-ek/n) Estonian free morpheme emphatic ep (< clitic pa) Estonian interrogative es (< clitics -ko and -s)

5. Discussion and concluding remarks According to Hopper and Traugott (1993:127), the prototypical direction of grammaticalization goes from lexical item to functional element. Traugott

94


(1982) proposes a more detailed model of grammaticalization which involves an increase in informational content of a grammaticalizing element before it fully grammaticalizes. An element travels from the Propositional Component optionally through the Textual and/or Expressive Component(s) toward grammatical-marker status. Prototypically, then, the grammaticalization process is unidirectional. Based on the counterexamples discussed in section 3, we propose viewing the grammaticalization process as part of a lexicalizationgrammaticalization continuum. The very notion of a prototypical process implies another process that is not prototypical. By positing a lexicalizationgrammaticalization continuum and appealing to the notions of regularity, iteration, and unpredictability as used in Chaos Theory (Chalice 1993), we can capture the facts that, overall, the emergence of functional elements from lexical items is more regular under iteration than the emergence of lexical items from functional elements. We can also capture the fact that any one development of a lexical item to a grammatical element or vice-versa is ultimately unpredictable. Thus, the approach we propose can handle general developmental tendencies, but also the development of individual items in language. The chaotic approach to grammaticalization, on the lexicalizationgrammaticalization continuum as outlined here, makes predictions regarding the form of an element’s development on the continuum, as well as the development of its meaning/function. Taking Traugott’s model in Figure 1 as a starting point, the prediction would be as follows: given sufficient iteration— that is, given a data base containing a large number of all types of change along the lexicalization-grammaticalization cline—a development from any stage to any other stage in the model would be theoretically possible. This prediction results from the notion of unpredictability of the semantic processes involved in grammaticalization and could be represented as in Figure 2. 10 In (29) and (30), we summarize the analysis of the data discussed so far based on the flow chart in Figure 2. (29) contains the lexical-to-grammatical examples, and (30) the grammatical-to-lexical examples. (29) Lexical Æ Grammatical Prop. C. Æ Text. C. (Engl.: stead “place” Æ (in)stead [adverb]) Prop. C. Æ Exp. C. (Ptg.: mesmo “same” Æ [emphatic marker]) Prop. C. Æ Gram. item (Engl.: will “want” Æ [future marker]) Text. C. Æ Exp. C. (Sp.: pues “after” Æ [emphatic marker], se “3sg.REFL” Æ [emphatic marker]) Text. C. Æ Gram. item (KCP: ja “already” Æ [past marker]) Exp. C. Æ Gram. item (no example found) 10 Some of the detail included in Figure 1, such as the development from more to less personal in each component, has been omitted here for purposes of exposition.


95

(30) Grammatical Æ Lexical Gram. item Æ Exp. C. (KCP: ja “already” Æ [emphatic marker]; G°~: -ná [habitual marker] Æ [emphatic marker]; Swiss French: ne [negator] Æ [emphatic marker]) Gram. item Æ Text. C. (no example found) Gram. item Æ Prop. C. (Uto-Aztecan: –wa [possessed marker] Æ “have”) Exp. C. Æ Text. C. (no example found) Exp. C. Æ Prop. C. (no example found) Text C. Æ Prop. C. (Engl.: up, down [adverbs] Æ ups, downs [nouns]; out [adverb] Æ [verb])

Lexical items

Propositional Component ––––––––––––––––––––––––– Grammatical Proposition Marker

Textual Component ––––––––––––––––––––––– Grammatical Text Marker

Æ

Æ

Expressive Component ––––––––––––––––––––––– Grammatical Expression Marker

Dummy marker

Reanalyzed marker or Ø

Figure 2: Possible pathways along the lexicalization-grammaticalization cline In the lexical-to-grammatical direction, we have been able to attest all but one possible development, and we have attested three of six possibilities in the grammatical-to-lexical direction. We believe, however, that with further research the missing possibilities will be attested. Regarding the data in (27) and (28), the question arises as to why certain developments occur rarely if at all. One possibility is that processes in general are unidirectional (Newmeyer 1998:262). Another possibility is that there are least-effort effects that influence the production of linguistic material, effects such as phonetic erosion. Moreover, frequency may play a central role in the downgrading of a lexical item (cf. Newmeyer 1998:252-260). However, it is


96 Lexical items

Propositional Component ––––––––––––––––––––––– Grammatical Proposition Marker

Textual Component –––––––––––––––––––––––– Grammatical Text Marker

Expressive Component ––––––––––––––––––––––– Grammatical Expression Marker

Dummy marker

indicates major pathway indicates minor pathway Reanalyzed marker or Ø

Figure 3: More commonly vs. less commonly attested pathways in the lexicalization-grammaticalization cline

clitic

free form

affix Figure 4: The possibilities for formal change between free forms, clitics, and affixes clear that the convergence of semantic change, downgrading, and phonetic erosion combined are constantly at work in tandem in the reshaping of language form and content.


97

After grammaticalization, however, this confluence of processes among the different components seems to weaken or does not occur at all. The development of a grammatical(ized) element or any element that is upgraded obeys, if not a different set of principles, then definitely a different set of processes. For example, we found three instances of the reanalysis of a redundant functional element into an emphatic marker. The process there seems apparent: If a grammatical function is marked redundantly and there are other factors at work (e.g., phonetic erosion, influence through language contact, typological universals, etc.), one of the markers is theoretically available to take on another function. Thus, although there is a certain developmental path followed by linguistic elements with high regularity (lexical Æ grammatical), as this development happens over and over again (i.e., it iterates), instances occur that counter the regularity. With sufficient iteration, there form less regular patterns along with those that are highly regular. Taken individually, the grammaticalization or lexicalization path that any given item follows in its development to some other stage is considered ultimately unpredictable in this system, although the range of developmental possibilities of a given item is finite. We find that this chaotic view of development within the lexicalizationgrammaticalization continuum allows us to define prototypicality in this case as the most regular developments within the flow chart, given sufficient iteration. This is shown by the lines in Figure 3, in which the more regular pathways are marked with thick lines, the less regular pathways with thin lines. In our remarks so far in this section, we have discussed only the development of the meaning or function of a given item. However, the same predictions apply regarding the development of the form of an item on a different continuum: the free form-clitic-affix cline. That is, assuming the possible forms of an item just mentioned, and shown in Figure 4, the prediction is that a development from any stage to any other stage in the model would be theoretically possible, given sufficient iteration. Again, this prediction follows from the notion of unpredictability of the formal processes involved in grammaticalization and lexicalization. The arrows in Figure 4 show all the possibilities. Indeed, all these possibilities are attested, as shown by the examples in (31). (31) a. affix ↔ clitic (examples: NSE clitic postposition taga < affix *pta-k-ek/n; KCP inflectional case-marking prefix p- < clitic p\r) b. clitic ↔ word (examples: Estonian clitic pa > free morpheme ep; Engl.’ll < will)c. affix ↔ word (NSE free form adverb taga < affix c. *-pta-k-ek/n; Spanish –á “3SG-FUTURE” < free form auxiliary ha “3SG-FUTURE”) 11 11

Given that ha (< Latin habet “have-3SG-PRESENT INDICATIVE”) has always carried stress on the penult syllable, one cannot speak in this case of a clitic (i.e., unstressed) phase in its evolution from a free form to an affix as a Spanish future marker.


98

To indicate the high and low regularity in the pathways, the most commonly attested ones are marked with thick arrows and the more infrequently attested ones are marked with thin arrows, shown in Figure 5. The regularity is impressionistic and is based on Newmeyer’s assessment that downgrading (e.g., lexical item Æ grammatical element, free morpheme Æ affix) is 10 times as common as upgrading (e.g., grammatical element Æ lexical item, affix Æ free morpheme).

clitic

free form

affix = major pathway

= minor pathway Figure 5: Major and minor pathways of change between free forms, clitics, and affixes Approaching grammaticalization with the use of a lexicalizationgrammaticalization continuum allows us to see the different processes at work in the two different directions. In the lexical-to-grammatical direction, there is, as noted by Newmeyer (1998:252-260), a confluence of processes that work together, such as production-related process, most notably least-effort effects, frequency of occurrence of an item, and phonetic erosion. In the grammaticalto-lexical direction, redundancy of markers has been shown to be factor in this study. It can combine with other processes such as phonetic erosion (in the case of Swiss French ne) and typological considerations, as in the case of KCP ja (cf. Clements 1990). What this suggests is that in one direction (lexical-togrammatical), one set of processes seem to be at work, whereas in the other direction this is not necessarily the case.


99

The data from this study also suggest that it is necessary to view the semantic-discourse-functional development of a form as ultimately separate from, though related to, its formal development. In future research, we hope to document all the logical possibilities predicted by the lexicalizationgrammaticalization cline. The hypothesis at this point is that many more different processes will be involved in the grammatical-to-lexical direction than vice-versa.

References Bates, Elizabeth & Judith C. Goodman. 1999. “On the emergence of grammar from the lexicon”. The emergence of language, MacWhinney (ed) 1999 (Mahwah, NJ: Lawrence Erlbaum). 29-79. Bruyn, Adrienne. 1996. “On identifying instances of grammaticalization in creole languages”. Changing meaning, changing functions. Papers relating to grammaticalization in contact languages, Baker & Syea (eds) 1996 (London: University of Westminster Press). 29-46. Chalice, Donald. 1993. “Chaos, models and templates”. Chaos and chaos theory—nonexistent objects, Milicic (ed) 1993 (Bellingham:Western Washington University). 11-30. Clements, J. Clancy. 1990. “Deletion as an indicator of SVO Æ SOV shift”. Language Variation and Change 2. 103-133. ----------. 1991. “The Indo-Portuguese Creoles: Languages in transition”. Hispania 74. 327-46. ----------. 1992. “Elements of resistance in contact-induced language change”. Explanation in historical linguistics, Davis & Iverson (eds) 1992 (Amsterdam: John Benjamins). 41-58. ----------. 1995. “O processo de gramaticalização e a teoria do caos”. Papia: Revista de Crioulos de Base Ibêrica 4. 32-45. ----------. 1996. The genesis of a language: The formation and development of Korlai Portuguese. Amsterdam: John Benjamins. ----------. 2001. “Word order shift and natural L2 acquisition in a Portuguese Creole”. Romance syntax, semantics and L2 acquisition, Wiltshire & Camps (eds) 2001 (Amsterdam: John Benjamins). 73-87. ----------. 2003. “The tense-aspect system in pidgins and naturalistically learned L2”. Studies in Second Language Acquisition 25. 245-281. ----------. In press. “The Portuguese-based creoles in Africa and Asia”. Berkeley Linguistic Society 29. ----------. Forthcoming. “Korlai Creole Portuguese (NŠ Ling)”. Comparative creole syntax, Holm & Patrick (eds) (London: Battlebridge Press). Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press.

100


Dalgado, Sebastião Rodolfo. 1906. “Dialecto Indo-Português do Norte”. Revista Lusitana 9. 142-66, 193-228. Fonseca-Greber, Bonnie. 2003. “The emergence of emphatic “ne” in conversational Swiss French”. Paper presented at Linguistic Symposium for the Romance Languages 33. Bloomington, IN, April 24-27. García Carrillo, Antonio. 1988. El español en México en el siglo XVI. Sevilla: Ediciones Alfar. Halliday, M.A.K. & Ruqaiya Hasan. 1976. Cohesion in English. London: Longman. Haspelmath, Martin. 1998. [Review of Historical syntax in cross-linguistic perspective, by Alice Harris & Lyle Campbell]. Linguistic Typology 2. 131-139. Heine, Bernd, Ulrike Claudi, & Friederike Hünnemeyer. 1991. Grammaticalization: A conceptual framework. Chicago: University of Chicago Press. Holm, John. 2000. An introduction to Pidgins and Creoles. Cambridge: Cambridge University Press. Hopper, Paul J. & Elizabeth C. Traugott. 1993. Grammaticalization. Cambridge: Cambridge University Press. Joseph, Brian & Richard Janda. 1988. “The how and why of diachronic morphologization and demorphologization”. Theoretical morphology, Noonan & Hammond (eds) 1988 (San Diego: Academic Press). 193-210. Labov, William. 1982. “Building on empirical foundations”. Perspectives on historical linguistics, Lehmann & Malkiel (eds) 1982 (Amsterdam: John Benjamins). 17-92. Langacker, Ronald. 1977. “Syntactic reanalysis”. Mechanisms of syntactic change, Charles Li (ed) 1977 (Austin: University of Austin Press). 57-139. Lewis, Marshall. 1990. “Aspect and figure/ground texture: Habitual marking on G°~ prepositions”. Paper presented at the International Pragmatics Conference. Barcelona, Spain, July 12. Mühlhäusler, Peter. 1986. Pidgin and creole linguistics. Oxford: Basil Blackwell. Muysken, Pieter & Tonjes Veenstra. 1995. “Haitian”. Pidgins and creoles: An introduction, Arends, Muysken, & Smith (eds) 1995 (Amsterdam: John Benjamins). 153-163. Myers, Kathleen & Amanda Powell. 1999. Study of the writings of Sor Maria de San José. Bloomington: Indiana University Press. Naro, Anthony. 1978. “A study on the origins of pidginization”. Language 53. 314-347. Nevis, Joel. 1986a. “Decliticization and deaffixization in Saame: Abessive taga”. Ohio State University Working Papers in Linguistics 34. 1-9. ----------. 1986b. “Declitization in Old Estonian”. Ohio State University Working Papers in Linguistics 34. 10-28.


101

Newmeyer, Frederick J. 1998. Language form and language function. Cambridge MA: MIT Press. Oxford English Dictionary. London: Clarendon Press. Páez Urdaneta, Iraset. 1982. “Conversational pues in Spanish: A process of degrammaticalization?” Papers from the 5th International Conference on Historical Linguistics, Ahlqvist (ed) 1982 (Amsterdam: John Benjamins). 332-340. Traugott, Elizabeth C. 1982. “From propositional to textual and expressive meanings: Some semantic-pragmatic aspects of grammaticalization”. Perspectives on historical linguistics, Lehmann & Malkiel (eds) 1982 (Amsterdam: John Benjamins). 245-71. ---------- & Bernd Heine (eds). 1991. Approaches to grammaticalization. (2 vol.). Amsterdam: John Benjamins. Ziegler, Debra. 1997. “Retention in ontogenetic and diachronic grammaticalization”. Cognitive Linguistics 8. 207-241. Zobl, H. 1982. “A direction for contrastive analysis: The comparative study of developmental sequences”. TESOL Quarterly 16. 169-183.

CREOLE TRANSPLANTATION A SOURCE OF SOLUTIONS TO RESISTANT ANOMALIES

John McWhorter Manhattan Institute

ABSTRACT: Creole genesis theory tends to focus on polygenetic scenarios, wherein a creole developed in the place it is spoken today as the result of sociohistorical factors specific to that location. Yet a great deal of linguistic and historical evidence shows that most creoles were brought to their current locations from elsewhere, or that a creole’s emergence was profoundly affected by the previous existence of another contact language already spoken on site. This paper demonstrates that this phenomenon allows principled explanations of various long-standing anomalies to creole genesis theory, such as the mysterious absence of creole continua in French plantation colonies and the paucity of Spanish-based creoles, as well as newer questions such as why a creole emerged in Hawaii among children who were being taught in English in school. 1. Introduction One of the most important aspects of creole studies is reconstructing how and why these languages developed. The dominant reconstruction has become what I will call the limited access conception. According to this model, creole languages emerged when learners of a language, most often plantation slaves, received only fragmentary input from that language as the result of social or psychological distance. In need of a full language nevertheless, they created one from this material via transfer from substrate languages and the operations of universals of various kinds, be this through an innate bioprogram or more cognitively general universals of second language acquisition. To be sure, there is a wide range of creole genesis theories, and limited access to a target is by no means the sum total of any of them. However, whether concentrating on the expression of a bioprogram in response to dilution of target input (Bickerton 1981, 1984), relexification of substrate languages (Lefebvre 1986 or, less schematically, Alleyne 1980; Boretzky 1983), ‘creative’ forging of a new interethnic communication vehicle (Baker 1995), or the development of ‘approximations of approximations’ of the lexifier as slave populations gradually increased (Chaudenson 1992), all creole genesis theories incorporate the basic limited access conception as a key component. Very few


104

working creolists could claim to consider the disproportion of black to white unimportant to the emergence of plantation creoles. Examples: Bickerton (1988:271-2): In the plantation situation, the preparatory phase of sugar colonization gave way...to the exploitative phase...requiring a rapid increase in the numbers of unskilled manual laborers;..dilution of the original model must have resulted...In fact, what took place...was second language learning with inadequate input.

Alleyne (1971:180): In the New World, on sugar plantations, production was organized on the basis of a kind of occupational stratification according to which field slaves were most numerous and were furthest removed from contact with Europeans. Social intercourse of the field slaves was almost exclusively confined within the group; and so, among them, linguistic forms showing a high degree of divergence...were able to crystallize and achieve the appearance of stability.

Singler (1995:219-20): The greater the proportion of people of color in a Caribbean colony during the period of genesis, the more ‘radical’ the creole that emerged, with ‘radicalness’ being measured as distance from the lexifier language...[this assumption], linking the proportion of people of color in the population to degree of radicalness, is apparently uncontroversial.

Corne (1995:12): If it is accepted that Mauritian Creole emerged and jelled in that formative period (roughly fifty years or three generations long) of the eighteenth century and in demographic conditions which by definition rendered access to French next to impossible for most people...

While lack of ready access to a dominant language, or indeed lack of social compunction to acquire it fully, certainly has been a factor in the birth of creoles, the concentration on this aspect has entailed the neglect of an equally important factor in explaining why we find creoles where we do today. Specifically, quite often, it can be shown that a creole is spoken in a given place not because it emerged there, but because a pidgin or creole had already existed in that location and served either as a template or as a direct ancestor for the new creole that developed. (Chaudenson 1981, 2000 has long noted this fact, but his work, written in French, has had only minor impact on Anglophone creolist work.). 1 1

Chaudenson first applied the concept to a hypothesis that Mauritian Creole is a direct continuation of Réunionnais French, a scenario that Baker (1996 and others) has rather decisively


105

In comparison to limited access-based scenarios, the development of a creole as a mere continuation of one imported from elsewhere seems at first rather mundane, which partly explains the general marginalization of this phenomenon even in studies of cases where its importance is clear. For example, no one comparing Haitian and Antillean French Creoles could deny that Haitian has obvious direct historical links to Antillean, to the extent that Haitians and Antilleans can manage conversation. Yet while this is occasionally acknowledged briefly (e.g., Baker 1987:73; Singler 1995:203-4), accounts of the genesis of Haitian largely imply that the language emerged in situ (e.g., Lefebvre 1998), with the implication that the influence of pre-existing creoles was at best adstratal. It is not difficult to see the reasons for this: The Language Bioprogram Hypothesis suggests that creoles will lend especial insight into Universal Grammar; relexificationist/substratist conceptions depict creoles as fascinating strategies of cultural hybridization and survival; the ‘creativist’ analysis uses science to help rid creoles of the Sisyphean taint; as so forth. There is more apparent fascination in processes like these than in simply tracing the birth of a creole to the fact that seasoned immigrants were already speaking one. In short, the limited access model simply seems more interesting. Yet the fact is that the emergence of certain creoles as continuations of preexisting ones, while seemingly a rather inert affair, in fact sheds vital insights upon a number of problems long unsolved in creole studies. This paper will show the benefits that more explicit acknowledgment and study of these historical links can yield to the discipline. 2. Hawaiian Creole English: Sine Pidgin Non? Bickerton (1981) argued that Hawaiian Creole English (HCE) developed when children of parents speaking various languages were exposed to a mixture of structurally minimal and irregular English interlanguages spoken by the adults around them, and were thereby forced to transform this farrago into a full language via an innate language bioprogram. The validity of Bickerton’s extension of this scenario to other plantation colonies has, of course, been hotly disputed. For the purposes of this paper, however, what is of interest to us is that Bickerton’s application of this limited access-based scenario to HCE has been refuted by recent work (Roberts 2000). Roberts’ work by no means invalidates Bickerton’s conception completely. For example, Goodman (1985), Holm (1986) and even I (McWhorter 1994) have argued that it was adults from the mid-1800s on, rather than children at the turn of the century, who created HCE. This conclusion was warranted by argued against; however, Chaudenson makes less controversial reference to the concept in Chaudenson (2000).

106


the evidence Bickerton brought to bear viewed against the other evidence available at the time. However, through assiduous examination of a lode of historical documents, Roberts has conclusively shown that while many adult immigrants in Hawaii did speak a range of pidgin Englishes, children were indeed central in creating the stabilized language we today recognize as HCE, including the development and/or conventionalization of the tense, mood, and aspect markers. However, it is equally important to note that these children created this creole in a context in which they were nothing less than bathed daily in rich input from full varieties of English, which would hardly have necessitated resort to a bioprogram. For example, a newspaper editorial of 1887 includes the following description of a fuller English spoken by adults (Reinecke et al. 1975:595, 609, cited in Goodman 1985:111-2): The colloquial English of Hawaii nei is even now sufficiently sui generis to be noticeable to strangers. It is not a dialect, but a new language with English as its basic element, wrought upon by the subtle forces of other languages, not so much in the matter of a changed vocabulary as a changed diction.

Goodman, Holm, and myself have argued that this suggests that HCE already existed by the 1880s, long before Bickerton reconstructed the creole as having been born. However, Roberts (2000) has convincingly shown that this author was in fact referring to a full indigenized Hawaiian English. Crucially, however, earlier in the editorial the author makes specific reference to a ‘pigeon’ English spoken by children. This is important in showing that HCE, or its precursor, was indeed already spoken in the 1880s, even if the excerpted passage above itself is not referring to it. In other words, children created HCE in the very context in which this ‘colloquial English’ was coin of the realm. With input of this richness, children would not have needed to create a language from the ground up. Furthermore—a fact markedly de-emphasized in Bickerton’s writings—the children who created HCE attended schools where they were taught in English. Alone, this fact eliminates any reason for supposing that their English input was limited to any jargon or pidgin. Thus, we see that second-generation immigrant schoolchildren created HCE not as a solution to inadequate English input, but as the linguistic expression of a new young Hawaiian identity. Many have been uncomfortable with Bickerton’s attempts to parse creole genesis as a mere matter of demographic ratios, an approach most explicitly espoused by Bickerton via his ‘pidginization index’ (1984:176-8). For those committed to a more sociologically rooted approach to the subject, it might be tempting here to suppose that expression of identity was the dominant factor in the birth of HCE. However, in broader view, this would raise a difficult question: If group identity was why multiethnic schoolchildren in Hawaii devel-


107

oped a creole despite being surrounded by full English, then why has this not happened throughout the world in multiethnic nations where a single official language is the medium of instruction in schools? For example, if HCE arose simply as a vehicle of identity, then why have multiethnic schoolchildren in urban schools in countries like Congo (formerly Zaire) or Tanzania not spontaneously developed creolized varieties of the language of teaching? As long as we find it unremarkable that no French creole has developed in Kinshasa or that no English creole has developed in Dar es Salaam (or even in schoolyards in American cities like Oakland, where children often speak as many as a dozen foreign languages natively), we must assume that something more was at work in the emergence of HCE. In this light, a factor that distinguishes turn-of-the-century Hawaii from Tanzania is that in Hawaii there was, in addition to the varieties of full English, the range of pidgin Englishes spoken by immigrant adults, as demonstrated by Roberts (1998). While hardly discounting the concomitant presence of pidgin Hawaiian at the time (Bickerton & Wilson 1987; Roberts 1995), pidgin Hawaiian was strongest on plantations; in towns and cities, pidgin English was a vigorous presence by the 1880s, and it is here that schoolchildren developed HCE. It is impossible that these children created HCE because of input deprivation when they spent five days a week being instructed in English. Nevertheless, it is plausible that for children surrounded by parents and other adults using pidgin English, as a vehicle of identity, the use of English with reduced morphology, simple syntax, and fluid lexical categorial boundaries was a natural choice. Where no such pidgin varieties of the school language are ‘in the air’, so to speak, even schoolchildren speaking a wide variety of home languages simply learn the full school language, with slang lexical items and expressions serving as the markers of in-group identity, rather than a vast reduction and reconstitution of the language’s structure per se. There is a ‘control case’ available supporting this analysis of Hawaii, in the form of the development of Singapore English (‘Singlish’). Like HCE, Singapore English was created in English-language schools by children speaking a range of first languages (Chinese dialects, Hindi, Bengali, Tamil, and Malay), and expresses a local identity much as HCE today does, with even fluent speakers of English retaining Singapore English as a casual alternative (Platt 1975). Yet Singapore English would not be considered a creole under any analysis. English inflections are somewhat reduced, but vigorous: Bare thirdperson singular verbs are found, but quite variably even in basilectal speakers, and zero past marking occurs largely in a systematic opposition to marked pasts to encode subtle semantic and discourse distinctions. Zero copula occurs variably, but there are no distinct equative and locative copulas, no complex of preposed tense and aspect markers, no restructuring of the NP determination

108


apparatus, and in general, none of the sharp departures from English syntax which distinguish Jamaican patois, Sranan, or Tok Pisin. (Ho & Platt 1993). Here is a typical sample: Then sekali (Malay “suddenly”) the teacher come in. We all chabot (Malay “run away”) off ah. Then when de Sir (teacher) want to sit down den she shout you know. She shout you know. Den de Sir tot (thought) what happen [y’]know. Den he se, ‘Got kum -- dose pin ah, kum (thumb) tacks ah.’ So den de Sir as(k) who lah (expressive particle) (Ho & Platt 1993:135).

Clearly, while this variety could be classified as a ‘creoloid’ or semi-creole (Platt 1975), in general it registers more as a ‘kind of English’ than as a completely different language. An important difference between the genesis contexts of HCE and Singapore English is that pidgin English was a much more marginal presence amidst the development of the latter in the 1940s. Rather than being the increasingly prevalent interethnic lingua franca that it was in late nineteenth-century Hawaii, pidgin English in mid-twentieth century Singapore was spoken only by some older immigrants when communicating on a utilitarian basis with whites. The interethnic lingua franca was Bazaar Malay (Platt 1975:364). As a result, pidginized English was a much less integral part of the linguistic context the creators of Singapore English lived in. We must now recall that in most countries where there is no pre-existing pidgin English model at all, multiethnic schoolchildren’s in-group language is not even as reduced as Singapore English. In this light, in seeking to attribute the different outcomes in Hawaii and Singapore’s schools to a systematic factor, we can hypothesize that the lesser influence of Singapore’s pidgin English resulted in an in-group language less reduced in comparison to English than HCE. In other words, it may be that Hawaii, Singapore, and Tanzania represent a cline of contexts in which the prevalence, marginality, and absence, respectively, of a pre-existent pidgin English have determined the extent of reduction and restructuring in the in-group speech developed by its multiethnic urban student bodies. It must be said that in attributing the birth of HCE to the presence of a pidgin spoken by adults, I am in no sense attempting to rescue or recast my earlier claim that HCE was developed by adults. On the contrary, Roberts’ work has convinced me, in a way that Bickerton’s nimble but insufficiently supported guess did not, that the grammar of what we know as HCE was essentially a child creation, and I find this fascinating. Part of what fascinates me, however, is that these children did create this language despite having full access to English, contra Bickerton’s reconstruction. Given that children of various ethnicities do not create creoles in schoolyards every day, and the availability of useful sources of comparison like


109

Singapore, I am suggesting that the adults’ pidgin varieties served less as a direct ancestor to HCE than as the spur for a degree of paradigmatic and structural reduction that would not otherwise have been the case. Specifically, Hawaii shows us that a creole can develop as an in-group code even amidst rich, directed input from its lexifier—but only when a pidginized variety of the lexifier is already spoken by some segment of the society to serve as a model and inspiration. HCE, under this analysis, did not simply spring into existence on its own—the pre-existent pidgin was a necessary condition for its birth. 3. The French continuum anomaly 3.1 The problem One of the strangest, most persistent problems in creole studies is that French plantation creoles do not exhibit a continuum of lects shading gradually into the local standard, as English Caribbean creoles do. To be sure, there are varieties of French creoles that lean towards the standard, such as varieties of Haitian which incorporate front rounded vowels. However, the fact remains that in Haitian, the various Antillean varieties, Guianese, and Seychellois, there is a distinct break between creole and French, such that there is no band of lects not readily classifiable as standard or creole as there is in, for example, Jamaica and Guyana. It is difficult to see how creole genesis theory, as currently configured, could ever provide an answer to this question. We must note that even an answer based on exquisitely detailed sociohistorical research on a single colony, read to suggest that the local conditions would have kept a standard and creole separate, would be inadequate because of the nature of our data set. Continua are oddly absent in colonies formerly run by one particular power, and thus our explanation must apply ‘productively’, so to speak, across the French Caribbean as well as in the Indian Ocean. Any sense that the French may have maintained more social distance between themselves and their slaves than the English appears belied by the ample documentation of mulatto classes of slaves in French Caribbean colonies. One must also ask how likely it would be that any such cultural tendency, even if nominally enshrined in codes noirs, would be upheld so consistently by every plantation supervisor in each of so many separate colonies. Clearly, something broader and more decisive was at work. That this problem has remained a puzzle for 30 years suggests that it will be useful to consider a complete change of lens in addressing it. For example, we might surmise that the French situation—where plantation societies led to two distinct varieties (standard and creole)—was not exceptional, but a norm. This would eliminate the French situation as a ‘problem’ and render it simply predictable, leaving us with the English continua as an anomaly to ‘explain’. The advantage is that, upon examination, there is promise in this approach.

110


3.2 The history of the English creole continua To see why, we must first briefly refer to my argument in McWhorter (1995a) that all of the Atlantic English-based creoles (AECs) trace back to a single ancestor distributed across and beyond the Caribbean. This argument is based on six features shared by most or all of these creoles which are distinct in being too idiosyncratic to have emerged in several separate locations, and can only have emerged in a single grammar. The equative copula da (with its cognates na, a and duh), the locative copula de, and the modal use of fu (or cognates fi or fuh) to mean “should” 2 are idiosyncratic grammaticalizations representing particular choices out of many possible, with it thus being impossible that the ‘roll of the dice’ would have come out so similarly in several separate colonies. The second-person plural pronoun unu, the past marker bin, and the use of self as “even” are, respectively, a borrowing, an etymological recruitment, and an adstrate calque which are all, again, particular choices out of many possible and thus also indicate a single ancestor (for example, would every AEC use an African word for a second-person plural pronoun specifically, and all from the same particular language, Igbo?). The influence of the polygenetic frame of reference is so powerful in the field that it is tempting to suppose that the creoles emerged independently, but that these features ‘diffused’ from one creole to another. Yet while diffusion can certainly explain some AEC commonalities, it is inapplicable to these six features. First, they are core grammatical items and thus relatively unlikely to be borrowed into a grammar; second, it would strain credulity to suppose that no fewer than six items, all already unlikely to diffuse, would end up distributed so uniformly across several creoles. Basic historical linguistics, bolstered by the rich sociohistorical relationships between the various colonies, makes the reconstruction of a single ancestor obviously the most economical choice. To wit, the ancestor of the AECs is most likely to have been distributed from Barbados, one of the earliest colonies settled by the English and the source of the colonization of many subsequent locations, such as Suriname, Jamaica, South Carolina, and Guyana (Cassidy 1983). 3 Settlers of new colonies often either brought slaves with them or later bought slaves from Barbados, thereby transplanting the ancestral pidgin or creole from one colony to another (Baker 1998). 2

The argument refers strictly to this modal usage; the uses of fu (and its cognates and allomorphs) as preposition and complementizer could easily have been incorporated from English in several separate locations and thus have no bearing upon my argument for a common ancestor (cf. Bickerton’s misinterpretation of my reasoning [1998:87]). 3 The jury currently remains out on Baker’s (1998) and Parkvall’s (1995b) argument that AEC traces even further back than Barbados to St. Kitts, settled slightly earlier. However, the importance of Barbados as a distribution point is not in question, and the actual formation of AEC on St. Kitts would have no bearing on my lines of argumentation here.


111

Today, even the Barbadian speech furthest from standard English is essentially mesolectal compared to the basilects of Jamaica or Guyana. However, even Bajan has four of the six ancestral AEC features (de, unu, self, and bin [the latter recorded in Rickford 1992]), with the absence of the particularly basilectal da and modal fu unremarkable. Moreover, sociological and historical evidence suggest that a more basilectal variety was once spoken in Barbados (Rickford & Handler 1994). Indeed, it is virtually impossible that this was not the case: The six features that AEC as a whole share must trace to a single source, and history demonstrates that this source must have been Barbados. Thus comparative reconstruction takes its place alongside the sociological and historical indications to make the former existence of a Barbadian basilect almost certain. Where this applies to the continuum issue is that while McWhorter (1995a) addresses basilectal AEC features, there is no a priori reason to assume that only the basilect was transplanted from one colony to another and, in fact, all reason to suppose that the basilect was transplanted along with the mesolect (cf. Bickerton 1996:324 for a similar observation). What suggests this is that the AEC mesolects, too, show signs of being traceable to one instance of restructuring rather than separate ones in each colony. First thought to have developed after emancipation, given that blacks had richer encounters with the standard, today the mesolect is thought by most creolists to have been in place at the outset of a colony’s history, developed by those slaves whose positions gave them more access to the standard than field hands (Baker 1990; Bickerton 1996). Whether we conceive of the continuum as having developed early or late, however, the implication of most authors is that this development would have been in a given colony itself. If this were true, however, we would not expect AEC mesolects to share certain features. Most strikingly, for instance, AEC mesolects have zero copula in equative and locative sentences in contrast to AEC basilects, which have overt copulas in such sentences. Guyanese copula examples are provided in Table 1 (adapted from Bickerton 1973). The facts are similar in Jamaican (Holm 1984; LePage & DeCamp 1960), Gullah (Geraty 1990), Bajan (Rickford 1992), and other AECs. Under the old ‘bottom up’ conception of continuum development, this was attributed to mesolectal speakers having ‘unlearned’ their basilectal copulas but not yet acquired the acrolectal ones (Bickerton 1975). While ingenious, this notion was already problematic (McWhorter 1995b:340-1), but under the current ‘top down’ conception it becomes utterly untenable: why would those speakers with more contact with the standard have zero copula, a hallmark of pidginization, while those with the least contact, likely to pidginize the lexifier to the greatest extent, have overt ones? Moreover, exactly what would explain mesolectal speakers not retaining the acrolectal copula, if their access to the target was relatively free?


112

Acrolect Mesolect Basilect

Equative iz ø a

Locative iz ø de

Table 1: The copula along the Guyanese creole continuum And finally and most importantly, whatever answers one could provide to those two questions, what is the likelihood that things would come out in these particular ways as a regular occurence, happening in Barbados, Jamaica, South Carolina, Guyana, Trinidad, Antigua, and so forth.? Thus of course, ‘things happen’ in language change and language contact; flukes are part of the territory, Grimm’s Law being a prime example. Certainly it is not unheard of for a mesolectal, or semi-creole, variety to have zero copula. Crucially, however, semi-creoles retain copulas as often as they drop them: Popular Brazilian Portuguese retains the copula, Singapore English can drop it; Afrikaans retains the copula, Réunionnais can drop it (Cellier 1985:73), and so forth. In other words, zero copula in a mesolectal variety is not something we would reasonably expect to have occurred in all AEC mesolects; what we would expect is for it to occur in some but not all. Thus its actual distribution suggests that mesolectal AEC developed once and was subsequently distributed. Another feature of mesolectal AEC suggesting this is the use of did as a past or anterior marker, in Guyanese, Jamaican, and Trinidadian (Winford 1993:64-66) and other AECs. As Bickerton (1996:315-6) notes, it is unlikely that this feature is simply a reflection of the pleonastic use of do in simple declarative sentences in earlier English, because this feature was passing out of general usage in most English dialects as early as the 1500s. This is especially important in reference to Guyana, where the English did not establish plantations until as late as the mid-1700s. While the feature held on somewhat longer in Irish and West Country English dialects, such speakers were hardly distributed so uniformly and predominantly across the English Caribbean that we would expect this particular feature to be incorporated by so many separate creoles. Once again, if these mesolects had developed independently we would expect more variation between them: One or two would have did, while another might have overgeneralized was (as one variety of Liberian Pidgin English does [Singler 1981]) or had (as African-American Vernacular English now does in narratives [Rickford & Théberge 1996]). Much more likely, then, is that this feature was incorporated once in one place, most likely from regional British dialect influence in the 1600s, and subsequently spread elsewhere. Thus basic principles of reconstruction lead us to the conclusion that not only the AEC basilect traces to a single ancestor, but that the AEC mesolectal band does as well. It must be emphasized that the argument is not that this basilect and mesolect remained pristine amidst these transplantations. The

113


basic template was obviously altered in each location it reached, according to the nature of the English dialects spoken, the substrate language mixture, degree of contact between slaves and whites, etc., the result being the various different creoles extant today. Yet while obviously distinct enough to be considered separate languages, all of these creoles share a particular core of features whose nature clearly reveals a direct historical relationship between all of them, and rendering it implausible that, for example, Gullah formed independently in Charleston with mere marginal ‘influence’ from Bajan. 3.3 Brief excursus: The reality of transplantation Some creolists may find themselves resisting this argumentation on the basis of a conviction that even if slaves from a former colony imported something identifiable as a grammar, that it would have been impossible for this grammar to survive transmission to massive numbers of new adult slaves (e.g., Bickerton 1998:85-6). This is an understandable intuition, but is, quite simply, refuted by the comparative linguistic data: the AECs are too closely similar in idiosyncratic ways for each imported variety to have been ‘pulverized’ by waves of new imports. If these languages had developed in any sense independently, even with correspondences in superstrates, substrates, and universal tendencies, we would expect much more variation than we in fact find. This is because of the vital role that chance plays in language contact, expressing itself even within the broad constraints that source languages and universals impose. For example, it is uncontroversial that Tok Pisin English Creole is the result of an encounter with English completely separate from that which created AEC. Here we contrast equivalent sentences in Tok Pisin and Sranan: (1) a. Tok Pisin Em mipela i bin kirap-im dispela wok. PRED ANT start-TRANS this work it we “It is us who started this work.” (Mühlhäusler 1985:352) b. Sranan Na wi di ben bigin a it-is we REL ANT start the “It is us who started this work.” (2) a. Tok Pisin Mi stap long haus. he COP LOC house “I am in the house.”

wroko disi. work this

114


b. Sranan Mi de na ini a he COP LOC inside the “I am in the house.”

oso. house

Some creolists might suppose that the differences between these two creoles are attributable to differences in the superstrate, substrate, or timing of demographic developments. This is definitely valid to an extent—but leaves a crucial realm of the evidence unaccounted for. For example, substratal influence is clearly one factor in the differences between the two creoles. The predicate marker and transitive marker in the Tok Pisin example (1a) are undisputed calques on Melanesian languages (Keesing 1988:119-27, 143-70), and the -pela marker is also possibly traceable to substrate behaviors (Keesing 1988:113, 137-9). Meanwhile, the choice of di (> disi “this”) as a relativizer in Sranan is firmly traceable to West African patterns (Bruyn 1995). However, other things are more fortuitous. It is a universal tendency for positional verbs to serve as locative copulas, cross-linguistically and in pidgins and creoles. However, while Tok Pisin observes this universal, Sranan happens not to do so, opting instead for an adverb there. It would be difficult to ascribe this to any parseable factor; for example, it cannot be traced to the substrate, since Sranan’s West African substrate languages use verbs, not adverbs, as locative copulas. The AEC choice of de was a matter of chance (see McWhorter 1995a:292-9). Similarly, there is a universal tendency in pidginization to utilize one or two items alone as prepositions. However, this universal allows a wide variety of etymological choices. What factor could ‘explain’ the choice of long (< along) in Tok Pisin as opposed to na (either from Portuguese “in + DEF” or the Igbo general preposition) in Sranan? Tok Pisin radically reinterpreted a preposition from English; Sranan chose a foreign item—one would search in vain for a ‘reason’ for the difference in outcomes. Because chance alone ensures such distinct outcomes in separate instances of pidginization even apart from differences in source languages, it is clear that the AECs must indeed be the result of the transplantation of a single system to various places, in each colony adapted, to be sure, but by no means rebuilt from the ground up. If they had developed in any sense separately, then the AECs would be much more different on all levels than they are. 3.4 Solution: The French colonies as default At this point, many readers will naturally suppose that even if the English continuum was transplanted from colony to colony, that if it had not been, then continua would have developed independently in each place anyway. Cru-


115

cially, however, the French colonies may be telling us that in fact, this is not what would have happened. To wit, the genetic relationship between the English creole continua on the one hand, and the odd absence of French creole continua on the other, can be subsumed under a coherent analysis if we propose the following. One interpretation of the current distribution of creole continua in the Caribbean is that the typical sociolinguistic result of plantation slavery was a creole distinct from the standard, with no unbroken continuum of lects in between; this result would be represented by the French colonies. Barbados, however, can be seen to have been uniquely suited to the development of such a continuum, and when the creole that developed there was transplanted to other colonies, the continuum of lects was transported along with it. This continuum thus took its place in colonies where, if AEC had not been imported and a creole had formed locally, the default situation—a creole distinct from the standard with no continuum, as in the French colonies—would have obtained. In evaluating this hypothesis, the first thing we must ask is whether there indeed was anything about colonial Barbados which would have made the development of a continuum particularly likely where it was not elsewhere. Presumably, what we seek is evidence of contact between white and black significantly richer than it was in other English colonies. In fact, as those familiar with English Caribbean history will immediately note, Barbados is just such a place. It is hardly accidental that today, Bajans are the most notoriously acrolectal English Caribbeans, with mesolectal registers available but hardly as readily elicited as in Guyana, Jamaica, or even the Sea Islands, and now spoken regularly only by isolated, elderly people. As in other plantation societies, blacks vastly outnumbered whites in Barbados by the mid-1600s; however, the proportion of whites nevertheless remained higher than in other English Caribbean colonies (Handler 1974:10), partly because indentured servitude was particularly prevalent. In general, as Roy (1986:143) notes, the small size, flat topography, and early construction of roads ensured a richer contact between black and white than was possible in, for example, much of mountainous Jamaica. Thus the absence of French creole continua, when viewed against the origin of the English continuum in Barbados, may suggest that plantation conditions alone were not sufficient to the development of a continuum, and that typical plantation societal conditions, such as, for example, those in Martinique, kept a standard and a creole distinct in the minds of speakers. Instead, it would appear that particularly rich contact between black and white was necessary. In support of this, it is significant that we also find a creole continuum, less assiduously documented than the English Caribbean ones but obviously similar, in Cape Verde (Bartens 2000), where plantations were always small

116


because of poor soil and thus black-white contact was always relatively intimate. However, no such continuum is found in São Tomé, where in its initial colonial period, large plantations were worked by Africans in relative isolation from whites. Similarly, the one New World French creole which has developed a mesolect is Louisiana French Creole, whose speakers have existed in a uniquely intimate relationship with similarly poor, isolated Cajun French speakers for centuries. Another useful example is that there has never existed a continuum between Spanish and Palenquero, despite its speakers having apparently always been bilingual in Spanish (Schwegler 2000). The social demarcation between Palenquero speakers and outsiders has prevented any such continuum from emerging. In sum, then, if we shift our focus from polygenetic conceptions to the seemingly mundane fact that most creoles are descendants of pre-existing ones, then one advantage is that we have a possible solution to a uniquely intractable anomaly which has otherwise remained unsolved for decades. 4. The Spanish creoles If we are positioned to focus as much upon pre-existing pidgin or creole models as upon demographics, then we are also prepared to constructively address another truly baffling problem in creole studies, the scarcity of Spanish-based creoles. The problem here is much more serious than most authors currently know. It is relatively well-known that the Spanish established plantation societies worked by Africans in Cuba, Puerto Rico, and the Dominican Republic. However, Chaudenson (1992:124-8) and myself (McWhorter 1995c:223-6) have shown that the absence of creoles in these locations can be attributed to timing of demographic developments; namely, slaves were used in small numbers for long periods during which they were able to learn relatively full Spanish; when massive importation of slaves finally began, these older slaves could pass their competence on to these new ones (cf. Baker & Corne 1982 for the introduction of this model). What is virtually unacknowledged in the field, however, is that the Spanish also established thriving plantation colonies worked by Africans in their mainland colonies, such as modern-day Colombia (particularly the Pacific lowlands), the Chota Valley of Ecuador, Mexico, Peru, and Venezuela. Most importantly, in these places Africans were imported in large numbers at the outset, with no société d’habitation “homestead society” phase at all. Furthermore, the Iberian racial tolerance and restraining influence of Catholicism often cited as having impeded the development of Spanish creoles was nowhere in evidence in these settings, where large-scale agriculture or mining engendered the same dehumanization of African laborers as obtained in betterknown colonies like Suriname and Haiti. I discuss this issue in detail in McWhorter (2000).


117

Why no creoles appeared in these colonies is an issue we will address in the following section. However, the first question to be resolved is why Spanish creoles did appear in Curaçao (Papiamentu) and El Palenque de San Basilio (Palenquero). In light of the past two sections, what is most interesting here is that Papiamentu and Palenquero share a certain factor: Specifically, there is concrete evidence that a pre-existing pidgin or creole was highly influential in the genesis contexts of both. Until recently, older people in Curaçao controlled a secret variety they called Guene (< Guinea, i.e., the Guinea Coast of Africa), which they considered to be the language spoken by their ancestors. Significantly, ‘Guiné’ is still what some native speakers of Guinea-Bissau Creole Portuguese call their language (Birmingham 1976:19), and in general Guene had features tracing it to Portuguese-based contact languages of West Africa, most strikingly the third person pronoun ine, a substrate borrowing also found in the Gulf of Guinea Portuguese creoles. If Papiamentu had emerged as a Spanish-based contact language, then we would expect any preserved ‘slave’ language to be Spanishbased, like the bozal Spanish similarly preserved as a ritual language among Afro-Cubans (Cabrera 1954). The relationship between a Portuguese creole and Papiamentu is further bolstered by well-documented clues in the language itself linking it to Portuguese creoles still spoken on the West African coast. For one, there are a healthy number of lexical items in Papiamentu derivable only from Portuguese rather than Spanish (most assiduously and rigorously identified by Grant 1996). In addition, on the grammatical level, the plural morpheme nan is also found in the Portuguese creole of Annobón (Birmingham 1976:22). Similarly, the parallel between the Cape Verdean Portuguese el taba ta kanta “he was singing” and Papiamentu e tabata kanta is striking (Birmingham 1976:20), since this usage is impossible to derive from any Iberian construction and is only one of many possible reconceptualizations of the lexifier material. Recently, Quint (1998) has documented an impressive range of parallels between Papiamentu and Cape Verdean Portuguese Creole which render the case for a direct relationship even stronger. (See McWhorter 1995c and 2000 for details.) Meanwhile, in the case of Palenquero, in reference to his long-term residence in Cartagena, near where Palenquero is spoken, the father Sandoval noted in 1627 that there were many slaves who had lived in São Tomé who used a “highly corrupt and backwards” version of Portuguese “which they call the language of São Tomé” (un género de lenguaje muy corrupto y revesado de la Portuguesa que llaman lengua de S. Thome) (cited in Schwegler 1998:229). This passage suggests that many of the originators of Palenquero already spoke a form of what is today São Tomense Creole Portuguese. The connection between São Tomense and Palenquero is supported by linguistic evidence. The third person plural subject pronoun is iné in São To-

118


mense, while other Portuguese-based creoles show reflexes of standard êles (élis in Guinea-Bissau Creole, for example [Kihm 1980:44]). Palenquero has ané, rather than a reflex of Spanish ellos. 4 Palenquero also has a postposed anterior marker -ba (ele kelé ba “he wanted”) which is also found in the Upper Guinea Portuguese creoles. The most likely source for this -ba is acabar “to finish,” found as kabá in other Caribbean creoles such as Sranan and Saramaccan. The particular elision to simply ba in both Palenquero and the Upper Guinea Portuguese creoles suggests yet another link between Palenquero and West African coastal Portuguese pidgins (although São Tomense lacks this particular feature). The Guene connection in Curaçao and the São Tomé connection in Cartagena are striking, in that it is precisely where we have these two historical indications of a contact language already in the air that Spanish creoles exist today. If in addition to the historical clues from Curaçao and northwestern Colombia, we also had documentation of Cape Verdean-speaking slaves in Venezuela and mentions of slaves from São Tomé in Mexico, and yet creoles were spoken only in Curaçao and northeastern Colombia, then there would be no reason to view these pre-existing contact languages as significant in the linguistic outcomes in various locations. However, what we have is historical citations of pre-existing creoles in Curaçao and northwestern Colombia, and creoles spoken today in—Curaçao and northwestern Colombia. What clinches the case is that both creoles have clear signs of Portuguese creole influence—if they had arisen simply via slaves’ limited access to Spanish, we would expect much less, if any, of this kind of influence. It would be unrealistic to analyze Papiamentu as a relexification of Cape Verdean itself, or Palenquero as a relexification of São Tomense. Nevertheless, the existence of Spanish creoles in just the places where Portuguese creoles are documented to have been imported by early slave populations suggests some causal link. In view of the role that we have seen that pre-existing contact languages can play in Hawaii, these facts suggest that the presence of forms of West African Creole Portuguese in Curaçao and Cartagena were crucial to the emergence of creoles in these two locations. It is plausible that the very prevalence of a reduced contact language, especially one based on Portuguese— even more closely akin to Spanish in the 1500s and 1600s than it is today— made the adoption and stabilization of a reduced version of Spanish as an interethnic communication vehicle seem natural. Just as pidgin English seems to have served as a model modus operandi for young Hawaiians if not as a 4

Schwegler (forthcoming) traces this to a Kikongo demonstrative meaning “those,” such that technically the two creoles could have incorporated the Kikongo pronoun independently. This, however, is unlikely: Why would both happen to choose an African etymon for the third person plural specifically, and why would both happen to recruit a demonstrative rather than personal pronoun in the function?


119

direct template, the pre-existing Portuguese creoles in Curaçao and Cartagena, while not serving as direct sources of relexification for the new creoles, obviously played an adstratal function. They left not only lexical items but had a grammatical impact, their analytic typology serving as a model for the new language developing. Most important, where such creoles are not documented to have existed in the genesis context, no creole emerged. Under limited access-based theory, the fact that a creole emerged in Curaçao but not in Venezuela is an eternal problem, upon which demographic tallies would be extremely unlikely to shed useful light (e.g., blacks and whites had rich contact in Curaçao and yet there is a creole, and little contact on Venezuela plantations and yet there is no creole). On the other hand, we see that once again, calling attention to pre-existing contact languages allows a potential solution. 5. The Afrogenesis hypothesis However, we are now left with the question as to why no creole formed in Venezuela and similar Spanish colonies. Along those lines, let us review these two statements from leading creolist thinkers, first quoted in the introduction. Alleyne (1971:180): In the New World, on sugar plantations, production was organized on the basis of a kind of occupational stratification according to which field slaves were most numerous and were furthest removed from contact with Europeans. Social intercourse of the field slaves was almost exclusively confined within the group; and so, among them, linguistic forms showing a high degree of divergence...were able to crystallize and achieve the appearance of stability.

Bickerton (1988:271-2): In the plantation situation, the preparatory phase of sugar colonization gave way...to the exploitative phase...requiring a rapid increase in the numbers of unskilled manual laborers;..dilution of the original model must have resulted...In fact, what took place...was second language learning with inadequate input.

These statements are nothing less than canon in creole studies, and yet in fact they touch upon what is in fact the most pressing discrepancy in the field. Namely, on sugar plantations in Venezuela, the “social intercourse of field slaves was almost exclusively confined within the group,” as Alleyne puts it with Jamaica in mind; and yet “linguistic forms showing a high degree of divergence” are nowhere in evidence; black Venezuelans simply speak Spanish even when living in Afro-Venezuelan communities (Megenney 1985). Similarly, when in seventeenth-century Ecuador there was “rapid increase in the numbers of unskilled manual laborers,” as Bickerton puts it with Guyana in mind, there was no appreciable “dilution of the original model” in terms of

120


slaves’ acquisition of Spanish. The descendants of slaves brought to work sugar plantations in Ecuador’s Chota Valley today speak a Spanish only minimally variant from the local standard, despite having always lived a ‘life apart’ from other Ecuadorans (Lipski 1986). Thus our crucial discrepancy is that vast disproportion of black to white, so confidently adduced as key to the development of creole languages, had no such effect throughout mainland Spanish American colonies. We have seen an explanation for why Spanish creoles did appear in two places; however, the question remains as to why they did not appear in the Chocó, Colombia, Chota Valley, Ecuador, Mexico, Peru, or Venezuela, all places where huge numbers of African slaves were imported to work under conditions in all ways akin to those obtaining under other colonial powers. None of the strategies used to explain the absence of creoles in other locations works for these cases. The enormity of this problem will become clear with one brief ‘case study’. In the late 1600s, the Spanish began importing massive numbers of West Africans speaking a wide variety of languages into the Pacific lowlands of northwestern Colombia to work their gold mines. Today, the descendants of the Chocó slaves live in the same lowlands, leading a subsistence existence via small-scale mining. Whites are a negligible presence and interethnic relations are distant (Rout 1976:243-9). Yet the Spanish of black Chocoanos is essentially a typical Latin American dialect of Spanish, easily comprehensible to speakers of standard Spanish varieties: (3) Esa gente som muy amoroso. Dijen que...dijeron COP very nice they say that they say-PAST that people que volbían sí...cuando le de su gana a ello vobe that they-return-IMP yes when to them give their desire to them return “Those people are really nice. They say that...they said that they would come back...when they felt like it.” (Schwegler 1991:99) The reason for this is not that blacks did not outnumber whites significantly. There were no fewer than 5,828 black slaves in the Chocó by 1778, while there were only about 175 whites—a mere 3% of the total population (West 1957:100, 108). The reason is not that slaves and whites had an unusual degree of contact. Slaves were organized into large teams, or cuadrillas, each formally supervised by a white overseer but actually directed by a black capitanejo (West 1957:131-2). Cuadrillas typically consisted of two hundred blacks or more, with ones as large as 567 reported (West 1957:115-6). The reason is not that there was a long société d’habitation “homestead society” phase where Africans could acquire relatively full Spanish and pass it on to later arrivals. The nature of mining is such that relatively large numbers of


121

slaves were needed from the outset. One of the earliest cuadrillas was established with forty slaves and was increased to sixty-five later that year (Restrepo 1886:77-8), and the sharp disproportion of black to white increased by leaps and bounds: There were 600 slaves in the Chocó in 1704, 2000 in 1724, and 7088 by 1782 (Sharp 1976:21-2). Finally, the reason was not that the Spanish were gentler masters. There was a slave code enacted in Colombia in 1789 to protect slaves from abuse, but besides the fact that its very institution suggests the brutality which had obtained over the century of mining beforehand, it was barely enforced (Sharp 1976:128). Catholic clerics were in short supply and were often slaveowners themselves (Sharp 1976:130-1). Flogging was a regular punishment for any perceived offenses (Sharp 1976:139-40), and the most unequivocal evidence of the misery of the slaves’ lot is the simple fact that escapes and revolts were common (Sharp 1976:140). The facts are similar in all of the colonies I have mentioned (see McWhorter 2000 for details); the Spanish American mainland stands as a stark contradiction to the very fundamentals of creole genesis theory, which despite their variety all stipulate that disproportion of black to white led to vast reduction and restructuring of a dominant language. A general distrust of mechanistic approaches to creole genesis reigns at this writing among creolists, sociolinguists, and anthropologists. In this light, a natural response to these observations might be to take them as evidence that creolization will not reduce to ratios and formulas, and to suppose that our solution will come from thorough study of the social history of each of these anomalous colonies. However, acknowledging the usefulness of such studies in themselves, our problem runs deeper than this approach could reach. An attempt to account for why a creole would not have formed in one of these colonies, no matter how assiduous in itself, would fail to address the heart of this problem, which is that creoles failed so consistently to appear under the Spanish. If idiosyncratic combinations of intra-colonial factors were the culprit here, then we would expect mysteriously absent creoles not in almost all Spanish colonies, but in maybe two, and also in, say, Guadeloupe and Antigua as well. Thus we seek not to speculate upon the operations of the local factors that happened to obtain in one given place, but a systematic factor which could determine the linguistic outcomes in several separate places sharing a common feature, in this case, proprietorship by one particular power. In this light, then, it is also highly implausible that creoles were once spoken in all of these former Spanish colonies but have since disappeared, pace Granda (1978) and Schwegler (1993, 1996). One might propose an account as to why a creole might have disappeared in one of these places, but properly speaking, the only valid address is one which explains why this would happen so oddly frequently only under the Spanish. Moreover, claims that prescriptiv-

122


ism is strong in Spanish culture must account for the fact that French creoles survive despite equally strong prescriptivist impulses in French culture. There is one identifiable factor which truly sets the Spanish apart from other powers. The Spanish were the only slaving power not to establish trade settlements on the West African coast, because of the Treaty of Tordesillas, under which the West African coast fell on the Portuguese side of a line dividing exploration rights between the two Iberian powers. Seemingly irrelevant to sociolinguistic outcomes across the Atlantic Ocean, this absence of trade forts becomes more interesting in that some creolists have surmised that English West African trade settlements were the birthplace of the pidgin which, transported across the ocean, became the AECs, while others have surmised a similar origin for French plantation creoles. 5 To outline the argumentation for these reconstructions would take us too far from the concern of this paper, but following and suggesting revisions to Hancock (1969, 1986), I have presented a case for the birth of AEC on the Ghanaian coast in McWhorter (1996), more fully limned in McWhorter (2000). Meanwhile, Parkvall (1995a) has been the most recent adherent of this kind of scenario for the French creoles. For our purposes, it will suffice to observe that a great many aspects of English and French creoles’ linguistic structure and history point in the direction of a West African pidgin origin. Oddly enough, these trade fort scenarios relate to our Spanish anomaly. If the AECs show signs of tracing to a West African trade fort pidgin, and the French creoles do as well, while the one power that had no such trade forts also left behind very few creoles, then one interpretation of the facts is that the trade forts were pivotal, rather than incidental, to the birth of the New World creoles. In other words, it may well be that New World creoles were born in West African trade forts and only imported to, rather than emerging on, plantations. If so, then the reason the Spanish left behind so few creoles would be because they had no trade forts where an ancestral pidgin could form and be spread throughout its New World colonies. Supporting this would be the very fact that the two Spanish plantation creoles that do exist both apparently trace their origin to imported contact languages, in these cases Portuguese pidgin or creole. It is unlikely to be accidental, as we have noticed, that we have Spanish creoles precisely where history or anthropology reveals these imported Portu-

5

I do not address the Dutch creoles in this paper because accidents of history make it difficult to draw conclusions from them. There have been only three Dutch-based plantation creoles. Berbice Creole Dutch has so heavy and deep-rooted a contribution from a single West African language, Eastern Ijo, that it in fact belongs more properly not to the creole, but the intertwined language class (e.g., Angloromani, Michif) (McWhorter 1999). This leaves just Negerhollands and Skepi Dutch. This already small basis of comparison is further restricted by the fact that both are now extinct, and Skepi was only sketchily documented before its death.


123

guese contact languages to have been spoken by early slaves, but none where we do not. McWhorter (2000) is a full-length presentation of the evidence for the Afrogenesis Hypothesis. If valid, it suggests another change of lens, along the lines of the one I suggested regarding the conditions for creole continua to emerge. Once again, we can account for the facts if we transform what we now take as a problem into the default situation, and recast what we now think of as ordinary as the ‘exceptional’ case. Namely, the mainland Spanish colonies appear to be telling us that ordinarily, slaves were capable of learning a relatively full version of the whites’ language on a plantation even when outnumbering them vastly. As counterintuitive as this seems to anyone familiar with the limited access paradigm that creole studies has settled upon, in fact this is precisely what we see again and again in the mainland Spanish colonies. It is obvious that most adults are not capable of full acquisition of second languages, and the sociological conditions on plantations and in mines would have exerted even more of a downward pull upon the competence of all but a few slaves. Most likely, it was children who had richer opportunities for acquisition than we might think. Hence, technically speaking, plantation conditions may well have limited adults’ access to Spanish, working in tandem with the lesser language-learning abilities of most people after their early teens. However, what we see is that such conditions did not limit access to such an extent that the only Spanish that slave children heard was the pidginized varieties spoken by their parents. Hopefully, future research will elucidate how these children acquired full Spanish. Most important for us to realize, however, is that there is a question only as to how the slaves did this; as we have seen, there simply is no question as to whether they did it—the Spanish-speaking isolated Afro-Hispanics in Mexico and South America admit no other interpretation.6 Why, then, if slaves could learn the superstrate, did slaves under the English and French develop creoles? Here is where we come back to the topic of this paper: I have hypothesized that castle slaves working in the Cormantin 6

Granda (1978:416-7) claims that a now-extinct Spanish creole was once spoken further south in the town of Uré by descendants of slaves who escaped from gold mines. Granda gives no data and was unable to visit the town. However, if this variety was indeed a creole, then despite the fact that there is no evidence tracing it to Portuguese pidgin or creole, it nevertheless leaves intact my claim that Spanish America is crucially problematic to limited accessbased creole genesis theories. My claim is not that limited access to a target language cannot drive creole genesis, but that plantations (and mines) did not limit access to Spanish to the extent that pidgin-level competence would be passed on to future generations, in contradiction to all expectations according to creole genesis theory. In relation to Uré, it is quite natural that adult maroons, most of whom had probably spent little time in the mines, would speak only pidginized Spanish and, if having no subsequent contact with whites, would creolize it. These, however, were conditions quite distinct from the plantation or the mine.

124


trade fort in Ghana and speaking a pidginized English would have been transported most likely to Barbados, serving as models for the sale slaves arriving alongside or after them. 7 Just as children in Hawaii, although going to school in English, developed a creole because a pidgin English was available as a model, plantation slaves, though quite capable of acquiring the lexifier itself— as the mainland Spanish colonies prove—could have developed a creole variety because an English pidgin variety was available as a model. Just as Hawaiian kids thought of themselves as developing a ‘kind of English’, plantation slaves would have thought of themselves as developing a ‘kind of English’, not constrained by the taxonomic boundaries around our modern conception of ‘English’ (note that today, Jamaican patois speakers are much less likely than creolists to consider their in-group speech a ‘different language’ than English). Some might object that pidgin-speaking castle slaves cannot have been numerous enough to have had such an impact upon colony-wide speech habits. Yet small numbers of people, with sufficient prestige, can have impact far beyond what their numbers would lead us to expect: At the turn of the 1900s in Papua New Guinea, the British stationed a squadron of just over 400 Papuan constables who spoke the pidginized Motu Hiri Motu in inland villages, with at most just two constables per village (Dutton 1985:72). Yet because of the prestige of these mere 400 constables, 150,000 people spoke Hiri Motu in 1971 and more do now (Dutton 1985:3). The idea that creoles formed on plantations as the European language receded from most slaves’ everyday contact has great intuitive appeal, and is in fact an almost inevitable interpretation based on the fact that, say, Jamaican patois, Gullah, Haitian Creole, Sranan, Cape Verdean, and Louisiana French Creole are all spoken by descendants of slaves in former plantation colonies. However, this genesis scenario, thoroughly reasonable in itself, was developed in neglect of mainland Spanish America, former plantation colonies where no creoles are spoken. It is natural to suppose that explanations for this discrepancy will turn up which leave the limited access idea intact, but as I have shown, if such an explanation is found, it is vastly unlikely to involve the tools and approaches currently at creolists’ disposal. The Afrogenesis Hypothesis is motivated not by a mere predilection for monogenetic scenarios, nor by a natural bent towards revisiting past approaches rather than participating in current ones. It is motivated purely by a 7

Hancock’s (1969, 1986) position has been that sale slaves learned the pidgin from castle slaves tending to them before shipment, and then brought the pidgin to the New World; he also reconstructs this as a regular occurrence, such that slaves would have been bringing pidgin English to all of the colonies of the New World over centuries’ time. This conception has been so well known for so long that there is a tendency for readers to assume that I am basically echoing it. In this light, I must emphasize that I suppose that castle slaves brought the pidgin across the Atlantic, and that they did this only once or twice during the foundation of Barbados, with AEC then disseminating from there.


125

desire to resolve the fascinating problem of the absence of Spanish creoles in almost all of mainland Spanish America in a falsifiable rather than ad hoc fashion. The facts suggest that this is one more problem whose solution may center upon the importance of pre-existing contact languages in determining the language choice of multiethnic groups seeking a lingua franca. 6. The Portuguese creoles A final area for which an acknowledgment of the role of pre-existing pidgins might shed light is the Portuguese creoles of Africa. Cape Verdean and the Gulf of Guinea dialect complex are generally thought of as ‘plantation creoles’ because they are spoken in former plantation colonies. However, we must recall that canonical plantation conditions failed to produce creoles in all Spanish colonies except in two cases where a pre-existing pidgin/creole is known to have been widely spoken in the early colony. In this light, it may be significant that Portuguese pidgin is known to have been an integral part of the genesis contexts of Cape Verdean as well as the Gulf of Guinea creoles. The prevalence of a Portuguese pidgin at the site of present-day GuineaBissau and its environs is uncontroversial (e.g., Boxer 1963:9-12). It is presently unknown whether Portuguese pidgin was incidental or crucial in the birth of the Guinea-Bissau creole. If it was crucial, then by implication we must trace the genesis of Cape Verdean ultimately to this pidgin, since Cape Verdean and the Guinea-Bissau creole are variants of the same language and thus must share a common source. On the other hand, some consider Cape Verdean to have been born first and later transported to present-day Guinea-Bissau (e.g., Quint 1998). However, Cape Verde was supplied with slaves from Guinea-Bissau, and the two regions were in fact administrated as a single colony until 1879. As such, it is unlikely that Portuguese pidgin played no role in the sociolinguistic context of slaves in Cape Verde. This is especially the case given that plantations were generally small on Cape Verde due to poor soil, and its main role was to supply passing ships with food. Much interaction between the slaves and passing sailors would likely have been conducted in Portuguese pidgin. Portuguese pidgin was also used in the slave trade conducted at São Tomé (although it is unlikely that this was the same pidgin used further north; Upper Guinea and Gulf of Guinea Portuguese creoles do not appear to trace to a single ancestor [Ferraz 1987]). Once again, it is possible that this trade pidgin had no effect on the contact languages developing on the sugar plantations. However, it is also possible that there was indeed influence, and this is suggested by correspondences between Portuguese pidgin recorded from the nearby Slave Coast and modern São Tomense, such as the idiosyncratic etymological source of copula sa, a combination of ser and estar (the author thanks Mikael Parkvall for this observation).

126


In both of these cases, my suggestion is not that the creole was necessarily a direct development of the pidgins. The available citations of Portuguese pidgin almost never correspond in any meaningful way to the modern creoles in the way that, for example, pidgin varieties of Tok Pisin were clearly the direct precursors to the modern creole. I suggest rather that the presence of reduced Portuguese ‘in the air’ in Cape Verde and São Tomé were important in leading slaves to develop a creolized version of Portuguese as their lingua franca. Of course, this speculation would be of questionable value if there were no explicit nécessité, as Chaudenson puts it (1992:44), for the hypothesis. Once again, I suggest this as a way of grappling with the anomaly of mainland Spanish colonies. Specifically, we find slaves having learned the full superstrate only in the contexts where (a) there is no documentation of early slave arrivals speaking a pidginized version of the superstrate (or a close relative) as there is in Curaçao or Cartagena, and where (b) linguistic and sociohistorical reconstruction do not suggest the existence of such pidgin ancestors, as they do in English, French, and Portuguese colonies. 7. Implications The implication of this paper is not meant to be that restricted contact with a dominant language plays no part at all in creole genesis, but rather that this aspect has come to play a more important part in our creole genesis scenarios than the actual distribution and histories of today’s creoles actually indicate. If the limited access conception had truly been as central to the birth of these languages as the literature suggests, then we would find that (a) every creole— or at least most creoles—had developed in situations in which access to the lexifier had been restricted, and (b) that in every situation—or at least almost every situation—in which current theory presumes that access to a dominant language would be limited, such as plantations and mines worked by massive African crews, the language had been creolized. This, however, is not what we find. Instead, on one hand we find many situations where people developed creoles even with rich access to the lexifier language, such as in Hawaii. Additional examples include Curaçao, where Papiamentu developed despite most slaves working in small groups in close proximity to whites, and Suriname, where Sranan developed among slaves working on small farms alongside equal numbers of English-speaking whites (McWhorter 1996, 2000). We find, on the other hand, many situations where people learned the dominant language fully in situations where current theory tells us that they could not have, such as slaves under the Spanish in Mexico and most South American colonies. The actual distribution of pidgins and creoles around the world suggests, then, that the drastic reduction and reconstitution associated with these lan-


127

guages occurred most often where learners had neither need nor inclination to fully acquire the dominant language. This seems an unremarkable statement, but in fact is not: Contrary to appearances, plantation colonies in themselves apparently were not situations where learners ‘had neither need nor inclination to fully acquire the dominant language’, at least not past the first generation. This is what the Spanish colonies, where today’s isolated Afro-Hispanics speak no creoles, show us. Rather, lack of need or inclination to acquire a superstrate fully manifested itself most clearly in the case of pidgins like Melanesian Pidgin English, which began as a work and trade vehicle between Australian Aboriginals and Englishmen (Baker 1993), and developed as a lingua franca between multiethnic labor crews in the South Pacific working under temporary contract (Keesing 1988), who thus had no inclination to acquire English as their new principal tongue. This pidgin has creolized in various locations where it has played a similar role; native languages thrive and English itself has been the primary language only for a small elite until recently (Papua New Guinea, Vanuatu, the Solomon Islands, Torres Strait). The pidginized Zulu Fanakalo arose similarly (Mesthrie 1989). Plantation slaves under the Spanish did not creolize Spanish so often that it begs explanation, and one way of doing this coherently is to read the independent evidence of West African pidgin ancestors for AEC and French creoles in the new way I have outlined. Namely, the introduction of these pidgins into English and French plantation colonies appears to have given the illusion that plantations themselves pidginized and creolized a dominant language. What in fact may have occurred is that the introduction of these West African pidgins—which had developed among castle slaves under conditions akin to the birth of Melanesian Pidgin English—into these colonies at their foundation provided models which made them, rather than the superstrate language, the model for an interethnic lingua franca. Thus just as immigrants’ children in Hawaii created HCE not because they had limited access to English but because they were surrounded by intimates whose English was at pidgin level, slaves in English and French colonies may have developed creoles not because learning English or French would have been impossible, but because they were offered a pidginzed version of the superstrate as a model. The Spanish colonies suggest that these slaves could well have learned actual English or French if such pidgins had not been introduced, and that under ordinary circumstances, in a context where slaves were forced to use a European language not as a mere temporary work jargon but as a primary language, the result was not pidginization and creolization, but relatively full acquisition. The introduction of West African pidgins was, in this sense, an extraordinary circumstance, and was one more demonstration of the central thesis of this

128


paper—that pre-existing contact languages played a crucial role in the establishment of creoles around the world. It is natural to question why slaves would ‘settle for’ or ‘stop at’ the creole if they were capable of acquiring the superstrate. Here is where Hawaii is again useful: as a demonstration not of the limited access model, which we have seen is inapplicable to it, but as a demonstration of the powerful impact that a pidgin register can have upon in-group vernaculars even when the superstrate language is readily available. This question also shows us the subtle effects of prescriptivist thinking even upon linguists. Despite the well-known overt denigration of creoles even by their speakers, on the covert level, creoles are perceived not as steps on the way to a standard, but as vehicles of group identity in their own right. Note that even today, Jamaican patois speakers are often perplexed or even taken aback to be told that they are not speaking ‘English’; Huber (2000:283) documents a similar tendency among speakers of early Sierra Leone Krio. This would have been magnified among slaves, for whom the evaluative norms of larger society would have played a distinctly more marginal role than among today’s postcolonial creolophones. Ascribing a central role to the limited access scenario in creole genesis leaves open a number of questions unlikely to be resolved within its framework: why Hawaiian children developed a creole, why there are English creole continua so consistently but almost never French ones, why the English creole mesolects are so similar to one another, why Spanish creoles almost never appeared on Spanish plantations, why the Spanish creoles that do exist emerged where they did. Allowing the seemingly mundane evidence of preexisting contact languages a greater role in our reconstructions presents a way of transforming these problems into predictions, which is, of course, the principal goal of scientific inquiry.

References Alleyne, Mervyn C. 1971. “Acculturation and the cultural matrix of creolization”. Pidginization and Creolization of Languages, Hymes (ed) 1971 (Cambridge: Cambridge University Press). 169-186. ----------. 1980. Comparative Afro-American. Ann Arbor, MI: Karoma. Baker, Philip. 1987. “Combien y a-t-il eu de genèses créoles à base lexicale française?” Études Créoles 10. 60-76. ----------. 1990. Column: “Off Target?” Journal of Pidgin and Creole Languages 5. 107-19. ----------. 1993. “Australian influence on Melanesian Pidgin English”. Te Reo 36. 3-67. ----------. (ed). 1995. From contact to creole and beyond. London: University of Westminster Press.


129

----------. 1996. “Pidginization, creolization and français approximatif (Review)”. Journal of Pidgin and Creole Languages 11. 95-120. ----------. 1998. “Evidence from old texts on the origin and diffusion of Kittitian and other Atlantic English Creoles”. St. Kitts and the Atlantic creoles: The Text of Samuel Augustus Matthews in perspective, Baker & Bruyn (eds.) 1998 (London: University of Westminster Press). 289-314. ---------- & Chris Corne. Isle de France creole: Affinities and origins. Ann Arbor, MI: Karoma. ---------- & Chris Corne. 1986. “Universals, substrata and the Indian Ocean Creoles”. Substrata versus universals in creole genesis, Muysken & Smith (eds) 1986 (Amsterdam: John Benjamins). 163-183. Bartens, Angela. 2000. “Notes on componential diffusion in the genesis of the Kabuverdianu cluster”. Language change and language contact in pidgins and creoles, McWhorter (ed) 2000 (Amsterdam: John Benjamins). 35-61. Bickerton, Derek. 1973. “The nature of a creole continuum”. Language 49. 640-649. ----------. 1975. Dynamics of a creole continuum. Cambridge: Cambridge University Press. ----------. 1981. Roots of language. Ann Arbor, MI: Karoma. ----------. 1984. “The language Bioprogram Hypothesis”. The Behavioral and Brain Sciences 7. 173-188. ----------. 1988. “Creole languages and the Bioprogram”. Linguistics: The Cambridge survey: Vol. II. Linguistic theory—Extensions and implications, Newmeyer (ed) 1988 (Cambridge: Cambridge University Press). 268-282. ----------. 1996. “The origins of variation in Guyanese”. Towards a social science of language: Papers in honor of William Labov—Vol. I. Variation and change in language and society, Guy, Feagin, Schiffrin, & Baugh (eds) 1996 (Amsterdam: John Benjamins). 311-327. ----------. 1998. “A sociohistoric examination of Afrogenesis”. Journal of Pidgin and Creole Languages 13. 63-92. ---------- & William Wilson. 1987. “Pidgin Hawaiian”. Pidgin and creole languages, Gilbert (ed) 1987 (Honolulu: University of Hawaii Press). 6176. Birmingham, John C., Jr. 1976. “Papiamentu’s West African cousins”. 1975 Colloquium of Hispanic Linguistics, Aid, Resnick, & Saciuk (eds) 1976 (Washington, DC: Georgetown University Press). 19-25. Boretzky, Norbert. 1983. Kreolsprachen, Substrate und Sprachwandel. Wiesbaden: Harrassowitz. Boxer, C.R. 1963. Race relations in the Portuguese colonial empire, 14151825. Oxford: Clarendon Press. Bruyn, Adrienne. 1995. Grammaticalization in Creoles: The development of determiners and relative clauses in Sranan. Amsterdam: IFOTT.

130


Cabrera, Lydia. 1954. El monte. Paris: Gallimard. Cassidy, Frederic.G. 1983. “Sources of the African element in Gullah”. Studies in Caribbean language, Carrington, Craig, & Dandare (eds) 1983 (St. Augustine, Trinidad: Society for Caribbean Linguistics). 76-81. Cellier, Pierre. 1985. Comparaison syntactique du créole réunionnais et du français. Saint Denis: Université de la Réunion. Chaudenson, Robert. 1981. Textes anciens en créole réunionnais et mauricien: Comparaison et essai d’analyse. Hamburg: Buske. ----------. 1992. Des îles, des hommes, des langues. Paris: L’Harmattan. ----------. 2000. “Créolisation du français et francisation du créole: Les cas de Saint-Barthélemy et de la Réunion”. Degrees of restructuring in creole languages, Neumann-Holzschuh & Schneider (eds) 2000 (Amsterdam: John Benjamins). 361-381. Corne, Chris. 1995. “Métchif, Mauritian and more: The “creolisation” of French”. The 1995 Samuel Weiner lecture delivered before the University of Manitoba. Winnepeg: Voices of Rupert’s Land. Dutton, Tom. 1985. Police Motu: Iena sivarai. Port Moresby: University of Papua New Guinea Press. Ferraz, Luis Ivens. 1987. “Portuguese Creoles of West Africa and Asia”. Pidgin and creole languages, Gilbert (ed) 1987 (Honolulu: University of Hawaii Press). 337-360. Geraty, Virginia. 1990. Porgy: A Gullah version. Charleston, South Carolina: Wyrick & Co. Goodman, Morris. F. 1985. [Review of Bickerton 1981]. International Journal of American Linguistics 51. 109-37. Granda, German de. 1978. Estudios lingüisticos hispánicos, afrohispánicos, y criollos. Madrid: Editorial Gredos. Grant, Anthony. 1996. “Zamboangueño, Papiamentu, and ‘Spanish-based Creoles’”. Unpublished manuscript, York University. Hancock, I. 1969. “A provisional comparison of the English-based Atlantic Creoles”. African Language Review 8. 7-72. ----------. 1986. “The Domestic Hypothesis, diffusion and componentiality: An account of Atlantic Anglophone creole origins”. Substrata versus universals in creole genesis, Muysken & Smith (eds) 1986 (Amsterdam: John Benjamins). 71-102. Handler, Jerome S. 1974. The unappropriated people: Freedmen in the slave society of Barbados. Baltimore: Johns Hopkins University Press. Ho, Mian-Lian & John T. Platt. 1993. Dynamics of a contact continuum: Singaporean English. Cambridge: Cambridge University Press. Holm, John. 1984. “Variability of the copula in Black English and its creole kin”. American Speech 59. 291-309.


131

----------. 1986. “Substrate diffusion”. Substrata versus universals in creole genesis, Muysken & Smith (eds) 1986 (Amsterdam: John Benjamins). 259278. Huber, Magnus. 2000. “Restructuring in vitro? Evidence from Early Krio”. Degrees of restructuring in creole languages, Neumann-Holzschuh & Schneider (eds) 2000 (Amsterdam: John Benjamins). 275-307. Keesing, R. 1988. Melanesian pidgin and the Oceanic substrate. Stanford: Stanford University Press. Kihm, Alain. 1980. Aspects d’une syntaxe historique: Études sur le créole portugais de Guiné-Bissau. Unpublished doctoral dissertation, Université de Paris III, Sorbonne Nouvelle. Lefebvre, Claire. 1986. The role of relexification in creole genesis revisited: The case of Haitian Creole. Substrata versus universals in creole genesis, Muysken & Smith (eds) 1986 (Amsterdam: John Benjamins). 279-301. Lefebvre, Claire. 1998. Creole genesis and the acquisition of grammar. Cambridge: Cambridge University Press. LePage, Robert B. & David De Camp. 1960. Jamaican Creole. London: MacMillan & Co. Lipski, John. 1986. “Sobre lingüística afroecuatoriana: El Valle del Chota”. Anuario de Lingüística Hispanica (Valladolid) 2. 153-176. McWhorter, John H. 1994. [Rejoinder to Derek Bickerton’s Reply to McWhorter 1992 “Substratal influence in Saramaccan serial verb constructions”], Journal of Pidgin and Creole Languages 9. 79-93. ----------. 1995a. “Sisters under the skin: A case for genetic relationship between the Atlantic English-based creoles”. Journal of Pidgin and Creole Languages 10. 289-333. ----------. 1995b. “Looking into the void: Zero copula in the creole mesolect”. American Speech 70. 339-360. ----------. 1995c. “The scarcity of Spanish-based Creoles explained”. Language in Society 24. 213-244. ----------. 1996. “It happened at Cormantin: Tracing the birthplace of the Atlantic English-based creoles”. Journal of Pidgin and Creole Languages 10. 289-333. ----------. 1999. “A Creole by any other name: Streamlining the terminology in creole studies”. Spreading the word: The issue of diffusion among the Atlantic Creoles, Huber & Parkvall (eds) 1999 (London: University of Westminster Press). 5-28. ----------. 2000. The missing Spanish Creoles: Recovering the birth of plantation contact languages. Berkeley: University of California Press. Megenney, William W. 1985. “Africa en Venezuela: Su herencia lingüística y cultura literaria”. Montalbán 15. 3-56.

132


Mesthrie, Rajend. 1989. “The origins of Fanagalo”. Journal of Pidgin and Creole Languages 4. 211-240. Mühlhäusler, Peter. 1985. “The scientific study of Tok Pisin: Language planning and the Tok Pisin lexicon”. Handbook of Tok Pisin (New Guinea Pidgin), Wurm & Mühlhäusler (eds) 1985 (Canberra: Australian National University). 595-664. Parkvall, Mikael. 1995a. A dual approach to Creole genesis. Unpublished master’s thesis, University of Stockholm. ----------. 1995b. “The role of St. Kitts in a new scenario of French Creole genesis”. From contact to creole and beyond, Baker (ed) 1995 (London: University of Westminster Press). 41-62. Platt, John T. 1975. “The Singapore English speech continuum and its basilect ‘Singlish’ as a ‘creoloid’”. Anthropological Linguistics 17. 363-374. Quint, Nicolas. 1998. Le créole de l’île de Santiago (République du Cap-Vert). Unpublished doctoral dissertation, University of Paris III Sorbonne Nouvelle. Reinecke, J.E., S.M. Tsuzaki, D. DeCamp, I.F. Hancock & R.E. Wood (eds). 1975. A bibliography of pidgin and creole languages. Honolulu: University Press of Hawaii. Restrepo, Vicente. 1886. A study of the gold and silver mines of Colombia. (Trans. by C.W. Fisher.) New York: Colombian Consulate. Rickford, John R. 1992. “The creole residue in Barbados”. Old English and new: Essays in language and linguistics in honor of Frederic G. Cassidy, Doane, Hall, & Ringler (eds) 1992 (New York: Garland Publishing, Inc). 183-201. ---------- & Jerome S. Handler. 1994. “Textual evidence on the nature of early Barbadian Speech, 1676-1835”. Journal of Pidgin and Creole languages 9. 221-255. ---------- & Christine Théberge. 1996. “Preterit HAD + V-ed in the narratives of African American preadolescents”. American Speech 71. 227-252. Roberts, Julianne. 1995. “Pidgin Hawaiian: A sociohistorical study”. Journal of Pidgin and Creole Languages 10. 1-56. ----------. 1998. “The genesis of Hawaiian Creole and diffusion”. Language 74. 1-39. ----------. 2000. “Nativization and the genesis of Hawaiian Creole”. Language change and language contact in pidgins and creoles, McWhorter (ed) 2000 (Amsterdam: John Benjamins). 257-300. Rout, Leslie B. 1976. The African experience in Spanish America: 1502 to the present day. Cambridge: Cambridge University Press. Roy, John D. 1986. “The structure of tense and aspect in Barbadian English Creole”. Focus on the Caribbean (varieties of English around the world,


133

Vol. G8), Görlach, & Holm (eds) 1986 (Amsterdam: John Benjamins). 141-156. Schwegler, Armin. 1991. “El habla cotidiana del Chocó (Colombia)”. América Negra 2. 85-119. ----------. 1993. “Rasgos (afro-) portugueses en el criollo del Palenque de San Basilio (Colombia)”. Homenaje a José Perez Vidal, Alayón (ed) 1993 (La Laguna, Tenerife: Litografía A. Romero S. A.). 667-696. ----------. 1996. “La doble negación dominicana y la génesis del español caribeño”. Hispanic Linguistics 8. 246-315. ----------. 1998. “El palenquero”. América negra: Panorámica actual de los estudios lingüísticos sobre variedades hispanas, portuguesas y criollas, Perl & Schwegler (eds) 1998 (Frankfurt: Vervuert). 219-291. ----------. 2000. “The myth of decreolization: The anomalous case of Palenquero”. Degrees of restructuring in creole languages, Neumann-Holzschuh & Schneider (eds) 2000 (Amsterdam: John Benjamins). 409-436. ----------. (forthcoming) El vocabulario africano de Palenque (Colombia). Segunda parte: Compendio de palabras (con etimologías). Sharp, William Frederick. 1976. Slavery on the Spanish frontier. Norman, OK: University of Oklahoma Press. Singler, John V. 1981. An introduction to Liberian English. East Lansing, MI: Michigan State University, African Studies Center/Peace Corps. ----------. 1995. “The demographics of creole genesis in the Caribbean: A comparison of Martinique and Haiti”. The early stages of creolization, Arends (ed) 1995 (Amsterdam: John Benjamins). 203-232. West, Robert C. 1957. The pacific lowlands of Colombia. Baton Rouge, LA: Louisiana State University Press. Winford, Donald. 1993. Predication in Caribbean English Creoles. Amsterdam: John Benjamins.

SECTION TWO: SOCIETY

CREOLES, CAPITALISM, AND COLONIALISM

Derek Bickerton University of Hawai‘i, Emeritus

ABSTRACT: Arguing from the Hawaiian case, which is the only site where pidginization and creolization processes in a plantation setting can be reconstructed empirically, Bickerton locates the origins of plantation creoles in the economy of early colonial capitalism, and in particular the homestead-toplantation transition. During this period of expansion, a prior original contact language (OCL) established between the earliest Europeans and nonEuropeans is swamped by massive importations of new slave labor, leading to a partial breakdown in language transmission. Examining three demographic conditions involving proportions of old hands to new hands among a hypothetical servile population, Bickerton argues that in the expansion phase, acquisition of the OCL is impossible, and pidginization inevitably results. He shows how one can use this model to predict the outcomes of sociolinguistic interactions in a developing colony. 1. Introduction The linguistic effects of the market forces driving creole formation have been under-appreciated. All plantation economies required a period of rapid expansion in order to recoup the considerable initial capital investment they involved. This entailed rapid importation of large quantities of labor as soon as infrastructure was available, and this in turn entailed a high proportion of linguistically naive workers during and immediately after the expansion stage. The large imbalance between these and the ‘old hands’ who might have had reasonable access to the superstrate ensured that, regardless of whatever linguistic developments took place in the earliest years, a structureless jargon must have emerged in all but the less successful economies and served as input for subsequent creolization. 2. Two views on the sociolinguistics of creolization Although for many years creolistics was treated as a subtopic of sociolinguistics, I noted over a quarter-century ago (Bickerton 1976) that hardly any studies of creole languages were in fact sociolinguistic studies. Around the same time, Alleyne (1980:5) wrote that:

138


since we agree that pidgins and creoles demonstrate so forcefully the influence of social context on language change, language structure and language usage, we should be concerned, if we wish to reconstruct the early language and language processes of pidgin and creole speakers, with the reconstruction of the sociolinguistic situation at the time of their formation, with the structure of the communicative network and the communicative needs of different sectors of the population....

But despite this call, and despite the recent growing consensus that creoles have no distinctive typology and can only be defined as a class on sociohistorical grounds (see Parkvall 2001 and references therein), there have been few studies of the type that Alleyne called for. In the resultant vacuum, some remarkable claims have been made; for instance, that “there are no fundamental differences between Pidgins and Creoles: that contact languages develop along the same lines and at similar speed” (Baker 1995:13) and that “creole vernaculars...emerged in contact settings where the development of pidgins would be inconsistent with the received doctrine that they are reduced systems for limited and specialized communication” (Mufwene 2001:9). It is not easy to reconcile these two claims with one another, and I shall not even try, since each in its own way flies in the face of what we should have learned from sociolinguistics, as well as what ought to have been learned from the only site (Hawaii) where pidginization and creolization processes in a plantation setting can be reconstructed from empirical evidence.1 In what follows, I shall briefly sketch (leaving to others the filling out of the details) the socioeconomic dynamics of creole development and show that these dynamics make a break in linguistic transmission almost inevitable. According to Mufwene (2001), the only socioeconomic change in creole societies that had any linguistic significance was the shift from a homestead phase (societe d’habitation, see Chaudenson 1992) to a plantation phase. During the homestead phase, numbers of Europeans and Africans differed little (in some places, such as Barbados, Europeans were in a majority for at least the first quarter-century of the colony’s life), and social differences were less acute than they later became. Consequently, African slaves had a good shot at acquiring at least a reasonable second language version of the dominant language. According to Mufwene, such versions would later be subjected to a gradual dilution (‘basilectalization’) of the contact language, as “the newly1

Baker (1995:13) claims that “the circumstances in Hawaii were so very different from those which existed on slave plantations...that we should not expect the linguistic consequences to have been the same then.” The only difference Baker mentions is that “formal education was provided for the children of plantation workers.” He does not, however, explain how formal education would have triggered creolization of an unstructured pidgin, and indeed, although many writers have dismissed the Hawaii evidence (usually for ideological rather than empirical reasons) not one of these writers has produced a reasoned explanation of why any differences between Hawaii and other plantation societies (differences easily exaggerated, I might add) should have led to significant linguistic differences.

CREOLES, CAPITALISM AND COLONIALISM

139

arrived Africans learned the colonial vernacular mostly from the creole and ‘seasoned’ slaves” (Mufwene 2001:51). Mufwene admits that the transition from homestead to plantation economy did not always take place “at the same speed” (Mufwene 2001:38), but fails to fully understand the significance of this fact. 3. Homesteading, monoculture, and the original contact language In order to understand that significance, we must understand the economic forces that drove colonialism in the Caribbean and subsequently in similar tropical areas. To Mufwene, Singler (1996), and others, the homestead– plantation transition is merely a fact of life, like climate change. In reality, polycultural homesteaders were driven off their land by monocultural planters, 2 not because homesteading was economically non-viable, 3 but because sugar-planting was immensely more profitable, at least for those who had, or could, amass the considerable capital outlay necessary.4 However, in order to profit from an investment in sugar, planters required a quick return on their profits. This meant that, as soon as the necessary infrastructure was in place, they had to maximize the amount of land under cultivation and consequently maximize, with the greatest speed possible, the number of slaves on each plantation (sugar being a highly labor-intensive crop). These conditions applied, to a greater or lesser extent (depending on variables such as area and richness of arable land, sugar and slave prices, current availability of new slaves, climate, distance from markets, and so on) wherever sugar was raised, and consequently in a large majority of creole-speaking colonies. All of these colonies experienced rapid rises in population over very short periods. In Tobago, for instance, the number of slaves tripled in nineteen years (from 1771 to 1790 it went from 4,716 to 14,170, according to Williams 1962:59). But this was a modest increase compared with that of Suriname, where after 27 years of settlement the number of slaves had increased by only 800 (200 to 1,000); twenty years later there were almost ten times as many (Migge 1998:221). But even these figures do not adequately represent the 2

The terms ‘polycultural’ and ‘monocultural’ are here used, of course, in their agricultural rather than their ‘culture wars’ senses. 3 “We found indigo planted, and so well ordered, as it sold in London at very good rates; and their cotton wool and fustick wood proved very good and staple commodities” (Ligon 1657). 4 Ligon (1657) describes an estate on Barbados: “500 acres of land, with a fair dwelling house, an ingenio placed in a room of 400 foot square, a boiling house, filling-room, cisterns and still house, with a carding house of 100 foot long and 40 foot broad, with stables, smith’s forge, and rooms to lay provisions of corn and bonavist, houses for Negroes and Indian slaves, with 96 Negroes and three Indian women with their children, 28 Christians, 45 cattle for work, 8 milch cows, a dozen horses and mares, 6 asinigoes.” To compete with establishments like this obviously required a capital outlay far exceeding the capacities of most European settlers and, equally obviously, those who could afford it expected a substantial and reasonably rapid return for their expenditure.

140


process, for during a similar period (1675 to 1699) some 15,000 slaves were brought into Suriname (Bickerton 1994:71), indicating an extremely high level of negative population growth. Let us consider the inevitable consequences of these very rapid population increases. It is assumed almost universally (and certainly by Mufwene 2001, Smith 1987, and Singler 1996 among others) that once some form of language had been established between the earliest European and non-European (I shall refer to this as ‘the original contact language’ or OCL throughout), 5 it would be passed on, with relatively little modification on the part of its speakers, to subsequent immigrants. Naturally, the OCL would tend to degrade progressively due to imperfect learning on the part of those subsequent immigrants, resulting in the gradual ‘basilectalization’ process assumed by Mufwene. However, there would not be at any time any kind of catastrophic ‘breakdown’ of language transmission, nor any consequent pidginization, contrary to what had been assumed in the universalist model. However this assumption is just that, an assumption. There was never any empirical evidence that supported it; indeed there was no evidence at all until quite recently when the data on the pidgin-creole cycle in Hawaii became available. Now, the evidence of this data goes strongly against the assumption that pidginization cannot occur once an OCL has been established. 4. An abstract model of linguistic interaction in colonial contexts I would like to dispose, here and now, of any suggestion that the Hawaii evidence is invalid because it deals with a ‘different’ situation to that of the majority of tropical colonies, where slaves rather than indentured laborers were involved. Unfortunately, most studies of creole languages are still at a taxonomic, pre-scientific level, where phenomena are merely described, without any distinction between scientifically significant and scientifically insignificant facts. One example of this kind of thinking is Mufwene’s remark that “there is no compelling reason for downplaying differences [between creole languages; DB] in favor of similarities...both are all equally significant” ([sic]; 2001:5152). Science is, of course, not in the least interested in either differences or similarities per se; such things are of interest only to descriptive taxonomists. Science is interested in the unexpected versus the expected. If science had not concentrated on the unexpected, we would still have the world-view of the early Greeks. It is precisely through examining what common sense, or the 5

For present purposes, it is immaterial whether the OCL was a native or a non-native language. The overall consequences of massive immigration would have been the same, in either case, except that if the OCL had nativized, its native speakers would be likelier to maintain it more or less unchanged among themselves and their children (though their inability to transmit it to new hands would not itself have been affected). Thus were upper mesolects, presumably, born.


141

prevailing theory, does not lead us to expect that science has progressed. And of course, what is expected may be similarity, or it may be difference, according to the facts of the situation. In the case of creoles, with a half-dozen or more superstrates and hundreds if not thousands of substrates, both common sense and most linguistic theories would predict difference, would predict no more resemblance between creoles than might be found in any set of randomly-chosen languages. The fact that they resemble one another to the (admittedly imperfect) extent that they do is the surprising, unexpected, unpredictable fact, and therefore the fact that any scientific approach should focus on. This is the fact that the universalist approach, accordingly, always has focused on. I shall therefore argue here, first, that the proximate cause of resemblances among plantation creoles must be some factor which (unlike the sets of ancestral languages) creoles all share, in this case, the economic processes of early colonial capitalism. I shall then argue that since social forces do not and of their nature cannot operate directly upon language, 6 this proximate cause merely made inevitable a situation—the rapid and massive outnumbering of any speakers of a prior, stabilized contact language—that in turn made inevitable a partial breakdown in language transmission. I shall begin by considering what the significant sociolinguistic variables are in the development of creole colonies, first dealing with those variables in general terms, and then seeing what their practical results were in Hawaii (the only place where we have detailed empirical evidence of those results). Any attempt at science involves a degree of idealization. When Newton was faced with the problem of predicting the tides, he did not get into a rowboat with a plumb-bob and measure the heights and depths of high and low tides in every corner of the world. He assumed a totally abstract and unrealistic model, an ocean of uniform depth covering the entire world, worked out mathematically what the tides would be for that, and then worked from that to predict real-world situations. For the present discussion, we need a similarly abstract model of sociolinguistic interactions. From such a model we must exclude all irrelevant variables. It is completely irrelevant whether the speakers in such a model are slave or free, African, European, or Asian, seventeenth, eighteenth, or nineteenth century, and whether the main lexifier for the OCL is a European language (as it was throughout the Caribbean, except perhaps for Berbice) or a non-European language (as it was in Hawaii). For that matter, it is irrelevant whether the locus is a creole colony or an established country with its own historic language. The only relevant variables are the relative numbers 6

Any with lingering longings to believe that social forces can and do operate in this way should turn to the kind of reductio ad absurdum found in some Marxist analyses of language— for example, Beaken’s (1996:165-167) claim that the English system of modal verbs arose as a direct response to social turmoil between 1350 and 1500.


142

of ‘old hands’ (speakers who control the OCL, be it a homestead contact language or the local indigenous language) and ‘new hands’ (recently arrived speakers who do not control that language). Thus there cannot be any legitimate grounds for ruling out the Hawaii evidence in the case at issue. Now let us consider the sociolinguistic forces involved in the cycle of capitalist development. There are three major forces, of which one is widely recognized while the other two are almost entirely ignored. The first of these forces is the need for newcomers to a region to acquire the language most widely used in that region. This force drives a belief in the following suppositions: • • •

Old hands will use the OCL without simplification to new hands. New hands will accept the OCL as their target and acquire it. At worst, only a ‘basilectalized’ version of the OCL will result.

The second force is the universal tendency for people to adapt their speech to that of their interlocutors (Ferguson 1968), which is particularly salient where the latter are acquiring a foreign language (Clements 2003), especially if asymmetrical social relations contribute to the difficulty of the task (Hinnenkamp 1984). The third is the necessity to communicate, which can certainly override the function of modeling the OCL for newcomers’ benefit. Together, these two forces may (especially in times of rapid immigration) overcome the first force. Consider why old hands would talk to new hands. They would have two primary purposes, one official (teaching the new hands their jobs) and one unofficial (explaining to the new hands how to survive in an extremely hostile environment). The quicker these purposes were achieved, the better it would be for all concerned. Bear in mind that to the newcomer, the OCL would initially be as opaque as any other language. If those purposes, therefore, could not be immediately achieved by use of the OCL, the old hands would pidginize their speech in order to make themselves understood. The relative strength of the three forces concerned will be determined by fluctuations in the ratio of old-hand/new-hand interactions to new-hand/newhand interactions. Old-hand/old-hand interactions are largely irrelevant to any process of acquisition by the new hands (except that the more frequent such interactions are, the likelier it is that the OCL is maintained as the primary vehicle among old hands); merely overhearing such conversations might enable new hands to pick up individual lexical items, but not the syntactic structures in which those items were embedded. Thus, the only opportunities new hands have of acquiring that language lie in new-hand/old-hand interactions. If these equal or outnumber new-hand/new-hand interactions, then new hands have a good chance of acquiring that language in something not too remote from its original form. If, on the other hand, new-hand/new-hand interactions


143

outnumber new-hand/old-hand interactions by a significant margin, then new hands must talk to new hands as best they can, by adopting the most primitive pidgin strategies (see Bickerton 1999 for a re-evaluation of the conventional wisdom on pidgins). Since most new-hand linguistic experience will consist of understanding and communicating with their fellow new hands, the capacity to acquire any pre-existing contact language (and even the utility of such a language, if it comes to that) will be limited, at best. In the homestead period, just as in most immigrant situations, old hands outnumbered new hands by a substantial margin. At one extreme, eight years passed between the initial settlement of Barbados and the first document that makes reference to a slave (Campbell 1993:95; there were six Africans on the first ship to land in Barbados but it is unclear whether they were slaves and not known whether they remained there); by that time, the white population of the island had swelled to several thousands. However, as noted above, the second phase of the plantation period soon caused new hands to outnumber old hands, sometimes (as in Suriname) to an overwhelming degree. During the last two decades of the seventeenth century, a contingent of new hands larger than the 1679 total of old hands arrived in Suriname every two years. Let us take as the model for our population a group consisting of 30 individuals, containing varying proportions of old and new hands. 7 Table 1 shows the results of three possible conditions: 20 old hands to 10 new hands; 15 old hands to 15 new hands; and 10 old hands to 20 new hands. In every society, people talk more to people with whom they have more in common; clearly, arrivals fresh off the boat, plunged into a society in which all are strangers and required to adjust rapidly to these wholly novel circumstances, will have far more in common with one another and far more interests and concerns in common, than they will with persons fully adjusted to the society who in many cases will occupy supervisory positions. Under these circumstances, some degree of spontaneous segregation between old and new hands is hardly avoidable. In addition to conditions that apply in all societies, there are linguistic conditions that apply in plantation societies. Either the old hands will be of the same ethnicity as the new hands, or they will not. If they are of the same ethnicity, then conversations can be conducted in the speakers’ native language, and such conversations are of course irrelevant when it comes to the formation of novel languages. If they are of different ethnicities, then the conversations must utilize some kind of contact language. Any contact language already developed by old hands will be, initially, no more comprehensible to the new 7

Before objecting that such a population is unrealistically small, readers should note that the numbers themselves are totally irrelevant: the figure of 30 is assumed for convenience only. Since only the proportions involved are significant, the reasoning that follows from this model would apply if the figure were 3,000, or 30,000, or 300,000 (as well, of course, as any intermediate figure).

144


hands than any other foreign language. Conversely, when talking to one another, all new hands will confront the same type of problem and will solve it in the same way—by using a structureless, macaronic type of contact language, of which there is abundant evidence from Hawaii. Consequently, let us make a fairly conservative assumption, given the differences in experience and capacity between the two groups: that new hands will talk to new hands and old hands to old hands twice as often as new hands and old hands talk together. Every speaker will speak with every other speaker once (if that other speaker belongs to a different class) or twice (if that other speaker belongs to the same class). Table 1 compares the inputs that will be received by individuals in each group under each of the three conditions (OH→OH = inputs received by old hands from other old hands; OH→NH = inputs received by new hands from old hands; NH→OH = inputs received by old hands from new hands; NH→NH = inputs received by new hands from other new hands). Condition A corresponds to the earliest stages of some tropical colonies (e.g., Barbados, where old hands outnumber new hands by two to one). Under Condition A, new hands receive slightly more input from old hands than they do from other new hands, while old hands receive nearly four times as much input from each other as they do from new hands. Under such circumstances, new hands have a good shot at picking up the OCL, while old hands will have relatively little pressure to repidginize in the presence of new hands. Inputs Condition A (20 OH, 10 NH) Condition B (15 OH, 15 NH) Condition C (10 OH, 20 NH)

OH→OH 38 28 18

OH→NH 20 15 10

NH→OH 10 15 20

NH→NH 18 28 38

Table 1: Ratios of old-hand/new-hand inputs under three demographic conditions Condition B corresponds to the expansion phase of the plantation period in colonies where immigration was less rapid. Under Condition B, new hands receive twice as much input from one another as they do from old hands; old hands, however, still receive more input from each other than from new hands. Under such circumstances, old hands may struggle to maintain the OCL, but new hands, with more limited exposure to it and more reinforcement from the primitive speech of others in the same boat as themselves, will have a harder time acquiring the OCL and will do so more slowly, if at all. Condition C corresponds to the expansion phase of the plantation period in colonies with massive and rapid immigration. Under Condition C, new hands receive input from other new hands almost four times as frequently as they receive it from old hands, while old hands are beginning to receive more input from new hands than they do from one another. Under such circumstances,


145

pidginization is unavoidable: old hands must severely reduce the structure of their speech to make themselves understood, and their own OCL may begin to erode (as the Hawaiian of native Hawaiian speakers began to erode in the late nineteenth century). 8 Under these circumstances, acquisition of the OCL becomes impossible: It is simply no longer being modeled by the old hands, and the new hands have no other way of accessing it. It must be emphasized that Conditions A, B, and C represent points on a continuum rather than discrete states. How far along that continuum any given community traveled would be determined by the rate of immigration and the shifting ratio between old and new hands. If the approach described here is correct, it should be empirically testable by comparison of careful analyses of all available demographic data with some kind of objective measure of the ‘radicalness’ of the resultant creole. Other things being equal (and of course other things were not always equal and might have to be factored in; degree of continued contact with the major lexical source language is one obvious example), either of the two data sets should be predictive of the other. 5. A closer look at Condition C Condition C corresponds to a stage actually reached by many creoles, certainly those of Suriname, Haiti, Mauritius, Hawaii, and Jamaica; in some of them (Suriname in particular) the ratio of Condition C may even have been exceeded at times. This stage typically took place in the expansion phase of a colony, discussed above, when planters sought by a rapid increase in the labor force to recoup their original heavy investments and begin to reap some of the profits that were their only motive for those investments. Under Condition C, new-hand/new-hand interactions would have outnumbered old-hand/new-hand interactions by a factor of almost 4 to 1. In other words, a new hand was four times as likely to be driven to basic communicative strategies as to have opportunities to improve his or her knowledge and use of the OCL. New hands would hardly remain deaf and mute in one another’s presence until old hands could spare enough time from their manifold other duties to teach them a pre-existing vernacular. To the contrary, the new hands would have had to resort to a primitive pidgin in order to communicate at all. One of the very few writers to appreciate this point is Plag (1993:31), who pointed out that the new hands must have “continually re-pidginized Pre-Sranan.” But we also have to ask what old hands would have done under Condition C. 8

Extremely relevant are two citations from Roberts (1995a:18, 23). The first, dated 1854, states that “There has long prevailed, between natives and foreigners, a corrupted tongue, which the former use only in speaking to the latter, but never among themselves” (emphasis added). The qualification has been removed less than 20 years later (1873), even before the major influx of new hands: “Some Hawaiians know very well how to speak correctly but through associating with foreigners they act as though they hardly know the language.”

146


The implication (it is seldom made explicit, but see the quote from Mufwene later in this section) in almost all studies is that old hands continued to speak the OCL to the new hands in a more or less unmodified form. I know of no empirical grounds for this assumption, and many things that make it inherently unlikely. First, it was not the speakers’ native language, 9 so they had no true competence in it, no loyalty to it, and therefore none of the confidence in one’s own choice of words and structures that natively speaking an established language confers. Second, as far as the new hands were concerned, the OCL would have been no more intelligible than the undiluted dominant language, so if the old hands were going to make themselves understood, those old hands were simply going to have to re-pidginize 10 their speech. Third, the old hands had to make themselves immediately understood by the new hands, for at least two purposes that would brook no delay. First, an official purpose: They had to instruct the new hands in a series of agricultural tasks of which none of the new hands had any prior experience (note that if, as one reviewer suggested, mute demonstration played the most significant role in this, modeling of the OCL to new hands would be less, not more). Second, an unofficial purpose: They had to instruct the new hands how to behave, how to get by, and to survive in an extremely repressive society that was, again, nothing like anything the new hands had previously experienced. These were tasks that would not wait on the niceties of grammatical detail; they had to be done immediately if they were to have the right effect and to make life a little easier for old hands and new hands alike. Communicating in the absence of any common code, by whatever means works, is of course the essence of the pidginization process (Bickerton 1999). Thus the belief that new hands “would have communicated regularly with the old-timers and would have had plenty of opportunities to approximate their speech,” as Mufwene (1996:363) avers, is almost certainly incorrect. For under such circumstances the pidginization of one’s speech, even of one’s own native language, would constitute the most natural reaction. Everyone does it. You and I do it if we have to explain something to a recent Hmong immigrant. First-generation Hawaii Creole speakers used to do it routinely to their pidginspeaking parents while these were still alive. But if old hands in Condition C did it, this placed the OCL well outside the reach of the new hands, and made it inevitable that an early-stage pidgin would be received as input by their 9

I am assuming that for practical purposes, we can treat speakers of the OCL as nonEuropeans (except in Hawaii) since by the onset of Condition C, increasing social distance had heavily reduced, if not altogether eliminated, any European role in contact-language formation. 10 Many scholars believe that there never was a pidgin phase in the formation of OCLs, rather a collection of interlanguages. In those colonies where Condition A obtained, I am inclined to think that they are right. For their benefit, I would therefore be quite happy with the word ‘pidginize’ rather than ‘repidginize’ in this context.


147

children. Once one pays attention to population ratios and to the economic forces that made those ratios inevitable, it becomes apparent that the scenario envisaged by the universal approach, a generation of children receiving chaotic and unstructured input, is equally inevitable. 6. Evidence from Hawaii However, most creoles have left no history of their development. The only place we can seek empirical confirmation for what has been proposed here is Hawaii, and the facts from Hawaii are fully consonant with these proposals. Hawaii experienced the same sequence of economic stages as did other tropical areas. Sugar began to be grown early in the nineteenth century, but the industry remained small and experienced many failures. Then the lowering of American tariff barriers in 1875 triggered rapid expansion, and thousands of immigrants arrived within a very few years. Prior to 1875, a large majority of the plantation workforce had consisted of Hawaiians, but that force was quickly augmented with many thousands of Portuguese, Chinese, and Japanese. Within less than fifteen years, the percentage of Hawaiians on plantations had fallen to ten percent or less (Roberts 1996; cf. also Table 1), an extreme case of Condition C. The linguistic consequence was the destabilization of Pidgin Hawaiian (PH). By the 1860s, Pidgin Hawaiian had stabilized and was the predominant medium of communication among all ethnic groups both on the plantation and off it (Bickerton & Wilson 1987; Roberts 1995a, b). It had reached a level of complexity illustrated by (1): (1)

Atahi runa pi mai hale wau la, iaia lawe mai One boss come DIR house I PRT he bring DIR nupepa newspaper “An overseer came to my house, bringing a newspaper.” (Roberts 1995a)

The Standard Hawaiian translation of this sentence, differing from the pidgin in little more than the presence of case-markers, articles, and other functional items, would read as follows: (2)

Ua hele mai kekahi luna i ko-u hale i lawe go DIR certain boss OM of-me house PFT take mai ai i ka nupepa DIR PRT OM the newspaper PFT

(DIR = directional; PRT = particle; PFT = perfective; OM = object marker)


148

But this stable and expanding pidgin was not transmitted to the workers who arrived after 1875, even though most of the old hands who instructed them spoke Hawaiian natively, as well as the pidginized version. Typical of the utterances found in the subsequent two decades are the following: 11 (3) a. Yu (E) nani (J) hanahana? (H) you what work-work “What work do you do?” b. Pehea (H) yu (E) kaitai (C) yu (E) hanahana (H) olsem (E) mi why you bastard you work-work all-same me hanamake (H) yu (E) work-dead you “Why, you bastard, if you do that again I’ll kill you.” c. Luna (H) san (J) mi (E) danburo (E) faia (E) de (J) boss HON me down-below fire OBJ mauka (H) ga (J) pilikia (H) mountain TOP trouble “Overseer, I have burning pains in my stomach and my head aches.” d. Inu (J) shinda (J) pake (H) mejishin (E) dog die Chinese medicine “My dog died, Chinese medicine killed him.”

koroshita (J) kill-do

(E = English; C = Chinese; H = Hawaiian; J = Japanese; HON = honorific; OBJ = object marker; TOP = topic marker) Of the 21 word-types (out of 26 tokens) that these sentences contain, seven are Japanese, seven Hawaiian, six English, and one Chinese. The basic strategy involved has been well described by an eyewitness of the process: (4)

11

So we use the Hawaiian and Chinese together in one sentence, see? And they ask me if that’s a Hawaiian word, I said no, maybe that’s a Japanese word we put it in, to make a sentence with a Hawaiian word. And the Chinese the same way too, in order for them to make a sentence to understand you. 12

For the sources of these sentences, see Bickerton 1999, fn. 4. From an interview conducted by the author in 1974. The subject, a Hawaiian woman in her eighties, was remembering conditions early in the twentieth century, when Condition C, a continuous steady influx of new hands outnumbering old hands, still obtained. 12


149

That strategy would seem to be an inevitable result of Condition C, where the need for new immigrants to communicate with one another outpaces the capacity of old hands to transmit any stable contact language that they may have previously developed. New hands simply use whatever can be best understood by their interlocutors, from whatever fragmentary knowledge of the various languages in contact that they may have picked up. But Condition C in turn results from inexorable market forces, which will operate wherever investors wish to maximize returns on their investment as quickly as possible (which means, wherever this is physically possible!). It follows that very many, probably most, plantation colonies went through Condition C or some approximation thereto at some stage of their development. Furthermore, the extent that the condition operates (that is, the degree of adverse ratio between old hands and new hands during the period of most rapid immigration) will determine the extent to which the prior contact language is destabilized and hence the degree of pidginization (i.e., communication via a structureless, macaronic medium; see Bickerton 1999) that will develop. 7. Conclusion As noted above, the traditional belief in a pidgin-creole cycle should probably be replaced, with regard to most plantation colonies, by an OCLpidgin-creole cycle, in which an original contact language, possibly not too far from the dominant language, served as an input to the succeeding Condition-C pidgin, rather than the dominant language itself serving as direct input to the pidgin phase. That there was a pidgin phase (outside of colonies like Reunion, Barbados and a few others where population ratios would not have allowed it) can hardly be doubted. As I have shown above, the plantation system was driven by economic forces that made almost inevitable a population imbalance between OCL and non-OCL speakers. An unfortunate development in creolistics has emerged, however, in recent work by DeGraff (2001a, 2001b); in striking contrast to the open-mindedness of that author’s previous work, his recent contributions treat pidginization and the consequent special status of creole languages as mere artifacts of analysis resulting from post-colonial racial prejudice and the old-fashioned, condescending ‘baby-talk’ approach to pidgins and creoles. 13 It must have escaped his attention that citizens of ancient empires such as China and Japan, that 13

Other authors have commented publicly on this deplorable trend. McWhorter (2002:35) observes that “the subjective plays a significant role in the reception of a thesis arguing that creoles are a structurally delineable class of languages...The ‘rub’ here is, at heart, a supposition that those delineating a Creole Prototype are designating creoles as ‘primitive’ languages, or ‘baby talk’, most explicitly asserted by DeGraff (1999a et al.).” Parkvall (2001:150) writes, “If creoles are claimed to be different from non-creoles, this is routinely interpreted as a claim that they are somehow inferior, and insinuating accusations of racism are constantly present (cf. DeGraff 2000a, 2000b).”

150


were civilized when the forefathers of most Caucasians were running around daubed in woad, behaved linguistically in exactly the same way as Africans did when subjected to similar conditions. Moreover, if in the future Western civilization collapses (not all that unlikely an event) and Europe is colonized by, say, Nigerians, speakers of English, French, German, and Swedish will pidginize just as readily as did speakers of Bambara, Twi, Yoruba, and Kikongo in the seventeenth and eighteenth centuries, if their conquerors subject them to similar sociolinguistic conditions. For one does not have to subscribe to dialectical materialism, one can even believe that capitalism is marginally the best of a bunch of bad economic systems and still accept that of its very nature, unfettered capitalism in a colonial context must produce the economic and demographic results described above. Given those results, pidginization and abrupt creolization follow as inevitably as night follows day.

References Alleyne, Mervyn C. 1980. “Introduction”. Theoretical orientations in creole studies, Valdman & Highfield (eds) 1980 (San Diego: Academic Press). 117. Baker, Philip. 1995. “Motivation in Creole genesis”. From contact to creole and beyond, Baker (ed) 1995 (London: University of Westminster Press). 1-15. Beaken, Mike. 1996. The making of language. Edinburgh: Edinburgh University Press. Bickerton, Derek. 1976. “Pidgin and Creole languages”. Annual Review of Anthropology, 5. 169-193. ----------. 1994. “The origins of Saramaccan syntax: A reply to McWhorter”. Journal of Pidgin and Creole Languages 9. 65-78. ----------. 1999. “Pidgins and language mixture”. Creole genesis, attitudes and discourse: Studies celebrating Charlene J. Sato, Rickford & Romaine (eds) 1999 (Amsterdam: John Benjamins). 31-44. ----------, & W.H. Wilson. 1987. “Pidgin Hawaiian”. Pidgin and Creole languages: Essays in memory of John E. Reinecke, Gilbert (ed) 1987 (Honolulu: University of Hawaii Press). 77-92. Campbell, P.F. 1993. Some early Barbadian history. Barbados: Caribbean Graphics & Letchworth Ltd. Chaudenson, Robert. 1992. Des îles, des hommes, des langues: Essais sur la créolisation linguistique et culturelle. Paris: L’Harmattan. Clements, J. Clancy. 2003. “The tense-aspect system in pidgins and naturalistically learned L2”. Studies in Second Language Acquisition 25. 245-281. DeGraff, Michel. 2001a. “On the origin of Creoles: A Cartesian critique of Neo-Darwinian linguistics”. Linguistic Typology 5. 213-310.


151

----------. 2001b. “Morphology in creole genesis: Linguistics and edeology”. Ken Hale: A life in language, Kenstowicz (ed) 2001 (Cambridge, MA: MIT Press). 54-121. Ferguson, Charles A. 1968. “St. Stefan of Perm and applied linguistics”. Language Problems of Developing Nations, Fishman, Ferguson, & Das Gupta (eds) 1968 (New York: John Wiley). Hinnenkamp, Volker. 1984. “Eye-witnessing pidginization? Structural and sociolinguistic aspects of German and Turkish foreigner talk”. Papers from the York Creole Conference (September 24-27, 1983) [York Papers in Linguistics, vol. 2], Sebba & Todd (eds) 1984 (Heslington, England: Dept. of Language, University of York). 153-166. Ligon, Richard. 1657. A true and exact history of the island of Barbados. illustrated with a mapp of the island, as also the principall trees and plants there, set forth in their due proportions and shapes, drawne out by their severall and respective scales. together with the ingenio that makes the sugar, with the plots of the severall houses, roomes, and other places, that are used in the whole processes of sugar-making. London: Printed for H. Moseley. McWhorter, John. 2002. “The rest of the story: Restoring pidginization to creole genesis theory”. Journal of Pidgin and Creole Languages 17. 1-48. Migge, Bettina. 1998. “Substrate influence in Creole formation: The origin of give-type serial verb constructions in the Surinamese plantation creole”. Journal of Pidgin and Creole Languages 13. 215-266. Mufwene, Salikoko. 1996. [Review of Plag 1993]. Journal of Pidgin and Creole Languages 11. 361-367. ----------. 2001. The ecology of language evolution. Cambridge: Cambridge University Press. Parkvall, Mikael. 2001. “Creolistics and the quest for creoleness: A reply to Claire Lefebvre”. Journal of Pidgin and Creole Languages 16. 147-151. Plag, Ingo. 1993. Sentential complementation in Sranan: On the formation of an English-based creole language. Tübingen: Niemeyer. Roberts, Julian M. 1995a. “Pidgin Hawaiian: A sociohistorical study”. Journal of Pidgin and Creole Languages 10. 1-56. ----------. 1995b. “A structural sketch of Pidgin Hawaiian”. Amsterdam Creole Studies 12. 97-126. ----------. 1996. “Pidgin usage on three Hawaiian plantations in the 1890s: Linguistic and sociohistorical evidence”. Linguistics and Language Teaching: Proceedings of the Sixth Joint LSH-HATESL conference (Technical Report #10), Reves, Steele, & Wong (eds) 1996 (Honolulu: University of Hawaii). 153-167. Singler, John V. 1996. “Theories of Creole genesis, sociohistorical considerations, and the evaluation of evidence: The case of Haitian Creole and the

152


Relexification Hypothesis”. Journal of Pidgin and Creole Languages 11. 185-230. Smith, Norval. 1987. The genesis of the creole languages of Suriname. Unpublished doctoral dissertation, University of Amsterdam, The Netherlands. Williams, Eric. 1962. History of the people of Trinidad and Tobago. Port of Spain, Trinidad: PNM Publishing.

A CURIOSITY OF MAURITIAN CREOLE NUMERICAL SLANG

Robert Chaudenson Université de Provence

ABSTRACT: Research that I conducted for my thesis approximately three decades ago unearthed an interesting use of numbers among speakers of Mauritian Creole, particularly among those living in urban areas. More specifically, the numbers 1-40 have slang forms, many of which participate in slang expressions. Although these innovations seem to be particularly tied to the gambling milieu, these numbers have been found to replace nouns, adjectives, and verbs in everyday speech. In this chapter, I will review the data that I collected for Mauritian Creole and compare it to the number formulas used in bingo games in Reunion Creole. As I will show, the two systems have distinct semiological, sociolinguistic, and linguistic characteristics. 1. Introduction About 30 years ago, while I was preparing my thesis on Reunion Creole, I sought to compare this variety to the other creoles of the zone (Mauritian, Rodrigan, and Seychelles) by conducting research on the islands where they are spoken. My goal was first to determine which terms and which structures of Reunion Creole were found in these languages. Some of the data collected at that time thus appear in my thesis (when common elements were discovered in the various research sites), but much remains unpublished; the Rodrigan data constitute an exception because I was able to use them in the (not yet completed) editing of the Atlas linguistique et ethnographique de Rodrigues, of which only one volume (paradoxically, the third) appeared, in 1992 (Chaudenson, Carayol, & Barat). In Mauritius, my research was conducted during multiple visits that I was able to make to that island. I pursued this research after the defense of my thesis (1972; see Chaudenson 1974). Indeed, I had thought at one time that I would do a comparative lexicon of the Indian Ocean creoles. At the end of the 1970s and the beginning of the 1980s, there were plans for a linguistic and ethnographic atlas of the creolophone archipelagos of the Indian Ocean, an early version of which was even sketched out within the framework of a program of the Agence de Coopération Culturelle et Technique (1979-1981).


154

2. Numbers in Mauritian Creole During my Mauritian research, my attention was drawn to a microsystem of coded language founded on the usage of numbers (niméro) that, in Creole, could replace the nouns, adjectives, or verbs of everyday speech. This language, of which only a half-dozen elements were well known to most Mauritians, was essentially urban (which was confirmed to me by surveys done in rural milieus). In the city, it was familiar above all to those who frequented gambling parlors, and it was through research carried out in this milieu— where I was guided by one of my friends, Nichols de Casanove—that I was able to establish a fairly complete list of these numbers. I did not, however, succeed in finding a meaning for every one of the ‘numbers,’ from 1 to 40, which is as high as the list goes (with the addition of the number 71). I was thus able to establish the inventory shown in Table 1, with, in some cases, two different meanings for the same ‘number.’ In order to facilitate reading, I have used a spelling that is easily readable for a francophone: the name of the number is pronounced according to the Creole system, but the differences are minimal (un “one” = ène; deux “two” = dé; neuf “nine” = nèf); it is sometimes preceded by the word “number” (niméro), especially for the numbers from one to nine.

1

Creole number ène

2

dé

3

trwa

4 5

kat senk

6 7 8 9 10

sis sèt uit nèf dis

• lamor • tig • ène tig • bèf • volèr • lavyane • la line • ma(r)tin

11 12

onz douz

• zanm • kano lo

13

trèz

• makro fam • Téréz va pa o soley

14

katòz

• koup sévé • kwafèr

Slang name • ène lipou • lirwa/lérwa lipou • zako • niméro zako • gran dilo • ène trwa zorey

English translation and commentary “a louse” “the king of lice” “monkey”

(lit., “big water”) “idiot” or “homosexual” (lit., “a three-eared person”) “idiot” or “homosexual” “death” “tiger” “five roupees” “bull”; “testicle”; fig., “idiot” “thief” “meat” “the moon” “species of bird (Acridotheres tristis)”; although ma(r)tin is the word used, the meaning is makro “fusilier (fish of the genus Caesio)”; ène dis = makro “legs” “boat on . . .” This is the probable meaning, though in Creole we would expect kanot instead of kano. “pimp” (lit., “Theresa doesn’t go out in the sun,” where Téréz [ter°z] is phonetically close to the pronunciation of the number thirteen, trèz [tr°z] [Editors’ note]. “hairdresser”

A CURIOSITY OF MAURITIAN CREOLE

15

Creole number kinz

16 17 18

sèz di sèt dizuit

19 20 21

diznèf vin vintéin

22 23 24

vintdé vintrwa vintkat

25 26 27 28 29 30 31 32 33 34

vintsenk vintsis vintsèt vintuit vintnèf trant trantéin trantdé trantrwa trantkat

35

trantsenk

36

trantsis

37 38 39 40

trantsèt trantuit trantnèf karant

71

swasantéonz

Slang name • tété • Get so kinz. ? • pyés • larzan • kas • Done mwa ène dizuit. ? • divin • vintéin, fer vintéin • vintéin • vintdé • soval • manzé • Mo fine vintkat. • dé kat (2 4) • légliz ? • vintsèt • vintuit • fèr titou • trant • sou • sinwa • ti baba • trantkat • ène trantkat swasantuit • trantsenk • Get trantsenk la. • gogot • zozo • zanfan ? ? • karant • katzéro (4 0) • kaka • swasantéonz

155

English translation and commentary “breasts” “Look at her breasts.”

“girlfriend”; “girl” “money” “a little bit of money” “Give me an eighteen (i.e., some money).” “wine” “to get going” “to drink” “to get going”; “shoes”; “breasts” (rare) “horse” “to eat” “I ate.” “to eat” “church” “the police” “to get going, to leave” “to urinate” “a bluff” “drunk” “Chinese” “baby” “clever” (lit., “a thirty-four sixty-eight” [< 2 x 34]) “a very clever person” “pretty girl” “Look at the pretty girl.” “penis” “child” “buttocks ” “to defecate” “lame, who limps”

Table 1: Mauritian numerical slang Given what I had learned from my research in Reunion, I immediately thought of comparing this system to that of the announcements at bingo games. In Reunion, as doubtless nearly everywhere in France, there are puns on the numbers called. Sometimes they follow the announcement of the number drawn, sometimes they are substituted for it. For readers who may not know bingo as it is played in Reunion, I should point out that the game involves filling out a grid or card on which there are three lines of five numbers (one line is called a quine). The first player who, during a random drawing of numbers from 1 to 90, fills up a quine wins that round of the game; the first player


156

to obtain a “full card” (carton plein) because his fifteen numbers are called first, wins the game. As the game progresses, the drawings are announced aloud to the players. It is during the announcing of these numbers that the associations and/or substitutions of formulas to which I am referring here are made. Examples of these Reunionese formulas are presented in Table 2. It may be seen that the system relies essentially on puns based on the following: • the phonetic form of the number (1, 6, 9, 10, 13); • an interpretation of the shape of the number (3, 11, 55); • allusions to cultural traits of Reunion (4: allusion to a local riddle; 14: the strong man here is the hero of Grimm’s tale “The Little Tailor” 1; 21: allusion to the draft age). Number 1

Creole name in [°~]

Slang name or expression tou s k in koson pé dir

3 4

trwa kat

zorey sat kat pat mon sat dan zanbrevat

6

sis

sizo fin

9

nèf

la zourné la poul

10

dis

Dis . . . pité vou pou lé gro sou d la kine.

11 13

onz trèz

14 21 32

katòz vintéin trantdé

békiy Térèz, mon ti sér Térèz lom for zène konskri ti kouto

55

senkantsenk

59

senkantnèf

maryaz sin zil an palankin baro la zol

English translation and commentary “anything a pig can say” (The form in [°~] represents both the numeral “one” and the indefinite article “a, an”) “cat’s ear” “four paws my cat in the pigeon pea” (Pigeon pea, also known as Congo pea or Angola pea, is the plant Cajanu indicus [Editors’ note].) “fine scissors” (In French and Creole, the first syllable of ciseau/ sizo is homophonous with the word for “six” six: [si] [Editors’ note]) “the day of the chicken,” i.e., “an egg” (In French and Creole, “an egg” is homophonous with “a nine”: un œuf ~ in èf / un neuf ~ in nèf [°~nœf] ~ [°~n°f] [Editors’ note]) “Fight over (lit., dispute) the large sums of the quine.” (The first syllable of dispité “to dispute, fight over” is homophonous with the word for “ten” [dis] [Editors’ note]) “crutch” “Theresa, my little sister Theresa” (See the entry for 13 in Table 1.) “the strong man” “young conscript” “small knife” (the Opinal n° 32 is the most common) “wedding in Saint-Gilles in a palanquin”

“prison gate” (The prison is located in Saint-Denis at 59 rue Juliette Dodu.)

Table 2: Examples of Reunionese bingo slang 1

In the French tradition the ‘little tailor’ boasts of having killed seven with one blow—without specifying that it was seven flies he killed; the Creole tale includes the line, “Kill seven, wound fourteen,” whence the reference in question.


157

We have here, then, a system that is very different from that found in Mauritius. The systems differ in three respects, the first of which is on the semiological level. Indeed, in Mauritius it is generally not possible to establish a clear relationship between the number and its coded meaning, whereas in Reunion the formulas that are substituted for numbers are always motivated. Mauritian exceptions to this generalization include the cases of 3 (for the allusion to ‘ear’), 11, 20, and perhaps 71, which are comparable to the examples from Reunion. Given the general tendency, I am inclined to believe that these Mauritian cases (to which we could probably add 13 Térèz va pa o soley, though only in part since there are two formulas for this number) might be contaminations of a Mauritian usage of numbers for the lottery comparable to that of Reunion. This hypothesis is strengthened by the fact that the Mauritian ‘numbers’ do not appear to go beyond 40 and that the presence of 71 is therefore inexplicable. The systems also differ on the sociolinguistic level. In Reunion, bingo is more widely known and practiced in rural areas. This is the inverse of the situation in Mauritius, where the numbers system is urban, as we have seen. Bingo seems, then, to have left only traces in Mauritius, more or less mixed in with the numbers of the other system. Finally, the systems differ on the linguistic level in that, in Reunion, the use of the formulas associated with the numbers is almost exclusively linked to bingo, whereas in Mauritius, the conventional formulas may be substituted for the numbers outside of a gambling context; in some cases, the relationship may be inverted and the image is used for the number (e.g., tig = “five” for “five rupees”). The formulas are also used as nouns, adjectives, and in some cases, ordinary verbs (with preverbal markers). When the Diksyoner kreol morisyen (Baker & Hookoomsing 1987) appeared, I enjoyed comparing it with my data (which had been gathered 20 years before), in particular to see if the system of numbers was still in use. Unfortunately, there is no synthesized presentation. Under nimero “number” one finds, with no particular explanation, very diverse elements for which no mention is made of this system of ‘numbers’: nimero ène “number one”: 1. Expert, specialist 2. Bedbug nimero dé “number two”: monkey nimero trwa “number three”: ear niméro kat “number four” > kat nimero senk “number five”: Muslim numero sis “number six”: homosexual nimero set “number seven”: thief [English translations provided by the Editors.]

Alongside these terms, which undoubtedly come for the most part from the system I am describing, other words that have nothing to do with it are noted. This is the case of niméro ène “number one” (in the sense of “expert”), of

158


course, but also of niméro vinkat “number twenty-four,” meaning “the smallest size of fish hook (nearly invisible to the naked eye).” This term is noted as having a Creole etymology, which underscores the error made: In reality, it has to do with the traditional numbering of hooks that used higher numbers for smaller hooks. Kat “four” is the only term of this group that is the object of a particular treatment, indicated by the cross-reference. Under kat we find: 1. Quatre (argot des chiffres). 2. [n] Mort. 3. Mourir [...] 1 < F; 2:4 ? infl K fer katal [décapiter], ? infl.< H cf. khat : a bier ...on which dead bodies are conveyed to the pile. Forbes 1859 [1. Quatre (numerical slang). 2. [n] Death. 3. To die [...] 1 < F; 2:4 ? infl K fer katal [decapitate], ? infl. < H cf. khat: a bier ...on which dead bodies are conveyed to the pile. Forbes 1859 (Editors’ translation of French portions of quote)].

The entry is not very clear. It makes reference to “numerical slang,” but at the same time suggests an Indian origin. This is hardly consistent, since it is hard to see how just one element of this numerical slang could have its origin in Hindustani lexemes, which obviously have no connection to this number (unless this same explanation could be given for all the other numbers, which does not seem to be the case). It is curious, moreover, that the qualifier “numerical slang,” abbreviated as “num. sl.,” which figures at the end of the volume in the table of illustrations, does not appear for any of the terms cited under nimero (cf. above) which obviously belong to this category. It does appear, on the other hand, for other terms that must be looked up one by one throughout the dictionary, which complicates the task. Table 3 presents a comparison of the two inventories; using my own data as a point of departure (RC), I examined the presence of each ‘number’ (1-40) in the 1987 dictionary as well as whether the meaning offered is the same or different from what I recorded through my research. Notice that there are cases where the meaning in the dictionary is the same as what I have found, but the authors have not mentioned numerical slang. There are also cases where the term is not identified as belonging to numerical slang, and the meaning is different from my data. Comparisons involving these different characteristics are indicated in the table. The dictionary’s inventory is thus somewhat less extensive than mine. Out of the theoretical list of 40 items, five are missing from my inventory, compared to twelve in the 1987 dictionary. Additionally, the 1987 dictionary does not mention that seven of the numbers belong to numerical slang, even though this appears obvious. If we add that, in eight cases, the meanings given in the two sources are not the same, we see that the two lists are substantially


159

1987 Dictionary

Meaning Creole name Missing Numerical slang Same Different 1 éne 9 9 2 dé 9 9 3 trwa (“ear”) 9 9 4 kat 9 9 9 5 senk (“Muslim”) 6 sis (“homosexual”) 9 9 7 sèt 9 9 8 uit 9 9 9 nèf 9 10 dis 9 11 onz (“feet”) 9 12 douz 9 13 trèz 9 14 katoz (“head”; “intelligence”) 9 15 kenz 9 16 sèz 9 17 disèt 9 18 dizuit 9 19 diznèf 9 20 vin (“to run”) 9 21 vintéin 9 22 vintdé 23 vintrwa 9 9 24 vintkat 25 vintsenk 9 26 vintsis 9 27 vintsèt 9 28 vintuit 9 29 vintnèf 9 30 trant 9 31 trantéin 9 32 trantdé 9 33 trantrwa 9 34 trantkat 9 35 trantsenk 9 36 trantsis (“Black”) 9 37 trantsèt 9 38 trantuit 9 39 trantnèf 9 40 karant 9 Note. RC provided a meaning for vintdé not noted in the 1987 dictionary. The slang meaning for some of the ‘numbers’ is provided in parentheses. A checkmark in the numerical slang column indicates that an item is included in the dictionary but not identified as numerical slang

Table 3: Comparison of two inventories of Mauritian numerical slang

160


different and that only the best-known items on them coincide (my own lacunae are also found in the dictionary’s inventory, and thus I learned nothing new when I consulted it). However, the presence of this numerical slang in the dictionary, 20 years after my own research, testifies to the vitality of this microsystem. What remains to be done is to describe its usage and determine its origin with greater precision. As we have seen, the Mauritian system goes far beyond what may be observed in Reunion for bingo numbers. In the latter, a formula is associated with, and sometimes substituted for, a number in the context of the game. In Mauritius, it is possible to substitute a ‘number’ for a lexeme in everyday conversation, but the inverse is also possible. A merchant questioned about a price by a customer he knows might answer “zako” (“two rupees”) or “tig” (“five rupees”). This system is also used with cryptic intent, for example, in order to give a price that will not be understood by a foreign tourist. This type of usage is, however, less frequent than the use of the best-known terms, whether or not they are preceded by the word niméro: Sa ène niméro set! = “He’s a thief!”; Get so kenz! = “Look at her breasts!”; Get so karant! = “Look at her butt!”; Get trantsenk la! = “Look at that chick!” The relatively secret nature of the formulation allows it to be used in the presence of the person in question, especially if she happens to be a tourist, which spices up the situation a bit. A notable point is that these numbers can sometimes serve as verbs. Of course, not all numbers are subject to this ‘verbalization’ and, depending on the case, the verb fer “to do, to make” may be added to them: “to drink” may thus be vintéin “twenty-one” or fer vintéin “make twenty-one”; “to leave” may be vintuit “twenty-eight” or fer vintuit “make twenty-eight”; “to urinate” may be vintnèf “twenty-nine” or fer vintnèf “make twenty-nine.” Sometimes, the term is used by itself: mo fine vintkat = “I ate.” According to the evidence I was able to gather, this microsystem might have had its origin in a Chinese card game. This appears plausible, considering the passion that the Chinese have, in Mauritius as elsewhere, for all forms of games of chance. Two facts seem to confirm this possibility: on the one hand, the limitation of the numbers to 40 (on the hypothesis that the cards or figures of the game would have been limited to this number), and, on the other, the meanings of the numbers, which appear to be found in pictures. In spite of all the research that I have done in this domain, I have not been able to identify the game that could be the source of this curious feature of Mauritian Creole. Perhaps a reader of this article might be able to give me the key to this enigma.


161

References Baker, Philip & Vinesh Hookoomsing. 1987. Diksyoner kreol morisyen/ Dictionary of Mauritian Creole/Dictionnaire du créole mauricien. Paris: L’Harmattan. Chaudenson, Robert. 1974. Le lexique du parler créole de la Réunion. Paris: Champion. Chaudenson, Robert, Michel Carayol, & Christian Barat. 1992. Atlas linguistique et ethnographique de Rodrigues (vol. 3). Paris: Agence de Coopération Culturelle et Technique.

THEORETICAL AND PRACTICAL CONDITIONS FOR THE EMERGENCE OF A KOINE AMONG FRENCH-LEXIFIED CREOLE LANGUAGES

Jean Bernabé Université des Antilles et de la Guyane

ABSTRACT: Creole language, like language itself, is both one and many. Its unity, considered a myth by some, is largely attributable to a shared lexifier. Mutual intelligibility between creoles is a function of the dialectal distance that separates them and of the speakers’ knowledge of the lexifier. Paradoxically, mutual intelligibility is becoming increasingly possible between speakers of different French-based creoles because of schooling in French, which makes French the native language of creole speakers from various nations. The French language as a common denominator and the structural similarity of the various French-based creoles increase the likelihood of mutual intelligibility. The creation of a veritable ‘linguistic market’ will, over time, allow a new inter-creole dynamic to develop. In this light, I examine institutional changes being brought about by the increased use of creole languages in the school setting.

1. Introduction Creole languages cannot be discussed without making reference to the nature of language itself. Language is an innate competence of human beings that reveals their capacity for symbolism. Paradoxically, language is simultaneously universal and diverse; it is universal in that it is shared by the entire human species, and it is diverse in that it is realized in specific languages. Each language is itself a set of rules that transcends the linguistic production of individuals while making that production possible. Language is a tool for communication, but it can also act as a barrier between different human communities that do not share a given form of speech. In other words, language is riddled with paradoxes. 2. Universality vs. diversity in creoles Interestingly, the way in which this opposition between unity (universality) and diversity cuts across creole linguistic reality means that linguists sometimes talk about creole and sometimes about creoles. Is the relationship of

164


creole to creoles of the same kind as the relationship of language to languages? Are particular creole languages concrete realizations of a universal system that transcends them? In other words, is creolization a process that stems directly from the human language faculty? And if so, where do creoles get this feature? And why then are not all of the languages of the world creole languages? These questions are worth asking even if, in the framework of the present chapter, no response can be attempted. It is not controversial to state that individual languages (which are instantiations in specific communities of the universal language faculty) are not readily mutually intelligible unless they happen to be slightly different variants of one another. This variance is the basis for translation, which illustrates the uniqueness of different languages and reduces their isolating effects. Without translation, people must rely on direct comprehension of the other language, and, generally speaking, to understand a language other than one’s own, it has to be learned. Sometimes the distance between the two languages is enormous. But this distance cannot be appreciated without asking whether the relationship between two languages is logical (typological, structural), or historical (by descent from a common ancestor or through a relationship of language contact, which facilitates borrowing and convergence). The Romance languages, for instance, share a common ancestor, Latin, which makes them closer to each other than they are to, say, African languages. Common ancestry does not, however, make them mutually intelligible; that is, a Spanish speaker does not understand a speaker of French or Italian unless he or she has at one time or another learned French or Italian. In other words, the genetic unity of the Romance languages hardly guarantees that their speakers will find them mutually intelligible. Where creole languages are concerned, the problem is both the same and different, which is precisely why the diversity of these languages is of particular interest. One of the distinguishing features of creole languages is the fact that they have resulted from contact between, on the one hand, various languages whose use had more or less been lost by their native speakers—generally African languages, due to the historical phenomena of slavery and the slave trade— and, on the other hand, a language with a more lasting presence that supplied the vocabulary—generally a European language, be it French, English, Portuguese, or Dutch, yielding what creolists refer to as French-based, Englishbased, Portuguese-based, or Dutch-based creoles. 1 In the group of Frenchbased creoles, two large geographical zones are distinguished: the American or Caribbean zone (Haiti, Guadeloupe, Guyana, Dominique, Martinique, and Saint Lucia, where the creole is still alive; Trinidad and Louisiana where it is 1

It should be noted that the Spanish language, for various reasons, generated practically no creoles apart from Palenquero, in Columbia. In Cuba, Puerto Rico, and the Dominican Republic, local varieties of Spanish are spoken.

CONDITIONS FOR THE EMERGENCE OF A KOINE

165

moribund; and Grenada where it is extinct); and the Indian Ocean zone (Mauritius and Rodrigue, Reunion, and the Seychelles, where it is also widely spoken). To use the metaphor of kinship terms, we can refer to (a) creole ‘sister languages’: those that share a common lexifier and are from the same zone; (b) creole ‘half-sister languages’: those that share the same lexifier but are from a different geographical zone; and (c) creole ‘cousin languages’: those that have different lexifiers. Thus Martinican Creole (MarC) and Guadeloupean Creole (GuaC) are sister languages, as are Reunion Creole (RC) and Mauritian Creole (MauC), whereas MarC and RC are half-sister languages. On the other hand, MarC and Jamaican Creole (JC) are cousin languages (Bernabé 2001). Among creole languages, mutual intelligibility can be quite high, but it can also be nonexistent. There is in fact no reason that a Martinican who does not know English should understand JC, or that a Jamaican who does not know French should understand MarC just because in both cases these are creole languages. Conversely, one can understand how mutual intelligibility may be possible when there is a shared lexifier. This mutual intelligibility depends on the speaker’s knowledge of the lexifier from which the creole in question is derived. For example, MarC and GuaC, both French-based, are fairly easily understood by Mauritians if the latter also understand French. But knowledge of the lexifier does not automatically ensure that a given creole will be intelligible to speakers of its sister or half-sister creoles; other criteria must be taken into consideration as well. For instance, MarC and GuaC share a higher degree of mutual intelligibility than do MarC and Haitian Creole; although the latter share the same lexifier (French), they are structurally further apart. Still, MarC and Haitian Creole are more mutually intelligible than are MarC and RC. The question of mutual intelligibility clearly concerns the dialect distance that separates two particular creole languages from each other. The amount of adaptation time required for a given creole speaker to understand another creole that shares the same lexifier certainly depends on this distance. It is, however, untrue that the distance is so great from one end of the spectrum of French- (or English-) based creoles to the other that it can be crossed only by gifted individuals or specialists in the languages in question. Those who make this claim believe that creoles are irremediably fragmented into separate dialectal varieties, as opposed to those who believe that a coming together of the great creole family scattered throughout the world is still possible. In this Internet age, such a coming together need not take place physically, nor does it necessarily have political ambitions. It is simply a human goal, one of fostering consciousness-raising that can lead creole-speaking peoples to affirm a shared identity.

166


3. Lexifiers as keystones In discussions of mutual intelligibility, the term ‘dialect’ is unavoidable, and so it is appropriate to discuss its application to creole languages. We must take into consideration what is fittingly called ‘areal linguistics,’ which concerns the configuration of given geographical areas, as is illustrated by the existence of the two geographical zones where French-based creoles are found. The various movements happening in creole languages must be studied both within and across these two zones. At present, however, opportunities for contact between the various French-based creoles are still rather limited, except in French Guyana, 2 where all of them are in extensive contact. Thus, it seems quite unlikely that a transregional creole ‘linguistic market’ will emerge unless the French school system encourages interdialectal exchanges between the Caribbean and Indian Ocean zones. Lexifier languages are, to be sure, the keystones of creole languages, but as mentioned above, a common lexifier will not necessarily enable speakers of one creole language to understand a sister creole. The transparent relationship between the lexifier and the creole can become so obscure to some speakers that mutual intelligibility is impossible. On the other hand, educational practices in the Antilles, Guyana, and Reunion may be helping to highlight the relationship between the lexifier and the Creole. Through the educational system, the French language, formerly a second language, has gradually become a mother tongue (and often the only mother tongue) for creole speakers in these areas. The phenomenon has not yet reached completion but it is inexorable. Let us now consider creole languages themselves, with their multitude of paradoxes. They may follow one of two evolutionary paths. The first is ‘decreolization,’ or the increasing similarity and ultimate merger with the lexifier (Hall 1966). Such a mechanism has the effect of increasing mutual intelligibility among the creoles themselves. The second path is ‘recreolization’ (by various means, notably literary creation or strong, collective influence on the mass media), or the establishment of an ever-increasing distance from the lexifier language. This process has the effect of reducing mutual intelligibility. The reduction of mutual intelligibility can result from two different developments. In the first, creole languages may follow their own evolution, with the most militant speakers trying to halt decreolization. But such an effort risks becoming a labor of Sisyphus, which in Creole is called chayé dlo an panyen (“carrying water in a basket”), because decreolization is an integral part of the lasting relationship between Creole and French, at least in an historical per2

French Guyana is a thinly populated land that has been a destination for immigrants who speak any of the various French-based creoles of the Caribbean. Today all of these creoles are therefore in contact and interacting in French Guyana.


167

spective. For instance, it is neither possible nor desirable to cut Martinicans off from French, a language to which they have a legitimate claim, and that is increasingly their first language. At the same time, it is also essential to introduce them to other Caribbean languages such as English and Spanish. Furthermore, despite their somewhat limited Amerindian antecedents, and their more significant African ones, French-based creole languages are no less neoRomance languages. Any program for directing their development that does not take this basic fact into account would be a purely ideological vision of the structural, lexical and, especially, terminological future of these creoles. In the second development, creole languages may enter a developmental path that is open to the other languages of the same geographical area (i.e., sister languages) or even of the other area (i.e., half-sisters). In short, the increasing use of Creole in schools and the maintenance of a single interdialectal teaching certification, the Certificat d’Aptitude pour l’Enseignement Secondaire (CAPES) (Bernabé & Confiant 2002), as a way of directing future developments towards a more normative Creole for educational uses, have established the conditions for a vast and new linguistic market. Such a market can take shape via an accelerated rate of cultural exchange bearing on Creole literature, theater, and music. It is interesting, in this light, to reflect on the problem of lexical creativity in Creole as well as on the phenomena of structural complementarity that concern these languages (cf. the ‘patchwork effect,’ discussed in section 5.2 below). 4. Naturalness vs. artificiality Although it has been contested on serious sociogenetic grounds by Chaudenson (1979, 2003), Prudent (1993), McWhorter (1997), and Mufwene (1997, 2002), the theory that sees creoles as nativized pidgins (Hall 1966; Valdman 1978 3) is not rejected by all creolists. The word ‘nativization’ (French: naturalisation) refers to the process of a pidgin becoming the mother tongue of a community that has been deprived of the use of its various original mother tongues and illustrates rather well the original meaning of the Latin word natura. The natura is etymologically the place where one was born (natus). Therefore the nativization of a creole sets up natura as a primary datum, unpremeditated, 4 collective, native, and primeval. Nativization is the opposite of artificiality, which situates language in the paradigm of what occurs after-the-fact, of what is premeditated, constructed, secondary. These oppositions highlight the clear difference between a language like French, or any historical creole language, and artificial constructs like Esperanto or Volapük. Naturalness and artificiality are two poles that preside over the evo3

Valdman (1977) defends a different point of view. Later, in 1978, he presents Hall’s point of view in a very clear, faithful and, in short, rather convincing way. 4 On the concept of language being unpremeditated, see Coursil (2000).


168

lution of languages, and it is possible to attribute this or that phenomenon of language development to one or the other of these poles. In this regard, we must distinguish (Bernabé 2002) between those phenomena associated with ‘spontaneous streamlining’ (French: la standardisation) and those associated with ‘normalization’ (French: la normalisation). By spontaneous streamlining I mean the natural movement of the language whereby its norms evolve in keeping with the subconscious representations of its speakers. Normalization, on the other hand, is a conscious activity that aims to codify the norms of the language in an authoritarian way. The field of language planning deals with the conflicts between these two processes. We should also differentiate spontaneous streamlining and normalization from ‘norming’ (French: la normation), 5 which is the activity of producing tools like grammars and dictionaries that promote a given set of norms and encourage their internalization. Neological processes are located at the very crossroads of naturalness and artificiality. Every language exhibits a number of lacunae (Geckeler 1974) or combinations that are permitted by the system of the language but never realized in actual usage. The following table from Connolly (1978) dealing with the derivational system of French illustrates this type of gap. Starting with the four roots listed at the top of each column, we see clearly that only a few combinations constitute actual words, even if the unattested forms are semantically valid in certain cases. (See Table 1.) [-mettre] [admettre] [commettre] [démettre] [*dismettre] [émettre] [*exmettre] [*immettre] [permettre] [remettre] [soumettre] [*sumettre] [transmettre]

[-porter] [apporter] [comporter] [déporter] [*disporter] [*éporter] [exporter] [importer] [*perporter] [reporter] [*souporter] [supporter] [transporter]

[-paraître] [apparaître] [comparaître] [*déparaître] [disparaître] [*éparaître] [*exparaître] [*imparaître] [*perparaître] [reparaître] [*souparaître] [*suparaître] [transparaître]

[-vaser] [*advaser] [*convaser] [dévaser] [*disvaser] [évaser] [*exvaser] [*imvaser] [*pervaser] [*revaser] [*souvaser] [*suvaser] [transvaser]

Table 1: Attested and unattested derivations in French Such a table allows us to get a sense of the lexical productivity of a given language. Without a doubt, a similar table for creoles would display a much larger number of lacunae, meaning that lexical derivation is much less produc-

5

Andrée Tabouret-Keller presented the difference between norming and normalizing, so dear to her, in her closing lecture at the “Colloque d’Amiens (Novembre 2001) sur les langues collatérales” (Tabouret-Keller 2004).


169

tive in creoles than it is in French. The task of filling these gaps, assigned to usage, might appear to be a matter of both naturalness as well as of artificiality: naturalness, because the realization of potential forms is involved, and artificiality, because usage has not yet produced the potential forms. The entire question of decreolization, and the remedies proposed to counter it, lie at the heart of this opposition. In this sense, the ongoing decreolization, which involves the spontaneous streamlining of Creole on the basis of the dominant model furnished by French, would fall on the side of naturalness. Furthermore, research indicates that many Creole speakers accept French influence on Creole as ‘natural’ and they readily accept the proposition that every language is subject to change. 6 These same people, however, remain unaware that linguistic evolution in this instance is proceeding unilaterally; borrowings go only from French to Creole because the H-language in the diglossic relationship serves as the lexical provider. 7 On the other hand, attempts at recreolization would involve normalizing efforts, which are often seen as utopian, dictatorial, and artificial. The facts of the debate raise a number of questions: Should the struggle against decreolization—a goal no less legitimate than creating a universal language from scratch—accept a measure of artificiality of the same kind that generated a language like Esperanto, or should it rather limit itself to what results from natural processes? Or should it combine the two approaches? The notion of a ‘koine,’ or interdialectal lingua franca, offers an answer to precisely these questions. But it is an entirely theoretical answer, because every lingua franca participates in natural, artificial, or both types of development. For example, the history of the mixed language resulting from dialect convergence in ancient Greece shows us that the language of the Iliad and the Odyssey is a composite, artificial language resulting from several dialect traditions joined together by Homer. Therefore Homer is transitional between a varied, diversified oral literature (French: oraliture) and a writing style that is traditionally characterized as koine. But in reality nothing should prevent such an assimilation, since both the Iliad and the Odyssey constitute a pooling of dialect forms that, if they are not melded into a single dialect, at least participate in the epics’ linguistic economy. Outside the Homeric rhapsody, 8 different historically attested koines are the result of contact between speakers of diverse language varieties. Thus, the conquests of Philip of Macedonia and his son 6

An epilinguistic study conducted by a group of master’s students enrolled at the Université des Antilles et de la Guyane in the school year 2000-2001 examined attitudes of Creole speakers toward decreolization. The results indicate that 90% of those interviewed believe that Frenchification of creoles is a natural direction of change, and that creoles, like all other languages, change over time. 7 Borrowing the economic terminology that is used in analyzing North-South relations, this implies a breakdown in the terms of the exchange. 8 In other words, the sewing together of various pieces. The Greek bards considered themselves rhapsodes.

170


Alexander facilitated contacts between Greek city-states that over the course of history led to linguistic unification. In this sense, the Hellenistic era opened the way past the original Greek heterogeneity towards the homogenization resulting in the Greek koine. Thus we need to evaluate the two paths represented by orality and writing without forgetting their effects on each other. That said, it appears clear that any sociohistorical mechanism that aims to produce linguistic exchange (in other words, sharing and not isolation) will generate a more extensive and unifying ‘linguistic market’ (Boyer 2001). From there it is not far to the natural development of a koine. That is why the creation of an interdialectal (GuaC, Guyanais, MarC, RC) CAPES as part of the mechanism in place for the regional languages of the French Republic constitutes the primary tool for establishing contact not only between languages but also, and especially, among linguistic usages (paroles), cultures, practices, and world visions. Such a tool seems to operate only under one condition: that an integrated educational program allow each of the Creole-speaking groups involved (a) to affirm itself through its language, (b) to remain open and in contact with the others, and (c) to be conscious of its role as co-creator of its own language. But the dynamics of a koine will not be effective unless other, more specific, factors are favorable as well. Let us now consider the essential characteristics of the context that would encourage the development of a possible creole koine. 5. The characteristics likely to facilitate the development of a Creole koine 5.1 Shared lexical stock The French-based creoles, mutatis mutandis, emerged from the same womb. In the case of the creoles of the Caribbean zone, rapid dialect differentiation did not take place until the nineteenth century. There is, therefore, a shared lexical stock attributable to the dialects of western France (Bernabé 2001). This factor is not unimportant to the extent that it ensures a certain coherence in the substrate of the different modern-day creoles. 5.2 The patchwork effect The present-day Caribbean creoles appear to operate by a mechanism that we could call the patchwork effect. One can observe a complementary distribution of grammatical forms that is visible only when all of the relevant dialects are considered. The following examples will serve to illustrate this effect. 5.2.1 The definite article. The definite article takes the following forms: in GuaC it is always la, 9 whether it follows a consonant or a vowel, a nasal or an oral segment; in Guyanais Creole (GuyC) it occurs as a after an oral segment (vowel or consonant), but as an after a nasal segment (vowel or consonant); 9

See also Hazaël-Massieux in this volume. [Editors’ note]


171

and in MarC it occurs as a after an oral vowel, an after a nasal vowel, la after an oral consonant and lan after a nasal consonant. These forms are summarized in Table 2. For this feature, MarC is the only variety which appears to present the entire spectrum of contextually varying forms, whereas GuaC is totally insensitive to these contextual variations. GuyC occupies an intermediate position; like GuaC it is not sensitive to the presence of a preceding vowel or consonant, but like MarC it is sensitive to the nasality of the preceding segment.

Context after oral V after nasal V after oral C after nasal C

GuaC lavi-la pon-la tab-la kann-la

GuyC lavi-a pon-an tab-a kann-an

MarC lavi-a pon-an tab-la kann-lan

English “life” “bridge” “table” “cane”

Table 2: Forms of the definite article in three sister creoles Context affirmative negative

GuyC a mo a pa mo

GuaC sé mwen a pa mwen

MarC sé mwen sé pa mwen

English “it’s me” “it’s not me”

Table 3: Forms of the emphatic focus particle in three sister creoles Word order Dem.+N+Def N+Def+Dem N+ Dem.+Def

GuyC sa tab-a

GuaC

MarC

tab-lasa tab-tala

English “that table” “that table” “that table”

Table 4: Forms of the demonstrative determiner in three sister creoles 5.2.2 Emphatic focus particle. The emphatic focus particle takes the form sé in MarC and the form a in GuyC, whether the context is affirmative or negative. However, in GuaC the focus particle takes the form sé in affirmative contexts and a in negative contexts. See Table 3 for an illustration of these forms. In this case, GuaC is the one that presents the full range of contextual variants whereas GuyC and MarC are insensitive to this type of contextual variation. 5.2.3 The demonstrative determiner. The demonstrative determiner takes the forms presented in Table 4: In this case, each dialect presents a different word order. As for the form of the signifier, GuyC and GuaC have the same form (sa), while MarC presents the form ta. 5.3 Lexical diversification, complementarity, and homogeneity The syntactic complementarity of these creoles accompanies a triple lexical characteristic: lexical diversification, complementarity, and homogeneity. As for lexical diversification, the different signifiers corresponding to the meaning

172


“thing” are well known emblematic terms. The forms in question are biten (GuaC), bet (GuyC), and bagay (MarC). In general, however, leaving aside phonology, the area of lexical overlap among the Lesser Antillean creoles is much larger than the area of lexical difference. This means that lexical diversification is far from extending across the entire vocabularies of these creoles. This lexical diversification has the following noteworthy features: complementarity, interdialectal borrowing and synonymy, and fusion rather than Frenchification. 5.3.1 Complementarity that fills lexical gaps. A given meaning may be expressed in one creole dialect by a signifier that differs little or not at all from French, whereas in another creole dialect the meaning in question, in the same range of uses, will be expressed by a completely innovative form vis-à-vis French. For example, the GuaC word kalanjé can correspond to all of the uses of the MarC word ézité (French hésiter “to hesitate”). Similarly, the GuaC word doukou is the basilectal equivalent of the GuaC and MarC konjonkti (French conjoncture “[economic] situation”). 5.3.2 Interdialectal borrowing. This lexical diversification also features an ease of interdialectal borrowing that flows from the phenomenon of ‘likely source dialect’ (vraisemblance des appartenances) and that facilitates the integration of loanwords into the borrowing dialect. Thus, for example, the GuaC word pal “help” is used more and more commonly by Martinicans whereas the MarC term obidjoul “adequate’ is used more and more frequently by Guadeloupeans. 5.3.3 Interdialectal synonymy. Lexical diversification also allows for interdialectal synonyms 10 to quickly become intradialectal. It is not necessary to furnish many examples illustrating the point that two lexical items having the same contexts of homologous syntactic and semantic usage, although they belong to two different dialects, can, if borrowed by the other dialect, enter into a relationship of synonymy with their homologue. In that case one can even have perfect synonymy, insofar as what passes for synonymy within a given language is often a loose semantic association. For many linguists, synonymy does not exist in natural languages. It will suffice to mention the equivalence emerging between GuaC konfedmanti and MarC asiré pa pétet “assuredly, most certainly” (French assurément, pour sûr).

10 The paradoxical nature of this assertion is evident, since the notion of synonymy cannot apply to lexical items that belong to two different linguistic systems even if one considers them two concrete subsystems of a single abstract system, which defines, on the one hand, the mutual relationships existing between the various creole dialects and, on the other hand, the relationship of each creole dialect with Creole (conceived of as a superstructure).


173

5.3.4 Fusion, not Frenchification. A fusion in the borrowing dialect can offer a credible alternative to systematic Frenchification or, at the very least, feed a productive competition with borrowings from French (that one could call ‘captive’), and in this way encourage a certain autonomy of creole languages. (It goes without saying that the Neo-Romance character of these creoles destines them to inevitable convergence with the lexical, and especially terminological, economy of the Romance languages.) By way of illustration, MarC and GuaC have borrowed so successfully from each other that speakers today may not realize that a given word started out in one dialect and was borrowed in the other. If one were to ask a group of Martinicans and Guadeloupeans under 30 years of age to identify the source dialect of the verb wouklé “to complain, to grumble” (in fact, of GuaC origin), they would be hard pressed to assign it to one or the other dialect. Assigning an origin to potalan “important,” one of the words most commonly used by the media due to the impetus of the GEREC-F 11 researchers, would give similar results. One could further mention plodari “speech,” djok “vigourous,” doukou 12 “[economic] situation,” matjoukann “heritage,” ladjoukann “servitude, slavery,” véyatiz “vigilance,” kanman “manner, style,” larel-fandas “fissure, faultline,” glorié “to pay homage to,” mofwazé “to metamorphose,” kouchal “pathetic-looking, ugly.” The case is the same for the word Béké, referring to a white Creole, which few young people know to be of Martinican origin. The assignment of an origin is all the more complicated in that in MarC, the form is palatalized to Bétjé, which does not correspond to the characteristics of GuaC phonology. That furthers the illusion that the former is the GuaC variant and the latter the MarC variant of a single transdialectal word, which, however, is not the case. It is no less true that lexical homogeneity is responsible for neutralizing source-dialect effects. From there a central notion emerges—that of ‘plausible creole.’ 5.4 Lexical neology The neutralization of source-dialect effects can be harmonized with the practice of lexical neology precisely because it goes hand in hand with the notion of plausible creole. Through lexical neology, creole terms that are unattested, not generated by derivational rules, are produced ex nihilo or almost ex nihilo. This practice could be called the James Joyce or the Henri Michaux syndrome, as both these writers—one a novelist, the other a poet— have a reputation as inventors of language. In this respect, the difference between langue and langage has never been so important. Leading not into 11

Groupe d’Etudes et de Recherche en Espace Créolophone et Francophone [Editors’ note] It is interesting to note that MarC dékou exists in a relatively independent way with the same meaning, but it is being supplanted by the increasingly dominant use of doukou whose GuaC origin is, mistakenly, not identified by those increasingly numerous Martinican writers who use it. This is despite their being under the influence of Creole media, inspired by the GERECF researchers. 12

174


langue but into langage, the work of Joyce and Michaux is certainly of the nature of a ‘fantasy linguistics’ (Auroux et al. 1985; Yaguello 1984) that engages in a process of fictional (as opposed to derivational) word creation. Although even with the help of context, comprehension (in the Cartesian sense) of the literary works concerned is partial and varies from one reader to the next, these works are no less significant. Furthermore, the idiosyncratic nature of the language of these works does not prevent them from being translated, which supposes that the translators must call upon cognitive and imaginative skills. A study of these skills would reveal much about this sort of writing. Without wishing to systematize such a practice, Creole-language writers and Creole linguists could take charge of it in a convergent and synergistic way, making it into a perfectly efficient, if peripheral, means of lexical enrichment of ‘nuclear’ creole vocabulary (Bernabé 1978), as opposed to a kind of creole that has undergone an intense ‘creolization’ and which one would not wish to eliminate in a spirit of purification. 13 Examples of this practice include the following terms: owonzon “approximately,” a kontrilanm “against the grain, the wrong way,” ranboulzay “revolution,” ranboulzayé “to revolutionize,” an blodal “violently,” an blipann “brutally.” Such words are progressively introduced to readers who, at present, recognize them only passively. 5.5 Use of Creole in the mass media Along with its increasing use in schools, which the current CAPES consecrates and legitimizes, Creole has entered into the world of the media, a privileged place for promoting a streamlined creole in planned delivery 14 that results not only from preparation but possibly also elaboration, and that takes account of the communicative constraints of this new setting. From the independent local radio stations launched at the beginning of the 1980s that used creole in an open and systematic way, television has taken over: Antilles Télévision in 1999 and Radio France Outre-Mer in 2000. There is clearly a desire to use a middle-of-the-road media creole, neither too Frenchified nor too 13 It scarcely bears repeating that emphasizing the unilateral character of the French-Creole relationship and trying to create positions that, if not capable of remedying it, can at least reorganize it, constitutes a healthy glottopolitical approach rather than a terrorist-inspired one, as those who have no special regard for the procedures of the GEREC-F systematically but unfruitfully repeat. Let me repeat this as forcefully as I can: normalizing authorities propose, the dynamics of spontaneous streamlining dispose. 14 Let me emphasize the two discourse strategies recommended by the GEREC-F (Bernabé 1999, 2002): on the one hand, the spontaneity associated with unplanned discourse and on the other hand, the elaboration associated with deferred (preplanned) speech. Both of these flow from communicative constraints. The former concerns the naturalness pole and spontaneous streamlining (even if that means on a Frenchifying basis); the second involves the pole of artificiality and a normalizing perspective. That said, it is important not to overlook the mutual effects of each of these strategies on the other as the new standard variety being targeted is acquired.


175

basilectal. Frequent use is made of terms that are a product of ‘plausible creole’ (with or without an attested stem, but not generated by any discernible derivational procedures) and that are broadcast by the GEREC-F researchers with some degree of success. Some of these terms enjoy a measure of success in formal situations outside of the mass media. Their surprisingly repetitive and stereotypic character might seem to suggest a limited lexical creativity on the part of those who make use of them on the air and, more generally, in the ‘marketplace’ of creole neology. In fact, there is an implicitly pedagogical significance to this neological practice that seeks to ease the transition from an artificial practice to a natural one. This neological practice seeks to bring the public gradually into the standard that, although still timidly emerging, brings daily pressure to bear on the public. Careful future research should explore interactional mechanisms at work (Gumperz 1989) between producers and receivers of this emerging norm. 5.6 Waiting period for new terminology In the coming together of the development of a koine and the artificial activity of neology, there is a period of waiting for the creation of terminological banks based on institutional practices that have been set up for French, both in France and in Quebec (since the field of terminology is the area of the lexicon that must be particularly defined). Nevertheless, one can scarcely forget the fate of some terminological inventions brought about by the dynamics of natural language change when looking at actual usage. For instance, despite its symbolic value, 15 the word bouteur, which was proposed to replace the American English (and universal) word “bulldozer,” has never been able to escape from the conceptual fortress in which it was created. 5.7 Contact among creoles in the educational system While working toward the interdialectal creole CAPES, the pupils of the collèges and lycées of each country come into contact with the oral and written productions of speakers of the other creoles in an institutional setting. This contact creates a need that can only encourage borrowing among the different varieties; over time, the varieties will no longer be considered xenolects or ‘heterolects,’ but ‘homolects.’ There will be a diversity of creoles, to be sure, but there will also be a creole complicity.

15 In the context of a symbolic attempt to fight against the invasion of franglais (in actual fact, in its American English variety), in denouncing which Etiemble became famous, what better word than the neologism bouteur which reminds one of Joan of Arc’s epic struggle to drive (bouter) the English out of France?

176


6. Conclusion The stakes implicit in the contradictory notions of the unity and the diversity of creoles are clear. These stakes are not merely descriptive; they have a broader scope involving not only the technical activity of language planning, but also language policies. For today’s youth, who lack fixed reference points and ideals, one can imagine the openness and the hope that are engendered by the possibility of concretizing the potential links that exist between the Frenchbased creoles in all areas of public and private life. There is no doubt that what are conventionally called French Overseas Departments (Départements d’Outre-Mer) will derive an increasingly coherent culture and identity from such links. Thus, one can imagine concretely bringing together the unity and the diversity of creoles. It bears repeating that the creation of an interdialectal creole CAPES (“un CAPES transversal de créole”), so-named by the ministerial authorities of the French Republic, following the desire of GEREC-F, cannot help but activate the means for deepening and appropriating a mosaic creole identity that is bursting with life.

References Auroux, Sylvain, Jean-Claude Chevalier, Nicole Jacques-Chaquin, & Christiane Marchello-Nizia (eds). 1985. La linguistique fantastique. Paris: Denoël. Bernabé, Jean. 1978. “Problèmes et perspectives de la description des créoles à base lexicale française”. Etudes Créoles 1. 11-20. ----------. 1999. “La relation créole-français: Duel ou duo? Implications pour un projet scolaire”. Langues et cultures régionales de France: Etat des lieux, enseignement, politiques. Actes du colloque (Université Paris V-René Descartes, 11-12 juin 1999), Clairis, Costaouec, & Coyos (eds) 1999 (Paris: L’Harmattan). 35-52. ----------. 2001. La fable créole. Petit-Bourg, Guadeloupe. Editions Ibis Rouge. ----------. 2002. “Lire et délires dans les créoles à base lexicale française”. Codification des langues de France : Ecrits divers–écrits ouverts: Actes du colloque les langues de France et leur codification (Paris, l’INALCO, 1931 mai 2000), Caubet, Chaker, & Sibille (eds) 2002 (Paris: L’Harmattan). 233-255. ---------- & Raphaël Confiant. 2002. “Le CAPES de créole: Stratégies et enjeux.” La France et les outre-mers: L’Enjeu multiculturel, Bambridge (ed) 2002 (Series: Hermès, vol. 32-33; Paris: Editions du CNRS). Boyer, Henri. 2001. Introduction à la sociolinguistique. Paris: Dunod. Chaudenson, Robert. 1979. “À propos de la genèse du créole mauricien: le peuplement de l’Île de France de 1721 à 1735”. Etudes Créoles 1. 43-57. ----------. 2003. La créolisation: Théorie, applications, implications. Paris: L’Harmattan.


177

Connolly, Guy. 1978. Linguistique descriptive: Méthodes et applications en phonologie et en morphologie. Montreal: Guérin. Coursil, Jacques. 2000. La fonction muette du langage. Petit-Bourg, Guadeloupe: Editions Ibis Rouge/Presses Universitaires Créoles. Geckeler, Horst. 1974. “Le problème des lacunes linguistiques”. Cahiers de Lexicologie 25. 31-45. Gumperz, John J. 1989. Sociolinguistique interactionnelle: Une approche interprétative. Saint-Denis de la Réunion: Université de la Réunion; Paris: L’Harmattan. Hall, Robert A., Jr. 1966. Pidgin and creole languages. Ithaca, NY: Cornell University Press. McWhorter, John. 1997. Towards a new model of creole genesis. Berlin: Peter Lang Mufwene, Salikoko. 1997. “Jargons, pidgins, creoles and koinès: What are they?”. The structure and status of pidgins and creoles, Spears & Winford (eds) 1997 (Amsterdam: John Benjamins). 35-70. ----------. 2002. “Développement des créoles et évolution des langues”. Etudes Créoles 25. 45-70. Prudent, Lambert-Félix. 1993. Pratiques langagières: Genèse et fonctionnement d’un créole. 3 volumes. Thèse d’Etat, Université de Rouen. Tabouret-Keller, Andrée (2004). “Pourquoi veut-on qu’un parler soit une langue?”. Des langues collatérales : Vol. 1. Problèmes linguistiques, sociolinguistiques et glottopolitiques de la proximité linguistique. Actes du colloque international réuni à Amiens, du 21 au 24 novembre 2001, Eloy (ed) 2004 (Paris: L’Harmattan). 77-90. Valdman, Albert. 1977. “Créolisation sans pidgin: Le Système des déterminants du nom dans les parlers franco-créoles français aux Antilles”. Langues en contact: Pidgins-créoles—languages in contact, Meisel (ed) 1977 (Tübingen: TBL-Verlag Narr). 105-136. ----------. 1978. Le créole: Structure, statut et origine. Paris: Klincksieck. Yaguello, Marina. 1984. Les fous du langage: Des langues imaginaires et de leurs inventeurs. Paris: Le Seuil.

FRENCH IN HAITI CONTACTS AND CONFLICTS BETWEEN LINGUISTIC REPRESENTATIONS

Corinne Etienne University of Massachusetts, Boston

“…je préfère parler en grec que parler en français.” (Participant 14) ABSTRACT: Defining a variety of French involves not only showing its specificity compared to other French varieties but also reporting on its speakers’ linguistic representations (Robillard 1993b). In this paper I investigate how French in Haiti is defined by its users and whether they perceive it as a specific variety of French. Data come from semi-structured interviews with French-Creole bilingual Haitians focusing on four questions: What is good French?; what is good Creole?; does Haitian French exist?; and how is it defined? Findings indicated that only half the participants recognized French in Haiti as a local variety of French. The definition of Haitian French appeared to be inextricably linked to Creole. Participants’ epilinguistic discourses revealed various conflicts between linguistic uses marked by French-Creole contact and participants’ French and Creole ideal norms. 1. Introduction Discussion of a local variety of French in Haiti involves bringing to light not only the linguistic but also the sociological and ideological processes at work in a language contact situation where two of the languages in contact, French and Creole, are closely related. As shown by Robillard (1993b), defining a variety of French entails not only showing its specificity compared to other French varieties, but also assessing its sociolinguistic functionality and reporting on its speakers’ linguistic representations. Linguistic practices and behaviors, or what speakers do with language, are commonly distinguished from linguistic representations, or what speakers say or think they do with the languages available in their linguistic repertoire. Linguistic representations imply that speakers express judgments and opinions about what they recognize as theirs or others’ practices and uses (Baggioni & Kasbarian 1993). These judgments and opinions are crucial to understanding the contacts and potential conflicts between varieties or languages in a polyglossic situation. Robillard (as cited in Dumont & Maurer 1995:101) aptly

180


said, “Language contact starts with contact between language representations.” 1 Following Robillard’s model and based on a study of lexical particularities in the Haitian press, I have endeavored to enrich my intralinguistic perspective—that of an outside linguist—of French in Haiti. Adopting an epilinguistic perspective, I explored native speakers’ reactions to and opinions about my linguistic observations. According to Robillard (1993a:21), the features of any French variety can be identified from several viewpoints: that of the speaker, who may say one thing (an explicit representation) but then behave in a way that contradicts what was said; that of people from other Francophone countries who recognize the variety as a Francophone variety; and that of linguists, who, through a purportedly objective analysis, may identify a variety when the speakers themselves do not recognize one, or claim not to recognize one. The model proposed by Robillard is particularly useful to show distortions that may exist between linguistic behaviors and representations within a given community. These distortions may then be explained by the specific social, economic, and cultural environment of the community. For Robillard, any lexicographic work should take into account any such distortions insofar as it “should aim to be more of a catalyst triggering consciousness than a ‘definite’ and ‘complete’ inventory, by presenting a linguistic community with an image of itself, in order to see if it recognizes itself or not in this work.” (1993a: 33) 1.1 Summary of the sociolinguistic study of the lexical particularities of French in the Haitian press (Etienne 2000) This article is based on an extensive sociolinguistic study of the lexical particularities of French in the Haitian press (Etienne 2000). For this study, I compiled a sample inventory of the lexical particularities 2 gathered from a corpus consisting of French language newspapers targeted toward FrenchCreole bilingual Haitians in the American diaspora and in Haiti. The particularities in this inventory from the Haitian press consisted of 49.6% creolisms, 28.7% neologisms, 8.4% anglicisms, and the remaining items or traits of unascertained or unknown origin. I examined 46 participants’ perspectives on French in the Haitian press by giving them two epilinguistic tasks. The tasks asked them to serve as editorsin-chief of a fictitious newspaper targeting a Haitian readership and a fictitious newspaper targeting a Francophone readership. Participants were FrenchCreole bilingual Haitians from Haiti and the diaspora. The material used in each task was a collage of press excerpts taken from the corpus that I had 1

All translations are the author’s. ‘Lexical particularity’ has been defined as any item or lexical trait that does not belong to referential French (RF). RF is French as it is described in the most commonly used dictionaries and grammars. For examples of lexical particularities, see Etienne (2002, 2005). 2

FRENCH IN HAITI

181

gathered for the lexicological study. All press excerpts contained lexical particularities and 29 particularities (creolisms, neologisms, and anglicisms) were common to both tasks. The goal of these two editing tasks was twofold: Task 1 (T1) revealed subjects’ tolerance or intolerance for Haitian lexical particularities in a newspaper targeting a Haitian readership; Task 2 (T2) measured subjects’ tolerance of these same lexical particularities in a newspaper targeting a Francophone readership. I expected participants to be less tolerant of Haitian lexical particularities in a newspaper targeting a Francophone readership than in a newspaper targeting a Haitian readership. In each task, participants received a score corresponding to the number of particularities they rejected. For instance, P1 received a score of (7,15), which means that she rejected 7 particularities in the newspaper targeting a Haitian readership and 15 in the newspaper targeting a Francophone readership. I considered her as relatively tolerant of Haitian lexical particularities, whereas I considered P3, who received a score of (12,16), as relatively prescriptive. Participants whose rejection rates in both T1 and T2 were equal to or higher than 8 were considered to be prescriptive. In order to be deemed tolerant, participants had to fulfill two conditions: Their rejection rates in T2 had to be equal to or higher than 8; the difference between their rejection rates in T1 and T2 had to be at least equal to 6. 3 When subjects had completed the tasks, I asked them to comment on their corrections or changes, thus collecting both quantitative and qualitative data. Their comments helped identify the criteria they used to reject or accept lexical particularities and thereby defined the limits of an emerging local norm. Results indicated that acceptance of the particularities in a newspaper targeting a Haitian readership varied according to participants’ place of residence (Haiti or the diaspora), the type of particularities (creolisms as opposed to all other types), and the semantic domain of the particularities. For further information, see Etienne (2005). In the last part of the study, I interviewed all participants. I followed a semi-structured questionnaire with five open-ended questions that left space for digressions and anecdotes. These questions were: • What is ‘good French’ for you? • What is ‘good Creole’ for you? • Is there such a thing as Haitian French? 4 • What is Haitian French? • Do you like the particular features of Haitian French? Do you use them?

3

For more explanations about these criteria, see Etienne (2000, 2005). When I refer to Haitian French, I imply that this is a variety of French recognized as such by its users. In other contexts, I refer to French in Haiti. 4

182


My objective was to compare what participants stated about French and Creole in general with their editing behaviors in T1 and T2. My hypothesis was that the participants who had shown themselves to be tolerant of lexical particularities in the newspaper targeting a Haitian readership, and therefore of a local norm, would recognize the existence of a specific variety of French in Haiti. 1.2 Research questions In this article, I analyze the answers to these five questions, focusing on 20 participants that reside in Haiti. I address the following questions: • What is the relationship between the local uses (lexical particularities) speakers have identified and reacted to in the newspaper targeting a Haitian readership and speakers’ perception of French in Haiti? In other words, does participants’ legitimation of a local norm mean that they have appropriated a local variety of French? • How do these speakers refer to and define French in Haiti? • How do these speakers define their linguistic identity as they talk about their languages? 1.3 Overview of the current sociolinguistic situation in Haiti In Haiti, four languages—Creole, French, English, and, to a lesser extent, Spanish—coexist at different levels, with different values and functions. Although Creole has been the native language of all Haitians, regardless of social class, since the colonial period, French was the only official language in Haiti until 1987. In 1987, after the departure into exile of president Jean-Claude Duvalier, a new constitution recognized Creole as an official language of the Haitian nation, along with French. French has a high status, which is not matched by a corresponding high use in oral communication. The introduction of Creole into domains where it was formerly excluded, its presence in nearly all oral communication contexts, and the influence of English have had an impact on the nature of the French used by Haitians in Haiti and in the diaspora communities in the United States. The written press, which has traditionally been a means of expression for and has targeted the elite who speak French, has not remained impervious to such influences. The political changes that have been taking place since 1986 have given the population wider access to expression in the media, regardless of their degree of competence in French or Creole. This article explores how this access, in turn, has influenced language attitudes and social categorization by language and how social barriers created and maintained in part through language differences have been shaken.

FRENCH IN HAITI

183

2. The study 2.1 Participants Only 10 to 15% of the population in Haiti (depending on the source of the statistics) is proficient in both French and Creole. These Haitian Francophones live mainly in Port-au-Prince and its residential suburbs. Since the early days of the Haitian Republic, French has been linked to power, wealth, and education. The French/Creole bilingual speakers in Haiti have always been members of the elite: They are in high positions; they have completed secondary education and often hold advanced college degrees. As such, they continue to play an essential role in the definition of public policy in all domains. Zéphir’s extensive work (1990, 1996, 1997) on the French/Creole bilingual community documents language uses and attitudes in Haiti. The 20 French/Creole bilingual Haitians who participated in the study were selected on the basis of a background questionnaire that indicated they were representative of the Haitian French/Creole bilingual elite. The criteria used were age, nationality, family and language background, language uses and proficiency, and education. 2.2 Epilinguistic interview As mentioned above, I asked participants about their ideal norms in French and Creole and their perceptions of French in Haiti. My goal was to have participants express themselves as freely as possible and to foster an interaction that would lead them to reflect on the languages they use in their everyday life. Like Dumont and Maurer (1995:103), I wanted to “capture speakers’ spontaneous linguistic representations, particularly the ways they articulated in discourse their various linguistic conflicts.” During the interview, I let participants choose the language (French, Creole, or English) they wanted to use at any given time. None of them used English. They alternated their uses of Creole and French and code-switched throughout the interview. As shown by Zéphir (1990), this is common among bilingual Haitians. Free use of participants’ languages throughout the interview was definitely an advantage, since I am a French speaker from France (the former colonial power in Haiti) and conducted the interviews. 3. ‘Good French’ and ‘Good Creole’ 3.1 What is ‘bon français’? When I asked participants to edit press excerpts as if they were the editorsin-chief of the newspapers, I avoided referring to ‘correct French’ or ‘good French’ in the instructions to prevent participants from thinking that I was evaluating their command of French. I repeatedly stressed that the first newspaper they looked at targeted a Haitian readership and the second a wider Francophone readership. However, in order to explore their linguistic represen-

184


tations, I thought it was essential to find out whether their notion of ‘good French’ could accommodate their norm of French in the Haitian press as it emerged in the editing process. Hence this question about ‘bon français’. When defining ‘good French,’ participants emphasized four qualities: French from France, the observance of rules, the absence of foreign elements, and clarity of expression. Most participants defined good French as ‘the French of the French’ or ‘French in France.’ Participant 7 (P7) said: Mais le bon français, c’est le français de la France, c’est le français des Français! (rires). Y a des normes d’écriture, les structures se construisent de telle ou telle façon, de même pour le créole. “But good French, this is French of France, French of the French! (Laughs). There are some rules when you write, structures are put together in a certain way, same for Creole.”

Although participants recognized that French is dynamic, most of them mentioned that change originates from France and should be overseen by relevant authorities. As P20 explained, Se franse ki soti nan gramè franse la e nivo vokabilè, plis pre diksyonè ki soti an Frans yo, règ yo e tipikman pou mwen mèm bon franse li pratikman soti an Frans. Sa vle di m pa we okenn posibilite evolisyon franse a andeyò Lafrans. “This is French as defined by French grammar. As for the vocabulary, it should be closer to that contained in dictionaries published in France, rules, and for me, good French comes from France. That means that I don’t see any possibility of French changing outside France.”

These comments indicate that participants perceive French as a borrowed language that does not belong to Haitians. 5 Participants emphasized the importance of and need for rules, principles, and standards that users of good French should observe. As P16 said, Le français, c’est la langue des Français, il y a certaines règles qui existent, on n’écrit pas n’importe comment en français; pour toutes les langues il y a des règles et je pense que les règles doivent être respectées. “French is the language of the French, there are certain rules, you cannot write whatever you want in French; for every language, there are rules, and those rules have to be observed.”

P1 echoed this concern:

5 Valdman (1984:82,84) reported on a study of language attitudes among monolingual Haitians from a rural area. Participants emphasized how alienated they felt from French: Franse se pa lang pa ou, se lang achte, “French is not our language, it is a bought tongue.” They considered that French was “the relic of an externally imposed social order.” My bilingual participants’ comments agree with these comments.

FRENCH IN HAITI

185

Le bon français: essayer de respecter les règles parlées et écrites. “Good French: trying to abide by the rules of spoken and written French.”

This emphasis on rules is not surprising since the French instruction these Haitians have received has revolved around the mastery of rules and the knowledge of French literature (Haitians are particularly good at quoting from French literature). Implicit in P16’s and P1’s comments is the sense of not feeling free to express oneself in French and of trying to apply rules but not always being successful. I will return to this aspect later as one factor that explains participants’ perceptions of French in Haiti. It appears that participants’ definition of good French is not linked to anything close to Haitians’ real use of language: French is defined and regulated elsewhere. Participants did not seem to believe that they could appropriate French and judge it according to their own standards. In contrast to most participants who immediately referred to France and the French, two participants (P15 and P13), had a more pragmatic, common-sense view of good French. They said that French is good when it is clear and simple. Le bon français, ça veut être concis et très explicite, le gars qui lit n’a pas besoin de se casser la tête pour comprendre. (P15) “Good French should be concise and very explicit, the guy who reads that should not need to make any effort to understand.” Le bon français, c’est difficile à dire. Quelqu’un qui parle bien le français, pour moi, d’abord c’est quelqu’un qui dit les choses simplement sans vouloir imiter les Français, parce qu’il y aussi cet aspect, parler français correctement, c’est parler comme les Français, c’est pas forcément une référence, c’est pas parler non plus comme dans les livres. (P13) “Good French, it is difficult to say. Someone who speaks French well, for me, first, this is someone who says things in a simple way, without trying to imitate the French, because there is also this aspect, speaking French correctly, this is speaking like the French, it is not necessarily a reference, nor does it mean speaking like a book.”

This criterion, clarity, is a reaction to what people in Haiti over the past two centuries have considered good French. Too often, what people wrote or said was not as important as the manner and words they used to say it. The more convoluted and elegant French sounded, the better the speaker was considered to be. More than a tool for communication, French was an instrument for indicating status and for proving how educated one was. The creation of opposition journals in the diaspora before the fall of the Duvalier dictatorship might have contributed to a new awareness and function of French. Instead of obfuscating thoughts and ideas through the use of bombastic words and intricate syntax similar to the language used in the books they read in school, writers

186


sought to do the opposite, that is, to communicate as directly as possible. Some people might achieve this clarity by using Creole words that everybody understands. P7 illustrated this idea by advocating the use of some creolisms in French: Si vous écrivez ‘bandits armés’ au lieu de ‘zenglendo,’ la minorité qui parle français vous comprend, mais si vous voulez que plus de gens vous comprennent, c’est ‘zenglendo’ qu’il faut utiliser, parce qu’on l’utilise en français et en créole. “If you write ‘armed bandits’ instead of ‘zenglendo,’ the French-speaking minority will understand you, but if you want more people to understand, you should use ‘zenglendo,’ because it is used in French and Creole.”

However, in participants’ comments, this emphasis on clarity and immediacy of communication clashed with a criterion of purity in good French. All participants cited purity as a guarantee of quality. They all maintained that good French should not be mixed with any other language. Comments echoing this concern abounded, even in the words of those participants who were the most adamant when they defended clarity: Pour moi, parler français correctement, c’est dire les choses en français dans la mesure du possible, c’est pas faire la moitié d’une phrase en anglais, l’autre moitié en créole...Pour moi, lorsqu’un Français parle et qu’il y a un abus de termes anglais, pour moi c’est pas parler français correctement, pour moi c’est une faute, parce qu’il y a moyen de dire ça autrement. (P13) “For me, to speak French properly, this is to say things in French to the extent it’s possible, this is not to make half a sentence in English, the other half in Creole.... For me, when a French person speaks and there are too many English words, for me, this is not speaking French properly, for me, this is wrong because there is another way to say that.”

Likewise, although P11 conceded that French had to evolve and that he was not opposed to using foreign words in order to name some technological innovations using foreign words, he stressed that: C’est dépouillé de tout anglicisme, de tout créolisme…c’est dépouillé de tout mot étranger. “This is stripped of any anglicism, any creolisms…stripped of any foreign word.”

Participants’ belief that foreign elements should be eliminated was often linked to their recognition of Creole’s influence. P2 said: Moi, je veux bien qu’on fasse l’essor du créole, qu’on parle créole, mais pas quelque chose entre les deux. “As for me, I don’t mind if we work for promoting Creole, if we speak Creole, but not something in between.”

From these comments, it seems that the proximity of Creole to French in Haiti (both lexical and functional) has exacerbated bilingual participants’ concern

FRENCH IN HAITI

187

about the ‘risk’ of mixing languages. In an ideal norm, they aspire to find a purity they do not always find in their own French. These criteria, except for the clarity criterion, define a norm, or ideal, of French that appears rigid, bookish, and ‘untouchable.’ Participants put French, as used by the French in France, on a pedestal and do not associate it with everyday communication, contacts, and influences. This is evident when participants insist on purity and lack of foreign influence. 3.2 What is ‘Bon Kreyòl’? Although this question may appear irrelevant to exploring the perceptions Haitians have of French in Haiti, participants’ answers to it shed light on how their perceptions of French and Creole were at once inextricably tied to one another and yet diametrically opposed. Two main trends appeared. First, some of the participants mentioned criteria similar to the ones they gave in defining good French (clarity and purity). For them, good speakers of Creole have the ability to make themselves understood in all contexts with all types of Creole speakers. Unlike French, which always has a limited audience in Haiti, Creole is perceived as the instrument meant for national unity and communication. Creole is seen to be all encompassing. Other participants insisted on the purity of the language—that is, the absence of French expressions in Creole. For several participants, the ideal speaker of Creole is the peasant who does not know any French. Several participants did not consider themselves good Creole speakers because their Creole is too influenced by their knowledge of French. In response to the question, P20 said: Yon kreyòl ki ta eseye ale pli pwofon nan ekspresyon, non vokabilè, pli pre bagay paysannerie yo, se limèm ki t ap yo pli bon kreyòl. 6 “A Creole that should strive to go deep in expressing, in vocabulary, closer to things related to peasants, this is what good Creole is.”

However, more than half the participants voiced their discomfort and their surprise when pondering this question about good Creole. Several of them realized they could not answer the question. Some of them claimed that the geographical variations observed in Creole, as well as the lack of rules and

6 Letters and words in boldface type highlight cases of codeswitching between French and Creole as well as some of the French features found in the Creole of bilingual Haitians. On the phonological level, mesolectal Creole spoken by urban bilingual Haitians has three additional vowels. However, these vowels may not be totally absent in basilectal Creole spoken by monolingual Haitians. These three additional vowels are u [y] often used instead of i [i], eu [ø] instead of é [e], and eù [œ] instead of è [ε]. Sometimes, bilingual Haitians inflect adjectives in Creole for gender although basilectal Creole speakers do not.

188


official standards, did not allow them to define good Creole. 7 P18 and P19 said: Dram kreyòl la, li pa toutafè yon lang paskeu pa genyen diksyonè, pa genyen gramè, chak moun ekri à sa façon. (P18) “The tragedy with Creole is that it is not really a language because it does not have a dictionary, everybody writes the way they want.” Je n’ai pas prêté attention à la syntaxe créole. Je la connais cette syntaxe mais je n’ai pas accordé d’importance à cette syntaxe, même si je parle la langue parce que dans le pays, on n’a pas fait de recherche en stylistique créole, donc je ne peux pas apprécier un texte créole. (P19) “I have not paid attention to the Creole syntax. I know this syntax, but it has not mattered to me, although I speak the language, because, in Haiti, nobody has done any study about style in Creole, so I can’t appreciate a Creole text.”

Other participants were perplexed by the question because they did not study Creole at school and did not, therefore, have any normative perspectives. Kreyòl ou pa bezwen apran l … je ne vois pas comment je peux définir ça. (P1) “You don’t need to study Creole … I don’t see how I could define that.”

P5 explained her inability to answer the question: Il y a cet aspect de liberté totale qu’on a l’impression d’avoir en créole et qui empêche d’avoir des inhibitions. “There is this sense of total freedom that we have the impression of having in Creole and that prevents any inhibition.”

Her comment nicely synthesized the main contrast between linguistic representations about French and Creole. Participants alternate between two linguistic universes, one where there are so many rules and constraints that perfection or adequacy seems unattainable, and one where there appear to be no rules and no judgments. Vernet (in preparation) aptly explains: “For Haitians, with the exception of a few avant-garde intellectuals, Creole has always been this nonlanguage that everybody speaks without knowing it.” 4. What is Haitian French? Table 1 presents participants’ answers to the questionnaire regarding their recognition of a Haitian variety of French (Haitian French?) and their personal use of that variety (Personal use?). In order to compare these answers with their acceptance or non-acceptance of a Haitian local norm, I report, in columns 2 and 3, on each participant’s results during previous stages of the study. In the column entitled Editing behaviors, I give each participant’s scores on the 7 Many bilingual speakers seem to be unaware that spelling in Creole is governed by rules and standards officially approved in 1979.

FRENCH IN HAITI

189

epilinguistic tasks (rejection rate of Haitian lexical particularities in a newspaper targeting a Haitian audience only and in a newspaper targeting a broader Francophone readership). In the third column (Local norm?), “yes” indicates that a participant tolerated lexical particularities in the newspaper targeting a Haitian readership. As mentioned above, participants were deemed tolerant, and thus accepting of a local norm, if they satisfied two conditions: (a) their rejection rate in T2 was equal to or higher than 8 and (b) the difference between the rejection rate in T1 and T2 was at least equal to 6. In addition, in the last column, I report whether participants said that they used Haitian French in writing, because they all emphasized the distinction between oral and written French. Table 1 shows that, although 13 out of 20 participants accepted a local norm in a Haitian newspaper, only 10 participants recognized the existence of Haitian French, only four admitted to using it personally, and only three believed that it can be used in writing. In addition, an examination of the results in Table 1 reveals several levels of ambivalence or inconsistency in participants’ answers. I expected participants with a high rejection rate in both Tasks not to recognize the existence of Haitian French. I expected participants with a low rejection rate in Task 1 (≤ 3) to recognize the existence of Haitian French. These two hypotheses were not confirmed by the results. I will now examine and analyze the reasons why participants rejected the notion of Haitian French, how participants defined it when they recognized its existence, and how participants’ inconsistencies in their epilinguistic discourse shed light on their perception of French in Haiti. 4.1 “No such thing as Haitian French” For various reasons, half the participants (n = 10) believed that there is no Haitian French. Some of them think that there are not enough differences between French in Haiti and French elsewhere to identify a specific variety. They mentioned the existence of a different accent and a few expressions dictated by context. They referred to French in Haiti as français haïtianisé (“Haitianized French”), but they did not think that it should be considered a variety of French. To them, there is no desire or need to consider French in Haiti as a specific variety. P17 said: Oui, il y a un parler haïtien, de la même manière qu’il y a un parler québécois. Il n’a pas nécessairement sa place dans la presse haïtienne. Il n’y a pas une véritable étiquette de français haïtien parce qu’en fait, quand on parle français, c’est comme si on entendait n’importe quel francophone parler français. […] Il y a des expressions, des mots haïtiens, mi-francisés, mi-créoles qui sont admis dans la conversation courante, dans le parler haïtien. “Yes, there is a Haitian way of speaking, just like there is a Quebec way of speaking. There is not necessarily room for it in the Haitian press. There is no real label of Hai-


190

tian French because, actually, when we speak French, it is as if we heard any other Francophone speak French. […] There are some phrases, some Haitian words, halfgallicized, half-Creole that are tolerated in everyday conversation, in the Haitian way of speaking.”

These participants saw evidence of this so-called Haitianized French in oral discourse only. They categorically denied its manifestations in writing, except as a specific choice to give local color or to create a specific speech style. P20 Participants N = 20 P1

Editing behaviors (7,15)

Local norm? Yes

Haitian French? Yes

Personal use? No

Written use? No

P2

(2,8)

Yes

No

No

No

P3

(12,16)

No

No

No

No

P4

(12,17)

No

Yes

Yes

No

P5

(8,15)

Yes

Yes

No

No

P6

(3,9)

Yes

Yes

Yes

Yes

P7

(0,8)

Yes

Yes

Yes

Yes

P8

(1,16)

Yes

No

No

No

P9

(1,10)

Yes

Yes

Yes

Yes

P10

(9,12)

No

No

No

No

P11

(0,12)

Yes

No

No

No

P12

(2,11)

Yes

No

No

No

P13

(9,16)

Yes

Yes

No

No

P14

(0,10)

Yes

No

No

No

P15

(5,9)

No

Yes

No

No

P16

(11,10)

No

Yes

No

No

P17

(0,9)

Yes

No

No

No

P18

(3,12)

Yes

No

No

No

P 19

(13,12)

No

Yes

No

No

P20

(9,7)

No

No

No

No

Yes = 13 No = 7

Yes = 10 No = 10

Yes = 4 No = 16

Yes = 3 No = 18

TOTAL

Note.

Boldfaced font = participants consistent across tasks and responses Italics = participants recognizing Haitian French

Table 1: Participants’ perceptions: The reality of Haitian French

FRENCH IN HAITI

191

mentioned exceptional written use of Haitianized French in propaganda documents, because it is more apt to tap into the Haitian imagination. Other participants categorically denied the existence of a Haitian variety of French. They attributed the differences between French in Haiti and their idea of standard French to unconscious, erroneous use by non-proficient speakers who have not completed their high school studies in good schools. 8 These participants claimed that the different ways of speaking French in Haiti correspond to different proficiency levels. They mentioned that, in general, speakers are not aware of the quality of the French they speak. They believe that at a high level of proficiency, there is no difference between French spoken in Haiti and French spoken in France, except for an accent. P1 and P18 said that Haitians are stricter about the rules than other Francophones and speak polished and refined eighteenth-century French. Other participants were less categorical and left the door open for a possible evolution of a Haitian variety of French. However, they did not want to legitimize Haitianized French or make it official, because they feared it would lead to mediocrity. P3 said: M pa di non de manière catégorique. M panse yon lang ap toujou anrichi tèt li à partir de tout yon seri de …eleman realite de chak pèp mais mwen mèm m pa vle itilize le terme français haïtien paskeu an Ayiti m a viv yon realite ke lè moun di franse ayisyen, yo vin konpran yo kap di nenpòt kèl bagay, e lè ou di franse ayisyen, se kòm si ce serait synonyme de médiocrité… “I am not saying no categorically. I think a language may always become richer through the addition of elements, from the reality of each people but I don’t want to use the term Haitian French because Haiti is such a place that when you tell people Haitian French, they are going to understand that anything goes and when you say Haitian French, it is going to mean mediocrity.”

It is interesting to look at what these participants did in the epilinguistic tasks. Of 10 participants who did not recognize the existence of Haitian French, only three appeared to be consistent. P3, P10, and P20 (highlighted in boldface type in Table 1) recorded higher than average rejection rates in both editing tasks, and stated that they did not recognize any Haitian variety of French. This statement was logical given their rejection of Haitian lexical particularities in the newspaper targeting a Haitian readership. Their prescriptive behaviors in the editing tasks fully coincided with their views on Haitian French, whose use they considered inappropriate and not needed whatsoever. On the contrary, as shown in Table 1, the remaining seven participants who did not recognize any Haitian variety of French recorded very low rejection 8

In Haiti, the private sector administers 80% of the schools. Catholic schools created in the nineteenth century after the Haitian government signed a concordat with the Vatican are the most prestigious schools and still have considerable influence today. Except for these schools, private schools in general are of very uneven quality.

192


rates in Task 1. This meant that although they tolerated Haitian lexical particularities in the Haitian press, they did not think the particularities warranted recognition of a Haitian variety of French (as indicated in their epilinguistic comments above). P14 is one of these inconsistent participants who tolerate a local norm in the press but deny the existence of Haitian French. He illustrated and gave rationale for this contradiction when answering the question about the existence of Haitian French: C’est fort mais il y a un français très haïtianisé, ça veut dire qu’on emploie des termes qui ne sont pas toujours réellement français, dans le langage courant. L’Haïtien en général, est très strict, il voudrait rester très français. Mais je ne dirais pas qu’il y a un français haitien. “This is strong, but there is a Haitianized French, this means that in everyday discourse we use terms that are not really French. Haitians, in general, want to remain very French. But I would not say there is a Haitian French.”

This participant clearly recognized that there are practices that violate the French norm. However, he hastened to pit these oral, current practices against the desire, principles, and aspirations for remaining very French. There seemed to be an impossible reconciliation of these practices and ideals. Note that P14 said that Haitians would like to remain French in their linguistic usage, suggesting that they don’t manage to remain French, in reality. Thus, his conclusion as to the non-existence of Haitian French is manifest in the impossibility for him to recognize and legitimate practices in his epilinguistic discourse. He further supported his position by explaining how, for ‘people’ (other bilingual speakers from the elite), accepting Haitian French would be equivalent to admitting that one does not speak French. Even orally, P14 is reluctant to use Haitianized French and monitors himself: Les gens prennent ça en considération: il faut parler en bon français. On peut faire des fautes en espagnol, en anglais, il n’y a pas de problèmes, mais pour le français on est exigeant, c’est le fait d’être colonisés. C’est la raison pour laquelle je préfère parler en grec que parler en français. “People take that into account: You have to speak good French. You may make mistakes in Spanish, in English, there is no problem, but for French, we are very strict, this is because of colonization. This is why I prefer speaking in Greek to speaking in French.”

His lack of appropriation of the features of French in Haiti, whether orally or in writing, and his fear of being caught making a mistake, lead him to wish he could avoid speaking the language in which he conducts most of his professional and intellectual life.

FRENCH IN HAITI

193

4.2 Defining Haitian French As shown in Table 1, 10 participants (in italics) recognized the existence of Haitian French. While these participants believed there is a specific way to use French in Haiti, eight of them agreed that there should be a clear line between spoken and written French. They did not believe that Haitian French is and should be used in writing. One of the most frequent reasons participants gave for not putting Haitian French in writing was their fear of being judged and of committing an error. To a lesser extent, this echoed P14’s reasons for not recognizing Haitian French. The words or phrases “error,” “mistake,” and “I would not dare to” often appeared in participants’ arguments and explanations. This showed that participants are still influenced by the way they learned French. The French they know is French in grammar books. They do not feel that they have permission to appropriate French. The more closely they adhere to the rules, the higher their proficiency in French is considered. In this regard, P5 was particularly eloquent: Moi, je n’ai pas la sensation de m’approprier le français comme étant ma langue parce que j’ai toujours peur de faire des erreurs, j’ai toujours peur de me tromper... J’ai vécu trois ans à Montréal. A Montréal, j’étais plus à l’aise parce que c’était un autre accent, c’était d’autres expressions disons, j’apprenais autre chose. Je me suis adaptée. A la rigueur, j’étais plus à l’aise à Montréal que je ne le suis quand je parle français en France. “I don’t have the feeling that I have appropriated French as my language because I am always afraid of making mistakes, I am always afraid of erring.... I lived for three years in Montreal. In Montreal, I felt more comfortable because the accent was different, idioms were different, you know, I was learning something different. I adapted. Actually, I was more comfortable in Montreal than I am when I speak French in France.”

When defining Haitian French, all these participants agreed that French used in Haiti is characterized by the introduction of words and phrases that are typically Haitian and borrowed from Creole. The influence of Creole and the use of creolisms were perceived as determining and inevitable elements. Participants emphasized that, in other respects, French in Haiti does not differ from French used anywhere else: Grammar rules (choice of tense, agreement, and morphology) are the same as in RF. Participants’ statements varied along a continuum. Participants at one end of the continuum equated Haitian French with creolized French. Participants at the other end of the continuum said that Haitian French is punctuated by selected creolisms. P6 and P19 illustrated these two trends: Le français haïtien, c’est le français bourré de créolismes, paskeu ayisyen, lè yo pale franse o lè yo ekri franse, à part quelques grands écrivains qui écrivent dans un français châtié, en général ayisyen an, se yon franse pale li ekri, se yon franse

194


Dayiti li ekri e li pale tou, ou kap wè yon diferans, au niveau lexical, au niveau accent. Li genyen yon originalite keu nou mete l nan lang lan. (P6) “Haitian French, this is French full of creolisms, because Haitians, when they speak French or when they write French, except for a few great writers who write in perfect French, in general what Haitians write is the French they speak, this is Haitian French that Haitians write and speak, you can see the difference at the lexical level, at the level of pronunciation. There is some originality that we are able to put in the language.” Le français haïtien, c’est du vocabulaire, des concepts qu’on accepte parce qu’on vit avec ça. (P19) “Haitian French, this is vocabulary, concepts that we accept because we live with them.”

Most participants, while recognizing the influence of Creole, voiced concerns about an excessive use of Creole in French. They believed that, because Haitian French is closely related to Creole, the boundary between Haitian French and bad French is easy to cross. For most participants, the main problem with Haitian French is its current lack of norms. Worth remembering is that the observance of rules was one of participants’ most important criteria when they defined good French. When describing Haitian French, most participants used derogatory terms. They often used the word mélange, conjuring up an image of Haitian French as a confused mix between Creole and French. Even though participants accepted this mix as part of the Haitian reality, they rarely valued it. Accepting Haitian French as a mélange clashed with participants’ criterion of purity when defining good French. As P5 explains: Je préfère parler créole ou parler français mais autant que possible un français qui soit correct. Je ne dis pas que je réussis. C’est vrai que j’ai tendance à me surveiller, ne pas trop mélanger. “I prefer speaking Creole or speaking French but as much as possible, a French that is correct. I am not saying I succeed. It’s true that I tend to monitor myself, to not mix too much.”

Many participants’ comments about Haitian French indicated that, above all, Haitian French was not as good as French used in France: Quand vous prenez n’importe quel journal haïtien et que vous prenez un journal français, la différence est claire, vous voyez qu’il y a une différence de niveau. “When you take any Haitian newspaper and a French newspaper, the difference is clear, you can see a difference in quality.”

Several participants indicated that Haitian French, as seen in the press, is a recent and unofficial phenomenon. The mélange has been occurring more often since 1986. After the departure into exile of president Jean-Claude Duvalier, there was a liberalization of public expression. Increasing numbers of people

FRENCH IN HAITI

195

from diverse educational backgrounds have gained access to the media. The urge to communicate ideas that had been under censorship for years led to less focus on the form than on the content of the written pieces. Other participants mentioned the need for an official recognition of Haitian French by an institution. They felt they needed official permission to relax the rules of French. Only then would they feel free to do so spontaneously. Although 10 participants recognized the existence of a Haitian variety of French, Table 1 points to several inconsistencies in their linguistic representations. As in the other group, only three participants—P6, P7, and P9— appeared to be fully consistent (in italics and boldface type in Table 1). They tolerated Haitian particularities in the newspaper targeting a Haitian readership, they recognized the existence of a Haitian variety, and they stated that they use this variety. Participant P6 said: Ma vision concernant kesyon kreyolis nan franse a m kwe se yon bagay obsève et peut-être que en fin de compte, c’est carrément yon enrichissement pour la langue e se de bagay à considerer paskeu se manifestasyon yon popilasyon par rapport à un zouti ki se yon lang. Donk se de bagay pou ou respekte. “My opinion on creolisms in French, I think that is something that has been observed, and maybe, eventually, it is really enriching language and it matters because this is the manifestation of one population with regard to a tool which is language. This is something worth respecting.”

The remaining 7 participants appeared to be inconsistent to a lesser or greater extent. Four of them (P4, P15, P16, and P19) were the most prescriptive speakers in the editing tasks. Their rejection rates indicated that they rejected a local norm in the Haitian press. However, despite their intolerance of a Haitian norm in the press, they recognized the existence of a Haitian variety. As shown in Table 1, P15, P16, and P19 resolved that contradiction by claiming not to use this variety, either orally or in writing. Participant P4 did not want to see or use this variety in writing. However, she recognized using Haitian French in oral interactions: P4: Oui, je crois que c’est un langage qui s’implante. Je crois qu’il y a plein de gens médiocres ou des gens qui ne devraient pas écrire qui, sous cette couverture, se lancent à écrire... “Yes, I think that this is a language that is gaining ground. I think that there are plenty of mediocre people or people who should not write who, by this token, set up to write in the press...” CE: Est-ce que tu considères qu’il existe un français haïtien? “Do you think that there is a Haitian French?” P4: Oui, il est ponctué de mots créoles. “Yes, it is punctuated with Creole words.” CE: Tu te reconnais dans ce français, tu parles français haïtien? “Do you recognize yourself in this French, do you speak Haitian French?”


196

P4: Oui, je me surveille un tout petit peu mais pas beaucoup. “Yes, I monitor a little bit but not much.”

P4 distinguished her Haitian French in oral interactions from the Haitian French as it appeared in the Haitian press, because she believed that everyday language is not marked by the same creolisms as political texts. Although P4 appropriated the particular features of French in Haiti in oral interactions, her reluctance to use them in writing echoed the concern for errors voiced by participants who did not recognize Haitian French at all. She confessed: Moi, je vais te dire, honnêtement parlant, quand je suis partie de la maison, je n’écrivais pas et je n’écrivais pas à cause de ça parce que je savais que mes soeurs allaient rire de l’écriture. Je n’écrivais pas de lettres. […] Je suis parfaitement d’accord qu’on se sente libéré et que je parle un français haïtien. […] je peux même maintenant l’écrire dans une lettre mais je prendrais peut-être la peine de mettre des guillemets, […] juste pour signaler que je sais que c’est pas proprement dit mais c’est ce que je veux dire. “I am going to be frank, when I went away from home, I did not write to my family and I did not write because of that, because I knew that my sisters were going to laugh at my written French. I did not write any letters. […] I fully agree that we should feel liberated and that I speak a French that is Haitian. Now I can even use it in a letter but I would pay attention to use quotation marks, just to let the reader know that I know it is not properly said but that this is what I want to say.”

The remaining three participants (P1, P5, and P13) were tolerant of Haitian particularities and thus of a norm in the Haitian press, and they recognized the existence of a Haitian variety. However, they also claimed not to use this variety. 5.

Conclusion In summary, my study has brought to light various types of conflicts in participants’ linguistic representations: • • • •

Participants may tolerate Haitian lexical particularities in the press, but they may not recognize French in Haiti as a Haitian variety of French. Participants may not tolerate Haitian lexical particularities in the press, but they may recognize that there is a Haitian variety of French. Participants may tolerate Haitian lexical particularities in the press and recognize that there is a Haitian variety of French. However, they may oppose its use in written contexts. Participants may recognize that there is a Haitian variety of French, but they may claim not to use it themselves, either orally or in writing.

I return to the three questions posed in the introduction. First, what is the relationship between the local uses (lexical particularities) participants identified

FRENCH IN HAITI

197

and reacted to in the newspaper targeting a Haitian readership and their perception of French in Haiti? The relationship is certainly more complex than my original hypothesis. I thought that tolerance of local uses would signal recognition of a local variety of French and vice-versa. However, participants’ editing behaviors were not a good indicator of their feelings about French in Haiti. There is a need to distinguish the tolerance of a local Haitian norm from the recognition of a Haitian variety of French. Editing tasks constitute a good tool to test the semantic and sociolinguistic characteristics of specific lexical items or syntactic structures in a particular context (see Etienne 2005). However, only by combining editing tasks with epilinguistic interviews is it possible to examine if some accepted uses are legitimized and appropriated. In this study, epilinguistic interviews, in combination with editing tasks, made participants aware of the factors that drive their linguistic choices. In the course of the interviews, several participants analyzed their own reactions to the question of Haitian French. They found reasons for these reactions, realizing that, to some extent, Haiti’s colonial past still defines their linguistic identity or that they are victims of a language education that overemphasizes academic uses and rules. In the course of the interviews, others became aware of their own discomfort as users of French, obsessed by the possibility that they may make mistakes and incur others’ judgments. The data resulting from the epilinguistic interviews revealed numerous conflicts among linguistic representations. The data also revealed conflicts between linguistic behaviors and linguistic representations: conflict between uses marked by French-Creole contact and French and Creole ideal norms defined by purity, conflict between the need for French and discomfort when using it, conflict between functionality of French in writing and lack of appropriation, conflict between ‘permissive and free’ Creole and regimented French, and conflict between recent linguistic turmoil and speakers’ desire for structure and rules. Regarding the second question about the definition of Haitian French, half the participants recognized its existence, but this recognition was accompanied by lukewarm and even negative feelings. Instead of emphasizing Haitian French’s originality, participants focused on how it lacked many of the attributes of ‘good French.’ Three elements contributed to participants’ definition of French in Haiti and accounted for their generally negative attitudes: their reluctance to use it in writing, their fear of being judged when they use it, and the influence of Creole. In general, participants defined Haitian French as creolized French and mentioned that its use was very natural and inevitable in conversation. While stressing how unavoidable its use was, they implied or insisted that they wished they could avoid using it, emphasizing the conflict between uses and ideal uses.

198


More than anything, participants could not reconcile the use of creolized or Haitianized French with their ideal of linguistic purity. Purity, both in French and Creole, was a recurrent theme. Participants’ somewhat naïve idealization of the peasants’ way of speaking Creole corresponds to a search for authenticity. In spite of a linguistic reality in which Creole and French are in constant contact, participants would like to keep the two languages separate. In their eyes, mixing Creole and French disparages both languages. On the one hand, participants saw contacts between French and Creole as a constant threat to their French, to who they are socially, and to how they feel they are expected to use French. On the other hand, participants felt that when creolized French was used, Creole was absorbed, distorted, and denied its own identity as a language by French. The tension between the reality of these unavoidable contacts and participants’ linguistic aspirations (or illusions, if I may be so bold to say) is even more acute because, in general, their definition of good French is fantasized, as if there were some “international French” (as labeled by one of the participants) in which no influence of any other language is present. As for my third question related to the definition of participants’ linguistic identity, 13 out of 20 participants (73%) recognized either a local use of French in the Haitian press or a Haitian variety of French, but they claimed not to use it or to monitor themselves in order not to use it. Zéphir (1990:244) describes this ambivalence in bilingual speakers’ conflicting attitudes when she explains that the speakers apply two standards: their own individual standard and ‘others’’ standard: ... there is a double-standard at play: one that is applied to individual speakers and one that is applied to ‘others’ or to the Haitian nation as a whole. . . . The Haitian bilingual speakers can be conceived as being two different entities. One is the individual who does not belong to the nation, and the other is a member of this nation.

Participants’ refusal to use Haitian French and their efforts not to use it might be a way to preserve their individual linguistic identity. It might be an attempt to oppose the linguistic changes that have taken place in Haiti during the last 20 years, changes that have created a new linguistic national identity in which Creole has more weight and value than before. All participants, as members of the French/Creole bilingual elite, have been schooled in very linguistically purist environments. For most of them, the use of Creole at school was prohibited and for all, a good mastery of French has been the key to professional and social success. Consequently, they are at a loss when it is time to judge and appropriate French as spoken or used in Haiti. Since the departure into exile of Haitian life-president Jean-Claude Duvalier almost 20 years ago, increasing numbers of people have gained access to public expression and to the written media. The Haitian press displays a French

FRENCH IN HAITI

199

lexicon in transition that has become more and more permeated with Creole. The growing influence of Creole, the freedom and immediacy of communication that Creole or the use of creolisms provides, clash with an ideal norm of French that is distant, bookish, and unattainable. Participants recognized linguistic changes, but they do not want to apply them to their own use of French. They do not seem able to reconcile a permissive climate in writing to years of rigid instruction in French. Since French is inseparable from education in Haiti, participants may not want to be associated with less educated speakers who have not received the same education as they did. Even if, politically, they are open to promoting Creole in numerous domains and to changing the rules of the political system, they cannot suddenly forget the values on which their class identity is built. In view of all these conflicts and tensions, calling French in Haiti ‘Haitian French’ would be imposing an illusory cohesion on a very divided and unstable linguistic picture, where French/Creole bilingual speakers constantly face an existential and linguistic dilemma. For my participants, what makes French in Haiti Haitian—its contacts with Creole—is what makes it unacceptable as French. As P16 remarked: C’est certain qu’il y a un français haïtien parce qu’il y a beaucoup de phrases, il y a beaucoup d’expressions que nous avons créées, qui n’existent pas en français. Ce n’est pas du tout, pas du tout du français. “For sure there is a Haitian French because there are many phrases, there are many sentences that we have created that do not exist in French. This is not French, not at all French.”

References Baggioni, Daniel & Jean-Michel Kasbarian. 1993. “La production de l’identité dans les situations de francophonie en contact”. Le français dans l’espace francophone, Robillard, Beniamino, & Bavoux (eds) 1993 (Paris-Genève: Champion-Slatkine). 854-869. Dumont, Pierre & Bruno Maurer. 1995. Sociolinguistique du français en Afrique francophone. Paris: Edicef/AUPELF. Etienne, Corinne. 2000. A sociolinguistic study of the lexical particularities of French in Haiti. Unpublished doctoral dissertation, Indiana University, Bloomington. ----------. 2005. “The lexical particularities of French in the Haitian press: Readers’ perception and appropriation”. Journal of French Language Studies 15. 257-277. Robillard, Didier de. 1993a. “Le Concept de particularité lexicale”. Inventaire des usages de la francophonie: nomenclatures et méthodologies, Latin, Queffélec, & Tabi-Manga (eds) 1993 (Paris: John Libbey Eurotext). 113135.

200


----------. 1993b. Contribution à un inventaire des particularités lexicales du français de l’île Maurice. Vanves, France: EDICEF. Valdman, Albert. 1984. “The linguistic situation of Haiti”. Haiti–today and tomorrow, Foster, & Valdman (eds) 1984 (New York: University Press of America). 77-99. Vernet, P. (in preparation). “La situation linguistique d’Haïti”. Dictionnaire Encyclopédique d’Haïti (Projet DEH – version 04-08-98) (Port-au-Prince, Haiti: Centre d’Etudes et de Culture Haïtiennes). Zéphir, Flore. 1990. Language choice, language use, language attitudes of the Haitian bilingual community. Unpublished doctoral dissertation, Indiana University, Bloomington. ----------. 1996. Haitian immigrants in Black America: A sociological and sociolinguistic portrait. Westport, CT: Bergin & Garvey. ----------. 1997. “The social value of French for bilingual Haitian immigrants”. The French Review 70. 395-406.

SECTION THREE:

VARIATION

ALBERT VALDMAN ON THE DEVELOPMENT OF CREOLES *

Salikoko S. Mufwene University of Chicago

ABSTRACT: Valdman (1991) is rare in questioning, justifiably, Ferguson’s (1968) stipulation that Haiti represents a typical case of diglossia unique among creole communities, which suggests that Haiti had no continuum comparable to what DeCamp (1971) described for Jamaica. However, a continuum does exist and it may have arisen due to attempts by speakers of the basilect to produce the acrolect of their community, to accommodations made by the acrolectral speakers to speakers of the basilect, or to both. Plantation colonies, including Haiti, developed their creoles through a basilectalization process, which was not uniform. I pick up the pieces from here to argue that, although the vast majority of Haitians are considered monolingual creole speakers, Haitian Creole is still a continuum of lects. Valdman’s (1991) position represents only one of the later contributions to the maintenance of a linguistic continuum that undoubtedly obtained when Haitian Creole was recognized as a separate language variety from its French lexifier or its present acrolect, sometime in the (early) eighteenth century. 1. Introduction Professor Albert Valdman’s name has not figured prominently in the recent debate over the development of creoles, especially regarding whether creoles developed from antecedent pidgins and whether they developed by any restructuring processes that set them apart from other cases of language evolution in general. Equally little cited have been his positions on the nature and significance of the terminus a quo of these vernaculars and on the role of substrate influence in their emergence. As a scholar centrally engaged in the debate, I must admit my own embarrassment for not paying particular attention to Professor Valdman’s views on these issues. I write this essay in part to correct my oversight. I focus on Valdman (1977), where I find positions very akin to Chaudenson (1979, 1992, 2001), who has influenced me the most over the past decade (Mufwene 1996, 2001) and has been widely cited in the debate, but to whom I will refer only in passing mostly to support my discussion of Valdman’s insights. * I am grateful to Alison Irvine and Michel DeGraff for feedback on an earlier draft of this chapter. I stand alone to blame for all the remaining shortcomings.

204


The neglect of Valdman’s contribution to the debate on the development of creoles is unfortunately so pervasive that Valdman (1977) is rarely cited, not even by himself—for instance, in his general introduction to the articles in La créolisation: À chacun sa vérité, the special thematic issue of Etudes Créoles (25.1) devoted to the state of the art on the subject matter, and in his own article in the same issue (Valdman 2002), in which he defends basically the same position discussed in 1977. I have seen this work cited by Manessy (1995) but neither by Chaudenson (2001) nor by Corne (1999), though they cite other publications of his. The oversight is shocking because the position they advocate is closely related to, if not the same as, what Valdman (1977) submits, adducing supportive data from the determiner system of French creoles of both the American-Caribbean region (ACR) and the Indian Ocean region (IOR). Below, I focus both on Valdman (1977), also expanded in Valdman (1983), and on Valdman (2002), hoping to show why they must be read by anybody who wishes to contribute new ideas to our understanding of the development of French creoles in particular. Although he assumes in respects that do not necessarily undermine his strengths that creoles have evolved from erstwhile pidgins, Valdman (1977: 105) addresses two questions. First, he asks what is the terminus a quo of creoles; in other words, what are the specific varieties that were targeted and subsequently restructured by the non-native speakers in the relevant plantation settlement colonies? Answers to this question should help us determine the extent to which French creoles have diverged either from the colonial koines from which he claims that they developed (via antecedent pidgins) or from other non-creole vernaculars that have evolved from the same, or similar, koines, such as the French varieties of St. Barths, Louisiana, and Québec. Second, Valdman addresses the question of how creoles developed. I discuss these questions in the same order below. 2. The terminus a quo of creoles Valdman addresses the first question in an informative way. While he does not deny substrate influence (118f, discussed below), he identifies the usage of phrases such as c’te type-là “that guy/man” and c’te femme-là “that woman” (my translations) as “tournures populaires.” Valdman points to the nonstandard varieties undoubtedly spoken by the French colonists (given the proletarian origins of the early settlers) to whose speech the non-Europeans were exposed, as described by Chaudenson (1992, 2001). Valdman (2002:123, 133) is more explicit in identifying that target as the “koine” based on “le français populaire et régional” (his own terms)—that is, non-standard French rather than the often unjustifiably-invoked standard variety. It is thus to a variety that was itself emergent that the African slaves, who were shifting from their ancestral languages, were exposed.

THE DEVELOPMENT OF CREOLES

205

Just as the creoles’ structures were internally variable in their earlier stages (Valdman 2002:137), we can assume that the koines themselves were internally variable, owing to the diverse regional backgrounds of the colonists whose dialects were then being brought into contact. We can also assume that in the same way that the languages of the slaves were not represented with equal demographic strengths in the various ACR and IOR plantation settings, the competing dialects of French were not represented with equal demographic strengths in the different colonies. As observed by Mufwene (2001), competition and selection operated not only between the European and the nonEuropean languages, but also among variant features of the targeted European language and at the level of substrate influence among the features of the nonEuropean languages. There must thus be a number of reasons why creoles vary among themselves, as the linguistic ecologies of their emergence do not seem to have been identical. Among the relevant factors is the demographic strength in the representation of structural variants in the colonies. These same considerations also account for why non-creole varieties of the same languages that emerged among (descendants of) European colonists are not structurally identical. In fact, once race segregation is factored into this ecological matrix, one can use the same considerations to account for the divergence of creole from noncreole varieties. Creoles can be characterized as those vernaculars that developed under heavier influence from non-European substrate languages, as they are spoken primarily by populations that are of non-European descent (Mufwene 2001). This perspective on language evolution in European settlement colonies enables us to derive another interesting consequence of Valdman’s (1977) identification of the source of French creoles’ structures (my generalization) as the colonial koine or le français populaire. In discussions on the development of creoles, it is necessary to make a distinction between, on the one hand, ‘source’ or ‘origin’ of a particular feature and, on the other, ‘influence’ exerted by some other system on the integration of that feature in the emergent or new vernacular (Mufwene 2001). Identifying the origins of creoles’ features that distinguish them from their non-creole kin in the relevant European language should not be a problem, in part because the (descendants of) non-Europeans who developed them targeted those European languages. As part of the adaptive process to their new ethnolinguistic ecologies, we can expect them as learners to have made every effort to speak the target language in the best way they could, and not to have deliberately decided to develop a new language out of it (pace Baker 1997). We can also surmise that the systems developed by these speakers have diverged from the target or from other, non-creole offspring varieties of the target despite their efforts not to do so. The reason is that those adults who

206


targeted the European language as a second language were subject to the substrate influence of the languages they had previously spoken (as also advocated by Baker 1997 and Corne 1999, among others). Since the colonial European language itself was also internally variable, we can expect the learners to have applied the principle of least effort and to have typically favored those variants that were the closest to options available in their substrate languages; this process was subject to the extent of structural homogeneity among the substrate languages themselves (Sankoff 1984; Singler 1988, among others). This is consistent with the complementary hypothesis, although in the original formulation its proponents (including Mufwene 1986) thought of substrate influence almost exclusively in terms of structures that the non-European languages did not share with the European target language; these authors sought to identify the contribution of those substrate systems to the creole’s system, next to those elements inherited from the target language. In the ecology-based version presented in Mufwene (2001), the contribution of the substrate languages was made very significant, to the extent that it functioned as a filter (Hazaël-Massieux 1993), influencing the selection of particular options that were available in the European language itself (bearing in mind that this target was typically non-standard). Identifying substrate influence is hardly straightforward (e.g., DeGraff 2001, 2002a), although Baker (1984, 1997) makes an interesting lexical case when he accounts for the re-analysis of French article + noun phrases into single words (e.g., du riz “some rice” > diri “rice”) on the model of the agglutination of noun prefixes to their nominal bases in Bantu languages (see below). Valdman (1977) is an illustration of the above congruence-oriented interpretation of substrate influence; in the context of French creoles, this position can be understood best when one also reads Sylvain (1936) from the complementary hypothesis perspective rather than from the point of view of the relexification hypothesis, with which she has been too hastily and unfairly associated. She provides many examples from West African languages that are congruent with structural features of non-standard French varieties, which she also cites extensively. In the end, heeding Chaudenson (2001), one must ask whether there is documentation of naturalistic language appropriation in which learners pay attention only to the vocabulary of the target language, completely ignore its morphosyntactic strategies, and use those of their native languages to speak with target vocabulary. Why don’t we have more instances of relexification in those domains where the substrate grammars are more drastically different from the target grammar? Why have such cases obtained only in Melanesian pidgins (Keesing 1988; Sankoff 1993; Sankoff & Brown 1976)? For instance, why do we not have evidence of noun classes and suffixal tenseaspect markers in any ACR or IOR creole, especially those in which there is extensive fusion of the French article and the noun into creole nouns, such as


207

dilo “water” (< de l’eau “water” or “the water”), nom “man” (< un homme “a man”), zorey “ear” (< les oreilles “the ears”), lari “street” (< la rue “the street”), lakaz “house” (< la case “the house”), and diri “rice” (< du riz “some rice”)? Most of the cases explained by substrate influence seem to involve (partial) congruence of structures of the target European language with those of substrate languages, as made abundantly obvious by Corne (1999). The fusion of (part of) the French article to the head noun is a case in point. The article (+ adjective) + noun sequence in French is partly congruent with the agglutination of the noun class prefix to the nominal base in Bantu not only because both the article and the prefix precede the noun they delimit but also because, phonetically, no pause intervenes between it and the noun, unless the latter is modified by an anteposed adjective. Thus what is written as les enfants, with two words, is spoken as [lzf], in one phonetic word. Likewise, in Bantu languages, the noun class prefix also makes one phonological word with the nominal base. What must be analyzed as ba-ána is spoken as bána, without a possible break between the prefix and the base morpheme. No modifying material can occur between them. 1 Baker (1997) suggests that the African slaves would have deliberately worked at developing their own “means of interethnic communication,” which they would have wanted to diverge from the colonial European koine. However, the evidence as presented by Valdman (1977) suggests correctly that the 1

As Michel DeGraff (personal communication, 19 May 2003) reminds me, the question arises of why fusion was not prevented in those cases where an adjective occurs between the article and the noun. This is a very interesting question, because the alternation between the article + noun and article + adjective + noun sequences would have made it obvious that the article is a separate constituent, even to Bantu speakers. Is it possible that the fusion occurred only in those cases when the adjective rarely occurred pre-nominally? Diachronic information about such constructions in seventeenth- and eighteenth-century French should help, because while du riz chaud is better formed than du chaud riz “(some) hot rice,” du bon riz is better formed than du riz bon “(some) good rice.” Such materials are not evidence against substrate influence. They might lend more support to Chaudenson’s (1992, 1993, 2001) hypothesis that the structural features that distinguish creoles either from the European koines from which they ultimately evolved or from their non-creole kin are later developments. The divergent features would have emerged when, as expressed in Mufwene (1996, 2001), (a) the slave population grew more by importation than by birth, (b) the proportion of non-native speakers exceeded that of native-speakers, and (c) the local vernacular was increasingly being transmitted from non-native to non-native speakers, while communication by slaves was also predominantly (if not exclusively) with other slaves. Although today’s French creoles provide constructions such as gran moun “big person, adult” and bel gato “good/beautiful cake,” it may be informative to find out what the dominant position was and is for adjectives that combine with those nouns that have fused with articles. I leave this question to the attention of those with expertise on French creoles. A modified version of DeGraff’s question can also be extended to the fact that the fusion affected only articles but not other pre-nominal determiners, especially demonstratives, which must have been commonly used too.

208


non-Europeans on the plantation settlement colonies learned the non-standard European koines to which they were exposed much better than they have been credited for in most theories of the development of creoles. These theories rely on the exclusive effect of substrate influence to account for the divergence of creoles’ structures from those of the standard varieties of the European languages with which they have typically been unjustifiably compared in the first place. The above considerations make more sense once one acknowledges that in the first place, language ‘acquisition’ is a reconstructive process (DeGraff 1999a, 1999b, 2002b; Hagège 1993; Meillet 1929; Mufwene 2001, 2004) and proceeds both gradually and piecemeal during which the elements inherited with modification from diverse sources (on the model of polyploidism in population genetics) are re-integrated in somewhat novel ways, despite a significant amount of similarities with various model idiolects (Mufwene 2001). The model idiolects can be native or xenolectal and their features are filtered through a selection mechanism whose operations are still not fully understood. In a contact setting of the kind where creoles developed, the vocabulary and structural features of one particular language have usually been favored by the fact that it is targeted. However, the target system is hardly ever faithfully reproduced, for a host of reasons discussed in Mufwene (2001), one of which is the inability of the learner to reproduce the whole range of variants that occur in the communal system. While we can recognize features of various model idiolects in the emergent idiolect, the latter in no way replicates a single one of the models faithfully. The other reason why the communal target variety is not replicated lies in the fact that substrate languages affect how units and principles of the target language are perceived and interpreted. They account for some of the modifications it undergoes during the appropriation of the target by the new cohort of speakers. That is precisely how substrate influence affects language appropriation; this influence works more pervasively than has been acknowledged, even in those cases where there is total or extensive congruence with a structural feature of the target language, because it makes it easier to acquire that particular aspect of the target system. Substrate languages can thus also account for divergence of creoles from the languages they have evolved from, because substrate languages bias the way some units and principles of that language evolve. Valdman’s (1977) study becomes very informative in this respect, because it provides or suggests details that help us determine the direction and extent of the divergence of French creoles’ determiner systems from the colonial French koines from which they developed. It helps explain how the determiner system has evolved into something closer to the system of some substrate languages.


209

Although the determiner is postposed in several African languages, in which it is typically a demonstrative (hence a deictic), Valdman shows that in “français populaire” the postposed là has a similar syntactic position and semantic function. In fact, Chaudenson (1993:19-22) observes that this determiner + noun + là morphosyntactic option seems to be preferred over the determiner + noun alternative in several varieties of français populaire, not only the colonial koines that were targeted by the African slaves but also the new French dialects that developed in places like Québec, Louisiana, and Missouri. There is no doubt that the post-posed là co-occurs with a preposed definite determiner in these non-creole varieties. Although this is no longer the case in French creoles such as Haitian, it still carries a definitizing function, in addition to its deictic function, just as in French. Markedness considerations that could be invoked as favoring the selection of là as a marker of definiteness and a deictic include its occurrence at the end of the noun phrase, in a position that makes it stressed and therefore salient, as well as the congruence between là and a functionally similar morpheme in several substrate languages. These factors must have favored it over the definite article or the demonstrative adjective, which are prenominal and have prevailed in standard French. Interestingly, the constructions with the postposed là are still used in standard French, although only with a specialized meaning. The structures of French creoles are individually not as drastically different from those of French as has been claimed. This is perhaps unsurprising given that French creoles are understood to have evolved from non-standard French koines, in which there were competing variants from different metropolitan non-standard dialects; additionally, different offspring of such varieties cannot be expected to have selected exactly the same subsets of features, a factor that accounts for differences among the creoles themselves. Differences between French creoles as a group and their non-creole counterparts are only enhanced by the particular combinations of features that the creoles or the non-creole vernaculars have selected. These selections depended on the particular nonEuropean substrate influence that the xenolectal speakers exerted on them and the extent of divergence they have effected relative to (some of) the other offspring varieties. Valdman (1977:111-115) shows more of this in his brief comparison of possessive constructions in French and creoles. As Chaudenson (1992, 2001) and Fattier (2002) also point out, French offers two alternative models, one in which the possessor precedes the possessed/head noun (mon papa “my father”) and a second, in which the possessor follows the head noun (papa à moi). In the latter option, there is variation regarding the preposition, as de can be used instead of à, when the possessor is nominal. However, in non-standard French and in child language, the alternative with the preposition à either prevails or is

210


well represented. One can safely assume that it may have also prevailed in colonial French koines. Valdman (1977, 2002) shows that in early French creoles of the late eighteenth century, both the preposed and the postposed possessor constructions were attested in the ACR and the IOR. He also shows that the changes that would ultimately affect this aspect of French are largely the consequence of internal evolution within the emergent creoles under substrate influence. First of all, the French distinction between the possessive and non-possessive forms of pronouns and personal pronouns was obliterated. For example, in the IOR creoles, mo “I/me” and to “you” came to be used in possessive and nonpossessive functions and in both pre- and postposed positions: mo ti ser “my younger sister” vs. mo get li “I look for him”; ki krié to “who called you” vs. to kodak “your camera”; m a pr Iš to “I will take your kids.” In the ACR creoles, it is the French tonic pronouns that prevailed in the postposed possessor construction: mwen isit “I am here” vs. li ban/bay mwen “he gave me” vs. ròb (a) mwen “my dress.” Regardless of whether or not there is a partial model for the obliteration of these distinctions in non-standard French, especially in its colonial koine varieties (e.g., Brasseur 1997), one can safely invoke substrate influence; for instance, in non-Bantu languages the same pronominal form is used in all syntactic functions, which are distinguished by position. 2 There are among substrate languages of West Africa many that mark possession only by adjoining the possessor noun phrase (regardless of whether it is nominal or pronominal) to the possessed. This would explain the omission of the preposition in, for instance, Haitian Creole (HC). What Valdman’s (1977, 2002) discussion also suggests is that the constructions that appear to diverge the most from French patterns are later developments (see also Chaudenson 1992, 2001), since the earlier, late-nineteenthcentury texts he cites have structures closer to French than those attested today in the same vernaculars. 3 Variation in the same community may have decreased with the gradual normalization of the creoles’ systems, with those of the ACR favoring the postposing option, while those of the IOR favored the preposed option. Whether or not, as argued by Chaudenson (1992, 2001) the difference in these selections has something to do with differences in substrate populations merits further investigation. Note that Baker (1984) suggests the same thing regarding the more extensive fusion of the article and noun in IOR creoles than in their ACR counterparts. Specifically, he points to the greater 2

Ironically, Gungbe, one of the languages of the Gbe cluster that have typically been invoked in relexification in the development of Haitian Creole, appears to make a case distinction for the first and second person singular (Aboh 1999). 3 Chaudenson (1993:23) also argues that older Mauritian and Réunionnais vernaculars (of the eighteenth and early nineteenth centuries) were more similar structurally than they are today. Their later demographic histories account at least partly for their present structural differences.


211

presence of Bantu speakers in the IOR than in the ACR, where there were more non-Bantu speakers during the plantation period. What is also worth noting here is the fact that the ACR creoles split into the Greater Antilles variety (e.g., HC, at least its southern variety), which does not use a preposition, and the Lesser Antilles varieties, which use a preposition, at least variably. It is not clear how substrate influence can be invoked to account for this divergence, though one can invoke the history of Haiti, which severed links with European speakers during the first half of the nineteenth century (a consequence of the Haitian Revolution) and created an ecology in which its creole could be subject to more extensive substrate influence than its ACR counterparts. Interestingly, Haiti was also the first New World polity to stop further importation of slaves, which means that its more extensive basilectalization compared to other ACR creoles has to do primarily with the demographic attrition of French in the country. This entailed that in the remaining non-European population, which was largely proletarian and basilect-speaking, the slaves who had been imported soon before the Revolution and were freed soon after were more likely than their counterparts elsewhere in the Caribbean to exert a more significant substrate influence on the emergent creole. A point that clearly emerges from this is that creoles seem to be an opportune window into colonial European koines. This is not because they can be misconstrued as fossils of that stage in the evolution of European languages, but rather because they reveal some of the early colonial koines’ features that were inherited from and through the earliest creole populations of the homestead phase. 4 The creoles naturally share some of these features with their noncreole non-standard kin, but they have also restructured some of them under the influence of the separate non-European substrate that was felt by the creoles during the plantation phase of the colonies. In some cases, there is no particular substrate influence to invoke in order to account for such divergent selections, which may simply be an accident of probability games. Valdman (2002) suggests this in his discussion of front rounded vowels in Northern HC. Their presence goes against the general trend of unrounding which is attested not only in other French creoles but also in a number of non-standard French vernaculars. The unrounding is of course one of those developments that can be easily explained by the congruence of evolutionary tendencies in the target and substrate languages, which does not contradict Chaudenson’s (1992, 2001) invocation of the extension of self-regulating tendencies in French itself. This consideration provides an ecological context for the process. However, Valdman also shows through his examination of historical texts that this marked retention of rounded vowels is the legacy of the kind of variation that obtained in the plantation settings themselves (2002:127-128). The demonstra4

The Creole populations were apparently not creole-speakers, because the evidence suggests that creole vernaculars may not have emerged before the eighteenth century.

212


tion is consistent with Fattier’s (2002:107) observation that “The creole of monolingual Haitians appears to be a repository of old French words and constructions or [morphosyntactic] strategies.” She also observes that many features of non-standard French are persistent in HC, despite the basilectalization process that has caused it to diverge from its colonial non-creole kin. The moral of this story is that we should be very cautious in our interpretation of the structural features that distinguish creoles from their non-creole kin. Based on the competition-and-selection model submitted by Mufwene (2001, 2002a), any variety must have diverged from the colonial koine by selecting its own subset of variants and integrating them in a new system. In this new system, some of the variants would be modified under the influence either of the internal dynamic of their coexistence with other selected features or of some substrate languages. In both his 1977 and his 2002 essays, Valdman clearly answers the first central question of this essay: Similarities between features of French creoles and those of non-standard French suggest that the terminus a quo of French creoles is certainly the non-standard French koines spoken by the European colonists. I submit that, contrary to the tradition that has overemphasized the innovative aspect of the development of creoles, French creoles provide us a glimpse of the structural features of colonial French. Below, I assess Valdman’s answer to the second central question. 3. How did creoles develop? Valdman (1977) considers a number of relevant matters here. The first is whether an antecedent pidgin ancestor to the ACR and IOR French creoles could have developed on the coast of West Africa. Anticipating Huber (1999), his answer is negative. One of the reasons is that the French had very few trade forts; this is in contrast to the Portuguese and the English, who covered more territories and had better established trade networks. Another important reason, which he shares with Huber (1999), is that a Portuguese pidgin was already well established in the region, and traders from other European nations must have used it to trade with the indigenous populations. He points out that the pidginized French varieties that have been reported in the seventeenth and eighteenth centuries (identified as baragouins) have been situated only in the Americas. Thus, according to Valdman, there is no compelling evidence for relating the ACR or the IOR creoles to a pidgin that would have developed on the West African coast (1977: 134). We should note that Huber (1999) also argues against assuming a widespread West African Pidgin English that would have provided the beginnings of Atlantic English creoles, noting that Nigerian and Cameroon Pidgin Englishes are much later developments. The evidence for the often-invoked West African Pidgin English or Guinea Coast Creole English since Dillard (1972)


213

and Hancock (1972, 1986) remains scant, indirect, and disputable. The available limited texts, cited from the eighteenth century, do not suggest conclusively that the varieties spoken in the mixed marriages of European lançados and African women, especially by their children, was either a pidgin or a creole. The Dillard-Hancock hypothesis is less plausible in regard to children: Even if their mothers, as adult naturalistic second language learners, would have spoken the kinds of approximations inferred by Chaudenson (1992, 2001) in the context of homestead colonies, the children must have been able to speak their father’s language natively, in addition to the indigenous vernacular. Once they became adults, some of the children must have become interpreters between the European and African traders, in addition to other privileges they must have enjoyed. This alternative insight into early trade-colonization history is in fact consistent with Berlin’s (1998) discussion of the role of creoles (extended also to Mulattoes) as power-brokers in both the Americas and Africa during that period. Moreover, the Mulatto children’s mothers cannot be expected to have spoken pidgins because they had regular contacts, not sporadic ones, in the forts where they lived with the European lançados. It is also necessary to re-examine another aspect of the Dillard-Hancock hypothesis. Because they relied heavily on grumettoes (including, most likely, adult children of the lançados) as intermediaries, the Europeans and the segment of African populations with whom goods were traded (before the slave trade intensified) need not have interacted directly with one another, and they probably did not. Thus, the Africans would not have felt much pressure to learn a European language. However, as intermediaries in the Euro-African trade, the grumettoes who were not children of the lançados may have spoken pidginized varieties, if not simply approximations, of the European languages. Even during the slave trade, the African captives would have had no compelling reason to learn the grumettoes’ pidgin as their lingua franca. There must not have been much two-way communication between the two groups. We simply do not know how the Africans communicated with each other on forts before the Middle Passage or during the Middle Passage itself, or how much communication they had with each other before they reached the ACR and IOR colonies. As far as language contact in Africa is concerned, I surmise that as contacts between Europeans and Africans increased by the nineteenth century and they became more and more direct and relied less and less on intermediaries or interpreters, the Africans would have increasingly learned the European lingua francas hitherto spoken by the grumettoes or interpreters. Thus, to use Chaudenson’s (1992, 2001) depiction of the process, “approximations of approximations” of models based on sporadic contacts would produce present-day pidgins on the West African coast.

214


Valdman (1977) leads us to conclude that both the ACR and IOR creoles must have developed in situ. He says nothing about whether the baragouins of the Americas exerted any influence on the development of the ACR creoles, though Chaudenson (1992, 2001) argues that the baragouins and creoles had separate developments. Based on his study of the determiner system, Valdman concludes that this particular subsystem must have originated in the kind of French that was spoken in the French plantation colonies during their formative periods: “Creole must have originated in the approximative varieties that diverged little from colonial French” (1977:132-133). Worth taking into account to make sense of the above conclusion is also an interesting distinction that Valdman makes between “créole conservateur” and “créole innovateur.” He applies the former term to those vernaculars that are close to the earliest approximations of colonial French, those that some creolists today would describe as mesolectal, and the latter to the later-stage developments, the end-state of the basilectalization process hypothesized by Chaudenson (1992, 2001) and Mufwene (1996, 2001). His “créoles innovateurs” thus contain innovations that distinguish them from other colonial offspring of French spoken by descendants of Europeans, not because the latter have not innovated at all, but because the innovations are different. In the case of creoles, one must invoke the non-European substrate influence as an important factor in the divergence. This is quite consistent with Chaudenson’s (1992, 1993, 2001) observation that if one did not factor in the African substrate languages, then it would be difficult to account for those respects in which creoles differ from their colonial non-creole kin, not because the origins of the features are necessarily substrate but because the substrate languages have influenced the particular ways in which some features would evolve in slave communities segregated from their white counterparts. Valdman’s arguments can also be used against the decreolization qua debasilectalization hypothesis. He claims that standard French, which was not spoken by the “Petits Blancs” (“homesteaders”), was introduced later, by the time the colonies had moved to large plantation systems. This period, which he does not identify explicitly, would correspond to the late eighteenth century. As plausible as the hypothesis sounds, the local standards cannot have developed independent of the influence of the local vernacular French spoken by the plantocratic families that started with small holdings and grew bigger and bigger over the years or centuries by buying their neighbors out. As suggested by Irvine (2004), the evolution of standard varieties in the colonies deserves just as much careful investigation as their non-standard counterparts. For now, however, we will settle for the observation that Valdman’s arguments speak


215

against the decreolization hypothesis and in favor of the basilectalization alternative. 5 Valdman returns to this question in his (1991) essay on decreolization in HC, in which he states that “considerable variation obtained during the formative stages of the language and continued during the colonial period” (86). 6 If HC can be said to be undergoing decreolization, it is only in the sense that a continuum of variation is being created between French and B[ilingual] H[aitian] C[reole] by the use, on the part of bilinguals, of mesolectal morphosyntactic features. (85)

According to him, it is especially at the lexical level that the boundary between BHC and French becomes blurred, but the grammars are still distinct. In quite an informative way, he adds that: When bilingual speakers affirm their Haitian identity and wish to signal their solidarity with their less affluent compatriots, they will shy away from mesolectal forms and model their speech on M[onolingual]HC. (85)

Otherwise, “because of the relative unavailability of direct French models, MHC is, as it were, protected from decreolization” (86). I need not discuss issues that I have addressed in detail in Mufwene (1994), especially since my position is akin to aspects of Valdman’s views that I highlight here. He claims that for BHC speakers, the population typically identified as French-speaking, Creole and French stand in complementary distribution, being allocated to different ethnographic domains (1991:85). Dejean (1993) disputes this claim, pointing out that: French is the preferred language of communication among parents and children of the so-called Haitian ‘elite’. One of the main characteristics of the linguistic behavior of literate Haitians is constant code switching in lengthy and animated discussions of topics like arts, literature, music, religion, politics, sciences, etc. (74) 5

To be sure, Valdman’s position on decreolization is not unproblematic. Valdman (1983) claims that the varieties spoken on St. Thomas and St. Barts are decreolized. The reason for this is his traditional assumption that the evolution from the European language to its creole offspring must have taken place via a pidgin, which is structurally reduced and quite removed from the original colonial koine to which the non-Europeans were exposed. Only extended contacts would have putatively brought the ensuing creoles closer to the colonial koines spoken by descendants of Europeans or to the local standard variety allegedly introduced later. This circuitous evolution is disputed by Valdman’s own claim that the origins of creoles’ structural features lie in the non-standard varieties spoken by the European colonists and by Chaudenson’s (1992, 2001) and Mufwene’s (1996, 2001) position that creoles have evolved directly, by gradual basilectalization, from the relevant European language. 6 This later statement is certainly more accurate than his earlier one in Valdman (1983:517): “from the early days of French colonization in Saint Domingue there has existed a decreolized variety of H[aitian]C[reole].” Based on what we now know about the development of creoles, there is no particular reason to invoke early decreolization during the very period when the creole was emerging by basilectalization.

216


Dejean’s main argument is that BHC speakers are native speakers of both French, which they prefer to use within their elite class, and HC, which they use fluently in communicating with the overwhelming mass of the population, on whose hard labor they depend. Thus there is definitely a wide range of overlapping settings, and topics can be handled linguistically in either HC or French, especially since, as he also points out: All civil servants in Haiti speak Creole predominantly in their everyday activities as members of the armed forces and the police, or as employees of the courts, the postal service, the internal revenue service, the departments of education, agriculture, health, etc. (74)

Still there exists a charitable misinterpretation of Valdman’s claim. In most if not all creole communities, as in most non-creole ones, a distinction is made between the non-standard or colloquial variety, which creoles fundamentally are, and its standard counterpart. Bilingual or bi(dia)lectal speakers tend to associate the creole or non-standard/colloquial varieties with vernacular topics and settings, and/or with people in the lower strata of their communities (who happen to be the vast majority). On the other hand, they typically reserve the acrolectal or standard variety for non-vernacular or formal topics and settings, and/or with people in the upper class. However, the boundaries of this distinction are quite fuzzy. It is not that Creole cannot be used in domains where the acrolect is used or vice versa, but rather the varieties tend to specialize, at least as targets, with either users or topics. For instance, science is typically learned in school and through the acrolect. Though a speaker may not be fluent in this variety, he or she would still target it in this particular ethnographic domain. 7 Otherwise, except in the fuzzy areas, there is really no competition between the standard and non-standard varieties; consistent with Dejean’s (1993) observations, members of the two segments of the population typically speaking one or the other variety socialize more among themselves than with the other socioeconomic group. As in the case of the coexistence of language varieties in general, it is those competing for the same ethnographic functions that vie for monopoly, typically with one of them gradually prevailing at the expense of the others. In other words, the topic of decreolization is not really different from that of language endangerment and loss by demographic and/or structural attrition (i.e., loss of speakers and/or structural changes in favor of acrolectal alterna7

We are actually now at a point where it is useful to distinguish the acrolect, as the variety spoken by the upper class (in the way Irvine, 2004, realistically characterizes it), from the local standard variety of the European language, which is restricted to formal settings. Dejean’s observations are more consistent with Irvine’s characterization of the acrolect, whereas Valdman’s may be more consistent with the identification of the acrolect with the standard variety, although one must wonder how many use it as their vernacular.


217

tives). Decreolization qua debasilectalization or shift away from the basilect can also be explained as loss of the basilect by the erosion of its system in favor of that of the acrolect. So far this process has not been documented, although, like any language, creoles have changed some of their features over time, as shown by Rickford (1987). It is not clear that the disappearance of the nineteenth-century Jamaican-like basilect in Barbados (Rickford & Handler 1994) has necessarily been in favor of the acrolect, since there is a nonstandard vernacular—which a creole really is—that is still spoken by many Bajans. 8 Prestige of the acrolect need not lead to the loss of the basilect if they do not share ethnographic functions. One of the reasons creoles all over the world have not died by the putative decreolization process is indeed the fact that they are typically used in domains where the local acrolect either is precluded or would be odd, such as in ordinary interactions in the socio-economic group whose members use Creole also as their identity badge, in private or rather informal settings even among acrolectal speakers (at least at some level of its mesolectal continuum), in the lyrics of vernacular music, and in folk narratives. Another reason that creoles have not disappeared due to decreolization, which Valdman reminds us of, is that monolingual Creole-speakers often constitute the (overwhelming) majority of their communities. According to Dejean (1993:76-78), 95% of Haitians are monolingual in Creole, many of them do not interact with French speakers (see also Valdman 1983:518), and members of the French-speaking elite also speak Creole. The fact that acrolectal speakers can speak some mesolectal or basilectal variety of the local Creole makes it unnecessary for Creole-speakers to learn the acrolect, 9 especially when they are socio-economically disenfranchised. Contrary to Valdman’s (1983: 521) claim that vernacular creole speakers are the expected agents of decreolization, those who develop bilingualism (or bidialectalism) in the basilect and the acrolect are under no pressure to give up the former as a vernacular, despite the socio-economic prestige and advantages that the acrolect is associated with.10 8

The evidence of that nineteenth-century variety is of course literary. It is not clear what proportion of the Bajan population spoke it, despite the growth of large plantations on the Island. Up to the late eighteenth century, its demographic disproportion between Africans and Europeans was not as overwhelmingly in favor of Africans as was the case in colonies such as Jamaica and Guyana around the same time. 9 Valdman (1983:518) also observes that “some educated Haitians have adopted an academic, puristic attitude toward the vernacular.” 10 Here too Dejean (1993:79-80) has some insights to contribute to the debate: Alongside a certain admiration, in the least favored social strata, for educated people who also speak French, no one has described an equally real phenomenon: that of French being perceived by the people as an object of ridicule. Far from admiring, as some think naively, the population at large mocks the language behavior of the bilin-

218


Urban Creole is different from rural Creole not because of decreolization but because the European language has not evolved in a uniform way among the non-Europeans in the plantation settlement colonies. The plantation settings were not the same as the urban settings, in which the European/nonEuropean population disproportions as well as the race-based linguistic divergence were less drastic (Stewart 2002). This setting difference is just like in North America, where English among (descendants of) Africans diverged less drastically on the tobacco and cotton plantations, where they had more contacts with (descendants of) Europeans, than in the rice fields, where Gullah developed (Mufwene 1999b, 2001). In places like Jamaica, other grass-roots cultural developments have in fact recently produced a new variety that diverges from both the traditional basilect and the acrolect: Dread Talk (Pollard 2000). This new variety is indeed based on the basilect rather than on the acrolect. Overall, Valdman provides useful insights about the lasting coexistence of Creole and the acrolect, despite some confusing statements he made about decreolization. 4. Conclusions Although he recognizes the role of substrate influence in the development of French creoles, Valdman (1977, 1983, 2002) clearly identifies the nonstandard French koines spoken by the early colonists as the varieties from which the new vernaculars of the descendants of Africans in the ACR and in the IOR have developed. Strangely, while he recognizes that varieties close to these colonial koines have been spoken among (descendants of) Africans since the beginnings of French creoles, he espouses a pidgin-to-creole evolutionary trajectory. The position is at odds with the rest of his arguments, which clearly suggest Chaudenson’s (1992, 2001) alternative hypothesis that creoles have evolved directly from the koines, with some of the non-European populations of the plantation phase increasingly producing varieties that are structurally divergent from the local terminus a quo, while a minority of others continued to speak near approximations of the koines. Based on Rickford (1990), the vast majority of the populations in creole communities speak something at some level of the continuum between the basilectal and acrolectal poles. Valdman is thus led to posit some early decreolized varieties to account for the less divergent creole varieties spoken on St. Thomas and St. Barts (in relation to a putatively common terminus a quo), while arguing at the same time that the basilect-to-acrolect continuum observable in Haiti is due to bilingual speakers who mix lexical and phonological materials from both poles. gual bourgeois. (...) Not to speak Creole in public is increasingly viewed as an overt expression of disdain and hostility toward the people. These attitudes and feelings on the part of the mass of the population, which Dejean reports so accurately (see also Mufwene 1999a), are among the factors that have prevented decreolization qua debasilectalization, to which I return below.


219

However, he does not consider this continuum to be evidence of decreolization. He is quite correct. There is no particular reason why the continuum per se should be evidence of it, because speakers of substrate languages shifting to the colonial koine as their vernacular were not engaged in this ethnographic process as a sports team. Aside from the fact that they did not all have the same (second) language learning skills, they were not always exposed to identical models, and did not all develop idiolects that were basilectalizing (Mufwene 2002b). As pointed out earlier by Chaudenson (1992, 2002) there were always some among them who continued to acquire and speak closer approximations of the terminus a quo or of the more recent varieties spoken by (descendants of) the European colonists. Valdman assumes that the critical agents of decreolization would be monolingual speakers, in whom he has detected no linguistic behavior similar to that of bilinguals, in part because most of them have no, or limited, interactions with acrolectal speakers. Whether or not monolingual speakers borrow no words today associated with the acrolect, Valdman is very useful in his highlighting of the ethnographic complementarity between it and the basilect. This complementarity of settings, functions, and speakers (as explained in section 2) turns out to be an important reason why decreolization remains a myth. In fact, creoles in most of the relevant territories continue to thrive even when they coexist with some variety of the European language from which they evolved, contrary to stipulations on the subject matter from Schuchardt (1914) and Bloomfield (1933) to DeCamp (1971) and his followers. It is also relevant to underscore Valdman’s position that creoles did not emerge in a uniform way, with the urban varieties being less divergent than the rural ones, though it is not certain that decreolization has anything to do with the condition of urban creole. Besides, it has been established for a while that there is regional variation even on the same island, as has been often reported about Haiti and Jamaica.11 Like Chaudenson, Valdman also makes a distinction between the creole spoken by the Petits Blancs and that spoken by (descendants of) the Africans, thus highlighting the importance of recognizing the variable ecology of the development of creole as a non-monolithic vernacular. One can undoubtedly point out today that some of Valdman’s positions are not fully developed, but this in no way diminishes the significance of the fact that Valdman counts among the first of today’s creolists to have articulated positions that appear to be consistent with the socio-economic histories of the relevant population of speakers.

11

In the case of HC, we can now benefit from Fattier’s (1999) extensive documentation of its regional variation, on the model of Carayol, Chaudenson, and Barat’s (1989) documentation of lexical variation in Réunionnais.

220


References Aboh, Enoch Oladé. 1999. From the syntax of Gungbe to the grammar of Gbe. Unpublished doctoral dissertation, University of Geneva. Baker, Philip. 1984. “Agglutinated French articles in creole French: Their evolutionary significance”. Te Reo 27. 89-129. ----------. 1997. “Directionality in pidginization and creolization”. The structure and status of pidgins and creoles, Spears & Winford (eds) 1997 (Amsterdam: John Benjamins). 91-109. Berlin, Ira. 1998. Many thousands gone: The first two centuries of slavery in North America. Cambridge, MA: Harvard University Press. Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart, & Winston. Brasseur, Patrick. 1997. “Créoles à base lexicale française et français marginaux d’Amérique du Nord: Quelques points de comparaison”. Contacts de langues, contacts de cultures, créolisation: Mélanges offerts à Robert Chaudenson à l’occasion de son soixantième anniversaire, M.-C. HazaëlMassieux & de Robillard (eds) 1997 (Paris: L’Harmattan). 141-166. Carayol, Michel, Robert Chaudenson, & Christian Barat. 1989. Atlas linguistique et ethnographique de la Réunion. Paris: CNRS. Chaudenson, Robert. 1979. Les créoles français. Paris: Fernand Nathan. ----------. 1992. Des îles, des hommes, des langues: Essais sur la créolisation linguistique et culturelle. Paris: L'Harmattan. ----------. 1996. “De l’hypothèse aux exemples: Un cas de créolisation—la formation des systèmes de démonstratifs créoles”. Etudes Créoles 16. 17-38. ----------. 2001. Creolization of language and culture. London: Routledge. Corne, Chris. 1999. From French to Creole: The development of new vernaculars in the French colonial world. London: University of Westminster Press. DeCamp, David. 1971. “Toward a generative analysis of a post-creole speech continuum”. Pidginization and creolization of language, Hymes (ed) 1971 (Cambridge: Cambridge University Press). 349-370. DeGraff, Michel. 1999a. “Creolization, language change, and language acquisition: A Prolegomenon”. Language creation and language change: Creolization, diachrony, and development, DeGraff (ed) 1999 (Cambridge, MA: MIT Press). 1-46. ----------. 1999b. “Creolization, language change, and language acquisition: An epilogue”. Language creation and language change: Creolization, diachrony, and development, DeGraff (ed) 1999 (Cambridge, MA: MIT Press). 473-543. ----------. 2001. “On the origin of creoles: A Cartesian critique of NeoDarwinian linguistics”. Linguistic Typology 5. 213-310. ----------. 2002a. “Relexification: A reevaluation”. Anthropological Linguistics 44. 321-414.


221

----------. 2002b. “On creole genesis and language acquisition: Some methodological and theoretical preliminaries”. Unpublished Mnauscript. Dejean, Yves. 1993. “An overview of the language situation in Haiti”. International Journal of the Sociology of Language 102. 73-83. Dillard, J.L. 1972. Black English: Its history and usage in the United States. New York: Random House, Vintage Books. Fattier, Dominique. 1999. Contribution à l’étude de la genèse d’un créole: Atlas linguistique d’Haïti, cartes et commentaires. Villeneuve d’Ascq: Les Presses Universitaires du Septentrion. ----------. 2002. “La créolisation du français en Haïti: Partir du produit pour penser le processus”. Etudes Créoles 25. 105-122. Ferguson, Charles. 1968. "Language development." In Joshua Fishman, Charles Ferguson and J. Das Gupta (eds), Problems of Developing Nations. New York: Wiley, 27-36. Hagège, Claude. 1993. The language builder: An essay on the human signature in linguistic morphogenesis. Amsterdam: John Benjamins. Hancock, Ian F. 1972. “A domestic origin for English-derived Atlantic creoles”. The Florida FL Reporter 10. 7, 8, 52. ----------. 1986. “The domestic hypothesis, diffusion and componentiality: An account of Atlantic anglophone creole origins”. Substrata versus universals in creole genesis, Muysken & Smith (eds) 1986 (Amsterdam: John Benjamins). 71-102. Hazaël-Massieux, Guy. 1993. “The African filter in the genesis of Guadeloupean Creole: At the confluence of genesis and typology”. Africanisms in Afro-American language varieties, Mufwene (ed) 1993 (Athens: University of Georgia Press). 109-122. Huber, Magnus. 1999. “Atlantic creoles and the Lower Guinea Coast: A case against Afrogenesis”. Spreading the word: The issue of diffusion among Atlantic creoles, Huber & Parkvall (eds) 1999 (London: University of Westminster Press). 81-110. Irvine, Grace Alison. 2004. “A good command of the English language: Phonological variation in the Jamaican acrolect”. Journal of Pidgin and Creole Languages 19. 41-76. Keesing, Roger M. 1988. Melanesian Pidgin and the Oceanic substrate. Stanford: Stanford University Press. Manessy, Gabriel. 1995. Créoles, pidgins, variétés véhiculaires: Procès et genèse. Paris: CNRS. Meillet, Antoine. 1929. “Le développement des langues”. Continu et discontinu, Chevalier (ed) 1929 (Paris: Bloud & Gay). 119-131. Mufwene, Salikoko S. 1986. “The universalist and substrate hypotheses complement one another”. Substrata versus universals in creole genesis, Muysken & Smith (eds) 1986 (Amsterdam: John Benjamins). 129-162.

222


----------. 1994. “On decreolization: The case of Gullah”. Language and the social construction of identity in creole situations, Morgan (ed) 1994 (Los Angeles: Center for Afro-American Studies). 63-99. ----------. 1996. “The Founder Principle in creole genesis”. Diachronica 13. 83134. ----------. 1999a. “The language bioprogram hypothesis: Hints from Tazie”. Language creation and language change: Creolization, diachrony, and development, DeGraff (ed) 1999 (Cambridge, MA: MIT Press). 95-127. ----------. 1999b. “North American varieties of English as byproducts of population contacts”. The workings of language: From prescriptions to perspectives, Wheeler (ed) 1999 (Westport, CT: Praeger Publishers). 15-37. ----------. 2001. The ecology of language evolution. Cambridge: Cambridge University Press. ----------. 2002a. “Développement des créoles et évolution des langues”. Etudes Créoles 25. 45-70. ----------. 2002b. “Competition and selection in language evolution”. Selection 3. 45-56. ----------. 2004. “Multilingualism in linguistic history: Creolization and indigenization”. Handbook of bilingualism, Bhatia & Ritchie (eds) 2004 (Oxford: Blackwell). 460-488. Pollard, Velma. 2000. Dread talk: The language of Rastafari. Mona, Jamaica: The University of the West Indies Press. Rickford, John R. 1987. Dimensions of a creole continuum: History, text, and linguistic analysis of Guyanese Creole. Stanford: Stanford University Press. ----------. 1990. “Number delimitation in Gullah: A response to Mufwene”. American Speech 65. 148-163. ---------- & Jerome S. Handler. 1994. “Textual evidence of the nature of early Barbadian speech”. Journal of Pidgin and Creole Languages 9. 221-255. Sankoff, Gillian. 1984. “Substrate and universals in the Tok Pisin verb phrase”. Meaning, form, and use in context: Linguistic applications, Schiffrin (ed) 1984 (Washington, DC: Georgetown University Press). 104-119. ----------. 1993. “Focus in Tok Pisin”. Focus and grammatical relations in creole languages, Byrne & Winford (eds) 1993 (Amsterdam: John Benjamins). 117-140. ---------- & Penelope Brown. 1976. “The origins of syntax in discourse: A case study of Tok Pisin relatives”. Language 52. 631-666. Schuchardt, Hugo. 1914. Die Sprache der Saramakkaneger in Surinam. Amsterdam: Johannes Muller. Singler, John V. 1988. “The homogeneity of the substrate as a factor in pidgin/creole genesis”. Language 64. 27-51.


223

Stewart, Michele. 2002. “The emergence of basilectal varieties in Kingston 1692-1865”. Paper presented at Xe Colloque International des Etudes Créoles, Corail, Réunion. Sylvain, Suzanne. 1936. Le créole haïtien: Morphologie et syntaxe. Wettern, Belgium: Imprimerie De Meester. Valdman, Albert. 1977. “Créolisation sans pidgin: Le Système des déterminants du nom dans les parlers franco-créoles”. Langues en contact: Pidgin -Creoles—Languages in contact, Meisel (ed) 1977 (Tübingen: Gunter Narr). 105-136. ----------. 1983. “Creolization and second language acquisition”. Pidginization and creolization as language acquisition, Andersen (ed) 1983 (Rowley, MA: Newbury House). 212-234. ----------. 1991. “Decreolization or language contact in Haiti”. Development and structures of creole languages: Essays in honor of Derek Bickerton, Byrne & Huebner (eds) 1991 (Amsterdam: John Benjamins). 75-88. ----------. 2002. “Comment distinguer la créolisation du changement linguistique ordinaire? ” Etudes Créoles 25. 123-141.

DIATOPIC VARIATION IN HAITIAN CREOLE ∗

Annegret Bollée & Pamela Nembach Universität Bamberg ABSTRACT: In his 1987 article, Valdman points to the importance of dialectal variation in Haiti and provides some examples showing that a “central form” occurring in the region of Port-au-Prince is opposed to “peripheral variants” in the North or the South. He cites Orjala (1970) who sets up the hypothesis that there are three discernible dialects in Haiti: northern, central, and southern. Fattier’s (1988) thesis, by far the most comprehensive description of Haitian Creole so far, makes it possible to verify this hypothesis; using this data, we intend to try to identify dialect boundaries, perhaps in the form of bundles of isoglosses which separate the southern peninsula or the northern peninsula, from the rest of the country. 1. Introduction In his epoch-making description of Haitian Creole, Valdman devoted a chapter to variation, including three pages on “geographic variation” (1978:286-289). 1 On the basis of two previous studies, those by Hyppolite (1949) and Orjala (1970), he identified three regional dialects (“groupes régionaux”): “the North (the Cap-Haïtien region in particular), the South (Jérémie and Les Cayes), and the Center (the Ouest department, including Port-auPrince)” (286). He goes on to say that “these three regional dialects can be distinguished primarily by morphophonological and lexical variables: there is little phonological and syntactic variation, at least to our current knowledge” [Editors’ translation]. The examples with which he demonstrates diatopic variation include, among others, forms of the postposed article, forms of the personal pronoun li “s/he, it,” forms of the progressive marker ape, and the use of a in possessive constructions. Other criteria for the identification of what Valdman calls dialect groups (“groupes dialectaux”) 2 are phonetic (e.g., swaf “thirst,” lay “garlic,” rad “clothes” in the Center; swèf, laj, had in the North and South) and lexical differences (e.g., kaderik “kettle,” chandèl “candle” in the North vs. bonm “kettle,” bouji “candle” in the Center and the South). The ∗

We thank Dr. Gabriele Knappe for a critical reading of our text and for polishing our English. Cf. also Valdman (1987:111-112) and (1991:76-77). 2 In this chapter, Valdman never uses the term dialect, but speaks of “groupes régionaux” or “groupes dialectaux” (286-287). Maybe he wants to suggest that the three ‘dialects’ show further internal variation. Cf. the “transitional zones” identified by Orjala (see below). 1

226


cases where coexisting variants in the North and South are opposed to a unique form in the Center can, in his view, be explained by the radiation of innovations from the center to the periphery (1978:288). Valdman’s description of dialectal variation, limited in scope because he had only Hyppolite (1949) and Orjala (1970) to rely on, is a good starting point for an attempt to discuss diatopic variation in Haiti on new empirical grounds. Fattier’s Atlas linguistique d’Haïti (ALH, 2000) provides a remarkable corpus of Haitian Creole as spoken by monolingual speakers in 20 villages or small towns spread fairly evenly over the Haitian territory. The data were elicited by means of a questionnaire of 2227 items, 1752 of which are presented on maps, the rest in text format in notes which accompany the maps or entries in the commentaries in volumes 1 and 2. The questions which will guide our analyses of the ALH are the following: Is it possible to identify dialect boundaries in Haiti which manifest themselves as bundles of isoglosses? Can the hypothesis underlying Valdman’s description be confirmed, namely, that there are three dialects in Haiti and that the central region, including the capital Port-auPrince, is a center from which innovations radiate towards the periphery? 3 Before discussing our findings, it may be helpful to show in which way the 20 localities or points of the ALH are related to the ‘dialects’ as described in the previous studies. Orjala identifies “three main dialect areas”—northern, central, and southern—and “at least six transitional zones marked by overlapping characteristics but belonging chiefly to one of the main dialect areas” (1970:28). The northern area corresponds to points 1 to 7 of the ALH (6 and 7 being transitional zones); the central area is represented by points 8 to 15 (8, 14, and 15 forming transitional zones); the southern area comprises points 16 to 20. “Common core items which are distributed throughout all the dialect areas of Haiti [are] referred to as General” by Orjala (1970:28). According to Valdman, the dialectal variables can be classified in two categories (1978:288): (1) Features which are found only in one region (“particularismes des parlers du Nord et du Sud”), including lexical items, morphophonological and morphosyntactic items, and phonetic features; (2) Traditional features coexisting in the North and the South with innovations ‘imported’ from the Center. We will present our data in accordance with this classification.

3

After discussing several examples from Hyppolite which, according to his inquiry, “do not offer sufficient evidence to identify as dialect markers,” Orjala (1970:39) remarks that “this may possibly offer some evidence to reinforce the impression that central dialects are actually invading both the north and the south.”

DIATOPIC VARIATION IN HAITIAN CREOLE

227

2. Specific features characterizing the dialect area of the North or the South 2.1 Lexical items In order to set off the northern dialect, Valdman chose seven lexical items (1978:287-288); five of these could be checked against maps in the ALH, 4 one of them (‘candle’) will be discussed in section 3.2. ‘straw bag’: djakout vs. makout (map 1398 ‘sac en latanier’). Djakout is attested in the North (except at point 1 showing djola/dyola, which also occurs at points 18 and 19), but does not form an isogloss, since it also occurs in the Center and the South (11, 14, 16, 17); makout(i) is mainly used in the Center (but 4 points have only a question mark) but is also found in the North at points 3, 5, and 7. ‘kettle, cooking pot’: kaderik vs. bonm (map 765 ‘faitout’). The typical name in the North, kaderik according to Valdman, is registered at two points only (4 and 5), whereas 1, 2, and 3 have bonm, which has been associated with the Center and the South. If kaderik was formerly the only or the dominant term in the North, this is no longer the case. It should be noted that point 4 (Fort Liberté), due to its geographic isolation, 5 often retains obsolescent terms or phonetic features (cf. the item ‘to joke’ below). ‘pot; tin can’: kanistè vs. mamit (map 764 ‘marmite’ and 1461 ‘marmite dite de cinq livres’). Valdman’s hypothesis could not be checked on map 764 because at all points the only designation is chodyè, with mamit as an alternative only at point 10. On map 1461, (gwo) mamit is attested virtually everywhere and kanistè only at 1, 5, and 8 together with mamit. This again may be indicative of a local feature which is gradually disappearing. ‘to joke’: jwe vs. badinen (map 330 ‘plaisanter’). The principal designations on the map are blage, fè blag, bay blag, bay istwa, etc.; jwe occurs once as one of four synonyms at point 4 (where obsolescent terms or forms are often retained, cf. ‘kettle, cooking pot’); badinen is not attested. An examination of five of the lexical items mentioned by Valdman revealed no discernible isoglosses on these maps. This discovery lead us to check all the other items of ‘lexical contrasts’ mentioned by Hyppolite or Orjala. Additionally, we examined a considerable number of maps, chosen unsystematically in the ALH in search of possible isoglosses. Dialect boundaries, however, could only be established in very few cases (cf. also section

4

The item ‘peanut’ was also found, but question 1660 (ALH 2:727-728) nowhere received the answer amizman, supposed to be typical of the north. The reason for this may be that the exact meaning of amizman is “salted peanut,” whereas the question asked for the ALH referred to the simple peanut. 5 The town, situated in a bay, is separated from the Center by mountains, and access by the sea requires a long trip along the northern coast.

228


3.2.). In all, we were able to identify five lexical items whose distributions suggest dialect boundaries. ‘winnowing tray’: layo vs. laye vs. bichèt. According to Orjala (1970:3738), layo is northern, laye is general, and bichèt is central and southern. His description is partly confirmed by map 772 ‘van’: with one exception (point 12), layo occurs only in the North (1 to 4), and bichèt only in the South (16 to 20); 6 laye is spread almost everywhere. Two isoglosses can be drawn on this map. ‘inga tree’: pwadou vs. sikren (map 1694 ‘Inga vera Wild’). An isogloss separates the northern area, where the tree is called pwadou (1 to 5), from the remaining territory, where its name is sikren. Point 6 proves to be a transitional zone where both terms are known. 7 ‘sesame’: jijiri vs. wowoli (map 1640 ‘sésame’). Jijiri is the northern term, wowoli (or roroli) occurs in the Center and the South. An isogloss can be traced either south of points 5 and 6 or including 7 and 8, where both terms are used. ‘arrow-root’: alawout, awout vs. sagou (map 1637 ‘arrow-root’). Sagou is confined to the South; the term was elicited at points 16 (alongside with alawout), 17, and 20, but is not known (or no longer known?) at 18 and 19. ‘gate’: pòtal vs. bayè (map 680 ‘porte à clôture’). With the exception of point 3 showing bayè, the northern points 1 to 4 have pòtal/pòtay; informants at 5 and 6 named both types. 2.2 Morphophonological and morphosyntactic items The examination of morphophonological and morphosyntactic items was, on the whole, more rewarding, and most of Valdman’s examples (1978:287) proved to be specific features of the northern or southern diatopic variety. Nasalization of the postposed article. Non-nasalized forms of the postposed article after nasal vowels and consonants (e.g. baton-la “the stick,” semèn-la “the week”) are, according to Valdman, a feature of the northern dialect. They are, however, not attested on the ALH maps 1973, 1974, and 1975; on map 1976 (mont-lan, bank-lan) the variant -la does occur, but more often in the Center and the South than in the North: bank-la at points 3, 4, 8, 10, 14, 16; mont-la at 6, 8, 10, 14, 16; sant-la at 10, ponch-la at 12.

6 We disregard one attestation at point 2, which may, however, represent the relic of a wider distribution also in the northern area. Since bichèt must be the older word (some informants mentioned that it was used “en milieu rural” “in rural areas” [ALH 1:353]), this is perhaps one of the cases where older variants are maintained at the peripheries, being gradually replaced by innovations coming from the Center (cf. section 3.2). 7 It is interesting to note that, according to his statement quoted in the information given for this item, the informant at this point is aware of the diatopic difference.


229

Pronoun li “s/he, it” and progressive marker ape. Map 1989 confirms the distribution of the personal pronoun “s/he, it” described by Valdman: i in the North (1 to 5) vs. li/l in the Center and South, with point 7 having both li and i. The variant pe of the progressive marker ape/ap/pe, supposed to be le trait méridional le plus connu (“the most well-known southern trait”), is indeed confined to the South on map 2016 (16, 18, 19, and 20), but is rivalled at every point by ap, which is spread over the whole territory. Isoglosses can be established on both maps. Possessive constructions. The use of a in possessive constructions (papaa-y “his father,” sè-a-mwen “my sister” vs. papa-li, sè-mwen) is typical of the northern variety, as confirmed by map 1985 (‘déterminants possessifs’), on which an isogloss can be drawn isolating points 1 to 5. Another specific feature of the North is kin a + pronoun: kin-a-m “mine,” kin-a-y “his/hers,” vs. pa + pronoun: pa-m “mine,” pa-l “his/hers” in the South (map 2007 ‘le mien/la mienne’; cf. Valdman 1978:208). The isogloss dividing the two constructions groups together points 1 to 5, to which 7, where both variants appear, may be added. Plural with la yo. To the features characterizing the North mentioned by Valdman, we add one which is typical of the South. The plural is marked with postposed yo in Haitian Creole; in the South, however, this marker is preceded by la (or sandhi-variants a, an, nan). These specific forms of the plural appear on maps 1971 to 1976 from points 11 to 19, but not always with the same distribution and never at 17 and 20. 2.3 Phonetic variation Valdman mentions only one example of phonetic variation opposing the North to the Center and the South: the forms of the word ‘water.’ dlo/djo/glo “water” (< French de l’eau “some water”). 8 On his map (based on Orjala 1970:54), glo is attested once, near Jacmel; djo, the traditional variant (diau is found in Ducœurjoly [1802, 2:312]), is dominant in the North, where it coexists with dlo, the only variant found in the Center and the South. The ALH does not confirm this. Words for “water” occur on maps 16 (‘l’eau stagne’), 17 (‘couler’; ‘saliver’), 26 (‘la rivière est en crue’), and 27 (‘verbes pour exprimer le bruit de l’eau...’). On all these, maps only dlo is attested, with one exception: lolo “saliva” on map 17, point 4. 3. Innovations in the Center, traditional forms in the North and the South 3.1 Phonetic features Valdman illustrates the second category of geographic variation with nine lexical items (1978:288-289). Seven of them contain phonetic variants which 8

In French dialects, yo [jo] is a frequent variant of eau (FEW 25:63a).

230


are not confined to these words but are representative of phonetic developments from French to Creole to be discussed below. We assume that this holds true also of deja vs. dija “already” (< French déjà), an item not found in the ALH. Within the limited framework of this contribution, we have refrained from looking for other examples with similar phonetic structures. bliye/bilye/blilye “to forget” (< French oublier). No pattern of sound change is represented by this item. The variant bilye is very likely a survivor from French dialects; 9 according to Orjala (1970:73) and Valdman, it is used in the North and the South together with bliye and blilye. A comparison of ALH map 548, where blilye does not occur at all and bilye only at 1, 8, 10, and 15 (at 8 alongside with bliye) with Orjala’s map reveals that the ‘modern’ form bliye must have made rapid progress in replacing the traditional variant. swaf/swèf “thirst” (< French soif ), pwav/pwèv “pepper” (< French poivre), bwat/bwèt “box” (< French boîte). The French diphthong [oi] (> [wε] in Old French, derived from Latin Ē, and other sources) underwent two developments beginning in the fourteenth century: [wε] > [wa] or [ε]. The latter form was accepted by grammarians of the sixteenth century only in some cases—for example, the endings of the imperfect and conditional (avait “had3RD-SING-IMP,” était “was-3RD-SING-IMP”). The pronunciation [wa] was confined to substandard varieties until the French Revolution. The colonies received all three variants: There are several examples of [ε] and [wε] in Ducœurjoly: crère “to believe” (< French croire), drèt, drouet (“et quelquefois droit”) “straight, right” (< French droit), endrèt “place” (< French endroit), frèt “cold” (< French froid), moue, toué “I, you” (< French moi, toi) (2:307,311,313,319,334,352). According to Valdman, who gives the examples mentioned above (boîte with a map on p. 289), the younger forms swaf, pwav, and bwat are found in the Center and coexist with swèf, pwèv, and bwèt in the North and the South. The ALH data (987 ‘soif’; 923 ‘poivre’; 721 ‘boîte’) do not confirm this: swaf, pwav, and bwat are dominant everywhere; swèf is found at three points (12, 14, and 15), prèv at one point (12), bwèt at two points (3 and 11). Map 279 (‘doigt’) is better suited to give support to Valdman’s hypothesis: Points 1 to 4 in the North and 8, 11, and 14 in the Center have only dwèt; 5, 7, 17, 18, and 20 have only dwat; 6, 10, 12, 13, 15, 16, and 17 have both forms. Maps 81 (‘obscurité’) and 54 (‘à droite’) have not a single form with [wε]; for map 81, only fenwa (< French faire noir “to make dark/black”) is found, and for map 54, this means that the forms drèt and drouet noted by Ducœurjoly seem to have disappeared completely. Map 88 (‘noms des étoiles’) has one single variant etwèl (point 10), and map 89 (‘étoile filante’) only shows etwal. Obviously, no isogloss dividing ‘dialects’ marked by [wε] or [wa] or both can be established from these maps, and it seems difficult to find 9

Cf. FEW 7:271a: “bess. obélié, Thaon ubelye, Mesnil-Az. oubélié, Tinch. Obellier” etc.


231

an explanation for the differences noted. Perhaps they have something to do with the meaning and frequency of the word in question: The low frequency of ‘poil du pubis’ (map 318) may account for the preservation of pwèl/prèl at all points except 5, 7, 9, and 18, which have only pwal and 8 and 14, which have both variants. Taken together, the data evidence a gradual shift from [wε] to [wa]; this can also be illustrated by map 152 (‘Noël’) showing the ‘hypercorrect’ form nwal (at 1, 5, 9, 10, and 20), which is not attested in France (cf. Fattier 2000, vol. 1:55). saj vs. say/saj “wise” (< French sage). There is no map for this word; therefore we checked the maps 445 (‘sage-femme’: famsaj), 679 (‘clôture’: lantouraj), 822 (‘ménage’), 1065 (‘corsage’), and 1139 (‘mariage’) in order to describe the development of the French ending -age. With the exception of famsay (map 445, points 1, 3, and 6) and lantouray (map 679, point 1), the North including points 1 to 4 and 6 has only forms ending in –aj. On the other hand, the Center and the South evidence a mixture of both forms, often quoted as alternatives. Tentatively, an isogloss can be established setting off points 1, 2, 3, 4, and 6 in the North, but more examples need to be checked. lay vs. lay/laj “garlic” (< French l’ail “the garlic”). No clear picture emerges from map 925 (‘ail’); lay is attested as often in the North as in the Center, and laj, contrary to Valdman’s analysis, is dominant in the South. No further examples of the development of French -ail were found. rad vs. had/rad “clothes” (< French hardes). French [h] (h aspiré “aspirated h”), which occurs chiefly in words of Germanic origin, became mute in the standard language in the seventeenth century, but has survived in dialects in the north of France, where a variant [r] is also sporadically attested. 10 The normal development of French [h] is [r] in Haitian Creole, but variants with [h] are maintained, according to Valdman, in the North and the South. The analysis of four maps has yielded very meagre results: (a) no forms with [h] on 902 (‘hacher’) or on 1016 (‘vêtement’ < hardes) 11 and (b) two variants of hanch with [h] on map 302 (‘hanche’; points 7 and 19), and two variants of haje with [h] (< hazier) on 1520 (‘plantes non cultivées’; points 16 and 19). uit vs. wit/uit “eight” (< French huit). Since ‘huit’ does not occur in the ALH, we examined maps 71 (‘nuit’), 82 (‘pleine nuit’), and 95 (‘nuage’) for the development of the French semivowel [] in Haitian Creole. Variants with [wi] are found more often in the North and in the Center than in the South, in different distributions on maps 71 and 82; for ‘nuage,’ beside nuaj/nuay and nwaj/nway, a third variant—nyaj—is attested at points 5, 6, 9, 17, and 20. No isogloss can be established. 10 Cf. Chaudenson (1974:844-845,853-854). There are several words with initial [r] derived from French words with h aspiré in Indian Ocean Creoles—for example, Réunion Creole rale “to pull” < haler; Mauritian Creole rise “to haul” < hisser. 11 The form had occurred 7 times on Orjala’s map 43.

232


3.2 Lexical items One of Valdman’s items meant to illustrate a feature of the North (‘candle’) as well as some of the lexical contrasts identified in Hyppolite and Orjala actually belong to Valdman’s second category. ‘candle’: chandèl vs. bouji (map 722 ‘bougie’). Chandèl occurs in the North from points 1 to 5 (at 3, 4, and 5 together with bouji), but also in the South (at 11, 14, and 20). Although the picture is blurred by a third lexical type—balen—attested as the only form at points 6, 7, 15, 17, and 18, and together with chandèl or bouji at 1, 5, 8, 9, 11, 12, and 19, an isogloss is visible in the North, but not in the South. ‘taro’: tayo vs. malanga (map 1636 ‘taro’). Isoglosses can mark the distribution of tayo, found in the North down to points 7 and 8, and in the South at 16, 17, and 20. However, malanga, the only lexical type in the Center, has invaded the peripheral areas, as is evidenced by two lexical items given at points 2, 3, 4, 7, 16, and 20. In the South a third synonym—kalayib—is registered at three points (16, 18, and 19). ‘garden’: plas is northern and jaden is general according to Orjala (1970:38). On map 1517 (‘parcelle cultivée’), plas is found only at point 3 in the North and in the South at 16, 18, and 19. In each case, plas is found together with jaden, which is indeed general. ‘kite’: kap vs. sèvolan (map 1309 ‘cerf-volant’). At the periphery, sèvolan has been maintained at points 3, 4, 5, 17, 19, and 20; at 3, 5, and 20, it has been rivalled by the central type kap. ‘twins’: jimo/jumo vs. marasa (map 1127 ‘jumeau’). Jimo is spread in the North down to points 7 and 8, where its rival marasa appears. At points 6 and 7, the two lexical items are semantically distinct: Jimo designates twins of the same sex, marasa a boy and a girl. In the South, jumo is found only once at point 19, synonymous with marasa. 4. Conclusion Starting with the items presented in previous studies of geographic variation in Haitian Creole, we scrutinized about 300 maps of the ALH. Our conclusions can, of course, be only tentative; further research is needed. We were able to establish 12 isoglosses in the North (8 lexical, 2 morphosyntactic, 1 morphophonological, 1 phonetic) and 7 in the South (5 lexical, 1 morphophonological, 1 morphosyntactic). Those in the North clearly form a bundle, between points 4 and 8, with points 7 and 8 sometimes proving to be a transitional zone. In the South no bundle emerges: Isoglosses appear between 10 and 11, 15 and 16, 18 and 19, and 17 and 19. While a limited number of lexical contrasts—often related to words of very low frequency—certainly does not suffice to establish distinct dialects or to impede communication, the morphosyntactic items are well suited to give support to the speakers’ impression that


233

the speech of people from the North or the South is different from that in the Center, which is considered as a kind of standard. Valdman’s hypothesis that lexical types and phonetic features of the central variety—never marked by traits which do not exist in the North or the South—invade the periphery, is fully confirmed. The differences between the pictures emerging from previous research and from the ALH are noteworthy. Two explanations seem possible: Either the results obtained by Hyppolite and Orjala are not sufficiently founded on an empirical basis, or diatopic variation is disappearing so rapidly that Fattier’s atlas, the fieldwork for which was conducted from 1982 to 1985, came too late to document the dialects in the form they existed before 1970.

References ALH: Fattier, Dominique. 2000. Contribution à l’étude de la genèse d’un créole: L’Atlas linguistique d’Haïti, cartes et commentaires. (6 vol.). Villeneuve d’Ascq: Les Presses Universitaires du Septentrion. Chaudenson, Robert. 1974. Le lexique du parler créole de la Réunion. Paris: Champion. Ducœurjoly, S.J. 1802. Manuel des habitans de Saint-Domingue (2 vol.). Paris: chez Lenoir. FEW: Wartburg, Walther von. 1922-2003. Französisches Etymologisches Wörterbuch: Eine darstellung des galloromanischen sprachschatzes. 25 vol. Bonn: Klopp, 1928; Leipzig-Berlin: Teubner, 1934 & 1940; Basel: Helbing & Lichtenhahn, 1946-1952; Basel: Zbinden, 1955-2003. Hyppolite, Michelson P. 1949. Les origines des variations du créole haïtien. Port-au-Prince: Imprimerie de l’Etat. Orjala, Paul R. 1970. A dialect survey of Haitian Creole. Unpublished doctoral dissertation, Hartford Seminary Foundation. Valdman, Albert. 1978. Le créole: Structure, statut, origine. Paris: Klincksieck. ----------. 1987. “Le cycle vital créole et la standardisation du créole haïtien”. Etudes Créoles 10. 107-125. ----------. 1991. “Decreolization or dialect contact in Haiti?” Development and Structures of Creole Languages: Essays in honor of Derek Bickerton, Byrne & Huebner (eds) 1991 (Amsterdam: John Benjamins). 75-88.

INTERROGATIVE PRONOUNS IN LOUISIANA CREOLE AND THE MULTIPLE GENESIS HYPOTHESIS

Kevin J. Rottet Indiana University

ABSTRACT: The complexity and incomplete documentation of the pronominal system of Louisiana Creole (LC) has been pointed to in the literature (Neumann 1985). This is potentially problematic, as creole interrogative systems have been of interest in recent years in addressing a variety of issues regarding creolization (e.g., Chaudenson 2003; Clements & Mahboob 1999; Muysken & Smith 1990). In this chapter, I undertake to shed some light on this system as concerns, for instance, the use of a focus particle. I also show how interrogative data can be brought to bear on a recent scholarly debate concerning the number of geneses of Creole in Louisiana. In particular, I will provide evidence for the position taken by Klingler (2000), by giving further linguistic evidence that the Mississippi variety of LC has more in common with Haitian Creole than does the Bayou Teche variety. 1. Introduction In this chapter my purposes are twofold. First, I will seek to characterize the system of interrogative pronouns in Louisiana Creole (LC) in order to clarify certain features of that system as compared with other French-based creoles. Certain facts about Creole interrogatives have been of interest in the recent scholarly literature, in particular interrogative focus particles (cf. Clements & Mahboob 1999), the taxonomy of Creole question word types (cf. Muysken & Smith 1990), and the question of a substrate model for bimorphemic interrogative patterns (Chaudenson 2003: 286-300; Lefebvre 1993). The inclusion of LC material in such studies has sometimes been problematic due to incomplete documentation of the LC system. My study, based on a compilation of all available data from 13 published LC sources, goes some way towards clarifying the descriptive facts of the complex system of LC pronouns. 1 1

Neumann (1985:333 fn. 1) wrote: “Since the system of interrogative pronouns in LC is fairly complex, it goes without saying that this list [of interrogative forms] is not exhaustive. More research is needed in this area, especially as concerns the distribution and uses of the various forms in LC and in Cajun varieties.” [My translation-KR]

236


My second purpose is to investigate how these data shed additional light on a recent scholarly debate concerning the number of separate geneses of Creole in Louisiana. This debate arose out of an attempt to account for differences between two regional varieties of LC: an eastern variety spoken along the Mississippi River including Pointe Coupee Parish 2 (henceforth MIS), and a western variety spoken along Bayou Teche, including St. Martin Parish (henceforth TEC). 3 Both of the main contemporary descriptions of LC— Neumann (1985) and Klingler (2003)—acknowledge that there are differences between these varieties, although neither apparently viewed their existence as anything other than typical dialect differentiation. Speedy (1994, 1995), on the other hand, argued that the kinds of linguistic features differentiating MIS and TEC (e.g., relative clause structures) are incompatible with an assumption of shared origin and require postulating at least partially independent creolization events; furthermore, Speedy suggested that TEC was influenced in its formative stage by Haitian Creole (HC) when over 3000 slaves from Saint Domingue (modern Haiti) were brought to Louisiana in 1809-1810 by slaveholders fleeing the slave insurrection there. Quoting Griolet (1986:70), Speedy (1994:105) maintained that “the vast majority” of these slaves settled in the Teche region of Louisiana. In his published reply, Klingler (2000) 4 challenges Speedy’s view, arguing for a single genesis of LC along the Mississippi River in communities above and below New Orleans. In Klingler’s view, LC was later carried from the Mississippi to the Teche region. He further provides evidence that in the majority of documented cases, the slaves imported from Saint Domingue stayed in New Orleans and in the Mississippi River communities. Therefore, if either variety of LC were to show HC influence it should be MIS rather than TEC. Klingler goes on to show how this claim is borne out by a variety of data. In this chapter, I will examine additional differences between the MIS and TEC varieties which have not heretofore been considered in this debate, and I will show how these data lend further support to the position taken by Klingler (2000). I will examine this evidence, from the interrogative system of LC, in section 3. First, though, it will be necessary to lay out the facts regarding the interrogative pronoun system of LC. It is to this preliminary task that I turn in the next section.

2

Counties are called civil parishes in Louisiana. For TEC, the primary sources are Lane (1935), Broussard (1942/1972), Morgan (1960), and Neumann (1985). The material from Wiltz (1991-1993) and Ancelet (1994) also comes from the TEC region. For MIS, the sources are Fortier (1884), Fortier (1895 /1972), Jarreau (1931), Klingler (1992, 2003), Lavergne (1930), Mercier (1881), Neumann-Holzschuh (1987), and Wogan (1931). 4 The topic is further explored in Klingler (2003:69-92). 3

INTERROGATIVE PRONOUNS IN LOUISIANA CREOLE

237

2. Interrogative pronouns in Louisiana Creole The descriptive sources on LC describe a large number of interrogative pronoun forms that are not functionally or semantically differentiated. 5 Nonetheless, certain patterns emerge from a compilation of all of the direct and indirect questions 6 using interrogative pronouns occurring in the following 13 sources: Ancelet (1994), 7 Broussard (1942/1972), Fortier (1884, 1895/1972), Jarreau (1931), Klingler (1992), 8 Lane (1935), Lavergne (1930), Mercier (1881), Neumann (1985), Neumann-Holzschuh (1987), Wiltz 9 (1991-1993), and Wogan (1931). These texts collectively furnished 325 questions containing an expression of “who” or “what.” Table 1 shows the distribution of these 325 occurrences in the following categories: Subject ([-/+human]); Direct Object ([-/+human]); Copulatives ([-/+human], e.g., ‘What is [an] X?’ or ‘Who is in the kitchen?’); and Other ([-/+human]). The category Other is a catch-all for the following: cases where the interrogative is a complete utterance by itself, as in (1); wh-in situ questions as in (2); cases where the interrogative occurs in final position as in (3); cases where an interrogative is the object of a preposition as in (4); and the nine occurrences of /sa ina/ “What’s wrong?” (cf. French Qu’est-ce qu’il y a?) as in (5): 10

5

(1)

Si to di li kechòz andan kreol l’ale monde twa ‘Kisa?’ (Neumann 1985:358) 11 “If you tell him something in Creole he’ll ask you, ‘What?’”

(2)

Se te ki? (Neumann 1985:394) “It was what?”

Some writers have just listed LC interrogative pronouns without seeking to determine how they might be differentiated. Others, such as Lane (1935:12), a description of Saint Martinville LC, suggest that there is some differentiation according to grammatical function and animacy. Some of his claims, however, (e.g., that kisa is used only for [+ human] objects), are too restrictive to account for all of the attested data. 6 Pronouns used in direct and indirect interrogatives are not differentiated in LC (Klingler 1992:214). 7 In Ancelet (1994), texts 3, 7, 9, 11, 12, 15, 30, 102, and 103 are in LC. (Some of the speakers in the collection who were identified as ‘Creole’ nonetheless told their tales in Cajun French rather than LC and were therefore excluded). 8 Only volume 1 (i.e., pages 1-252). 9 Wiltz (1991-1993) refers to a series of short articles entitled “La Leson Kreyol” which appeared in CREOLE Magazine. I only had access to issues of the magazine for a two year period in 1991-1993. 10 I group these various uses together as Other because there are too few tokens to identify potential patterns. 11 I normalize the spellings from Neumann (1985) and Klingler (1992) to conform to the system used in Valdman, Klingler, Marshall, and Rottet (1998). I do not alter the spellings of published texts which used a Frenchified orthography.


238

(3)

Ye te met le zonyon la epi ki? (Neumann 1985:394) “They put in onions and what else?”

(4)

Kote ki moun to ye isi? (Klingler 1992:217) “Who are you living with/Whose place are you staying at here?”

(5)

Me sa ina ek twa, Bouki? (Neumann 1985:436) “But what’s wrong with you, Bouki?”

The data in the columns of Table 1 labeled “Subject” and “Direct Object” suggest greater differentiation of pronouns according to grammatical function than according to animacy. With few exceptions, the same pronouns are used with both animate and inanimate referents, the interpretation as ‘who’ or ‘what’ being determined from context. 12 Grammatical function is generally indicated by the presence or absence of a final element ki. The ki-final forms sa ki, ki-sa ki, ki-se-sa ki, and ki ki are used as interrogative subjects. These forms are illustrated in (6) through (11). Note that the vowel of this final /ki/ is elided before a following vowel-initial word as seen in items (7) and (10). (6)

Ça qui tchué li? (Wogan 1931:8) “What killed him?”

(7)

Sa k’ale monje le chat? (Neumann 1985:412) “Who is going to eat the cat?”

(8)

Qui ça qui senti si bon dans chaudière la, Compair Lapin? (Fortier 1895/1972) “What smells so good in the kettle, Brer Rabbit?”

(9)

Li mandé so fame ki ça ki mangé so pichetache. (Fortier 1887) “He asked his wife who ate his peanuts.”

(10) Ki-se-sa k’ape pele? (Neumann 1985:430) “Who is calling?”

12

There may well be occasional instances where the pronominal form is truly ambiguous, even in context, but I have not found any such instances in this corpus of examples. I should point out that most of the texts used here included a published translation, either in French or in English, and in such cases I have always followed the published translation. In no case did I disagree with the translator’s decision regarding animacy.


Subject [+/– human] [– ] [+]

Pronoun

Direct object [+/– human] [– ] [+]

239

Copula [+/– human] [– ] [+]

Other [–/+]

No final ki

sa

0

0

ki-sa

0

0

ki-se(-sa)

0

0

0

3% (1)

61% (14) 26% (6) 4% (1) 4% (1) 0

4% (1) 23

ki moun Final ki sa ki

ki-sa ki

ki-ki

ki-se(-sa) ki

ki moun ki Bare ki ki N=325

Note.

76% (124) 16% (27) 2% (4) 0

10% (1) 20% (2) 20% (2) 10% (1)

7% (4) 19% (10) 20% (11) 0

38% (10) 8% (2) 8% (2) 8% (2) 19% (5)

1% (1) 0

10% (1) 0

19% (10) 0

0

0

0

10% (1) 0

0

0

7% (4) 0

15% (4) 26

5% (8) 164

20% (2) 10

27% (15) 54

0

0

29% (6) 9% (2)

33% (9) 13 11% (3) 4% (1) 4% (1)

5% (1) 5% (1) 9% (2) 19% (4) 5% (1)

0

19% (4) 21

48% (13) 27

0

0

0

0

TEC data is from Ancelet (1994), Broussard (1942/1972), Lane (1935), Neumann (1985), Wiltz (1991-1993); MIS data is from Fortier (1884, 1895 /1972), Jarreau (1931), Klingler (1992), Lavergne (1930), Mercier (1881), Neumann-Holzschuh (1987), Wogan (1931). The shaded areas represent apparent violations of the principles of focus particle distribution.

Table 1: Interrogative pronouns by function in 13 LC sources (11) Ki ki grouy la? (Neumann 1985:185) “Who/What is moving there?” 14 Unlike the above forms, the pronoun ki moun (sa) ki is restricted to [+human] reference. It only occurs in the MIS variety, not in TEC: (12) yé té conain qui moun qui té cognain (Neumann-Holzschuh 1987:125) “They knew who had struck.”

13 These nine tokens are all occurrences of the question Ça ina? “What’s wrong?” (cf. French Qu’est-ce qu’il y a?) 14 Neumann provided both translations for this example, which occurs in her discussion of LC interrogatives. The source of some of her metalinguistic examples, including this one, is not specified (i.e., whether they occurred in actual speech or were elicited in some way).

240


Direct objects are expressed with pronominal forms that lack final ki. Ki-sa and ki-se-sa can have either [+human] or [-human] referents: (13) Qui ça t’olé? (Broussard 1942/1972:10) “What do you want?” (14) Kisa to wa? (Lane 1935:12) “Who did you see?” (15) Ki-se-sa to wa yer? (Neumann 1985:186) “Who did you see yesterday?” Object sa, by far the most common LC interrogative, is overwhelmingly [-human]: 15 (16) Ça to gaignin dan panier là? (Wogan 1931:14) “What do you have in the basket?” Ki moun (sa) is restricted to [+human] uses, but again it only occurs in MIS, not in TEC: (17) Qui moune ça la palé marié? (Wogan 1931:18) “Who is she going to marry?” In copula questions, including equative, locative, and predicative uses, appearance of the focus particle ki seems to be quite variable (33 of these items lacked ki and 23 had it): (18) Pa konen ki moun se. (Klingler 1992:214) “I don’t know who it is.” (19) Qui moun ça qui là? (Wogan 1931:18) “Who’s there?” (20) Ki-sa ki nouzot kandida? (Neumann-Holzschuh 1987:99) “Who is our candidate?” (21) Ki-se-sa so bon non? (Neumann 1985:354) “What is his real name?” 15

Neumann (1985:334) notes that the bare form sa is pejorative in reference to people; her illustrative examples on the same page include the only instance of [+human] direct object sa that I have found.


241

(22) Hey, ga, qui ça ça yé, to cou corché. (Neumann-Holzschuh 1987:121) 16 “Hey, look, what is that, your neck is scratched.” It is possible that there are some generalizations to be made about copulatives, but more data would be needed to discover any such patterns. 17 LC also makes use of a pattern that I refer to as ‘bare ki,’ referring to instances where the pronoun ki functions entirely by itself. This usage is attested both for animates and inanimates, and in all grammatical functions: (23) Ki apé tué mo piti? (Mercier 1881:170) “Who is killing my child?” (24) Me Bouki, ki sa va som, gen en gro dine kom sa e nou p ale gen bal? (Neumann 1985:396) “But Bouki, what will it look like if we have a big dinner like that and we don’t have a dance?” Bare ki is also used in tag, echo, and wh-in situ questions and after a preposition, again with both [+human] and [-human] reference: (25) Ye te met le zonyon la epi ki? Se te ki? (Neumann 1985:394) “They put in onions and what else? What was it?” (26) Mais pou qui to prend moin, Jean Sotte? (Fortier 1895/1972) “But who do you take me for, Foolish John?” (27) On ki to mèt li? (Lane 1935:12) “What did you put it on?” (28) To konen mo non? ... Ki mo ye? (Klingler 1992:188) “Do you know my name? Who am I?”

16

In analyzing the questions ki sa ye?, sa sa ye?, and ki sa sa ye? I took the sa immediately in front of the copula ye to be its subject, and therefore not part of the interrogative pronoun. (The subject pronoun appears to be obligatory in this construction; I have not found *Ki ye?) Therefore, I counted occurrences of ki sa ye? as tokens of ki, occurrences of sa sa ye? as tokens of sa, and occurrences of ki sa sa ye? as tokens of ki sa. 17 Some varieties of French also show variation in some of these patterns—for example, Qu’est-ce que c’est (qu’) une aubergine? “What is an ‘aubergine’?”

242


The use of a focus particle in creoles has been of some interest in recent years. For instance, Clements and Mahboob (1999) made predictions about what interrogative patterns (including use or non-use of a focus particle) would be expected in various creole languages by comparing interrogative features in lexifier and substrate languages. Their prediction for LC was that it should have such a particle. Given the brief presentation of relevant data in their LC source (Neumann 1985:333-338), however, they concluded that LC did not appear to have such a particle and was therefore a counterexample to their prediction. 18 The data examined in my study, based on a much larger pool of sources, show that LC in fact does have the predicted focus particle. Function subjects direct objects

Final -ki 98% (43) 2% (3)

No final -ki 2% (1) 98% (161)

Table 2: Correlation of grammatical function with final -ki In fact, the correlation between use of ki with subjects and absence of ki with direct objects is very high in the data (see Table 2): 19 98% (43 out of 44 tokens) of subject interrogative pronouns were ki-final, whereas only 2% of direct objects (3 out of 164 tokens) had final ki. (The four deviant tokens out of 208 total will be examined below.) In its use of a focus particle (ki) with subjects, LC patterns with a number of other French-based creoles and even with French dialects (e.g., Koopman 1982 on HC, Lefebvre 1982 on Montreal French). Nevertheless, there are four exceptions to this pattern in the data examined. These are the non-zero figures in the grey areas of Table 1. One of these is an occurrence of an interrogative subject without final ki: (29) Ki moun ki gen peye mon, ki moun gen peye vou? (Klingler 1992:214) “Who will pay me, who will pay you?”

18

The data Neumann presents in her metalinguistic discussion, which she gives side by side with frequently occurring types, provide some of the few attested counterexamples to my analysis. For instance, she gives Ki/ki-se/ki-se-sa/ki-ki/sa/sa-ki to wa yer? ‘Qui as-tu vu hier?’ (“Who did you see yesterday?”). This is in fact the only occurrence of sa as a [+human] direct object, and the ki-ki and sa-ki examples account for two of the three counterexamples in my corpus to the otherwise robust distribution in which no focus particle occurs with direct objects. For an explanation for these counterexamples, see below. 19 The uses of bare ki are excluded from the counts in Table 2 since bare ki appears in every function, as noted above. That is, it is able to function both like ki-final and like non ki-final forms.


243

It is noteworthy that the first occurrence of ki moun above has the expected ki whereas the second does not. Its ellipsis in the second occurrence may simply be a matter of economy. Item (29) is therefore not a compelling counterexample. Conversely, there are three occurrences in the corpus of the focus particle ki with a non-subject. They are given below: (30) Mo pa gen di li sa ki vou fe dan la kizin. (Klingler 1992:137) “I won’t tell her what you do in the kitchen.” (31) Ki ki to wa yer? (Neumann 1985:334) “Who did you see yesterday?” (32) Sa ki to wa yer? (Neumann 1985:334) “Who did you see yesterday?” There is a plausible explanation for these counterexamples. Basilectal nineteenth century LC typically used a Ø relativizer with direct object antecedents and a Ø focus particle with non-subject interrogatives. Mesolectal or Cajuninfluenced varieties of LC, however, sometimes have an overt object particle, which takes the form /ke/ or occasionally /ki/ (Neumann 1985:162). In the present corpus, for instance, there is one occurrence of ké in an object interrogative: (33) Ki ça ké vou gagné pou dézolé com sa? (Neumann-Holzschuh 1987:81) “What is wrong with you (lit. “what do you have”) that you are despairing so?” Klingler (2000:14-15) also gives examples of an overt object relativizer, and in all three of the examples he provides, it occurs as /ki/: (34) Ye petet di mo de paròl an kreyòl ki mo pa konpròn. (Klingler 2000:14) “They might tell me some words in Creole that I don’t understand.” Apparent counterexamples such as (30)-(32) above would appear to be instances of the mesolectal object particle /ke/ ~ /ki/. Indeed, Neumann (1985) lists ki ki/ke and sa ki/ke, implying that the focus particle could take the form ki or ke interchangeably. This hesitation between /ke/ and /ki/ may reflect uncertainty among speakers of Cajun-influenced varieties of LC about when to use a focus particle or which one to use. This would be a kind of hypercorrection, in

244


which the speaker is aware that introduction of a focus particle accompanies movement along the linguistic continuum toward French (i.e., the vernacular French of Louisiana, see Rottet 2004), but has not fully internalized their distribution. Apart from the four apparent counterexamples discussed above, the interrogative data from LC suggest that this language does use a focus particle ki with subjects but not with direct objects, thus bringing this variety in line with the prediction of Clements and Mahboob (1999). Having laid out the basic descriptive facts regarding the interrogative system of LC, we now proceed to examine how these data can shed further light on the question of the number of geneses of Creole in Louisiana. 3. The influence of slaves from Saint Domingue on LC At the beginning of this chapter I referred to a recent scholarly debate between Speedy and Klingler, based on observed differences between the two regional varieties of LC (MIS and TEC), which have implications for the number of geneses of a French-based creole in Louisiana. Various differences between MIS and TEC have been discussed in that debate, although interrogative data have not been brought to bear on the question. In my examination of the interrogative pronouns of LC, an additional difference between these varieties emerged—namely the fact that MIS has a form ki moun, unattested in TEC, which can be used for [+human] reference. Ki moun also, of course, occurs in HC, and therefore it is another important feature which MIS and HC have in common but which TEC does not share. As such, it constitutes further evidence for the position in Klingler (2000), who provided demographic data showing that the majority of slaves imported in 1809-1810 from Saint Domingue stayed in New Orleans and in communities along the Mississippi River, rather than settling in the Teche region as Speedy (1994, 1995) had claimed. Therefore, Klingler reasoned, if either variety of LC were to show HC influence it would be MIS, not TEC. It is not the mere occurrence of ki moun in MIS which I take to be significant. A much more concrete reason to see this form as a borrowing from HC into MIS is that it is an isolated form in LC. Ki moun is unusual in being the only clear example in LC of a bimorphemic interrogative in the sense usually intended in creole studies. 20 In bimorphemic interrogatives, a question particle is combined with a second element, the QSU or ‘questioned semantic unit’ (Muysken and Smith 1990). The question particle takes the form we, wi, or wa in English-based creoles, ke in Portuguese-based creoles, and ki in Frenchbased creoles. The QSU following this particle is typically a noun such as an equivalent of ‘thing,’ ‘person,’ or ‘time.’ By way of illustration, the bimor20 See Muysken and Smith (1990) and Clements and Mahboob (1999) for some recent discussions.


245

phemic interrogatives (adverbs and pronouns) of HC are as follows (Koopman 1982): (35) ki moun ki bagay 21 ki kote ki bò ki lè ki jan

“who” “what” “where” “where” “when” “how”

(Q + ‘person’) (Q + ‘thing’) (Q + ‘side’) (Q + ‘side’) (Q + ‘time’) (Q + ‘manner’)

The above system is illustrative of the radical restructuring involved in the evolution of HC. Only HC konbyen “how much/how many” and kouman, a variant of ki jan “how,” are obvious reflexes of French interrogatives. In LC, however, interrogative adverbs are all derived rather transparently from their French equivalents (standard or dialectal): (36) eyou, ayou ekan, kan ki lè, kèl œr kofèr pourkwa kòman konbyen

“where” “when” “at what time” “why” “why” “how” “how much/many”

(cf. French où) (cf. French quand) (cf. French à quelle heure) (cf. French dialect [pour] quoi faire) (cf. French pourquoi) (cf. French comment) (cf. French combien)

The LC interrogative adverbs which are bimorphemic, such as ki lè “at what time,” are inherited from French phrases of the same pattern. 22 Despite the clear French origins of the interrogative adverbs of LC, it might be tempting to view the LC pronoun ki sa as a bimorphemic interrogative by analogy with HC ki bagay or Guadeloupean Creole (GC) ki biten “what,” consisting of a question particle ki plus a QSU meaning “thing.” In my view, such an analogy would be mistaken. First of all, in clear cases of bimorphemic interrogatives, it is the QSU which supplies the interpretation of the expression; the interrogative particle simply serves to make it a question word. Yet in LC, ki is attested alone with precisely the same meanings (i.e., “what” and

21

HC also has ki sa “what,” which is often shortened to sa. Although this form is probably seen as a bimorphemic interrogative synchronically in HC, I take it that its origin is probably the same as for ki sa in LC as discussed below. HC ki sa also occurs in pouki sa “why.” 22 Harris (1973) considered the form kofèr “why” an African borrowing. This view is rejected by Chaudenson (2003:295-296) who compellingly points to French dialect sources for quo’faire and related forms.

246


“who”) that ki sa has as a unit (see the examples of ‘bare ki’ discussed early).23 It therefore does not need sa to supply its interpretation. Furthermore, sa is attested in LC after other interrogative words besides ki: (37) Où ça mo té jamin bésoin dolo? (Fortier 1884:104) “Where did I ever need water?” (38) Komon sa t’ape ongrese? (Neumann 1985:434) “How can it be that you’re gaining weight?” (39) Qui moun ça qui là? (Wogan 1931:18) “Who’s there?” Lane (1935:12) even gives /usa/ as the usual word for “where” in the LC of Saint Martinville. It appears therefore that, at least originally, sa was an element used to reinforce a number of LC interrogatives; its distribution was not limited to pronominal forms meaning “what” and therefore it is not analogous to the QSU elements bagay and biten of HC and GC. A reinforcing use of ça, as a popular variant of c’est, was common in the French of the colonial period and later (Chaudenson 1989:124) and can be seen in the following nineteenth century literary example: (40) Qui ça qui m’a vu? (Maupassant, La ficelle) “Who saw me?” Réunionnais preserves this usage clearly in that most of its interrogatives have incorporated sa—for example, ki sa? “who?”; ko sa? “what?”; ou sa? “where?” (Chaudenson 1989:109). A similar usage is still common in modern spoken French with interrogatives used as complete utterances by themselves: Qui ça? “Who?” Où ça? “Where?” (see Ball 2000:76-77 for such usage in modern colloquial French). It is true that sa can be used alone as “what” in LC. This can probably be understood as an extension of the use of sa as a relative pronoun meaning “that which, what” (cf. Cajun French ça qui, ça que) as suggested by Goodman (1964:52), and illustrated in (41): (41) Cé pa in capon comme toi ka fé moin largué ça mo tchombo. (Mercier 1881:298) “It’s not a coward like you who can make me drop what I’m holding.” 23

Muysken and Smith (1990:897) similarly conclude that Krio we is not “a true Q element, since it occurs on its own, as well as with a QSU element deriving from English ‘thing’.”


247

The ability of sa to function as an interrogative pronoun by itself would then be a later development, facilitated by the homophonous relative pronoun sa illustrated in (41). 24 Let us come back then to the one clearly bimorphemic interrogative in LC which is not based on a French model, 25 namely ki moun “who” (cf. moun “person”). One could speculate that in earlier periods, LC may have had a fuller set of bimorphemic interrogatives, and that ki moun is the only survivor from that period. However, the possibility that it entered MIS via borrowing from the HC-speaking slaves who settled in communities along the Mississippi River (in keeping with the view of settlement patterns argued for in Klingler 2000) is a more plausible explanation, and one that offers the additional advantage of accounting for the occurrence of ki moun in MIS but not in TEC. 26 4. Conclusions In this chapter I have tried to do several things: first, to set forth some numerical data concerning the distribution and use of interrogative pronouns in LC in order to go some way toward clarifying what is clearly a complicated and incompletely understood system; secondly, to clearly document the existence of a focus particle ki used with subject but not direct object interrogatives as predicted, but not found, by Clements and Mahboob (1999); and thirdly, to show how the form ki moun, occurring in the MIS but not the TEC variety of LC, provides a further piece of evidence for the position taken in Klingler (2000) who argued for a single genesis of LC with later diffusion, and the settlement of a large contingent of slaves from Saint Domingue in the MIS rather than in the TEC region.

24

Unfortunately, the first attestation of a question with “what” in LC does not occur until 1846 (see Neumann-Holzschuh 1985:24). By that time LC had existed for a century or more, so it is impossible to say exactly when sa emerged as an independent interrogative pronoun. It does occur in the 1846 text. 25 Although LC forms like ki sa (ki) may not at first glance look like obvious reflexes of French forms, a closer examination of interrogatives in Louisiana French suggests otherwise. In Rottet (2004), I discuss the geographical distribution of inanimate interrogatives using quoi and those using qui, and I argue that the data strongly support the conclusion that inanimate qui forms (e.g., qui c’est qui or qui ça qui “what” as subject) were present in the speech of the early eighteenth century settlers in Louisiana who spoke what is usually called Colonial French. 26 There is one additional feature of the interrogative system of MIS which is shared with HC but not with TEC: the form of the interrogative adjective. In TEC, this takes the forms ke or kel (Broussard 1942/1972:10; Lane 1935:12; Morgan 1960:14; Neumann 1985:338)—for example, Ke mezon t’ap reste òndon? “What house do you live in?” (Neumann 1985:338). In MIS, the usual form is ki (Klingler 1992:216-17)—for example, Dan ki mwa nou ye astè la? “What month are we in now?” Ki is also the form of the interrogative adjective in HC (Koopman 1982:222).

248


References Ancelet, Barry. 1994. Cajun and Creole folktales. Jackson: University of Mississippi Press. Ball, Rodney. 2000. Colloquial French grammar. Oxford: Blackwell. Broussard, James Francis. 1972. Louisiana Creole dialect. Reprint. Port Washington, NY and London: Kennikat Press (Original work published 1942). Chaudenson, Robert. 1989. Créoles et enseignement du français. Paris: L’Harmattan. ----------. 2003. La créolisation: Théorie, applications, implications. Paris: L’Harmattan. Clements, J. Clancy & Ahmar Mahboob. 1999. “WH-words and question formation in Pidgin/Creole languages”. Language change and language contact in pidgins and creoles, McWhorter (ed) 1999 (Amsterdam: John Benjamins). 459-497. Fortier, Alcée. 1884. “The French language in Louisiana and the Negro French dialect”. Publications of the Modern Language Association of America, vol. 1. 96-111. ----------. 1887. “Bits of Louisiana folklore”. Transactions of the Modern Language Association of America III. 100-168. ----------. 1972. Louisiana folktales in French dialect and English translation. New York: Kraus Reprints (Original work published 1895). Goodman, Morris. 1964. A comparative study of Creole French dialects. The Hague: Mouton. Griolet, Patrick. 1986. Cadjins et Créoles en Louisiane. Paris: Payot. Harris, M. Roy. 1973. “‘Kofè’ ‘Pourquoi’, un africanisme parmi d’autres en créole louisianais”. Revue de Louisiane/Louisiana Review 2. 88-102. Jarreau, Lafayette. 1931. Creole folklore of Pointe Coupee Parish. Unpublished master’s thesis, Louisiana State University, Baton Rouge. Klingler, Thomas A. 1992. A descriptive study of the creole speech of Pointe Coupee parish, Louisiana with focus on the lexicon. Unpublished doctoral dissertation. Indiana University, Bloomington. ----------. 2000. “Louisiana Creole: The Multiple Genesis Hypothesis reconsidered”. Journal of Pidgin and Creole Languages 15. 1-35. ----------. 2003. If I could turn my tongue like that: The creole language of Pointe Coupee parish, Louisiana. Baton Rouge: Louisiana State University Press. Koopman, Hilda. 1982. “Les questions”. Syntaxe de l’haïtien, Lefebvre, Magloire-Holly, & Piou (eds) 1982 (Ann Arbor, MI: Karoma). 204-241. Lane, George. 1935. “Notes on Louisiana French II: The Negro-French Dialect.” Language 11. 5-16. Lavergne, Remi. 1930. A phonetic transcription of the Creole Negro’s medical treatments, superstitions, and folklore in the parish of Pointe Coupee. Unpublished master’s thesis, Louisiana State University, Baton Rouge.


249

Lefebvre, Claire. 1982. “‘Qui qui vient? ou Qui vient? Voilà la question”. La syntaxe comparée du français standard et populaire: Approches formelle et fonctionnelle, I, Lefebvre (ed) 1982 (Quebec: Office de la langue française, Gouvernement du Québec). 47-101. ----------. 1993. “The role of relexification and syntactic reanalysis in Haitian Creole: Methodological aspects of a research program”. Africanisms in Afro-American varieties, Mufwene (ed) 1993 (Athens: The University of Georgia Press). 254-279. Mercier, Alfred. 1881. L’habitation Saint-Ybars ou maîtres et esclaves en Louisiane. New Orleans: Imprimerie Française. Morgan, Raleigh. 1960. “The lexicon of St. Martin Creole”. Anthropological Linguistics 2. 7-29. Muysken, Pieter & Norval Smith. 1990. “Question words in pidgin and creole languages”. Linguistics 28. 883-903. Neumann, Ingrid. 1985. Le créole de Breaux Bridge, Louisiane. Hamburg: Helmut Buske. Neumann-Holzschuh, Ingrid. 1987. Textes anciens en créole louisianais. Hamburg: Helmut Buske. Rottet, Kevin J. 2001. Language shift in the coastal marshes of Louisiana. Berlin: Peter Lang. ----------. 2004. “Inanimate interrogatives and settlement patterns in Francophone Louisiana”. Journal of French Language Studies 14. 169-188. Speedy, Karin 1994. Mississippi and Teche Creole: A demographic and linguistic case for separate geneses in Louisiana. Unpublished master’s thesis, University of Auckland, New Zealand. ----------. 1995. “Mississippi and Teche Creole: Two separate starting points for Creole in Louisiana”. From contact to creole and beyond, Baker (ed) 1995 (London: University of Westminster Press). 97-111. Valdman, Albert & Thomas A. Klingler. 1997. “The structure of Louisiana Creole”. French and Creole in Louisiana, Valdman (ed) 1997 (London: Plenum Press). 109-144. ----------, Thomas A. Klingler, Margaret M. Marshall, & Kevin J. Rottet. 1998. Dictionary of Louisiana Creole. Bloomington: Indiana University Press. Wiltz, Herbert. 1991-1993. “La Leson Kreyol”. CREOLE Magazine, Lafayette, LA. Wogan, Marguerite B. 1931. Cancans Kisinieres. New Orleans: Rogers Printing Co.

GENDER IN FRENCH CREOLES THE STORY OF A LOSER ∗

Ingrid Neumann-Holzschuh Universität Regensburg

ABSTRACT: The grammatical category of gender has captured the attention of linguists interested in a wide variety of languages. Despite the fact that creoles are generally counted among the genderless languages, creolists continue to consider the role of gender in creoles. In this chapter, natural gender marking in a number of French-based creoles will be discussed. This description will also take into consideration possible paths of language change, and I will reflect upon why gender lost out in the complex process of restructuring the nominal morphology that occurred during creolization. I will end by turning my attention to a recent and speculative account of a special relationship between gender and number. 1. Gender in creole languages “Gender is the most puzzling of the grammatical categories”—this is the first sentence of Corbett’s (1991) stimulating book on gender, a comprehensive survey of gender and gender systems with data from over 200 languages. Creolists, however, will look in vain for a paragraph on creole languages, which is not really surprising given the fact that creoles are considered genderless languages. No introductory work on creoles or creole grammar fails to hint at the difference between creoles and their respective base language as far as gender marking is concerned. Holm (2000:216) remarks “[l]ike distinctions of number, distinctions of grammatical gender in the European lexical source languages were not maintained by inflections on creole nouns or adjectives. […] Sometimes the creoles have preserved natural gender oppositions in specific nouns, but this is on the level of the lexical item rather than inflectional morphology.” Valdman (1978:149) notes that “compared to French, one of the salient grammatical features of creole is the absence of gender in the noun system.” Heil (1999:75) adds, “grammatical gender does not exist in the French-based creoles,” along with Bollée (1977:27) who states, “as in the other ∗ I am indebted to Karolin Heil, whose master's thesis Die Kategorie Genus und ihr Status in den französischen Kreolsprachen (Regensburg 2001) was an important source for this article.

252


French-based creoles, the category of grammatical gender does not exist in Seychellois Creole” [Editors’ translations]. It is common knowledge that in creole languages there is no classification of nouns according to masculine and feminine, although all creoles irrespective of their base language actually do express natural (biological) gender, which is most often indicated on the level of the lexical item rather than by inflectional morphology. The situation appears to be quite simple, so why discuss gender in creole languages when there is no problem? The aim of this article is twofold. First, I will briefly sketch natural gender marking in French creoles and discuss possible paths of language change. From a more general perspective, I will then ask why gender was given up during creolization, and I will briefly allude to the special relationship between gender and number against the background of more recent, typologically oriented studies on gender. Obviously the answer to the question of whether creoles have gender or not is directly linked with the definition of gender. 1 If one accepts the traditional definition of Hockett (1958:231) that “[g]enders are classes of nouns reflected in the behavior of associated words,” which is shared by Corbett, 2 French creoles certainly do not have gender; in basilectal creoles, there is no such phenomenon as congruence or agreement between noun and determiner, noun and adjective, or noun and participle. In French creoles, nouns and adjectives are uninflected items; the determiner is postnominal -la for singular nouns and either postnominal ye or yo or preposed se or bann for plural nouns. The personal pronouns li or i (singular) and ye (yo) and zot (plural) for both masculine and feminine are the equivalents of French il/elle “he/she” and ils/elles “they3 MASC/they- FEM,” respectively. Thus, natural gender expression is primarily a lexical issue, not a syntactic one. 4 According to Wurzel (1986:77), however, nominal classes can also be established without taking agreement into consideration: “Classification of nouns means the division of the nouns of a language into a limited number of classes with class membership formally taking effect on the nouns itself [sic] and/or 1

Cf. for example Claudi (1985), Corbett (1991), and Unterbeck, Rissanen, Nevalainen, and Saari (2000) for a comprehensive theoretical discussion of the status and function of this nominal category in various languages. Corbett (1991) also gives an overview of the different theories concerning the genesis of gender as a nominal category, which originally meant ‘kind or sort’ and had nothing to do with sex. 2 Cf. Corbett (1994:1348): “To demonstrate the existence of a gender system evidence is required from agreement, that is outside the noun itself.” Cf. also Claudi (1985:13ff.) and Dahl (2000a:113). 3 The plural marker bann and the plural pronoun zot are restricted to the Indian Ocean creoles; cf. Stein (1986) for a more precise overview on the distribution of the abovementioned forms in the different French creoles. 4 Valdman (1978:151) already emphazised “the lexical, rather than grammatical, nature of the indication of sex for nouns referring to human beings” [Editors’ translation].

GENDER IN FRENCH CREOLES

253

beyond the nouns in at least certain contexts” (Unterbeck 2000:xvi). 5 This view seems to be shared by Dahl (2000a:106) for whom gender appears “on one hand as a property of a noun (as a lexical item), on the other, as a property of a noun phrase (as an occurrence).”6 Given the fact that in creoles the expression of (natural) gender is above all lexically inherent in the nouns, this more encompassing view will certainly be more compatible with the creole data than definitions based solely on agreement. Furthermore, if one accepts that gender “is reserved for such noun marking systems where sexual gender is transparent, although not necessarily all-encompassing” (Hurskainen 2000:665), 7 creoles are certainly not completely genderless. Since they mark biological gender, creoles just about fulfil the basic requirements for “gender systems proper, i.e., systems showing sex differentiation within nominal classification” (Unterbeck 2000:xxvi). Thus, they differ considerably from the African substrate languages, the majority of which are characterized by complex nounclass systems (cf. Claudi 1985; Hurskainen 2000).

5

Cf. Claudi (1985:14): “By gender system I mean […] any categorizing classification of nouns that is expressed through a marking on the noun itself, on other sentence constituents, or both; the criteria for the classification may or may not be of a semantic nature (e.g., animate or inanimate, human or nonhuman, male or female, etc.)” [Editors’ translation]. 6 This view is shared by most of the recent works on gender, cf. inter alia Claudi, who emphasizes that agreement plays a much more important role with the other nominal categories (e.g., number and case), whereas gender is more of a lexical category: “Gender as a ‘lexical category’ […] is, in contrast, inherent in nouns, due to semantic and/or lexical restriction; it conveys no information beyond that which the noun itself conveys” [Editors’ translation] (1985:31). Cf. also Unterbeck (2000). 7 Hurskainen continues, “It is important to make a distinction between gender systems and noun class systems since there are languages which apply both of these systems simultaneously” (2000:665). Cf. also Claudi (1985:14): “‘Gender’ languages are, in more or less clearly recognizable ways, associated with natural differences in sex and exhibit two or three ‘genders’ (masculine, feminine, and in some cases, neuter). In contrast, ‘class’ languages may also more or less clearly be linked to classification criteria such as ‘human,’ ‘animal,’ ‘plant,’ ‘fruit,’ etc., and accordingly possess a greater number of ‘classes,’ as a rule more than five and up to twenty” [Editors’ translation]. Others use the term gender in a broader sense meaning all kinds of noun class distinctions, cf. for example Corbett (1994:1348): “Gender systems may have sex as a component, as in languages with masculine and feminine genders, but equally sex may be irrelevant, as in Algonquian languages where distinction is between animate and inanimate.”

254


2. Natural gender marking in French Creoles 8 In all French creoles, many personal nouns are gender neutral; that is, they do not provide any formal indication of the referent’s sex. Words that may be used to refer to both female and male referents usually go back to the French masculine forms and include common terms such as Louisiana Creole (LouCr): amèriken “an American (FEM or MASC),” endyen/zendyen “an Amerindian (FEM or MASC),” blon “a white (FEM or MASC),” koupèr dekon “cane cutter (FEM or MASC).” 9 Whereas in French it is possible to express gender by means of the determiner, the creole determiner -la is gender neutral. “In Creole [=Guadeloupe Creole (GuaCr)] there is no masculine or feminine gender; ‘le’ (masc.) et ‘la’ (fem.) are always expressed by ‘la’ (det.) after the noun: machann-la ‘la marchande = the (female) merchant’ et van-la ‘le vent = the wind’” (Poullet & Telchid 1990:1) [Editors’ translation]. In creoles, the lack of overt morphological gender marking on the noun itself as well as on other word classes leads to ambiguities which can only be resolved from the context. However, in all French creoles, natural gender can be indicated either by certain lexicalized items with inherent gender or by the semantically transparent compounding of a gender-neutral noun with a gender-marked lexical item. Exceptionally and only in some creoles, gender can be expressed by means of certain nominal and adjectival suffixes. Note, however, that the range of natural gender expression by inflectional or lexical means varies from creole to creole. Thus, overt gender marking and agreement phenomena are much more frequent in LouCr and Reunion Creole (ReuCr) which, due to their specific history and sociolinguistic situation, are closer to French than the so-called ‘deep’ creoles like Haitian Creole (HaiCr). One can furthermore assume that for some creoles, the introduction of a written form will have consequences for gender marking, a problem that cannot be dealt with in this article. 2.1 Inherently gender-marked terms In all French creoles, the labelling of natural gender is confined to a couple of basic antonymic pairs. Gender distinction is obligatory when referring to 8

The data are taken from various sources: Guadeloupe Creole: Bernabé (1983, 1994), Ludwig (1996); Louisiana Creole: Neumann (1985); Haitian Creole: Valdman (1978), DeGraff (2001); Seychelles Creole: Bollée (1977), Corne (1977); Reunion Creole: Chaudenson (1974). Spelling that varies from source to source presents an important problem. I decided to keep the spelling in those cases where few diacritics are involved; in the case of complicated, phonetically oriented spellings, Louisiana Creole data are written according to Valdman, Klingler, Marshall, and Rottet (1998), Haitian Creole according to Valdman (1996). 9 Cf. Valdman (1978:151): “But the differentiation of a feminine form and a masculine form for nouns and adjectives is exceptional in Creole. In cases in which two distinct forms are available, the masculine form often serves as the undifferentiated one” [Editors’ translation]. I will not deal here with the question of whether these nouns are mostly interpreted as referring to male referents solely due to the higher prominence of men as opposed to women in social life (cf. Migge 2001).


255

members of the family, since most kinship terms or address forms are inherently specified for referential gender. The following examples are mostly taken from LouCr but they are also attested in other French creoles although there may be slight differences as to pronunciation and morphophonemic shape: nòm–fam “man–woman,” frèr–sèr “brother–sister,” garson–fiy “boy–girl,” neve–nyes “nephew–niece,” kouzen–kouzin “male cousin–female cousin,” gron-popa–gron-momon “grandfather–grandmother,” misje–madam “mister– madam,” nonk–tont “uncle–aunt,” paren–marenn “father–mother” etc. 10 Professional titles exist in every creole whose predominant association is with either female or male referents (for example HaiCr bòn “maid” and kizinyèz “cook” for female referents; jeran “custodian” and ougan “voodoo priest” for male referents; Frérère 1974:92f.). 11 Additionally, most French creoles differentiate higher animals as to natural gender: bef–vach “bull–cow,” kòk–poul “rooster–hen,” matou–chat “tomcat–female cat.” Again these examples are taken from LouCr, but they are found in other French creoles as well. 12 2.2 Expression of gender by suffixation Although inflectional or morphological marking was not transferred to pidgins and creoles, most French creoles have preserved a certain amount of productive and systematic morphological marking of natural gender, especially by means of the suffixes -èz and -ès (cf. Valdman 1978:150). 13 It is essential, in this instance, to differentiate between (a) lexicalized items and (b) genuine products of word formation. While the former have equivalents in French, having been transferred as such from the base language, the latter are genuine creole neologisms. With respect to the lexicalized items, again many of the following items from LouCr are well documented in several French creoles 14: dansèr–dansèz “dancer,” tretèr–tretèz “folk healer,” nèg–negrès “black person,” mèt–mètrès “teacher,” and terms of nationality like fronse–fronsèz “French.” Furthermore, 10

Cf. Frérère (1974:90ff.) for HaiCr, Chaudenson (1974:351-352) for ReuCr and Ludwig, Montbrand, Poullet, and Telchid (1990) for GuaCr, to name just three other sources. 11 Cf. Fattier (1998b:627-628) for the HaiCr lexem koutriyèz “tailor (FEM)” (French couturière) which is probably an archaism. 12 For Seychelles Creole Corne remarks: “While the contrast kok ‘rooster’ vs pul ‘hen’ (...) occurs in basilectal SC, such pairs as seval ‘horse’ vs. zima ‘mare’ (...) seem to be due to French influence” (1977:23). 13 As to the distribution of the different suffixes, Poullet and Telchid (1990:4) note: “In Creole, though the feminine is infrequent, it nevertheless exists in some rare cases. It is then marked by èz : branbrannèz (“who likes to wear trinkets”), dousinèz (“cuddly, tender”), èskandalèz (“scandalous”), soutirèz (“[female] pimp”) or by is: sélibatris (“single, unmarried”). There are no rules governing the use of èz or is: It is through usage that one comes to know these feminine forms” [Editors’ translation] (1990:143). 14 Note that it is not my aim here to investigate the exact derivation process of the respective pairs of words.

256


most creoles have some feminine forms in -in: koken–kokin “thief” (LouCr), kadjen–kadjin “Cajun” (LouCr), kaf–kafrin “Kaffir” (ReuCr), fouben–foubinn “care-free” (HaiCr). As for the derived forms, Haitian words like wangatè/wangatèz “man/woman who believes in and/or makes magical fetishes” (DeGraff 2001), kòmèsèz “business woman” (Fattier 1998b:628), kabalèz “chatty woman” (ibid.:135), abitèz “female occupant” (ibid.:524) or Martinique Creole (MarCr) kenbousèz “female sorcerer” (Valdman 1978:150) are neologisms and thus genuine derivations. According to Marie-Christine Hazaël-Massieux (personal communication), the suffix -èz may have a pejorative connotation with certain nouns: ou byen doktèz, Mamzèl “You act like you’re smart, young lady, but you aren’t!” 15 Two points seem to be of special interest in this context. First, in French creoles the suffix -èz can only mark biological, not grammatical, gender. 16 As in French, personal nouns ending in -èz can be derived from objects or activities: GuaCr: branbrannèz “someone who wears kitschy things” (< branbrann “knick-knacks”), kankannèz “gossipy woman” (< kankan “gossip”), boujèz “fickle woman” (< boujé “move, leave”) (Ludwig 1996:192-193). Second, according to Ludwig (1996:194), Fattier (1998b:596), and DeGraff (2001), examples such as wangatè–wangatèz and branbrannèz prove that gender marking is productive in HaiCr and GuaCr. “The last pair […] wangatè/wangatèz, is one of the Haitianisms that testify to the productivity of HC gender marking: the stem wanga is etymologically Bantu, from Kikongo [...], and the inflected suffix –è/èz is etymologically French” (DeGraff 2001:74). 17 2.3 Formation of gender-marked nouns by composition In some French Creoles, gender-neutral personal nouns may be marked for sex of the referent by compounding them with a gender-marked lexical item. 18 Thus, in Seychelles Creole (SeyCr) and the Antillean creoles, there is the possibility to mark the feminine by combining a noun [+animate, +human] with the item fam [-male] for the sake of clarity: SeyCr: ban aviater fam “women pilots” (Corne 1977:28). 19 The masculine is usually unmarked. Ac15

Hazaël-Massieux also commented that examples such as chantè/chantèz or kinbwazè/ kinbwazèz are not frequent in GuaCr. 16 Cf. Ludwig (1996:192). 17 For DeGraff, these examples also prove that creoles have morphology. In HaiCr, “the obligatory/optional marking of gender is related to both the semantic class of the predicate head and the structure of the predicate projection: Manman m se yon Ayisyèn/*Ayisyen ‘my mother is a Haitian woman,’ Manman m se Ayisyèn/Ayisyen ‘my mother is Haitian’” (DeGraff 2001:74). 18 In French creoles, the ability to form gender-marked personal nouns from non-personal ones doesn’t seem to exist. In Ndjuka, for example, words like seliman “trader” (< to sell) or kolokuman “a lucky person” (< koloku “luck”) are frequent (cf. Migge 2001). 19 The construction ban aviater zom “men pilots” (cf. Corne 1977:28) does not exist in SeyCr (Annegret Bollée, personal communication).


257

cording to Marie-Christine Hazaël-Massieux (personal communication), on profésè-fanm “a woman professor” is much more likely than on profésè-nom “a man professor.” This strategy for overt gender marking is commonly employed in contexts in which the sex of the referent is under discussion or plays an important role in the context. If the referents (and their sex) are known to the interlocutors or if the referent’s sex is irrelevant, overt gender marking is generally omitted. As for nouns referring to animals, all French creoles form gender-specified nouns by combining mal “male, man” and femel “female, woman” with a non gender-specified noun. Apparently this strategy is mainly used with (higher?) animals (cf. Chaudenson 1974:797). In such nominal compounds the gendermodifying noun precedes the main noun: HaiCr: mal bourik–femèl bourik “male donkey–female donkey” (Frérère 1974:90); mal chen–femèl chen “male dog–female dog” (Frérère 1974:91); GuaCr: mal-béf–fimèl bèf “bull–cow” (Poullet & Telchid 1990:162); ReuCr: mal kabri–femel kabri “billy goat–goat” (Chaudenson 1974:352). In order to indicate natural gender in animals, French creoles also make use of the lexemes maman “mother” (or, less frequent, mer “mother”) and papa “father.” Note, however, that femel and maman/mer are not really synonyms; while femel is the general term, maman/mer are used for animals that have already given birth (Fattier 1998b:751-752). 20 With respect to the origin of these constructions, substrate influence cannot be completely excluded, since this strategy to indicate the sex of the animate noun is frequent in West African languages (cf. Boretzky 1983:85-86; Holm 2000:119-120). However, in French the juxtaposition of a gender-marked noun and a gender-unmarked noun also exists; thus, convergence seems to be a plausible explanation (cf. Chaudenson 1974:797). 21 2.4 Gender marking on adjectives Since the majority of adjectives are morphologically invariant, there is no gender agreement with adjectives (Caid-Capron 1996:168; Holm 2000:216). As with nouns, however, gender distinctions on adjectives have been preserved for most French creoles in a couple of frequently used adjectives referring to nouns with the marker [+animate]. Thus, Bernabé (1994:38) cites fou–fol “crazy” and soutirè–soutirèz “extortioner” for GuaCr; for ReuCr, Chaudenson (1974:366) discusses gra–grad “big,” lur/lurd “heavy,” and sèk–sès “dry,” which are exclusively used in fixed expressions ([ló lu:rd] “muddy water” or

20

Moreover, maman/papa are used as augmentatives, thus maman-bèf means “a big cow.” Migge (2001) attests to the same structure for Ndjuka. 21 A comparable differentiation exists in English, where pronominal elements can occur with [+animate] referents: he-bear, she-bear, boy-friend, girl-friend. This is a strategy that is also frequent in various African languages (cf. Vogel 1996:147).


258

the toponym [la ravin sès] “dry ravine”). For SeyCr, Corne (1977:23) and Bollée (1977:33) quote fu–fol “crazy,” ere–erez “happy,” and vje–vjey “old.” 22 In those creoles which have always been in close contact with their base language such as LouCr, the frequency of gender agreement with adjectives is much higher. Note that agreement is also made with animate as well as with inanimate referents: enn bonn metres “a-FEM good-FEM school teacher-FEM” (Neumann 1985:108), la fam-laba-la li vyey [lit. the woman there she old-FEM] “that woman is old” (Neumann 1985:143), enn gros vyey rob nwar [lit. a-FEM big-FEM old-FEM dress-FEM black] “a big, old, black dress” (Neumann 1985:144), enn tit mezon “a-FEM small-FEM house-FEM” (Neumann 1985:143). Moreover, in Louisiana the frequency with which gender agreement is made varies from speaker to speaker and from region to region: “Basilectal varieties of LC [LouCr], for example, that are spoken by many blacks in Pointe Coupée do not show any gender distinction. All specifiers occur in a single grammatically undifferentiated form” (Valdman & Klingler 1997:116). 2.5 Gender marking on pronouns In basilectal French creoles, gender is not indicated on personal pronouns: li/i “he, she,” ye or zot “they.” In the Atlantic creoles, however, natural gender can exceptionally be expressed on the pronominal level if the reference is not unambiguous. Thus, in GuaCr and MarCr, the nouns manzel/madam–misyé “lady–man/gentleman” can—although only sporadically—replace i “he, she” in order to emphasize natural gender if the referent is [+animate], without conveying a pejorative connotation to the sentence (Bernabé 1994:36). An example is provided in (1). In accordance with the animacy hierarchy, these items mostly refer to human beings, although (higher?) animals can be referred to. As for the plural, the forms sé mésyé “those men (or gentlemen)” and sé manzèl “those young ladies” can replace the plural pronoun yo “they” in GuaCr (Bernabé 1983:909), as in the example (2). (1)

Pyè vini, misyé di mwen bonjou. “Pierre has come, he said good day to me.” (Bernabé 1994:36)

(2)

Sé mésyé ka matébis. “They are playing hookey [from school].” (Bernabé 1983:908)

The use of madam/misyé “lady/man” as (quasi-) pronouns is also observed in HaiCr. On this matter, Fattier (1996:229) notes: Pronouns do not vary in gender. This is not a problem for the 1st and 2nd persons, whose sexual identities are visible. For the 3rd person, on the other hand, it is more 22

According to Bollée (1977:33), these forms are a marker of the “refined creole.”


259

problematic: Creole uses special lexical morphemes (msyé/mésyé pl. and madam < monsieur, madame) that replace or regularly alternate with 3rd person or other noun phrases (with [+human or humanized] referents). These markers make it possible to specify the sex when the situational or textual referent is unrecoverable or unstated, or to personalize the dialogue when the referent is included in the speech situation, probably by deference.” 23 [Editors’ translation]

Interestingly, Bernabé (1994:39-40) also lists two examples, shown in (3) and (4), in which manzel and misyé refer to inanimate objects in GuaCr, which, however, should not be interpreted as a sign of the introduction of grammatical gender into creole, but rather as cases of anthropomorphism (1994:39). (3) Lapòt-la tonbé anlè pyé mwen, manzel fè mwen mal toubannman. “The door fell on my foot. It really hurt me badly.” (4) Mato-a konyen lanmen mwen, man té anvi fouté misyé an razyé. “I hit my hand with the hammer, I feel like flinging it away.” Bernabé (1994:40-41) adds: In fact, the features [+male] and [+female] involved here are nothing more than an expression on the semantic level of grammatical (syntactic) features that come from the respective traits of the French words marteau ‘hammer’ and porte ‘door’, that is [+masculine] and [+feminine]. Here we have a completely atypical example of feature transfer (from French to Creole, from syntactic category to semantic category). This phenomenon was made possible only through the contact of the two languages within the same linguistic ecosystem…. [Editors’ translations]

According to Marie-Christine Hazaël-Massieux (personal communication) and Michel DeGraff (personal communication), this use of misyé and manzel/ madam to refer to inanimates is extremely rare in everyday speech in Guadeloupe and Martinique, as well as in Haiti, and is usually restricted to humoristic or fairy-tale contexts. Although the above-mentioned constructions with misyé/madam/manzel (especially those that refer to [+animate] nouns) are interesting from a typological point of view, their marginality does not permit the assumption that they could eventually become some sort of dynamic starting point for a gender system as it is found in English, where grammatical gender does not exist but where natural gender is signaled by pronouns. 24 In mesolectal LouCr, speakers differenciate between li “he” and el “she” in the singular if the referent is [+animate] to make the sex difference quite clear 23

According to Michel DeGraff (personal communication), madam/misyé can be used for all sorts of proper names with animate referents in HaiCr, including names with non-French origins. 24 A possible grammaticalization of the items mentioned would of course imply specific phonetic and semantic processes. Note in this context that, according to Greenberg (1963), gender systems could very well have their origins in the pronominal system.

260


(Neumann 1985:170). Given the fact that these forms did not exist in the historical texts, one can certainly attribute these examples to decreolization. 2.6 Gender agreement on determiners The French definite article has not survived in French creoles except as an agglutinated article, the definite determiner in most French creoles being postposed -la. Only ReuCr and mesolectal LouCr have prenominal determiners that can be compared to the French system. In LouCr, the use of the preposed articles l, le [lə]) “the-MASC” and la “the-FEM” as well as the distinction between masculine and feminine forms with respect to the indefinite and possessive determiner (en “a-MASC,” mon “my-MASC,” etc.; enn “a-FEM,” ma “my25 FEM,” etc. ) is definitely due to decreolization (cf. Neumann 1985:107-112): la fiy “the-FEM daughter,” l kouto “the-MASC knife,” la bonn vyòn de kochon [lit. the good-FEM meat-FEM of pig] “good pork” (Neumann 1985:144); enn gronn fom “a-FEM great-FEM hunger,” enn gros char “a-FEM big-FEM car-FEM” (Neumann 1985:138), en gro bal “a-MASC big-MASC ball,” ma mezon “my-FEM house” (Neumann 1985:128). Note that the use of gender-marked forms is unsystematic and that basilectal gender-unmarked forms coexist with the gender-marked equivalents. To a large degree, ReuCr has preserved the French determiners le [lə] and la: le bug “the guy, the man,” la lin “the moon” (Chaudenson 1974:355ff.). With respect to the indefinite article, the use of the forms [e] and [èn] does not seem to follow strict rules. 26 A brief look at maps 57 (Une petite pluie fine [lit. “a small rain fine”), 121 (Une bosse “a hump, a bump”), and 122 (Une ampoule “a blister”) of the linguistic atlas of Reunion (Carayol, Chaudenson, & Barat 1984) testifies to the high amount of linguistic variation: We find the same speaker using the masculine and the feminine determiner randomly with the same noun, or different speakers using different articles with the same lexical item. 27 3. Gender—a loser category This overview of gender-marking in the French creoles has revealed that all these languages have only elementary gender distinctions based on animacy 25

In LouCr èn [εn] is rare among basilectal speakers, enn [εn] being the unmarked form for both genders. 26 According to Chaudenson (1974:357), the distribution of these forms “does not correspond to that of [le] and [la], which is generally determined by those of the former definite articles of French” [Editors’ translation]. According to the same author the possessive is not gendermarked in ReuCr (1974::360): The form ma “my-FEM” only appears in fixed expressions as in, for example, the exclamation ma fiy “my daughter” (cf. also Bollée 1977:42). 27 Pierozak (in press) emphasizes the sociolinguistically determined variation with respect to the use of gender-marked forms in ReuCr. A detailed analysis of the linguistic atlas of Reunion will certainly help to better understand the functioning of gender-marking in this creole.


261

and sex difference, the “minimal building blocks that gender systems are made of” (Dahl 2000b:577). Although the basic and historically oldest criterion for noun classification is animacy (Dahl 2000b:590), the major criterion, however, for attributing animate nouns to different genders in creoles as well as in other languages is sex. 28 Trudgill (1999:138) notes that “[i]t is much less surprising that human languages have gender distinctions for human beings than that they have grammatical gender, since the distinction between male and female is the most fundamental one there is between human beings.” According to Dahl (2000a:101), “in any gender system, there is a general semantically-based principle for assigning gender to animate nouns and noun phrases”, but the domain of this principle “may be cut off at different points of the animacy hierarchy: between humans and animals, between higher and lower animals, or between animals and inanimates.” In creoles, the cut-off point seems to be between humans and everything else. If gender is morphologically assigned at all, it will be assigned to humans and to higher animals by means of the lexical and affixal strategies mentioned above. Without any doubt, grammatical gender was the big loser in the complex process of restructuring nominal morphology that took place during creolization (cf. Trudgill 1999). As a matter of fact, this category has never existed in creole languages and the very few signs of residues of grammatical gender in the early texts do not really permit one to suppose a gradual erosion of this category. 29 Why has grammatical gender been given up so radically in pidgins and creoles? And why has gender not been reintroduced in those creoles that are in close contact with their base language (e.g., ReuCr or GuaCr/MarCr, which only show marginal decreolization phenomena in this respect)? The answers to these questions are complex and cannot be dealt with in detail here. However, we will address some of the relevant factors, including certain de28

Cf. Dahl (2000a:102): “The pervasiveness of sex as gender criterion is striking. There are many possible ways of classifying animates, in particular human beings, that might be used as a basis for gender, such as social status, ethnic origin, profession, age, hair color etc., but none of them except perhaps age seems to play any important role in gender assignment. The frequency of sex-based gender distinctions to some extent hides the importance of animacy.” 29 As for the Passion selon St-Jean “Passion according to Saint John” (probably written before 1760), Hazaël-Massieux (1996:195-196) notes an “éventuel maintien d’une opposition de genre: sur vingt-cinq [sic] attestations de déterminant indéfini, vingt-six sont la forme ‘ïon’, deux la forme ‘ïone’ et sept ‘ïoune’” (“a possible maintenance of a gender opposition: out of 35 attestations of the indefinite determiner, 26 have the form ‘ïon,’ 2 ‘ïone,’ and 7 ‘ïoune’” [Editors’ translation]) whereby ïon always precedes French masculine nouns. As for the definite article, he observes for several early texts “the disappearance of a significant number of verbal markings of definite or indefinite determination” [Editors’ translation] (189), the consequence being the destruction of the system of determination marking of French” [Editors’ translation] (191). Cf. also Fattier (1998a:263) who confirms that there was no gender differentiation on the 3sg pronoun in the text of Ducœurjoly from 1802: li “he/she,” yo “they,” and no gender marking in nouns (Fattier 1994:59).


262

velopments in the base languages themselves, peculiarities of the language acquisition process in plantation societies, and, above all, functional considerations. In his discussion of gender in ReuCr, Chaudenson (1974:350) underlines that spoken French exhibits several signs of gender weakening and he justifiably asks, “If we take spoken French as our reference point, is the distance between French and Creole so large?” (Editors’ translation.) (cf. also Bollée 1977:35). It is a well known fact that, due to a series of sound changes, the gender system is less transparent in spoken than in written French. 30 As for popular French, although the gender opposition is still rather strong, there are signs of neutralization. Feminine plural nouns can be replaced by the masculine plural pronoun ils, as in Ils sont où les fleurs? [lit. they-MASC are where the flowers-FEM] “where are the flowers?”; in the plural ils “they-MASC” and elles “they-FEM” can be replaced by gender neutral ça “that” or eux-autres “those”; the 3sg pronouns il “he” and elle “she” are often pronounced [l] before a vowel; and predicative adjectives tend to be invariable (Gadet 1992:5859;63-64). In the so-called marginal varieties of French (français marginaux) like Cajun or Acadian French (Chaudenson, Mougeon, & Beniak 1993), this evolution is even further advanced, although grammatical gender is still strong. If one accepts the hypothesis that the français marginaux allow one to draw certain conclusions about the language of the settlers,31 one can possibly assume a reasonable amount of linguistic uncertainty as to gender assignment as well as first signs of category reduction. In Cajun, for example, agreement of adjectives is frequently lacking, as shown in the examples (5)–(7).

30

(5)

sa mère était mort 3sg. POSS-FEM mother was dead-MASC “His/her mother was dead.” (Stäbler 1995:42)

(6)

elle est froid she is cold-MASC “She’s cold.” (Stäbler 1995:153)

“French orthography preserves clues about gender which are lost in the spoken language” (Corbett 1991:58). Cf. also Tucker, Lambert, and Rigault (1977), Härmä (2000), and Dubois (1965:65). 31 Cf. Remarks in Chaudenson et al. (1993:44) on the French of the overseas diasporas (D1: Acadia, Quebec, Lesser Antilles, Bourbon etc.; D2: Missouri, New England, Ontario, The West of Canada, Louisiana): “These D1 varieties, cut off from the French of France and subject to little ‘normative pressure,’ thus provide evidence of features that have survived from earlier stages of the language. The D2 varieties, insofar as they are linked to migratory movements having taken place during the initial stages of French colonization, and insofar as they established themselves in contexts where normative pressure was virtually absent […], are doubtless even more ‘archaic’ than the D1 varieties and make it possible to add even greater time depth to research on French spoken during the colonial period” [Editors’ translation].


(7)

le gros maison à the-MASC big-MASC house-FEM POSS “Cecil’s big house” (Stäbler 1995:181) 32

263

Cecil Cecil

In Cajun as well as in Acadian French, the indefinite article is frequently not gender-marked. This is particularly true before nouns beginning with a vowel. 33 Additionally, a certain uncertainty as to the choice of the correct article in seen in these two varieties: le chose-là “the-MASC thing-FEM there” (Stäbler 1995:41), la office “the-FEM office-MASC” (Stäbler 1995:180); le machine “the-MASC machine-FEM” (Wiesmath 2000:163), le poêle “the-MASC frying pan- FEM” (Wiesmath 2000:65); moreover, gender confusion is seen in un/une arbre “the tree” (Wiesmath 2000:13) and une/un serpent “a snake” (Wiesmath 2000:34, 35). As far as the personal pronouns are concerned, gender distinction in Cajun is highly endangered. In the singular, il “he” and al “she” are often replaced by indefinite ça “that”; in the plural, the gender distinction has already been lost. Specifically, the equivalents of French ils/elles “they-MASC/they-FEM” are replaced ils “they,” eux-autres “those,” eusse “those,” or ça “that,” with the latter two becoming more and more frequent. Some examples are given in (8)(10).

32

(8)

C’est sa vieille soeur qui cuit. Ça, ça parle bien français it’s her older sister that cooks. that, that speaks well French comme nous-aut’. like us “It’s her older sister who cooks. She, she speaks French well, like us.” (Smith 1994:201)

(9)

Mon neveu ça veut s'acheter un tit truck, un Toyota my nephew that wants to buy himself a little truck, a Toyota “My nephew, he wants to buy himself a little truck, a Toyota.” (Smith 1994:49).

Cf. Bollée and Neumann-Holzschuh (1998:183). For St. Barths French/St. Thomas French cf. Corne (1999:138). 33 Cf. for Acadian French Péronnet (1989:40) and for Cajun French Conwell and Juilland (1963:130).


264

(10) Quand eux-autres sont comme dans le village de Houma, when they are like in the town of Houma … eux-autres assaye de parler français et si la personne they attempt to speak French and if the person qu’ eusse après parler [à] les comprends pas, là ça who they talk to them understand NEG, there that va parler en anglais is going to speak in English “When they are like in the town of Houma, they attempt to speak French, and if the person who they talk to doesn’t understand them, then they start to speak in English.” (Rottet 1995:237) Of course, these examples should not be overinterpreted. That slaves obviously came into contact with grammatical gender in the acquisition process can be shown by the numerous nouns having a fossilized article as part of the word (e.g., lo “water” < Fr. l’eau “the water”). Apparently, slaves were unable to understand the function of the preposed articles because they didn't have anything comparable in their native languages, which have a complex noun class system instead of a gender system (see note 8 and Hurskainen 2000). French articles were thus reinterpreted as part of the nouns (cf. Detges 2000). 34 More important, however, seems to be the fact that grammatical gendermarking is largely unnecessary for human communication. Grammatical gender affixes do not mark any real-world entity, they are devoid of semantic substance, and do not serve any communicative need. 35 Compared to other categories such as case, number, tense, aspect, or person, grammatical gender is a relatively marginal category, whose function is “to a considerable extent obscure” (Trudgill 1999:138). Since languages can very well do without special syntactic means that help with reference tracking, 36 Trudgill (1999) and McWhorter (2001) classify grammatical gender marking as one of those linguistic phenomena that have an explanation but no function and that form part 34 It is questionable whether noun classes can be established phonologically, basing them on the reanalyzed forms containing articles. As in many other French creoles, different groups of nouns can be distinguished in LouCr, depending on whether one of the articles has been reanalyzed as part of the word and, if so which one: di- in disik “sugar” (< French du sucre “some sugar”), l- in lide “idea” (< French l’idée “the idea”), n- in nom “man” (< French un homme “the man”), z in zozo “bird” (< French les oiseaux “the bird”) (cf. Neumann 1985:150ff.). Apart from MauCr and SeyCr, the number of nouns containing a reanalyzed article is rather small in most French creoles. 35 Frei (1929/1971:73) already considered gender to be “a largely useless category from a semantic perspective” [Editors’ translation]. 36 Gender is by no means a universal category, cf. Corbett (1991) and McWhorter (2002:4): “This kind of marking is obviously not specified by Universal Grammar, something made clear by the fact that so many of the world’s languages lack it: it is merely the fossilized remainder of diachronic drift. As such, it is ornamental to a natural language.”


265

of the unnecessary historical baggage most ‘older’ languages still have. 37 From a different perspective, the superfluousness of gender for communication is also stated by Leiss (2000:242). This author even doubts that gender can be characterized as a grammatical category at all, because contrary to number or case, it does not permit selection of a paradigm: “In German and other IndoEuropean languages today, gender is an opaque leftover category which may well indeed exhibit only a leftover function, namely, that of producing grammatical agreement.” Given the low functional load of grammatical gender it is not surprising that in language-learning processes there is a clear acquisition order: number > case > gender. Wegener shows for German that number markers are acquired first, then the markers for case, and finally gender markers: “The difficult and slow acquisition of gender thus proves ex negativo that children search for functions for the forms and, failing to discover them, they have difficulty in learning meaningless elements” (Wegener 2000:538). Similar observations have been made as to the acquisition of French as a first and second language (Clark 1985:705-707; Véronique 1994). It is also not surprising that in the reverse process—language death—gender is among the first to be given up. Since dying languages undergo a certain number of evolutionary processes which are also characteristic of pidginization and creolization, they tend to simplify complex morphophonemic paradigms and aim at a maximum amount of semantic transparency (cf. Dorian 1989; Neumann-Holzschuh 2000a; Sasse 1992). Against this background, gender loss in creoles can be convincingly explained as a phenomenon caused by the interplay of cognitive and functional as well as acquisitional factors. In comparison with ‘older’ languages that continue to drag along with them a certain amount of historical baggage, creoles indeed display less complexity having given up a lot of their morphological and syntactical encumbrances (cf. McWhorter 2001). 38 Since grammatical gender is one of the least important categories for coding and communicating knowledge, its abandonment was complete. This is in opposition to natural 37

Borrowing a term coined by Lass (1997:13), they consider grammatical gender-marking (analytical or affixal) as an example of ‘linguistic male nipples’. “They are phenomena which, as biologists would say, have an explanation but no function. Whether or not it is clear why such grammaticalization processes take place, it is clear that their motivation is not originally to divide nouns into agreement classes, or to aid with reference-tracking or disambiguation” (Trudgill 1999:149). 38 I do not believe, however, that (at least French) creoles arose from ground zero. McWhorter (2001:155) states “Creole languages are unique in having emerged under conditions which occasioned the especial circumstance of stripping away virtually all of a language’s complexity [...], such that the complexity emerging in a creole is arising essentially from ground zero, rather than alongside the results of tens of thousands of years of other accretions. As such, creoles tend strongly to encompass a lesser degree of complexity than any older grammar.” This hypothesis is also strongly refuted by DeGraff (2001).

266


gender, whose preservation certainly has to do with semantic transparency and the anthropologically determined wish to distinguish between male and female. 39 4. Gender and number in French Creoles Finally, I would like to refer briefly to an aspect of the recent gender debate which, as regards creole languages, is admittedly rather speculative. Recent diachronic research on gender emphasizes the relationship between the categories of number and gender in its hypothesis that the original function of gender was the perspectivation of plurality. 40 Whereas number has always marked the distinction between singular and plural, feminine gender was frequently used to express the collective plural and thus supplied a different perspective on a multitude of entities (Unterbeck 2000:xxxi). 41 Be that as it may, it is taken for granted that number is the primary, underlying category and gender is the category built ‘on the shoulders’ of number (Corbett 2001; Leiss 1994, 2000; Unterbeck 2000: xxxiv). As Leiss (1994:288) notes: If a language has the category of gender, it always has the category of number (Universal No. 36 in Greenberg 1963:95). But the reverse does not necessarily hold. The rule therefore says that the existence of the category of number implies the category of gender. Thus gender presupposes the category of number. Gender needs the category of number, so to speak, in order to develop. This indeed hints at an affinity in contents between the grammatical meanings of the categories of gender and number. But it is still unclear what these grammatical meanings are (translation by Unterbeck 2000:xxxi).

Where, however, is the link to creoles? First of all, the creole data seem to confirm this hierarchy: All creole languages have number but not all of them have gender. Although plural marking, which is marked by free morphemes in

39

Within the renewed discussion on whether creolization necessarily presupposes pidginization and thus a fundamental process of reduction of the overt grammatical apparatus (cf. McWhorter 2001, 2002), gender might illustrate that the ‘syntax-internal conception’ which stresses the gradual evolution of creole grammar complements the pidginization-based framework. Gender was indeed one of the most prominent victims of the restructuring process which led to the creation of creole languages, although it was certainly already a weakened category in the spoken French of the settlers. Moreover, what is important in this connection is that the restructuring process in the individual creole languages took place at completely different speeds and even affected the individual grammatical categories with different degrees of thoroughness (cf. Neumann-Holzschuh 2000b). 40 Cf. Leiss (1994:293): “The function of the gender category consisted in making available different possibilities of perspective in relation to a variety of elements” [Editors’ translation]. 41 Although “we are still some way from understanding how gender systems arise” (Corbett 1991:310), the fact that gender originally had something to do with the view on plurality could be a clue (Leiss 1994, 2000).


267

French creoles, 42 does not follow the same rules as in French, the singular/plural distinction (which is not made if plurality is clear from the context) must be made under certain conditions: “In general, the plural marker is used only when the noun of a noun phrase has a specific reference; when it was previously mentioned in speech, when it has a deictic reference, etc.” [Editors’ translation] (Valdman 1978:199). Indefinite and non-specified nouns are not generally marked for plurality in French creoles. 43 However, in some creoles, at least, there are signs that the category number is gaining ground in the sense that plural markers are also used when the referent is non-specific (Bollée 2000; Neumann 1985). The second link seems to be animacy, one of the major criteria for noun-classification (Dahl 2000a:100): In earlier stages, the pluralizer was most frequently used with animate nouns, although there have always been exceptions (Bollée 2000; Holm 2000:215). The third point is more problematic. Apparently at one stage in the evolution of creoles, number markers had something to do with the perspectivation of plurality. Whereas abstract and mass nouns express a collective plural and usually appear without a pluralizer in all creoles, count nouns (‘individuativa’) form a distributive plural with the aid of a plural marker (providing they have the marker [+specific]). 44 Apart from some fairly recent developments, 45 number in creole languages seems to function in a similar way to gender in early Indo-European. Mass nouns and count nouns are distinguished by means of the respective number markers, the same basic bipartite distinction which was decisive for gender assignment in early Indo-European languages (Weber 2000:505). To postulate that creoles will eventually establish a gender system once number has become a more solidly established grammatical category would certainly be daring and would have to be classified as linguistic speculation. As Trudgill and Corbett show, there are quite a number of languages which can do perfectly well without grammatical gender because it is more or less afunctional. What could eventually happen, however, is a gradual expansion of natural gender marking, which is widespread in French creoles and is even expressed by inflectional means. So far, however, items like misyé “sir” and manzel “madam” in Antillean creoles are far from being grammaticalized and gender agreement on adjectives and determiners are predominantly signs of decreolization. However, if sex is indeed so pervasive as a gender criterion, it 42 Creole nouns are not inflected to indicate number, although some creole words contain fozzilized remnants of plural inflections from their base languages. 43 Cf. also Holm (2000:216) and Alleyne (1996:148ff.) for a diachronically oriented analysis of the different realizations of plural marking in French creoles. 44 Sometimes the plural marker appears with mass nouns, in which cases the noun is no longer interpreted as collective or generic but as [+countable] (cf. Bollée 2000). 45 There are signs that the category number is gaining ground in the sense that the respective markers bann, -ye, or se are also used if the referent is non-specific (cf. Bollée 2000; Neumann 1985).

268


may not be completely out of the question that one day natural gender marking in French creoles will also be made on pronouns.

References Alleyne, Mervyn C. 1996. Syntaxe historique créole. Paris: Karthala. Bernabé, Jean. 1983. Fondal-natal: Grammaire basilectale approchée des créoles guadeloupéen et martiniquais (3 vol.). Paris: L'Harmattan. ----------. 1994. “Genre grammatical, genre naturel: Contrastes, oppositions et interférences en anglais, en créole et en français”. Espace Créole 8. 29-59. Bollée, Annegret. 1977. Le créole des Seychelles: Esquisse d’une grammaire, textes, vocabulaire. Tübingen: Niemeyer. ----------. 2000. “La restructuration du pluriel nominal dans les créoles de l’Océan Indien”. Études Créoles 22. 25-39. ---------- & Ingrid Neumann-Holzschuh. 1998. “Français marginaux et créoles”. Français d'Amérique: Variation, créolisation, normalisation, Brasseur (ed) 1998 (Avignon: Université de Provence). 181-203. Boretzky, Norbert. 1983. Kreolsprachen: Substrate und Sprachwandel. Wiesbaden: Harrasowitz. Caid-Capron, Leila. 1996. “La classe adjectivale en créole mauricien et réunionnais”. Matériaux pour l’étude des classes grammaticales dans les langues creoles, Véronique (ed) 1996 (Aix-en-Provence: Université de Provence). 163-192. Carayol, Michel, Robert Chaudenson, & Christian Barat. 1984. Atlas linguistique et ethnographique de la Réunion (vol. 1). Paris: CNRS. Chaudenson, Robert. 1974. Le lexique du parler créole de La Réunion (2 vol.). Paris: Honoré Champion. ----------, Raymond Mougeon, & Édouard Beniak. 1993. Vers une approche panlectale de la variation du français. Paris: Didier. Clark, Eve V. 1985. “The acquisition of Romance, with special reference to French”. The crosslinguistic study of language acquisition (vol. 1), Slobin (ed) 1985 (Mahwah, NJ: Lawrence Erlbaum). 687-782. Claudi, Ulrike. 1985. Zur Entstehung von Genussystemen: Überlegungen zu einigen theoretischen Aspekten, verbunden mit einer Fallstudie des Zande. Hamburg: Buske. Conwell, Marilyn J. & Alphonse Juilland. (1963). Louisiana French grammar, I: Phonology, morphology and syntax. The Hague: Mouton. Corbett, Greville C. 1991. Gender. Cambridge: Cambridge University Press. ----------. 1994. “Gender and gender systems”. The encyclopedia of language and linguistics (vol. 3), Asher & Simpson (eds) 1994 (New York: Pergamon Press). 1347-1353. ----------. 2001. Number. Cambridge: Cambridge University Press.


269

Corne, Chris. 1977. Seychelles Creole grammar: Elements for Indian Ocean Proto-Creole reconstruction. Tübingen: Narr. ----------. 1999. From French to Creole: The development of new vernaculars in the French Colonial World. London: Westminster University Press. Dahl, Östen. 2000a. “Animacy and the notion of semantic gender”. Gender in grammar and cognition: Part I. Approaches to gender, Unterbeck (ed) 2000 (Berlin: Mouton de Gruyter). 99-116. ----------. 2000b. “Elementary gender distinctions”. Gender in grammar and cognition: Part II. Manifestations of gender, Unterbeck, Rissanen, Nevalainen, & Saari (eds) 2000 (Berlin: Mouton de Gruyter). 577-594. DeGraff, Michel. 2001. “Morphology in creole genesis: Linguistics and ideology”. A life in language, Hale (ed) 2001 (Cambridge, MA: MIT Press). 53121. Detges, Ulrich. 2000. “Two types of restructuring in French creoles: A cognitive approach to the genesis of tense markers”. Degrees of restructuring in creole languages, Neumann-Holzschuh & Schneider (eds) 2000 (Amsterdam: John Benjamins). 135-162. Dorian, Nancy. (ed). 1989. Investigating obsolescence: Studies in language contraction and death. Cambridge: Cambridge University Press. Dubois, Jean. 1965. Grammaire structurale du français: Nom et pronom. Paris: Larousse. Fattier, Dominique. 1994. “Un fragment de créole colonial: Le Manuel des habitans de Saint-Domingue de S.J. Ducœurjoly, 1802. Réflexions sur l’apprentissage et la créolisation”. Créolisation et acquisition des langues, Véronique (ed) 1994 (Aix-en-Provence: Université de Provence). 53-73. ----------. 1996. “Regards sur les pronoms de l’haïtien (dans une perspective comparative)”. Matériaux pour l’étude des classes grammaticales dans les langues creoles, Véronique (ed) 1996 (Aix-en-Provence: Université de Provence). 213-242. ----------. 1998a. “Créolisation et changement: A propos des ‘pronoms personnels’ (de l’antillais ancien à l’haïtien)”. St. Kitts and the Atlantic Creoles, Baker & Bruyn (eds) 1998 (London: University of Westminster Press). 255-270. ----------. 1998b. Contributions à l'étude de la genèse d'un créole: L'Atlas linguistique d'Haïti, cartes et commentaires. Aix-en-Provence: Publications de l'Université de Provence. Frei, H. 1971. La grammaire des fautes. Genève: Slatkine Reprints (Original work published 1929). Frérère, Gerard Alphonse. 1974. “Haitian Creole sound-system, form-classes, texts”. Dissertation Abstracts International 35 (08) [Feb 1975], 5377A (University Microfilms No. AAT 7502726).

270


Gadet, Françoise. 1992. Le français populaire. Paris: Presse Universitaire de France. Greenberg, Joseph H. 1963. Universals of language: Report of a conference held at Dobbs Ferry, NY, 13-15 April 1961. Cambridge, MA: MIT Press. Härmä, Juhani. 2000. “Gender in French: A diachronic perspective”. Gender in grammar and cognition: Part II. Manifestations of gender, Unterbeck, Rissanen, Nevalainen, & Saari (eds) 2000 (Berlin: Mouton de Gruyter). 609-619. Hazaël-Massieux, Guy. 1996. “L’expression de la détermination en créole de la Caraïbe”. Les créoles : problèmes de genèse et de description, HazaëlMassieux. (ed) 1996 (Aix-en-Provence : Publications de l’Université de Provence). 185-205. Heil, Anett. 1999. Grammatische Reduktion in Frankokreolsprachen und Plansprachen. Frankfurt am Main: Lang. Heil, Karolin. 2001. Die Kategorie Genus und ihr Status in den französischen Kreolsprachen. Unpublished master’s thesis, Universität Regensburg. Hockett, Charles F. 1958. A course in modern linguistics. New York: Macmillan. Holm, John. 2000. An introduction to pidgins and creoles. Cambridge: Cambridge University Press. Hurskainen, Arvi. 2000. “Noun classification in African languages”. Gender in grammar and cognition: Part II. Manifestations of gender, Unterbeck, Rissanen, Nevalainen, & Saari (eds) 2000 (Berlin: Mouton de Gruyter). 665-687. Lass, Roger. 1997. Historical linguistics and language change. Cambridge: Cambridge University Press. Leiss, Elisabeth. 1994. “Genus und Sexus. Kritische Anmerkungen zur Sexualisierung von Grammatik”. Linguistische Berichte 152. 281-300. ----------. 2000. “Gender in Old High German”. Gender in grammar and cognition: Part II. Manifestations of gender, Unterbeck, Rissanen, Nevalainen, & Saari (eds) 2000 (Berlin: Mouton de Gruyter). 237-258. Ludwig, Ralph. 1996. Kreolsprachen zwischen Mündlichkeit und Schriftlichkeit. Zur Syntax und Pragmatik atlantischer Kreolsprachen auf französischer Basis. Tübingen: Narr. -----------, Danielle Montbrand, Hector Poullet, & Sylviane Telchid. 1990. Dictionnaire créole français. Paris: Servedit/Editions Jasor. McWhorter, John. 2001. “The world's simplest grammars are creole grammars”. Linguistic Typology 5. 125-166. ----------. 2002. “The rest of the story: Restoring pidginization to creole genesis theory”. Journal of Pidgin and Creole Languages 17. 1-48. Migge, Bettina. 2001. “Eastern Maroon Creole: Communicating gender in the Eastern Maroon Creole”. Gender across languages: The linguistic repre-


271

sentation of women and men (3 vol.), Hellinger & Bußmann (eds) 2001 (Amsterdam: John Benjamins). 85-104. Neumann, Ingrid. 1985. Le créole de Breaux Bridge, Louisiane: Étude morphosyntaxique–textes–vocabulaire. Hamburg: Buske. Neumann-Holzschuh, Ingrid. 2000a. “Español vestigial y francés marginal en Luisiana: Erosión lingüística en isleño/bruli y en cadjin”. Boletín de Lingüística 15. 36-64. ----------. 2000b. “Restructuration dans un créole ‘conservateur’: Le Cas du créole louisianais”. Degrees of restructuring in creole languages, Neumann-Holzschuh & Schneider (eds) 2000 (Amsterdam: John Benjamins). 383-407. Péronnet, Louise. 1989. Le parler acadien du Sud-Est du Nouveau-Brunswick. Éléments grammaticaux et lexicaux. Berlin: Peter Lang. Pierozak, Isabelle. In press. “Eléments de reflexion à partir du ‘genre’ en créole réunionnais: Grammaticalisation, variation et écriture. Etude de pratique discursives dans quelques écrits contemporains”. Grammaticalisation et réanalyse: Approches de la variation créole et française, S. Kriegel (ed) (Aix-en-Provence: Université de Provence). Poullet, Hector & Sylviane Telchid. 1990. Le créole sans peine (guadeloupéen). Chennevières-sur-Marne: Assimil. Rottet, Kevin J. 1995. Language shift and language death in the Cajun French-speaking communities of Terrebonne and Lafourche Parishes, Louisiana. Unpublished doctoral dissertation. Bloomington: Indiana University. Sasse, Hans-Jürgen. 1992. “Theory of language death”. Language death: Factual and theoretical explorations with special reference to East Africa, Brenzinger (ed) 1992 (Berlin: Mouton de Gruyter). 7-30. Smith, Jane S. 1994. “A morphosyntactic analysis of the verb group in Cajun French”. Dissertation Abstracts International, 55 (08) [Feb 1995], 2370A (University Microfilms No. AAT 9434353). Stäbler, Cynthia. 1995. La vie dans le temps et asteur: Ein Korpus von Gesprächen mit Cadiens in Louisiana. Tübingen: Narr. Stein, Peter. 1986. Kreolisch und Französisch. Tübingen: Niemeyer. Trudgill, Peter. 1999. “Language contact and the function of linguistic gender”. Poznań Studies in Contemporary Linguistics 35. 133-152. Tucker G. R., W.E. Lambert, & A. A. Rigault. 1977. The French speaker’s skill with grammatical gender: An example of rule-governed behaviour. The Hague: Mouton. Unterbeck, Barbara. 2000. “Gender: new light on an old category: An introduction”. Gender in grammar and cognition: Part I. Approaches to gender, Unterbeck (ed) 2000 (Berlin: Mouton de Gruyter). xv-xlvi.

272


Unterbeck, Barbara, Matti Rissanen, Terttu Nevalainen, & Mirja Saari. (eds) (2000). Gender in grammar and cognition: Part II. Manifestations of gender. Berlin: Mouton de Gruyter. Valdman, Albert. 1978. Le créole: Structure, statut et origine. Paris: Klincksieck. ----------. 1996. A learner's dictionary of Haitian Creole. With R. Jean-Baptiste & C. Pooser. Bloomington, IN: Indiana University Creole Institute. ---------- & Thomas A. Klingler. 1997. “The structure of Louisiana Creole”. French and Creole in Louisiana, Valdman (ed) 1997 (New York: Plenum). 109-144. ----------, Thomas A. Klingler, Margaret Marshall, & Kevin J. Rottet. 1998. Dictionary of Louisiana Creole. Bloomington: Indiana University Press. Véronique, Daniel. 1994. “Naturalistic adult acquisition of French as L2 and French-based creole genesis compared: Insights into creolization and language change?” Creolization and language change, Adone & Plag (eds) 1994 (Tübingen: Niemeyer). 117-138. Vogel, Petra M. 1996. Wortarten und Wortartenwechsel: Zu Konversion und verwandten Erscheinungen im Deutschen und anderen Sprachen. Berlin: Mouton de Gruyter. Weber, Doris. 2000. “On the function of gender”. Gender in grammar and cognition: Part I. Approaches to gender, Unterbeck (ed). 495-509. Wegener, Heide. 2000. “German gender in children's second language acquisition”. Gender in grammar and cognition: Part I. Approaches to gender, Unterbeck (ed) 2000 (Berlin: Mouton de Gruyter). 511-544. Wiesmath, Raphaële. 2000. Enchaînement des propositions dans le français acadien du Nouveau-Brunswick/Canada: Place de ce parler parmi d’autres variétés d’outre-mer. Unpublished dissertation. Freiburg i.Br: Universität Freiburg. Wurzel, Wolfgang Ulrich. 1986. “Die wiederholte Klassifikation von Substantiven: Zur Entstehung von Deklinationsklassen”. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 39. 76-96.

TENSE, MOOD, AND ASPECT AND THE DEIXIS ORDERING PRINCIPLE

Anand Syea University of Westminster

ABSTRACT: Members of the auxiliary system follow a general ordering pattern in Creoles in which tense markers linearly precede modal and aspect markers and modal markers in turn linearly precede aspect markers. It is interesting that creole languages display this pattern with very little variation and, naturally, questions arise as to why this particular pattern is found and why there is so little variation across the creole languages. This paper is about these questions and it will propose, on the basis of a detailed study of TMA markers in Mauritian and other creoles, that the ordering is independently determined by principles of syntax. It will further seek to explain why T is the closest to C. The paper will show how current syntactic ideas about clausal structure and functional projections can provide a plausible alternative to semantic and historical explanations of TMA ordering. 1. Introduction This paper is concerned with the ordering of the tense, mood, and aspect (TMA) markers. The data come mainly from Mauritian Creole (MC). As was first noted by Thompson (1961) and Taylor (1971), when more than one of these markers occurs in a creole language, they end up fixed in their position in relation to one another. Thus, T precedes M and A, and M in turn precedes A. Muysken (1981) offers an explanation of this restriction by appealing to Woisetschlaeger’s (1977) TMA Ordering Principle and proposing a principle of Universal Grammar to the effect that aspect is interpreted before mood and mood before tense. Thus, the fixed linear ordering can be seen to reflect the sequence of interpretation. In this paper, I propose an alternative in terms of selection and a general organizational principle based on the intuitive idea that where a deictic and a non-deictic element occur, the deictic element takes precedence and occupies a c-commanding position. This principle has application beyond TMA, for instance, in ordering functional categories inside Determiner Phrases (DP). It derives, I will suggest, from a locality constraint on binding of deictic elements. It will thus be seen that the constraint on TMA ordering can be accounted for by making use of principles that are needed elsewhere in the grammar, which makes it therefore unnecessary to invoke

274


Woisetschlaeger’s TMA ordering principle and Muysken’s principle of interpretation. The discussion is for the most part couched within the Principles and Parameters framework (Chomsky 1989). The paper is structured as follows. Section 1 presents the list of TMA markers in MC and comments on their historical sources and ordering. Section 2 discusses properties of TMA markers while section 3 examines their status. It is proposed on the basis of the properties they display that they should be treated as heads. Section 4 discusses TMA ordering first in relation to the verb (or Verb Phrase [VP]), then in relation to negation, and finally in relation to one another. It is suggested that the position of TMA in relation to verbs (VPs) falls out of the Head Parameter while their position in relation to negation falls out of the assumption for clause structure in Universal Grammar. On the other hand, their position in relation to one another is shown to be a consequence of an ordering condition on the distribution of deictic elements and selection. Section 5 concludes the discussion. 2. Data 2.1 Inventory and development of TMA markers in MC Tense, mood, and aspect are expressed in MC (as in most creoles) by a small set of lexical items, which derives historically from French words. Interestingly, all these words belong to the [-N] category (i.e., verb and preposition). The TMA markers (with their historical sources) are given in Table 1. The examples that follow the table illustrate the earliest attestations of the French words as markers of tense, mood, and aspect in MC. 1 TMA markers Source ti [past] été [V] (or one of its phonological variants) (v)a [future] va(s) [V] pu [future] pour [P] (f)in [completive] fini [V] (a)pe [progressive] après [P] Note. P = Preposition; V = Verb

Table 1: MC TMA markers and their historical sources (1) a. moy n’apa été batté ça blanc là (1779) NEG T hit that whiteman I “I didn’t hit that white man” b. Toi va paye moi ça (1777) you M pay me that “You’ll pay for this” 1 Examples (1a, b, d, and e) are from Chaudenson (1981). Example (1c) is from Freycinet (1827). After each citation, the year in which it was attested is given in parentheses.

TMA MARKERS AND ORDERING

275

c. L’her sole levé, mo pour alle prend mon poste … (1827) when sun rise I M go take my position “When it’s morning I will go and stand guard …” d. Moy fini mouri (1734) I A dead “I’m about to die” e. ... li trouve sartié aprè mette souval carosse lé Roi (1850) …he find driver A put horse coach the king “He saw the coach driver getting the King’s coach ready” A brief comparison of the TMA markers with their historical roots shows that the bi-morphemic French source words have undergone phonological reduction; all the TMA markers are now mono-morphemic, which is an expected consequence of the process of grammaticalization. An interesting observation here is that the TMA markers and their historical sources can actually cooccur, as shown in (2). (2) a. kot zot ti ete? where they T be “Where were they?” b. mo fin fini manze I A finish eat “I’ve finished eating” c. zot pe apre nu they A chase us “They are chasing us” d. nu pu pu pa kont we M for NEG against “We will be for not against” Ete has been replaced by ti as a tense marker and apre is both a preposition (its original meaning) and a verb (through reanalysis), meaning “to chase.” However, it is arguable that apre was already reanalysed as a verb in the input, given its verbal use in Poitevin (a dialect of French) as in “Les enfants sont après jouer” (Rickard 1974). Like many other Creoles, MC does not have a marker of the present (habitual) tense. It is reasonable to assume in this instance a zero (phonologically null) marker in paradigmatic relation with ti, the past tense marker. The difference between (3a) and (3b) is thus simply one of tense (habitual present versus past).

276


(3) a. mo buar dilo I drink water “I drink water” b. mo ti buar dilo I T drink water “I drank water” It follows from this that verbs are in effect non-finite, tense being expressed by independent morphemes. The items listed in Table 1, together with the zero marker of the habitual present, thus realize the categories tense, mood, and aspect in MC. Interestingly, Haitian Creole (HC) also displays a more or less similar inventory of TMA markers, as shown in Table 2 (adapted from Magloire-Holly 1982). 2 The only difference between MC and HC here is the use of ap in HC as both an aspect and a mood marker. The striking similarity in the choice of lexical items to realise the grammatical categories of tense, mood, and aspect in the two languages is particularly intriguing given the geographical distance between them and given the fact that they have different substrates (East African Bantu languages for MC and Kwa for HC). What is also useful to note is the similarity in their distribution. The same ordering restrictions appear to hold on TMA in both languages. TMA Tense Mood Aspect

Markers t(e) [past]; ‘zero’ [present] (a)va, pu, ap fin [completive]; ap [progressive]

Table 2: HC inventory of TMA markers These similarities raise interesting questions for the genesis of French Creoles. Lefebvre (1996) argues that the TMA system in Haitian owes everything except its phonology to Fon, a substrate Kwa language. However, the striking similarity of the two systems and the fact that they have different substrates do raise questions and seem to suggest that something else must be implicated in the development of their TMA systems. 2.2 Order of TMA markers Given the 3 categories (T, M, and A), there are in theory 6 logical ordering possibilities, as is the case with clause constituents (Subject Verb Object [SVO]). However, unlike clause constituents, TMA ordering is rigid and fixed. For instance, of the following six possible orderings (TMA, TAM, ATM, AMT, MAT, and MTA), only the first (i.e., TMA) is generally acceptable. 2

Lefebvre (1996:239) includes apral(e), which is a prospective non-completive aspect marker.


277

(4) a. zot ti/Ø pu/(a)(va) fin/pe mãze they T M A eat b. *zot pu/(a)(va) ti fin/pe mãze they M T A eat c. *zot pu/(a)(va) fin/pe ti mãze they M A T eat d. *zot fin/pe pu/(a)(va) ti mãze they A M T eat e. *zot fin/pe ti pu/(a)(va) mãze they A T M eat f. *zot ti fin/pe pu/(a)(va) mãze they T A M eat This ordering constraint holds not only in MC but also across most creoles, as noted in Muysken (1981). The question of course is what rules out the other five possibilities. Before discussing the ordering constraint, let us briefly consider some of the properties that TMA markers have. 3. Properties of TMA markers First, TMA markers tend to cluster in a single position between the subject Noun Phrase (NP) and the VP, as shown in (5). (5) a. zot ti pe mãze they T A eat “They were eating” b. zot ti pu fin arive sa ler la they T M A arrive this time “They would have arrived by now” Where negation is present, TMA markers follow the negative element but still precede the verb, as in (6). (6) a. zot pa ti pe mãze they N T A eat “They weren’t eating” b. *zot ti pa pe mãze c. *zot ti pe pa mãze Second, TMA markers cannot be separated from one another by intervening adverbial elements, suggesting some kind of adjacency condition holds between any two of them at any one time. An adverbial, if it occurs, has to follow the TMA markers.

278


(7) a. *zot pu deza fin mãze they T already M eat “They would soon eat.” b. zot pu fin deza mãze (8) a. *nu ti ãkor pe dormi we T still A sleep “They were still sleeping” b. nu ti pe ãkor dormi (9) a. *li ti byẽ pe dormi he M well A sleep “He was sleeping soundly” b. li ti pe byẽ dormi Third, they never occur in the absence of a verb. This is clear from the fact that the VP that follows them cannot be ellipted. (10) a. mo pa ti ale, li ti *(ale) NEG T go he T go I “I didn’t go, he did.” b. li pa fin mãze, nu fin *(mãze) he NEG A eat we A eat “He hasn’t eaten, we have” c. zot pa pu vini, nu pu *(vini) they NEG M come we M come “They won’t come, we will” In this respect, they contrast with modals such as kapav (“can/may”) and bizẽ (“must”) which do allow VP ellipsis. (11) a. mo pa kapav fer sa, li, li kapav (fer sa) NEG M do this him, he M do this I “I can’t do this, he can.” b. toi, to pa bizẽ ale, moi, mo bizẽ (ale). you you NEG M go me I M go “You needn’t go, I must.” This contrast is further illustrated in the following exchange. (12) A: to ti fer sa? you T do this “Did you do this?”


279

B: non, mo pa ti *(fer sa) no I NEG T do this “No, I didn’t do this.” (13) A: to kapav fer sa? you M do this “Can you do this?” B: non, mo pa kapav (fer sa) no I NEG M do this “No, I can’t do this.” VP ellipsis, while possible with modal verbs, is impossible with TMA markers. A similar contrast obtains in HC (see Lefebvre 1996). Note also that unlike TMA markers, but like verbs, modal verbs allow modification by a degree adverbial in MC. (14) a. to byẽ kapav fer li to-mem (cf.*to byẽ ti kone [“know”]) you well M do it you-self “You can jolly well do it yourself” b. mo byẽ bizẽ ale I well M go “I really must go” Summarizing, TMA markers appear to have the following properties: (a) they cluster in a single position in the clause; (b) they do not allow intervening adverbial elements; (c) they do not allow VP ellipsis (i.e., they cannot occur independently of a verb); and (d) they do not allow modification by a degree adverbial. What does this cluster of properties tell us about TMA markers? I suggest below that they all derive from the fact that they are basically heads. 4. Status of TMA markers TMA markers present an interesting analytic problem. On the one hand, they are clearly free words (or free standing morphemes) given that they can be separated from a following verb by an adverbial element (as was shown in (7)(9)). On the other, however, they display characteristics that one normally associates with bound morphemes (i.e., affixes). For instance, they cannot occur independently of a verb as we saw in (10) and (12). Secondly, they are phonologically reduced (and therefore weak) items, which means they cannot bear phonological stress, another characteristic of affixes. In this respect, they are more akin to contracted auxiliaries of English (e.g., ’ll, ’ve, etc.) except that they are not affixed to a host. Note that contracted English auxiliaries also have

280


the property of not being able to license VP ellipsis, which we said characterizes TMA markers. (15) a. John won’t buy this computer, Bill will. b. *John won’t buy this computer, Bill’ll. (16) a. John isn’t an electrician, Bill is. b.*John isn’t an electrician, Bill’s. This dual nature of TMA markers (i.e., being both free and bound) seems to suggest that they should be treated as clitic-like elements, which usually combine properties associated with free and bound morphemes. There is here an interesting parallel with the definite articles ‘the’ in English or le (and its allomorphs) in French, which also display properties of clitics. These determiners share some of the properties we associate with TMA markers. For instance, they can be separated from a following noun (the element to which they assign definiteness) by an adjective; at the same time, they cannot occur independently of it, as shown in (17) and (18). Note also that they cannot normally be phonologically stressed. (17) a. the red bus b. the *(bus) (18) a. la

belle fille pretty girl b. la *(fille) DET

As we see, an adjectival element can intervene between the determiner and the noun ([17a] and [18a]) but the determiner cannot stand on its own ([17b] and [18b]). A further parallel between determiners and TMA markers is that they are both the loci for morphosyntactic features (tense, aspect, mood, definiteness, etc.), features we associate with clitics (Zwicky 1985). Given the properties of TMA markers and given that they share some of these properties with determiners, it would not be unreasonable to treat them, at least syntactically, in the same way as we treat determiners. Determiners are heads of DPs (Abney 1987). Given the similarities between determiners and TMA markers, it seems reasonable to suggest that TMA markers are also heads. Assigning head status to TMA markers has the advantage of accounting for the properties that they have. Consider first the fact that they cannot occur independently of a VP. One way of capturing this is to say that as heads, TMA markers select VP complements. However, this by itself


281

does not explain why they cannot occur in the absence of a VP. After all, as we saw in (11), modal verbs and other auxiliaries (which are also heads) can occur without a VP complement (i.e., they allow VP ellipsis). The contrast displayed in (15) and (16) is instructive here. What it tells us is that VP ellipsis is blocked in the context of contracted auxiliaries. As noted earlier, TMA markers share with contracted auxiliaries the property of being phonologically reduced. In other words, they are prosodically deficient heads (i.e., phonologically weak) and cannot therefore license a VP deleted site, anymore than determiners can an empty NP slot (see examples (17b) and (18b)). As far as the possibility of an adverbial adjunct occurring between the TMA markers and the VP complement is concerned, this does not pose any problem. In fact, it is expected on the assumption that adverbial elements are adjuncts to VPs (Pollock 1989). Adverbial adjuncts, as was noted, cannot however intervene between any two TMA markers. This too is consistent with TMA markers being heads of the clitic type. As the examples below show, adverbial adjunction, while possible with full auxiliaries, is not with contracted auxiliaries. (19) a. They sadly have misunderstood the teacher. b. *They sadly’ve misunderstood the teacher. What this suggests is that adjunction to a functional phrase with a phonologically deficient head may not be allowed. The cluster of properties associated with TMA markers can therefore be given a principled account if, as we have argued here, they are heads. Further evidence for the head status of TMA markers can also be seen in their interaction with verbs. Consider the following examples. (20) a. li travay byẽ he work well “He works well” b. *li byẽ travay c. li fin byẽ travay (21) a. zot buar byẽ they drink well “They drink well” b. *zot byẽ buar c. zot fin byẽ buar What these examples show is that an adverbial can precede the verb only when a TMA marker is present. Why should this be? The contrast observed here

282


makes sense if we assume that the verb has to raise as is the case in French where similar facts seem to obtain. (22) a. il travaille bien he work well “He works well” b. *il bien travaille c. il a bien travaillé he has well worked Here too an adverbial to the left of the verb is allowed provided an auxiliary is also present. The similarity between French and MC can be captured if we assume that verbs raise to a functional head position. The (b) sentences are then ruled out if we assume that raising is obligatory. The (c) sentences suggest however that raising is obligatory only when it can apply. That is, a verb must raise if it can. It clearly cannot in the (c) sentences because raising yields ungrammatical sentences in French (23a-b) and either sentences with doubtful grammaticality or different meanings in MC (24a-b). 3 (23) a. il a bien travaillé. b. *il a travaillé bien. (24) a. li fin byẽ travay b. *li fin travay byẽ These differences follow in a straightforward manner if it is assumed that verbs raise and that auxiliaries and TMA markers are heads. The interaction between verb and TMA markers in such cases therefore provides further evidence that TMA markers must be head elements. Verb raising in MC, of course, seems an odd suggestion if it is indeed driven by morphology (as assumed in Pollock 1989). Still, there is some indication that some sort of verb raising must be assumed to account for the interaction between TMA markers and verbs in the context of a degree (intensifying) adverbial. Interestingly, there are a few cases where verbs do occur to the left of negation just as in French. 3

With some verbs, the adverbial byẽ can only occur in pre-verbal position, as shown in the following. i. zot fin byẽ maze/buar/amize they A well eat/drink/enjoy “they have eaten/drunk/enjoyed themselves a lot” ii. * zot fin mãz/buar/amiz byẽ. With other verbs such as travay “work” the adverbial seems to take on a ‘manner’ interpretation and is therefore acceptable in post-verbal position.


283

(25) a. mo pãs pa li pu vini I think NEG he M come “I don’t think he will come” b. mo kruar pa li pu vini I believe NEG he M come “I don’t believe he will come” c. fode pa zot kriye ar li M NEG they shout at him “You mustn’t shout at him” However, this is not possible with other verbs. (26) a. *mo buar pa divẽ I drink NEG wine “I don’t drink wine” b. *zot van pa puasõ they sell NEG fish “They don’t sell fish” (27) a. *buar pa/ pa buar drink NEG “Don’t drink” b. *koz pa/ pa koze speak NEG “Don’t talk” Given that these verbs can raise over an adverbial, as we saw in (21) for instance, the impossibility of (26) and (27) can be accounted for if negation, like TMA markers, is a head element. (26) and (27) then simply violate the Head Movement Constraint (Chomsky 1989; Travis 1984). The difference between (25) on the one hand and (26) and (27) on the other suggests that pa functions as an adverbial adjunct with a few verbs (a residue of its function in French) but as a head of Negation Phrase (NegP) with others. 4 The discussion in this section leads us to conclude that TMA markers are best analysed as heads. This explains the properties they have and share with other heads (D for instance) and their interaction with verbs in the presence of certain adverbial adjuncts. The question which arises now is how treating

4

The reanalysis of pa from specifier to head of NegP might have been triggered by the loss of the French finite auxiliary a and its replacement by the existential y-en-a (subsequently to become ena “have/be”) and the loss of ne.

284


TMA markers as heads allows us to explain the constraints on their ordering not only in relation to one another but also in relation to verbs and negation. 5. TMA Ordering This section examines the ordering of TMA markers (a) in relation to V (VP), (b) in relation to negation, and (c) in relation to one another. 5.1 TMA markers and V (VP) TMA markers, as we have seen, linearly precede V (VP). Both French and English have auxiliaries in preverbal position, as expected in SVO languages (preverbal auxiliaries are a property of VO languages). Given that the pidgins/Creoles derived from them are also SVO languages, the expectation is that TMA markers too will be to the left of verbs. The following are some of the earliest examples in MC displaying TMA markers. 5 (28) a. Toi va paye moi ça (1777) you M pay me that “I will make you pay for this.” b. Qui encore va gagne? (1805) what more M get “What more will one get?” French has both the analytic form ‘va + V’ and the synthetic form ‘V + future inflection.’ The analytic form probably developed as a result of word order change (Harris 1978). (29) a. moy fini mouri (1734) me A dead “I’m dead” b. Papa, votre femme fini mort, moi tue ly (1784) Oldman your wife A dead I kill her “Old man, your wife’s dead, I killed her.” (30)

moy n’apa été batté ça blanc la (1779) me NEG T hit that white “I didn’t hit that white man.”

So the preverbal position of TMA markers is what we expect in grammars which have set their head parameter to head-initial. 5

Example (28b) is from Pitot (1805) while (28a), (29), and (30) are cited in Chaudenson (1981).


285

5.2 TMA markers and negation TMA markers follow the negative marker in MC (and other Creoles). This is at first surprising given that in the lexifier languages finite auxiliaries (and verbs in French) precede the negative marker (‘not’ in English and pas in French). The difference is illustrated in Table 3. Note that verbs in MC are only non-finite, finiteness being expressed by free standing tense markers. Note also that non-finite auxiliaries in French can occur in both pre- and postnegative position. The pattern that emerges is that MC is like French and English insofar as the distribution of non-finite verbs and auxiliaries is concerned but unlike them insofar as the distribution of finite auxiliaries is concerned (these occur after negation in MC but in pre-negation position in French and English). Looking at the distribution of finite auxiliaries, MC is more like Italian, Spanish, and a few other Romance languages. The following examples show finite verbs and auxiliaries following negation in Italian. (31) a. Maria non parla molto Maria NEG speak much “Maria doesn’t talk much” b. *Maria parla non molto (32) a. Maria non ha parlato molto Maria NEG have speak much “Maria hasn’t spoken much” b. *Maria ha non parlato molto The question here then is why TMA and auxiliaries in Romance languages (French excepted) follow negation but precede it in French and English.

V[+FIN] V[-FIN] AUX[+FIN] AUX[-FIN]

Pre-negation English French MC No Yes — No No No Yes Yes No No Yes No

Post-negation English French MC No No — Yes Yes Yes No No Yes Yes Yes Yes

Table 3: Patterns of negation in English, French, and MC Analyses of the surface position of finite verbs and auxiliaries in relation to negation (for instance, Pollock 1989) crucially assume that verbs and auxiliaries are base-generated to the right of negation but end up on its left following the application of a rule of verb movement, which, when it applies, raises the verb to the head of Tense Phrase (TP) (Pollock 1989) or Agreement Phrase (AgrP) (Belletti 1990), as illustrated in (33).

286


The verb moves to Agr through T and Neg. Given the Head Movement Constraint (Chomsky 1989; Travis 1984), it is assumed by Pollock (1989) that in French negative constructions, pas is not the head of NegP but an adverbial element in the spec of NegP. Similarly, Rizzi (1990) treats ‘not’ in English as the specifier of NegP. The difference between French and English, on the one hand, and Italian and MC, on the other, can then be attributed, it seems, to a difference in the status of the negative particle—namely that it is an adverbial adjunct in the former but the head of NegP in the latter. Verb movement past NegP is therefore possible in the former but not in the latter. Zanuttini (1996) provides arguments to the effect that the negative marker non in Italian is the head of NegP and DeGraff (1993) similarly argues that pa in HC is the head of NegP. (33)

AgrP

Spec

Agr’

Agr

Spec

NegP

Neg’ Neg Spec

TP T’

T

VP V

The occurrence of TMA, finite auxiliaries, and verbs to the right of negation (in other words, their failure to raise over negation to Agr in MC and Italiantype languages) could be construed as a straightforward consequence of the negative particle functioning as the head of NegP. It is interesting to note in this connection that both MC and Italian-type languages allow negative concord, whereas French and English have double negation. DeGraff (1993) attributes the difference between HC and French with respect to negative concord to a difference between pa in HC and pas in French. HC pa is the head of NegP, whereas French pas is the specifier of NegP. Negative quantifiers can enter a spec-head relation with pa but not pas because the specifier position is available in the former but filled in the latter. This means we end up with


287

concord/agreement in HC but double negation in French. 6 Given that similar concord facts obtain in MC, the conclusion that pa is the head of NegP is further supported. Thus, verb raising and the Head Movement Constraint together explain why we get the difference in the relative order of negation and finite auxiliaries (and verbs) in French and English on the one hand and Italian and MC on the other. The relative position of finite auxiliaries (and verbs) in relation to negation has implications for the structure of clauses. If finite auxiliaries, finite verbs, and TMA markers occur to the right of negation, as they do in Italian-type languages and MC, it suggests that TP is lower than NegP in these languages but higher than NegP in French and English, where finite auxiliaries precede negation. Zanuttini (1996) in fact derives the order of negation in relation to finite auxiliaries and verbs by proposing that TP is a complement selected by negation, the head of NegP. The presence of negation, in other words, implies the presence of TP. This proposal extends straightforwardly to the observed word order in MC and other Creoles. However, there are two objections. First, with NegP above TP, the only position available to the subject of the sentence is Spec of NegP. This creates a problem because this position is for those elements which can check the features on the head of NegP (namely negative quantifiers). If this is correct, a further functional projection needs to be posited above NegP—something like AgrP (as in Belletti 1990). But note that AgrP, while plausible for Italian and other languages which display verbal agreement properties, is not plausible for MC and other Creoles. (But see DeGraff 1993 and Deprez & Vinet 1997 where AgrP is assumed.) An alternative approach might be the following. Suppose that clauses are universally TPs (as suggested in Pollock 1989 and Chomsky 1995). Then NegP will be lower than TP but above VP. Let us assume that TMA markers (and auxiliaries in general) are in fact verbs with the feature [+Aux]. We saw earlier that several of the TMA markers in MC (and HC) have French verbs as their historical source. If TMA markers are verbs, as we are proposing, then a sentence like (34a) will have a structure like (34b). (34) a. li pa ti ale he NEG T go “He did not go” 6

The following example in MC have only one instance of negation. That is, they display negative concord and not double negation (cf. the French examples). i. mo *(pa) fin truv person (*Je n’ai pas vu personne.) I NEG A see no-one “I haven’t seen anyone” ii. person *(pa) fin truv mua (*Personne ne m’a pas vu.) no-one NEG A see me “No one saw me”

288


b.

TP

T’

Spec

NEGP

T

Spec

NEG’

VP

NEG pa

Spec

V [+AUX] ti

V’

VP

V

ale Following this approach, a question of course arises as to how the tense feature on T gets checked. Raising of ti to T is impossible given that the head of NegP is not empty. Because pa, like the TMA markers, is a clitic-like element, we could propose an analysis that draws on the analysis of Italian non (Belletti 1990) and French ne (Pollock 1989). In this way, we would argue that pa in (34b) adjoins to T (giving us pa + T); ti then raises to substitute for T (resulting in pa + ti) and has its tense feature checked. While this works and solves the ordering problem, it does not seem to be conceptually attractive, particularly because movement is sometimes adjunction and sometimes substitution. As noted in Lasnik, Depiante, and Stepanov (2000), adjunction simply creates positions to order—adjunction positions are not there in the same way that specifiers, heads, and complements are. If we rule adjunction out on conceptual grounds, then there is no way we could arrive at the surface word order on the basis of substitution alone. The alternative must be that there is no overt movement at all in (34b). All that is needed to derive (34b) is to assume that the tense features on ti and on T are checked covertly. In the case of Italian-type languages, the tense and agreement features are also checked covertly (or overtly, following Taraldsen’s 2000 suggestion of overt feature movement).


289

Placing TP above NegP in MC is at first sight at odds with the linear ordering we find in the examples given. But we can show that it is not only theoretically motivated, as shown above, but also empirically (from historical perspective) motivated. Early texts on MC display examples in which pa(s) (the negative marker) co-occurs with na (from French ne or ‘ne + avoir’). 7 (35) a. mô n’a pas zardin donc? (1805) NEG garden then I “Don’t I have a garden?” b. Mais, papa, vous n’a pas simize, donc, vous n’a kilote: But oldman you NEG shirt then you NEG shorts n’a rien? (1805) NEG nothing “But, old man, don’t you have shirts, shorts, nothing?” c. Ah! hé! n’a pas l’optal, n’a pas sourzin (1805) NEG hospital NEG sorcerer “Ah! hé! Aren’t there hospitals, aren’t there sorcerers?” It can be argued that negation in these examples have not only a negation force but a verbal force as well, lexicalized by the -a- (French avoir) auxiliary to which n(e) is affixed. This verbal negation must have been in use until the alternative y-en-a (ena in modern MC) came to replace it. Indeed there is one example where both occur side by side. 8 (36)

Quand moi zène moi té-bête, N’a pas y-en-à la raison (1831) when I young I T stupid NEG have reason “When I was young, I was stupid; I didn’t have any brain”

The n’a in (36) and the n’a in (35) must be interpreted differently. In (36), n’a no longer has a verbal force; instead, the verbal force is expressed by the phrase y-en-a (subsequently to become ena). The loss of this verbal meaning led to the fusion of n’a and pas into a single morpheme before n’a was lost through phonological reduction, a process which also affected ete (becoming te and ti), fini (becoming fin) and apre (becoming pe), as noted earlier. It does not seem unreasonable to claim that n’a was at an early stage in the development of MC a combination of the clitic ne and the auxiliary a and was located in T, with pas in Spec, NegP, along the line of the analysis proposed for French pas in Pollock (1989). Thus, taking TP to be above NegP in MC seems to have some historical justification.

7 8

Examples (35a-c) are from Pitot (1805). Example (36) is from Chrestien (1831).

290


Finally, the alternative proposal is in accord with the Tensed C-command Condition of Laka (1990). According to this condition, T must c-command all propositional operators (Neg for instance) in a clause. The surface word order of negation preceding TMA markers in MC (and other languages) is possible provided TMA markers are treated as verbs and tense features are covertly (or overtly) checked by feature movement. 5.3 Ordering TMA markers From the data discussed above, we saw that TMA markers are strictly ordered. Tense markers precede mood and aspect markers, and mood markers in turn precede aspect markers. Another way of looking at the ordering of these elements is to say that aspect markers occupy the rightmost position while tense markers occupy the leftmost position, with mood markers placed in between. The question that arises is this: Is this ordering accidental (i.e., an historical accident) or is there some underlying principle that determines which marker goes where? Given the remarkable similarity observed across the creole languages with respect to how TMA markers are ordered, it seems unlikely that the ordering would be an historical accident. Rather, and more plausibly, it seems to be determined by some independent principle(s). As noted earlier, Muysken (1981) suggests that the TMA order is determined by a principle of Universal Grammar—which stipulates that aspect is interpreted before mood and mood before tense—working alongside Woisetschlaeger’s (1977) TMA ordering principle. (37)

TMA Ordering Principle If a and b are each syntactic formatives representing a verbal category, and a is ‘closer’ to the verb stem than b, then the translation (interpretation) rule for a precedes that for b.

The principle of Universal Grammar does not on its own guarantee that aspect should be ‘closer’ to the verb than mood or tense. Woisetschlaeger’s ordering principle is needed to ensure that of the three elements, aspect is the closest to the verb. Together, they provide an explanation for the observed ordering. However, they do not appear to be independently motivated. It seems that two specific principles are needed simply to account for word order facts to do with TMA. The alternative that I put forward here attributes the surface word order to a discourse ordering principle (38). (38)

Deictic Ordering Principle (DOP) Where α (a deictic element) and β (a non-deictic element) co-occur within a functional projection, α c-commands β.

As expected, this principle applies beyond the domain of TMA marking (i.e., TP). In fact, we will see that it applies inside DPs where deictic elements also


291

occur. Importantly, (38) appears to subsume Laka’s (1990) Tensed Ccommand Condition which states that at s-structure, tense must c-command all propositional operators in a clause. Tense is a deictic element and by (38) we expect it to c-command other operators (such as negation, aspect, etc.). 5.3.1 Ordering of constituents inside TP and DP. An interesting parallel exists between the ordering of TMA markers inside TP and the ordering of elements inside DP. In particular, elements expressing deixis can be seen to c-command those that do not. Within DPs, adjectives are ‘closer’ to the noun than determiners are. Reordering these two constituents results in ill-formed phrases in the same way that reordering TMA markers results in ungrammatical sentences. The ordering constraint on determiners and adjectives in a few languages is illustrated in (39) and (40). (39) a. b. c. d. e.

the big ball le gros ballon (French) la granda palla (Italian) een grote bal (Dutch) der grosse ball (German)

(40) a. b. c. d. e.

*big the ball *gros le ballon *granda la palla *grote een bal *grosse der ball

Of the two pre-nominal elements, only determiners are deictic. They locate an entity in the spatial domain. Adjectives on the other hand are non-deictic. They simply assign properties or attributes to the entity denoted by the noun they modify. What the contrast in (39) and (40) clearly illustrates is that determiners (the deictic elements) must take precedence over adjective modifiers (the nondeictic elements). Recent discussions of the structure of DPs have led to other functional projections (such as Number Phrase) being proposed. But even here, D will c-command the other functional heads by virtue of being deictic. Turning to TMA markers, note that an interesting parallel exists between adjectives and aspect markers. Like adjectives, aspect markers too are nondeictic elements (Lyons 1977). Interestingly, like adjectives, they seem to assign (or can be seen to assign) properties, in this case to an event denoted by the verb. For instance, they can either describe the internal nature of an event (i.e., whether it is habitual, repetitive, etc.) or the external nature of an event (i.e., whether it is completed or ongoing, etc.).

292


On the other hand, tense, like determiners, is deictic except that it locates events in a temporal domain (i.e., past, present, or future; or in those languages that have a bipartite tense system: non-past and past or future and non-future). Given the parallel between determiners and tense on the one hand and adjectives and aspect on the other, it is expected that they should be ordered in the same way (i.e., deictic precedes non-deictic). This is in fact what we find. Like determiners, tense takes precedence over aspect. We attribute this ordering constraint to the DOP (38). However, there is another reason why tense and aspect are ordered in the way they are: Aspect, a head, selects a verb phrase as its complement (Zagona 1988). In this case, aspect has to be adjacent to its complement to satisfy the structural adjacency condition on heads and complements (i.e., complements are sisters to heads). As far as tense is concerned, it has a link with Complementizer (C) (Stowell 1981). So, for instance, ‘that’ selects a finite verb while ‘for’ selects a non-finite verb. (41) a. John said that/ *for Mary has resigned b. John prefers *that/ for Mary to resign Such a link can be formalized by having TP selected by C. That is, C takes TP as its complement. Given the structural constraint on heads and complements, it follows that TP (hence T) will be ‘closer’ than aspect to C. The ordering of tense and aspect markers can therefore be explained by invoking selection and an adjacency requirement on heads and complements. One problem here however is that while the c-commanding position of T in relation to A is guaranteed in embedded clauses by the embedded C, it is not clear whether this is also true of matrix T. Matrix T is not in an obvious sense selected by the matrix C. This problem also arises with D in DP. Unlike T (embedded), D is not selected regardless of its distribution. So selection cannot guarantee that D will be leftmost, although it might be argued that this falls out of the head parameter. Still, these problems do not seem to arise with an approach that includes the DOP (38). The DOP captures the ordering restriction in TP and DP in a straightforward manner. However, as it is, it amounts to no more than a descriptive statement. The question that obviously arises is what lies behind this principle. I propose that (38) derives from a constraint on the binding of deictic elements. Deictic elements are not inherently referential; they take their reference from elsewhere. This means they have to be discourse bound or ‘anchored’ (as suggested for tense by Enç 1987). They have to be bound at Logical Form by discourse operators (a temporal index on C in the case of tense, as suggested by Enç, and spatial index on Topic [Top] in the case of determiner). I assume further, following Enç (1987), that ‘anchoring’ is constrained by locality (i.e.,


293

local head or feature movement). This means that T and C on the one hand and D and Top on the other must be ‘close’ (‘local’) to each other and this entails that T and D must be leftmost in their respective functional domains, as shown in (42). The DOP thus provides a fairly simple account of the ordering constraints not only inside TP but also inside DP. Note incidentally that this principle applies not only at the level of syntax but also at the level of morphology. According to Bybee (1985), verbal inflections across languages tend to be ordered as shown in (43) with voice closest to the verb stem and agreement features furthest from it. It is clear from (43) that deictic elements (agreement and tense features) are ordered to follow aspect. Given Baker’s (1985) Mirror Principle—in a given word, the respective order of affixes reflects the syntactic derivation of the word—it follows that Agr (or T more appropriately) will be leftmost and therefore the highest c-commanding functional head in the syntactic tree. So its position in (43) is predicted by the DOP. Seen in this way, the DOP provides an alternative explanation for the ordering in (43), which Bybee explains in terms of conceptual distance (agreement features have the least to do with verbs and are therefore the least relevant of the categories associated with verbs). (42) a.

b.

TopP

CP

AspP

Ti

Asp (43)

DP

Topi

TP

Ci

Di

VP

NumP

Num

N

voice < aspect < tense < mood < person/number

5.4 The distribution of mood markers Mood in MC (and other Creoles) is located between T and A. If aspect and verb, on the one hand, and tense and complementizer, on the other, have to be contiguous for independent reasons, then the only position available to mood is the position between T and A. If mood and aspect or mood and tense were to be reordered, then locality imposed on discourse binding (as in the case of T and C) or on head-complement relation (i.e., sisterhood between aspect and verb) would be violated.

294


Tense and mood (modality) semantically have to do with propositions (sentences). Of the two, tense is clearly deictic. The DOP predicts that where tense and mood (modality) co-occur, tense will c-command mood. This is certainly true in the TMA systems of the Creole languages. However, (43) stands as a potential counterexample to the DOP. This may be because the line between tense and mood (modality) is far from clear. Thus, future tense is used not only for making statements or predictions but is also used in utterances involving supposition, wish, inference, intention, and desire (Lyons 1977). Similarly, exponents of modality are very often at the same time exponents of Tense. Thus, ‘will’ in English indicates not only future tense but also modality (e.g., certainty). Similarly, va in MC was initially an exponent of futurity but subsequently developed the meaning of ‘indefinite future’ (possibly as a result of the emergence of pu (the definite future)). Future tense is clearly modal (nondeictic) and must therefore follow the past tense marker (a deictic element), as predicted by the DOP. The ordering of the mood marker in relation to the aspect marker (both non-deictic) falls out of the adjacency condition between aspect and its complement (the VP). I will assume that where cases like (43) exist, tense is still somehow accessible to discourse binding. However, as far as the position of mood in MC (and other Creoles) is concerned, it seems to be determined jointly by DOP (being non-deictic, it cannot precede tense) and the relation that obtains between aspect and its complement. 6. Conclusion The fixed order displayed by TMA markers is a phenomenon that characterizes MC and other Creoles (and possibly other languages). I have argued that TMA markers are syntactically heads. They cluster on the left of verbs because pidgins and Creoles emerge as head-initial languages. The position of T in relation to M and A, I have argued, is a consequence of a Deictic Ordering Principle according to which where a deictic and a non-deictic element coexist, the deictic element takes linear precedence or asymmetrically ccommands the non-deictic one. The ordering of A in relation to M (both nondeictic elements) falls out of the nature of the structural relation between A and its complement (VP) as well as an adjacency condition on relation between property-assigning elements and property-receiving elements. A therefore has to be string adjacent to VP. The DOP operates on TP and DP, thus capturing some interesting parallels between them, at the levels of both syntax and morphology. This principle, I have argued, derives from a locality constraint on the binding of deictic elements by discourse operators. On this approach there is no need to postulate principles which apply solely to TMA ordering facts.


295

References Abney, Steven. 1987. The English noun phrase in its sentential aspect. Unpublished doctoral dissertation. Massachusetts Institute of Technology, Cambridge, MA Baker, Mark. 1985. “The mirror principle and morphosyntactic explanation”. Linguistic Inquiry 16. 373-415. Belletti, Adriana. 1990. Generalized verb movement, aspects of verb syntax. Turin: Rosenberg & Sellier. Bybee, Joan L. 1985. Morphology: A study of the relation between meaning and form. Amsterdam: John Benjamins. Chaudenson, Robert. 1981. Textes créoles anciens (La Réunion et île Maurice). Hamburg: Buske. Chomsky, Noam. 1989. “Some notes on the economy of derivations and representations”. MIT working papers in Linguistics 10. 43-74. ----------. 1995. The minimalist program. Cambridge, MA: MIT. Chrestien, F. 1831. Les essays d’un bobre africain. Port Louis: Déroulléde. DeGraff, Michel. 1993. “A riddle on negation in Haitian”. Probus 5. 63-93. Deprez, Viviane & Marie Thérèse Vinet. 1997. “Predicative constructions and functional categories in Haitian Creole”. Journal of Pidgin and Creole Linguistics 12. 203-235. Enç, Mürvet. 1987. “Anchoring conditions for tense”. Linguistics Inquiry 18. 633-657. Freycinet, Louis Claude Desoulses. 1827. Voyage autour du monde, entrepris par ordre du roi... Paris: Pillet ainé. Tome 2(2). Harris, Martin. 1978. The evolution of French syntax: A comparative approach. London: Longman. Laka Mugarza, Miren Itziar. 1990. Negation in syntax: On the nature of functional categories and projections. Unpublished doctoral dissertation. Massachusetts Institute of Technology, Cambridge, MA. Lasnik, Howard, Marcela Depiante, & Arthur Stepanov. 2000. Syntactic structures revisited: Contemporary lectures on classic transformational theory. Cambridge, MA: MIT Press. Lefebvre, Claire. 1996. “The tense, mood and aspect system of Haitian Creole and the transmission of grammar in creole genesis”. Journal of Pidgin and Creole Linguistics 11. 231-311. Lyons, John. 1977. Semantics (Vol 2). Cambridge: Cambridge University Press. Magloire-Holly, Hélène. 1982. “Les modaux: auxiliaires ou verbes?” Syntaxe de l’Haïtien, Lefebvre, Magloire-Holly, & Piou (eds) 1982 (Ann Arbor, MI: Karoma). 92-121. Muysken, Pieter. 1981. “Creole Tense/Mood/Aspect systems: The unmarked case?” Generative studies in creole languages, Muysken (ed) 1981 (Dordrecht: Foris). 181-191.

296


Pitot, T. 1805. “Quelques observations sur l’ouvrage intutilé Voyage à L’ille de France par un officier du roi”. Révue historique et literaire de L’ille Maurice 2. 272-274. Pollock, Jean-Yves. 1989. “Verb movement, universal grammar, and the structure of IP”. Linguistic Inquiry 20. 365-424. Rickard, Peter. 1974. A history of the French Language. London: Hutchinson. Rizzi, Luigi. 1990. Relativized minimality. Cambridge, MA: MIT Press. Stowell, Timothy. 1981. Origins of phrase structure. Unpublished doctoral dissertation. Massachusetts Institute of Technology, Cambridge, MA. Taraldsen, Knut T. 2000. “The que/qui alternation and the distribution of expletives”. Subjects, expletives and the EPP, Svenonius (ed) 2000 (Oxford: Oxford University Press) 29-42. Taylor, Douglas. 1971. “Grammatical and lexical affinities of creoles”. Pidginization and creolization of language, Hymes (ed) 1971 (Cambridge: Cambridge University Press). 293-296. Thompson, Robert W. 1961. “A note on some possible affinities between Creole dialects of the Old World and those of the New”. Proceedings of the 1959 Conference of Creole Language Studies, Le Page (ed) 1961 (London: Macmillan). 107-113. Travis, Lisa. 1984. Parameters and effects of word order variation. Unpublished doctoral dissertation. Massachusetts Institute of Technology, Cambridge, MA. Woisetschlaeger, Erich F. 1977. A semantic theory of the English auxiliary system. Unpublished doctoral dissertation. Massachusetts Institute of Technology, Cambridge, MA. Zagona, Karen T. 1988. Verb phrase syntax. Dordrecht: Kluwer. Zanuttini, Raffaela. 1996. “On the relevance of tense for sentential negation”. Parameters and functional heads, Belletti & Rizzi (eds) 1996 (Oxford: Oxford University Press). 181-207. Zwicky, Arnold. 1985. “Heads”. Journal of Linguistics 21.1-29.

INDEX A

B

Acadian French, 262, 263 Acadians, 12 acquisition order, 265 acquisition process, 208, 262, 264, 265 acrolect, 36, 111, 112, 115, 216, 217, 218, 219 acrolectal speakers, 217, 219 adjacency condition, 277, 292, 294 adjacency requirement, 292 adjectival suffixes, 254 adjectives, 19, 84, 86, 92, 154, 157, 187, 207, 251, 252, 254, 262, 280, 291, 292 demonstrative, 209 gender marking on, 58–63, 257, 258, 267 possessive, 19 predicate, 262 adjunction, 281, 288 adjuncts, 281, 283, 286 affixal strategies, 261 affixation, 78 affixes, 4, 77, 78, 80, 87, 90, 91, 92, 93, 96, 97, 98, 264, 279, 293 African languages, 3, 30, 39, 48, 79, 114, 122, 164, 206, 209, 214, 253, 257, 276 African slaves, 67, 116, 120, 126, 138, 139, 141, 143, 150, 204, 207, 209, 211, 213, 218, 219, 236, 244, 247, 264 African traders, 213 African-American Vernacular English, 112 Afro-Cubans, 117 Afrogenesis Hypothesis, 119–25 analogy, 49, 52–55, 245 anglicisms, 180, 181, 186 animacy, 237, 238, 260, 261, 267 animacy hierarchy, 258, 261 animates, 238, 241, 253, 256, 257, 258, 259, 261, 267 apical [r], 13 archaisms, 18, 25, 255 authenticity of early texts, 3, 12, 17, 26, 29, 36–40 auxiliaries, 22, 77, 78, 97, 273, 279, 281, 282, 284, 285, 287, 289 finite, 285, 286, 287 non-finite, 285

Bajan, 111, 113 Bajans, 115, 217 Bambara, 150 basilect, 24, 27, 36, 111, 112, 217, 218, 203, 219 basilectal creole, 18, 19, 20, 25, 26, 27, 111, 172, 175, 187, 217, 218, 219, 243, 252, 255, 258, 260 basilectal speakers, 107 basilectalization, 138, 140, 142, 203, 211, 212, 214, 215 Batticaloa Creole Portuguese, 79 Berbice Creole Dutch, 122, 141 bilingualism, 217 bimorphemic interrogatives, 235, 244, 245, 247 bioprogram, 103, 105, 106 Bioprogram Hypothesis, 105 bound morphemes, 92, 279, 280 C Cajun French, 13, 19, 116, 243, 246, 262, 263 Cane River, 12, 13, 15, 18, 20, 23, 24, 25, 26 Cape Verde, 115, 125, 126 Cape Verdean Portuguese, 117 Cape Verdean Portuguese Creole, 117, 118, 124, 125 CAPES, 167, 170, 174, 175, 176 c-command, 273, 290, 291, 292, 293, 294 Chaos Theory, 94 clitics, 78, 79, 83, 86, 91, 93, 96, 97, 98, 280, 281, 289 reflexive, 84 TMA, 78, 288 Wackernagel-type, 78 code switching, 25, 183, 215 Colonial French, 247 complementarity, 167, 171, 172, 219 complementary hypothesis, 206 compounding, 58–63 congruence, 206, 207, 208, 209, 211, 252 continuum, 3, 78, 79, 80, 94, 103, 109, 111, 114, 115, 116, 123, 128, 145, 193, 203, 217, 218, 219, 244 continuum, creole, 109–13

298


cousin languages, 165 creole genesis theory, 103, 109, 121, 123 creole neologisms, 255 creoles linguistic diversity of, 165, 175, 176 creolisms, 37, 180, 181, 186, 193, 194, 195, 196, 199 creolization, 82, 87, 121, 127, 137, 138, 164, 174, 235, 236, 252, 261, 265 abrupt, 82, 150 D deaffixation, 92 debasilectalization, 214, 217, 218 decreolization, 166, 169, 214, 215, 216, 217, 218, 219, 260, 261, 267 decreolization hypothesis, 215 definite article, 31, 171, 209, 260, 261, 280 definite determiner, 30, 34 preposed, 209 definite determiners postposed, 31, 260 deictic, 294 Deictic Ordering Principle, 290, 291, 294 deictics, 31, 32, 81, 82, 209, 267, 273, 274, 290, 291, 292, 293, 294 deixis, 82, 291 demographic attrition, 211, 216 demonstrative adjectives, 209 demonstrative determiners, 209 demonstratives, 32, 77 derivation, 58, 168, 174, 293 derivational morphemes, 92 derivational procedures, 175 derivational rules, 173 derivational suffixes, 92 derivational system, 168 derivations, 256 determiner system, 30, 204, 208, 214 determiners, 31, 32, 34, 254, 280, 281, 291, 292 definite, 30, 34 postposed, 31 preposed, 209 gender agreement on, 260, 267 ordering constraint on, 291 possessive, 17, 24 gender marking on, 20, 18–21 postnominal, 252 postposed, 209

dialect boundaries, 6, 225, 226, 227, 228 dialect differentiation, 236 dialect mixture, 33, 34 dialectal variation, 36, 225, 226 dialects British, 112 English, 112, 113 French, 65, 205, 209, 230, 231, 242 Gulf of Guinea, 125 Haitian, 225, 226 diglossia, 169, 203 discourse anchoring, 292 discourse binding, 293, 294 discourse strategies, 174 Dutch, 164, 291 Dutch-based creoles, 122, 164 E echo questions, 241 ecology, 25, 206, 211, 219 emphatic focus particle, 171 emphatic marker, 84, 86, 87, 88, 89, 92, 93, 94, 95, 97 enclitics, 92 English-based creoles, 110, 164, 244 epilinguistic discourse, 179, 189, 192 epilinguistic perspective, 180 Estonian, 93, 97 Modern, 92 Old, 92 expressive component, 80, 82, 83, 84, 85, 87, 93, 94 expressive grammatical marker, 84, 85, 89 extension, 52 F fantasy linguistics, 174 finiteness, 285 first language acquisition, 265 fixity (lack of), 41 focus particle, 242 Fon, 276 form and function clines, 91 free form-clitic-affix cline, 97 free standing morphemes, 279 free standing tense markers, 285 French triangle (in Louisiana), 11, 16 French-based creoles, 72, 163, 164, 165, 166, 167, 170, 176, 235, 242, 244, 251, 252

INDEX

Frenchification, 169, 172, 173 Frenchified orthography, 237 functional categories, 273 functional domains, 293 functional elements, 80, 86, 87, 89, 91, 93, 94, 97, 147 functional phrase, 281 functional projection, 287, 290, 291 functional–semantic components, 80 fusion, 172, 173 future marker, 24, 83, 94, 97 G gender acquisition of, 265 grammatical, 261, 262, 264, 265, 267 loss of, 265 morphologically assigned, 261 natural, 251, 266, 267, 268 gender agreement on adjectives, 267 on determiners, 260, 267 gender confusion, 263 gender distinctions, marking, 17 gender marking difference between creoles and base language, 251 grammatical, 261 natural, 252, 253–60 on adjectives, 58–63 on indefinite articles, 263 on possessives, 18–21, 25, 26 on pronouns, 258–60, 263, 268 gender systems, 251, 252, 253, 259, 261, 262, 264, 266, 267 gender weakening, 262 grammatical indeterminacy, 33 grammaticalization, 30, 39, 77, 78, 79, 80, 94, 97, 98 apparent, 79 idiosyncratic, 110 instantaneous, 79, 82 ordinary, 79 prototypical direction of, 93 grammaticalization cline, 78, 86 Grimm’s Law, 112 Guadeloupean Creole, 7, 30, 32, 33, 42, 43, 165, 170, 171, 172, 173, 245, 246, 254, 256, 257, 258, 259, 261 Guinea-Bissau Creole, 117, 118, 125 Gulf of Guinea creoles, 117, 125

299 Gulf of Guinea dialect complex, 125 Gullah, 111, 113, 124, 218 Guyanais Creole, 170, 171, 172 Guyanese, 109, 111, 112 Guyanese creole continuum, 112 H Haitian Creole, 21, 79, 124, 165, 203, 212, 216, 225, 226, 235 as a continuum, 203 bimorphemic interrogatives in, 245, 246 decreolization in, 215 determiners in, 34 diatopic variation in, 225–33 gender marking in, 256 in complementary distribution with French, 215 negative concord in, 286, 287 plural marking in, 229 possessive construction in, 210, 211 possible influence of on Louisiana Creole, 236, 244 reflexes of French [h] in, 231 restructuring in evolution of, 245 rounded front vowels in, 211 semivowel [] in, 231 substrate languages for, 276 TMA markers in, 276, 279, 287 Haitian French, 47, 179, 181, 182, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199 Haitian lexical particularities, 181, 189, 191, 192, 196 Haitianisms, 256 half-sister creoles, 165 half-sister languages, 165, 167 Hawaii, 103, 106, 107, 108, 109, 118, 124, 126, 127, 128, 138, 140, 141, 142, 144, 145, 146, 147 colloquial English of, 106 Hawaiian, 145, 148 Hawaiian Creole English, 105, 106, 107, 108, 109, 127 Hawaiian English, 106 Head Movement Constraint, 283, 286, 287 Head Parameter, 274, 284, 292 heterolects, 175 historical creole texts, 17, 29, 33 H-language (in diglossia), 30, 169 homestead colonies, 213

300


homestead contact language, 142 homestead phase (of colonial development), 138, 139, 143, 211 homogeneity lexical, 171, 173 structural, 206 homolects, 175 hypercorrection, 65, 243 I idiolects, 208, 219 inanimates, 238, 241, 253, 258, 259, 261 inflectional morphology, 251, 252 interdialectal borrowing, 172 interrogative adverbs, 245 interrogative focus particles, 235 interrogative pronouns, 235, 237, 241, 235–47, 247 interrogatives, 77 creole, 235 creole interrogative systems, 235 Estonian, 92, 93 French, 245 inanimate, 247 isoglosses, 225, 226, 227, 228, 229, 230, 231, 232 isolates, 17 J Jamaican Creole, 111, 112, 165 Jamaican patois, 108, 124, 128 K Kikongo, 118, 150, 256 koine, 169, 170, 175 Korlai Creole Portuguese, 79, 82, 84, 88, 99 Krio, 128, 246 Kwa, 89, 276 L language change, 112, 138, 175, 252 language death, gender loss in, 265 language endangerment, 216 language shift, 13, 27 lects, continuum of, 109, 115, 203 Lesser Antillean creoles, 21, 43, 172, 211 lexical derivation, 168 lexical diversification, 171, 172

lexical neology, 173–74 Lexical particularity, 180 lexical variables, 225 lexical variation, 47, 48, 219 lexicalization, 97 lexicalization-grammaticalization cline, 80, 91, 93, 94, 95, 96, 99 lexicalization-grammaticalization continuum, 94, 97, 98 lexifier, 103, 104, 109, 111, 117, 124, 126, 141, 163, 165, 167, 203, 242, 285 Liberian Pidgin English, 112 likely source dialect, 172 lingua franca, 125, 213 creolized Portuguese, 126 European, 213 interdialectal, 169 interethnic, 108, 127 linguistic isolates, 17 linguistic market, 163, 166, 167, 170 L-language (in diglossia), 30 local norm, 181, 182, 188, 189, 192, 195 Louisiana Creole, 116, 124 bimorphemic interrogatives in, 244, 245, 247 demographic center of, 15 early texts in, 12, 17, 18, 19, 20, 21, 24, 25, 26 ecology of, 25 gender marking in, 19, 259 gender marking on possessive determiners in, 18–21 historical representations of, 16 in contact with other French-related varieties, 12, 13, 16, 25, 27, 116 interrogative adverbs in, 246 interrogative focus particle in, 240, 244 interrogative pronouns in, 235, 244 long and short verbs in, 25 mesolect of, 116 origin and development of, 12, 15, 16, 25, 26, 124, 236 possible influence of Haitian Creole on, 236, 247 pronominal system of, 235 pronouns in, 235 regional varieties of, 236, 244 single genesis of, 236, 247 speakers' ethnicity, 12, 13, 14 speakers' shift to English, 15 verb 'to have' in, 18 verbal system of, 21, 26

INDEX

Louisiana French, 247 Louisiana Regional French (LRF), 12, 13, 14, 15, 16, 19, 20, 25, 26 M Mahim Creole Portuguese, 79, 87 markers affixal grammatical relation, 92 anterior, 112, 118 aspect, 273 aspectual, 85, 86 case, 90, 147, 265 deictic, 209 discourse, 78 dummy, 81 emphatic, 84, 86, 87, 88, 89, 92, 93, 94, 95, 97 expressive grammatical, 84, 85, 89, 90 future, 83, 94, 97 gender, 265 grammatical, 77, 78, 80, 82, 86, 87, 89, 90, 92 habitual, 93, 95 modal, 273 negative, 93, 289 number, 265, 267 object, 147, 148 object pronoun, 82 of definiteness, 32, 209 of emphasis, 89 of in-group identity, 107 of negation, 90 of turn taking, 82 past, 79, 87, 88, 89, 93, 94, 110, 112, 275 plural, 229, 267 possessed, 93, 95 pragmatic, 85 predicate, 114 progressive, 225, 229, 274, 276 pronominal object, 79 redundancy of, 98 tense, 86, 90, 273 textual, 82 TMA, 21, 22, 78, 79, 106, 107, 206, 273, 274, 276, 277, 294 order of, 284–94, 284–94 status of, 279–84 topic, 148 transitive, 114 zero, 275, 276

301 Martinican Creole (MC, MarC) definite article in, 171 definite determiner in, 30, 32, 34, 35 demonstrative determiners in, 171 distinctiveness from Guadeloupean Creole, 42 emphatic focus particle in, 171 gender marking in, 261 identifying features of, 33 mutual intellibigility with other creoles, 165 possessives in, 30 reading and writing of, 43 stabilization of, 30 Mauritian Creole early texts on, 289 finite auxiliaries in, 285 negative concord in, 286, 287 negative markers in, 285 negative particle in, 286 non-finiteness of verbs in, 285 numerical slang in, 153–161 origin and development of, 104 reanalyzed articles in, 264 relationship to Reunion Creole, 165 TMA markers in, 273–94 verb raising in, 282 Melanesian Pidgin English, 127 Melanesian pidgins, 206 mesolect, 20, 111, 112, 116, 128, 140 mesolectal continuum, 217 mesolectal features, 215, 243 mesolectal speakers, 111 mesolectal speech, varieties, 111, 112, 115, 187, 214, 217, 243, 259, 260 metaphor, meaning change by, 56–57 metonymy, meaning change by, 55–56 Mirror Principle, 293 modality, 294 Mon Louis Island, Alabama, 12, 15, 18, 20, 23, 24, 25 N nasalization of postposed article, 228 Natchitoches, 12 Natchitoches Parish, 12, 13, 14 nativization, 167 natural gender, 251, 252, 253–60 naturalistic language appropriation, 206 naturalness, 167–70, 174 negation

302


as a head element, 283 discontinuous, 89 double, 286, 287 in relation to auxiliaries, 287 in relation to TMA markers, 274, 277, 284, 287, 290, 285–90 in relation to verbs, 282, 285, 286, 287 with verbal force, 289 negation particle, 89 negative concord, 286 negative markers, 286, 289 in lexifier languages, 285 in Mauritian Creole, 285 negative particle, 286 negative quantifiers, 286, 287 Negerhollands, 122 neological practice, 175 neologisms, 175, 180, 181, 255, 256 neology, 175 new hands (in colonial contexts), 140, 142, 143, 144, 145, 146, 148, 149 norm, local, 181, 182, 188, 189, 192, 195 normalization, 168, 210 norming, 168 Northern Saame Enontekiö, 91 Noun Phrase gender marking in, 19 nouns gender marking on, 261 O old hands (in colonial contexts), 137, 142, 143, 144, 145, 146, 148, 149 onomatopoeic origin, words of, 57 orthographic norm (absence of), 41 P Palenquero, 116, 117, 118 Papiamentu, 117, 118, 126 patchwork effect, 167 pidgin English in Singapore, 108 pidgin Hawaiian, 107, 147 pidginization, 82, 111, 114, 127, 138, 140, 145, 146, 149, 150, 265, 266 pidginization index, 106 plantation economy, 139 plantation phase (of colonial development), 138, 211, 218

Plaquemines Parish, 12, 13, 14, 15, 18, 20, 23, 24, 25 plausible creole, 173–74, 175 plural markers, 43, 252, 267 plurality, 266, 267 pluralizer, 267 Portuguese-based contact languages, 117 Portuguese-based creoles, 118, 164, 244 possessive constructions, 209, 225, 229 possessive determiners, 17, 20, 24, 31, 32, 34, 260 gender marking on, 18, 20, 18–21, 26 possessive forms, 42 possessive locutions, 91 possessive pronouns, 30, 39, 210 possessive system, 30 postposed anterior marker, 118 postposed articles, 225 nasalization of, 228 postposed definite determiner, 31, 32 -la/-là, 31, 32, 209, 228,260 yo, 229 postposed possessor constructions, 210 prescriptivism, 122, 128, 191, 195 preverbal future marker, 24 progressive marker, 225, 229, 274, 276 pronominal systems, 30, 79, 235, 259 pronouns anaphoric and cataphoric, 81 gender marking on, 258–60, 263, 268 interrogative in Louisiana Creole, 235–47 possessive, 30, 39, 229 reflexive, 85 relative, 246, 247 second person plural, 110 third person, 117, 225, 229, 252, 258, 262, 263 tonic, 210 propositional component, 80, 81, 82, 84, 86, 92, 93, 94 propositional operators, 290, 291 Q questioned semantic unit, 244, 245, 246 R recreolization, 166, 169 reflexive pronouns, 85, 86 relative clause structures, 236

INDEX

relative pronouns, 246, 247 relativizers, 81, 114, 243 relexification, 103, 118, 119, 206, 210 relexification hypothesis, 206 rememberers (in situations of language shift), 15, 18 Reunion French-language education in, 166 linguistic atlas of, 260 Reunion Creole as half-sister language of Martinican Creole, 165 as sister language of Mauritian Creole, 165 mutual intelligibility with other creoles, 165 S Saint Domingue at the end of the eighteenth century, 47–72 slaves brought to Louisiana from, 236, 247 slaves' influence on Louisiana Creole, 244–47 São Tomense Creole Portuguese, 117, 118, 125 Saramaccan, 118 second language acquisition, 87, 103, 120, 123, 127, 142, 145, 265 second language learning, 104, 119 semantic extension, meaning change by, 49–50 semantic transparency, 265, 266 semi-speakers, 15, 18, 20 Seychelles Creole, 109, 165, 252, 254, 255, 256 Seychellois, 109 Seychellois Creole, 252 Sierra Leone Krio, 128 Singapore English, 108 as a non-creole language, 107 as casual alternative to English, 107 creation of, 107, 108 development of, 107 genesis of, 108 zero copula in, 112 sister creoles, 87, 165, 166, 171 sister languages, 165, 167 Skepi Dutch, 122 specifiers, 258, 283, 286, 288

303 spontaneous streamlining, 168, 169, 174 Sranan, 79, 108, 113, 114, 118, 124, 126, 145 St. Tammany Parish Louisiana Creole in, 12, 13, 14, 15, 18, 20, 23, 24, 25, 26, 27 streamlined creole, 174 structural attrition, 216 substitution, 288 substrate borrowing, 117 substrate grammars, 206 substrate influence, 114, 203, 204, 205, 206, 207, 208, 209, 210, 211, 214, 218, 235, 257 substrate languages, 103, 113, 114, 141, 170, 205, 206, 207, 208, 209, 210, 211, 212, 214, 219, 242, 253, 276 substrate populations, 210 substrate systems, 206 suffixal past marking, 88 suffixation marking of gender by, 255–56 suffixes adjectival, 254 caritive, 91 case, 92 derivational, 92 gender-marking, 255, 256 lative, 91 possessed, 91 tense-aspect, 206 superstrate languages, 113, 114, 123, 126, 127, 128, 137, 141 Suriname, 110, 116, 126, 139, 140, 143, 145 Swiss French, 89, 90, 93, 95, 98 synonymy, 172 interdialectal, 172 perfect, 172 T tag questions, 241 Talasri Creole Portuguese, 87 Tensed C-command Condition, 290, 291 textual authenticity, 17, 36 textual component, 80, 81, 82, 83, 84, 85, 86, 93, 94 textual markers, 82 TMA ordering, 273, 274, 276, 284–94 TMA Ordering Principle, 273, 274, 290 Tok Pisin, 79, 108, 113, 114, 126

304


U

X

unity (of the Romance languages), 164 unity, universality (of creole languages), 163, 176 Universal Grammar, 105, 264, 273, 274, 290 universal language, 169 universal language faculty, 164 universal system, 164 universalist approach, 141, 147 universalist model, 140 universals, 97, 103, 113, 114, 142 Upper Guinea creoles, 125 Uto-Aztecan, 90, 91, 92, 93, 95 uvular [ë], 13

xenolectal speakers, 209 xenolects, 175, 208

V variation contextual, 171 continuum of, 215 dialectal, 36, 225, 226 diatopic, 225–33 geographic, 187, 225, 229, 232 in historical texts, 41–43 in plantation settings, 211 in possessives, 209, 210 lexical, 47, 48 linguistic, 260 mesolectal, 20 phonetic, in Haitian Creole, 229 phonological, 225 regional, 219 syntactic, 225 verb raising, 282, 287 verb systems, 24, 30 VP ellipsis, 278, 279, 280, 281 W West African Creole Portuguese, 118 West African Pidgin English, 212 West African pidgins, 118, 122, 127, 212, 213 West African trade settlements, 122 West Indies, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 222 wh-in situ questions, 237, 241

Z zero copula, 107, 111, 112 zero markers, 275, 276 zero past marking, 107

In the series Creole Language Library the following titles have been published thus far or are scheduled for publication: 29 DEUMERT, Ana and Stephanie DURRLEMAN (eds.): Structure and Variation in Language Contact. Expected November 2006 28 CLEMENTS, J. Clancy, Thomas A. KLINGLER, Deborah PISTON-HATLEN and Kevin J. ROTTET (eds.): History, Society and Variation. In honor of Albert Valdman. 2006. vi, 304 pp. 27 ESCURE, Geneviève and Armin SCHWEGLER (eds.): Creoles, Contact, and Language Change. Linguistic and social implications. 2004. x, 355 pp. 26 MOUS, Maarten: The Making of a Mixed Language. The case of Ma’a/Mbugu. 2003. xx, 322 pp. 25 MIGGE, Bettina: Creole Formation as Language Contact. The case of the Suriname Creoles. 2003. xii, 151 pp. 24 MÜHLEISEN, Susanne: Creole Discourse. Exploring prestige formation and change across Caribbean Englishlexicon Creoles. 2002. xiv, 332 pp. 23 SMITH, Norval and Tonjes VEENSTRA (eds.): Creolization and Contact. 2001. vi, 323 pp. 22 NEUMANN-HOLZSCHUH, Ingrid and Edgar W. SCHNEIDER (eds.): Degrees of Restructuring in Creole Languages. 2001. iv, 492 pp. 21 McWHORTER, John (ed.): Language Change and Language Contact in Pidgins and Creoles. 2000. viii, 503 pp. 20 RICKFORD, John R. and Suzanne ROMAINE (eds.): Creole Genesis, Attitudes and Discourse. Studies celebrating Charlene J. Sato. 1999. viii, 418 pp. 19 SPEARS, Arthur K. and Donald WINFORD (eds.): The Structure and Status of Pidgins and Creoles. Including selected papers from meetings of the Society for Pidgin and Creole linguistics. 1997. viii, 461 pp. 18 ESCURE, Geneviève: Creole and Dialect Continua. Standard acquisition processes in Belize and China (PRC). 1997. x, 307 pp. 17 THOMASON, Sarah G. (ed.): Contact Languages. A wider perspective. 1997. xiii, 506 pp. 16 CLEMENTS, J. Clancy: The Genesis of a Language. The formation and development of Korlai Portuguese. 1996. xviii, 282 pp. 15 ARENDS, Jacques, Pieter MUYSKEN and Norval SMITH (eds.): Pidgins and Creoles. An introduction. 1994. xv, 412 pp. 14 KIHM, Alain: Kriyol Syntax. The Portuguese-based Creole language of Guinea-Bissau. 1994. xii, 310 pp. 13 ARENDS, Jacques (ed.): The Early Stages of Creolization. 1996. xvi, 297 pp. 12 BYRNE, Francis and Donald WINFORD (eds.): Focus and Grammatical Relations in Creole Languages. Papers from the University of Chicago Conference on Focus and Grammatical Relations in Creole Languages. 1993. xvi, 329 pp. 11 BYRNE, Francis and John HOLM (eds.): Atlantic Meets Paciﬁc. A global view of pidginization and creolization. 1992. ix, 465 pp. 10 WINFORD, Donald: Predication in Caribbean English Creoles. 1993. viii, 419 pp. 9 BYRNE, Francis and Thom HUEBNER (eds.): Development and Structures of Creole Languages. Essays in honor of Derek Bickerton. 1991. x, 222 pp. 8 BAILEY, Guy, Natalie MAYNOR and Patricia CUKOR-AVILA (eds.): The Emergence of Black English. Text and commentary. 1991. x, 352 pp. 7 FABIAN, Johannes (ed.): History from Below. The “Vocabulary of Elisabethville” by André Yav: Text, Translations and Interpretive Essay. With the assistance of Kalundi Mango. With linguistic notes by W. Schicho. 1990. vii, 236 pp. 6 SINGLER, John Victor (ed.): Pidgin and Creole Tense/Mood/Aspect Systems. 1990. xvi, 240 pp. 5 JACKSON, Kenneth David: Sing Without Shame. Oral traditions in Indo-Portuguese Creole verse. 1990. xxiv, 257 pp. 4 LIPSKI, John M.: The Speech of the Negros Congos in Panama. 1989. vii, 159 pp. 3 BYRNE, Francis: Grammatical Relations in a Radical Creole. Verb Complementation in Saramaccan. With a foreword by Derek Bickerton. 1987. xiv, 293 pp. 2 SEBBA, Mark: The Syntax of Serial Verbs. An investigation into serialisation in Sranan and other languages. 1987. xv, 218 pp. 1 MUYSKEN, Pieter and Norval SMITH (eds.): Substrata versus Universals in Creole Genesis. Papers from the Amsterdam Creole Workshop, April 1985. 1986. vii, 311 pp.

History, Society And Variation: In Honor of Albert Valdman (Creole Language Library)

Structure And Variation in Language Contact (Creole Language Library)

Mbugu (Creole Language Library)

Creolization and Contact (Creole Language Library)

Language Universals and Variation

Language Universals and Variation

Creole Formation as Language Contact: The Case of the Suriname Creoles (Creole Language Library)

Roots of Creole Structures: Weighing the contribution of substrates and superstrates (Creole Language Library)

Pedagogical Norms for Second and Foreign Language Learning and Teaching Studies: Studies in Honor of Albert Valdman (Language Learning & Language Teaching)

Creole and Dialect Continua: Standard Acquisition Processes in Belize and China (Prc) (Creole Language Library)

Kriyol Syntax: The Portuguese-based Creole Language of Guinea-bissau (Creole Language Library)

Kriyol syntax: the Portuguese-based Creole language of Guinea-Bissau (Creole Language Library)

Language Variation in South Asia

Language Change and Variation in Gibraltar (Impact: Studies in Language and Society, 23)

Creole Discourse: Exploring Prestige Formation and Change Across Caribbean - English-lexicon Creoles (Creole Language Library)

Noun Phrases in Creole Languages: A multi-faceted approach (Creole Language Library)

The Handbook of Language Variation and Change

Deconstructing Creole (Typological Studies in Language)

Contact Languages: A Wider Perspective (Creole Language Library)

Contact languages: a wider perspective (Creole Language Library)

Variation in the Caribbean: From creole continua to individual agency

The Making of a Mixed Language: The Case of Ma'a Mbugu (Creole Language Library)

Creoles, Contact, and Language Change: Linguistics and Social Implications (Creole Language Library)

Language, Technology, and Society

The Indonesian Language: Its History and Role in Modern Society

The Indonesian Language: Its History and Role in Modern Society

The Indonesian Language: Its History and Role in Modern Society

Variation in second language acquisition, Volume 50

Language in exile: three hundred years of Jamaican Creole

Language, Society and Power

The Dynamics of Linguistic Variation: Corpus Evidence on English Past and Present (Studies in Language Variation)

History, Society And Variation: In Honor of Albert Valdman (Creole Language Library)

Structure And Variation in Language Contact (Creole Language Library)

Mbugu (Creole Language Library)

Creolization and Contact (Creole Language Library)

Language Universals and Variation

Language Universals and Variation

Creole Formation as Language Contact: The Case of the Suriname Creoles (Creole Language Library)

Roots of Creole Structures: Weighing the contribution of substrates and superstrates (Creole Language Library)

Pedagogical Norms for Second and Foreign Language Learning and Teaching Studies: Studies in Honor of Albert Valdman (Language Learning & Language Teaching)

Creole and Dialect Continua: Standard Acquisition Processes in Belize and China (Prc) (Creole Language Library)

Kriyol Syntax: The Portuguese-based Creole Language of Guinea-bissau (Creole Language Library)

Kriyol syntax: the Portuguese-based Creole language of Guinea-Bissau (Creole Language Library)

Language Variation in South Asia

Language Change and Variation in Gibraltar (Impact: Studies in Language and Society, 23)

Creole Discourse: Exploring Prestige Formation and Change Across Caribbean - English-lexicon Creoles (Creole Language Library)

Noun Phrases in Creole Languages: A multi-faceted approach (Creole Language Library)

The Handbook of Language Variation and Change

Deconstructing Creole (Typological Studies in Language)

Contact Languages: A Wider Perspective (Creole Language Library)

Contact languages: a wider perspective (Creole Language Library)

Variation in the Caribbean: From creole continua to individual agency

The Making of a Mixed Language: The Case of Ma'a Mbugu (Creole Language Library)

Creoles, Contact, and Language Change: Linguistics and Social Implications (Creole Language Library)

Language, Technology, and Society

The Indonesian Language: Its History and Role in Modern Society

The Indonesian Language: Its History and Role in Modern Society

The Indonesian Language: Its History and Role in Modern Society

Variation in second language acquisition, Volume 50

Language in exile: three hundred years of Jamaican Creole

Language, Society and Power

The Dynamics of Linguistic Variation: Corpus Evidence on English Past and Present (Studies in Language Variation)

Recommend Documents